Modern Fracturing CD

Modern Fracturing
Enhancing Natural Gas Production
Michael J. Economides
University of Houston
Tony Martin
BJ Services
ET Publishing
Houston,TX
BJ Services Company 2007
BJ Services Company
P.O. Box 4442 [77210-4442]
4601 Westway Park Blvd.
Houston, TX 77041
Graphic design and production: Jay Clark

Production manager: Alexander M. Economides
Copy Editor: Stephanie Weiss
Cover Art: Armando Izquierdo
Published by: Energy Tribune Publishing Inc.

820 Gessner Rd.-Ste. 920
Houston, TX 77040

(713) 647-0903
(713) 647-0940 (fax)
for orders and customer service enquires contact: info@energytribune.com
All rights reserved. No part of the publication may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means, except under the expressed
permission of BJ Services Company, Designs and Patents Act 1988.
ISBN 978 1 60461 688 0
Printed and bound by Gulf Publishing Co.

Contents
Preface XV
Foreword XVII
Contributing Authors XVIII
Acknowledgements XIX
Chapter 1
Introduction to this Book
1-1 Introduction 3
1-2 Natural Gas in the World Economy 3
1-3 Russia: A Critical Evaluation of its Natural Gas Resources 5
1-3.1 The Resource Base 7
1-3.2 Russian Natural Gas Production 8
1-4 Alaska, its Natural Gas Resources and their Impact on US Imports 8
1-4.1 Alaskan Reserves and Production 9
1-4.2 The Uncertain Destiny of the North Slope of Alaska Natural Gas 10
1-4.3 Alaska in the Context of the United States and Canadian Natural Gas 11
1-5 Qatar Natural Gas 12
1-5.1 North Field Characteristics and Development 13
1-6 Fracturing for the Efficient use of Existing
Resources and for Increasing Recovery Factor 13
Chapter 2
Natural Gas Production
2-1 Introduction 19
2-2 Idiosyncrasies of Dry Gas, Wet Gas and Gas Condensates 19
2-3 Inflow from Natural Gas Reservoirs 20
2-3.1 Fundamentals of Non-Darcy Flow in Porous Media 20
2-3.2 Transient Flow 20
2-3.3 Steady State and Pseudosteady State Flow 21
2-3.4 Horizontal Well Flow 22
2-4 Effects of Turbulence 23

2-4.1 The Effects of Turbulence on Radial Flow 23
2-4.2 Perforated and Cased Well in a High-Rate Gas Reservoir 24
2-5 Production from Hydraulically Fractured Gas Wells 25
2-5.1 Unique Needs of Fracture Geometry and Conductivity 26
2-5.2 Turbulence Remediation in High- and Low-Permeability Wells 26
2-5.3 Multi-fractured Horizontal Gas Wells 28
2-6 Well Deliverability, IPR and Well Flow Performance 33
2-7 Forecast of Well Performance 34
2-7.1 Gas Material Balance and Forecast of Gas Well Performance 34
2-8 Correlations for Natural Gas Properties 35
2-8.1 Pseudocritical Pressure, ppc and Pseudocritical Temperature, Tpc 35
2-8.2 Gas Viscosity 35
2-8.3 Gas Deviation Factor, Z 35
Chapter 3
Gas Well Testing and Evaluation
3-1 Introduction 41
3-2 Background Theory 42
3-3 Radial Flow Solutions 44
3-4 Superposition 45
3-5 Model Development 46
3-6 Hydraulically Fractured Wells 47
3-7 Specialized Plots 48
3-8 Type Curves and the Log-Log Derivative Plot 49
3-9 Flow Regime Identification 51
3-10 Derivatives A Few Cautionary Remarks 54
3-11 PTA Interpretation Methods 56
3-12 Difference Between High and Low Permeability Analysis Techniques 57
3-12.1 High-Permeability Wells 57
3-12.2 Low-Permeability Wells Pre-Treatment Evaluation 59
3-12.3 Example 3-1, PID Test 60
3-12.4 Low-Permeability Wells Post-Treatment Evaluation 61
3-12.5 Example 3-2, Low-Permeability Well, Infinite-Conductivity Fracture 62
3-12.6 Example 3-3, Low-Permeability Well, Finite-Conductivity Fracture 65
3-13 Non-Darcy Flow 66
3-13.1 Example 3-4, Non-Darcy, High-Permeability Well, Finite-Conductivity Fracture 68
3-13.2 Example 3-5, Non-Darcy, Low-Permeability Well, Finite-Conductivity Fracture 69
3-14 Production Analysis 70
II
3-15 Heterogeneity 76
3-15.1 Dual Porosity 76
3-15.2 Anisotropy 76
3-16 Multiphase Flow 77
3-16.1 Gas Condensates 78
3-16.2 Fracture Fluid Cleanup 79
3-16.3 Example 3-6, Fracture Fluid Cleanup Case 79
3-17 Closure Analysis 81
3-18 Deconvolution 86
Chapter 4
Hydraulic Fracture Design for Production Enhancement
4-1 Introduction to Hydraulic Fracturing 93
4-1.1 Brief History of Fracturing and Qualitative Description of Process 93
4-1.2 High Permeability vs. Low Permeability 94
4-1.3 Near-Well Flow Enhancement vs. Reservoir Stimulation 94
4-1.4 Acceleration vs. Increase of Reserves 95
4-2 Description of the Process 95
4-2.1 One of the Most Energy- and Material-Intensive Industrial Activities 95
4-2.1.1 Understanding the Significance of Pressure 96
4-2.1.2 Different Types of Pressure 96
4-2.1.3 Net Pressure 97
4-2.1.4 Effects of Tortuosity and Perforation Friction 98
4-2.1.5 Fluid Leakoff and Slurry Efficiency 101
4-2.1.6 Dimensionless Fracture Conductivity 102
4-2.1.7 Nolte-Smith Analysis Predicting Fracture Geometry from Pressure Trends 103
4-2.1.8 Step Rate Tests 104
4-2.1.9 Minifracs 106
4-2.2 The Role of Advanced Technology in Design, Execution and Evaluation 109
4-2.2.1 Recent Advances and Breakthroughs 109
4-2.2.2 Pressure Matching 112
4-2.2.3 Getting Closer to Understanding Fracture Geometry 115
4-2.2.4 Real-Time Analysis 115
4-2.3 From Fracturing a Single Vertical Well to Complex Well-Fracture Architecture 116
4-3 Rock Mechanical Characteristics 116
4-3.1 Basic Definitions 116
4-3.1.1 Stress and Strain 116
4-3.1.2 The Poissons Ratio 116
III
4-3.1.3 Youngs Modulus 117
4-3.1.4 Other Rock Mechanical Characteristics 118
4-3.1.5 Hookes Law 119
4-3.1.6 Failure Criteria and Yielding 119
4-3.2 In-Situ Stress and Fracture Orientation 121
4-3.2.1 Overburden Stress 121
4-3.2.2 Horizontal Stresses 121
4-3.2.3 The Effect of Pore Pressure 122
4-3.2.4 Fracture Orientation 122
4-3.2.5 Stress Around a Wellbore and Breakdown Pressure 123
4-3.3 Fracture Shape 125
4-3.3.1 Two-Dimensional (2-D) Fracture Geometry 125
4-3.3.2 Elliptical Fracture Geometry 125
4-3.3.3 Limitations to Fracture Height Growth 126
4-3.3.4 Complex Fracture Geometry 127
4-3.4 Fracture Propagation, Toughness and Tip Effects 127
4-3.4.1 Linear Elastic Fracture Mechanics 127
4-3.4.2 Significance of Fracture Toughness 129
4-3.4.3 Complexity at the Fracture Tip 130
4-3.5 Measuring Rock Mechanical Characteristics 132
4-3.5.1 Introduction 132
4-3.5.2 Methods of Measurement 132
4-3.5.3 Core Selection/Sample Preparation Considerations 134
4-3.5.4 Deducing Elastic Properties without Core 135
4-4 Fluid Rheological Characteristics 137
4-4.1 Viscosity 137
4-4.1.1 Shear Rate, Shear Stress and Viscosity 137
4-4.1.2 Measurement of Viscosity 137
4-4.2 Fluid Behavior 138
4-4.2.1 Newtonian Fluids 138
4-4.2.2 Non-Newtonian Fluids 138
4-4.2.3 Apparent Viscosity 139
4-4.3 Flow Regimes 140
4-4.3.1 Plug, Laminar and Turbulent Flow 140
4-4.3.2 Reynolds Number 140
4-4.4 Fluid Friction 141
4-4.4.1 The Influence of Flow Regime 141
4-4.4.2 Predicting Pressure Loss due to Friction 141
4-5 Optimum Treatment Design 141
4-5.1 Dimensionless Productivity Index and Dimensionless Fracture Conductivity 143
IV
4-5.2 Optimum Dimensionless Conductivity 144
4-5.3 Optimum Length and Width 144
4-5.4 Treatment Sizing and Proppant Placement Efficiency 145
4-5.5 Taking Into Account Operational Constraints 145
4-5.6 Using Fracture Propagation Models 146
4-5.6.1 Height containment 146
4-5.6.2 2-D models 147
4-5.6.3 3-D models 149
4-6 Predicting Production Increase 150
4-6.1 Pseudo-radial Concepts: Equivalent Wellbore Radius, Fracture Skin 150
4-6.2 Finite Reservoir Concepts, Folds of Increase 150
4-6.3 Combining Productivity Index and Material Balance 151
4-6.3.1 Pseudo-steady state 151
4-6.3.2 Combined transient and stabilized flow 151
4-6.4 Reservoir Simulation and Nodal Analysis 152
4-7 Fracturing Under Specific Circumstances 153
4-7.1 Tight Gas 153
4-7.1.1 The Importance of Inflow Area 154
4-7.1.2 Effective vs Actual Propped Length 154
4-7.2 High-Rate Gas Wells 155
4-7.2.1 Non-Darcy Flow 155
4-7.2.2 Wellbore Connectivity 155
4-7.3 High-Permeability Wells 155
4-7.3.1 The Importance of Fracture Conductivity 156
4-7.3.2 The Tip Screenout 156
4-7.4 Unconsolidated Formations 156
4-7.4.1 Re-Stressing the Formation 156
4-7.4.2 The Frac-Pack Treatment 157
4-7.5 Skin-Bypass Treatments 157
4-7.6 Condensate Dropout 158
4-7.6.1 Description of Phenomena 158
4-7.6.2 Mitigating the Effect of Dropout 158
4-7.7 Shale Gas and Coal Bed Methane 158
4-7.7.1 Gas Shales 158
4-7.7.2 Coal Bed Methane 158
4-7.8 Acid Fracturing 159
4-7.8.1 Description of Process 159
4-7.8.2 Estimating Fracture Conductivity 159
4-7.8.3 Use of Diversion Techniques 160

Chapter 5
Well Completions
5-1 Wellbore Construction 169
5-1.1 Effects of Uncertainty in Reservoir Description 169
5-1.2 Fitting Well Design to the Reservoir Potential 169
5-1.3 Well Design 170
5-1.4 Other Well Equipment 171
5-1.5 Well Integrity 171
5-2 Gas Well Cementing 172
5-2.1 General Objectives for Gas Well Cementing Operations 172
5-2.2 Gas Well Zonal Isolation 173
5-2.3 Review of Fundamental Cement Placement Practices 174
5-2.4 Predictive Wellbore Stress Modeling 174
5-2.5 Cement Slurry Criteria for Hydraulically Fractured Gas Wells 176
5-2.5.1 Slurry Criteria for Optimized Placement 176
5-2.5.2 Slurry Criteria for Anti-Gas Migration 177
5-2.5.3 Slurry Criteria for Long-Term Zonal Isolation 178
5-2.6 Fracturing Constraints Required to Maintain Long-Term Zonal Isolation 179
5-3 Identifying Gas Pays, Permeability and Channels 179
5-3.1 Pay and Water Zone Logging Methods 179
5-3.2 Effect of Formation Clays and Micro-porosity 180
5-3.3 Wellbore Deviation and Resultant Logging and Flow Problems 181
5-3.4 Completion Considerations for Naturally Fractured Reservoirs 181
5-3.5 Formation Characterization for Well Completions 182
5-4 Sizing the Completion 183
5-4.1 Initial Design Considerations 183
5-4.2 Flow Factors for Tubing Design 184
5-4.3 Tubing Selection 185
5-4.4 Multi-Phase Flow and Natural Lift 185
5-4.5 Multiphase Flow and Flow Correlation Options 186
5-4.6 Critical Lift Factors 187
5-4.7 Liquid Hold-up and Back Pressure 188
5-4.8 Lift Options for Gas Wells 188
5-5 Completion Design for Flow Assurance 188
5-5.1 Completion Design for the Prevention of Gas Hydrates 188
5-5.2 Formation Damage in Gas Wells, Completion Damage and Scales 190
5-5.3 Organic Deposits and Condensate Banking 190
5-5.4 Effects of H2S and CO2 on Corrosion 191
VI
5-6 Sand Control for Gas Wells 192
5-6.1 Why is the Sand Flowing? 192
5-6.2 Is Sand Flow All Bad? 192
5-6.3 Establishing and Monitoringa Sand-Free Rate 193
5-6.4 Sand Control Methods for Gas Wells 194
5-6.5 Reliability of Sand Control Completions 194
5-6.6 Repairing and Restoring Productivity in Wells hat Flow Sand 194
Chapter 6
Fracture-to-Well Connectivity
6-1 Introduction 201
6-2 Completion Techniques and Their Impact on Well Connectivity 202
6-2.1 Cased-Well Isolation Techniques 202
6-2.2 Open-Hole Completions 205
6-2.3 Open-Hole and Uncemented Liner Fracture Treatment Diversion 205
6-3 Perforating in General 206
6-4 Perforating for Fracturing 206
6-4.1 Oriented Perforations 206
6-4.2 Deviated and Horizontal Well Perforating 208
6-4.2.1 Production Impairment from Inefficient Fracture-to-Wellbore Contact 209
6-4.3 Underbalanced vs. Extreme Overbalanced Perforating 211
6-5 Near-Wellbore Fracture Complexity 213
6-5.1 Near-Wellbore Complexity 214
6-5.2 Diagnosing and Quantifying Near-Wellbore Complexity (Tortuosity) 215
6-5.3 Minimizing the Effects of Tortuosity 217
6-6 Mid- and Far-Field Fracture Complexity 218
6-6.1 An Introduction to Complex Fracture Growth 219
6-6.2 Evidence of Complex Fracture Growth 220
6-6.3 Consequences of Complex Fracture Growth 220
Chapter 7
Fracturing Fluids and Formation Damage
7-2 Fracturing Fluid Function 228
7-2.1 Fracture Initiation 228
VII
7-2.2 Proppant Transport 229
7-3 Fracturing Fluid Rheology 230
7-3.1 Pressure Loss Gradient in the Fracture 232
7-3.2 Rheology in the Presence of Proppant Material and its Relation to Settling 234
7-3.3 Impact of Fluid Rheology on Fluid Loss 235
7-3.4 Calculation of Pressure Loss in the Wellbore Using Rheological Parameters
and the Virk Maximum Drag Reduction Asymptote 235
7-3.5 Advanced Rheology 235
7-3.6 Foam Rheology 236
7-3.7 Effect of Proppant on Rheology 237
7-3.8 Laboratory Rheology Measurements 239
7-4 Types of Fracturing Fluids 242
7-4.1 Water-Based Fluids 243
7-4.1.1 Low-Viscosity Fluids 243
7-4.1.2 Crosslinked Fluids 243
7-4.1.3 Borate Crosslinked Fluids 244
7-4.1.4 Metallic Ion Crosslinked Fluids 244
7-4.1.5 Delayed-Crosslink Systems 245
7-4.1.6 Function of Breakers in Water-Based Fluids 246
7-4.1.7 Water-Based Fluids in Gas Wells 246
7-4.2 Oil-Based Fluids 247
7-4.3 Energized fluids 248
7-4.4 Foams and Emulsions 249
7-4.5 Unconventional Fluids 250
7-4.5.1 Viscoelastic Surfactant Fluids 250
7-4.5.2 Viscoelastic Surfactant Foams 251
7-4.5.3 Emulsion of Carbon Dioxide with Aqueous Methanol Base Fluid 251
7-4.5.4 Crosslinked Foams 251
7-4.5.5 Non-Aqueous Methanol Fluids 252
7-4.5.6 Liquid CO2-Based Fluids 253
7-4.5.7 Liquid CO2-Based Foam Fluid 254
7-4.6 Acid Fracturing Fluid 254
7-5 Fracturing Fluid Additives 254
7-5.1 Additives for Water-Based Fluids 254
7-5.1.1 Friction Reducers 254
7-5.1.2 Gelling Agents 255
7-5.1.3 Biocide 257
7-5.1.4 Buffers 259
7-5.1.5 Crosslinkers 259
7-5.1.6 Breakers 260
VIII
7-5.1.7 Clay Stabilizers 262
7-5.1.8 Surfactants 262
7-6 Fluid Damage to Fractures and Sources of Productivity Impairment 262
7-6.1 Example Calculation of Productivity Impairment from Fracture Damage 264
7-6.2 Formation Damage from Saturation Changes 265
7-6.2.1 Fluid Retention 265
7-6.2.2 Rock/Fluid Interactions 267
7-6.2.3 Fluid/Fluid Interactions 267
7-6.2.4 Wettability Alterations 267
7-6.3 Formation Damage from Production 268
7-7 Fracturing Fluid Selection 268
7-7.1 Mineralogical Evaluation 269
7-7.1.1 X-Ray Diffraction (XRD) Analysis 269
7-7.1.2 Scanning Electron Microscopy (SEM) 270
7-7.1.3 Immersion Testing 271
7-7.1.4 Capillary Suction Time Testing 271
7-7.1.5 Core Flow Analysis 271
7-8 Selection of Fracturing Fluids for Applications in Gas Wells 273
Chapter 8
Proppants and Fracture Conductivity
8-1.1 Overview 283
8-1.2 The Evolution of Proppants 283
8-1.3 Fracture Conductivity 285
8.2 Conductivity Impact on Fractured Well Production Potential 286
8-2.1 How a Propped Fracture Benefits Well Flow Rate 287
8-2.2 Steady-State Solutions 288
8-2.3 Transient Solutions 288
8-3 Proppants 289
8-3.1 Sands 289
8-3.1.1 Ottawa Sands 290
8-3.1.2 Brady Sands 290
8-3.2 Ceramic Proppants 291
8-3.2.1 Sintered Bauxite 291
8-3.2.2 Intermediate Strength Ceramic Proppant 291
8-3.2.3 Lightweight Ceramic Proppant 292
8-3.3 Resin-Coated Proppants 292
IX
8-3.4 Ultra-Lightweight Proppants 294
8-4 Proppant Properties, Testing Protocols, and Performance Considerations 295
8-4.1 Proppant Testing Procedure Standards 295
8-4.2 Proppant Sampling 296
8-4.3 Grain Size and Grain Size Distribution 297
8-4.3.1 Proppant Size Testing 297
8-4.4 Proppant Shape 298
8-4.4.1 Proppant Shape Testing 299
8-4.5 Proppant Bulk Density and Apparent Specific Gravity 299
8-4.5.1 Proppant Bulk Density and Specific Gravity Testing 300
8-4.6 Proppant Quality 300
8-4.6.1 Acid Solubility Testing 300
8-4.6.2 Turbidity Testing 301
8-4.7 Proppant Strength 301
8-4.7.1 Proppant Crush and Fines Generation 302
8-4.7.2 Crush Testing 302
8-4.8 Proppant Concentration 303
8-5 Proppant Placement 305
8-5.1 Effects on Fluid Rheology 305
8-5.2 Convection 305
8-5.3 Proppant Transport 305
8-6 Fracture Conductivity 308
8-6.1 API Short-Term Testing Procedure 308
8-6.2 ISO Long-Term Testing Procedure 309
8-6.3 Non-Darcy Flow Testing 310
8-6.4 Multiphase Flow Tests 311
8-6.5 Gel Damage 312
8-6.6 Other Factors 313
8-7 Proppant Flowback 314
8-7.1 Proppant Flowback Control 314
8-7.2 Curable Resin-Coated Proppant 315
8-7.3 Proppant Flowback Control Additives 315
8-7.3.1 Tackifiers 315
8-7.3.2 Fibers 315
8-7.3.3 Deformable Particles 315
8-8 Proppant Selection 316
8-8.1 Productivity Potential 317
8-8.2 Flowback Control 317
8-8.3 Availability 317
8-8.4 The Cost-Value Proposition 318

Chapter 9
Execution of Hydraulic Fracturing Treatments
9-2 Function of Equipment 324
9-2.1 High-Pressure Pumping Equipment 324
9-2.2 Blending Equipment 325
9-2.3 High-Pressure Treating Lines and Manifolds 326
9-2.4 Nitrogen and Carbon Dioxide Pumping 326
9-2.5 Treatment Control Vans and Cabins 327
9-3 Equipment Quality Control 328
9-3.1 How Much Horsepower and What is the Pressure Rating? 328
9-3.2 How Many High-Pressure Lines and Suction Discharge Hoses to Use? 329
9-3.3 Standby Pumping and Blending Equipment 329
9-3.4 Absolute Essentials for Every Job 329
9-4 Quality Control for Fracturing Fluids 330
9-4.1 Quality Control of Water-Based Fracturing Fluids Before Arriving on Location 330
9-4.2 Fracture Fluid Blending Methods 334
9-4.3 Quality Control of Water-Based Fracture Fluids on Location 334
9-4.4 Quality Control of Other Fluid Systems 335
9-5 Quality Control of Propping Agents 336
9-5.1 Quality Control Guideline for Propping Agents 338
9-6 Quality Control and Execution of Acid Fracturing 338
9-6.1 Quality Control for Acid Fracturing 339
9-7 Multi-Stage Fracturing and Isolation Methods 342
9-7.1 Diverting Agents 342
9-7.2 Ball Sealers 342
9-7.3 Limited Entry 343
9-7.4 Multi-Stage Fracturing with Mechanical Isolation 344
9-7.5 New Multi-Stage Fracturing Technology 346
9-7.6 Horizontal Well Multi-Stage Fracturing 347
9-8 Pre-Fracture Diagnostics and Fracture Evaluation Tests 347
9-9 Real-Time Pressure Interpretation 350
9-9.1 Nolte-Smith Plot (see also Section 4-2.1.7) 350
9-9.2 Surface Treating Pressure as a Tool 351
9-9.3 The Effects of Perforations on Surface Treating Pressure 353
9-9.4 The Effects of Pipe Friction on Surface Treating Pressure 354
9-10 Fracturing Fluid Recovery (Flowback) 355
XI
Chapter 10
Fracturing Horizontal Wells
10-2 Production from Transversely Fractured Gas Horizontal Wells 365
10-2.1 A Calculation for Transversely Fractured Gas Horizontal Wells 366
10-3 Open-Hole Horizontal Well Completions 369
10-3.1 Perforating 370
10-3.2 Zonal Isolation 370
10-4 Open-Hole Fracturing 371
10-4.1 Acid Fracturing Execution 372
10-4.2 Proppant Fracturing Execution 372
10-4.3 Cleanup 373
10-5 Cased-Hole Completions 373
10-5.1 Cementing Horizontal Wells 373
10-5.2 Perforating Cemented Completions 374
10-5.3 Zonal Isolation in Cased Completions 375
10-6 Fracturing of Cased-Hole Completions 376
10-6.1 Acid Fracture Execution 376
10-6.2 Proppant Fracturing Execution 377
10-7 Rationale and Conditions of Fracturing
Horizontal Wells in Gas Formation 377
Chapter 11
Unconventional Gas
11-2 Description of Unconventional Reservoirs 383
11-3 Production Mechanisms 385
11-3.1 CBM (Coalbed Methane) 385
11-3.2 Shale Gas Reservoirs 385
11-3.3 Shale Gas Reserves 386
11-4 CBM Reservoirs 387
11-4.1 Coalbed Description 387
11-4.2 CBM Fractured Systems 388
11-4.3 Adsorption/Desorption 390
11-4.4 Stimulation Techniques 391
11-4.5 Alternate Completions and Enhanced Production Techniques 393
XII
11-4.6 Fracture Modeling of CBM Wells 396
11-4.7 Fracturing Treatment Evaluation of CBM Wells 397
11-4.8 Estimation of Reserves and Production Data Analysis 398
11-5 Shale Gas 400
11-5.1 Shale Description 400
11-5.2 Thermogenic and Biogenic Systems 401
11-5.3 Ft. Worth Basin Barnett Shale 402
11-5.3.1 Barnett Shale Slickwater Treatment Design Considerations 404
11-5.4 Barnett and Woodford Gas Shale, Delaware Basin 406
11-5.5 Fayetteville Shale in Arkansas 409
11-5.5.1 Treatment Design Considerations Fayettville Shale 409
11-5.6 Woodford/Caney Shale, Arkoma Basin 410
11-5.7 Floyd Shale/Conasauga Shale, Black Warrior Basin (Alabama) 412
11-5.8 Mancos and Lewis Shales 412
11-6 Shale Treatment Design and Evaluation 413
11-6.1 Stimulation and Treatment Design for Shale Reservoirs 413
11-6.2 Fracture Modeling 416
11-6.3 Summary 416
Chapter 12
Fracturing for Reservoir Development
12-2 Impact of Fracturing on Reservoir- or Drainage-Wide Production 428
12-2.1 Example Application of Infield Drilling and Fracturing of Gas Wells 429
12-2.2 Transient Flow of Fractured Gas Wells 430
12-3 Forecasting Natural Gas Well Performance and Recovery 431
12-3.1 A Case Study for Reservoir Recovery Using Unfractured and Fractured Wells 431
12-3.2 Field Development Strategy 432
12-4 Impact of Fracture Azimuth on Well Planning 434
12-4.1 Determination of Fracture Azimuth 435
12-4.2 Considerations Regarding Directional Permeability in the Reservoir 435
12-4.3 Barnett Shale Case Study 437
12-5 Data Mining Techniques 441
12-5.1 Purpose of Data Mining 441
12-5.2 Data Sources 441
12-5.3 Data Preparation 442
12-5.4 Selected Data Mining Tools 442
12-5.5 Data Mining Case History 443
XIII
Chapter 13
Technologies for Mature Assets
13-1.1 Definition of a Mature Asset 455
13-1.2 Minimum Cost & Maximum Value 456
13-1.3 Motivation for Fracturing 457
13-1.4 New Technologies/Approaches 458
13-1.5 Reducing Treatment Costs 462
13-2 Candidate Selection 464
13-2.1 Regional Considerations 464
13-2.2 Neighborhood Considerations 465
13-2.3 Localized Considerations 466
13-2.4 Risk Ranking and Data Manipulation 467
13-2.5 Case Histories and Results 468
13-3 Fracture Design in Mature Fields 469
13-4 Depletion Considerations 470
13-4.1 Pore-Pressure Considerations 470
13-4.2 Fracturing Fluid Selection 472
13-4.3 Proppant Selection 473
13-4.4 Cleanout and Flowback 474
13-4.5 Mechanical Deployment 476
13-5 Re-Fracturing Operations 479
13-5.1 Re-Fracturing Case Histories 480
13-5.2 Candidate Selection for Re-Fracturing 481
13-5.3 Re-Fracture Re-Orientation 481
13-5.4 Improved Treatment Design 483
Nomenclature 491
Index 503
XIV
Preface
I
It is with great pleasure that I welcome you to Modern Fracturing: Enhancing Natural Gas Production. BJ Services
Company is proud to be involved in developing and publishing this work. We hope you find this book to be
instructive, informative and interesting.
This book is intended for use by all industry professionals, not just those who are already familiar with the engineering
concepts and field practices of hydraulic fracturing. The pages within comprise a state-of-the-art engineering manual
for planning, preparation, performance and evaluation of hydraulic fracture treatments in natural gas reservoirs. We
envision industry professionals throughout the world benefiting from the information in this book.
Hydraulic fracturing is already the completion method of choice for most natural gas wells in North America. As
global dependence upon natural gas increases, it seems likely the application and popularity of this completion
method will only increase further and spread farther. The techniques described within this book are applicable to all
gas reservoirs, not just to the low permeability formations typically developed in North America. We firmly believe
fracturing is the best possible completion technique for each and every gas reservoir throughout the world.
A wide range of knowledgeable authors from throughout the industry have come together to produce this book. On
behalf of BJ Services, I want to thank them for their sharing their experience and knowledge, as well as for their hard
work and dedication in completing such an ambitious project. We feel certain that in the years to come, each author
will continue to be proud of his or her involvement in this undertaking. We also trust that readers like you will
continue to improve best practices in developing natural gas resources worldwide with the insights derived from
this significant work.
Dave Dunlap
Executive Vice President and Chief Operating Officer, BJ Services
XV
Foreword
I
I was very pleased when my friend Michael Economides asked me to write the Preface to his new book. BJ Services
Company should be complimented for sponsoring this effort and for attracting some of the worlds top experts to
contribute. I know many of the contributors, and I am sure the result will be lasting and useful for years to come.
I am even more pleased that this specific book is put together for three reasons. The first is that natural gas will shortly
become the premier fuel of the world economy. Second, hydraulic fracturing, already the most important production
enhancement technique for oil wells, is absolutely indispensable for natural gas wells. Third, the existing know-how and
skill sets of the fracturing community are dreadfully inadequate, especially in management.
Fracturing in the petroleum industry is no longer an experimental or daring activity by some hot-shot, brash engineers,
often working against the established old thinking and even worse, conservative managers who still believe that economics
equal cost reduction, ignoring the benefit from improved well performance. When enhanced production and injection
performance is the motivation, nothing can compete with properly integrated fracturing.
Often, people are confused about the real impact from this well completion and stimulation technique. Most often, any
improvement in production compared to what a well did before fracturing is considered a success. In reality, we already
know how much a well should be producing after fracturing by using the concept of maximizing the JD, the dimensionless
productivity index. Anything less than that should be considered a performance gap and managed as such. We have to
push the limits and manage the completion and execution community to deliver what we know can be done.
All activities in a company must be integrated with hydraulic fracturing. We are by definition can-do people. So the idea
that ultra-high production targets are unrealistic and theoretical should be replaced by developing and implementing
the know-how and skill sets to deliver maximum performance.
Consider this: When my associates and I (including Michael) were working in Russia, in a five-year period we managed
to double a companys production, increasing by 20% per year to almost 2 million barrels per day while shutting-in 50%
of the original well stock. Most of this success occurred by pushing the limits of hydraulic fracturing and integrating the
other parts of the production system. And despite this success, we were constantly enhancing materials and increasing job
sizes to push the calculated performance limits. We established two management rules:
1. All new wells and workovers must be fractured unless top management approves otherwise.
2. All frac jobs must be designed and executed to perform at the peak of the NPV bell curve unless top management
approves otherwise.
The point is that many companies require approval to do it right but delegate enough financial authority, no approval
required, to do it wrong. We reversed this by giving enough authority (no approval required) to do it right and required
top management approval to do it wrong.
It is not so difficult to reproduce the same performance everywhere else. Just look at current worldwide well performance,
and one can easily see huge gaps, including the largest and best-known multinational oil companies. Fracturing can go a
long way to correct this obvious problem. Not only will the benefit to companies be immediate and large, but silly talk
about peak oil and twilight in the desert will go away.
Joe Mach - February 2007
XVII
Contributing Authors
Editors
Michael J. Economides, University of Houston

Tony Martin, BJ Services
Authors
Bob Bachman, Taurus Reservoir Solutions David Mack, Marathon Oil

Steve Baumgartner, BJ Services Mark Malone, BJ Services
Harold Brannon, BJ Services Tony Martin, BJ Services
Andronikos Demarchos, Hess Corporation C. Mark Pearson, Golden Energy, LLC
Michael J. Economides, University of Houston David Ross, InTuition Energy Associates Ltd.
John Ely, Ely & Associates, Inc. Martin Rylance, BP
Satya Gupta, BJ Services Gary Schein, BJ Services
Robert Hawkes, BJ Services Peter Valk, Texas A&M University
Barry Hlidek, BJ Services Leen Weijers, Pinnacle Technologies
George King, BP Xiuli Wang, BP
Randy Lafollette, BJ Services Don Wolcott, Aurora Oil and Gas
XVIII
Acknowledgements
F
First and foremost, the editors would like to express their sincere gratitude to JC Mondelli, who has been the
champion of this book within BJ Services from its initial conception, all the way through to printing and publication.
Without his perseverance and vision, this publication would never have come about. We would also like to thank
the senior management of BJ Services for providing funding and, especially, for allowing a great number of highly
dedicated people to put their time and energy into writing chapters, in spite of their busy schedules.
Our thanks to Joe Mach for gracing the book with his Preface and endorsement, and who also, in his unique
style, reminded all of us why doing this book mattered in the first place.
Writing this book was an added task both for our BJ Services colleagues and those from other companies and
institutions, and the result is a testament to their dedication and professionalism. Putting together a multi-authored,
multi-edged book is never an easy task and to no small measure, the authors deserve particular praise for persevering
and having to respond to suggestions and editorial interference by two admittedly highly demanding and opinionated
Editors. Compliments and credit are deserved by all of them, without whom this project would not have been
possible.
Special thanks go to Greg Salerno who shepherded many of the logistical tasks and kept a level-headed approach
on the day-to-day management of the project. Thanks also to Garth Gregory and Margaret Kirick for their invaluable
help with the organisation and administration of this undertaking.
The copy-editor Stephanie Weiss served a key role in the final version of the book. She is a highly experienced
and exceptional technical copy editor, a formidable vacuum cleaner for cleaning up deficiencies, omissions and
errors. Her work reminded all that adherence to detail and perfection are essential in elevating a professional book to
a different level. She was a rare find.
Alexander M. Economides and his staff in the Energy Tribune, headed by Jay Clark and the publication assistants
Alex Lewis and George Song, did a spectacular job in producing the book. They deserve special praise.
Michael J. Economides and Tony Martin - September 2007
XIX
Michael J. Economides is a professor at the Cullen College of Engineering, University of Houston, and the
managing partner of a petroleum engineering and petroleum strategy consulting firm. His interests include
petroleum production and petroleum management with a particular emphasis on natural gas, natural gas
transportation, LNG, CNG and processing; advances in process design of very complex operations, and
economics and geopolitics. He is also the editor-in-chief of the Energy Tribune. Previously he was the
Samuel R. Noble Professor of Petroleum Engineering at Texas A&M University and served as chief scientist
of the Global Petroleum Research Institute (GPRI). Prior to joining the faculty at Texas A&M University,
Economides was director of the Institute of Drilling and Production at the Leoben Mining University
in Austria. Before that, he worked in a variety of senior technical and managerial positions with a major
petroleum services company. Publications include authoring or co-authoring 14 professional textbooks and
books, including The Color Of Oil, and more than 200 journal papers and articles. Economides does a wide
range of industrial consulting, including major retainers by national oil companies at the country level and
by Fortune 500 companies. He has had professional activities in over 70 countries.
Tony Martin is business development manager for international stimulation at BJ Services Company. Since
graduating from Imperial College, London, with an honors degree in mechanical engineering and a master's
degree in petroleum engineering, Martin has spent 17 years in the oil industry and has completed engineering
assignments around the world. Martin's primary interest has been hydraulic fracturing and stimulation, and
he has been involved in production enhancement projects in more than 25 countries. He teaches fracturing,
acidizing and sand control both in-house and externally. A constant theme in this teaching is the need
to de-mystify the world of hydraulic fracturing, in an attempt to make the process more accessible and
less intimidating. He is the author or co-author of numerous SPE papers and has served on the technical
committees for several SPE events. He is also the author of BJ Services Hydraulic Fracturing Manual.
Chapter 1 mainstay of the developed world such as the United
States and Europe, can be revitalized through
Introduction to this Book this process (Chapter 13).
Michael J. Economides, University of Houston and Before the technical issues are addressed it is
Tony Martin, BJ Services essential to look at natural gas in the world economy,
why it is becoming increasingly important and what
are the reasons for all the excitement surrounding its
1-1 Introduction enhanced production.
This is a book about enhancing natural gas production 1-2 Natural Gas in the World Economy
using one of the most important and widespread well
completion technologies hydraulic fracturing. Although natural gas, with some 23% of all world energy
The book addresses the way that natural gas is demand in 2005, is still slightly behind coal (25.6%) as
produced from natural reservoirs (Chapter 2) and then the worlds third-largest source of primary energy (oil
describes diagnostic techniques that can pinpoint whether still dominates at 38%), it is poised to move up because
the well is producing as it should or whether intervention of significantly emerging new trade. Member countries
should be undertaken (Chapter 3), which is the central in the Organization for Economic Co-operation and
theme of this book. Development (OECD) and the USA, specifically,
Hydraulic fracturing is introduced as the solution consume about 51% and 22% respectively of global
of choice, showing the idiosyncratic nature of natural natural gas, now comprising about 103 Tcf (2.9 Bm3) per
gas wells compared to oil wells (Chapter 4). The year (Energy Information Administration, EIA, 2007).
subsequent two chapters address important peripheral
issues whose successful or failed resolution may affect
the well performance with equal or even more serious
consequences than the fracture treatment itself. These
issues include well completions (Chapter 5) and the
extremely important well-to-reservoir (and fracture)
connectivity (Chapter 6).
The next two chapters deal with materials for
fracturing: fluids and proppants (Chapters 7 and 8). Their
selection is essential to the successful execution of the
treatment. The execution itself becomes the next chapter,
and practical issues are addressed there (Chapter 9).
Then some modern applications are described.
One chapter deals with fracturing horizontal wells,
increasingly an important option among reservoir
exploitation strategies (Chapter 10). Not only new
well architecture but also newer reservoir targets are Figure 1-1 The top 12 holders of natural gas reserves:
Russia, Iran and Qatar dominate (EIA, 2006, BP
opening up, and natural gas demand points towards Statistical Review, 2006, ET, 2007)
unconventional sources, namely coalbed methane
(CBM), shale gas and very low-permeability formations. There are several obvious benefits to the use of
Technology makes their exploitation possible, and this natural gas. First, it is the cleanest-burning fossil fuel
is the subject of the next chapter (Chapter 11). and produces fewer emissions and pollutants than either
Finally, two issues round out the book: oil or, especially, coal. Second, the resource is becoming
Fracturing is employed in the full development of increasingly diverse. Since the early 1970s, world reserves
reservoirs (Chapter 12); and how mature fields, a of natural gas have been increasing steadily, at an annual

Modern Fracturing
Table 1-1 Top 25 Countries Ranked According to Proved Natural Gas Reserves and identifying the proved
reserves-to-production ratio (R/P) for each country
Proved Natural Gas Reserves at January 1, 2006
Reserves /
Trillion Cubic Trillion Cubic Cumulative
Country Share of Total Production (R/P)
Feet (Tcf) Meters (Tm3) Share of Total
Years
1 Russian Federation 1688 47.8 26.6% 26.6% 80.0
2 Iran 944 26.7 14.9% 41.5% >100
3 Qatar 910 25.8 14.3% 55.8% >100
4 Saudi Arabia 244 6.9 3.8% 59.6% 99.3
5 United Arab Emirates 213 6.0 3.4% 63.0% >100
6 USA 193 5.5 3.0% 66.0% 10.4
7 Nigeria 185 5.2 2.9% 68.9% >100
8 Algeria 162 4.6 2.5% 71.5% 52.2
9 Venezuela 152 4.3 2.4% 73.9% >100
10 Iraq 112 3.2 1.8% 75.6% >100
11 Kazakhstan 106 3.0 1.7% 77.3% >100
12 Turkmenistan 102 2.9 1.6% 78.9% 49.3
13 Indonesia 97 2.8 1.5% 80.5% 36.3
14 Australia 89 2.5 1.4% 81.9% 67.9
15 Malaysia 88 2.5 1.4% 83.2% 41.4
16 Norway 85 2.4 1.3% 84.6% 28.3
17 China 83 2.4 1.3% 85.9% 47.0
18 Egypt 67 1.9 1.1% 86.9% 54.4
19 Uzbekistan 65 1.9 1.0% 88.0% 33.2
20 Canada 56 1.6 0.9% 88.8% 8.6
21 Kuwait 55 1.6 0.9% 89.7% >100
22 Libya 53 1.5 0.8% 90.5% >100
23 Netherlands 50 1.4 0.8% 91.3% 22.3
24 Azerbaijan 48 1.4 0.8% 92.1% >100
25 Ukraine 39 1.1 0.6% 92.7% 58.7
Total World 6347.79 179.82 100% 65.1
Sum of Top 25 Countries 5885 166.7 92.7%
Rest of World 463 13.1 7.3%
rate of some 5%. Similarly, the number of countries primarily power generation. There is yet little overlap
with known reserves has also increased from around 40 between the use of natural gas and oil in all large
in 1960 to about 85 in 2005. The distribution among markets. However, certain developments on the
those countries, dominating the global proved reserves of horizon, including the electrifying of transportation,
natural gas, is shown in Fig. 1-1 and Table 1-1. will push natural gas use to ever higher levels.
One reason for anticipated increase in demand for Although potential natural gas supplies abound
natural gas is the public concern over environmental throughout the world, facilities and infrastructure
issues. Furthermore, forecasts of rapid increase in to receive and distribute the product to market are
natural gas demand over the next two decades, in the expensive to build, and their development can easily be
biggest market of all, the United States, have been hindered by geopolitics. These reasons have historically
exacerbated by forecasts of declining production. inhibited natural gas from reaching its full potential in
Declining production forecasts have been extended to the worlds energy markets. Natural gas is transported
Canada, a reliable provider to the US thus far (EIA, either by pipeline (73% of internationally traded gas
Annual Energy Outlook, 2007). in 2005, EIA 2007), mainly across land masses, and
Although natural gas demand is expected to by liquefied natural gas (LNG) transportation across
increase, such an increase in the near future will the oceans (the remaining 27%). The rapid expansion
be driven by additional demand from current uses, of LNG infrastructure worldwide in the past decade is

Chapter 1 Introduction to this Book
enabling natural gas to penetrate many more markets in its liquid state becomes economic. But the supply
through the development of many remote reserves once chain consists of expensive and specialized facilities
considered to be stranded and uneconomic to develop. both upstream and downstream, and generally requires
Ongoing construction and plans to expand and build dedicated marine vessels.
new LNG receiving terminals in North America The LNG industry is set for a large and sustained
(Canada, Mexico and the United States) are opening expansion as improved technology has reduced costs
up rapidly growing gas imports, destined to support and improved efficiency along the entire supply chain
many new LNG supply chains worldwide. European during the past decade. This shift in the dynamics of
and Asian markets are also hungry for LNG. the natural gas market will further commoditize and
But beyond the usual energy-demanding markets, diversify the natural gas globally. New LNG carriers
China and India have both emerged from the developing are 1000 ft long and require a minimum water depth
world to become globally significant economies in their of 40 ft when fully loaded. The global fleet of LNG
own right, both requiring massive energy imports to carriers reached 217 by the end of 2006 (Wood et al.,
sustain future economic growth. But their approaches are 2006) with more than 11 million tons of LNG capacity.
very different; China is focused on manufacturing, India The order book for new LNG marine carriers to 2010
more on services. However, both have large populations is some 120 firm and 32 proposed, meaning the future
with aspirations to lead high-energy consuming lifestyles. fleet may exceed 370 vessels by the end of 2010. The fleet
Together, they are promoting globalization that is putting was just 90 vessels in 1995 and 127 vessels in 2000. The
pressure on the worlds energy resources and existing current fleet transports more than 140 million metric
supply chain, traditionally directed to serving the OECD tons of LNG every year (converted to 7 Tcf), about
world. The rapid growth in China and India over the 23% of gas trade internationally and about 6.5% of
last few years has precipitated huge increases in demand total gas consumed worldwide.
for all energy sources, because of their lack of sufficient Below is a discussion of the state of natural
indigenous energy resources. This has left the rest of gas in three of the most important countries/
the world scrambling for the same sources of energy, regions of the world which, for different reasons,
including natural gas. The US is hampered by the myriad are defining the present and future of natural
permit approvals required and public opposition to siting gas in the world economy.
of LNG receiving terminals. Nevertheless, major US
companies and others are investing heavily in building 1-3 Russia: A Critical Evaluation of its
new LNG liquefaction infrastructure in Qatar, several Natural Gas Resources
countries in West Africa and Russias Sakhalin Island.
Transportation is an essential aspect of the gas The dissolution of the Soviet Union in 1991 and its
business because gas reserves are often quite distant replacement by the Commonwealth of Independent
from the main markets. Gas is far more cumbersome States (CIS), prominent among which was the Russian
than oil to transport, and the majority of gas is Federation, was a significant geopolitical event, affecting
transported by pipeline. There are well-developed the subsequent development of Russian resources
networks in Europe and North America and a relatively particularly natural gas. Contrary to widely held beliefs,
adequate one in the former Soviet Union. However, if current trends continue, Russia likely will have a severe
in its gaseous state, natural gas is quite bulky for the natural gas shortfall by 2010 (Moscow Institute of Energy
same time, a high-pressure pipeline can transmit only Research, 2007). This prediction is astonishing, given
about one-fifth of the amount of energy that can be that Russia has more gas reserves than any other country,
transmitted in an oil pipeline of the same size, even and one of the largest reserves-to-production ratios.
though gas travels much faster. When gas is cooled to One of the reasons for the looming gas shortfall is
160 C it becomes liquid and much more compact, that over the past several years, Russia has not invested
occupying 1/600 of its standard gas volume. Where sufficiently and lacks the technology to develop new gas
long overseas distances are involved, transporting gas fields to replace its rapidly depleting ones.

Figure 1-2 Russian gas resources, infrastructure, pipelines and future plans
There are complicated reasons behind Figure 1-1 compares Russian gas reserves with those
the state of Russias natural gas industry. A of the other major gas producing countries. Table 1-2
thorough understanding of the industry and its lists the 13 largest gas fields in the world. As is shown,
history is required before we can discuss its future Russia owns two-thirds of them (ET, 2007, EIA, 2007,
(see Section 1-3.2). www.Gazprom.com, and BP Statisitcal review, 2006).
Next, we examine Russias natural gas reserves, Gazprom, tracing its origins to the Soviet Gas
production and transportation. Ministry, is the dominant gas company in Russia.
Fig. 1-3 shows Russias total gas production and
1-3.1 The Resource Base consumption and Gazproms contribution from 2000
to 2005, which accounts for about 80%. Gazprom
Russia has the worlds largest proven natural gas is not only Russias largest gas producer, it also owns
reserves, estimated at 1,680 Tcf (EIA, 2007), about the entire gas pipeline infrastructure in Russia all
double those of Iran, the next largest. Russia is also the 155,000 km of it, along with the compressor stations.
largest gas producer and exporter. In 2004, Russias gas In addition, Gazprom controls the sole means of
production exceeded 22.4 Tcf and exports totaled 7.1 getting gas to domestic and export markets.
Tcf. In addition, the gas industry plays a significant
24
role in the Russian economy, contributing about
26% of total GDP in 2004 (ET, 2007). Fig. 1-2 is an
22
annotated map of Russia with all important natural
gas-related information (EIA, 2007, www.Gazprom. 20
Total Production
com, and BP Statistical review, 2006).
Tcf/year
18
Table 1-2 The Worlds Largest Natural Gas Gazprom's Share

16
Reservoirs (EIA, 1994-2004, Interfax,
2005,www.gazprom.com, ET, 2007)
14
Total Consumption
*Urengoy had been the worlds
largest gas field for years until the 12
North Dome was discovered. 2000 2001 2002 2003 2004 2005
Umm Shaif/Abu el-Bukush
450
Figure 1-3 Russian gas production and consumption and
400
Gazproms contribution (EIA, 2004-2006, Interfax, 2005,
Urengoy*
350 www.gazprom.com, ET, 2007)

North Dome, 1,200 (Tcf)
Shtokman
300
Orenburg
Yamburg
Zapolyarnoye
Panhandle-Hugoton
The reason that Russia has given Gazprom control

Kharasevey
250
Bovanenko
over its natural gas is the so-called social obligation.

Medvezhye
Hassi RMel
South Pars
200
Through Gazprom, the Russian government subsidizes
150
its inefficient domestic industries with low-priced
100
natural gas. Gazprom sells most of its gas to domestic
50
customers at a considerable discount. The wholesale
0 price of 1,000 m3 of gas for a Russian household is
Rank Field Reserves Location
1 North Dome 1,200 Qatar/Iran around $15.90 (about $0.45/Mscf ). For industrial
2 Urengoy 275 Russia
3 Yamburg 200 Russia users, gas costs around $24.20 ($0.69/Mscf ). By
4 Orenburg 200 Russia
5 Shtokman 200 Russia
comparison, in the European Union, household tariffs
6
7
Umm Shaif/Abu el-Bukush
Zapolyarnoye
175
150
Abu Dhabi
Russia
range from Finlands $159 ($4.50/Mscf ) to Denmarks
8 Kharasevey 150 Russia $735 ($20.82/Mscf, ET, 2007). Clearly, Gazprom
9 Bovanenko 125 Russia
10 Medvezhye 100 Russia is losing large amounts of money on domestic sales,
11 Hassi RMel 100 Algeria
12 South Pars 100 Iran compared to international market prices, and must rely
13 Panhandle-Hugoton 80 U.S.A.
on export revenues for the difference.

Modern Fracturing
Gazproms major challenge is the aging of its major and Sakhalin. Obviously, implementing these solutions
producing gas fields. Production from these fields is will require a substantial investment that Gazprom has
declining and studies project steep declines in Russias not yet been able to make.
overall natural gas output between 2008 and 2020. One scenario for the potential contribution
According to projections from the Moscow-based Institute of independent producers shows a net increase of
of Energy Research (2006), Russia will face a gas shortfall 100 Bm3 per year by 2010 (Moscow Institute of
of about 100 Bm3 by 2010. Considering that Russia owns Energy Research, 2006).
the largest gas reserves in the world and one of the largest
600
reserves-to-production ratios (81.5 years compared to
Algerias 55.4 and Canadas 8.8, for example, from EIA, Gazp
ro ms
fo rec
2007, calculated by ET, 2007), the future of Russian ast
p
500 rod
uct
natural gas production efforts is important globally. io nd
ec
lin
e
1-3.2 Russian Natural Gas Production 400
Combined Bcm/year
Gazprom holds about one-third of the worlds natural
gas reserves and produces about 80% of Russias 300
natural gas. The remaining percentage comes from
independent producers. The company operates 155,000
km of natural gas pipeline and 43 compressor stations. 200
As the worlds largest producer and exporter, Russia
is also a huge consumer of natural gas. The country
produces an annual 21 Tcf, consuming 14.5 Tcf and 100
exporting the rest (2002 numbers from EIA and ET,

2007). Despite the countrys huge reserves, natural
gas production has remained essentially flat over the 0
2004 2010 2015 2020
past several years, with a mild production increase
Others Zapadno-Tarkosalinskoye
(1.3%) forecast for 2008. In contrast to the natural gas Orenburg Komsomol'skoye
Astrakhan Zapolyarnoye
stagnation, oil production has flourished. Urengoyskoye(achimov)
Ety-Purovskoye
Medvezhye
Aner'yakhinskoye
The immediate future of natural gas production Yuzhno-Russkoye
Vyngayahinskoye
Kharvutinskoye
Yabburgskoye
in Russia does not allow for much optimism. The Pestsovoye
Yubileynoye
En-Yakhinskoye
Urengoyskoye
overall production decline forecast for Gazprom is Figure 1-4 Gazproms production decline forecast (Moscow
quite steep, as shown in Fig. 1-4 (Moscow Institute of Institute of Energy Research, 2007)
Energy Research, 2006).
Considering that Russias domestic consumption 1-4 Alaska, its Natural Gas Resources
is increasing by 2.5% annually, the current demand and their Impact on US Imports
in Europe, Turkey and the Commonwealth of
Independent States (CIS) for up to 325 Bm3 (ET, It has been known for many decades that Alaska has
2007), and Chinas demand for 38 Bm3 (Moscow prolific hydrocarbon resources, first with the discovery
Institute of Energy Research, 2006) its clear that of oil in the south central part (Cook Inlet) in the
additional sources of natural gas must be found if 1960s and then with the 1969 discovery of Prudhoe
Russia wants to play a major role in the future natural Bay, the USs largest field. Oil has been successfully
gas market. Its equally clear that the problem of Russias commercialized in Alaska since the 1970s construction
looming gas shortage can only be solved by optimizing of the Trans-Alaskan pipeline that stretches from the
existing fields and through the rapid development and North Slope to Southern Alaska. From there, oil is
production of major fields such as Yamal, Shtokman shipped to the lower 48 states.

Despite the success of Alaskan oil production, Clearly, Cook Inlet gas production is on decline while
and although it is widely known that natural gas North Slope gas production remains stable with its
exists in large quantities in the state, two important market limited to the local market without a natural
questions have always arisen: 1) in what kind and gas export pipeline to larger markets.
size of reservoirs is the gas trapped and 2) how can
600
it be commercialized? Furthermore, after 30 years of Cook Inlet North Slope
Alaskan oil production and almost 15 years after its 500
production peak, substantial natural gas exploitation 400

from the state is still not forthcoming.
Bcf/Year
300
We are convinced that Alaska has a very large
200
natural gas resource base, larger than commonly
accepted. Beyond the conventional gas reserves 100
on the North Slope (about 100 Tcf ) and Cook 0

Inlet (at least 30 Tcf ), perhaps as much as 1000
1958
1964
1970
1976
1982
1988
1994
2000
2006
2012
2018
2024
Tcf are in the form of coalbed methane and, at
Figure 1-5 Historic and forecast gas production (Alaska
least, 500 Tcf as natural gas hydrates (Anchorage Department of Natural Resources, 2006)
Chamber of Commerce, 2005).
Economic and technical obstacles abound. The The forecast in Fig. 1-5 is only for the current proved
cost for exploiting conventional reserves, with or reserves of natural gas. If we consider the unconventional
without government subsidies, has been a hindering resources in Alaska, the natural gas resource base grows
factor, but other factors such as the emerging much larger. However the technology and economics
large LNG trade are having an impact. The most for developing the unconventional resource base are
important question is whether Alaskan gas will be major blockers. The two main unconventional gas
commercialized any time in the foreseeable future, reservoirs that capture a lot of attention are coalbed
and we shall discuss this issue in detail. This has major methane and natural gas hydrates.
implications on the future of the state, the USA and It is estimated that coalbed methane is prevalent
the natural gas trade into the country. in the northern and southern parts of the state, shown
on the map in Fig. 1-6 (Alaska Department of Natural
1-4.1 Alaskan Reserves and Production Resources, 2006).
There are two major hydrocarbon producing areas Guadalupe Sonora

Barrow 15
in Alaska today: the Cook Inlet region in south-
Russia Hermosillo
central Alaska and the Prudhoe Bay complex on the
North Slope. The proved gas reserves for the Cook Alaska
Fort Yukon
Inlet and the North Slope are 2 Tcf (6% of total) Empalme
Guaymas
and 27 Tcf (94% of the total), respectively (EIA, Nome Fairbanks
2007). Currently all the gas produced on the North Ciudad Obregon
Canada
Navojoa
Slope is re-injected for pressure maintenance except Anchorage Huatabampo
Bering Sea
for the gas needed to maintain field operations Bituminous & Higher Rank Cordova
Los Mochis
and fuel the local villages. Subbituminous
Ciudad Constitucion
Figure 1-5 shows the historical production and Lignite
the prediction of natural gas production to 2025. Rual Sites with Sufficient
Data for Drill Testing 1
Baja California Sur La Paz
As can be seen, the 2006 production from the two of Coalbed Methane
Potential Pacific Ocean
areas is approximately 490 Bcf per year of gas and
is expected to decrease to 240 Bcf per year by 2025 Figure 1-6 Location of potential coalbed methane
(Alaska Department of Natural Resources, 2006). reservoirs (Alaska Department of Natural Resources, 2006)

Modern Fracturing
Alaskas estimated coal resources exceed 5.5 A gas pipeline from the North Slope through Canada
trillion tons and may contain up to 1,000 Tcf of gas to the Lower 48 states.
(Alaska Department of Natural Resources, 2006). An All-Alaska gas pipeline from the North Slope to
In 1994 the Alaska Div. of Oil and Gas drilled the Valdez, where the gas would be converted into LNG
states first coalbed methane test well near the town of and taken to markets outside Alaska in LNG tankers.
Wasilla, located in the northern portion of Cook Inlet A spur line to take natural gas from one or more
Basin. The well was drilled to a total depth of 1245 ft; off-take points on the main gas pipeline (whichever
coal was continuously encountered, with the thickest route it takes) and deliver that gas to customers and
seam measuring 6.5 ft and a net coal thickness of 41 users in Alaska.
ft. Thirteen seams were sampled for gas content. The
results were encouraging, but as elsewhere they are All Alaska
likely to suffer from the standard CBM problems: LNG shipped from All Alaska
Y-Branches from All Alaska
low permeability, water disposal and difficult and Northern Route Sonora
Guadalupe
expensive application of hydraulic fracturing and Southern Route Barrow 15
horizontal well technologies. Hermosillo

Our current assessment of the total resource base Alaska
for natural gas in Alaska, derived from a number of Fort Yukon
Empalme
references, is shown in Fig. 1-7. Nome Fairbanks
Guaymas
Ciudad Obregon
Cook Inlet North Slope Canada
Navojoa
Conventional, Conventional,
Anchorage Huatabampo
30, 2% 100, 6%
Bering Sea
Cordova
Los Mochis
Juneau
Ciudad Constitucion
Guasave
North Slope Sinaloa
Prince
Hydrates, Rupert
Unalaska 1
529, 32% Baja California Sur La Paz
Pacific Ocean
CBM, 1000,
60% Figure 1-8 Potential Alaskan natural gas pipeline routes
Figure 1-7 Natural gas resource base in Alaska (Williams

There are two variations on the gas pipeline to the
et al., 2005, Meyers, 2005, Hite, 2006, and Kornfeld, 2002) Lower 48 states proposal: the Northern Route and the
Southern Route. The Northern Route, also referred to as the
It is clear the 2006 resource assessment shows ARC over-the-top route (ARC is for the Artic Resources
the majority of potential reserves are locked in Company that first proposed such a gas pipeline in
unconventional reservoirs. For these plays to be the early 1980s), would start from Prudhoe Bay, move
developed, investment and technology hurdles will offshore into the Beaufort Sea and run parallel to the
need to be overcome. coastline eastward into Canada to the Mackenzie River
Delta, where up to 20 Tcf of natural gas reserves are just
1-4.2 The Uncertain Destiny of the North Slope of waiting to be produced. From there, if Canadians have
Alaska Natural Gas already built a pipeline to transport the Mackenzie River
reserves to Alberta, the Alaskan Northern Route would
Methods to deliver natural gas to market from the North simply reach and merge with it.
Slope of Alaska have been studied and proposed for over On the other hand, if Canadians havent started
30 years. The various schemes can be grouped into three yet to exploit the Mackenzie Delta reserves and
major categories, with variations in each (Anchorage a pipeline to Alberta is not available, the Northern
Chamber of Commerce, 2005). See Fig. 1-8. Route pipeline would be extended to Alberta, and
10
in all probability it would still offer the prospect to successful for many years, Canada can no longer
carry also the Mackenzie gas along with the North be relied upon to single-handedly secure the future
Slope gas. After the Alaskan natural gas is delivered in of US natural gas supply.
Alberta, there still would be the open issue of how to A declining conventional natural gas resource
carry it down to the rest of the United States. has pushed Canada into investing in arctic, CBM and
The two systems currently discussed and proposed tight gas plays. To date however, those unconventional
to accomplish this goal are: resources have contributed a very small percentage to
that countrys overall production of natural gas.
Rerouting Prudhoe Bay natural gas in the Canadian As is apparent in Fig. 1-9, the conventional
pipelines network that currently delivers Alberta gas natural gas supply in Canada is predicted to
to markets in Canada and the Lower 48 states. decline by roughly 35% from 2005 to 2020, while
Building a dedicated pipeline that would transport the production of unconventional/stranded gas is
Prudhoe Bay natural gas straight to the Northern expected to increase dramatically by 2012 (CAPP,
Midwest pipeline network. 2006a). This assumes in part the construction of the
Mackenzie pipeline to get arctic gas to the south as
The Southern Route, also known as the Alaska well as an expectation that CBM will be economic
Natural Gas Transportation System or as the Foothills, to produce within the next two decades.
would also start from Prudhoe Bay, but it would go
25
south half-way the length of Alaska, just south of
Fairbanks, and then cross into the Yukon and north 20
eastern British Columbia.
Production, Bcf/D
15
The All-Alaska route, also known as the Yukon
Pacific LNG Proposal, would start from Prudhoe Bay, run 10
for 805 miles parallel to the Trans-Alaska oil pipeline to
5
Valdez and then turn to the east to Anderson Bay. Conventional Gas
Nova Scotia
Coalbed Methane
Mackenzie Delta
A final (but tremendously important) part of the 0
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
All-Alaska proposal would be the construction of a
liquefaction and shipment plant in the Anderson Bay, to
Figure 1-9, Canadian natural gas production forecast
enable shipping as LNG the natural gas coming from the (CAPP, 2006a)
North Slope to Asian markets (Japan mainly) and potential
terminals along the Canadian and US West Coast. The amount of gas Canada will have left over
to export to the US remains in question, and this
1-4.3 Alaska in the Context of the United States and is what may push the building of the North Slope
Canadian Natural Gas pipeline. The first issue is that Canadian natural gas
consumption is expected to increase by 1.6% per year.
The current situation of the oil and gas industry in This equates to a demand of almost 12 Bcf per day by
Canada adds substantial reasons for considering the 2020 (Stringham, 2006.). However, this consumption
over-the-top Northern Route (the green line on Fig. does not include the gas that will be needed to produce
1-8) the most suitable option for the whole North- the Canadian tar sands.
American continent. That Canada expects to be producing about 4
Canada has been a net exporter of natural gas million barrels a day by 2020 (CAPP, 2006b, Fig. 1-
for many years, and all of that exported gas has been 10) means more of Canadas natural gas will be used
imported into the United States. This gas comprises for this purpose. In fact, the 0.5 Mcf of gas needed to
about 90% of the natural gas imported into the process each barrel of this crude equates to at least 2
US and about 17% of the total US natural gas Bcf per day natural gas needed to meet the production
consumption. Although this relationship has been forecast for Canadas oil sands.
11
Modern Fracturing
5000
4500 1-5 Qatar Natural Gas
4000
3500
Production, Mbbl/D
Qatar is a small, independent nation on the western

3000
2500
coast of the Persian Gulf. The country has good
2000 relationships with its Middle Eastern neighbors
1500 like Iran, and it has been leading the region
1000
in democratic reforms.
500
0
Before the discovery of its vast hydrocarbon
reserves, dominated by natural gas, Qatar was a poor
01
03
05
07
20 9
10
12
14
16
18
20
0
20
20
20
20
20
20
20
20
20
20
Conventional Oil Sands country. However, by 2006 Qatar had achieved one of
the worlds highest per capita gross domestic products
Figure 1-10 Prediction of Canadian heavy oil sands
growth (CAPP, 2006b) (Central Intelligence Agency, 2006).
Figure 1-11 shows that compared to its neighbors
This, of course, causes some concern because the in the Middle East, Qatar is a leader in natural gas
total natural gas production from Canada in 2020 is reserves. Iran and Qatar have comparable amounts of
expected to be about 18 Bcf per day, and Canada will gas reserves. This is because Qatars super giant North
be using 14 Bcf per day for its needs. This leaves 4 Bcf Field and Irans super giant South Pars Field overlie on
per day suitable to be exported to the US. However, the broad Qatar arch. The Qatar arch subdivides the
the demand in the US over the next several years is far Khuff formations into two basins located northwest
greater than what Canada can provide. (North Field) and south east (South Pars). The North
The over-the-top Northern Route is surely not the Field reservoir boundary is the political boundary
ultimate solution to the constantly growing hunger for between Iranian and Qatari waters as shown in Fig.
natural gas in North America. The over-the-top pipeline 1-12. (Note: The names of the fields in Fig. 1-12, at
may never be built because of competition from LNG times cause confusion. Qatars North Field is north of
imports, which are expected to boom in the next several Qatar but south of the Iranian demarcation boundary.
years if additional terminals can be built. The Iranian field known as South Pars is actually
Our assessment of the Alaskan gas resources, and in southern Iranian waters but north of Qatars
in particular the North Slope basin, indicates some North Field. The two fields constitute essentially
opportunities to develop a sustained market for natural a single geological structure, one of the largest gas
gas with the U.S. Lower 48 states and Eastern Asian accumulations in the world.)
destinations (mainly Japan, South Korea and Taiwan)
via LNG shipments. This motivates all the projects 1000
proposed by several groups of advocates for transporting 900
the natural gas produced in the North Slope into the 800
Lower 48 states market, as well as Eastern Asia.
Proven Reserves, Tcf
700
Nevertheless, a wide set of reasons leads us to believe 600
that these projects cannot even be considered marginally 500
competitive to LNG, especially when compared to the 400
economically superior LNG shipped from the recently 300
developed fields and facilities in countries such as Qatar, 200
Russia, Australia and Indonesia. In fact, as is usual for 100
large construction projects, the technical feasibility of 0

ria
en
an
it
ia
n
tar
yp
wa
Ira
Ira
UA
North Slope natural gas exploitation must be weighed

rab
Yem
Om
Sy
Qa
Eg
Ku
iA
against the inexorable balance of the economics. This

ud
Sa
is the bottleneck where all the advocated Alaskan gas

Figure 1-11 Dominant natural gas producers in the
pipeline schemes become difficult to justify. Middle East (after EIA, 2006)
12
QatarGas to deliver 15 Mta of LNG to the UK market.

A year later, in June 2005, Shell signed a SPA for 7.8
Mta of LNG for Europe and North America. The
South Pars contracts for LNG have been progressively getting
bigger and bigger since the first SPA with Japan.
North Field RasGas was founded in 1993. In 1995, an SPA
with KOGAS, a Korean company, was agreed upon.
Al Manamah Two years later the SPA was increased to 4.9 Mta, and
in April of 1999 the first LNG cargo left for Korea.
The delivery time of LNG to KOGAS was four years,
Doha Persian Gulf
like the 4-year delivery time between QatarGas and
Chubu Electric. Also, an SPA with Petronet of India
was signed to deliver 5 Mta of LNG. The delivery
time for this order was five years, and the first LNG
Figure 1-12 The North Field extends off the coast of
Qatar and is divided from Irans South Pars Field by a cargo left for India in 2004. RasGas also signed a 25-
political boundary year SPA for 3.5 Mta of LNG with Edison Gas of
the United States. The SPA agreement was altered to
1-5.1 North Field Characteristics and Development increase the LNG volume to 4.6 Mta in 2003. RasGas
signed an agreement with ExxonMobil to deliver 15.6
The North Field is the largest non-associated gas Mta of LNG to the United States. In February of
field in the world with estimated reserves of 900 2005, an SPA with Distrigas of Belgium was signed to
Tcf of gas. Al-Siddiqi and Dawe (1999) explain that deliver 2.07 Mta of LNG (EIA, 2007).
the North Field produces from four intervals in the It is interesting to note the disparity in development
Khuff formation. These zones are Permian dolomite between Qatar and Iran. Qatar and Iran have comparable
carbonates located at depths of 10,000 to 13,000 ft gas reserves. Despite its sizeable gas reserves, Iran remains
with thickness ranging from 1,300 to 2,000 ft. The a net importer of natural gas. According to Wood et al.
gas produced is rich in condensates. (2006), Irans surging internal demand for natural gas and
Given the tremendous size of natural gas reserves, stiff gas market competition from Russia and Azerbaijan
major investments for the production and transportation will present Iranian leadership with difficult hurdles to
of natural gas have followed. overcome in order to externally market those reserves.
QatarGas was founded 13 years after the North While Iran is relatively isolated politically, Qatar has
Field was discovered. Eight years later, in 1992, the first been busy forging relationships with the major natural
customer, Chubu Electric of Japan, signed a sales and gas consumers such as Japan, the United Kingdom, and
purchase agreement (SPA) with QatarGas for 4 million the United States. The Qatari civil reforms, natural gas
metric tons per year (Mta) of LNG. Two years later, resource development, and good political relationships
Chubu Electric and other buyers signed a second SPA have culminated in its enormous success.
for 2 Mta of LNG. Two years later, in January 1997,
the first LNG ship delivered gas to Japan. Efficient 1-6 Fracturing for the Efficient use of
production, processing, refrigeration, storing, loading Existing Resources and for Increasing
and shipping processes for LNG established by Recovery Factor
QatarGas have allowed it to deliver 100 loads of LNG
to Japan every year since 1997 (EIA, 2007). Since its advent in the 1950s, hydraulic fracturing has
In October of 2002, BP signed an SPA with proven to be a very robust technology, lending itself
QatarGas for 0.75 Mta of LNG to deliver to Spain. to many different types of reservoirs. Additionally,
To exploit the tremendous demand for natural gas although fracturing is a very complex process, it
in Europe, ExxonMobil signed an agreement with remains for the most part extremely forgiving of the
13
Modern Fracturing
industrys overall general lack of expertise. These two One reason for this is the relative size, immaturity
factors have led to fracturing becoming the most widely and prolific productivity of the gas reservoirs outside
used completion process. North America (see earlier discussions in this Chapter).
Fracturing has its roots firmly planted in the gas Another reason is that the USA is the only country in the
production industry. Even with the widespread use of world where the landowners often own the mineral rights
fracturing for oil and injection wells, gas well fracturing under their land. In every other country, the government
is still the largest sector of the industry, by a wide margin controls the mineral resources and decides how they are
(see Fig. 1-13). The majority of gas reserves in North exploited. Consequently, in the US there is often a very
America are only produced as a result of hydraulic fragmented approach to the depletion of a reservoir,
fracturing. However, apart from a few specific locations habitually concentrating on wellbore tactics, whilst
(such as China, Argentina, Australia and to a lesser elsewhere gas companies are more inclined towards the
extent Russia), the global gas industry has failed to big picture, allowing more focus on field development
embrace this technology to even a fraction of the extent strategies. Canada sits somewhere in the middle, having
it is used in North America (see Fig. 1-14). inherited the British system of Crown ownership of all
mineral rights, while at the same time being heavily
influenced by the activities of the US gas industry. In
any case, small operators, eager to maximize short-term
Unconventional Gas cash flow, have always been the driving force behind the
28%
popularity of fracturing in the US.
Outside the US, Canada, China, Argentina and
possibly Russia, fracturing has failed to reach the critical
Tight Gas
mass that has allowed the easy exploitation of its potential
Oil
42% 25% in these countries. Operating companies often complain
that service companies do not have the infrastructure
and expertise necessary for the cost-effective execution of
Other fracturing operations in a specific geographic area. At the
5%
same time, service companies complain that operators do
not provide enough work to economically justify building
Figure 1-13 Targets of Fracture Treatments Performed in up suitable equipment and personnel resources. This is a
the USA in 2006 (BJ Services, 2006)
Catch-22 situation that can only be overcome by a)
field development projects that are large enough to justify
the introduction of a complete fracturing operation,
and b) having an operating company (or companies)
Canada with sufficient confidence in the fracturing process to
17%
proceed with fracturing-dependent field development.
Outside the above-mentioned countries, there are very
Rest of the World
(excl. China) 13%
few companies with sufficient institutional confidence
USA
70% in the fracturing process to make this happen. Even
companies based in North America with considerable
experience in fracturing seem to be unable to translate
this confidence internationally.
However, confidence in the fracturing process is
required if many countries and companies are to fully
Figure 1-14 Estimated Proportion of Fracturing exploit their gas resources. It is hoped that the processes
Treatments Performed in the USA and Canada,
compared to the Rest of the World, excluding China and experiences described in this book will help
(BJ Services, 2007) significantly with this process.
14
Ultimately, producing hydrocarbons from a reservoir far further into the reservoir, providing much greater
comes down to efficient management of the pressure in depletion at the drainage perimeter. This effect can
the reservoir. Pressure, which is stored energy (or more be maximized if the fracture azimuth is known. Wells
accurately, energy per unit volume), lies at the heart of can be drilled further apart in the direction of fracture
everything we do. The basic principle of hydrocarbon propagation and closer together in the perpendicular
production is the fact that liquids and gases will move direction, allowing maximum depletion of resources.
from a region of high energy (or pressure) to a region of Such a strategy significantly reduces the localized
low energy, if a flow path exists. When we drill a well, or pin-point depletion caused by the wellbores
we are creating a region of low pressure at the wellbore, and spreads the effects of the depletion much more
and the conductive path is provided by the formations evenly across the reservoir.
permeability. If we are lucky, there is sufficient energy left Finally, it must be remembered that although
in the liquids and gases to reach the surface, once they fracturing can be very effectively used to redevelop a
have arrived at the wellbore. In many cases, however, extra mature field (see Chapter 13), it reaches maximum
energy has to be supplied via pumps or gas lift systems, effectiveness when applied to a new reservoir:
in order to achieve flow to the surface. Ultimately, the
efficient production of a reservoir is all about getting the 1. After the fracture azimuth has been obtained,
maximum amount of oil and gas out, while using the the placing of wells can be planned to allow for
minimum energy to do so. increased drainage efficiency in the direction of
In gas reservoirs, it is difficult to provide extra energy fracture propagation. This could easily result in the
after the gas reaches the wellbore. Although the density of need for fewer wells.
the gas means that far less energy is required to reach the 2. Wellbores can be planned to facilitate fracturing.
surface, often there is insufficient energy to produce the As discussed in Chapter 5, the wellbore can be
gas at sufficient rates. completed in such a fashion as to make fracturing
In its most basic form, fracturing can be thought of easy and reliable (whereas the completion often
as a process that minimizes the energy required for the does just the opposite). In addition, perforations
gas to reach the wellbore. This has several benefits: can also be planned to maximize the effectiveness
of fracturing operations (see Chapter 6). Of all
1. It leaves more energy available for bringing the the things under our control, the perforations will
gas to the surface. have the single biggest effect on the outcome of any
2. It can reduce the minimum energy (i.e. pressure) individual treatment. Finally, multiple intervals can
required in the reservoir to achieve economic flow be more effectively and efficiently stimulated on new
to the wellbore, thereby extending production wells than on existing wells (see Chapter 9).
beyond reserve levels that might otherwise be 3. Surface facilities also can be planned to facilitate
considered depleted. In gas reservoirs, pressure fracturing, especially with regard to fluid recovery
is reserves, and so minimizing energy losses during and handling of returned proppant.
production can significantly increase the ultimate 4. Long-term relationships can be built between
recovery from the reservoir. operating companies and service providers. This
3. It minimizes secondary pressure-dependent effects allows for building and retaining experience
such as water production (and associated problems and expertise in both operational and technical
such as scale deposition, fines migration and hydrate personnel. This also improves project economics
formation), retrograde condensation within the due to efficiencies of scale and a greater ability to
reservoir, and non-Darcy flow. plan for the long term.
Fracturing effectively allows the wellbore to achieve Hydraulic fracturing of gas wells is no longer a
a significant size in comparison with the reservoir. luxury instead, it is now a necessity. For economic,
This allows the wellbores localized depletion to spread environmental and political reasons, operating
15
Modern Fracturing
companies and national operating companies have Supply, 2006a.

an obligation to maximize the recovery from their CAPP: Canadian Crude Oil Supply and Forecast
resources, while doing this as efficiently as possible. 2006-2020, 2006b.
There is no question that hydraulic fracturing will Central Intelligence Agency: Fact Book, 2006.
continue to be a major tool for achieving these goals. Energy Information Administration: Annual Energy
Fracturing will only increase in importance as reserves Outlook, 2007.
become more depleted and harder to exploit. Energy Information Administration, 2007 http://www.
Hydraulic fracturing remains an inherently complex eia.doe.gov/pub/international/iealf/table18.xls
process, and as a result is viewed with suspicion by many Energy Tribune, Various articles, February, 2007.
resources owners and asset managers. However, the Hite, D.M.: Cook Inlet Resource Potential Missing
reality is that fracturing is no more complex than any Fields Gas (and oil) Distributive/Endowment A
number of widely accepted practices, such as drilling Log-Normal Perspective, presented at the South
deviated wellbores, performing pressure transient Central Alaska Energy Forum, September 2006.
analysis, studying petrophysics and stimulating the Kornfeld, S.: Alaska North Slope Gas Task Force,
reservoir. Yet these techniques are widely practised Presentation to the US Department of Energy,
and trusted throughout the world, whereas hydraulic April 2002.
fracturing remains a largely unexploited technique Meyers, M.D.: Alaska Oil and Gas Activities,
outside of North America. presentation to The House Special Committee on
Consequently, the authors of this book hope its Oil and Gas, January 2005.
publication will have two profound effects. First, we Moscow Institute of Energy Research: Russias Natural
hope this book will help to improve the techniques Gas Future, 2006 (in Russian).
and practices employed by those who are already Stringham, G.: Canadian Natural Gas Outlook,
familiar with hydraulic fracturing. Secondly, we hope presentation by CAPP, October 2006.
this book will increase the utilization of fracturing Williams, T.E., Millheim, K., and Liddell, B.: Methane
technology in reservoirs and geographic areas that Hydrate Production from Alaskan Permafrost,
have hitherto failed to appreciate the potential of this Final Report, (March 2005).
reservoir development technique. Wood, D., Mokhatab, S., and Economides, M.J.:
Iran Stuck in Neutral, Energy Tribune
(December 2006).
Wood, D., Mokhatab, S., and Economides, M.J.:
References Global Trade in Natural Gas and LNG Expands
and Diversifies, Hydrocarbon Processing, 2007.
Alaska Oil & Gas Report, Alaska Department of www.interfax.com, 2006
Natural Resources, Div. of Oil and Gas, Anchorage, www.Gazprom.com, 2007
Alaska (May 2006).
Al-Siddiqi, A., and Dawe, R.A.: Qatars Oil and
Gasfields: A Review, Journal of Petroleum Geology
(October 1999) 22, 4, 417.
Anchorage Chamber of Commerce: Natural Gas and
Alaskas Future, 2005.
BJ Services Company: Internal Marketing Information
(2006).
BJ Services Company: Internal Marketing Information
(2007).
BP Statistical Review, 2006
Canadian Association of Petroleum Producers (CAPP):
Canadian Natural Gas, A stable Source of Energy
16
Dr. Xiuli Wang is a petroleum engineer with BP in Houston, currently functioning as a completion engineer
with worldwide responsibilities. She serves as the project leader of a major companywide project in injection
well completions and sand control. She has more than seven years of service with BP, from work as a
reservoir engineer to full-field modeling work. She supported the completion team as a petroleum engineer,
developing flux models and guidelines for minimizing erosion of producer well screens. Finally, she was the
lead production engineer for a major field in the continental shelf. Before immigrating to the United States,
Wang earned a MS degree from Chinas premier technical university, Tsinghua University, followed by six
years of work with one of Chinas major petroleum companies, Sinopec. She joined BP after earning a PhD
in chemical engineering, with a number of professional publications in the fundamentals of multi-phase
and complex flow through porous media. She was recently featured in a major journal as an exemplary
representative of Chinese-born engineers employed by the US based petroleum industry. In 2007, she was
named the US 2007 Asian American Engineer of the Year.
Chapter 2 petroleum engineers have traditionally examined oil
field hydrocarbons in the context of phase behavior,
Natural Gas Production separating the mixture into liquid and gas. Fig. 2-1
Michael J. Economides, University of Houston and shows a two-phase envelope with a pseudocritical
Xiuli Wang, BP point (C) separating the bubble-point curve (AC)
from the dew point curve (BC) at a constant
composition. Emanating from the pseudocritical
2-1 Introduction point are equal saturation quality curves (DC,
EC) inside the two-phase envelope. To the right of
The natural gas we use in everyday life - as a the pseudocritical point is the maximum possible
source of space heating after combustion, for power temperature, called the cricondentherm.
generation even as industrial feedstock - is primarily Natural gas reservoirs whose pressure and
methane. Such fluid has been stripped of higher-order temperature lie to the right of the cricondentherm
hydrocarbons. This is not how natural gas appears just are known as dry gas reservoirs. If fluids from
one or two steps before its ultimate use. these reservoirs stay outside of the two-phase
At the present time there are two main sources for envelope in traversing a pressure and temperature
natural gas as a petroleum production fluid. path from the reservoir to the wellhead,
First, gas is found in association with oil. Almost they will produce only dry gas.
all oil reservoirs, even those that in-situ are above their If the path from reservoir to surface carries
bubble point pressure, will shed some natural gas, which the fluid into the two-phase envelope below the
is produced at the surface with oil and then separated cricondentherm wet gas is produced.
in appropriate surface facilities. The relative proportions
of gas and oil produced depend on the physical and
thermodynamic properties of the specific crude oil
system, the operating pressure downhole, and the
pressure and temperature of the surface separators.
The second type of gas is produced from reservoirs
that contain primarily gas. Usually such reservoirs are
considerably deeper and hotter than oil reservoirs. We
will deal with the production characteristics of these
reservoirs in this chapter.
There are other sources of natural gas, one of which
(coalbed methane desorbed from coal formations) is Figure 2-1 Phase diagram showing regions of retrograde
condensate
already in commercial use. This process is described
in relative detail in Chapter 11 of this book. In the far Between the critical point and the cricondentherm,
future, production from massive deposits of natural liquid emerges as the pressure declines below the dew
gas hydrates is likely, but such eventuality is outside point value (at a constant temperature) from point 1 to
the scope of this book. point 2, shown in Fig. 2-1. As pressure decreases from
point 2 to point 3, the amount of liquid in the reservoir
2-2 Idiosyncrasies of Dry Gas, Wet Gas increases. Further pressure reduction causes liquid to re-
and Gas Condensates vaporize. This is the region of retrograde condensation
(McCain, 1973). Many natural gas reservoirs behave in
Petroleum fluids found in nature, are always multi- this manner. During production from such reservoirs, the
component mixtures of hydrocarbons. Characterizing pressure gradient formed between the reservoir pressure
these fluids is difficult both from a scientific/laboratory and the flowing bottomhole pressure may result in liquid
point of view and in production operations. Thus, condensation near the wellbore (Wang, 2000).
19
Modern Fracturing
One way to prevent condensate formation is

dp g
to maintain the flowing well bottomhole pressure = vg + g g vg 2 , (2-2)
dx k g
above the dew point pressure. This is often not
satisfactory because the reservoir pressure drop may where g is the gas density and g is the effective non-
not be sufficient to achieve economic production Darcy coefficient to gas. The condensate liquid may
rate. An alternative is to allow condensate to form flow if its saturation is above the critical condensate
but occasionally to inject methane gas into the saturation, Scc (Wang and Mohanty, 1999a). Additional
producing well. The gas dissolves and sweeps the condensate dropout because of the further reduced
condensate into the reservoir. The well is then put pressure will aggravate the situation. Therefore,
back in production. This approach is repeated several two phenomena emerge Non-Darcy effects and a
times in the life of the well. It is known as gas cycling substantial reduction in the relative permeability to
(Sanger and Hagoort, 1998). gas. Because of the radial nature of flow, the near well
bore region is critical to the productivity of a well.
2-3 Inflow from Natural Gas Reservoirs This is true in all wells, but it becomes particularly
serious in gas-condensate reservoirs.
2-3.1 Fundamentals of Non-Darcy Flow Forchheimers equation describes high-velocity,
in Porous Media single-phase flow in isotropic media. Many naturally-
occurring porous media are, however, anisotropic
Fluid flow is affected by the competing inertial and (Wang et al., 1999). A direct understanding of
viscous effects, combined by the well-known Reynolds multiphase non-Darcy flow behavior in porous media
number whose value delineates laminar from turbulent that are anisotropic at the pore-scale is studied elsewhere
flow. In porous media the limiting Reynolds number is (Wang, 2000, Wang and Mohanty, 1999b).
equal to 1 based on the average grain diameter (Wang
and Economides, 2004). 2-3.2 Transient Flow
Because permeability and grain diameter are
well connected (Yao and Holditch, 1993), for small To characterize gas flow in a reservoir under transient
permeability values (e.g., less than 0.1 md) the production conditions, the combination of the generalized Darcys
rate is generally small, flow is laminar near the crucial law (rate equation) and the continuity equation
sandface and it is controlled by Darcys Law: can be used. Thus:
dp g k
= vg , (2-1) = p , (2-3)
dx k g t
where x represents the distance, p the pressure, vg where is porosity, and in radial coordinates:
the gas velocity, g the gas viscosity, kg the effective
1 k p
permeability to gas. A small amount of connate water = r . (2-4)
t r r r
is almost always present besides the gas. The water
saturation is often small and it does not affect the gas Because gas density is a strong function of pressure (in
permeability significantly. Therefore, kg is often equal contrast to oil, which is considered incompressible),
to k, the single-phase permeability. the real gas law can be employed:
Non-Darcy flow occurs in the near-wellbore
m pM
region of high-capacity gas and condensate reservoirs = = , (2-5)
V ZRT
as the flow area is reduced substantially, the velocity
increases, inertial effect becomes important, and the and therefore
gas flow becomes non-Darcy. The relation between
p 1 k p
pressure gradient and velocity can be described by the
= rp . (2-6)

t Z r r Z r
Forchheimer (1914) equation:
20
Chapter 2 Natural Gas Production
In an isotropic reservoir with constant permeability, The solution of Eq. 2-13 would look exactly
Eq. 2-6 can be simplified to: like the solution for the diffusivity equation
cast in terms of pressure. Dimensionless time is
p 1 p p
(2-7) (in oilfield units):
= r .

k t Z r r Z r
0.000264kt
tD =
2
, (2-14)
( ct )i rw
Performing the differentiation on the right-hand
side of Eq. 2-7 - assuming that the viscosity and gas and dimensionless pressure is
deviation factor are a small functions of pressure -
kh[m( pi ) m( pwf )]
and rearranging gives:
pD = . (2-15)
1424qT
p 2 2 p 2 1 p 2
(2-8) Equations 2-13 to 2-15 suggest solutions to
= 2
+ .
kp t r r r natural gas problems (e.g., well testing) that are
For an ideal gas, cg = 1/p and, as a result, Eq. 2-8 leads exactly analogous to those for an oil well, except
to: now it is the real gas pseudopressure functions
that needs to be employed. This function is
2 p 2 1 p 2 c p 2
(2-9) essentially a physical property of natural gas,
+ = .
r 2 r r k t dependent on viscosity and the gas deviation
This approximation looks exactly like the classic function. Thus, it can be readily calculated for
diffusivity equation for oil. The solution would look any pressure and temperature by using standard
exactly like the solution of the equation for oil, but physical property correlations.
instead of p, the pressure squared, p2, should be used, By analogy with oil, transient rate solution under
as a reasonable approximation. radial infinite acting conditions can be written as:
Al-Hussainy and Ramey (1966) used a far more
appropriate and exact solution by employing the real kh[m( pi ) m( pwf )]
q=
gas pseudo-pressure function, defined as: 1638 T
1
k
p p log t + log 3.23 + 0.87 s ,

m( p ) = 2
dp, (2-10)
2
( ct )i rw (2-16)
po Z
where po is some arbitrary reference pressure (usually where q is gas flow rate in Mscf/d, pi is reservoir
zero). The differential pseudo-pressure, m(p), pressure, pwf is the flowing bottomhole pressure, is
defined as m( p) m( pwf ), is then the driving force porosity, ct is the total compressibility of the system,
in the reservoir. and s is the skin effect.
Equation 2-16 can be used to generate transient
Using Eq. 2-10 and the chain rule: IPR (Inflow Performance Relationship) curves
for a gas well.
m( p ) m( p ) p
= . (2-11)
t p t
2-3.3 Steady State and Pseudosteady State Flow
Similarly,
Starting with the well known Darcys law equation
m( p ) 2 p p
= . (2-12) for oil inflow,
r Z r
kh( pe pw )
Therefore, Eq. 2-9 becomes q= ,
(2-17)
r
141.2 B[ln( e ) + s ]
rw
2 m( p ) 1 m( p ) ct m( p ) (2-13)
+ = .
r 2 r r k t and recognizing that the formation volume factor,
21
Modern Fracturing
B, varies greatly with pressure, then an average Similarly, the same coefficient can be employed
expression can be used as shown by Economides to the more rigorous expression using the real-gas
et al. (1994): pseudopressure. As an example, for pseudo-steady
state with q in Mscf/d:
0.0283ZT
Bg = . (2-18)
( p e + p wf )/ 2
kh[m( p ) m( pwf )]
q= .
With relatively simple algebra, and introducing the 1424T [ln(0.472re / rw ) + s + Dq ] (2-24)
gas rate in Mscf/d, Eq. 2-17 and 2-18 yield:
2-3.4 Horizontal Well Flow
141.2(1000 / 5.615)q (0.0283) Z T
pe pwf =
[( pe + pwf ) / 2]kh Analogs to Eq. 2-23 (for steady state) and 2-24 (for
re (2-19) pseudo-steady state) can be written for a horizontal
[ln( ) + s ],
rw well. Allowing for turbulence effects, the inflow
performance relationships for a horizontal well in a
and, finally: gas reservoir are for the steady state:
1424q ZT r
pe2 pwf
2
= [ln( e ) + s ], (2-20) k H h( pe2 pwf 2
)
kh rw q= ,

I ani h I ani h

1424ZT Aa +
+ Dq (2-25)

which re-arranged provides the steady-state ln
L
rw ( I ani + 1)

approximation for natural gas flow, again showing a where
pressure squared difference dependency.
A similar expression can be written for pseudo- a + a 2 ( L / 2) 2

steady state: Aa = ln ,

L/2

22

kh( p pwf )
q= . (2-21) and for pseudo-steady state:
0.472re
1424ZT [ln( ) + s]
rw k H h( p 2 pwf2
)
q= ,

I ani h I ani h 3

All expressions given thus far in this chapter have
1424ZT Aa +
+ Dq
ln
ignored one of the most important effects in natural L rw ( I ani + 1) 4

gas flow: turbulence. (2-26)
One of the simplest and most common
ways to account for turbulence effects is where Iani is a measurement of vertical-to-horizontal
through the use of the turbulence coefficient, permeability anisotropy given by:
D, which is employed by adding a component
kH
to the pressure drop, as shown below for the I ani = . (2-27)
kV
steady-state equation:
In Eqs. 2-25 and 2-26, a is the large half-axis of the
2 1424ZT
2 r drainage ellipsoid formed by a horizontal well of
p p =
e wf [ln( e ) + s ]q
kh rw length L. The expression for this ellipsoid is
1424ZTD 2 (2-22)
+ q , 0.5 0.5

L
kh r
4

a=
0.5 + 0.25 + eH

2
L / 2

which rearranged, provides the well-known:
L
for < 0.9reH , (2-28)
kh( pe2 pwf
2
) 2
q= .
re (2-23) where reH is the equivalent radial flow drainage radius.
1424ZT [ln( ) + s + Dq ]
rw
22
2-4 Effects of Turbulence For an isotropic formation, k equals the horizontal

permeability. For an anisotropic formation, k is defined
The effects of turbulence have been studied as the equivalent permeability,
by a number of investigators in the petroleum
kV k 1
literature, pioneer and prominent among which keq = [1 log( )]( V ) 3 k H , (2-31)
kH kH
have been Katz and co-workers (Katz et al., 1959;
Firoozabadi and Katz, 1979; Tek et al., 1962). In where kV is the vertical and kH the horizontal
their work they suggested that turbulence plays permeability.
a considerable role in well performance showing To demonstrate the effects of turbulence on
that the production rate is affected by itself: The natural gas production, a number of calculations are
larger the potential rate, the larger the relative shown here, using the Katz et al. (1959) approach for a
detrimental effect would be. One interesting means range of permeabilities. Table 2-1 contains the well and
to account for turbulence was proposed by Swift reservoir data; Table 2-2 presents the results.
and Kiel (1962), who presented Eq. 2-22, which
when rearranged gives Eq. 2-23. Table 2-1. Well and Reservoir Characteristics
Equation 2-23 is significant because it suggests pe 3000 psi Case 1 Case 2
that turbulence effects can be accounted for by a rate-
re 660 ft pwf 1500 psi 2500 psi
dependent skin effect, where the turbulence (at times
referred to as the non-Darcy) coefficient, D, has the rw 0.359 ft 0.0162 cp 0.0186
units of reciprocal rate. One of the implications is h 50 ft Z 0.91 0.9

that in testing a high-rate gas well, a calculated skin T 710R
effect must be construed as apparent, rather than the
g 0.7
real damage skin. Among the procedures suggested
for testing test gas wells are multi-rate testing with
subsequent determination of apparent skins at Table 2-2 Turbulence Effect at Different Permeabilities
and Different Drawdowns
each rate, and straight-line construction graphing
of s+Dq vs q. The graph allows field determination Case 1: p = 1500 psi
k,
of s, the skin not affected by turbulence, from md q (=0, s=0) q (>0, s=0) q (>0, s<0)
s
MMscf/d MMscf/d MMscf/d
the vertical axis intercept, and D from the slope
1 3.0 2.9 8.1 -5.7
(Economides et al., 1994). 5 15.1 13.0 24.6 -5.1
25 75.3 51.9 71.7 -4.3
2-4.1 The Effects of Turbulence on Radial Flow 100 301.2 151.2 179.1 -3.7
Case 2: p = 500 psi

k,
Katz et al. (1959) have presented an explicit md q (=0, s=0) q (>0, s=0) q (>0, s<0)
s
relationship for the radial flow of gas into a well, using MMscf/d MMscf/d MMscf/d
natural gas properties and by providing correlations 1 1.1 1.1 3.7 -5.7
5 5.4 5.1 12.2 -5.1
for the coefficient, :
25 27.0 23.0 37.9 -4.3
1424ZT r 100 108.1 75.5 100.1 -3.7
pe2 pwf
2
= [ln( e ) + s ]q
kh rw
The first two columns of Table 2-2 show
1 1
3.16(10)12 g ZT ( ) the expected production rates for two flowing
rw re
+ q2 , (2-29) bottomhole pressures, for laminar and turbulent
h2
where conditions, and for permeabilities from 1 to 100 md.
At low permeability, as expected, the rate reduction is
2.33(10)10
= . (2-30) negligible; however, at 100 md and pwf = 1500 psi the
k 1.201
reduction is almost 50%.
23
Modern Fracturing
Turbulence effects, viewed as an apparent skin, result the negative skin (-3.7) and the zero skin is far less (1.2),
in values of 0.3, 1.2, 3.4 and 7.5, for the 1-, 5-, 25- and but again, these production ratios do not paint the true
100-md cases, respectively (and pwf = 1500 psi). effect of fracturing, which will be addressed later.
Because the range of 5 to 25 md is perhaps the
most likely to be encountered in emerging natural gas 2-4.2 Perforated and Cased Well
fields, the ratio of actual to ideal (without considering in a High-Rate Gas Reservoir
turbulence) rates is perhaps the most telling. For the two
different drawdowns, these ratios are 0.86 and 0.94 (for The previous section deals with the flow reduction in
the 5-md case) and 0.68 and 0.85 (for the 25-md case). an open-hole well and could also be considered as a
These results, plotted on Fig. 2-2, show the effect of both reasonable approximation for a slotted liner.
the permeability value and the drawdown. The ratio For a cemented and perforated well, in the absence of
between turbulence-affected production and production turbulence, a configuration skin effect can be envisioned
calculated under the assumption of laminar flow declines and added to the denominator of the deliverability
precipitously as reservoir permeability and drawdown relations. Karakas and Tariq (1988) have published a
(and, hence, production rate) increase. method to calculate this skin effect which depends on
the length of the perforation tunnel, the perforation
1.0
diameter, the phasing (degrees among adjoining planes of
0.9
%p = 500 psi perforations) and, especially, the perforation density, i.e.,
0.8 how many perforations per unit net thickness, measured
q actual / q ideal
in shots per foot (SPF). They also quantified the effect

0.7
of vertical-to-horizontal permeability anisotropy: The
0.6
lower the vertical permeability, the larger the value of
%p = 1500 psi
0.5 the skin effect would be. Finally, they showed that if the
0.4
perforation tunnel lengths end outside a damage zone,
1 10 100 rather than inside, the composite damage/perforation
Permeability, md
skin effect is substantially reduced.
Figure 2-2 Turbulence effects for a permeability range
and different drawdowns Using the Karakas and Tariq (1988) model, one
important conclusion is that in a permeability-isotropic
An additional interesting issue is the question formation without near-well damage, 4 SPF of typical
of the negative skin effect. This will become even tunnel length and diameter result in a perforation skin
more pronounced in a later section of this chapter effect equal to zero; i.e., this configuration may be
presenting the expected production rate from construed as open-hole equivalent.
hydraulic fracturing. For now, a hypothetical negative Ichara (1987) used a similar approach, constructing
skin effect is used to represent, for example, matrix a numerical model for a perforated natural gas well
acidizing of carbonate rock. In Table 2-2, the listed and accounting for turbulence effects. He showed that
negative skin effects result from fracturing. What perforations add a production impediment because of the
will become apparent is that hydraulic fracturing in increase in turbulence. Fig. 2-3 presents some of Icharas
natural gas wells has a much larger effect than merely results, which show the effect of permeability anisotropy
imposing a negative skin because of the extraordinary and perforation tunnel length. One observation is that
reduction in turbulence effects. long perforations are useful, making a well with reasonable
When turbulence effects are insignificant, the perforation density (4 SPF) near the performance of an
negative skin effect is very large. In the 1-md case with open-hole well (still affected by turbulence). From Fig.
1500-psi flowing bottomhole pressure, the production 2-3 it can be concluded that a gas well with 4 SPF and
ratio between the negative skin (-5.7) and the zero skin a typical 8-in. tunnel length in a sandstone reservoir
is nearly 3. Conversely, when the turbulence effects are (kv /kh = 0.1) will perform at about 85% of an open-hole
great, as in the 100-md case, the production ratio between well. Fig. 2-3 is for 0 perforation phasing.
24
1.2
2-5 Production from Hydraulically
1.1
Open Hole
Fractured Gas Wells
Productivity Ratio
1.0
Hydraulic fracturing has been established as the
0.9
kv / kh
premier production enhancement procedure in the
4 SPF
0.8 1.0 p = 1500 psi petroleum industry. For the first 40 years since its
0.1
0.7
0.01 0 Phasing Angle inception, hydraulic fracturing has been for primarily
low-permeability reservoirs; in the last two decades
0.6
it has expanded into medium- to high-permeability
0 3 6 9 12 15 18 21
formations through the tip screenout (TSO) process
Perforation Length, in
(see section 4-7.3.2). For natural gas wells, a reservoir
Figure 2-3 Productivity ratio vs. perforation length for
kV /kH = 0.01, 0.1, and 1.0 (after Ichara, 1987) above 0.5 md should be considered a medium-
permeability reservoir. Above 5 md it should be
Figure 2-4 (for an isotropic formation) suggests considered a high-permeability formation. In all
that improving the perforation phasing to 90 and, high-permeability cases, the fracture should be a TSO
especially, increasing the perforation density to 8 or treatment (Economides et al., 2002a). Even in many
12 SPF may render the cased and perforated well medium-permeability formations with relatively small
an even better performer than an open hole. This elastic moduli, TSO is the indicated method.
is because of the penetration of the flow channels Valk and Economides and co-workers as in Romero
beyond the sand face. In short, high perforation et al. (2002) introduced a physical optimization technique
density of long-penetrating tunnels will reduce to maximize the productivity index of a hydraulically
turbulence effects. For example, from Fig. 2-4, fractured well. They call it the Unified Fracture Design
for 8 SPF of 18 in. perforation tunnels, the well (UFD, Economides et al., 2002a) approach. They
performance would be about 10% larger than that introduced the concept of the dimensionless Proppant
of an open-hole well. Number, Nprop, given by:
(Note: Icharas work assumes that all perforations
are open and undamaged. This is of course rarely true 4k f x f w 4k f x f whp
N prop = I x2C fD = 2
=
and the results presented here should be considered kx
e kxe2 hp
an upper limit. Turbulence effects would be enhanced 2k f V p (2-32)
with damaged or partly open perforations.) = ,
kVr
1.2
where Ix is the penetration ratio, CfD is the
SPF
1.1 12 dimensionless fracture conductivity, Vr is the reservoir
8
Open Hole drainage volume, Vp is the volume of the proppant in
4
1.0 the pay (the total volume injected times the ratio of the
Productivity Ratio
12
90 8 net height to the fracture height), kf is the proppant
4
0.9
Phasing Angle
pack permeability, k is the reservoir permeability,
xe is the well drainage dimension, hf is the fracture
0.8
0 height and h is the reservoir thickness. The proppant
permeability for gas wells will have to be adjusted
0.7
because of turbulence effects. This adjustment will be
shown in a later section.
0.6
0 3 6 9 12 15 18 21
Valk and Economides also found that for a
given value of Nprop there is an optimal dimensionless
Perforation Length, in
fracture conductivity at which the productivity index
Figure 2-4 Effect of shot density and phasing angle on is maximized.
productivity ratio for kV /kH = 1
25
Modern Fracturing
2-5.1 Unique Needs of Fracture 2-5.2 Turbulence Remediation

Geometry and Conductivity in High- and Low-Permeability Wells
At low Proppant Numbers, the optimal conductivity, In the case of a potentially high-rate natural gas well, the
CfD = 1.6. The absolute maximum dimensionless effective proppant pack permeability used to calculate
productivity index (see Section 4-5.1), JD , is 6/ = 1.909 the Proppant Number and the dimensionless fracture
(the productivity index for a perfect linear flow in a conductivity depends on the production rate because of
square reservoir). When the propped volume increases the non-Darcy flow effects.
or the reservoir permeability decreases, the optimum Economides et al. (2002b) presented an iterative
dimensionless fracture conductivity increases somewhat. procedure combining the UFD method with the Gidley
Valk and Economides (1996) also presented (1990) adjustment to proppant pack permeability and
correlations for the maximum achievable dimensionless the Cooke (1973) correlations for flow in fractures.
productivity index as a function of the Proppant The procedure starts with correcting the effective
Number (see Eq. 2-33 at the bottom of the page). permeability using the in-situ Reynolds number by:
Similarly, correlations were presented for
k f ,n
the optimal dimensionless fracture conductivity k f ,e =
, (2-36)
1 + N Re
for the entire range of Proppant Numbers (see
Eq. 2-34, at bottom). where kf,n is the nominal fracture permeability.
After the optimal dimensionless fracture conductivity First a Reynolds number is assumed. A good first
is known, the optimal fracture length and width can be value is Reynolds number equal to zero. Then from
readily determined from: Eq. 2-32 and the adjusted proppant pack permeability
the Proppant Number is calculated from which the
0.5 0.5
k f V f C kV maximum JD (Eq. 2-33) and optimum dimensionless
x fopt = and wopt = fD ,opt f

.
(2-35) conductivity (Eq. 2-34) are calculated. The latter
C fD ,opt kh k f h
allows the determination of the indicated fracture
where Vf is the volume of one propped wing, dimensions using Eq. 2-35.
Vf = Vp/2. From the dimensionless productivity index and
The idea of using the maximized dimensionless drawdown, the actual production rate is calculated,
productivity index to 1) design a hydraulic fracture which in turn is used to obtain the Reynolds number.
treatment and 2) to evaluate the subsequent well The procedure ends when the assumed and calculated
performance against a benchmark and indeed any Reynolds numbers are close enough.
other well configuration allows a generalized approach The Reynolds number for non-Darcy flow
to production engineering. It becomes important to is given by
rationalize sub-standard performance and a constant k f ,n
effort to push the limits (Economides et al., 2001.) N Re = , (2-37)

1
if N prop 0.1
0.990 0.5 ln N prop
J D max ( N prop ) = (2-33)
6 0.423 0.311N prop 0.089( N prop ) 2
exp 2

if N prop > 0.1
1 + 0 . 667 N prop + 0. 015( N prop )
1.6

if N prop < 0.1

0.583 + 1.48 ln N prop

C fD ,opt ( N prop ) = 1.6 + exp
if 0.1 N prop 10 (2-34)

1 + 0 .1
1 42 ln N prop

N prop

if Nprop >10
26
where kf,n is the nominal permeability (under Darcy

Table 2-4 Results from Hydraulically Fractured Well
flow conditions) in m2, is in 1/m, v is the fluid ( kf = 60,000 md)
velocity at reservoir conditions in m/s, is the
Case 1: pwf = 1500 psi
viscosity of the fluid at reservoir conditions in Pa.s k, md s
and is the density of the flowing fluid in kg/m3. The q, MMscf/d kf,e , md xf , ft
value of is obtained from: 1 -5.7 13.1 9251 218

5 -5.1 43.5 7950 91
b 25 -4.3 160.3 6670 36
= (1108) , (2-38)
(k f ,n )a 100 -3.7 524.0 5525 16
where a and b are obtained from Cooke (1973). Some Case 2: pwf = 2500 psi
k, md s
values are given in Table 2-3.
q, MMscf/d kf,e , md xf , ft
1 -5.7 5.8 12493 250
Table 2-3 Constants a and b
Prop Size a b 5 -5.1 18.9 10770 108
8 to 12 1.24 17,423 25 -4.3 69.2 8980 44
10 to 20 1.34 27,539 100 -3.7 224.0 7494 20
20 to 40 1.54 110,470
40 to 60 1.60 69,405
Table 2-5 shows even more prolific fractured wells if
The velocity, v, is determined as the volumetric flow premium proppants are used (kf = 600,000 md),
rate in the fracture near the well divided by the fracture pushing the limits of hydraulic fracturing (Demar-
height times the fracture width (both determined from chos et al., 2004).
the design in each iteration.) For a detailed approach
and example see Economides et al. (2002b). Table 2-5 Results from Hydraulically Fractured Well
Table 2-4 presents the results for the fracture ( kf = 600,000 md)
designs and expected production rates for the four Case 1: pwf = 1500 psi
k, md s
permeabilities used earlier for the non-fractured wells q, MMscf/d kf,e , md xf , ft
presented in Table 2-2. These designs assumed sand as 1 -6.1 19.9 38300 375
a proppant with kf = 60,000 md. 5 -5.9 59.2 32050 182
There are some very important implications 25 -5.4 202.0 27110 75
in comparing the results in Tables 2-2 and 2-4. 100 -4.8 637.0 22410 35
At 5 md the non-fractured well would deliver 13
Case 2: pwf = 2500 psi
MMscf/d (with pwf = 1500 psi). If the fracture- k, md s
induced skin of -5.1 is assumed the production q, MMscf/d kf,e , md xf , ft
rate would be 24.6 MMscf/d, approximately a two- 1 -6.1 8.8 51600 456
fold increase (see Table 2-2) This production ratio 5 -5.9 26.3 44150 211
increase would be expected in an oil well flowing 25 -5.4 88.4 37020 91
under laminar conditions. However, the implicit 100 -4.8 270.0 31720 41
reduction in turbulence effects (because of the flow
profile modification in going from converging radial In summary, turbulence affects are the dominant
flow to fracture flow) leads to a considerable further features in the production of high-permeability (>5
increase in the production to (in this example) md) gas wells. Turbulence may account for a 25 to 50%
43.5 MMscf/d, a more than three-fold increase reduction in the expected open-hole production rate
(see Table 2-4). For higher-permeability wells, the from such wells, if laminar flow is assumed. Cased and
resulting folds of increase are similar, albeit in perforated wells may experience further turbulence-
actual production rates the achievable results are induced rate declines, which can be alleviated somewhat
spectacular (see Fig. 2-5). with long-penetrating perforation tunnels and large
27
Modern Fracturing
perforation densities (e.g., 8 to 12 SPF). However, Depending on the well orientation with respect to
nothing can compete with hydraulic fracturing. In the state of stress, either a longitudinal or a transverse
higher-permeability gas wells, the incremental benefits fracture may be created in a horizontal well (Soliman
greatly exceed those of comparable permeability oil wells, and Boonen, 1997; Mukherjee and Economides, 1991;
exactly because of the dramatic impact on reducing the Soliman et al., 1999). The longitudinal configuration
turbulence effects beyond the mere imposition of a is generated when the well is drilled along the expected
negative skin. It is fair to say that any gas well above fracture trajectory. The performance of such well is
5 md will be greatly handicapped if not hydraulically almost identical to a fractured vertical well when both
fractured. Indeed, pushing the limits of hydraulic have equal fracture length and conductivity. Therefore,
fracturing by using large quantities of premium existing solutions for vertical well fractures can be applied
proppants will lead to extraordinary production rate to a longitudinally fractured horizontal well (Economides
increases (Wang and Economides, 2004). et al., 2002a; Soliman et al., 1999; Villegas et al., 1996;
and Valk and Economides, 1996).
1000.0
Fractured Well
(Premium) Fractured
Well
100.0
q, MMscf/d
10.0 Negative
Skin Radial
Flow
1.0
Radial flow
1 10 100
Permeability, md
Figure 2-5 Comparison of gas production rates from non-

fractured wells, wells with negative skin and fractured wells
2-5.3 Multi-fractured Horizontal Gas Wells

Hydraulically fractured vertical well
As discussed in the previous section, in vertical gas wells
Figure 2-6 Configurations of radial flow and fractured
turbulence can be greatly reduced through hydraulic
vertical well
fracturing because the flow pattern (shown in Fig. 2-
6) through the hydraulic fracture towards the well is Almost all reported applications of fractured
different than for radial flow (Wang and Economides, horizontal wells are for transverse fractures (Crisby et al.,
2004). The same is not necessarily true for transversely 1998; Emannuele et al., 1998; Eirafie and Wattenbarger,
fractured horizontal gas wells (see Section 10-2). 1997; Minner et al., 2003; and Fisher et al., 2004). A
Because turbulence effects are enhanced in the latter transverse hydraulic fracture is created when the well
(due to the very small contact area between the well is drilled normal to the expected fracture trajectory
and the fracture), the conclusion is more nuanced. The (Valk and Economides, 1996; Soliman et al., 1999;
limited communication between the transverse fracture and Economides et al., 1994). The configuration of a
and the wellbore generates an additional pressure drop transversely fractured horizontal well is demonstrated
and a choking effect for all transversely fractured in Fig. 2-7. The cross section of the contact between a
horizontal gas wells. This also increases turbulence, transverse fracture and a horizontal well is 2 rww where
which precludes application to essentially any well w is the width of the fracture (which can be obtained by
whose permeability is 1 md or more and, perhaps, to using a design procedure such as the Unified Fracture
even much lower values of permeability, depending on Design approach) and rw is the radius of the horizontal
project economics (Wei, 2004). well. In this case, the flow from the reservoir into the
28
fracture is linear; the flow inside the fracture is converging Calculation Method and Theory
radial (Economides et al, 1994). This combination of for Transversely Fractured Gas Well
flows results in an additional pressure drop that can To study the performance of a transversely fractured
be accounted for by a choke skin effect, denoted as sc horizontal gas well, it is essential to account for
(Mukherjee and Economides, 1991). The horizontal turbulence effects, which are likely to be large
well is assumed to be in the vertical center of a reservoir because of high gas-flow velocity. Economides et
(see Fig. 2-7) and the flow is from the reservoir into the al. (2002b) have developed an iterative procedure
fracture and then from the fracture into the wellbore to account for turbulence effects in a hydraulic
(Mukherjee and Economides, 1991). The produced fluid fracture. The main steps and the correlations used
enters the wellbore only through the fracture, regardless are described below.
of whether the remaining part of the well is perforated. 1. Assume a Reynolds number, NRe , and
In this study, this assumption is also valid. calculate the effective fracture permeability kf,e
using Eq. 2-36.
2. Using kf,e , calculate the Proppant Number, Nprop
, from Eq. 2-32.
3. With Nprop , calculate the maximum productivity
index, JDmax , and optimal dimensionless fracture
conductivity, CfDopt , from Eq. 2-33 and 2-34,
respectively.
4. With CfDopt , calculate the indicated optimum
fracture dimension xfopt and wopt from Eq. 2-35.
5. With the known kf,e and wopt, calculate the choke
skin factor by:
kh h
sc = ln . (2-39)
Side view, fluid flow from reservoir to the fracture k f w 2rw 2
6. With the calculated JD,max and sc, calculate the

dimensionless productivity index of transversely
fractured horizontal oil well JDTH (neglecting
turbulence effects for now), JDTH :
1
J DTH =
1
+ s (2-40)
J c
DV
where JDV is the dimensionless productivity

index of the fractured vertical well calculated
using the procedure described by Wang and
Economides (2004).
Top view, fluid flow from the fracture to the wellbore 7. With JDTH and drawdown, the actual production
Figure 2-7 One transverse fracture intersecting a
rate can be obtained using Eq. 2-41. With this
horizontal well production rate, a new Reynolds number NRe
can be calculated with Eqs. 2-37 and 2-38, and
In the following section, the theory and calculation the flow velocity v obtained from the cross-
method for transversely fractured horizontal sectional area of flow.
gas well are described. Then some results and
kh( p 2 pwf
2
)
discussions are presented. q= J DTH . (2-41)
1424ZT
29
Modern Fracturing
8. Compare NRe calculated in Step 7 with the The details of the fracture design are omitted here.
assumed NRe in Step 1. If they are close enough, What are presented are fractured well performance
the procedure can be ended. If they are not, repeat results, summarized in Table 2-6.
from Step 1 until they are close enough. It should be noted that the skin choke effect,
The calculated results are optimum, which means sc , (from Eq. 2-39) is inversely proportional to the
that at a given Proppant Number the dimensionless proppant-pack permeability. Thus, choosing high-quality
productivity index is the maximum at the optimum proppant would decrease sc and benefit the dimensionless
dimensionless fracture conductivity (Demarchos et productivity index, JDTH (Eq. 2-40), and the Equivalent
al., 2004). However, this optimization often must Number of Vertical wells, X (Eq. 2-42).
be tempered by physical and logistical constraints
(Economides et al., 2002a) Table 2-6 Results for kf = 600,000 md
To compare the performance of fractured vertical kf = 600,000 md, 150,000 lbm mass,
and transversely fractured horizontal gas well, the single transverse fracture
Equivalent Number of Vertical Wells, X, is defined as: k, md JDV JDTH w, in. sc kf,e
1 0.739 0.121 0.35 4.64 1002

J DTH 5 0.457 0.056 0.69 13.3 871
X= . (2-42)
J DV 10 0.389 0.036 0.86 22.3 832
25 0.324 0.018 1.04 48.7 794
Assume the formation permeability is the same 50 0.288 0.013 1.48 69.1 783
throughout and n transverse fractures are generated 100 0.255 0.009 2.07 100 774
intersecting a horizontal well (Fig. 2-8). JDTHt is the total

dimensionless productivity index (sum) for n transverse The results in Table 2-6 show the value of JDTH is
fractures. JDTH1 is the dimensionless productivity very small (compared to that of the vertical well, JDV)
index of one isolated zone for a transversely fractured and decreases dramatically with increasing formation
horizontal well. Therefore: permeability. It is obvious that turbulence effects
influence the performance of a transversely fractured gas
J DTHt = nJ DTH 1. (2-43) well so much that even with the most premium proppant
(permeability 600,000 md), the results are unacceptable.
The comparison of production between a fractured
vertical well, a transversely fractured horizontal well and
laminar flow open-hole well (the ideal case in Wang and
Economides, 2004) is summarized in Fig. 2-9. The top
solid curve (qv /qideal) represents the ratio of the fractured
vertical well production to that from a laminar-flow,
open-hole vertical well. The solid bottom curve shows
Figure 2-8 Multiple transverse fractures intersecting a the ratio of a transversely fractured horizontal well (one
horizontal well fracture) with the same laminar-flow, open-hole vertical
well (qTH /qideal). Results clearly show that because the
Results and Analysis fracture in the vertical well changes the flow pattern in
for Formation Permeability from 1 to 100 md the near-wellbore area and alleviates the non-Darcy effect
A case study is presented here for the multiple fracturing the qv /qideal is considerably larger than 1. Conversely,
of a horizontal well in a gas reservoir with h = 50 ft, the qTH /qideal is much smaller than 1 even at reservoir
g = 0.7, reservoir pressure of 3000 psi and flowing permeability equal to 1 because of the choke skin and
bottomhole pressure of 1500 psi. non-Darcy effects. The dashed line in Fig. 2-9 shows that
Assume a single transverse fracture is generated in even with four transverse fractures, the productivity ratio
the horizontal well and the mass of proppant is 150,000 of a fractured horizontal well to an ideal open-hole is still
lbm. Proppant-pack permeability, kf , is 600,000 md. less than 1 for permeability larger than 10 md.
30
Would increasing the mass of proppant improve The JDTH is smaller than JDV when other parameters
the performance? The answer is no. The reason is are the same.
that the main factor that makes JDTH so low is the The JDTH decreases with increasing formation
converging skin effect, sc , which cannot be reduced permeability regardless of proppant-pack permeability,
by increasing the mass of proppant (see Eq. 2-39). as expected.
For example, for the 1-md formation, doubling the When reservoir permeability is less than 0.1 md,
mass of proppant to 300,000 lbm (with all other proppant-pack permeability has slight impact on sc.
variables kept the same) increases the JDTH only to When reservoir permeability increases, sc increases
0.122, almost the same as that for the 150,000-lbm and X decreases.
mass case, where JDTH is 0.121.
These results further suggest that for high- and
6 even moderate-permeability reservoirs, a transversely
fractured horizontal gas well is not attractive because of
5
the production impediment from turbulence effects and
4
converging skin effect. For low-permeability (k 0.5 md)
qv / q ideal
reservoirs, the results should be attractive if multiple
q / q ideal
3 fractures intersecting a horizontal well are generated (and

if the project economics are attractive.)
2 4q TH / q ideal
2.0
Open Hole
1 1.8
qTH / q ideal 1.6
1.4
0
JDV and JDTH
1.2
1 10 100
1.0
k, md JDV (kf =60,000 md) JDV (kf =600,000 md)
0.8
0.6
Figure 2-9 Turbulence effect on fractured vertical and 0.4
0.2 JDTH (kf =600,000 md)
transversely fractured horizontal wells JDTH (kf =60,000 md)
0
0.01 0.1 1 10
k , md
The conclusion from this part of the study is that
hydraulic fracturing is essential for both stimulating Figure 2-10 (a) JDV, JDTH, vs. k for different proppants
and reducing the strong turbulence effects in higher-

0.8 40
permeability vertical gas wells, but the same is not 0.7 35
necessarily true for transversely fractured horizontal gas 0.6 30
X (k f =600,000 md)
wells. Transversely fractured horizontal gas wells are not 0.5 sc (kf =60,000 md) 25
attractive in terms of productivities for moderate and 0.4 20
Sc
X
higher formation permeability (e.g. k > 1 md). 0.3 X (kf =60,000 md) 15
0.2 sc (k f =600,000 md) 10
Results and Analysis 0.1 5
for Formation Permeability from 0.01 to 10 md 0

0.01 0.1 1 10
0
k, md
A second study presents results for a much lower
permeability range (0.01 to 10 md). Designs assume the Figure 2-10 (b) sc, X vs. k for different proppants
use of 150,000 lbm mass of proppant with proppant-
pack permeabilities of 60,000 md and 600,000 md. Because JDV and sc are functions of the mass
Drainage radius is 660 ft. of proppant and proppant-pack permeability, it is
A single transversely fractured horizontal gas well is worth performing a parametric study to show the
calculated. The results are plotted in Figs. 2-10 (a) and effect of important reservoir and treatment variables
2-10 (b). The obvious trends from these results are: on JDTH , JDV and X.
31
Modern Fracturing
Impact of Fracture Treatment Size a horizontal well are generated. Thus, it is useful to study
To find the impact of the mass of proppant on JDTH, a range how the number of isolated zones affects the Equivalent
of proppant mass from 75,000 to 300,000 lbm is used. Number of Vertical Wells.
The proppant-pack permeability used in this study is sand, Assume the total drainage radius is 1320 ft, the
with permeability 60,000 md, and the drainage radius is proppant-pack permeability kf is 60,000 md and mass
660 ft. The results are summarized in Table 2-7. of proppant is 150,000 lbm. The number of isolated
zones and, thus, the number of transverse fractures
Table 2-7 Impact of Mass of Proppant on X and JDTH intersecting a horizontal well vary from 1 to 4.
75,000 lbm
The results, plotted in Fig. 2-11, show that
k, md JDTH X sc when the number of transverse fractures is more
0.01 0.786 0.531 0.16 than four for low permeability (k < 0.5 md), X
0.05 0.465 0.481 0.45
0.1 0.31 0.35 0.86 becomes more than 1, which makes transversely
0.5 0.105 0.198 5.63 fractured horizontal gas wells attractive. The lower
1 0.067 0.152 10.5
the formation permeability is, the more attractive
5 0.029 0.092 28.6
10 0.017 0.06 51.3 the transverse fracture configuration is (subject to
150,000 lbm overall economic considerations). If the formation
k, md JDTH X sc permeability is larger than 1 md, the transverse
0.01 1.075 0.589 0.08 configuration does not appear attractive. For
0.05 0.345 0.487 0.43 example, X is only 0.280 for k = 10 md formation
0.1 0.323 0.294 0.91
0.5 0.106 0.162 5.66 with four transverse fractures generated.
1 0.067 0.127 10.6
5.0
5 0.029 0.08 28.7
10 0.018 0.058 51.3 4.5 k =0.01 md
300,000 lbm 4.0
k, md JDTH X sc 3.5
0.01 1.42 0.755 0.19 3.0

0.05 0.314 0.518 0.46 2.5 k =0.05 md
X
0.1 0.332 0.235 0.95 2.0 k =0.1 md

0.5 0.107 0.138 5.69
1 0.068 0.116 10.6 1.5
5 0.029 0.071 28.9 1.0 k =0.5 md
10 0.018 0.053 51.3 k =1 md
0.5 k =5 md
k =10 md
0.0
1 2 3 4 5
It is apparent that increasing the mass of proppant n
has impact on the results for the low-permeability
(k 0.1 md) formation but virtually no impact in higher Figure 2-11 Impact of number of fractures, n, on X
permeabilities. The reason is that increasing the mass In summary, turbulence effects have a great
of proppant, while it may increase the dimensionless impact on transversely fractured horizontal gas wells
productivity index, also increases the skin factor sc (see due to the small cross-section of the contact between
Table 2-7). The one effect nullifies the other. Thus, the well and the fracture. Although a vertical fractured
there is no need to increase the mass of proppant. A gas well in the permeability range of 1 to 100 md
modest treatment is sufficient. may perform very well, turbulence effect procduce
in unacceptable results in transversely fractured
Impact of the Number of Isolated Zones horizontal gas wells in the same permeability range.
on Equivalent Number of Vertical Wells, X For low permeability (k < 0.5 md), the results are
As mentioned earlier, for low-permeability (k 0.5 attractive if a fracture stimulation treatment generates
md) reservoirs, fracture stimulation results will not be multiple fractures intersecting a horizontal well.
attractive unless multiple transverse fractures intersecting However, if the permeability is larger than 0.5 md,
32
the configuration appears unattractive. For relatively a basic requirement is the value of the wellhead
higher-permeability (0.1 md < k <10 md) formations, pressure, which determines the necessary value of the
increasing the mass of proppant has very small or bottomhole pressure for the desired rate.
virtually no effect on results. A medium mass of Another important task of production engineering
proppant is enough for this permeability range. is to design the well tubulars, primarily the tubing
diameter, to try to accommodate the combination of
2-6 Well Deliverability, IPR pressures in the well and rate.
and Well Flow Performance A plot of the well flow performance, for a given
wellhead pressure, is then done with rate on the horizontal
All expressions relating gas well production rate axis and the required flowing bottomhole pressure on the
with the reservoir and bottomhole pressures can lead vertical axis. Such a plot is convenient because superposing
readily to an inflow performance relationship, IPR. the IPR with the well flow performance provides exactly
Graphically, an IPR is plotted traditionally with the the unique solution for well deliverability.
bottomhole pressure on the vertical axis and the Well deliverability is a central theme of petroleum
production rate on the horizontal axis. production engineering. Calculations and solutions
The IPR of a single-phase oil well forms a straight of various scenarios seek to predict and optimize well
line. The IPR of a gas well is a parabola because of performance. Changing the IPR can be easily related
the pressure-squared dependence, even without the by plotting the original and new IPR and predicting
additional complication of the turbulence term, e.g., well deliverability in combination with the well
Eq. 2-23 (for vertical well, steady state) or 2-26 (for flow performance. At times, one well intervention
horizontal well, pseudo-steady state.) (e.g., hydraulic fracturing of a gas well) may not
An IPR describes a very simple and very useful provide maximum benefit without changing the
picture of well performance. It says that for any given well performance curves. The existing tubing may be
flowing bottomhole pressure there is a corresponding restricting the rate (tubing-limited). Changing the
rate. Intersection of the IPR with the vertical axis is the tubing size can remedy the situation.
driving reservoir pressure, i.e., the initial pressure for The relationship between upstream and
transient conditions, the outer boundary pressure for downstream gas pressures flowing in a vertical
steady state and the average reservoir pressure for pseudo- or inclined tubing, is given in oilfield units, by
steady state. The intersection with the horizontal axis at (Economides et al., 1994):
pwf = 0 is the well absolute open-hole potential.
f ( ZTq ) 2 s
One of the main tasks of production engineering p22 = e s p12 + 2.685103 (e 1), (2-44)
sin Dtbg 5
is modifying IPR. Indeed, all well intervention actions
aim to enhance IPR. Damage removal, re-perforating, where is the well deviation from the horizontal and e
matrix stimulation and hydraulic fracturing all modify is the natural exponent derived from:
the IPR. For a given reservoir pressure, all of these
g0.0375 sin L
production enhancement techniques shift the IPR s=
. (2-45)
ZT
to the right. Reservoir pressure decline also de-facto
changes the IPR, lowering the vertical axis intersection For horizontal flow the relationship is:
and shifting the curve downward. g f ZTq 2 L
What the reservoir can deliver into the well at p12 p22 = 1.007 104 . (2-46)
Dtbg 5
the bottomhole must be tempered by what the well
hydraulics (e.g., tubing diameter, depth, restrictions, To obtain the friction factor, ff , the Reynolds number
valves etc.) can allow on the way to the surface. must be calculated, and the Moody friction factor chart
Thus, for any given production rate there is a unique can be used:
combination of flowing bottomhole and wellhead g q
N Re = 20.09 .
pressures that would produce that rate. In practice, Dtbg (2-47)
33
Modern Fracturing
In Eq. 2-44 through 2-47, p is in psia, q is in Mscf/

Table 2-8 Resevoir Data for Gas Well Deliverability
d, Dtbg is in in., L is in ft, is in cp, T is in R, and all Example
other variables are dimensionless.
k = 0.3 md h = 78 ft re = 1490 ft p = 5100 psi
The indicated calculation procedure is iterative.
If either pressure is known, usually the well head or
g = 0.65 Tres = 185F Ttf = 140F rw = 0.328 ft
downstream pressure, then the other pressure can be
calculated by iteration. The reason is that natural gas
H = 9800 ft Dtbg = 2.259 in. = 90 ptf = 300 psi
physical properties are functions of the pressure value
to be calculated. Given one pressure value (upstream
or downstream) the other pressure is assumed, the
properties are calculated and then the unknown pressure 2-7 Forecast of Well Performance
is calculated using Eq. 2-44 to 2-47. The assumed
and calculated values are compared and the procedure The final step of natural gas production engineering and
repeated until a desired convergence is achieved. one that infringes on reservoir engineering is the forecast
of well performance, adding the necessary element of time.
Example of Gas Well Deliverability Any economic calculation requires this step. To make such
Figure 2-12 shows the results for two skin effects (s=0 and calculation it is necessary, of course, to know the drainage
-5). Table 2-8 contains the well and reservoir variables for area of the well, which can be physically delineated from
this example. For the first case the resulting flow rate q = geologic information, augmented by well testing analyses,
4415 Mscf/d at pwf = 945 psi. For the stimulated well, q or surmised by allocating drainage areas through well
= 10,500 Mscf/d at pwf = 2010 psi. Changing the tubing spacing in a fully developed field. Material balance is
diameter to 3 in. results in (for s =0) q = 4560 Mscf/d at essential, and basic elements are presented below.
pwf = 630 psi, not a great impact on well performance
but for s = -5, the results are more pronounced: q = 2-7.1 Gas Material Balance and Forecast of Gas
12,800 Mscf/d at pwf = 1225 psi, both of which are Well Performance
desirable. For higher reservoir permeability, the impact
of tubing size will be even more significant. This example If Gi and G are the initial and current gas-in-place within
demonstrates the importance of considering both the a drainage area, the cumulative production from a gas
well IPR and well flow performance in reaching the reservoir, considering the expansion of the fluid, is
desirable well deliverability.
Bgi
G p = Gi G = Gi Gi
, (2-48)
6000 Bg
IPR at s = 0 where Bgi and Bg are the corresponding formation
5000 IPR at s = -5
V LP
volume factors.
Equation 2-18 provides Bg in terms of pressure,
4000
temperature and gas deviation factor. Substitution in
Eq. 2-48, assuming isothermal operation throughout,
pwf , psi
3000
and rearrangement results in:
2000 p / Z
G p = Gi 1 . (2-49)
pi / Z i
1000
This expression suggests that if Gp, the cumulative
0 production, is plotted against o/Z, it should form a
0 2000 4000 6000 8000 10,000 12,000 straight line. Usually, the variable is plotted on the
q, Mscf/d
vertical axis and the cumulative production on the
Figure 2-12 Gas well deliverability example horizontal axis. At Gp=0, then, o/Z=pi /Zi, and at p/Z=0
34
then Gp=Gi. For any value of the reservoir pressure yH2S are the mole fraction of nitrogen, carbon dioxide
(and associated Z), there exists a corresponding Gp. and hydrogen sulfide, respectively, in the gas mixture.
Coupled with the gas well IPR expressions presented Therefore the pseudocritical temperature Tpc1 and
earlier in this chapter, a forecast of well performance pressure ppc1 for the whole mixture are:
versus time can be developed readily.
Tpc1 = ATpc + 227.3 yN 2 + 547.6 yCO2 + 672.4 yH 2 S
2-8 Correlations for Natural Gas p pc1 = Ap pc + 493.0 yN 2 + 1071 yCO2 + 1306 yH 2 S . (2-53)
Properties
Finally, the actual corrected critical properties must
Natural gas properties are used throughout production include the Wichert and Aziz correction factor given by:
engineering, from reservoir inflow to well flow 0.90 1.6
performance. In the past, graphical correlations have = 120 ( yCO2 + yH 2 S ) ( yCO2 + yH 2 S )
been used and can be found in almost all production 0.5
+ 15 ( yH 2 S ) ( yH 2 S ) .
4 (2-54)
and reservoir engineering textbooks. This book uses
explicit correlations which can be readily programmed to And thus, the corrected pseudocritical temperature and
calculate physical and thermodynamic properties. pressure are
2-8.1 Pseudocritical Pressure, ppc Tpc* = Tpc1

and Pseudocritical Temperature, Tpc p pc1 * Tpc*
p*pc = .
Tpc1 + yH 2 S (1 yH 2 S ) (2-55)
The critical temperatures and pressures are derived
from correlations for pseudocritical temperatures, Tpc 2-8.2 Gas Viscosity
and pressures, ppc, (Brown et al., 1948), with g= gas
gravity to air = 1. Gas viscosity is calculated with the Lee et al. (1966)
For condensate fluids, g < 0.00001: correlation:
Tpc = 187 + 330 g 71.5 g2 (2-50) = K exp X Y , (2-56)

p pc = 706 51.7 g 11.1 g2 .
where
For miscellaneous fluids, g 0.00001:
(9.4 + 0.02M )T 1.5
Tpc = 168 + 325 g 12.5 2 K=
g (2-51) 209 + 19 M + T
p pc = 677 + 15.0 g 37.5 g2 . 986
X = 3.5 + + 0.01M
T (2-57)
Equations 2-50 and 2-51 were developed for Y = 2.4 0.2 M
gases with no contaminants such as H2S, CO2 and pM
= ,
N2. If any of these contaminants are present, the ZRT
pseudocritical values must be corrected using the
Wichert and Aziz (1972) correction. and where M is molecular weight, is gas density, and
The gas gravity, g in Eq. 2-50 and 2-51, is replaced T is temperature in R.
by hc as given by:
2-8.3 Gas Deviation Factor, Z
g 0.967 yN 2 1.5195 yCO2 1.1765 yH 2 S
hc = .
A (2-52) The gas deviation factor, Z is calculated using either the
Hall and Yarborough (1974) or the Dranchuk-Purvis-
where A=(1 yCO2 YN2 yH2S), and yN2, yCO2 and Robinson (1974) correlations:
35
Modern Fracturing
Hall and Yarborough Correlation. The method is valid within the following ranges of
pseudo-pressure and temperature:
0.06125 p pr t
Z= exp 1.2(1 t ) 2 , (2-58) 1.05 < Tpr < 3.0
Y
0.2 < ppr < 3.0
where ppr is the pseudo-reduced pressure, t is the reciprocal
of the pseudo-reduced temperature, (i.e., Tpc /T), and Y is
the reduced density that can be obtained from:
References
Y +Y 2 +Y 3 +Y 4
F (Y ) = X 1 + 3
(1 Y ) (2-59) Al-Hussainy, R., and Ramey, H.J.: Jr., Application of
2
( X 2)Y + ( X 3)Y X4
=0 Real Gas Theory to Well Testing and Deliverability
Forecasting, JPT (May 1966) 637-642.
where Brown, G.G., Katz, D.L., Oberfell, C.G., and Alden,
R.C.: Natural Gasoline and the Volatile Hydrocarbons,
X 1 = 0.06125 p pr t exp 1.2(1 t ) 2
NGAA, Tulsa, OK (1948).
X 2 = (14.76t 9.76t 2 + 4.58t 3 ) Cooke, C.E., Jr.: Conductivity of Proppants in
X 3 = (90.7t 242.26t 2 + 42.4t 3 ) Multiple Layers, JPT (Sept. 1973) 1101-1107.
X 4 = (2.18 + 2.82t ). Crosby, D.G., Yang, Z., and Rahman, S.S.: Transversely
Fractured Horizontal Wells: A Technical Appraise
The Dranchuk-Purvis-Robinson correlation of Gas Production in Australia, paper SPE 50093,
1998.
1 + T1r + T2r2 + T3r5 + Demarchos, A.S., Chomatas, A.S., Economides, M.J.,
T4 r2(1+ A 2 ) exp( A 2 ) T5 = 0, (2-60) Mach, J.M. and Wolcott, D.S.: Pushing the
8 r 8 r
r Limits in Hydraulic Fracture Design, paper SPE
with 86483, 2004.
Dranchuk, P.M., Purvis, R.A. and Robinson, D.
A A B.: Computer Calculation if Natural Gas
T1 = A1 + 2 + 33
Tpr Tpr Compressibility Factors Using the Standing and
Katz Correlation, Institute of Petroleum, IP 74-
A
T2 = A4 + 5 008, 1974.
Tpr
Economides M.J.: Hydraulic Fracturing-a short course
T3 = A5 A6 / Tpr by Prof. Michael J. Economides, 2004.
T4 = A7 / Tpr3 Economides, M.J., Hill A.D., and Ehlig-Economides,
C.A.: Petroleum Production Systems, Prentice Hall,
T5 = 0.27 p pr / Tpr . NY (1994).
Economides, M.J., Valk, P.P. and Wang, X.: Recent
The coefficients A1 through A8 have the following Advances in Production Engineering, JCPT
values: (October 2001) 35-44.
Economides, M.J., Oligney, R.E., and Valk, P.P.: Unified
A 1 = 0.31506237 A 2 = -1.0467099 Fracture Design, Orsa Press, Houston (2002a).
A 3 = -0.57832720 A 4 = 0.53
3530771 Economides, M.J., Oligney R.E., and Valk, P.P.:
A 5 = -0.61232032 A 6 = -0.10488813 Applying unified fracture design to natural gas
A 7 = 0.68157001 A 8 = 0.68446549 wells, World Oil (Oct. 2002b), 52-62.
Eirafie, E.A., and Wattenbarger, R.A.: Comprehensive
Evaluation of Horizontal Wells with Transverse
36
Hydraulic Fractures in the Upper Bahariyia Romero, D.J., Valk, P.P., and Economides M.J.:
Reservoir, paper SPE 37759, 1997. Optimization of the Productivity Index and the
Emanuele, M.A., Minner, W.A., Weijers, L., Fracture Geometry of a Stimulated Well With
Broussard, E.J., Blevens, D.M., and Taylor. B.T.: Fracture Face and Choke Skins, paper SPE 73758,
A Case History: Completion and Stimulation 2002.
of Horizontal Wells with Multiple Transverse Sanger, P.J., and Hagoort, J.: Recovery of gas
Hydraulic Fractures in the Lost Hills Diatomite, condensate by nitrogen injection compared with
paper SPE 46193, 1998. methane injection, SPEJ. (1998) 3(1), 26.
Fisher, M.K., Heinze, J.R., Harris, J.R., Davidson, B.M., Soliman, M.Y., and Boonen, P.: Review of Fracturing
Wright, C.A., and Dunn, K.P.: Optimization Horizontal Wells Technology, paper SPE 36289,
Horizontal Completion Techniques in the Barrnett 1997.
Shale Using Microseismic Fracture Mapping, Soliman, M.Y., Hunt, J.L., and Azari, M.: Fracturing
paper SPE 90051, 2004. Horizontal Wells in Gas Reservoirs, SPEPF (Nov.
Forchheimer, P., Hydraulik, Teubner, Leipzig and Berlin 1999)14 (4).
(1914) 116-118. Swift., G.W., and Kiel, O.G.: The Prediction of Gas-
Firoozabadi, A., and Katz, D.L.: An Analysis of High- Well Performance Including the Effects of Non-
Velocity Gas Flow Through Porous Media, JPT Darcy Flow, JPT (July 1962) 791-798.
(Feb. 1979) 211-216. Tek, M.R., Coats, K.H., and Katz, D.L.: The Effect
Gidley, J.L.: A Method for Correcting Dimensionless of Turbulence on Flow of Natural Gas Through
Fracture Conductivity for Non-Darcy Flow Porous Reservoirs, JPT (July 1962) 799-806.
Effects, paper SPE 20710, 1990. Valk, P., and Economides, M.J.: Performance of a
Hall, K.R. and Yarborough, I.: A New Equation of Longitudinally Fractured Horizontal Well, SPEJ
State for Z-Factor Calculations, Oil & Gas J., (March 1996) 11-19.
June 18, 1973, 82-92. Villegas, M.E., Wattenbarger, R.A., Valk, P., and
Ichara, M.J.: The Performance of Perforated Completions Economides, M.J.: Performance of Longitudinally
in Gas Reservoirs, paper SPE 16384, 1987. Fractured Horizontal Wells in High-Permeability
Karakas, M., and Tariq, S.: Semi-Analytical Production Anisotropic Formations, paper SPE 36453,
Models for Perforated Completions, paper SPE 1996.
18247, 1988. Wang, X.: Pore-Level Modeling of Gas-Condensate
Katz, D.L., Cornell, D., Kobayashi, R., Poettmann, Flow in Porous Media, PhD dissertation,
F.H., Vary, J.A., Ellenbaas, J.R., and Weinang, C.F.: University of Houston (May 2000).
Handbook of Natural Gas Engineering, McGraw- Wang, X., and Economides, M.J.: Aggressive Fracture
Hill, NY (1959). Slashes Turbulence in High-Permeability Gas
Lee, A., Gonzalez, M.H., and Eakin, B.E.: The Well, (July 2004).
Viscosity of Natural Gases, JPT (Aug. 1966) 997- Wang, X., and Mohanty, K.K.: Critical condensate
1000. saturation in porous media, J. Coll. & Interf. Sci.
McCain, W.D., Jr.: The Properties of Petroleum Fluids, (1999a) 214, 416.
Petroleum Publishing Co., Tulsa (1973). Wang, X., and Mohanty, K.K.: Multiphase non-Darcy
Minner, W.A., Ganong, B.L., Demetrius S.L., and flow in gas-condensate reservoirs, paper SPE
Wright, C.A.: Rose Field: Surface Tilt Mapping 56486, 1999b.
Shows Complex Fracture Growth in 2500 Laterals Wang, X., Thauvin, F., and Mohanty, K.K.: Non-
Completed with Uncemented Liners, paper SPE Darcy flow through anisotropic porous media,
83503, 2003. Chem. Eng. Sci. (1999) 54, 1859.
Mukherjee, H., and Economides, M.J.: A Parametric Wei, Y.: Transverse Hydraulic Fractures From A
Comparison of Horizontal and Vertical Well Horizontal Well, MS thesis, University of Houston
Performance, paper SPE 18303, 1991. (November, 2004).
37
Modern Fracturing
Wichert, E., and Aziz, K.: Calculation of Zs for Sour

Gases, Hydrocarbon Processing (1972) 51(5).
Yao, C.Y., and Holditch, S.A.: Estimating Permeability
Profiles Using Core and Log Data, paper SPE
26921, 1993.
38
Bob Bachman is currently at Taurus Reservoir Solutions in Calgary, Alberta, Canada. Areas of expertise
include all aspects of traditional reservoir engineering, reservoir simulation and analyzing the performance
of hydraulically fractured wells. He graduated from the University of Manitoba in 1976 and 1978 with
BSc and MSc degrees in civil engineering. He began his career as a reservoir engineer at Mobil Oil Canada,
looking at hydrocarbon miscible floods. He currently has over 27 years consulting experience. Companies
he has worked for include Intercomp, SSI, Simtech and Duke Engineering.
Robert Hawkes is team leader of reservoir services for BJ Services Company Canada, overseeing petrophysical
and reservoir engineering services. He graduated from the Southern Alberta Institute of Technology in 1979
with a diploma in petroleum engineering. He began his career with Esso Resources in Calgary as a reservoir
technologist in the testing department and went on to become a senior well testing specialist with Fekete
Associates Inc. Hawkes is well-published in CIM and SPE and is a distinguished lecturer for the SPE 2007-
2008 Distinguished Lecture Program. He is a member of SPE, the Petroleum Society of CIM, Society of
Petroleum Petrophysicists and Well Logging Analysis (SPWLA) and the Alberta Society of Engineering
Technologists (ASET).
Chapter 3 2. How does the actual treatment, with respect to
the propped fracture properties, compare to
Gas Well Testing what was designed? Here the analyst is trying to
and Evaluation compute the fracture length and the conductivity
Bob Bachman, Taurus Reservoir Solutions and of the proppant pack. This would quantify if there
were any positive skin left after the treatment
Robert V. Hawkes, BJ Services
and whether the stimulation was successful in
With special contributions on closure analysis from removing it. Results of the analysis may determine
Bob Barree, Barree and Associates
if future design changes need to be made. This
in turn could lead to recommended changes in
3-1 Introduction fracture geometry (propped length and width), the
materials in use (proppant type and stimulation
Pressure transient testing of wells is now a commonly fluid) and the size of the treatment.
accepted sub-discipline of reservoir and production
engineering. The analysis of such a test is called Tests are most often done immediately after drilling,
pressure transient analysis (PTA). Conducting such casing and perforating, although open-hole DST tests are
tests is costly in both an absolute sense and in possible sometimes performed. This allows stimulation decisions
temporary loss of production. The benefits must to be made. Tight gas wells (those with permeability
therefore outweigh the costs if the test is to be done. below 0.1 md) may not flow without being stimulated.
Nevertheless, pressure transient testing is performed In these cases other methods, such as closed chamber
around the world because it is one of only a few testing, may be done to determine permeability. These
processes that help to quantify well behavior. Within alternative tests will be discussed in Section 3-12.2.
the context of hydraulic fracturing it is especially Another important area for pressure transient
useful in answering the following questions: testing is determining the average well pressure and
1. What are the reservoir properties so that how this changes with time from well to well in the
optimal hydraulic fracturing treatments can pool. This in turn helps determine the contacted gas
be designed? In designing fracture treatments in place and, indirectly, well reserves, in addition to
(as presented throughout this book and, in connectivity among various wells in the pool.
particular in Section 4-5), knowledge of reservoir A large number of commercially available software
permeability is crucial. This property is one of the packages are available to help the user in the design
most important findings of a pressure transient and analysis of PTA tests. This removes the burden of
test and it is most unambiguously determined data manipulation and problem set-up from the user.
before a treatment. Other properties that can However by making problem set-up so easy and the
be obtained from the test include the initial mathematics largely hidden by various plotting and
pressure pi, and the damage around the well. This calculation techniques, the user may not understand
damage is referred to as skin, s, and is normally what is going on. The focus of this chapter will be in
considered a short-distance phenomenon. A giving background PTA theory and more importantly,
positive skin indicates that damage has occurred guidance about accepted practices in interpreting test
and will result in production impairment results. This chapter is intended as a supplement to
and additional pressure loss around the well. the already excellent books on the subject of PTA
A negative skin corresponds to a production (e.g., Dake, 1978; Horne, 1997; Lee et al., 2003; and
enhancement compared with the normal state. Houz et al. (2007)) and will have a particular focus
The goal of a hydraulic fracture treatment is on natural gas testing.
to create an equivalent negative skin situation, Before proceeding, pressure transient test must
which is the result of the fracture half-length be defined. The classical definition is a process whereby
and conductivity (see Chapter 2). the production from the well is controlled, usually
41
Modern Fracturing
at the wellhead, and the resulting pressure measured the assumption of constant density, viscosity and
either at the wellhead or, preferably, at bottomhole compressibility is appropriate. If, in addition, constant
conditions. The overwhelming majority of tests have permeability is assumed then:
a sequence of production rates followed by a build-
2 p 1 p ct p
up. A build-up occurs when the well is shut-in and + = , (3-1)
r 2 r r k t
produces at a zero rate. Interpreting test results is most
easily performed when rates are held constant, and a where ct is the total compressibility and p is pressure.
well being shut-in is the easiest constant rate condition This is a linear equation, as none of the coefficients
to obtain operationally. In the literature, there is an are a function of pressure. This is known as the radial
overwhelming focus on the proper interpretation of diffusivity equation. The left side of this equation
the last build-up. The authors feel strongly that all flow represents the flow term in the reservoir. The right side
periods need to be reviewed and analyzed, in order to represents the accumulation of material at a given point
reduce non-uniqueness issues that frequently occur. over time. For gas reservoirs, it is possible to linearize
The typical duration of a test is less than one month, the left side of Eq. 2-4 by the process described in
often less than 10 days and in some cases as little as 24 Section 2-3.2, using the definition for real gas pseudo-
hours of flow, followed by a 24-hour build-up. pressure, m(p). From now on this will be referred to as
The time duration of a test gives a natural limit as just pseudo-pressure. This gives:
to how far into the reservoir properties can be inferred.
From the analysts point of view, all tests appear to be 2 m( p ) 1 m( p ) ct m( p )
(2-13)
2
+ = .
of insufficient length, as one is always on the verge of r r r k t
determining properties further into the reservoir at any The right side terms (viscosity) and especially
time. The realities of the oil and gas business dictate ct (total compressibility) are strong functions of
that compromises be made; enough time is granted pressure. Initial attempts to get around this non-
to conduct a test so that operational and stimulation linearity assumed the product of these terms was
decisions can be made, and no more. constant at the initial pressure value, which is only
approximately true. Agarwal (1979) introduced an
3-2 Background Theory additional term called pseudo-time, ta, which has
been slightly modified to be:
There is a tremendous body of literature dedicated to
t dt '
PTA, summarized in a review paper by Gringarten ta = (ct )i
, (3-2)
0 ct
(2006). Equations describing flow in porous media
involve combining the continuity equation (conservation where the subscript i represents initial conditions. This
of mass) with Darcys law (relationship between flow definition allows pseudo-time to retain units of time.
rate and pressure drop) as was shown in Chapter 2. This Subsequently, Lee and Holditch (1982) gave a strong
results in a second-order partial differential equation. theoretical justification for the approach. Pseudo-time
For vertical wells in radial coordinates, where vertical accounts for the large change in gas compressibility
and angular properties do not vary, this results in a one- that occurs at low pressures (early time in a build-
dimensional linear differential equation: up). It has little effect on the late time data, and was
introduced for build-ups only. For drawdowns, Lee
1 k p
r = , (2-4) and Holditch (1982) set the pressure equal to the

r r r t
initial pressure in Eq. 3-2 for the entire flow period.
where r is the radial distance, t is time, is the fluid Implicit in this calculation is that pressure must be
density, k is the permeability, is the porosity and known as a function of time.
the viscosity. This equation is appropriate for both Which pressure should be used? The pressure
oil and gas because no constitutive assumptions typically measured is at the well, so it seems a logical
have been made about the density term. For liquids, choice, and indeed this is what Lee and Holditch
42
Chapter 3 Gas Well Testing and Evaluation
(1982) imply. Production Analysis (or PA) techniques where tD is the dimensionless time, pD is the
use the average pressure so that material balance dimensionless pressure, r w is the wellbore radius, T
can be rigorously applied. This requires the size of the reservoir temperature, q is the rate, m(pi) is the
the container (i.e. the drainage area) to be specified pseudo-pressure at the initial pressure pi and m(pwf) is
before the analysis can begin, or else an iteration the flowing pseudo-pressure at the sandface pressure
to be performed upon it as a part of the solution pwf. The difference between Eqs. 2-14 and 3-4 is
process. Well test programs typically have not that the time function is different. For hydraulically
required this in their analytical engines. For software fractured wells, the definition of dimensionless time
using the bottomhole pressure in the pseudo-time changes by replacing the wellbore radius by the
calculation, long-term forecasts will not honor fracture half-length, xf:
material balance (Houz, 2002).
0.000264kta
When Eq. 3-2 is substituted into Eq. 2-13, and
tD = . (3-5)
( ct )i x 2f
higher-order terms are dropped, then one has:
Frequently one wishes to solve the problems where pwf
2 m( p ) 1 m( p ) ( ct )i m( p )
+ = , (3-3) is constant and the rate declines with time. For this case
r 2 r r k ta
the pD definition in Eq. 2-15 is not appropriate and
which is linear. This final transformation recasts the solutions are expressed in terms of the dimensionless
gas equation into the same form as the oil equation. rate, qD:
Mathematically there is now no difference between
1 kh[m( pi ) m( pwf )]
the two formulations. For gas problems, care = . (3-6)
qD 1424qT
must be taken as to how the compressibility term
is entered into commercial PTA software. Most Although the right side is apparently identical to Eq.
software codes require the user to specify only the 2-15, it must be remembered that in Eq. 2-15 q is
rock compressibility term, as the gas compressibility constant, while in Eq. 3-6 pwf is constant.
is automatically calculated from the specified PVT Two terms that play a significant role in PTA
properties. Current software is formulated in terms interpretation are the diffusivity () and storativity
of pseudo-pressure and possibly pseudo-time. For coefficients (). These are defined as:
pseudo-time-based software, rock compressibility k
will be added to the pressure-dependent gas
= ,
(c t )i (3-7a)
compressibility derived from the specified or
correlated PVT properties. For software using the and
concept of constant ct, viscosity and compressibility
would be specified at initial conditions, and the
= (ct )i h, (3-7b)
user may be expected to enter the combined
rock and gas compressibility. where h is the total net pay. The velocity of a
By applying the additional transformation of pressure disturbance from a given rate change is
dimensionless time (Eq. 2-14) and dimensionless proportional to the diffusivity coefficient. The
pressure (Eq. 2-15) into Eq. 3-3, a standard form is higher the permeability and the lower the system
achieved to report solutions to well test equations for compressibility, the faster the disturbance travels, and
radial problems with a constant producing rate: the sooner boundaries will be encountered. Pressure
disturbances typically travel much faster in oil wells
0.000264kta
tD = , (3-4) than gas wells due to the significant differences in
( ct )i rw2
compressibility. The storativity is less important,
and especially in short-term tests, but it can be significant
when defining heterogeneities within a system or
kh[m( pi ) m( pwf )]
p D = , (2-15) when depletion occurs.
1424qT
43
Modern Fracturing
3-3 Radial Flow Solutions m( pi ) m( pwf ) 0.000528kta

1422T r 3
= + ln e + s ,
q kh ( c ) r 2
rw 4
Van Everdingen and Hurst (1949) were the first to t i e
solve Eq. 3-1. The initial cases considered a constant (3-10)
well rate at the inner boundary. Various outer boundary
conditions, such as no flow and constant pressure, where re is the outer drainage radius. Fig. 3-1 shows
were also solved. They extended the solutions to a comparison of the transient and pseudo-steady
include wellbore storage, and in subsequent work (van state cases. When wellbore storage dominates
Everdingen, 1953; and Hurst, 1953) solutions included at the earliest times:
both wellbore storage and skin effects, but only gave m( pi ) m( pwf ) 0.420082T
partial solutions. Raghavan et al. (1970) extended the = ta . (3-11)
q i C
solution to all circumstances.
Wellbore storage is the phenomenon that occurs where C is the wellbore storage coefficient, which is a
when the rate is controlled at the surface rather than measure of how the compressibility of the well and its
the sand face. This results in a delay in the production contents affects the early time pressure response.
response at the sand face. Wellbore storage occurs
due to the compressibility of the wellbore fluids or a Transient
changing fluid level in the well (Horne 1997). Even
for the relatively simple case of no wellbore storage Pseudo-Steady State
or skin, the mathematical solution as a function

of time remains formidable. pe
For all flow problems, and not just the
p
homogeneous radial flow solution, there are two

special asymptotic cases. The first occurs at early Time
time, but not so early that wellbore storage affects

the results. This is called the transient solution, where
there is no influence from any outer boundaries.
r
This has already been given in Chapter 2 as Eq. Figure 3-1 Comparison of transient and pseudo-steady
2-16 in terms of t. In general: state cases
The complexity of the van Everdingen and

Hurst (1949) solution comes from the so-called
late transient term, where there is a transition from
pure transient flow to pseudo-steady state flow. For
(3-8) radial flow, the late transient flow period is of short
duration. For more complicated geometries, some
T is temperature and rw is wellbore radius. This can be but not all boundaries are felt for significant periods
expressed in dimensionless terms as: of time, making the late transient period significant.
This late transient time presented great difficulties for
1
pD = (ln t D + 0.80907) + s, (3-9) the first generation of theoreticians in PTA.
2
After a well has reached pseudo-steady state
The second asymptotic solution occurs at late time when conditions, it is not possible to decouple permeability
all the boundaries are felt. For the special case of a constant and skin effects. Instead, one can only determine
rate with a no-flow outer boundary, this is referred to as the overall productivity index (J) of the well, which
the pseudo-steady solution, and pseudo-pressure at all represents a combination of permeability and skin.
points in space declines at a constant rate with time: For cylindrical drainage this equation is:
44
kh q flow, which is similar to pseudo-steady state flow.

J=
= Bounded flow assumes all the reservoir boundaries
1422T [ ln(re / rw ) (3 / 4) + s ] m( pavg ) m( pwf )
have been encountered and the well is flowing at
(3-12) constant pwf, with a declining rate.
For a typical gas well the range of permeability is
where m(pavg) is the pseudo-pressure at the average between 0.001 to 10 md, or 4 orders of magnitude.
reservoir pressure pavg. The average pressure changes with Permeability is typically highly variable across a
time according to the material balance equation. For group of wells. For a given constant pressure drop
hydraulically fractured wells, an overall negative skin the calculated rate is proportional to permeability,
can represent the fracture after radial flow is achieved. indicating a possible rate variation of 4 orders of
The treatment itself does not affect the permeability magnitude, as given by Eq. 3-12. Independent of
of the well, only the skin. For wells draining differing permeability, a well normally has a positive skin after
reservoir shapes, Eq. 3-12 can be appropriately the initial completion. In practice this varies from +25
modified, by using the Dietz shape factor. The time to (highly damaged) to 0 (no damage). After a fracture
get into bounded flow has been determined by Dietz stimulation treatment the best skin factor that can
(1965) and is also dependent on the drainage area be expected is about 6. For an initially undamaged
shape. For convenience, an additional dimensionless well (s = 0), a stimulation treatment can result in a
time function, tDA, is defined as: four-fold change in rate by Eq. 3-14. Permeability
dominates the resultant production rate, making the
0.000264kt
t
DA = , (3-13) success or failure of a treatment hard to quantify if
( ct )i A
permeability is unknown.
where A represents the drainage area of the well. For a The only way to determine the effectiveness of
well in the center of a square or circular drainage area, a fracture treatment is to separate the permeability
the time to pseudo-steady state flow is tDA = 0.1. For wells and skin effects. This can only be accomplished
with the vast majority of drainage area shapes, Brown from an analysis of the transient portion of the flow.
and Hawkes (2005) developed a simple relationship for Skin can then be related to the fracture geometry
comparing the benefit of a stimulation treatment based and proppant conductivity, from data prior to the
on a ratio of the pre- and post-stimulated skin effect, establishment of radial flow.
which is a simplification of Eq. 3-12. They called it the
rule of eight, and it is a slight modification of earlier work 3-4 Superposition
by Golan and Whitson (1991):
During actual pressure tests, rate variation with time has
[8 + so ]
qf =
[ qo ], (3-14) to be dealt with. Because the fundamental mathematical
8 + s f
equations are linear, superposition (sometimes called
where so and sf represent the skin effects before and rate convolution) can be applied to develop more
after fracture treatment, respectively. Similarly qo and general solutions from the building block of the
qf represent the boundary-dominated flow rates before constant rate solution, as shown by Odeh and Jones
and after fracture treatment. (1965). By definition, a new transient is introduced at
Evaluating the effectiveness of fracture stimulation each rate change. If the last flow rate was a shut-in,
treatments for wells with long-term production data one then has a pressure build-up test. Transient flow is
onlywithout knowledge of permeability or skin guaranteed for at least some period of time. Assuming
is difficult. These wells are assumed to be in bounded a single constant rate and build-up, Theis (1935) and
1422Tq t pa + ta 1 4t
m( pi ) m( pwf ) = 1.151 log + pD (t pa ) ln pa , (3-15)
kh t
a
2
45
Modern Fracturing
Horner (1951) determined a special superposition time where qn is the gas rate in the nth flow period. Equation
function (Horner time) appropriate for the build-up 3-16a is the classical definition in the textbooks (Dake,
portion of the test (see Eq. 3-15), where is a constant 1978, Eq. 8.49), and the superposition time function,
equal to 1.781. ta is delta pseudo-time or alternatively tasuper1, would plot on a linear scale. Equation 3-16b is
t
the pseudo-time since shut-in, tpa is the pseudo-time at just a mathematical trick because ta sup er 2 = 10 . The
a sup er 1
shut-in and pD(tpa) is the dimensionless pressure at tpa second superposition time function plotted on log
and is an unknown constant. scale is identical in form to the Horner plot. Even
If radial flow is present, then a semi-log plot of where radial flow is not strictly true, or boundaries
m(pwf) versus (tpa + ta)/ta (the Horner plot) should are felt in some of the earlier flow periods, the radial
be linear. The permeability can be calculated from the superposition definition is used.
slope of this plot, and the y-intercept gives the false Some well test software no longer allows the direct
pseudo-pressure (or m(p)*). An auxiliary equation plotting of the original Horner plot, preferring to use
eliminates the unknown term pD(tpa) and gives the the more general superposition time concept. Dake
skin. (In Section 3-12.1, Fig. 3-16 will illustrate the (1978 and 1994) strongly objected to analyses based
use of this plot.) m(p)* can be interpreted as m(pi) on superposition time because it is based on infinite
if minimal production from the field has occurred. acting radial flow occurring throughout the entire test,
Matthews et al. (1954) established ways to determine and not just the last flow period. He argued that if this
m(pi) from m(p)* where significant production has assumption were incorrect, the test would be inherently
occurred. Curvature in the Horner plot indicates misinterpreted. Until fairly recently there was no
other flow regimes or a breakdown in the assumptions adequate answer to his objections. This point will be
related to the development of the plotting function, deferred to Section 3-11.
such as a boundary. In practice it has proven very When superposition is used to generate plots in
difficult to pick the correct straight line on the Horner the manner described so far, one can only analyze the
plot, and a number of plausible and significantly equivalent of the time within one flow period. This puts a
different results could be generated. This point will limit on the volume of the reservoir for which properties
be addressed in Section 3-11. can be determined. While the rate information is
One can extend the Horner time concept for an incorporated into the superposition time term, causing an
arbitrary series of rates, assuming infinite acting radial adjustment in the pressure response, limited information
flow (Earlougher, 1977). The resulting time function about earlier rates seeing a possible boundary is available.
is commonly called superposition time. The arbitrary This is an additional reason the build-up is the preferred
rate solution is presented in two forms: flow period to match: Direct control over the length of
this flow period can be exercised, independent of normal
field operating conditions. These limitations are only now
being circumvented by a process called deconvolution,
which will be discussed in Section 3-18. Deconvolution
allows the transformation of the entire well test, with
(3-16a) all its rate variations, into an equivalent constant rate
drawdown, covering the full length of the test.
and
3-5 Model Development
As work progressed from the 1950s, the solutions to

ever more complicated cases with a variety of boundary
conditions were being published (see Gringarten,
2006, for an overview). The most significant cases in
(3-16b) modern software include:
46
1. Non-centered wells in a rectangle. These formulae are used as simple checks to more
2. Multiple radial composite rings (composite in detailed analyses. Gringarten et al. (1974) developed
diffusivity (Eq. 3-7a) and storativity (Eq. 3- full solutions for the cases of infinite fracture
7b) coefficients, the normal variation being in conductivity and uniform flux into a hydraulic
permeability only). fracture within a rectangular drainage area. Cinco-
3. Multi-layered wells, which are only in pressure Ley et al. (1978) and Cinco-Ley and Samaniego
communication through the wellbore. (1981a) developed solutions for the case of a finite-
4. Dual-porosity or naturally fractured wells. conductivity fracture (see also Section 4-6). The
Typically, these wells have a high-diffusivity, dimensionless fracture conductivity, CfD (see Section
low-storativity media connected to the wellbore. 4-2.1.6) of a fracture is given by:
A second media, with low diffusivity and high
k w
storativity, feeds fluid into the first media. The
C fD =
f
(3-18)
time delay in the low-storativity media becoming kx f
dominant is what makes this system unique. where kf w is the fracture conductivity. It is made
5. Infinite-conductivity hydraulically fractured wells up of the fracture permeability kf and the fracture
within a rectangle/circle. width w. Values of CfD > 30 are normally considered
6. Finite-conductivity hydraulically fractured wells to be infinite conductivity. This concept is further
within a rectangle/circle. discussed in Section 4-2.1.6. In Cinco-Ley and
7. Horizontal wells. Samaniego (1981a) four possible transient flow
regimes are identified for hydraulically fractured
The difference between infinite- and finite- wells. These are shown in Fig. 8-4.
conductivity fracture behavior will be discussed in The first of these is fracture linear flow. No
the next section. A wide variety of inner boundary production pressure response from the formation
conditions (constant rate or constant pressure with has occurred at this stage. This is of extremely
or without wellbore storage and skin) and outer short duration and is in practice not visible due to
boundary conditions (no flow, constant pressure or wellbore storage effects. The second flow regime is
leaky fault) have all been solved. either bilinear flow (for finite-conductivity fractures)
or formation linear flow (for infinite-conductivity
3-6 Hydraulically Fractured Wells fractures). For either case, this is followed by
elliptical flow, which is sometimes referred to as
The analysis of hydraulically fractured wells is of pseudo-radial flow, and finally radial flow (infinite
particular significance, because it is possible to quantify acting radial flow or IARF).
in-situ fracture properties. PTA is one of the best ways Bilinear flow, which should be expected in the
to determine the effectiveness of the treatments placed vast majority of hydraulic fractures, and formation
in the field. Prats et al. (1962) performed one of the linear flow have simple pseudo-pressure relationships
first analyses of hydraulically fractured wells. For an with time given by:
infinite-conductivity fracture, no significant pressure
drop occurs in the fracture. Laterafter radial flow for bilinear flow (3-19a)
has been achievedthere is a relationship between
the fracture half-length, xf, and the effective wellbore with
radius, rw. Knowing the effective wellbore radius, the
skin effect can be directly calculated from: 443.3581qT (3-19b)
(k f w)0.5 k 0.25 = ,
h 0.25 (ct )i0.25 mbf
1
rw = x f ,
(3-17a) and for formation linear flow,
2
rw = rwes or s = ln(rw / rw ). (3-17b) and (3-19c)
47
Modern Fracturing
Table 3-1 Duration of Flow Regimes in Hours

Flow Regime Equation Value of tD or tDA
1.53
a) t D = 0.0205 (C fD 1.5)
1.6 CfD 3.0, first use Eq. a) to find tD
End of Bilinear Flow
0.000264kt then use tD to solve Eq. b)
b) tD =
( ct )i x 2f
0.000264kt
End of Formation Linear Flow tD = 0.016
( ct )i x 2f
0.000264kt 3.0
Start of Radial Flow tD =
( ct )i rw2 slightly lower value for CfD 3
0.000264kt 0.1
End of Infinite Acting Behavior t DA =
( ct )i A circular bounded reservoir
0.000264kt 0.1
Start of Pseudo-Steady State t DA =
( ct )i A circular bounded reservoir
rw 0.5
40.9521 = f (C fD ) = .
x f k 0.5 = 0.5
, (3-19d) x f C fD
1.01
(3-20)
h (ct )i0.5 mlf 1 +
1.7
where mbf is the bilinear flow slope and mlf is the linear Work by Cinco-Ley and Samaniego (1981b) introduced
flow slope. Horne (1997) has provided comprehensive the idea of a damaged hydraulically fractured well. Two
formulae to determine when these and other flow possible cases were quantified: the so-called choked
regimes start and end (Table 3-1). fracture (Fig. 7-39) and a fluid loss damage case (Fig.
Cinco-Ley and Samaniego (1981a) presented 7-40). In practice it is not possible to determine from
a chart, Fig. 3-2, showing how, after radial flow has pressure data alone which of these is occurring.
occurred, rw /xf relates to fracture conductivity. The An analog to the radial superposition plot (a.k.a. the
skin effect can be calculated from Eq. 3-17b. generalized Horner plot) can be made for both bilinear
and linear flow, accounting for rate changes over the
1.00
production history. These are sometimes referred to as
the tandem quad-t and tandem root-t plots, respectively.
Like the radial superposition plot, they are generated
assuming transient flow occurs through all flow periods.
rwa/xf
0.10
3-7 Specialized Plots
Returning to the general case, specialized plots can

0.01 be constructed versus an appropriate time function
0.01 0.1 1 10 100
CfD
for a flow regime. These plots allow the calculation
of reservoir properties if the correct straight lines are
Figure 3-2 Relationship between effective wellbore
radius and CfD (from Cinco-Ley and Samaniego, 1981a) chosen. For radial flow, a plot of m(pi) m(pwf) or m(pwf)
versus log ta the so-called MDH plot from Miller et
Subsequently, Barree et al. (2003) used a curve fit to al. (1950) should give a straight line. From Eq. 3-
put Fig. 3-2 into a simple equation: 8, the slope is inversely proportional to permeability.
48
Analogous plots for bilinear (m(pwf) versus ta0.25) and 3-8 Type Curves and the Log-Log
linear flow (m(pwf) versus ta0.5) will allow one to determine Derivative Plot
combinations of reservoir properties involving either (kf
w)0.5 k0.25 or xf k0.5, as given in Eqs. 3.19b and 3-19d and All of the analysis procedures presented so far allow
shown in Figs. 3-3 and 3-4, respectively. one to analyze distinctive flow regimes such as radial or
formation linear flow. They do not allow one to analyze
CfD = 1.6 transitions to other flow regimes or the late transient
CfD > 1.6 flow period. The inability to pick an unambiguous
straight line for a hypothesized flow regime from the
plots was an additional frustration. To address both
m(pwf)
CfD < 1.6 of these issues, Ramey (1970) introduced the idea of
type curves. Gringarten et al. (1979) further extended
Slope Proportional to the type curve concept with a judicious choice of
1/((kf w)0.5k0.25) variables to be used in the matching process. The
well test equations were re-formulated in terms of a
ta0.25
pressure difference defined as:
Figure 3-3 Specialized bilinear flow plot (m(pwf) versus
ta0.25) for the determination of (kfw)0.5 k0.25

m( p ) = m( pwf , ta = 0) m( pwf , t a ) , (3-21)
where the vertical bars represent the absolute value

and m(pwf , ta = 0) represents the pseudo-pressure
End of Linear Flow at the beginning of the appropriate flow period (or
m(pwf)
build-up). Fig. 3-5 illustrates the concept for both a

drawdown and build-up.
pi
Slope Proportional to
1/(xf k0.5)
ta0.5
%p
Figure 3-4 Specialized linear flow plot (m(pwf) versus ta0.5) %p
for the determination of xf k0.5
pwf
Without radial flow, where permeability can be
uniquely determined, no decoupling of these terms is tp %t
possible. For bilinear flow there is no information with
q1 q2 = 0
respect to fracture length; in this case, length can only be
determined indirectly by history matching. Figure 3-5 Drawdown and build-up plots showing
definition of p and, by analogy, m(p) for each case
For multi-rate tests, one simply ignores the effect
of all previous flow periods for these specialized
plots, plotting variables in terms of ta, ta0.25 or The distinguishing feature of any type curve is that
ta0.5 as appropriate. In the real world, determining both axes are plotted in log coordinates. All reservoir
the correct straight line (or determining whether a properties can be directly obtained from an overlay
straight line even exists) is difficult, just as in the case match. For build-ups, there are problems in developing
of the generalized Horner plot. The MDH plot, in type curves because they depend on the length of time
particular, is notorious for showing false straight lines, of the drawdown. Empirically it was shown that if the
or in some situations where radial flow is actually flow period is long, then the build-up matches the
occurring, no straight line at all. drawdown type curve, until bounded flow is observed.
49
Modern Fracturing
Agarwal (1980), assuming transient radial flow, applied capability at late times and large m(p) is poor,
superposition consistent with the definition of m(p) making it almost impossible to match different
in Eq. 3-21. A time function called equivalent pseudo- middle to late time flow regimes.
time, tae, was derived to handle these arbitrary flow In 1983, the biggest breakthrough in PTA occurred
rate variations. This allowed the creation of an adjusted with the introduction of the logarithmic pressure derivative
drawdown type curve for all flow periods. For radial technique of Bourdet et al. (1983). This coincided with
flow, the solution for any flow period is: the industry adaptation of high-resolution electronic
gauges capable of a virtual avalanche of recorded data
values. Because numerical differentiation of data is
inherently a dispersive process, the pressure derivative
could not be successfully applied to older, less accurate,
pressure measurements. The logarithmic derivative (often
(3-22) just called the derivative) for gas wells is defined as:
d (m( p )) d (m( p ))
where qn represents the flow rate for flow period m( p ) =

d (ln tae )
= tae
d tae (3-24)
n. Equation 3-22 is analogous to Eq. 3-8. Similar
analogous expressions for wellbore storage, bilinear and For various flow regimes this results in:
radial flow were developed as:
m( p ) 1637T (qn1 qn )
= a1tae n , (3-23a) m( p ) = (radial flow), (3-25a)
qn1 qn 2.303kh
or
0.420082T
a1 = (n = 1), (3-23b)
i C
m( p ) = nm( p ), (3-25b)
443.3581qT
a1 = (n = 0.25), (3-23c) where n = 1 for wellbore storage, n = 0.25 for bilinear
h 0.25 ( ct )i0.25 (k f w)0.5 k 0.25
flow and n = 0.5 for linear flow.
and For build-ups qn = 0. The standard plot is a combined
Gringarten/derivative type curve alternatively called the
40.9521
a1 = (n = 0.5), (3-23d) log-log derivative type curve and is shown in Fig. 3-6.
h ( ct )i0.5 x f k 0.5
0.5
102
where n is 1.0 for wellbore storage, 0.25 for bilinear CDe2S 1060
flow and 0.5 for formation linear flow, respectively. 1060 1020
1010
Different equivalent pseudo-time expressions, which
pD and (tD/CD) pDa
101 1020 104

are flow regime-specific, could be developed for these 1010
102
3 x 10-1
last three flow regimes. In practice most software uses 10 4
the radial equivalent pseudo-time definition for all 1 102
cases because it typically does not appear to make much CDe2S 3 x 10

-1
difference. Even when the condition of transience is

10-1 -1
violated as often happens in PTA the equivalent 10 1 10 102 103 104
tD/CD
pseudo-time function is still used.
The merits of type curve matching at early Figure 3-6 Radial flow with wellbore storage type curve
(after Bourdet et al., 1983)
times and small m(p) are obvious. The log-log
plot magnifies the data in this region. In practice At early times, while in wellbore storage, m(p)
the transition from wellbore storage to radial flow has unit slope. There is then a transition into pure
could more or less be determined. The resolution radial flow, which is different for various skin values.
50
Once radial flow is achieved m(p) is constant and In practice one can use t, defined as the time from
is inversely proportional to permeability. At times the end of the previous flow period, instead of ta as an
beyond those shown in the figure, the character would x-axis plotting variable for the majority of cases.
change depending upon whether the test was a build- Derivative plots are useful primarily for build-
up or a drawdown. For build-ups m(p) approaches ups because even very small changes in production
zero, while a drawdown approaches pseudo-steady rate destroy the character of the derivative plot in
conditions and m(p) has unit slope, much like the drawdown periods. Another useful definition is
wellbore storage curve. the normalized derivative:
For the other three flow regimes referenced in Eq. m( p )
3-25b, the derivative curves are parallel to the m(p) m( p ) q =
, (3-26a)
qn1 qn
curve and displaced by n. In Section 3-12.1, Fig. 3-
17 illustrates the concept for a finite conductivity or
fracture. Fig. 3-17 also shows how skin affects a m( p ) m( p )

m( p) q = or

finite-conductivity fracture. This skin can be either as qn1 q
(3-26b)
a result of a choked fracture or fluid damage. Note the
flattening of the slope of the derivative curve coming for build-ups, qn = 0, and qn-1 is the flow rate before the
out of wellbore storage, before bilinear flow. build-up. For the previously identified flow regimes
The process of computing the derivative the normalized derivative is independent of rate. For
is clear; derivatives are taken with respect to complicated multi-rate tests with a number of build-
equivalent pseudo-time, tae. What time function ups, overlaying the various normalized derivatives for
is used when plotting log-log derivative type curves? all the build-ups in a single plot is a necessary quality
Anderson et al. (2002) showed that the correct control step (Fig. 3-8). Differences in permeability from
time function to plot on the x-axis is the delta one build-up to the next may be a result of reported rate
pseudo-time, ta. The equivalent pseudo-time, tae, errors. Changes in skin may be an indicator of non-Darcy
should not be used. This ensures that the shape effects. For hydraulically fractured wells any changes in
of the build-up type curve is as close as possible properties may be a result of post-fracture cleanup.
to the drawdown type. Figure 3-7 shows how a
1000
hydraulically fractured well with finite conductivity
( slope) may be misinterpreted as an infinite
q1 > q 2 > q3
conductivity fracture ( slope) when the wrong time 100
function is used on the x-axis plot.
%m(p)a/q
q1
q2
1000 10 q3
%tae
%ta
Correct Stabilized Derivative Slope
100 1
mIf = 0.0001 0.001 0.01 0.1 1 10
%m(p)a
%ta
mbf =
Figure 3-8 For a multi-rate test, overlay of rate
10 normalized derivatives for different build-ups
mc = 1
3-9 Flow Regime Identification
1
0.001 0.01 0.1 1 10 100 1000
Time The log-log derivative type curve is the flow
regime identification/diagnostic tool for PTA. It
Figure 3-7 Possible misinterpretation of finite
conductivity fracture ( slope) as infinite conductivity
is uniquely capable of identifying all flow regimes
fracture ( slope) when the wrong time function is used from early time through the late transient period
51
Modern Fracturing
into boundary-dominated flow. The plot can These type curves are more or less appropriate
also be used as a property calculator after the for build-ups if the procedure for calculating and
appropriate lines for a flow regime are overlaid on plotting them (outlined in Section 3-8) is followed.
the plot. Alternatively or in addition, superposition In boundary-dominated flow the build-up type curves
plots or specialized plots can be used as property significantly differ from the drawdown type curves,
calculators after identifying from log-log but this behavior is readily identified. A summary of
derivative type curves the appropriate time ranges the most significant derivative responses is given by
over which they apply. Economides et al. (1994) and Economides and Nolte
A library of derivative responses for various types (2000), and represents a generalization of the material
of reservoir models and flow regimes is shown in Fig. presented in Section 3-8.
3-9 (The annotated letter below correspond to the The late transient flow regime is a particularly
graphs in Fig. 3-9). interesting topic. Except for the case of a single bounding
fault, this flow period was not analyzable before the
advent of the log-log derivative type curve. The late
1. Early Transient Time
transient period is characterized by irregularities in the
Wellbore storage: 1 slope A derivative. Because of the present emphasis on matching
the last build-up in a test, the current generation of
Spherical flow: slope B
analysts has gone to great lengths to match these irregular
Infinite-conductivity derivative responses. This is helped by powerful non-
slope C
fracture:
linear regression algorithms, which are part of most
Finite-conductivity current software systems.
slope D
fracture: To match the late transient time, one of two models
2. Middle Transient Time is typically chosen: the non-centered well in a rectangle
or the multi-composite radial model. A maximum of
Radial flow: 0 slope A
three boundaries can be felt at varying times in the
3. Late Transient Time non-centered well model before bounded flow occurs.
0 slope line (stabi- Each boundary gives half the value of the apparent
lizes at double the permeability on the derivative plot, by the equivalent
Single bounding fault: E
value of a previous of doubling the slope on the Horner plot (Earlougher,
0 slope line 1977). This model can only work if the derivative is
increasing with time at the end of the test.
Well producing from a monotonically in-
The radial multi-composite model is less restrictive.
corner of a rectangle: creasing derivative
One can put radial rings of different diffusivity (normally
Composite by changing permeability) and storativity. With this
wavy derivative
permeability: model a match of any derivative signature is possible,
dip and then climb whether the derivative goes up or down. Are such models
Dual porosity: H to be believed, or is this just a mathematical exercise? It
in derivative
is clear that in most circumstances the analyst is only
Channel flow (long
slope G accomplishing the latter. Kelly (1996) describes the
narrow rectangle):
absurdity of this approach to PTA matching, where the
4. Bounded Flow geological framework is completely ignored.
Drawdown: 1 slope Dake (1994) recounts a story where a number of
software vendors were asked to present solutions to well
derivative falls
test problems at a comparative solution forum. These
Build-up: downward off the F
problems exhibited a variety of late transient effects.
plot
Unfortunately the vendors were given only the pressure
52
Log-Log
Diagnostic Homer Plot Specialized Plot
10.0 1.00
Wellbore Storage
101
Infinite-acting
A 100
5.0 0.50 radial flow
C From specialized plot
10-1 -1 0.0 0.00

10 100 101 102 103 100 101 102 103 104 105 106 0.0 0.50 1.00
%ta (tpa%ta) %ta %ta
102 8.0
101 Wellbore Storage

Partial penetration
B 10 0 4.0
Infinite-acting
10 -1
radial flow
10-2 0.0
100 101 102 103 104 105 100 101 102 103 104 105 106
%ta (tpa%ta) %ta
4.0 0.50
101
Linear flow to an
100
infinite-conductivity
C 2.0 0.25 vertical fracture
10-1 xfk0.5 From specialized plot
10-2 -4 0.0 0.00

10 10-3 10-2 10-1 100 101 100 101 102 103 104 105 0.0 0.01 0.02
%ta (tpa%ta) %ta %ta0.5
101 3.0 0.50
10 0 Bilinear flow to a
finite-conductivity
D 10 -1 1.5 0.25 vertical fracture
10-2 (kfw)0.5 k0.25 From specialized plot
10-3 0.0 0.00

10-5 10-4 10-3 10-2 10-1 100 101 100 101 102 103 104 105 106 0.0 0.25 0.50
102 10.0
Wellbore storage
101
Infinite-acting
E 100
5.0
radial flow
Sealing fault
10-1
0.0
101 102 103 104 105 106 100 101 102 103 104 105
%ta (tpa%ta) %ta
102 10.0
101
Wellbore storage
F 100 5.0
No-flow boundary
10-1
10-2 -5 0.0
10 10-4 10-3 10-2 10-1 100 101 102 103 104 105
%ta (tpa%ta) %ta
40
60
102
Wellbore storage
101 Linear channel flow
G 20 30
100 bk 0.5
From specialized plot
10-1 -5 0 0
10 10-4 10-3 10-2 10-1 100 100 101 102 103 104 105 106 10-1 10-1 10-1 10-1 10-1 10-1 10-1
10.0
10 2
101 Wellbore storage

H 10 0 5.0 Dual-porosity matrix
to fissure fow
10-1
(pseudosteady state)
10-2 0.0
10-1 100 101 102 103 104 105 100 101 102 103 104 105 106
%ta (tpa%ta) %ta
Pressure Slopes Data Points
Figure 3-9 Derivative responses for various types of reservoir models and flow regimes (from Economides et al., 1994)
53
Modern Fracturing
data without the framework of a geological setting. Not build-up. What are the consequences of simplifying
surprisingly, a wide variety of incompatible solutions rate variation in a PTA? Gringarten (2006) has studied
were presented. PTA cannot exist in isolation from a this problem. Figure 3-10 shows an example of the
geological model because uniqueness issues inherently derivative of the last build-up for various simplifications
plague the analysis process. After a PTA is performed, it of the actual rate history. This shape might mistakenly
has to be revisited in the terms of the actual production be interpreted as boundary effects. The bad news is that
performance of the well. Too often reports are simply the distortion occurs primarily at late time in the build-
filed away and never looked at again. If the analyst is up, a non-intuitive finding. Rate errors particularly
truly interested in reservoir characterization the occur for post-fracture tests that are conducted shortly
discipline of integrating various disparate information after treatment because cleanup of fracturing fluid
sources everything must be subject to continuous may go unreported or be considerably simplified.
review as more information becomes available. Only in The effect of oversimplification on the pressure
this way will the industry have a hope of generating derivative also holds for the Horner, superposition and
better production forecasts. specialized plots. The current paradigm in well test
analysis is to match every variation in the derivative
3-10 Derivatives A Few Cautionary of the last build-up to a high level of accuracy,
Remarks while completely ignoring the drawdown. This is
fundamentally flawed, and more attention must be
Log-log derivative plots have truly revolutionized PTA. paid to the drawdown both in terms of describing
However, the analyst needs to be aware of a number of its rate variation and in matching the pressure. PTAs
issues when applying the technique. major objective is to provide a production forecast,
Although the derivative plot significantly reduces which involves a drawdown. To achieve higher-quality
the uncertainty of flow regime identification, it can also matches, a linear pressure plot versus time throughout
generate false information in two circumstances: the whole test has to be used as a guideline to any
1. If a drawdown is terminated in the middle of a late match. This can help resolve inevitable uniqueness
transient flow regime, the subsequent build-up type issues in matching field data.
curve will have a different shape than the drawdown
1000
type curve. This is usually not significant but can
result in either a false diagnostic or an inappropriate
reservoir property value being calculated. 45% Error in
%m(p) and %m(p)a
100 Produced Volume in

2. A more serious problem is the effect of rate variation the Rate History.
on the build-up type curve. Traditionally in PTA
the rate profile variation has been de-emphasized
Correct
because the rate information has historically been 10
Stabilzation Line
of poor quality. As a result, since the time of the %m(p)
%m(p)a
introduction of the Horner plot, the well test Model
community has adjusted producing time by the 1
0.001 0.01 0.1 1 10 100 1000
known cumulative production divided by the %ta
last measured rate, giving one big flow period. Figure 3-10 Example of the derivative response of the last
Additionally, the pressure data during the drawdown build-up for various simplifications of the actual rate history
at the best of times shows scatter, and the derivative
technique fails during these flow periods. The dispersive nature of taking pressure derivatives
has already been mentioned. Each commercial software
As a result, industry practice is to worry about only package makes default decisions about how derivatives
one pressure in the drawdown period the last point will be calculated for observed pressure. The software
before shut-in and match exclusively on the last vendor is in a quandary as to the degree of smoothing
54
that is applied. With too little smoothing, the plots look When one is within of the end of the flow period,
bad from an aesthetic point of view; with too much, the calculation uses only 2 points because it cannot look
some artifacts of the smoothing algorithm may look like upstream. It is too dangerous to use only the last point
real reservoir effects. in the flow period. Therefore there is a loss of accuracy
In most cases the user will have some control at the end. Trying to attribute reservoir affects to these
over how the derivative is calculated (i.e. the amount last few points in time is not recommended.
of smoothing that will take place). Therefore, it is As a start, it is recommended that raw unfiltered
imperative that the analyst understand the ideas behind data be used and be made equal to the smallest value
these algorithms. The simplest and, perhaps best, way permissible. This will force the use of successive points
of calculating derivatives was presented by Bourdet et for derivative calculations. This will create maximum
al. (1987). It is a 3-point calculation, and as the name dispersion of the derivative (minimal smoothing).
implies, three values are used. The central point is at Bands of unit slope emanating into a widening cloud
the time value in question. The other two points are (trumpeting) will result in some cases, as seen in a
selected at a set distance (delta or ) in each direction comparison of Figs. 3-12 and 3-11. This is an indication
(one earlier and one later in time). Because the plots are that either limits on gauge resolution or the sampling
logarithmic, the distance is a set fraction of a log cycle. algorithm used for reporting data from the gauge are
This can be thought of as the window for the derivative being met: The pressure data is being gauged off.
calculation. This has been called the Bourdet derivative Careful consideration is required before attributing
or "moving window" derivative and is available in most any fluctuations in the derivative plot after this time to
software. By default most vendors give a of 0.1 (1/10 reservoir effects, regardless of the smoothing algorithm
of a log cycle) as shown in Fig. 3-11 one ultimately chooses for presentation purposes.
There is a danger that the smoothing algorithm may
introduce artifacts (Mattar, 1996).
After resolving the issue of which data to use, the
analyst will inevitably see variations in the derivative
curve. The question the analyst faces is what should be
matched and what can safely be ignored? Brown and
Hawkes (2005) addressed this issue: If a derivative
transition concludes in less than one-half a log cycle,
it is not a reservoir effect. Transition is defined as a
change from one flow regime to another.
On a related manner, Mattar and Zaoral (1992)
Figure 3-11 Consequences of using a derivative
smoothing window of 0.1 have studied the issue of distinguishing between
wellbore effects and reservoir effects. Due to the
nature of the diffusivity equation it is intuitively
obvious that after a rate change there will be an
associated pressure change, the effect of which will
diminish with time. Mathematically this is equivalent
to stating that the primary pressure derivative with
respect to time (or the PPD), defined as dp/dt, must
decrease with time for a build-up. If it increases with
time at any point, then this is not a reservoir effect
and is instead related to something going on in the
wellbore. One should not attempt to match non-
Figure 3-12 Consequences of using a derivative reservoir effects, and as a precaution such data should
smoothing window of 0.0001 when the data is at the be removed from the observed data. Unfortunately
limits of the gauge resolution
55
Modern Fracturing
non-reservoir effects cannot be seen in the log-log One common error in well test interpretation
derivative. As a result, a PPD diagnostic plot should is to indiscriminately draw a straight line through
be plotted for all build-up flow periods at the pre- the last few points of the Horner plot to determine
screening stage of the interpretation. k and m(p)*. Theory only supports extrapolation in
the time range where radial flow is identified. In the
3-11 PTA Interpretation Methods case of a hydraulically fractured well, where clearly
defined bilinear or formation linear flow occurs at the
One might think that with proper implementation of end, extrapolation is not permitted. Each of these flow
the log-log derivative type curve, all other plots would be regimes occurs before radial flow is achieved. However,
rendered obsolete. This is not the case. The determination extrapolation is often done to see what k and m(p)*
of initial pressure for early tests still requires the would be if radial flow occurred at the instant the test
generalized Horner plot extrapolation to m(p)*. Issues was finished. This is a bounding case where k would
related to where to draw the straight line on a Horner be at a maximum and m(p)* would be at a minimum.
plot are resolved by determining the times associated When late transient effects are occurring, as in Fig. 3-
with radial flow from the log-log derivative plot. This 14, extrapolating the last few points of the Horner plot
is shown by comparing both Figs. 3-13 and 3-14. The is completely unjustified.
log-log derivative plot has therefore resolved Horner plot Another reason that log-log pressure derivative
ambiguity that plagued early analysts. plots cannot be used exclusively for matching is that
for large values of m(p) on the log-log derivative
1000
plot, the m(p) values are highly compressed. The
m(p) values are also not so sensitive to an absolute
error difference in the calculated pressure. Therefore
100 plots where pressure is on a linear scale on the y-axis
%m(p) and %m(p)a
must be included in the matching process. Specialized

Correct Semilog Slope plots including the generalized Horner plot are also
acceptable, although interpreting straight lines on
10
specialized plots can be fraught with danger.
It is recommended that four plots be included in
%m(p)
%m(p)a any match:
1
0.0001 0.001 0.01 0.1 1 10 100
1. Linear plot of pressures versus time for all flow
%ta periods
Figure 3-13 Onset of semi-logarithmic (Horner) straight 2. Log-log derivative plot
line identified on log-log derivative plot 3. Superposition (or Horner) plot for radial flow
700
4. Specialized plot (depending upon the flow regime)
650 At this point the reader may feel skeptical of

Cor
600
rec
t Sem
everything that has been discussed. The authors have
ilog
Slo
pe
presented some complex mathematics and a number
m(p)
550 of plotting techniques, all of which appear to be

500
flawed. It seems to be impossible to verify whether
the underlying simplifying assumptions are correct or
450
not. If this were the mid 1970s, as Gringarten (2006)
400 implies, the reader would have a right to feel this way.
1 10 100 1000 10000 100000 1000000
By their very nature, straight-line analyses based on the
Horner Time
recommended plots may lead to incorrect flow regime
Figure 3-14 Correct semi-logarithmic straight line on
Horner plot as identified by log-log plot identification with false reservoir properties.
56
Things are not as bad as they seem. The following include a rate term. If a flow and build-up test were
steps are recommended and form the basis of all current performed on a high- and low-permeability well with
interpretation techniques: flow periods of the same duration, with otherwise
1. Perform flow regime identification via the log-log identical properties, the depth of investigation into
derivative plot and calculate reservoir properties the reservoir would be much smaller for the low-
from any plot as if the appropriate underlying permeability case. Of course for the low-permeability
assumptions are correct. well the absolute drawdown at the well would be
2. Use forward simulation to calculate a pressure higher if the rates were the same in the corresponding
profile. In the forward simulation step all tests. As a result, low-permeability reservoirs will take
unknowns have been defined. A rigorous pressure much longer to achieve radial flow. For this reason,
versus time calculation is made based on the hydraulically fractured low-permeability wells rarely
selected flow model and rate superposition. exhibit radial flow. The amount of time allocated for
3. Overlay the calculated pressure profile on the testing will simply not allow it to happen. The lack of
various plots. If the underlying assumptions a distinct period of radial flow leads to interpretation
used in the straight-line analysis are correct, the difficulties, which in turn directly affects the analysts
computed pressure will overlay the straight-line ability to generate accurate production profiles, or
analysis. This in turn will overlay the observed indeed optimize any future wells treatment.
pressure data. Even in the best of circumstances,
there will be some mismatch. 3-12.1 High-Permeability Wells
4. If one is close, then fine-tune the selected model
by either manually adjusting reservoir properties High-permeability gas wells will flow even when
or by using non-linear regression. heavily damaged. As a result it is possible to do a proper
5. If the flow regime is completely off, the calculated pre-treatment test, so that fracture optimization
pressure profile may be significantly different than according to the principles outlined in Section 4-5
anticipated. One then needs to select another can be performed. When high-permeability wells
plausible flow model and repeat the process. have been tested before fracturing, one is typically
looking for an estimate of permeability, skin and
A natural consequence of this procedure is reservoir pressure. Techniques outlined in Section 3-
that any incorrect assumptions will be picked up by 9 work for most cases because radial flow normally
inconsistencies in the resulting forward simulation occurs quite quickly. For wells exhibiting classic
pressure match. The process has a feedback loop, which wellbore storage, positive skin and radial flow, the
is self-checking, and alleviates Dakes (1994) previously match is done with a combination log-log derivative
mentioned objections. type curve and the superposition plot.
Complications occur only when non-Darcy effects
3-12 Difference Between High and Low become large (see Section 3-13) or late transient effects
Permeability Analysis Techniques complicate the derivative interpretation. For the latter
case, instead of trying to reconcile all of the characteristics
Permeability controls two aspects of reservoir in the build-ups derivative, it is recommended to keep
response: The first aspect is the rate at which a well the models fairly simple and develop a reasonable
can be produced. The higher the permeability, the range of reservoir properties. Reservoir drainage areas,
higher the production rate will be. The second aspect shapes and possibly heterogeneities should be estimated
is the velocity at which the pressure signal travels based on offset well production performance data
through the reservoir. Velocity is proportional to when available, and not on late time characteristics in
the diffusivity coefficient given in Eq. 3-7a, which the derivative plot of the subject well. In the authors
also shows velocity is proportional to permeability. experience, long-term production profiles rarely match
It is important to note that Eq. 3-7a does not early complex PTA-derived models.
57
Modern Fracturing
In other cases a post-fracture well test is performed. particular match value of value for m(p) and ta
For high-permeability wells, radial flow will likely be to compute xf k0.5). Using the permeability from
evident at the end of the build-up. The interpretation of Step 1, calculate xf.
such a test is relatively straightforward for an infinite- or 4. Compute the wellbore storage coefficient from
finite-conductivity fracture. The log-log plot in Fig. 3- early time data.
15 and the corresponding superposition plot in Fig. 3-16 5. Perform forward simulation with nonlinear
illustrate the case for an infinite-conductivity fracture. regression to improve the match.
1000
%m(p)
The case for a finite-conductivity fracture is
%m(p)a illustrated in Fig. 3-17 and is slightly different:
Step 1
1. Determine permeability from the derivative or
%m(p) and %m(p)a
100
superposition plot.
mrf= 0
2. Calculate m(p)* from an extrapolation of the
n= superposition plot.
10
Step 4 3. Compute the permeability/fracture conductivity
mc = 1 Step 3: From term (kf w)0.5k0.25 from the derivative plot by
Position of Slope
mlf= Line Determine xf k0.5 overlaying a quarter-slope interpretation line on the
1
derivative curve. (Use Eqs. 3-23a, 3-23c and 3-25b
0.0001 0.001 0.01 0.1 1 10 100 1000
%ta at a particular match value of value for m(p) and
Figure 3-15 Log-log analysis procedure for high- ta to compute (kf w)0.5k0.25).Using the permeability
permeability, infinite-conductivity fracture from Step 1 calculate kf w.
1000 4. Make an initial estimate of xf based upon
m(p)* gives m(pi)
side information.
900
Steps 1 and 2
5. Compute the wellbore storage coefficient from
800 early time data.
700 6. Perform forward simulation with nonlinear
m(p)
Slope Inversely
Proportional to kh regression, on amongst other variables, xf, to
600
improve the match.
500
1000
400
300 The Early Time "Hump"

100 101 102 103 104 105 106 107 in the Derivative is an Indicator Step 1
100 of Fracture Face Skin.
%m(p) and %m(p)a
Horner Time
n= mrf= 0
Figure 3-16 Superposition plot procedure for high- Step 3: From
permeability, infinite-conductivity fracture Position of
10 Step 5 mbf= Slope Line
Determine
The interpretation process for an infinite mc = 1 (kfw)0.5k0.25
conductivity fracture is as follows: %m(p)

%m(p)a
1. Determine permeability from the derivative or 1
0.0001 0.001 0.01 0.1 1 10 100 1000
superposition plot. %ta
2. Calculate m(p)* from an extrapolation of the Figure 3-17 Log-log analysis procedure for low-
superposition plot. permeability, finite-conductivity fracture
3. Compute the permeability/fracture half-length
term xf k0.5 from the derivative plot by overlaying Six months or so after the well has been stimulated
a half-slope interpretation line on the derivative and put on production, the well tests should be reviewed
curve. (Use Eqs. 3-23a, 3-23d and 3-25b at a in light of actual production data.
58
3-12.2 Low-Permeability Wells Pre-Treatment Open-hole wireline tests, commonly referred to

Evaluation as repeat formation tests (RFTs), are another way of
obtaining reservoir pressure and permeability. However
Low-permeability (tight) gas wells normally do not flow RFTs have proved problematic in low-permeability
before being hydraulically fractured. Additionally, post- gas reservoirs because of the possibility of pressure
treatment evaluation of low-permeability wells can be supercharging. This occurs when excess pressure is
difficult to analyze as will be discussed in Sections 3-12.5 locked into the formation as a result of the drilling
and 3-12.6. To circumvent this problem one can run process. Low reservoir permeability prevents the excess
pre-fracture tests that do not require flow to the surface. pressure from bleeding off into the formation. Because
Alternatively, mini-frac tests can be performed. They will the container volume of the RFT tool is small, one
but will be addressed separately in Section 3-17. may not be able to get beyond the supercharged region
The types of tests that do not require flow to during the test. Obviously interpreting such a test for pi
surface are Perforation Inflow Diagnostic (PID) tests, may lead to wrong or ambiguous values. In cased-hole
impulse tests, back-surge tests and closed chamber CCT, the volume of the tubing/wellbore combination
tests (CCT). These are all similar in methodology. is on the order of barrels, and the test data sees well
This section will concentrate on PID testing, a beyond any supercharged zone.
subset of general CCT technology. The procedure for The modern interpretation process for closed
measurements during perforating testing is: chamber testing was introduced to industry by
1. Remove liquids and gases from the wellbore Alexander (1977). Rate, although not measured, is
in order to create underbalanced conditions calculated from the wellbore pressure using the wellbore
before perforating. compressibility equation:
2. Perforate the well.
dp
3. Record surface and/or subsurface pressures in = Vwcw
q . (3-27)
dt
a closed-chamber environment continuously
throughout the wellbore fill-up period. This is where Vw is the wellbore storage and cw is the
accomplished by having the surface valve closed compressibility of the wellbore fluid. The term Vw
throughout the test. may change with time as a result of an intervention,
such as placement of a plug that changes the wellbore
The range of tool configurations is wide and can be as volume. The rate typically declines rapidly, with a
elaborate as a tubing-conveyed perforating (TCP) system possible early time plateau if sonic velocity occurs
using a drill stem test (DST) string through to a rigless through the perforations.
operation with an E-line perforating truck and surface For analysis purposes the reservoir is assumed
pressure recorders. Production rate into the wellbore is to always be in transient radial flow. Rahman et al.
not measured. The PID test can be repeated up-hole as (2005) have solved this problem for all times, when
necessary with a plug being placed between successive the wellbore volume is constant. Early and late time
perforation events, possibly with a recorder below the solutions are respectively:
plug to monitor earlier perforation events. Alternatively,
because the wells are usually producing at such a low rate, kh(m( pi ) m( pwf 0 ))ta
m( pwf ) = m( pwf 0 ) +
the wellhead valve can be re-opened before the second 603.63Vw (ct )i S
perforating event and the well blown down. Then one
has commingled flow from the second perforating event (early time), (3-28a)
and the still-flowing first event to interpret. If at any time
the build-up in pressure is occurring too slowly for the 301.82Vw (ct )i (m( pi ) m( p f w0 ))
m( pwf ) = m( pi ) +
time allotted for the test, various plugs can be placed in khta
the tubing via wireline to reduce the tubing volume and
correspondingly reduce the wellbore storage. (late time), (3-28b)
59
Modern Fracturing
where m(pwf0) is the initial pseudo-pressure of the skin, which contains the true skin and permeability.
wellbore. The logarithmic derivative of both of these Equation 3-30 has the form of a straight-line y
functions can be calculated from Eq. 3-25b. The early = b m x when permeability and skin have been
time solution, Eq. 3-28a is identical in form to the estimated properly. The right side of Eq. 3-31a is
standard wellbore storage equation and will plot as a the rate convolution build-up time function and is
unit-slope line on a log-log derivative plot. The late time plotted as the x-axis with the left side of the equation
solution from Eq. 3-28b will have a 1-slope on the as the y-axis The slope m is inversely proportional
same log-log derivative plot. Permeability can be directly to formation flow capacity, kh. For a full PID
determined from this plot after the 1-slope radial flow evaluation, four plots are generated, each of which
regime has been identified. A hump on the derivative has diagnostic and calculation capabilities. These will
plot is an indicator of skin damage, analogous to that on be reviewed in Example 3-1.
a PTA derivative plot. Rahman et al. (2005) use the so-
called impulse derivative of Kuchuk (1999) to develop a 3-12.3 Example 3-1, PID Test
slightly different form of the log-log derivative plot. The
impulse derivative is defined as: This shallow gas well was perforated underbalanced (dry),
and pressures were obtained using only surface recorders
d (m( p ))
m( p )im = ta2 . (3-29) for 16 hours. Surface pressures were converted to sandface
dta
conditions using the wellbore and gas properties shown
In their approach, wellbore storage is represented by a line in Table 3-2. Figure 3-18 is the interpreted pwf and
of slope of 2, and the radial flow regime has a slope of 0. calculated gas rate response during the fill-up period. The
The late time solution from Eq. 3-28b is of the log-log derivative plot is shown in Fig. 3-19.
same form as one given by Soliman (1986). The impulse
solution is the limiting case when the flow time is small Table 3-2 Wellbore and Reservoir Properties for
compared with the build-up time. A plot of m(pwf) Example 3-1
versus 1/ta will be a straight line when radial flow is Gas Specific Gravity 0.56
occurring. Extrapolation to 1/ta = 0 will give the initial Reservoir Temperature, F 72
pressure. To handle the most general case of changing Expected Reservoir Pressure, psi 690
wellbore conditions, such as would occur if a plug were Casing Volume, ft 3
387
lowered in the hole during the test, one requires rate Perforation Interval, ft TVD 2086 - 2101
convolution (superposition). Based upon the work Shots Per Foot 5
of Simmons (1986) and Meunier et al. (1985), the
working equation for gas then becomes: Radial flow begins at 1.65 hours. If a hump in the
derivative occurs, indicated by the region having a slope

log (ta ) +
m( pwf ) = m( pi ) m n + q (ta ) s ,
(3-30) steeper than 1, there is positive skin. This does not
q ref q ref
happen in Fig. 3-19. The m(pwf) plot in Fig. 3-20 allows
where qref represents the calculated rate at time zero via the calculation of an initial pressure by extrapolating to a
Eq. 3-27, immediately after the closed chamber test 1/ta of 0. The rate convolution plot used for calculating
is initiated, m
is a slope term and s is an apparent permeability is given in Fig. 3-21.
n2 q
j +1 q j
= ((t an t aj ) log (t an t aj ) (t an t a ( j +1)) log (t an t a ( j +1)))

t j +1 t j
j =1
n

+ 0.434 (q 0 q n ) + (q n q n1) log (t an t a ( n1))
(3-31a)
and
k
s = 1.151 s log
2
+ 2.09 .
(3-31b)
(ct )i rw
60
500 200 Experience with using these tests indicates that

450 180
400 160
the following guidelines should be applied:
350 140 1. CCTs are particularly suited to low-permeability
300 120 (k 10 md) wells because of the relatively
q, Mscf/d
Pressure
pwf, psi
CCT Rate
250 100
small wellbore volume. Performing the test in
200 80
150 60
wells with higher permeability will nevertheless
100 40 provide an excellent estimate of reservoir
50 20 pressure because the pressure builds up rapidly,
0 0
0 2 4 6 8 10 12 14 16 overcoming wellbore storage effects.
Time Since Perforating, hrs
2. In ultra low-permeability wells (k 0.01 md),
Figure 3-18 For Example 3-1, observed pressure data downhole shut-in tools may be required to
and calculated gas rate response during fill-up
achieve practical testing times.
100
3. Confidence levels on permeability estimates
can be low on some ultra low-permeability
10 Approximate reservoirs due to such factors as limited
%m(p) and %m(p)a
Start of Radial
Flow @ 1.65 hrs entry perforating and perforation efficiency.
1
Confidence level for reservoir pressure, on the
mrf = -1
other hand, remains high.
0.1 4. Typically CCTs are run underbalanced to allow
%m(p)
%m(p)a
perforation clean-up and often result in negative
0.01
0.001 0.01 0.1 1 10 100
skin effects (see Section 6-4.3).
ta
5. In overbalanced wells (pore pressure gradient
Figure 3-19 Log-log derivative plot for Example 3-1 is greater than hydrostatic gradient of wellbore
fluid) surface pressures recorders typically
2.5
provide excellent results.
m(p)*gives p*=478 psi 6. A pressure rate of change of 7 psi/day can be
2.0
used as End of Test for field engineers when
Approximate Start of
diagnostic log-log plots are not utilized to ensure
m(p) x 107
Radial Flow @ 1.65 hrs

1.5
the end of storage dominated flow.
1.0
7. Because the skin term in Eq. 3-31b is a
function of both skin and permeability,
0.5
several iterations may be necessary to calculate
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
1/ta
the formation properties. Typically an iterated
skin range of 2.5 will provide a satisfactory
Figure 3-20 For Example 3-1, an inverse time plot
straight line when using the rate convolution
2.5 plotting technique.
s = - 0.2
2.0 s = +1.2
s = +2.2 3-12.4 Low-Permeability Wells
Match Properties Post-Treatment Evaluation
m(p) x 107
1.5
kh = 1.8 md-ft
s = +1.2
1.0 Interpretation difficulties for hydraulically fractured
wells primarily arise because radial flow does not
0.5
occur by the end of the test. This happens in the
0 majority of lower-permeability wells. Many low-
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Rate Convolution Time permeability gas wells are tested immediately after
Figure 3-21 For Example 3-1, a rate convolution time being stimulated because it is a convenient time
plot
61
Modern Fracturing
and regulatory authorities often require that all gas value for permeability. It is the equivalent of
wells be tested. It is convenient because the pipeline assuming that the well immediately goes into
is probably still being built, and the well cannot radial flow at the end of the test. They also
produce anyway. An interpreter must be aware that conclude that if initial pressure is known from an
such wells may not have sufficiently cleaned up, and independent source, a correct interpretation of
the assumption of single-phase flow may be violated. reservoir properties is possible. Hategan and Hawkes
Monitoring of the liquid production rate under such (2006) showed that permeability variation from
circumstances is critical. This topic will be reviewed one plausible interpretation to another caused the
in detail in Section 3-16.3. biggest difference in rate forecasts and hence reserves
Low-permeability wells are typically in estimates. This is extremely discouraging because
formation linear flow or bilinear flow at the end the value of any well test is greatly diminished if
of the test because it takes a very long time to meaningful forecasts cannot be generated.
get into radial flow. Operators are normally not
willing to run tests for the length of time required 3-12.5 Example 3-2, Low-Permeability Well,
to achieve radial flow. Because there is no radial Infinite-Conductivity Fracture
flow, permeability and initial reservoir pressure are
hard to determine (if a pretreatment test was not The goal of this example is to re-examine the Garcia
performed). Without initial pressure, it is difficult et al. (2006) conclusions with a comparable problem:
to decide the ideal well spacing for the field. The a well with the reservoir properties as shown in Table
trend in the industry is to drill low-permeability 3-3 and flow rate history shown in Table 3-4.
gas wells on lower and lower well spacing. At some
point, depleted areas are bound to be encountered, Table 3-3 Well Initial Properties for Example 3-2
and efficient drainage of reserves will not occur. Initial Pressure, pi, psia 3000
Pay Thickness, ft 30
Without the initial pressure, it is not possible to
Porosity, % 10
make rational well spacing decisions. Additionally, Water Saturation, % 25
without a reasonable value for permeability, initial Permeability (k), md 0.03
production forecasts will be significantly off, Propped Fracture Half-Length (xf), ft 150
CfD >30
diminishing any ability to diagnose the success or
rw, ft 0.3
failure of the actual treatment. xf k0.5, ft-md0.5 25.98
Garcia et al. (2006) assigned typical well CD 5000
and hydraulic fracture properties from a North
American gas well. They considered the case of an Table 3-4 Rate Schedule for Example 3-2
infinite-conductivity fracture. They then modeled Duration, Gas Rate,
Commentary
hours MMscf/d
a typical well test on such a well and tried to Cleanup post fracture
120 0.5
interpret the calculated results in the normal treatment
manner, i.e., they tried to determine k, xf and 2 0 Run gauges into well
Resume flow for a period 3
pi. At the end of the test the well test showed the length of the interruption.
formation linear flow with no indication of a 6 0.5
This is an application of S.C.
Swifts three times rule as
transition to radial flow. Their conclusion is described by Brown and
that one can get visually appealing matches of Hawkes (2005)
24 0 Build-up
the build-up in a variety of ways, leading to an
unacceptably broad range in values for k, xf and pi. If
radial superposition time is used (i.e. a Horner plot) The well properties and rate schedule are typical of a
and the last few points are used to perform an Western Canada low-permeability, post-stimulation gas
extrapolation, initial pressure is always under- well test. Forward simulation is performed to calculate
estimated. This will also give the largest possible BHP as a function of time (Fig. 3-22).
62
3000 3000 Preliminary values of initial pressure and

Pressure
2900
Gas Rate
2500
permeability are calculated from the superposition
plot, assuming radial flow occurs over the last few
2800 2000
points. This results in pi = 2832 psi and k = 0.175 md.
q, Mscf/d
This permeability is the maximum value possible. The
p, psi
2700 1500
fracture length is determined from slope analysis

2600 1000
on the log-log derivative plot, where the constraint
2500 500 equation (xf k0.5= 27.74) applies. The initial estimated
2400 0
fracture length is 66.3 ft, using k = 0.175 md. This is a
0 20 40 60 80 100 120 140 160 lower bound on the possible fracture length.
t, hrs
A typical analysis would proceed in the following
Figure 3-22 For Example 3-2, synthetic pressure
response for low-permeability fractured well with infinite- manner. A forward simulation based on the estimated
conductivity fracture reservoir properties would be made. The analyst
would then refine the match by simultaneously
A conventional PTA is then performed, regressing with k and xf. Pressure points to be used
treating the model calculated BHPs as observed in the regression analysis would be from the final
data to be matched. The reservoir properties to be build-up. This process would make the modeled final
determined are k, xf and pi. build-up look visually perfect on the superposition
The last flow periods final BHP was used as an and log-log derivative plots. As will be shown, this
anchor point for the analysis. Radial superposition and leads to an almost random set of reservoir properties
the log-log derivative plot for the build-up are shown in being presented as a match.
Figs. 3-23 and 3-24, respectively.
Table 3-5 Fracture half-length, xf ,matrix for Infinite
630 Conductivity Fracture from Example 3-2 with
m(p)* gives p* = 2832 psi Acceptable Build-up Solutions Highlighted
610
xf k0.5, ft-md0.5
590
k = 0.175 md k, md
570 24 25 26 27 27.74 28 29
m(p)
550 0.001 758.9 790.6 822.2 853.8 877.2 885.4 917.1

530 0.005 339.4 353.6 367.7 381.8 392.3 396.0 410.1
510 0.010 240.0 250.0 260.0 270.0 277.4 280.0 290.0

0.020 169.7 176.8 183.8 190.9 196.2 198.0 205.1
490
0.030 138.6 144.3 150.1 155.9 160.2 161.7 167.4
470
100 101 102 103 104 105 106 0.040 120.0 125.0 130.0 135.0 138.7 140.0 145.0
Horner Time
0.050 107.3 111.8 116.3 120.7 124.1 125.2 129.7
Figure 3-23 For Example 3-2, the Horner build-up plot
0.060 98.0 102.1 106.1 110.2 113.2 114.3 118.4
100 0.070 90.7 94.5 98.3 102.1 104.8 105.8 109.6
mrf = 0
mlf = 0.080 84.9 88.4 91.9 95.5 98.1 99.0 102.5
xfk0.5 = 27.74 ft-md0.5
10 k = 0.175 md 0.090 80.0 83.3 86.7 90.0 92.5 93.3 96.7
%m(p) and %m(p)a
Direction of 0.100 75.9 79.1 82.2 85.4 87.7 88.5 91.7

Decreasing
1 Permeability 0.110 72.4 75.4 78.4 81.4 83.6 84.4 87.4
0.120 69.3 72.2 75.1 77.9 80.1 80.8 83.7
0.130 66.6 69.3 72.1 74.9 76.9 77.7 80.4
0.1
0.140 64.1 66.8 69.5 72.2 74.1 74.8 77.5
%m(p)
%m(p)a 0.150 62.0 64.5 67.1 69.7 71.6 72.3 74.9
0.01
0.001 0.01 0.1 1 10 100 1000 0.160 60.0 62.5 65.0 67.5 69.4 70.0 72.5
%ta
0.170 58.2 60.6 63.1 65.5 67.3 67.9 70.3
Figure 3-24 For Example 3-2, the log-log build-up 0.175 57.4 59.8 62.2 64.5 66.3 66.9 69.3
derivative plot
63
Modern Fracturing
An alternative technique called the matrix method J and, in particular, k can be correlated to ultimate
is proposed. The technique requires a matrix, shown gas recovery. The figure shows that permeability
in Table 3-5. Columns are acceptable ranges in values dominates the productivity index. The corresponding
for the parameter xf k0.5. Rows are possible permeability change in fracture half-length required to match the
values. The entries in the matrix are xf calculated from build-up pressure does not change the productivity
the constraint equation xf k0.5 = value. The low-end nearly as much. Even knowing the initial pressure may
permeability has a natural lower bound. This is based not be enough for a unique solution because at low
upon a maximum possible value for xf, based on modeling permeability values, the calculated average reservoir
the treatment with a fracture simulator. After the matrix pressure is insensitive to permeability. For intermediate
is filled in, forward simulation is then performed on these permeability, knowing the initial reservoir pressure
permeability and fracture half-length values. reduces uncertainty in permeability determination.
Values colored in grey represent solutions where Without any knowledge of initial pressure, no amount
the build-up match on the superposition and log-log of nonlinear regression and related visual perfection
derivative plots is acceptable. All other solutions are on the build-up match can sort out this uniqueness
rejected. What is the consequence on productivity from problem. Any forecasts that have such a wide range of
this acceptable family of solutions? Using a combination uncertainty are effectively useless.
of Eqs. 3-12, 3-14, 3-17a, 3-17b and 3-20 for the case To make further progress, one must look at the
where both the permeability and skin vary: drawdown data. The drawdown character is significantly
different for each of these cases (Fig. 3-26). Increasing
f (C fD1 ) x f 1 permeability causes the calculated pressures to move
8 ln
J 2 k2 rw to the left and down, and causes pi to drop. The only
J ratio = = ,
J1 k1 f (C fD 2 ) x f 2 (3-32a) acceptable match is the input case. Mathematically it
8 ln appears that when radial flow is not achieved, the build-
rw
up can solve for at most two of the three unknowns. The
and drawdown allows the solution of the third unknown.
0.5 x f 1
8 ln 3200

J 2 k2 r Variation in Calculated Initial Input Case
w Pressure k=0.001
J ratio = = , (3-32b) 3100 k=0.005
J1 k1 0.5 x f 2

k=0.01
8 ln 3000 k=0.02
r k=0.04
w 2900
k=0.05
k=0.06
p, psi
k=0.07
for both CfD1, CfD2 30, 2800
Direction of Increasing
Permeability
2700
where Jratio is the productivity index ratio and the 2600
subscripts 1 and 2 refer to different interpretations. 2500
1.0 3150 2400

0 20 40 60 80 100 120
0.9
t, hrs
0.8 3100
0.7 Figure 3-26 For Example 3-2, pressure drawdown
0.6 3050 response profiles
pi , psi
J ratio
Jratio
0.5
pi
3000
0.4 For real problems, it is important to put in as detailed
0.3
0.2 2950 a rate profile in the drawdown portion of the test as is
0.1 reflected in the data set. The corresponding observed
0.0 2900
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 BHP will be noisy, so the matching process should be
k, md
performed on a linear pressure versus linear time plot
Figure 3-25 For Example 3-2, Jratio and pi that come in the drawdown portion. When the well test is run
from each match of the matrix method for the infinite
conductivity case after cleanup has been completed, reasonably unique
64
matches are theoretically possible. For such a case, if The overall pressure and rate history is given in
drawdown behavior cannot be matched regardless of Fig. 3-27. A log-log derivative diagnostic plot showing
permeability and fracture half-length combination, this bilinear flow, with either late time superposition effects
may be an indicator that the proppant pack is behaving or boundaries at late time, is shown in Fig. 3-28.
differently under drawdown compared with build-up.
1000
This could be an indicator of non-Darcy flow in the Late Time Superposition
fracture. If the well is not cleaned up before the test, the Effects or Boundary Effects.
match becomes more complicated.
%m(p) and %m(p)a

100
3-12.6 Example 3-3, Low-Permeability Well,

mc = 1
Finite-Conductivity Fracture mbf =
(kfw)k0.5 = 83.3 ft-md1.5
10
The procedure outlined in the previous section will be

%m(p)
re-examined for the case of a finite-conductivity fracture. %m(p)a
Solutions for a finite conductivity fracture have one 1
0.001 0.01 0.1 1 10 100 1000
additional degree of freedom: fracture conductivity. Basic %ta
reservoir properties for the well are given in Table 3-6.
Figure 3-28 For Example 3-3, log-log derivative plot for
The well was fracture stimulated with 132,000 lbs of low-permeability fractured well with finite-conductivity
25/50-mesh pre-cured, resin-coated proppant. Total fluid fracture
pumped was 1080 bbl of oil-based fracturing fluid, at 25
bpm. The maximum proppant concentration during the Some late time superposition effects or boundary
job was 8.3 lbs/gal. After an initial clean-up period of effects are evident. At no time is there any evidence
33 hours, fracture fluid recovery was close to 100%. The of radial flow. The constraint equation for bilinear
well was then shut in for 16 hours. Pressure gauges were flow is (kf w)k0.5 = 86.9 ft-md1.5, as evident from Fig.
then run in via wireline, and flow was resumed. A rate 3-28. This constraint equation is the square of the
decline was evident over the remaining 34 hours of the fundamental term (kf w)0.5k0.25 in Eq. 3-23c and has
drawdown. A 413-hour build-up followed. more convenient units.
The well properties to be determined are k, xf, kf w
Table 3-6 Well Properties for Example 3-3 and pi. These values will be determined by an extension
Pay Thickness, ft 26 of the matrix method to finite conductivity wells:
Porosity, % 8
1. Assume radial flow occurs at the very end
Water Saturation, % 30
rw, ft 0.3 of the test.
2. Use the radial superposition plot to determine k,
4000 8000 m(p)* and then pi. This results in k = 0.196 md
3500 7000 and pi = 3591 psi.
3000 6000 3. Build a matrix of plausible solutions based upon
the value (or plausible range of values) for (kf
q, Mscf/d
2500 5000
p, psi
2000 4000 w)k0.5 as shown in Table 3-7. The rows of the

1500 3000 matrix will be the possible permeability values,
1000 2000 k, and the columns the corresponding values for
500 Pressure 1000 fracture conductivity, kf w.
Gas Rate
0 0 4. Use non-linear regression in conjunction with
0 50 100 150 200 250 300 350 400 450 500
t, hrs forward simulation to adjust xf and the fracture
face skin for each element in the matrix, to get
Figure 3-27 For Example 3-3, pressure and rate
history for low-permeability fractured well with finite-
the best possible match of the build-up. The last
conductivity fracture flow periods final BHP was used as an anchor
65
Modern Fracturing
point for the analysis. This analysis showed Table 3-8 and Fig. 3-29 summarize the build-up
that the late time effects shown in Figure 3- match results with the appropriately determined
28 were superposition related and not as a reservoir properties. Plotting these solutions for the
result of a boundary. drawdown would result in a very similar character
5. Highlighted in grey (Table 3-7) are acceptable to the corresponding infinite-conductivity example
matching cases of the matrix for the build-up. shown in Fig. 3-26. Knowledge of pi from an
independent source would help the matching process,
Table 3-7 Fracture Permeability width, kfw, matrix except at the lowest permeability range. At the lower
for Finite Conductivity Fracture for Example 3-3 with values of permeability the required fracture length
Acceptable Build-up Solutions Highlighted
appears to be becoming unreasonably long, based
(kf w)k0.5 ft-md1.5
k, md
upon preconceived ideas about maximum attainable
75.5 78.7 82.0 85.3 88.6 91.9 95.1 fracture half-lengths. Figure 3-30 shows the best
0.040 377.3 393.7 410.1 426.5 442.9 459.3 475.7 drawdown match, with k = 0.16 md. As in the infinite-
0.060 308.1 321.5 334.8 348.2 361.6 375.0 388.4 conductivity case, increasing permeability causes the
0.080 266.8 278.4 290.0 301.6 313.2 324.8 336.4 calculated drawdown pressures to move to the left and
0.100 238.6 249.0 259.4 269.7 280.1 290.5 300.9 down in Fig. 3-30 and causes pi to drop.
0.120 217.8 227.3 236.8 246.2 255.7 265.2 274.7
0.140 201.7 210.4 219.2 228.0 236.7 245.5 254.3 3100
0.160 188.6 196.9 205.1 213.3 221.5 229.7 237.9
2900
0.170 183.0 191.0 198.9 206.9 214.8 222.8 230.8 Match (k=0.16 md)
2700 Obs Pressure
2500
Table 3-8 Build-Up Match Results for Example 3-3
p, psi
2300
k, Jratio kf w, xf , p i, Fracture
CfD
md md-ft ft psi Face Skin 2100
0.040 0.4417 426.5 567.9 18.78 3704 0.009
1900
0.060 0.5360 348.2 439.3 13.21 3701 0.010
0.080 0.6268 301.6 367.1 10.27 3702 0.013 1700
0.100 0.7140 269.7 321.3 8.40 3650 0.018 1500
0.120 0.8041 246.2 294.3 6.97 3624 0.022 45 50 55 60 65 70 75 80 85 90
t, hrs
0.140 0.8735 228.0 259.1 6.29 3614 0.024
0.160 0.9605 213.3 247.7 5.38 3594 0.030 Figure 3-30 Drawdown match for Example 3-3
0.170 1.0000 206.9 241.1 5.05 3588 0.034
3-13 Non-Darcy Flow
1.0 3780
Jratio
pi 3760 Section 2-3.1 gives the necessary background
0.9
3740 material on the Forchheimer (1914) equation, which
0.8
3720 is the porous media analog of the fluid mechanics
3700 transition from laminar to turbulent flow. The
pi, psi
Jratio
0.7 3680 modern formulation of this law applied to the field

3660 of reservoir engineering was given in Section 2-
0.6
3640 5.2 for flow in a fracture. However it is equally
0.5
3620 applicable for radial flow:
3600
k
0.4 3580 ke = k = , (3-33)
0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 1 + N Re
k, md
Figure 3-29 For example 3-3, Jratio and pi that come where
from each match of the matrix method for the finite-
conductivity case k
N Re =
, (3-34)

66
where ke is the effective permeability, is the non- automated techniques so that accurate solutions are
Darcy correction factor to permeability, is the non- possible for both radial and non-Darcy hydraulic
Darcy coefficient, v is the Darcy velocity and NRe is the fracture flow. This allows rigorous solutions to non-
Reynolds number. It is apparent that a nonlinearity Darcy problems and eliminates the need for the small
has been introduced because the velocity is dependent local region assumption. Using such models, Warren
upon the variables that are being solved for. All the (1993) showed that non-Darcy effects exhibit a unique
nonlinearities are combined into the term. signature for radial flow during build-ups: Wells have
Because non-Darcy effects are severe, it is not a high apparent positive skin, which rapidly dissipates,
readily apparent how to solve Eq. 2-6 where ke replace leading to a sharp drop in the derivative, which is
k. For radial flow it can be shown that although the controlled by after-flow effects resulting from the link
non-Darcy effects are severe, the zone of influence between wellbore storage and skin. Spivey et al. (2004)
is small. Consequently, the nonlinearity in the extended this concept and developed radial flow type
permeability term was dropped and the non-Darcy curve solutions shown in Fig. 3-31. One consequence of
pressure drop was included as an instantaneous effect their work is that non-Darcy flow can now be analyzed
related to velocity and included in the skin term. An for single-rate flow and build-up tests.
effective skin is then defined as:
104
s = s + Dq.
(3-35) 103
CDe2s= 101000
CDe2s= 10500
pwD or tD(dpwD / dtD)
CDe2s= 10250
CDe2s= 10125
in Eq. 3-33 is set to 1.0, and the term s (the rate- 102 CDe2s= 1050
CDe2s= 1030
dependent skin) then replaces s in all the skin-related CDe2s= 1015
CDe2s= 106
101
equations in Chapters 2 and 3. The term D, which
is usually derived from PTA, can be related back to 100
as given in Dake (1978).
Originally it was felt that non-Darcy effects were 10-1
100 101 102 103 104 105 106 107
only important in high-rate gas wells, which were rarely tD/CD
stimulated. For such wells one typically performs a
Figure 3-31 Build-up curves for high-permeability well
multi-rate test, such as a modified isochronal test. with non-Darcy skin (from Spivey et al., 2004)
The goal of the test is to determine s during each
drawdown period and plot it as a function of rate. The For hydraulically fractured wells, flow is
application of Eq. 3-35 will allow the effective skin to predominantly in the fracture, and non-Darcy effects
be split into its two components. The pressure signature occur down the entire fracture length. Guppy et al.
for a multi-rate test on a high-permeability well will (1982) provided the first technique for incorporating
tend to have a stair-step profile with little additional classical PTA based methods to analyze well tests.
pressure drop occurring during a given flow period. s is Similarly Economides et al. (2002a and 2002b) have
very easy to determine for such a test. extended the theory of Unified Fracture Design in
Further progress in the interpretation of non-Darcy a similar manner. On a related topic, Settari et al.
effects has been made by modeling well tests with (2002) quantified the effect of proppant factor
reservoir simulators. Reservoir simulation, by its nature, on well productivity index. Correlations were
is not restricted to solving linear problems. Settari et al. developed relating this to various reservoir and
(1990 and 2000) formulated and solved models where fracture properties. Miskimins et al. (2005) also
the term in Eq. 3-33 was directly included in the quantified the effects of non-Darcy fracture flow on
flow equations. The historical problem associated with long-term production performance.
simulating a well test has involved errors associated It has also become apparent that non-Darcy flow
with gridding and calculating the productivity index. can occur over a much broader range of cases than
Puchyr (1991) and Settari et al. (1990) have developed just high-rate gas wells. It has been observed in high-
67
Modern Fracturing
rate oil wells and, more importantly in the context Table 3-10 UFD Design for Example 3-4
of this book, in low-permeability hydraulically Property Value Type
fractured gas wells. Barree et al. (2005) studied such k, md 5 Input
a case. The usual problem when conducting a PTA Fracture Height (hf), ft 60 Input
is that fracture lengths appear to be significantly Proppant Porosity (p), % 32 Input
shorter than designed. Barree et al. (2005) examined kf , md 100,000 Input
lower-permeability gas wells where PTA indicated Proppant Number (Nprop) 0.036 Calculated
longer fracture lengths than could be justified by CfD,opt 1.6 Calculated
production analysis. These wells had estimated Optimal xf , ft 197 Calculated
permeability between 0.1 and 1 md and fracture Optimal w,in. 0.19 Calculated
lengths of 300+ ft. Production analysis indicated
lengths of less than 30 ft. The length discrepancies At this point, non-Darcy effects have been
could not be attributed to the finite-conductivity ignored. The treatment is assumed to be pumped and
nature of the PTA lengths alone. Using an iterative placed as per the design. Subsequently a flow and
production history-matching program, Barree et al. build-up test is done on the well with the flow rate
(2003) incorporated a variety of laboratory-measured history given in Table 3-11.
damage mechanisms on proppant pack conductivity.
They concluded that non-Darcy effects dominate the Table 3-11 PTA Flow Rate History for Example 3-4
production response. When the well is shut in, the Gas Rate, Flow Period 1, MMscf/d 5
non-Darcy effects quickly dissipate. PTA properties Duration, Flow Period 1, days 3
based on the build-up, therefore, over-estimate the Gas Rate, Flow Period 2, MMscf/d 0
actual production response of the well. Duration, Flow Period 2, days 3
3-13.1 Example 3-4, Non-Darcy, High-Permeability If everything went according to plan, the derivative
Well, Finite-Conductivity Fracture plot of the build-up would look like Fig. 3-32: The well
comes out of wellbore storage, has a short period of
In this example it will be assumed that reservoir bilinear flow, and finally goes into radial flow.
properties are known, as shown in Table 3-9.
108
%m(p)- Darcy Case, kf=100,000 md
Table 3-9 Well Properties for Example 3-4 %m(p) - Non-Darcy Case, kf=100,000 md
%m(p) - Non-Darcy Case, kf=250,000 md
Reservoir Temperature, F 120
Gas Specific Gravity 0.65
Pay Thickness, ft 30
%m(p)
Porosity, % 10 107
mbf =
Water Saturation, % 20
pi, psi 3000
rw , ft 0.25
Drainage Area, miles 2
0.25
Wellbore Storage Coefficient, rb/psi 1 106
0.01 0.10 1.00 10.00 100.00
%ta
For a pre-determined, fixed amount of proppant Figure 3-32 For example 3-4, derivative plots comparing
to be pumped and a known reservoir permeability, k, Darcy and non-Darcy flow in a high-permeability
fractured well with finite-conductivity fracture
one can determine an optimum stimulation design by
applying the UFD design approach given in Section If non-Darcy flow is occurring, then the effective
4-5. This results in the fracture treatment design permeability of the proppant can be significantly
parameters set out in Table 3-10. reduced. Assuming that the factor is given by Eq. 2-
68
38, one finds that = 6.72 104 ft-1. Substituting this 0.1 in., corresponding to a revised fracture length of 673
into the reservoir simulation model of the well test, one ft. Fracture design parameters can now be calculated
finds that 0.21 in the fracture just before shut-in. and are given in Table 3-12.
In the simulation model, actually varies in both space
and time in the fracture, and the value given represents Table 3-12 UFD Design for Example 3-5
an average at the selected time. corresponds to the Property Value Type
reduction factor for permeability of the proppant pack k, md 0.1 Input
at these flowing conditions. At higher rates, the value Fracture Height (hf), ft 60 Input
for would be much lower. Now instead of a designed Proppant Porosity (p), % 32 Input
CfD = 1.6, a value of CfD = 0.34 has been achieved. The kf , md 100,000 Input
derivative plot for this case is also shown in Fig. 3-32. Proppant Number (Nprop) 3.22 Calculated
The fracture is not evident at all in this build-up. The CfD,opt 12.4 Calculated
plot looks very similar to a standard radial model with Optimal xf , ft 673 Calculated
a positive skin. Radial flow is achieved at an earlier Optimal w,in. 0.1 Input
point in time than in the Darcy-flow fractured case.
A steep fall-off in the derivative plot is not evident, At this point non-Darcy effects have been ignored.
as occurs in the pure radial non-Darcy case of Warren The treatment is then assumed to be pumped and
(1993). This indicates that the high-velocity region is placed as per the design. Subsequently a flow and
not as concentrated near the well as in the radial case. build-up test is done on the well with the flow rate
There is therefore no clear diagnostic of non-Darcy history given in Table 3-13, which is at a slightly
flow from the build-up data alone. lower rate than the previous example.
To perform a proper UFD design with non-Darcy
flow, Economides et al. (2002b) have shown that the Table 3-13 PTA Flow Rate History for Example 3-5
designed xf must be decreased and the designed w Gas Rate, Flow Period 1, MMscf/d 3
increased. In this example, the 100,000-md proppant Duration, Flow Period 1, days 3
(assumed to be Ottawa sand) is replaced with Gas Rate, Flow Period 2, MMscf/d 0
intermediate-strength proppant (ISP) with permeability Duration, Flow Period 2, days 3
of 250,000 md. Increasing the permeability also reduces
the factor via Eq. 2-38. The new value is = 1.64 104 The derivative plot of the build-up portion of the
ft-1. All other properties remain the same. Re-running test would look like Fig. 3-33. The well comes out of
the well test through the reservoir simulator results in a wellbore storage into a transition period and then goes
pressure response very similar to the 100,000-md Darcy into linear flow. No radial flow is evident.
flow case, as is seen in Fig. 3-32 with a corresponding
109
0.29 in the fracture just before shut-in. Fracture
Response
False Radial Delayed
Flow
3-13.2 Example 3-5, Non-Darcy, Low-Permeability
108
Well, Finite-Conductivity Fracture
%m(p)
mlf =
As in the previous example, all the reservoir properties 107

will be assumed to be known and identical to the
high-permeability properties in Table 3-9, but with %m(p)- Darcy Case, kf = 100,000 md
%m(p) - Non-Darcy Case, kf = 100,000 md
reservoir permeability of k = 0.1 md instead of 5 md. 106
0.01 0.1 1 10 100 1000
Following through on the UFD design approach given
%ta
in Section 2-5.1, xf = 1313 ft and w = 0.051 in. for
Figure 3-33 For example 3-5, a derivative plot comparing
the kf = 100,000-md proppant. This width is felt to be Darcy and non-Darcy flow in a low-permeability fractured
unreasonably small; a more realistic minimum width is well with high-conductivity fracture
69
Modern Fracturing
If non-Darcy flow is occurring then = 6.72 decline (previously called constant percentage decline),
104 ft-1 based upon Eq. 2-38. Substituting this into the hyperbolic decline and harmonic decline. The unification
reservoir simulation model of the well test, one finds parameter is called b, and it ranges from 0 (exponential
that 0.16 in the fracture just before shut-in. Now, decline) to 1 (harmonic decline). Intermediate values
instead of a designed CfD = 12.4, a value of CfD = 1.98 of b represent hyperbolic decline. These models form
occurs. The derivative plot for this case is also shown in the basis for reserves calculations to this day. They are
Fig. 3-33. After coming out of wellbore storage, there entirely empirical in nature. The three most useful
is a long period of apparent radial flow, followed by a curves are semi-log rate versus linear time, linear rate
transition into linear flow. The interpretation of a real versus linear cumulative, and semi-log rate versus linear
test exhibiting this behavior would lead the analyst to cumulative. The first two of these curves plot as straight
assume the fracture was not there and that radial flow lines for exponential decline, and the latter plots as a
had been achieved. This would result in a significant straight line for harmonic decline.
over-estimation of permeability. Late time behavior From the late 1940s onward, rigorous solutions to
would be interpreted as boundary or composite single-phase flow problems with different boundary
reservoir effect. From the authors experience, this false conditions were becoming available. For a slightly
radial flow is evident in other types of tests as well, as compressible fluid with a constant permeability in
will be shown in Section 3-16.3. Clearly non-Darcy a radial bounded container producing at a constant
effects can lead to complications in PTA interpretation pwf, Raghavan (1993) has shown that the solution
of hydraulically fractured wells. is exponential decline at late time. Arguments have
been made that the hyperbolic and harmonic decline
3-14 Production Analysis models account for real world heterogeneities and
multiphase flow effects that are not accounted for
Production analysis (or PA) has a tradition dating in simple analytical solutions (Laustsen, 1996; and
back even further than PTA and involves forecasting Ahmed and McKinney, 2005). Alternatively gas
future performance. The equations and mathematics compressibility effects associated with declining
are identical to PTA; only the boundary conditions pressure in gas wells cause decline behavior to become
at the well are different. Classical PA assumes a well more hyperbolic than exponential.
is producing at a constant pwf, with the rate declining Using Arps decline models for production
with time. This restriction has now been removed, forecasting implicitly assumes that bounded flow has
and modern PA allows very general well behavior occurred. Transient radial solutions to the constant pwf
to be analyzed. One underlying assumption is that case have also been developed:
rate changes are smooth and not too rapid, so that
1 1
full rate superposition can be replaced with simpler + s = (ln t D + 0.80907). (3-36)
qD 2
techniques. One difference between PA and PTA is
in the nature and quality of the data. A PTA can be Fetkovich (1980) developed a type curve of rate
viewed as a controlled experiment with the rate and versus time that tied the transient radial solution
BHP data being precisely known. PA involves taking (Eq. 3-36) to the Arps boundary-dominated flow
field data as-is and doing the best analysis possible. solutions. This results in a type curve appropriate
Field data, by its nature, is extremely noisy, and the for all times (Fig. 3-35).
pressure data is typically much less precise. A general type curve of rate versus time will be
The most commonly used forecasting technique referred to as a Fetkovich plot. Such a plot has a
uses Arps (1945 and 1956) decline curves. These are number of distinguishing features. When a well is in
illustrated in Fig. 3-34. radial flow, the slope is much shallower than when
By implication, these curves apply to wells the well is in bounded flow. Highly damaged wells
with constant pwf. Arps unified earlier decline curve (wells with high positive skins) have very shallow
methodology into three related models: exponential slopes approaching zero. Wells with negative skins
70
show initially steeper slopes with a maximum value a necessary condition for doing Arps decline is to
of , which flattens out to approximately at first generate a Fetkovich plot of production data to
late transient time. With the transition to bounded determine when bounded flow starts. Arps plots can
flow, the slope gets steeper. For harmonic flow, the then be constructed, ignoring data in the transient
slope becomes unit slope. For exponential decline, the portion. In the past this process may not have been so
slope eventually becomes infinite. Hyperbolic flow is important because only high-permeability reservoirs
between these two extremes. were developed. Bounded flow occurs quite quickly
For an infinite-conductivity hydraulically for high-permeability reservoirs because the time to
fractured well, the Fetkovich type curve was developed boundary of a pressure wave is proportional to the
by Locke and Sawyer (1975), and is shown in Fig 3- permeability (or, more correctly, the diffusivity of Eq.
35. The production declines on a slope. This is an 3-7a). With the development of lower-permeability
indicator of formation linear flow analogous to the gas reservoirs throughout the world, transient flow
slope of on the log-log derivative type curve being can occur for many years before bounded flow
an indicator of formation linear flow for a well test. occurs. As a result, the assumption of bounded flow
After a well begins radial flow, the slope becomes very needs to be verified. An attempt to fit a traditional
shallow, approximately . This figure also shows decline curve through transient data can lead to
that when bounded flow occurs, the type curve drops hyperbolic coefficients greater than 1.0 (b > 1.0). This
off. By analogy, for a finite conductivity fracture one has been termed super-harmonic decline and is not
would see an initial slope of . justified (Fetkovich et al., 1987).
Fetkovich et al. (1987) have shown that the Arps The Fetkovich plot is to production analysis what the
decline curves cannot be applied to wells in transient log-log derivative plot is to well test analysis. It gives flow
flow. The problem with traditional Arps curves is that regime identification capabilities and can be extended
they contain no flow diagnostic information within to variable rate conditions by normalizing the rate
them; they result in only a forecast. Because the onset (replacing q with q/(m(pi) m(pwf)) on the y axis). This
of bounded flow is clear from the Fetkovich plot, is sometimes called a rate-normalized Fetkovich plot.
Coordinate Semilog Log-Log

3040 3040 3040
1000 1000 II
I
Rate, q
2000 I
Log Rate, q
Log Rate, q
II III
I
III
1000 II 100 100
III
0 0 0
0 120 0 120 0 10 100 120
Time, t Time, t Time, t
Coordinate Semilog Log-Log

3040 3040 3040
I I
1000 1000
Log Rate, q
II
Log Rate, q
2000
Rate, q
II III
I
1000 100 III 100

II
III
0 0 0
0 100,000 0 100,000 0 1000 10,000 100,000
Cumulative Cumulative Cumulative
Figure 3-34 Decline curve plots (from Arps, 1956) (I. Exponential Decline b = 0, II. Hyperbolic Decline 0 < b < 1,
III. Harmonic Decline b = 1)
71
Modern Fracturing
For future discussions, no distinction will be made as a function of time to perform the normalized time
between the two types of Fetkovich plots. Unlike calculation. In practice this requires iteration on the
the log-log derivative plot it cannot be used directly drainage volume. In transient flow, the average pressure
to calculate reservoir properties for gas wells. This is does not change, and iteration is not required.
a consequence of the change in gas compressibility Palacio and Blasingame (1993) further generalized
during late time, resulting in unacceptable errors. the Fraim and Wattenbarger (1987) technique by
Unfortunately, the Fetkovich plots shape is sensitive to considering the variable-rate, variable-pwf case. They
shut-ins and significant rate changes due to operational showed that with a new definition of time called
issues. If these occur, it is recommend that one re-zero material balance time, tm, in days, and given by:
the time axis at the beginning of the high-rate event. It
(ct )i q t
is imperative that before production forecasts are made for tm =
0 dt ,
q
(3-37)
any well, the target well and offset wells production data ct
be reviewed with a Fetkovich plot. then one can show
When using analytical solutions to the flow (m( pi ) m( pwf )
equations at late time, as opposed to the Arps decline = bpss + ms tm , (3-38a)
q
curve model, one can theoretically determine all reservoir
properties from a type curve match. For gas wells, this where
requires circumventing the assumption of constant
compressibility. Carter (1985) was the first to modify the (3-38b)
Fetkovich curves for such a case. Fraim and Wattenberger
(1987) extended this work. They showed that instead and
of using time, one should use a new function called
1000
normalized time. This is quite similar to the pseudo-time ms =
, (3-38c)
Gi cti
definition given in Eq. 3-2, except the average pressure
is used in the integration. With this definition of time, and where bpss and ms are the y intercept and
analytical solutions for the variable compressibility slope, respectively, on the material balance time
case plot as exponential decline, on the Fetkovich plot. plot, Bgi is the initial gas formation volume factor
The practical disadvantage of the technique is that the in rcf/scf, Gi is the original gas in place in scf, and
average reservoir pressure has to be known or estimated CA is the Dietz shape factor. As in the Fraim and
100
- Slope Region
Infinite Conductivity Fracture
re/rwa=10
10
re/rwa=30 Transient Boundary Dominated
1
re/rwa=1000
Direction of Increasing Skin
qDd
0 < slope <= -

0.1
0.01
b = 0.0 b = 0.5 b = 1.0

0.001
0.0001 0.001 0.01 0.1 1 10 100 1000
tDd
Figure 3-35 Fetkovich (1980) type curve including the transient radial solution with Arps (1956) decline curves
72
Wattenbarger (1987) technique, the drainage volume The advantage of the derivative function as defined
(or alternatively Gi) needs to be first estimated and by Eq. 3-40b is that it is considerably smoother than
then iterated to get a solution. more traditional derivative formulae. It therefore has
Reformulating Eq. 3-38a in a form consistent with diagnostic capability even when data are scattered.
the Fetkovich variables produces: The transition from transient flow to bounded flow is
1 evident at much earlier times than from the Fetkovich
qDd = 1 + t , (3-39a) plot alone. The biggest difficulty with this technique is
mDd
that identifying the initial flow regime is not possible
qbpss
qDd =
, (3-39b) with this plot, as the x- axis is a function of Gi, which
(m( pi ) m( pwf )
initially is not known. One is immediately involved in
and history matching well performance by changing reservoir
properties to fit the data. Flow regime identification is
ms tm
tmDd =
(3-39c) best performed with a Fetkovich plot.
bpss
A number of alternative approaches have been
where tmDd is the normalized material balance pseudo- developed for PA. One class of techniques is based
time and qDd is the normalized rate. Equation 3-39a is on the concept of a well having constant productivity
the harmonic decline formula and will plot as a unit index when in bounded flow. Adjustments are made
slope on the Fetkovich plot at late times. to the original gas in place and the average pressure
To improve interpretations, Palacio and Blasingame is recalculated from material balance until a constant
(1993) used a derivative technique first described by productivity index is achieved at late time. If one
McCray (1990). New functions are defined as: is not truly in bounded flow, these techniques will
underestimate Gi. West and Cochrane (1995), Agarwal
1 tm
q
qDdi =
tm 0 (m( p ) m( p )) dtm and
i wf
(3-40a) et al. (1999), and Mattar and Anderson (2003) give
variations of this idea and a number of different
straight-line plotting techniques. It is recommended
1 d 1 tm q that this class of techniques be used in conjunction with

qDdid = t dt t
m m
0 (m( p ) m( p )) dtm , (3-40b)

the Palacio and Blasingame (1993) technique, which is
m
i wf
better at determining the onset of bounded flow.
where qDdi is the integrated normalized gas rate and A third class of PA techniques concentrates on the
qDdid is the derivative of integrated normalized gas transient flow period. In transient flow, the average
rate. Plotting the new rate and time functions plus pressure does not change. As a result there is no need
the derivative function on a type curve produces the for either iteration on drainage area or for special time
Blasingame type curves shown in Fig. 3-36. functions. Reitman (1995) looked at wells in radial
transient flow producing at a constant pwf. Equation 3-
10-4 10-3 10-2 10-1 100 101 102 103
101 101 36 describes the flow of such wells. A specialized plot of
1/q versus log t should be a straight line if radial flow is
100 100 occurring. In a direct analog to PTA, picking the proper
qDdi and qDdid
straight line may be difficult. As a result, a Fetkovich

10-1 10-1 plot should be done first to identify the flow regime. The
Fetkovich plot will quantify the time interval for which
10-2 10-2 radial flow is occurring. The slope of the 1/q versus log
t plot can be used to determine permeability. At the
10-3 10-3
10-4 10-3 10-2 10-1 100 101 102 103 end of radial flow, there is significant upward curvature
tmDd in the plot. Arevalo-Villagrain (2001) extended this
work to deal with the variable-rate, variable-pwf case.
Figure 3-36 Blasingame type curves for production
analysis (from Palacio and Blasingame, 1993) The specialized plot then becomes (m(pi) m(pwf))/q
73
Modern Fracturing
versus log tdays. By choosing a time to the end of radial 107 10,000
q
flow from the specialized plot, and the slope of the line
while in radial transient flow, one can determine the
permeability, the Gi and the drainage area from: m = - tel = 6630
%m(p)/q
q,Mscf/d
Days
10 6 1,000
1640T
k=
, (3-41a) m=
m
cpr h
163T ter %m(p)/q

Gi = 100
cpr and (3-41b) 105
(ct Bg )i m 10 100 1,000 10,000 100,000
t, days
Bgi Gi Figure 3-37 Reciprocal Fetkovich plot of a hydraulically

A = , (3-41c) fractured well (from Arevalo-Villagran et al., 2006)
h
where m cpr is the slope of the specialized plot and ter is This well shows slope production decline for
the time to the end of radial flow. Gi can be determined 6630 days (18.2 years). Hydraulic fracture treatments are
without knowledge of porosity, thickness, h, or drainage designed to have fracture half-lengths of at most 1500 ft.
area, A. The reason for this is that the transition from Using the formula in Table 3-1 for the end of formation
linear flow to bounded flow is quite sharp. Table 3- linear flow, one would calculate unreasonably large
1 shows that this occurs at a value of tDA = 0.1. The fracture lengths for any permeability range. What is going
definition of tDA has been given in Eq. 3-13 and involves on here? From a PTA viewpoint one normally sees fracture
permeability, porosity and drainage area, among other lengths appearing to be much shorter than designed; this
variables. Arevalo-Villagrain (2001) found that the initial production analysis shows the exact converse.
transition out of radial transient flow starts a little earlier,
10,000
at tDA = 0.065. Using tDA = 0.065 in the flow equation,
one can then derive the expression for Gi shown in Eq.
3-41b and eliminate the other variables. Gas Rate
q, Mscf/d
This is a powerful technique for comparing 1000

production results to volumetric estimates. If the end of
radial flow has not been observed, then the last available
time is used and Gi represents the currently contacted
100
gas in place. This would be expected to increase with 0 5000 10,000 15,000 20,000
t, days
time until bounded flow occurs. Forecasting with
Figure 3-38 Semi-log rate versus time production
this technique is not possible after bounded flow has
analysis plot (Arevalo-Villagran et al., 2006)
been achieved. A general theory of forecasting for all
conditions will be given later in this section. The cause of long-term linear flow in this case
With the increase in efforts to produce gas reserves is either the presence of channels or, alternatively,
from lower-permeability basins in North America, it is anisotropic permeability. Anisotropic permeability is
becoming evident that many wells do not achieve radial discussed in more detail as a case history in Section 3-
flow and instead exhibit long-term linear flow. From the 16.3. Both reservoir descriptions result in identical
authors experience, long-term linear flow for tight gas slope decline on the Fetkovich plot and can be shown
reservoirs is very common. Stright and Gordon (1983) to be mathematically equivalent. Determining which
were some of the first to identify this behavior. Linear mechanism is appropriate requires geological input.
flow can be identified by a slope on the Fetkovich The transition from formation linear flow as a result
plot. Figure 3-37 is a reciprocal Fetkovich plot (mirrored of the hydraulic fracture to channel linear flow can, in
along the y axis) of a single-zone, hydraulically fractured practice, be near impossible to detect because the slopes
well from Arevalo-Villagran et al. (2006). are the same. One then must rely on well test theory
74
equations (Table 3-1) to decide when formation linear

flow should end. A well having long-term linear flow
xf
has more rapid production decline than a well in the
Formation
center of a square drainage area. Linear Flow ye
Fig. 3-38 is one of the main plots for traditional
Arps decline curves analysis, a semi-log rate versus linear
time plot. An attempt to fit a traditional decline curve
through all this data leads to hyperbolic coefficients
greater than 1.0, or so-called super harmonic decline. As
stated previously, it is not justified because the Fetkovich
plot (Fig. 3-37) shows the well is still in transient flow
up to 6630 days. Despite this, industry still tries to use
semi-log rate versus linear time plots for forecasting,
even when transient flow occurs. Fig. 3-38 shows that
xe
after a long time a semi-log rate versus linear time plot
will appear to exhibit straight-line decline behavior. In Figure 3-39 Fractured well in a long channel, (from
Arevalo-Villagran and Wattenbarger, 2001b)
Canada this combination of an initially steep and finally
a straight-line decline on a semi-log rate versus time Long-term linear flow results for all flow times until
plot is referred to as hockey stick decline. People in boundary effects occur. A specialized plot of (m(pi)
the industry often talk of time to stabilization to this m(pwf))/q versus tdays0.5, called the Neal-Wattenbarger plot,
constant decline rate. From that perspective, Fig. 3-38 will be a straight line if linear flow is occurring. Upward
shows that the well reaches stabilized decline at about curvature occurs in this plot after boundary effects begin.
5000 days. This is close to the time associated with the A Fetkovich plot should first be plotted to ensure that the
beginning of bounded flow, when Arps decline curves linear flow regime is properly identified.
are justified. Forecasting production profiles at earlier Arevalo-Villagran et al. (2006) have identified
times on this plot is impossible. a dimensionless time function when linear flow
Neal and Mian (1989) developed a forecasting ends. Picking the actual time to the end of linear
technique for wells producing under long-term formation flow and determining the slope of the straight
linear flow at a constant pwf. Reservoir properties can line on the Neal-Wattenbarger plot allows one to
be directly related to the decline curve parameters. In calculate reservoir properties:
their case the key reservoir properties were permeability 0.5 315.43T
xe k = ,
and fracture half-length. Their technique involved a ( ct )i0.5 m
cpl h (3-42a)
specialized plot of 1/q versus t0.5. This should be a straight 0.5
ktel
line if linear flow is occurring. If the well is in formation ye = 0.1591 , (3-42b)
( ct )i
linear flow as a result of the hydraulic fracture, the slope
of the plot is proportional to 1/(xf k0.5). This is the same
grouping of variables as in infinite conductivity flow in 200.74T tel0.5 and (3-42c)
Gi =
PTA, as seen in Eq. 3-19d in Section 3-6. (ct Bg )i m
cpl
Wattenbarger et al. (1998) and Arevalo-Villagran
et al. (2006) have extended this solution to the A = 4 xe ye ,
(3-42d)
variable-pwf case and have looked at various special
cases to preserve the one-dimensional character of the where m cpl is the slope of the specialized plot, tel is the
solution. One case they consider is a well located in time to the end of linear flow, xe is the channel half-
a long, narrow channel with an infinite conductivity length in the short direction and ye is the channel half-
hydraulic fracture extending to the boundary in the length in the long direction. As in the radial case, Gi can
short direction (Fig. 3-39). be determined without knowledge of porosity, thickness
75
Modern Fracturing
or drainage area. Other reservoir properties are coupled. transient flow occurs for typical wells, leading to a most
Note that the coupling between permeability and channel likely forecast. Stotts et al. (2007) have extended the
half-width is of the same form as the permeability and technique by considering infinite-conductivity fractured
the fracture half-length in PTA. If permeability, porosity wells in arbitrary locations within a rectangle.
and thickness are known from independent data, the
complete channel geometry can be calculated. When the 3-15 Heterogeneity
production data from the hydraulically fractured well
from Arevalo-Villagran et al. (2006) already plotted in Anyone who has looked at core or log data recognizes
Figs. 3-37 and 3-38 is plotted on a Neal-Wattenbarger that real reservoirs show a remarkable variation of
plot one gets Fig. 3-40. Gi can be calculated directly from reservoir properties on a very fine vertical scale. Why
this plot from Eq. 3-42c. is layered heterogeneity not reflected in well tests?
The short answer is that by its nature the underlying
16
8 diffusivity equation, Eq. 2-4, acts like a huge low-pass
14
7 tel= 6630 days (18.2 years) filter. Most high-frequency information vertical
12
6 variations in horizontal permeability included just
%m(p)/q x 106
Cum Gas, Bcf

10
5
Cum Gas
gets averaged out. As an example, if one generates
4 8
an infinite-acting two-layer reservoir model with all
3 6
properties being identical except permeability, and
2 4
~ fixes the average permeability of the system to be 0.1
1 mcpl = 25 x 103 2
%m(p)/q md, even a 1000-fold contrast in permeability among
0 0
0 20 40 60 80 100 120 140 the layers will not affect the log-log derivative plot.
t0.5, days0.5 Complications occur when layers experience boundary
Figure 3-40 Neal-Wattenbarger plot for a fractured well effects at different times during the test.
in a long channel with long-term linear flow (from Arevalo-
Villagran et al., 2006) 3-15.1 Dual Porosity
As in the radial flow case, if the end of linear flow So what other kinds of heterogeneity are seen in well
has not been observed, then Gi is calculated using the last tests? The dual-porosity model where there is a low-
available time. This result is the currently contacted gas in storativity, high-diffusivity media connected to a high-
place. For either the radial or linear transient case, forecasts storativity, low-diffusivity matrix shows a distinct
can be made assuming that this currently contacted gas signature on the derivative plot (Fig. 3-9). This is a
in place represents the actual Gi. This is equivalent to consequence of the time delay in the lower-permeability
stating that bounded flow begins immediately at the matrix contributing to the production response of
end of the observed data. This forecast should have a the system. Gringarten (1984) has shown that a high-
100% probability of occurring (P100 case) and could be permeability contrast, two-layered model is equivalent
performed with simple material balance type models as to a dual-porosity model. Dual-porosity reservoirs are
used by West and Cochrane (1995) because J is now extremely complicated, and the reader is referenced to
constant. Of course, economic limits and the possible Aguilera (1995) for a thorough review of the topic.
onset of liquid loading in the wellbore, among other
things, may constrain this forecast. Continuing the 3-15.2 Anisotropy
transient case forever will represent the zero probability
case (P0 case). These two bounding cases are easily A type of heterogeneity that has been largely neglected
constructed for a well still exhibiting transient flow. This in the well test literature is permeability anisotropy. This
technique can also be used to forecast the increasing value discussion will consider only permeability anisotropy
of the contacted Gi versus time. Offset wells production in the horizontal direction (i.e., kx ky). The neglect
may be used as a guideline for determining how long is understandable in one sense because an isotropic
76
medium can always be mapped mathematically into an (Xnew and Ynew) are not determinable. However the
equivalent homogenous medium by a process known fracture length is still affected by Eq. 3-43d. If all
as conformal mapping. The problem is solved in the PTA matching is performed in homogeneous space,
homogenous space and then projected back onto the as is recommended, and then mapped backward into
original co-ordinates. Time is not involved in this anisotropic space, the fracture length will grow in
process. Anisotropic well test models will thus have a anisotropic space. This is essentially moving from the
functionally equivalent solution in homogeneous space. right side of Fig. 3-41 to the left.
By an application of Occams razor, why bother? Occams
razor basically says: do not unnecessarily complicate a
problem if a simple solution works just as well with the xe
Ynew
available data. The answer is that additional information
not normally part of the usual PTA process may force ye
the analyst to complicate the problem. The additional
information in question is the long-term production
data normally collected long after the well test.
Geometry for Equivalent
Earlier it was stated that channel linear flow and xf
Isotropic Reservoir
anisotropic linear flow are mathematically identical.

Actual Anisotropic Geometry
Collins (1961) and Raghavan (1993) developed
the mapping functions from anisotropic space to
homogeneous space and vice versa: Xnew
Figure 3-41 Equivalent geometries of drainage and

0.5
k = (k x k y ) , fracture: isotropic case (right) and anisotropic case (left)
(3-43a)
(from Arevalo-Villagran et al., 2006)
k
0.5
ky
0.25 The long fracture in anisotropic space may
X new = X e = X e
, (3-43b) represent the actual treatment pumped; the short
k x k x
fracture in homogeneous space represents the PTA
k
0.5
k
0.25 interpreted length. This is one mechanism whereby
Ynew = Ye = Ye x and (3-43c) PTA-derived fracture lengths could appear shorter
k y k y
than actual treatment lengths. This will also have an
k
0.5
ky
0.25 effect on the associated production forecast, so this
x fnew = x f = x f
, (3-43d) action needs to be justified.
k x k x
where kx and ky are the x and y direction permeabilities, 3-16 Multiphase Flow
Xe and Ye are the distances in the x and y direction from
the well to the boundary in anisotropic space. Xnew and Multiphase flow effects as they pertain to PTA have
Ynew are the corresponding distances in homogeneous their origin in the work of Perrine (1956). The essence
space. xfnew is the equivalent fracture half-length in of the technique is that it determines the total mobility
homogeneous space. Equation 3-43d only applies of the system as opposed to a single-phase permeability.
when the fracture is in the x direction and is of infinite Perrines (1956) technique relies on the relative mobility
conductivity. Figure 3-41 illustrates the mapping of the fluids not changing significantly with time.
function for severe anisotropy associated with widely This approach is of limited use because the
spaced directional natural fractures, where kx >> ky. two multi-phase problems of most interest are gas
The hydraulic fracture is shown parallel to the natural condensates and the cleanup of fracture treatment
fractures because, logically, they would be oriented in fluids immediately after the treatment. They both
the same direction due to the stress regime of the well. can show significant changes in mobility over
For an infinite-acting system, the outer boundaries relatively short periods of time.
77
Modern Fracturing
3-16.1 Gas Condensates Bachman et al. (2006) were involved in a study

of Saudi Arabian gas condensate wells that had been
Gas condensates have been briefly described in Chapter successfully acid fracture stimulated. The wells had
2. As pressure drops below the dew point, hydrocarbon PTA tests immediately after stimulation and cleanup,
liquid drops-out. By its nature, this will occur around prior to the BHP dropping below the dew point.
the producing wells, where the drawdown is largest. They consistently showed good infinite-conductivity
Relative permeability effects restrict the gas mobility fracture behavior. Long-term production data
around the well. As a result there is a significant loss in showed a significant loss in productivity. After about
well productivity. This can be severe and is especially one year of production, additional flow and build-
prominent in wells producing in radial flow. The most up tests were performed. The goal of the study was
successful means of combating this productivity loss to determine the reason for the loss of productivity.
is hydraulic fracturing. One consequence is that the Possible damage mechanisms included liquid
pressure drop is dispersed more uniformly away from dropout effects, collapse of the stress-sensitive acid
the well. Hydraulic fracturing does not remove the fracture or stress-dependent reservoir permeability.
problem; it just makes the best of a bad situation. For Comparisons were made of the first and subsequent
a stimulated or unstimulated well, GOR is typically PTAs. The reservoir simulation models pressure
almost constant, or slowly increasing with time. response is shown in Fig. 3-42.
After a gas condensate well begins production
1010
below the dew point, three distinct flow
regimes quickly occur, as described by Fevang
and Whitson (1996): 109
%m(p) and %m(p)a
1. The inner region is nearest the well and has

both mobile gas and oil (oil and condensate 108
are the same fluid). Oil saturation is above %m(p) - Initial build-up
residual oil saturation. 107 %m(p)a - Initial build-up
%m(p) - Second build-up
2. The intermediate region has the oil phase present,
%m(p)a - Second build-up
but only gas is mobile. The oil saturation is 106
0.0001 0.001 0.01 0.1 1 10 100 1000 10000
therefore below residual oil saturation. %ta
3. The outer region has only gas present (single
Figure 3-42 Derivative response of fractured gas
phase), with solution condensate. condensate well before and after liquid drop-out (from
Bachman et al., 2006)
These regions are dynamic in that they change
shape with time. The problem to be solved from a PTA The humping behavior is identical to what
perspective is a flow-induced composite permeability Yadavalli and Jones (1996) identified as a type 2
system with moving boundaries. condensate response. A simple procedure was then found
PTA techniques that use extended analytical to convert the long-term pressure match to relative
solutions ignore the moving boundary. For radial skin changes associated with the identified damage
flow, Raghavan et al. (1999) compared a number of mechanisms. The vast majority of the skin increase
traditional approaches to more detailed numerical resulted from condensate liquid dropout.
simulation cases to get a feel for what was going on. Gringarten et al. (2006) give a comprehensive review
The first attempt to perform PTA with numerical of the complex issues associated with gas condensate
reservoir simulators for hydraulically fractured PTA and present numerous field cases. Bozorgzadeh
wells was done by Yadavalli and Jones (1996). The and Gringarten (2007) give further examples related to
effect on the derivative curve for a well in which a PTA. Mohan et al. (2006) have considered non-Darcy
significant bank has formed was dramatic (as will be flow, in addition to other effects. The whole field is
shown in more detail later). complex and is an active area of research.
78
3-16.2 Fracture Fluid Cleanup Table 3-14 Well Initial Properties for Example 3-6
pi, psia 2960
In North America, wells are usually hydraulically Pay Thickness, ft 26
fractured immediately after drilling. Cleanup and well Porosity, % 10
testing tend to be lumped together while tie-in of the Water Saturation, % 25
pipeline is occurring. This is logistically an ideal time Estimated Permeability Range, md 0.1 to 0.3
to do a test because production is not lost. From a test Breakdown Pressure, psi 4325
interpretation viewpoint, however, this is hardly ideal Pumped Fracture Half-Length, ft 700
because the effects of fracture fluid cleanup are readily (Best Estimate from Treatment Model)
apparent during early flow periods. However, this must Total Treatment Fluid, bbl 2420
be dealt with as well as is possible. Fluid Recovered to First PTA, bbl 1040
In the authors experience, fluid cleanup effects Fluid Recovered to First PTA, % 43
on test interpretation are often profound. Fracture Proppant Type ISP
lengths often appear to be much shorter than what Proppant, lbs 225,0000
has been designed. This inevitably leads to questions Max Concentration, lbs/gal 7.5
about where did the fracture go? Barree et al. (2003)
looked at a number of factors associated with There are extensive shales around the producing zone,
conductivity and length, including: and fracture height containment was not an issue (see
1. Non-Darcy effects (see Section 3-13) Section 4-3.3.3).
2. Packing of the proppant in the fracture
1011
3. Embedment and spalling in soft formations Superpostion Effects or
4. Filter cake and gel residual migration into the Boundary Effects?
proppant pack
%m(p) and %m(p)a
1010
mlf =
5. Gel yield effects from the concentration of gel xf = 52.5 ft
residue at the fracture tip, reducing its effective mrf = 0
producing length 109
6. Multiphase flow of stimulation and formation %m(p) - Observed

%m(p)a- Observed
fluids in the proppant pack 108
0.01 0.10 1.00 10.00 100.00 1000.00
%ta
Items 3 through 6 are directly related to
Figure 3-43 For Example 3-6, log-log derivative plot for
multiphase flow (if a solid can be considered a phase).
first build-up
The proppant pack relative permeability curves shown
by Barree et al. (2003) have been used by Sullivan et The initial PTA interpretation (Fig. 3-43), done
al. (2006) in sensitivity work to see how they would shortly after the treatment was completed, looks like
affect various modeled PTA cases. a textbook example of a short, infinite-conductivity
fracture proceeding into radial flow. The fracture
3-16.3 Example 3-6, Fracture Fluid Cleanup Case half-length was interpreted to be 52.5 ft. Production
forecasts were generated assuming the well drained
In the Deep Basin area of Western Canada an a square drainage area, following standard industry
operator wanted to assess the effectiveness of past practice. Because the drainage area is unknown, a
hydraulic fracture treatments. As is the norm for number of what ifs based on reasonably estimated
this field, a standard well test was performed with drainage areas were made.
gauges in the hole after the fracture treatment. When the well was put on production, the forecast
The wells were still cleaning up at this point. The almost immediately dropped below expectations, even
wells were all single-zone completions. Table 3-14 for pessimistic drainage areas. Fig. 3-44 shows actual
shows the reservoir properties. long-term production data.
79
Modern Fracturing
10,000 1. The fracture treatment (injection period)

mlf = - 2. All the cleanup flow periods before the first PTA
3. The first PTA
q, Mscf/d
1,000 4. The long-term production data

5. The second PTA
100 The keys to this study centered on two independent

1 10 100 1000
t, days issues. The first was explicitly modeling cleanup after
Figure 3-44 For Example 3-6, Fetkovich plot for long- the fracture treatment, circumventing any simplifying
term production data assumptions with respect to ignoring cleanup on the PTA
interpretation. The second was identifying the source of
What is evident is that linear flow is occurring for the long-term linear flow. The depositional environment
a long time. Calculations as to when formation linear of the pool did not support channel linear flow. Because
flow should end (Table 3-1) indicate that it should be of extensive faulting in the field, permeability anisotropy
within a few days of the start of production. Clearly seemed plausible. The simulation model was run on
linear flow is continuing because of reservoir effects. BHP control throughout the long-term production
A review of offsetting wells production showed history with varying levels of anisotropic permeability.
that about 50% of the interpretable production The match was then done on gas production rate. The
data from a group of over 50 wells indicated long- magnitude of the calculated gas production was most
term linear flow. For the well in this example, at sensitive to the overall permeability, and the shape of
the 280-day mark, a fluid level-based well test was the decline on the Fetkovich plot was sensitive to the
performed. The log-log derivative plot for this second anisotropy ratio. This simplified the process of dialing in
test is shown in Figure 3-45. the correct amount of anisotropy. Table 3-15 gives final
reservoir properties.
1010
Table 3-15 Final Well Properties for Example 3-6

Overall Permeability (kx ky)0.5, md 0.175
%m(p) and %m(p)a
109 kx/ky ratio 400

Propped Fracture Half-Length, ft 525
Fracture Conductivity, md-ft 833
mlf =
CfD 9.1
108
Homogeneous Equivalent Fracture
117
%m(p) - Observed Half-Length, ft (from Eq. 3-43d)
%m(p)a- Observed Gi, scf 1.97 x 109
10 7
Drainage Area, acres 120
0.01 0.1 1 10 100 1000
%ta
1011
Figure 3-45 For Example 3-6, log-log derivative plot for %m(p) - Observed
second build-up, after long production period %m(p)a- Observed
1010 %m(p)- History match
%m(p) and %m(p)a
%m(p)a- History match

This test does not indicate radial flow and shows
109 Observed Data Showing
slope until late time. This can be interpreted as evidence False Radial Flow
that a long, highly conductive hydraulic fracture was
108
indeed placed. Cleanup issues appear to mask the
fracture in the first PTA test. 107
It was decided to run a multiphase numerical 0.001 0.01 0.1 1 10 100 1000
%ta
simulation model and attempt to match all
data simultaneously. This involves modeling the Figure 3-46 For Example 3-6, log-log derivative match of
following distinct events: first PTA test
80
The final fracture half-length is ten times the and could easily result in interpreting too high a
estimate from the first PTA. Figures 3-46 and 3-47 permeability and too short a fracture half-length.
show the match of the two PTA tests. Figure 3-48 Because permeability dominates the productivity
gives the treatment fluid saturation profile around index calculation, forecasts would be optimistic. With
the well after the fracture treatment at the end the lower-mobility fracture fluid around the well,
of the first well test. it simply takes longer to make the transition into
formation linear flow. This is evident in comparing
1010
the time to the start of formation linear flow in the
second test to the corresponding same time in the
first test, where linear flow is not occurring.
%m(p) and %m(p)a
109
With the final history-matched model it is possible
to test the concept that a bigger fracture job is better.
108
Traditional gas forecasting techniques based upon
%m(p) - Observed
%m(p)a- Observed square drainage areas and homogeneous properties
%m(p)- History match
%m(p)a- History match usually indicate this. With anisotropy the benefit of a
107 bigger treatment is reduced.
0.01 0.10 1 10 100 1000
%ta It is concluded that the treatment was pumped
Figure 3-47 For Example 3-6, log-log derivative match of properly and the desired high-conductivity fracture
second PTA test was created. This analysis highlights the impact of
two apparently reasonable assumptions: Neglecting
Wellbore 525 ft fracture half length
cleanup and assuming a square drainage area can
result in forecasts that are unrealistically optimistic.
0.25 In mature producing areas, a Fetkovich plot of any offset
5.3 ft 0.30
invasion 0.35
wells is an absolute necessity in helping identify field-
0.40 wide production flow regimes.
0.45
0.50
0.55
0.60
3-17 Closure Analysis
0.65
0.70
0.75 Industry has been involved in analyzing pressure
Water behavior of wells during pumping of the fracture
Saturation
Legend treatment or from preliminary fracture data for a long
time. Nolte and Smith (1981) pioneered the process
of interpreting fracture geometry from the shape of
Figure 3-48 Saturation profile around the well after the the pressure response. The types of analysis have been
fracture treatment at the end of the first PTA test split into two categories:
1. Before Closure Analysis The pressure data during
There is still a lot of fluid surrounding the pumping is examined. The value of the treating
fracture, with a significantly different mobility from pressure minus the closure pressure (piw pc) is the
the gas. For the first build-up match, the early key parameter in determining the geometry and
flattening of the derivative in Fig. 3-46 is modeled leak-off characteristics of the fracture.
quite well. The flattening is associated with a skin- 2. Closure Analysis After the treatment is
like effect, as was similarly shown in Figs. 3-17 and complete, the pressure begins to fall off, much like
3-33. It is easy to mistake this as the onset of radial a conventional PTA fall-off test. Complications
flow. Hategan and Hawkes (2006) warn about the exist because there is a fracture in the media, which
appearance of false radial flow in fractured wells that is very much in a dynamic state. Over time this
are cleaning up. This example quantifies its effect open fracture will close. Before closing the fracture
81
Modern Fracturing
may continue to grow as a result of energy stored a change in flow regime. Nolte (1986) identified this as
during the treatment. The analysis of pressure the closure time, from which pc could be determined.
to closure will give an indication of the leak-off From the closure time alone, it is then possible to
behavior of the combined fracture/porous media determine the fluid efficiency, , which is the amount
(see Section 4-2.1.9). After the fracture is closed, of fluid remaining in the fracture at closure, normalized
continued analysis of the pressure signature may to the total pumped volume:
give information about permeability and other Gc
reservoir properties for the well.
, (3-45)
2+G c
Both categories of analysis critically require where Gc is the G function at closure. This is
knowledge about or the determination of the independent of fracture geometry and is a key
closure pressure, pc. The various pressures involved parameter in hydraulic fracturing design because it
in hydraulic fracturing are more fully defined in will directly affect the size of the pad. An additional
Section 4-2.1.2. This section will deal exclusively plot used in the industry is the t plot, where pressure,
with pressures after pumping has stopped. Further p, is plotted against t. It is an empirical plot, based
restrictions are that the analysis will be applied to on the general idea of linear flow, as in PTA. However
clean, ungelled fluid without proppant. because the point of initiation of leak-off for fracturing
Nolte (1986) introduced a dimensionless fluid is different for each point in the fracture, the G
function called the G function, or sometimes just function is rigorous, while t is not.
called G time. Castillo (1987) recognized that when In practice, picking closure time from a G
the G function is plotted against the fall-off pressure function plot has proved to be problematic. As with
after a fracture injection test, under ideal conditions a specialized plots in PTA, multiple interpretations are
straight line would form and the slope would lead to often possible. Additionally, a very significant number
the calculation of the leak-off coefficient. of wells have pressure-dependent leak-off (PDL) or
The development of the G function was based other nonlinearities that invalidate Noltes (1986) basic
on the assumption that the leak-off coefficient during assumptions. Approaches and solutions to the PDL
fracturing is constant. Different forms of the G problem began with Castillo (1987) and Mukherjee et
function are possible for low- and high-leak-off cases. al. (1991). Barree and Mukherjee (1996) developed a
The reader is referred to Section 4-2.1.9 for a more diagnostic plot, which will be called the combination
detailed explanation. In practice, there is little difference G function plot, that involves plotting three quantities
between these two extreme cases, and the low-leak-off p, dp/dG and G[dp/dG] on the same plot. The
definition will be used: combination G function plot is the equivalent of the
log-log derivative plot in PTA and can be used to
4
g (t D ) = (1 + t D )1.5 t D1.5 ,
(3-44a) identify flow regimes and to choose pc.
3
Other specialized plots have been developed by
Nolte et al. (1997) and Talley et al. (1999) for two
4 4
G (t D ) = g (t D ) and (3-44b) special cases after the fracture has closed. The first
3
case assumes the well is in radial flow after closure and
t t the second that the well is in linear flow after closure.
i
t D =
, (3-44c) Soliman et al. (2005) have pointed out that the well
ti
can also be in bilinear flow after closure, and give
where ti is the time to the end of injection, tD is ways to determine reservoir properties. These flow
dimensionless time, g is an intermediate variable, and regimes are all possible because it is unclear whether
G is the G function. When the leak-off coefficient is the fracture will completely heal or not.
constant, a plot of pressure versus G should be a straight For radial flow Talley et al. (1999) give the following
line. Deviation from such a straight line would indicate flow equations:
82
1. Combination G function plot

p (t ) pi = mR FR (t , tc ),
(3-46a) 2. Combination pressure, dp/dt and t[dp/dt]
versus t plot
3. Log-log derivative plot
1 t
FR (t , tc ) = ln 1 + c , with t > t , (3-46b) 4. Log-pressure and derivative versus log FL2 plot
4 t tc c
5. If radial flow, a) Horner plot, b) Pressure
versus FR plot
16
= 1.6 , and (3-46c)
2
The concepts will be illustrated for two cases:
Vi constant matrix leak-off and pressure-dependent leak-
kh
= 251, 000
, (3-46d) off. Additional cases considered in Barree et al. (2007)
mR tc
but not covered due to space limitations include
where FR is the Nolte radial after-closure leak-off fracture tip extension after shut-in and fracture height
function, tc is the closure time, Vi is the injection recession during closure.
volume and mR is the slope of the radial flow The constant matrix leak-off case is shown in
after closure analysis plot. For linear flow the Figs. 3-49 through 3-52. Figure 3-49 shows the
corresponding equations are: combination G function plot. Constant leak-off is
shown by the straight-line character of the observed
p (t ) pi = mL FL (t , tc ),
(3-47a) G[dp/dG] data through the origin.
10500 2000
2 t 1
FL (t , tc ) = sin 1 c , with t > t , and
1800
(3-47b) 10000
t c pp 1600
1400
9500 Fracture Closure
G(dp/dG)
1200

P, psi
mL = CL k c
, (3-47c) 9000 G(dp/dG) 1000
800
8500
dp/dG 600
where FL is the Nolte linear after-closure leak-off 8000
400
function, CL is the combined leak-off coefficient and 200

7500 0
mL is the slope of the linear flow ACA plot. In these 0 5 10 15 20 25
formulae , refers to the far-field viscosity, i.e. the G time
gas viscosity. For the majority of cases pi must also be Figure 3-49 Combination G function plot, constant matrix
considered an unknown. Various diagnostics plots to leak-off case
be introduced will rely on special derivatives of these Closure time is picked at deviation from the straight
functions. Although the value p(t) pi depends upon line. Figure 3-50 shows the combination t plot.
the possibly unknown pi, derivative plots do not.
10500 1000
Therefore diagnostic plots will be presented that rely on 1
900
p
first identifying the correct flow regime and secondly 10000
800
Fracture Closure
determining reservoir properties including pi. More 9500 700
Derivatives
recently Craig and Blasingame (2006) came up with a t(dp/dt) 600

p, psi
9000 500
very general theory that accounts for all possible flow dp/dt 400
regimes after the fracture has closed. 8500
300
Now that the underlying theory has been reviewed, 8000 200
Barree (1998), and Barree et al. (2007) describe a 100
7500 0
comprehensive methodology utilizing multiple plots to t
ensure a consistent interpretation of the closure process.
Figure 3-50 Combination t plot, constant matrix leak-
The following plots need to be constructed: off case
83
Modern Fracturing
This plot has frequently been misinterpreted changing fracture/wellbore storage. The deviation
when picking pc, even for this simple case. from constant separation suggests fracture closure
Originally it was felt that departure from the and is consistent with the earlier plots pick of closure
straight line, as in the combination G function time. After closure, the 1 slope line is an indicator
plot, should give pc. This is incorrect and leads to of radial flow. If the slope had been this would
a later closure time and lower apparent closure indicate after-closure linear flow. Figure 3-52 shows
pc. The correct pick occurs at the inflection point, the log-pressure and derivative versus FL2 plot.
as shown in Fig. 3-50. The best way to pick the Time increases to the left in this plot. After closure,
inflection point is from the local maximum of the dp/ the unit slope line indicates radial flow. If after closure
dt curve. In Fig. 3-50 the dashed vertical line is at linear flow occurred, the line would have a slope.
the same time in both the combination G function Because radial flow is evident, a Horner plot and a
plot and the combination t plot. Figure 3-51 shows pressure versus FR plot could also be constructed to
the log-log derivative plot with the G function compare k and pi values.
closure time superimposed. The corresponding pressure-dependent leak-off
(PDL) example is shown in Figs. 3-53 to 3-56. Figure
BH ISIP = 9998 psi 1
3-53 shows the combination G function plot.
%p and %t(d%p/d%t)
1000 m = 0.632
Radial Flow
10,500 1000
1 2
10,250 900
m = -1
p 800
10,000 Fracture Closure
100 700
9750
600
G(dp/dG)
%t(d%p/d%t) 9500 G(dp/dG)
p, psi
%p 500
Fracture Closure 9250
10 400
0.1 1 10 100 1000 9000
300
%t
8750 dp/dG 200
Figure 3-51 Log-log derivative plot, constant matrix leak- 8500 100
off case with G function closure time superimposed 8250 0
0 2 4 6 8 10 12 14 16 18 20
G, Time
10000
Figure 3-53 Combination G function plot, pressure-
dependent leak-off (PDL) case
%p vs FL 2
10250 500
%p and FL2 d%p/dFL2
1000 Start of Radial Flow False Closure 2

10000
p Fracture Closure 400
9750
9500 t(dp/dt) Derivatives
m=1 300
p, psi
100
FL2 (d%p/dFL2) 9250
9000 200
8750
dp/dt 100
10 8500
0.001 0.01 0.1 1
FL2 8250 0
Figure 3-52 Log-pressure and derivative versus FL2 plot t
Figure 3-54 Combination t plot, PDL case

It is common for the pressure difference
and logarithmic derivative curves to be parallel Pressure versus G shows concave upward
immediately before closure. In many cases a near curvature during PDL. The G[dp/dG] plot shows a
perfect slope line is evident, strongly suggesting hump above the characteristic straight line. The end
linear flow in the open fracture. In this case the slope of the hump corresponds to the end of PDL. Fracture
is greater than , suggesting linear flow coupled with closure is still exhibited by departure of the G[dp/dG]
84
function through the origin, as one is back to the the fracture must still be highly conductive. Later,
constant matrix leak-off situation. The combination a 1 slope occurs, indicating a transition to radial
t plot is shown in Fig. 3-54. flow. The log-pressure and derivative versus log-FL2
As in the combination G function plot there plot in Fig. 3-56 shows the linear and radial flow
is a hump in the t[dp/dt] plot. Incorrect picks of features at the appropriate times.
closure time occur when the dp/dt curve is used to Barree et al. (2007) also found that for
pick a flat derivative at early time. Instead one should either the constant matrix leak-off case or PDL
use the separation of t[dp/dt] from the straight line case, one can develop a correlation between
thorough the origin to eliminate this apparent closure permeability, k, and Gc:
time. The correct closure time occurs with a later flat
derivative showing the correct behavior on the t[dp/ 0.0086 f 0.01 pz
dt] curve (line 2 in Fig 3-54). This plot shows that the k= 1.96
, where (3-48a)
G Er
multiple-plot approach is far better than relying on a ct c p
0.0038
single diagnostic plot. It is not recommended to use the
combination t plot in isolation. The log-log derivative pz = pISI pc
(3-48b)
plot in Fig. 3-55 shows near perfect slope flow with
the pressure difference and logarithmic derivative curves and where f is the minifrac fluid viscosity, pz is the
showing a separation of . process zone pressure, and E is Youngs modulus.
The fracture fluid viscosity, f , is normally set to 1.0.
BH ISIP = 10000 psi 1 The storage ratio rp, represents the amount of excess
1000 Linear
%p Flow Radial fluid that needs to be leaked off to reach fracture
(m = -0.5) Flow
%p and %t(d%p/d%t)
closure when the fracture geometry deviates from the

(m = 0.5) normally assumed constant-height planar fracture.
(m = -1)
100 For the constant matrix leak-off and PDL cases, it is
1.0. This permeability correlation allows design work
%t(d%p/d%t) when all other data is lacking. Figure 3-57 is a plot of
Fracture the permeability function, with the PDL example case
Closure
10
0.1 1 10 100 1000 permeability estimate shown.
%t
100
Figure 3-55 Log-log derivative plot, PDL case
rp 1
10000 10 G 0.08 V/V
ct 6.0 x 10-5 psi-1
Start Linear Flow
E 5.0 x 106 psi
%p and FL2 d%p/dFL2
End Linear Flow 1 N 1 cp

Gc
k,md
1000 2.9
%p
Pz 841.0 psi
Start Radial Flow 0.1
(m = 1)
100 (m = 0.5)
0.01
FL2 (d%p/dFL2)
3 2 1 0.001
10 0 1 2 3 4 5 6 7 8 9 10
0.001 0.01 0.1 1 Gc
FL2
Figure 3-57 Permeability versus G function (Eq. 3-48a)
Figure 3-56 Log-pressure and derivative versus FL2 plot, for the PDL case (from Barree et al., 2007)
PDL case
This is a strong indication of fracture linear Conway et al. (2007) have also developed a
flow before closure. The slope immediately after correlation relating initial expected post-fracture rate
closure indicates linear flow, so some remnants of versus Gc, as shown in Fig. 3-58.
85
Modern Fracturing
100,000 of the same basic idea. Both procedures require user

specification of control parameters, so the algorithms
10,000
Higher Reservoir Pressure cannot be treated like a black box.
Post-fracture Rate, Mcf/d
For the average person the best description of

technique is given by Houz et al. (2007). They include
1000
a number of thought-provoking examples, with mixed
Lower Reservoir Pressure results. The problem is not with the algorithm of von
100 Schroeter et al. (2004) but with nonlinear regression in
general. To quote Houz et al. (2007): Any engineer
Post-fracture Rate
10 that has been using non-linear regression on analytical
0 1 2 3 4 5 6 7 8
models knows how this process is impressive when it
Gc
succeeds and pathetic when it fails.
Figure 3-58 Expected post-fracture rate vs. Gc (from
Conway et al., 2007)
The demands for a successful deconvolution
study are high. The data must be self-consistent. For
The rate is defined as the rate 24 hours after instance, if a well has multiple build-ups throughout
burnable gas was recovered at surface. In general, each its producing life, they all must show quite similar
field should have its own correlation, and the best-fit character at each time. If they do not, the analysis is
line represents an optimal stimulation. Having said doomed; therefore, the technique cannot be used on a
that, Fig. 3-58 was originally developed for the low- well with both pre- and post-fracture tests. A necessary
permeability sub-thrusted Deep Morrow in Oklahoma, first step in any study with multiple build-ups would
but the same plot has proved useful in the Moxa Arch be to overlay the normalized derivative plots on top of
Frontier in Wyoming and the Deep Basin area in each other as was shown in Fig. 3-8.
Canada. Evidence indicates that Gc values greater than
8 may be uneconomic to stimulate. 100
FP 06 (Exploration Build-up)
10 FP 35 (Production Build-up)
3-18 Deconvolution Deconvolved Derivative

Rate Normalized Derivatives
1 for Various Smoothness Values

Dimensionless Unit Rate
Pressure & Derivative Curves
The idea behind deconvolution is not new; it is 0.1
essentially the same principle of superposition 0.01

discussed in Section 3-4. The promise of the
0.001
technology is this: Variable-rate pressure history can be
converted mathematically into an equivalent constant- 0.0001
Maximum Flow Period
Durations Test Duration
rate drawdown over the entire duration of the data. 0.00001
10-4 10-3 10-2 10-1 100 10 1
10 2
10 3
10 4
10 5
106 107 108
The standard log-log derivative type curves could Time, hrs
then be used to analyze the test. This would provide
Figure 3-59 Normalized log-log derivative plot and
a means to see all flow regimes that occur over the
deconvolution derivative (from Gringarten et al., 2006)
production history of the well in a type curve plot
that is already understood. On a last note, a simple example from Gringarten
The reason this technology has proven to be et al. (2006) will be shown. This well has only BHP data
elusive is that the numerical process is inherently on an initial DST and on a PTA test two years later.
unstable. A general algorithm proved beyond reach Other than that, only production data was available.
until von Schroeter et al. (2004) found a stable way Figure 3-59 shows the normalized log-log derivative
to implement the mathematics. Their technique was plot for the two tests along with the deconvolution
radically different and converted the problem to a derivative. For this test it may not seem necessary to
standard nonlinear least-squares optimization problem. use deconvolution as the drainage area could be treated
Subsequently Levitan (2005) published a variation as a history matching parameter on its own right. This
86
is true and is indeed how the industry has and will Bachman, R.C., Settari, A., Walters, D.A, Rahim,
continue to solve such problems, but it shows that an Z., and Ahmed, M.S.: Decoupling Damage
alternative approach is now available. Deconvolution Mechanisms in Acid-Fractured Gas Condensate
reveals character in the late transient period not seen in Reservoirs, paper SPE 100570, 2006.
any one-flow period. This feature alone is of great value. Barree, R.D: Applications of Pre-Frac Injection/Falloff
The technology is only in its infancy and it remains to Tests in Fissured Reservoirs Field Examples,
be seen where it will go. paper SPE 39932, 1998.
Barree, R.D., Barree, V.E., and Craig, D.P.: Holistic
Fracture Diagnostics, paper SPE 107877, 2007.
References Barree, R.D., and Mukherjee, H.: Determination
of Pressure Dependent Leakoff and its Effect on
Agarwal, R.G.: Real Gas Pseudo-TimeA New Fracture Geometry, paper SPE 36424, 1996.
Function for Pressure Buildup Analysis of MHF Barree, R.D., Cox, S.A., and Barree, V.L.: Realistic
Gas Wells, paper SPE 8279, 1979. Assessment of Proppant Pack Conductivity for
Agarwal, R.G.: A New Method to Account for Material Selection, paper SPE 84306, 2003.
Producing Time Effects when Drawdown Type Barree, R.D., Cox, S.A., Gilbert, J.V., and Dobson,
Curves are Used to Analyze Pressure Build-up and M.: Closing the Gap: Fracture Half-Length
Other Test Data, paper SPE 9289, 1980. from Design, Build-up, and Production Analysis,
Agarwal, R.G., Gardner, D.C., Kleinsteiber, S.W., and SPEPF (November 2005).
Fussell, D.D.: Analyzing Well Production Data Bourdet, D.P., Whittle, T.M., Douglas, A.A., and
Using Combined-Type-Curve and Decline Curve Pirard, Y.M.: A New Set of Type Curves Simplifies
Analysis Concepts, SPEREE (October 1999). Well Test Analysis, World Oil (May 1983).
Aguilera, R.: Naturally Fractured Reservoirs, 2nd Ed., Bourdet, D.P., Ayoub, J.A., and Pirard, Y.M.: Use of
PennWell Books, Tulsa, Okla., 1995. Pressure Derivative in Well Test Interpretation,
Ahmed, T., and McKinney, P.D.: Advanced Reservoir paper SPE 12777, 1987.
Engineering, Elsevier, Burlington, Mass., 2005. Bozorgzadeh, M., and Gringarten, A.C.: Estimating
Alexander, L.G.: Theory and Practice of the Closed- Productivity-Controlling Parameters in Gas/
Chamber Drillstem Test Method, paper SPE Condensate Wells from Transient Pressure Data,
6024, 1977. SPEREE (April 2007).
Anderson, D., Jordan, C.L., and Mattar, L.: Why Plot Brown, L.P., and Hawkes, R.V.: Rules of Thumb in
the Equivalent Time Derivative on Shut-in Time Well Testing: What Works and Doesnt Work
Coordinates? paper SPE 75703, 2002. and Why, JCPT (May 2005).
Arevalo-Villagran, J.A.: Analysis of Long-Term Carter, R.D.: Type Curves for Radial and Linear Gas-
Behavior in Tight Gas Reservoirs: Case Histories, Flow Systems: Constant Terminal-Pressure Case,
Ph.D Thesis, Texas A&M University, 2001. paper SPE 12917, 1985.
Arevalo-Villagran, J.A., and Wattenbarger, R.A.: Long- Castillo, J.L.: Modified Fracture Pressure Decline
Term Linear Flow in Tight Gas Wells, Internal Analysis Including Pressure-Dependent Leakoff,
Document, Texas A&M University, 2001. paper SPE 16417, 1987.
Arevalo-Villagran, J.A., Wattenbarger, R.A., and Cinco-Ley, H., Samaniego-V. F., and Dominguez,
Samaniego-V., F.: Some History Cases of Long- N.: Transient Pressure Behavior for a Well with
Term Linear Flow in Tight Gas Wells, JCPT a Finite Conductivity Vertical Fracture, SPEJ
(March 2006). (August 1978).
Arps, J.J: Analysis of Decline Curves, Trans. AIME, Cinco-Ley, H., and Samaniego-V., F.: Transient
1945. Pressure Analysis for Fractured Wells, SPEJ
Arps, J.J: Estimation of Primary Oil Reserves, Trans. (September 1981a).
AIME, 1956. Cinco-Ley, H., and Samaniego-V., F.: Transient
87
Modern Fracturing
Pressure Analysis: Finite Conductivity Fracture Fraim, M.L., and Wattenbarger, R.A.: Gas Reservoir
Case Versus Damaged Fracture Case, paper SPE Decline-Curve Analysis Using Type Curves with
10179, 1981b. real Gas Pseudopressure and Normalized Time,
Collins, R.E.: Flow of Fluid Flow through Porous paper SPE 14238, 1987.
Materials, Reinhold Publishing Co., New York, Garcia, J.P., Pooladi-Darvish, M., Brunner, F., Santo,
1961. M., and Mattar, L.: Well Testing in Tight Gas
Conway, M.W., Lindeman, S.W., and Barree, R.D.: Reservoirs, paper SPE 100576, 2006.
Selection of Stimulation Fluids and Treatment Golan, M., and Whitson, C.H.: Well Performance,
Design for Low-Permeability Reservoirs, paper Prentice Hall, Englewood Cliffs, N.J., 1991.
SPE 106085, 2007. Gringarten, A.C.: Interpretation of Tests in Fissured
Craig, D.P., and Blasingame, T.A.: Application of a and Multilayered Reservoirs With Double-Porosity
New Fracture-Injection/Falloff Model Accounting Behavior: Theory and Practice, JPT (April 1984).
for Propagating, Dilated, and Closing Hydraulic Gringarten, A.C.: From Straight lines to Deconvolution:
Fractures, paper SPE 100578, 2006. the Evolution of the State of the art in Well Test
Dake, L.P.: Fundamentals of Reservoir Engineering, Analysis, paper SPE 102079, 2006.
Elsevier, Amsterdam, 1978. Gringarten, A.C., Bourdet D.P., Landel, P.A., and
Dake, L.P.: The Practice of Reservoir Engineering, Kniazeff, V.J.: A Comparison between Different
Elsevier, Amsterdam, 1994. Skin and Wellbore Storage Type-Curves for Early-
Dietz, D.N.: Determination of Average Reservoir Time Transient Analysis, paper SPE 8205, 1979.
Pressures from Build Up Surveys, JPT (August Gringarten, A.C., Bozorgzadeh, M., Daungkaew, S.,
1965). and Abdolnabi, H.: Well Test Analysis in Lean
Earlougher, R.C. Jr.: Advances in Well Test Analysis, Gas Condensate Reservoirs: Theory and Practice,
Monograph Vol. 5, SPE, Dallas, 1977. paper SPE 100993, 2006.
Economides, M.J., Hill, A.D., and Ehlig-Economides, Gringarten, A.C., Ramey, H.J. Jr., and Raghavan, R.:
C.A.: Petroleum Production Systems, Prentice Unsteady-State Pressure Distribution Created by a
Hall, NY, 1994. Well with a Single Infinite Conductivity Fracture,
Economides, M.J., and Nolte, K.G.: Reservoir SPEJ (August 1974).
Stimulation, 3rd Ed., John Wiley and Sons, West Guppy, K.H., Cinco-Ley, H., Ramey H.J. Jr., and
Sussex, U.K., 2000. Samaniego-V., F.: Non-Darcy Flow in Wells with
Economides, M.J., Oligney, R.E., and Valk, P.P.: Finite-Conductivity Vertical Fractures, paper SPE
Unified Fracture Design, Orsa Press, Houston, 8281, 1982.
2002a. Hategan, F., and Hawkes, R.V.: The Importance
Economides, M.J., Oligney, R.E., and Valk, P.P.: of Initial Pressure For Tight Gas Completions
Applying Unified Fracture Design to Gas Wells, and Long-Term Production Forecasting, paper
World Oil (October 2002b). 2006-171, Canadian International Petroleum
Fetkovich, M.J.: Decline Curve Analysis Using Type Conference, 2006.
Curves, JPT (June 1980). Horne, R.H.: Modern Well Test Analysis, A Computer-
Fetkovich, M.J., Vienot, M.E., Bradley, M.D., and Aided Approach, 2nd Ed. Petroway, Palo Alto,
Kiesow, U.G.: Decline Curve Analysis Using Calif., 1997.
Type Curves Case Histories, SPEFE (December Horner, D.R.: Pressure Build-ups in Wells, Proc.
1987). Third World Pet. Cong., 1951.
Fevang, O., and Whitson, C.H: Modeling Gas- Houz, O.P.: Why We Should Stop Using
Condensate Well Deliverability, SPERE Pseudopressures and Other Good Old Well Test
(November 1996). Interpretation Tools After So Many Years of Good
Forchheimer, P.: Hydraulik, Teubner, Leipzig and Service, paper SPE 77619, 2002.
Berlin 116-118 (1914). Houz, O.P., Vituart, D., and Fjaere, O.S.: Dynamic
88
Flow Analysis, The Theory and Practice of Pressure Miller, C.C., Dyes, A.B., and Hutchinson, C.A. Jr.:
Transient and Production Analysis, Kappa The Estimation of Permeability and Reservoir
Engineering, Online Book, 2007. Pressure From Bottom Hole Pressure Build-Up
Hurst, W.: Establishment of the Skin Effect and its Characteristics, Trans. AIME, 1950.
Impediment to Fluid Flow into a Well Bore, Pet. Miskimins, J.L., Lopez-Hernandez, H.D., and Barree,
Eng. (October 1953). R.D.: Non-Darcy Flow in Hydraulic Fractures:
Kelly, R.T.: Non-uniqueness in Well Test Analysis: Does It Really Matter? paper SPE 96389, 2005.
Case History, paper SPE 36528, 1996. Mohan, L., Pope, G.A., and Sharma, M.M.: Effect
Kuchuk, F.J.: A New Method for Determination of of Non-Darcy Flow on Well Productivity of
Reservoir Pressure, paper SPE 56418, 1999. Hydraulically Fractured Gas/Condensate Wells,
Laustsen, D.: Practical Decline Analysis Part 1 Uses paper SPE 103025, 2006.
and Misuses, JCPT (November 1996). Mukherjee, H., Larkin, S., and Kordziel, W.: Extension
Lee, J., Rollins, J.B., and Spivey, J.P.: Pressure Transient of Fracture Pressure Decline Curve Analysis to
Testing, SPE Textbook Series Vol. 9. Richardson, Fissured Formations, paper SPE 21872, 1991.
Texas, 2003. Neal, D.B., and Mian, M.A.: Early-Time Tight Gas
Lee, W.J., and Holditch, S.A: Application of Pseudotime Production Forecasting Technique Improves
to Build-up-Test Analysis of Low Permeability Reserves and Reservoir Description, SPEFE
Gas Wells With Long-Duration Wellbore Storage (March 1989).
Distortion, JPT (December 1982). Nolte, K.G.: A General Analysis of Fracturing Pressure
Levitan, M.M.: Practical Application of Pressure/Rate Decline Analysis with Application to Three
Deconvolution to Analysis of Real Well Tests, Models, SPEFE (December 1986).
SPEREE (April 2005). Nolte, K.G., and Smith, M.B.: Interpretation of
Locke, C.D., and Sawyer, W.K.: Constant Pressure Fracturing Pressures, JPT (September 1981).
Injection Test in Fractured Reservoir History Nolte, K.G., Maniere, J.L., and Owens, K.A.: After
Match Using Numerical Simulation and Type Closure Analysis of Fracture Calibration Tests,
Curve Analysis, paper SPE 5594, 1975. paper SPE 38676, 1997.
Mattar, L.: Critical Evaluation and Processing of Data Odeh, A.S., and Jones, L.G.: Pressure Drawdown
Before Pressure-Transient Analysis, SPEFE (April Analysis, Variable Rate Case, JPT (August 1965).
1996). Palacio, J.C., and Blasingame, T.A.: Decline-Curve
Mattar, L., and Anderson, D.: A Systematic and Analysis Using Type Curves Analysis of Gas Well
Comprehensive Methodology for Advanced Performance Data, paper SPE 25909, 1993.
Analysis of Production Data, paper SPE 84472, Perrine, R.L.: Analysis of Pressure Build-up Curves,
2003. Drill. & Prod. Prac., API, 1956.
Mattar, L., and Zaoral, K.: The Primary Pressure Prats, M., Hazebroek, P., and Strickler, W.R.: Effect
Derivative (PPD) A New Diagnostic Tool In Well of Vertical Fractures on Reservoir Behavior
Test Interpretation, JCPT (April 1992). Compressible-Fluid Case, SPEJ (June 1962).
Matthews, C.S., Brons, F., and Hazebroek, P.: A Puchyr, P.J.: A Numerical Well Test Model, paper
Method for Determination of Average Pressure in SPE 21815, 1991.
a Bounded Reservoir, Trans. AIME, 1954. Raghavan, R.: Well Test Analysis, Prentice Hall, N.J.,
McCray, T.L.: Reservoir Analysis Using Production 1993.
Decline Data and Adjusted Time, M.S. Thesis, Raghavan, R., Al-Hussan, R., and Ramey, H.: An
Texas A&M University, 1990. Investigation of Wellbore Storage and Skin Effect
Meunier, D., Wittmann, M.J., and Stewart, G.: in Unsteady Liquid Flow: I. Analytical Treatment,
Interpretation of Pressure Buildup Test Using paper SPE 2466, 1970.
In-Situ Measurements of Afterflow, JPT (January Raghavan, R., Chu, W.C., and Jones, J.R.: Practical
1985). Considerations in the Analysis of Gas-Condensate
89
Modern Fracturing
Well Tests, SPERE (June 1999). Tight Gas Wells, paper SPE 98329, 2006.
Rahman, N.M.A., Pooladi-Darvish, M., and Mattar, Talley, G.R., Swindell, T.M., Waters, G.A., and Nolte,
L.: Development of Equation and Procedure for K.G.: Field Application of After-Closure Analysis
Perforation Inflow Test Analysis (PITA), paper of Fracture Calibration Tests, paper SPE 52220,
SPE 95510, 2005. 1999.
Ramey, H.J. Jr.: Short-Time Well Test Data Theis, C.V.: The Relationship between the Lowering of
Interpretation in The Presence of Skin Effect and the Piezometric Surface and the Rate and Duration
Wellbore Storage, JPT (January 1970). of Discharge Using Ground-Water Storage, Trans.
Reitman, N.D.: Determining Permeability, Skin Effect AGU, 1935.
and Drainage Area from the Inverted Decline van Everdingen, A.F.: The Skin Effect and its Influence
Curve (IDC), paper SPE 29464, 1995. on the Production Capacity of a Well, Trans.
Settari, A., Bale, A., Bachman, R.C., and Floisand V.: AIME, 1953.
General Correlation for the Effect of Non-Darcy van Everdingen, A.F., and Hurst, W.: The Application
Flow on Productivity of Fractured Wells, paper of the Laplace Transformation to Flow Problems in
SPE 75715, 2002. Reservoirs, Trans. AIME, 1949.
Settari, A., Puchyr, P.J., and Bachman, R.C.: Partially von Schroeter, T., Hollaender, F., and Gringarten, A.C.:
Decoupled Modeling of Hydraulic Fracturing Deconvolution of Well-Test Data as a Nonlinear
Processes, SPEPE (February 1990). Total Least-Squares Problem, SPEJ (December
Settari, A., Stark, A.J., and Jones, J.R.: Analysis of 2004).
Hydraulic Fracturing of High Permeability Gas Wattenbarger, R.A., El-Banbi, A.H., Villegas, M.E., and
Wells to Reduce Non-Darcy Skin Effects, JCPT Bryan, J.: Production Analysis of Linear Fractured
(May 2000). Tight Gas Wells, paper SPE 39931, 1998.
Simmons, J.F.: Interpretation of Underbalanced Surge Warren, G.M.: Numerical Solutions for Pressure
Pressure Data by Rate-Time Convolution, paper Transient Analysis, paper SPE 26177, 1993.
SPE 15477, 1986. West, S.L., and Cochrane, P.J.R.: Reserves
Soliman, M.Y.: Analysis of Buildup Tests With Short Determination Using Type-Curve Matching and
Producing Time, SPEFE (August 1986). EMB Methods in the Medicine Hat Shallow Gas
Soliman, M.Y., Craig, D., Bartko, K., Rahim, Z., and Field, paper SPE 28609, 1995.
Adams, D.: After-Closure Analysis to Determine Yadavalli, S.K., and Jones, J.R.: Interpretation of
Formation Permeability, Reservoir Pressure, and Pressure Transient Data from Hydraulically
Residual Fracture Properties, paper SPE 93419, Fractured Gas Condensate Wells, paper SPE
2005. 36556, 1996.
Spivey, J.P., Brown, K.G., Sawyer, W.K., and Frantz,
J.H.: Estimating Non-Darcy Flow Coefficient
From Build-up-Test Data With Wellbore Storage,
SPEREE (August 2004).
Stotts, G.W.J, Anderson, D.M., and Mattar, L.:
Evaluating and Developing Tight Gas Reserves-
Best Practices, paper SPE 108183, 2007.
Stright, D.H., and Gordon, J.I.: Decline Curve
Analysis in Fractured Low Permeability Gas Wells
in the Piceance Basin, paper SPE 11640, 1983.
Sullivan, R.B., Rushing, J.A., Bachman, R.C., Settari,
A., Ji, L., Conway, M.W., and Barree, R.D.:
Evaluation of Nonlinear Relative Permeabilities
and Their Impact on Water-Frac Performance in
90
Tony Martin is business development manager for international stimulation at BJ Services Company. Since
graduating from Imperial College, London, with an honors degree in mechanical engineering and a master's
degree in petroleum engineering, Martin has spent 17 years in the oil industry and has completed engineering
assignments around the world. Martin's primary interest has been hydraulic fracturing and stimulation, and
he has been involved in production enhancement projects in more than 25 countries. He teaches fracturing,
acidizing and sand control both in-house and externally. A constant theme in this teaching is the need
to de-mystify the world of hydraulic fracturing, in an attempt to make the process more accessible and
less intimidating. He is the author or co-author of numerous SPE papers and has served on the technical
committees for several SPE events. He is also the author of BJ Services Hydraulic Fracturing Manual.
Dr. Peter P. Valk is a professor of Petroleum Engineering and holder of the L.F. Peterson Professorship
at Texas A&M University. A native Hungarian, Valk holds BS and MS equivalent degrees in chemical
engineering and technical mathematics from Veszprm University, Hungary. He received his PhD
("Candidate of Sciences") in chemical engineering from the Institute of Catalysis, Novosibirsk, Russia, in
1981. Before joining Texas A&M in 1993, he was an adjunct professor at the Mining University in Leoben,
Austria; worked for the Hungarian Oil Company (MOL); and was a faculty member at Etvs University,
Budapest, Hungary. A Steering Committee member of the 1999 and 2006 Forum Series and previous
member of the Editorial Board of SPE Journal, Valk is an active participant in the SPE. At Texas A&M,
he teaches advanced hydraulic fracturing, well completion and stimulation, petroleum numerical methods
and general engineering courses. His research interests include design and evaluation of hydraulic fracture
stimulation treatments, rheology of fracturing fluids, performance of advanced and stimulated wells and
the underlying numerical methods. He published several dozen research papers in well-known journals.
Chapter 4 were pumped, one for each zone, with a primitive packer
being employed for isolation. The fluid used for the
Hydraulic Fracture Design for treatment was war-surplus napalm, surely an extremely
Production Enhancement hazardous operation. However, 3000 gals of fluid were
Tony Martin, BJ Services and pumped into each formation.
Peter Valk, Texas A&M University
With special contributions from
Dr Russell Maharidge, BJ Services
4-1 Introduction to
Hydraulic Fracturing
4-1.1 Brief History of Fracturing and Qualitative
Description of Process
The first attempts at fracturing formations for the

purpose of improving production were not hydraulic
Figure 4-1 Klepper Gas Unit No. 1, Hugoton field, Kansas:
in nature they involved the use of high explosives to
The very first frac job
break the formation apart and provide flow channels
from the reservoir to the wellbore. There are records Although post-treatment tests showed that the gas
indicating that this took place as early as 1890. This injectivity of some zones had been increased relative to
type of reservoir stimulation reached its ultimate others, the overall deliverability from the well was not
conclusion with the experimental use of nuclear increased. It was therefore concluded that fracturing would
devices to fracture relatively shallow, low-permeability not replace acidizing for limestone formations. However,
formations in the late 1950s and early 1960s (Howard by the mid-1960s, propped hydraulic fracturing had
and Fast, 1970 and Coffer et al., 1964). replaced acidizing as the preferred stimulation method in
By the late 1930s, acidizing had become an accepted the Hugoton field. Early treatments were pumped at 1 to
well development technique. Several practitioners 2 bpm with sand concentrations of 1 to 2 ppga.
observed that above a certain breakdown pressure, At the time of this writing, tens of thousands
injectivity would increase dramatically (Grebe and of these treatments are pumped every year, ranging
Stosser, 1935). It is probable that many of these early from small skin-bypass fracs at less than $20,000, to
acid treatments were in fact acid fractures. In 1940, massive fracturing treatments that end up costing well
Torrey recognized the pressure-induced fracturing of over $1 million. Many fields produce only because of
formations for what it was. His observations were based the hydraulic fracturing process. In spite of this, many
on squeeze cementing operations. He presented data to industry practitioners remain ignorant of the processes
show that the pressures generated during these operations involved and of what can be achieved.
could part the rocks along bedding planes or other lines Hydraulic fracturing involves the injection of special
of sedimentary weakness. Similar observations were fluids into the formation. As the flow rate increases, the
made for water injection wells by Grebe in 1943 and by pressure differential also increases. Pressure and stress
Yuster and Calhoun in 1945. are essentially the same thing (see Section 4-3.1.1), so
The first intentional hydraulic fracturing process that as the fluid flow generates a pressure differential, it
for stimulation was performed in the Hugoton gas field also creates a stress in the formation. As flow rate (or
in western Kansas, in 1947 (see Fig. 4-1). The Klepper viscosity) increases, so does the stress. If we can keep
Gas Unit No. 1 well was completed with four gas- increasing the rate, eventually a point will be reached
producing limestone intervals, one of which had been where the stress becomes greater than the maximum
previously treated with acid. Four separate treatments stress that can be sustained by the formation and the rock
93
Modern Fracturing
physically splits apart. It is important to remember that High-permeability fracturing is quite different.
it is pressure and not rate that creates fractures (although Usually, production is constrained by effects around the
we often use rate to create this pressure). wellbore, caused by the dual phenomena of skin damage
Pressure and stress is stored energy, or more (van Everdingen and Hurst, 1949, and van Everdingen,
accurately stored energy per unit volume. Energy is 1953) and the increasing constriction of flow as it
what hydraulic fracturing is all about. In order to create approaches the wellbore. Fractures in high-permeability
and propagate a fracture to useful proportions, we have formations are designed to be short and highly conductive,
to transfer energy to the formation. Producing width which in turn usually means maximizing propped fracture
and physically tearing the rock apart at the fracture width (see Section 4-2.1.6).
tip both require energy. Overcoming the often highly However, every formation has a specific combination
viscous frac fluids resistance to being pumped also of propped length and propped width that must be
takes energy. So the key to understanding the hydraulic designed for (more of this in Section 4-5). For now, it is
fracturing process is to understand the sources of energy enough to simply realize that for every combination of
gain, such as the frac pumps and the wells hydrostatic formation and proppant permeability, there is a specific
head, and the sources of energy loss and use. The sum combination of propped length and propped width
of these is always equal to zero. that must be designed for. Therefore, understanding
A great deal can be learned about a formation by the formation permeability is vital to maximizing post-
studying the pressures and pressure profile produced treatment production. If the formation permeability is
by a treatment. The product of the pressure and the unknown the usual situation, unfortunately then
flow rate gives us the rate at which energy is being stimulation will be less efficient (although still potentially
used, i.e. work. This is usually expressed as hydraulic highly successful in economic terms).
horsepower. The analysis of the behavior of fracturing
pressures is probably the most complex aspect of the 4-1.3 Near-Well Flow Enhancement
entire process for most frac engineers. vs. Reservoir Stimulation
4-1.2 High Permeability vs. Low Permeability Hydraulic fracturing has evolved into a technique
suitable to stimulate most wells under extremely
Just as formation permeability has a huge effect on varying circumstances. Originally suggested
production from an unfractured formation, it also for low-permeability gas, it still plays a crucial
has a similar impact on post-treatment production role in developing low-permeability sandstone
from a hydraulically fractured interval. Moreover, the formations, and is increasingly used to produce from
formation permeability also affects the size and shape low-permeability carbonates, shales and coal seams.
of the fracture required to maximize the production In general, a vertical well drilled and completed
increase obtained from fracturing. Put simply, a fracture in a tight gas reservoir must be successfully
has to be designed for specific formation permeability. stimulated to produce at commercial gas-flow
This design process is covered in Section 4-5. rates and produce commercial gas volumes. In
Low-permeability formations require stimulation fact, Holditch (2006) considers the best definition
because the permeability of the formation just isnt high of tight gas as follows: A reservoir that cannot be
enough for the well to produce naturally at economic produced at economic rates nor one can recover
rates. Although the reservoir may contain significant from it economic volumes of gas without large-
reserves, sufficient production cannot be obtained scale hydraulic fracturing treatment or advanced
simply by drilling wells, no matter how well-engineered (horizontal, multilateral) wellbores. Although in
they are. In low-permeability formations, it is easy to some naturally fractured tight gas reservoirs horizontal
produce a fracture that is many times more conductive wells are successful, often they also need fracture
than the formation. Fractures in low-permeability stimulation. In this sense hydraulic fracturing is truly
reservoirs are designed for length. a reservoir stimulation technique.
94
Chapter 4 Hydraulic Fracture Design for Production Enhancement
However, in higher-permeability soft formations pressure differential causes additional stress around the
(for instance in the Gulf of Mexico) hydraulic fracturing wellbore. Eventually, as the rate increases, this pressure
is primarily a near-wellbore flow enhancement (well differential will cause stresses that will exceed the
stimulation) technique, and often its side effects such stress needed to break the rock apart, and a fracture
as sanding prevention might be the primary reason will form. At this point, if the pumps are shut down
of application. In recent years high-permeability or the pressure is bled off, the fracture will close again.
fracturing has become as significant in the economic Eventually, depending on how hard the rock is and the
sense as low-permeability fracturing. magnitude of the force acting to close the fracture, it
will be as if the rock had never been fractured. By itself,
4-1.4 Acceleration vs. Increase of Reserves this would not necessarily produce any increase in
production (although there are specific circumstances
As we move along a hypothetical line from tight when this may increase productivity temporarily, but
gas to high permeability, the recovery acceleration generally these cannot be relied upon).
aspect of the technique becomes increasingly evident. However, if we pump some propping agent, or
However, one should recall that the gas produced proppant, into the fracture and then release the pressure, the
through the year 2000 from US tight gas reservoirs fracture will stay propped open, providing the proppant is
is estimated to be 58 Tcf, and most of this gas would stronger than the forces trying to close the fracture. If this
be still underground without the development and proppant also has significant porosity and permeability,
application of hydraulic fracturing technology then under the right circumstances a path of increased
(Holditch, 2006). The US experience lets the industry conductivity has been created from the reservoir to the
believe that in virtually every basin worldwide wellbore. If the treatment has been designed correctly,
producing substantial volumes of natural gas from this will lead to an increase in production.
conventional resources, large volumes of gas are still Generally, the process requires pumping a highly
to be recovered using hydraulic fracturing and other viscous fluid into the well at high rate and pressure.
up-to-date technology. Pumping at high rate and high pressure requires
Therefore, one should not use some artificial horsepower, and lots of it. This is why the process
classification boxing hydraulic fracturing into generally involves large trucks or skids with huge
either of the above categories and settle for design diesel engines and massive pumps. A typical frac pump
recommendations that were developed as rule of will be rated from 700 to 2700 hydraulic horsepower
thumb in a specific geographical area. The design (HHP). To put this in perspective, 1300 HHP is
methodology presented in Sections 4-5 and 4-6 is approximately equal to 1 MW, enough electricity to
based on a more rigorous approach and emphasizes the power ~500 homes in Western Europe.
importance of optimization with respect to the actual In order to create the fracture, a fluid stage known as
reservoir (or drainage volume) that is usually neither the pad is generally pumped first. This is then followed
clearly tight or high perm but rather represents by several stages of proppant-laden fluid, which carries
some kind of transition between these extremes. the proppant into the fracture. Generally, the fluid and
proppant is mixed into a slurry on-the-fly, and pumped
4-2 Description of the Process downhole using equipment specifically designed to
pump a mixture of liquids and solids. Finally, the
4-2.1 One of the Most Energy- whole treatment is displaced to the perforations. These
and Material-Intensive Industrial Activities stages are pumped consecutively, without any pauses.
After the displacement has finished, the pumps are
As fluid is pumped into a permeable formation, a shut down and the fracture is allowed to close on the
pressure differential between the wellbore pressure and proppant. The frac engineer can vary the pad size,
the original reservoir pressure is generated. As the rate proppant stage sizes, number of proppant stages,
increases, this pressure differential also increases. This proppant concentration within the stages, the overall
95
Modern Fracturing
pump rate and the fluid type in order to produce the 4-2.1.1 Understanding the
required fracture characteristics. Typically, a plot of Significance of Pressure
treatment variables will look like Fig. 4-2.
Understanding the sources of energy gain and energy
loss (or use) is fundamental to understanding the
fracturing process (see Table 4-1). Nolte-Smith analysis
Pressure, Rate, Proppant Concentration
BHTP (Nolte and Smith, 1981) is based on analyzing the

gradient of the bottomhole treating pressure plot
during fracturing operations. On a standard job plot,
STP
where pressure is plotted against time, the gradient of
Slurry Rate this curve is energy divided by time, or work. Thus, an
analysis of the pressure plot gradient can indicate how
much work is being performed by the fracturing fluid
Proppant
Concentration on the formation, or vice versa.
Time
Figure 4-2 Job plot from a typical hydraulic fracture Table 4-1 Sources of Energy Gain and Energy Use
treatment During Fracturing Operations
Hydraulic fracture treatments place hundreds of Energy Gain Energy Use
thousands of pounds of proppant into fractures hundreds

Conversion of mechanical Wellbore friction
of feet long, using tens of thousands of gallons of energy into pressure and Perforation friction
fracturing fluid and thousands of hydraulic horsepower. rate by frac pumps Tortuosity
Hydrostatic head Fluid friction in fracture
Fig. 4-3 shows an idealized fracture with two Overcoming in-situ stresses
symmetrical, elliptically-shaped fractures positioned Fluid leakoff
Producing fracture width
on either side of the wellbore. In reality, the situation Splitting rock at the fracture tip
is often quite a bit more complex. However, this figure
does demonstrate the basic fracture characteristics of
length, xf , height, hf , (assumed to be at a maximum at 4-2.1.2 Different Types of Pressure
the wellbore, for this idealized case) and width, wf (again,
assumed to be at a maximum at the wellbore). In hydraulic fracturing it is common to refer to a large
number of different pressures encountered during
operations and analysis. Each has its own name (or
more usually, several common names) which refer to
either where the pressure is being measured or what
the pressure is doing:
xf
Injection Pressure, pinj . Also referred to as

wellhead pressure (WHP), surface treating pressure
(STP) or simply treating pressure. It is the
hf
pressure at the wellhead, against which the frac
pumps must act.
wf
Hydrostatic Head, phead . Also referred to as HH,

hydrostatic pressure (ph) or fluid head. This is the
Figure 4-3 Idealized elliptical fracture showing length, pressure exerted by the wellbore fluid due to its depth
height and width and its density.
96
Pipe Friction Pressure, ppipe friction. Also referred to as minimum horizontal stress, allowing for the effects
tubing friction pressure or wellbore friction pressure. of pore pressure (see Sections 4-3.2.2 and 4-3.2.3).
This is the pressure loss due to friction effects in the Otherwise, pc is the result of some natural averaging
wellbore as fluids are injected. process involving all the layers. For distinctly multi-
layered formations, it is possible to observe more than
Bottomhole Injection Pressure, piw. Also referred to one closure pressure.
as bottomhole treating pressure (BHTP) or bottomhole
pressure (BHP). This is the downhole pressure, in the Extension Pressure, pext. Also known as fracture
wellbore, in the center of the interval being treated. extension pressure. This is the pressure required
BHTP can be calculated from surface data as follows: inside the fracture in order to make the fracture grow.
By definition, pext > pc as the fracture has to be held
piw = pinj + phead ppipe friction
(4-1) open before it can gain length, height and width.
Extension pressure is not a constant and will vary
Perforation Friction Pressure, ppf. This is the pressure with fracture geometry.
lost as the fracturing fluid passes through the restricted
flow area of the perforations. Perforation friction Fracturing Fluid Pressure, pf. Although used in a
pressure can be calculated by: variety of situations, strictly speaking the fracturing
fluid pressure is the pressure of the fracturing fluid
q 2
ppf = 0.2369 2 s2 2 ,
(4-2) inside the main body of the fracture, after it has passed
N perf D p Cd
through the perforations and any tortuosity. Fracturing
where is the slurry density (ppg), q is the total flow rate fluid pressure may not be constant over the entire
(bpm), Nperf is the number of perforation (so that q/Nperf fracture due to friction effects.
is the rater per perforation) and Dp is the perforations
diameter (inches) and Cd is the discharge coefficient. 4-2.1.3 Net Pressure
Tortuosity Pressure, ptort. Also known simply as Net pressure, pnet, is the excess pressure in the fracturing
tortuosity, this is the pressure lost by the fracturing fluid fluid inside the fracture, above that required to simply
as it passes through a region of restricted flow between keep the fracture open (i.e. pc). In other words, it is the
the perforations and the main fracture or fractures. energy in the fracturing fluid available for propagating
the fracture and for producing width.
Near-Wellbore Friction, pnear wellbore. This is the Net pressure, as used in analyzing fracture geometry,
total pressure loss due to near-wellbore effects and is is immediately beyond the wellbore and just inside the
equal to the sum of perforation friction pressure plus fracture and can be calculated as follows:
tortuosity pressure.
pnet = p f pc , and (4-3)
Instantaneous Shut-In Pressure, pISI. Also known as
ISIP or instantaneous shut-down pressure (ISDP). This pnet = piw p pf ptort pc . (4-4)
is the bottomhole injection pressure immediately after
the pumps have been shut down, so that the effects of It is difficult to over-emphasize the importance of
all the fluid friction-based pressure losses (ppipe friction, ppf net pressure during fracturing. Virtually all analysis
and ptort) have gone to zero. involving fracture geometry uses net pressure as the
common variable linking all parts of the mathematical
Closure Pressure, pc. This is the pressure exerted by model. The net pressure, multiplied by the volume
the formation on the proppant. It is also the minimum of the fracture, gives us the total quantity of energy
pressure required inside the fracture in order to keep available at any given time to make the fracture
it open. For a single layer, pc is usually equal to the grow. How that energy is used (generation of width,
97
Modern Fracturing
splitting of rock, fluid loss or friction loss) is decided Given that fluid is continually leaking off into the
by the individual fracture model being employed to formation (see Section 4-2.1.5), the volume of the
simulate fracture growth. fracture will start to decline if the leakoff volume is
Net pressure also defines the fracture width. For any not replaced. In practical terms, this means that as
given real (i.e. positive) net pressure, there is a specific the fluid leaks off, the pressure declines and the width
fracture width that will be generated by a specific net decreases. To stop the width decreasing, the fluid lost
pressure. For an elliptical fracture, the maximum width into the formation has to be replaced. Pumping into
is defined as follows: the fracture at the same rate that the fluid is leaking
off will maintain constant fracture geometry. In order
2 pnet d (1 2 )
wmax = , (4-5) to increase fracture volume, fluid has to be pumped
E
into the fracture faster than it is leaking off. As extra
where is the Poissons ratio (see Section 4-3.1.2), E volume is introduced, the width will increase in
is the Youngs modulus (see Section 4-3.1.3) and d is accordance with the rise in pnet. If fluid is pumped in
the minimum dimension of the ellipse, such that for at sufficient rate, pnet will rise to a point where it is
a fracture with good height containment, d = hf. The greater than the current pext, and the fracture will gain
term E/(1 2) is often abbreviated to E, the plane height and length in addition to width. In order to
strain modulus (see Section 4-3.1.3). continue fracture growth, fluid has to be pumped into
Net pressure also defines the propagation of the the fracture faster than the sum of the fluid lost to
fracture (the physical splitting of the rock at the leakoff, the volume gained by expanding width and
fracture tip) to produce height and length. For the the volume gained by expanding length and height.
fracture to propagate, the condition pnet > pext must Otherwise, pnet will decline below current pext and the
be satisfied, which is the same as saying that the net fracture will cease to grow, even if pnet is sufficient
pressure has to be high enough to induce a critical to maintain width. This, then, defines a minimum
stress in the formation sufficient to split the rock. treatment rate for fracture propagation.
Determining the value for the pext and hence the critical
stress is not simple; it varies with fracture geometry. 4-2.1.4 Effects of Tortuosity
In addition, fractures often propagate through layers and Perforation Friction
of rocks with varying rock mechanical characteristics
and hence different values for fracture toughness (see Near-wellbore friction (NWBF) is a catch-all term
Section 4-3.4). Therefore, given that in-situ stress and designed to encompass a number of effects that act to
Youngs modulus may also vary from rock layer to rock restrict the flow path between the wellbore and the main
layer, it is often easier for the fracture to propagate in body of the fracture or fractures. Usually, near-wellbore
one particular layer than in any of the others. Like friction is the sum of the effects caused by tortuosity and
everything else in nature, fractures follow the path of perforation friction (see Section 6-6). There are different
least resistance. Hopefully, this path of least resistance techniques for dealing with each of these, and a step-
is in the reservoir rock, leading to the maximum down test may be required to decide which phenomenon
designed length and good height containment within is causing the problem (see Section 4-2.1.8).
the zone of interest. NWBF manifests itself as a frictional pressure drop.
Consequently, net pressure falls into three In fracturing, the surface treating pressure is driven
regimes: by the bottomhole treating pressure, which in turn is
driven by the pressure of the fluids inside the fracture.
pnet 0 Fracture is closed, no propagation possible. Near-wellbore friction adds to the fluid pressure
0 < pnet pext Fracture is open with wf pnet. No inside the fracture, making the BHTP (and hence
propagation possible. the STP) higher than it would normally be. Because
pnet > pext Fracture is open with wf pnet, and pressure measurements of closure pressure are unaffected by
generates sufficient stress to propagate fracture. NWBF (as they are taken under static conditions,
98
when all friction pressure is zero), NWBF can have the 2. The active perforations will erode and increase in
effect of making pnet seem higher than it really is. This diameter throughout the treatment. Proppant is
is illustrated in Fig. 4-4, showing the shut-down at the highly abrasive and can significantly increase the
end of a minifrac. Note the pressure immediately drops size of a perforation opening as the treatment
by approximately 170 psi as soon as the rate is stopped progresses (Cramer, 1987).
(this is an accurate figure as the BHTP is being taken 3. Some perforations are more equal than others.
from a downhole gauge). Without allowing for NWBF, Perforating guns will lie along the low side of the
the engineer (and the simulator) would believe that well; consequently, holes on the low side will be larger
the net pressure was 170 psi higher than it actually is. than holes on the high side. This puts into question
In this particular case, with closure pressure measured the uniformity of the perforation diameters.
at 3335 psi, the actual pnet is about 100 psi, whereas 4. Some perforations may be blocked or partially
if NWBF had not been allowed for, the figure would blocked. Blockages may be due to materials
have been 270 psi. Consequently, the engineer would produced from the formation (scales, asphaltenes,
have believed that the fluid needed more energy for waxes, fines, produced sand), debris from
creating a fracture than it actually had. Accordingly, the perforating, damaging materials introduced from
treatment would have been designed with significantly the wellbore (such as fluid loss additives), or a
more proppant than the fracture could actually hold, combination of these.
leading to a premature screenout.
Most fracture simulators will automatically calculate
3600 50
perforation friction and back this out of any net pressure
3500 +/- 170 psi Near-Wellbore Friction 40 calculations. However, this calculation is only as good as
the data input. The engineer must be aware of this and
Slurry Rate, bpm
Pressure, psi
3400 Gauge BHTP 30

also be wary of behavior as seen in Fig. 4-4.
3300 20 Fixing poor-quality perforations usually requires
some kind of intervention such as a ball-out (with or
3200 Slurry Rate 10
without acid) or re-perforating.
3100 0
15 16 17 18 19 20 21 22 23 24 25
Elapsed Time, mins
Figure 4-4 Minifrac shutdown showing approximately

170 psi near-wellbore friction pressure
Perforation friction is easy to visualize. Perforations

consist of a number of small holes, through which a
Figure 4-5 Diagrammatic illustration of the restricted
highly viscous fluid is pumped at very high rate. Flow Paths between the Perforations and the Main
Therefore, it is almost a foregone conclusion that there Fracture(s) that Cause Tortuosity
is a pressure drop. However, four things act to make
this pressure drop complex and hard to predict: Tortuosity begins where the perforations end. In
an ideal world, there would be a straight, smooth and
1. The treatment will not be pumped through all wide connection between the perforations and the main
the perforations. Even over a long perforated body of the fracture or fractures. However, sometimes
interval with 200 to 400 perforations, it is this is not the case and the flow path is restricted.
likely that only 10 to 20 of the perforations will Sometimes this flow path can be so restricted that it
actually be taking the injected fluid. This has significantly affects the treatment (see Section 6-6). In
been demonstrated many times: The formation fact, tortuosity is probably the single biggest cause of
breaks down at the weakest point, not over the premature screenouts. Figure 4-5 shows a diagrammatic
entire perforated interval simultaneously. illustration of tortuosity. Because fracture-to-well
99
Modern Fracturing
connectivity is crucial in higher-permeability reservoirs, diameter holes (usually, but not always, associated with
alleviating tortuosity is essential; for lower-permeability deep-penetrating charges) produce a smaller stress
formations it is less critical but still desirable. cage around the perforation tunnel and thus make
Tortuosity was first documented by Palmer and fracture initiation easier and less complex. This effect
Veatch (1990), although many others had been observing is greatest in harder, lower-permeability formations.
this phenomenon for some time without realizing Wellbore deviation. Fractures tend to propagate along
exactly what it was. Cleary et al. (1991) were able to a vertical plane (see Section 4-3.2.4) as this is the path
further describe tortuosity and begin to account for it in of least resistance and controlled by the formations
fracture modelling. In 1993, Cleary et al. were the first in-situ stresses. However, in the region around the
to document field procedures for mitigating and even wellbore, because of its cylindrical shape, the stresses
curing the effects of tortuosity. are no longer far-field stresses. In addition, the act of
Not all wellbores are affected by tortuosity. Not all drilling the wellbore (which will have almost certainly
formations are susceptible to tortuosity. However, in have been drilled and cemented overbalanced) and
some formations it can be almost impossible to place any of pumping into the formation to create the fracture
proppant, in spite of the fact that it is relatively easy to further change the stresses around the wellbore. This
create and propagate the fracture itself. The following are means that for non-vertical wellbores, with an azimuth
some factors that affect tortuosity, some of which can be away from one of the principal stress orientations, the
controlled and some of which cannot: fracture will probably initiate on a different plane to
that determined by the far-field stresses. However,
Length of perforated interval. Every perforation is at some point away from the wellbore (within a few
a potential source of fracture initiation. The more wellbore diameters) the far-field stress regime will
perforations there are, the greater the chance of having become again dominant and the fracture will change
both multiple fractures and complexity between plane often quite suddenly. This increases fracture
the perforations and the fracture(s). Reducing the complexity and pressure losses.
number of perforations helps to keep things simple. Rock mechanical characteristics of the formation (see
While this may aid in fracture execution, it may lead Section 4-3). Formations that are hard and brittle (i.e.
to choked fractures during production, a considerable have high Youngs modulus (see Section 4-3.1.3) and
impediment in higher-permeability formations. low fracture toughness (see Section 4-3.4.1)) tend to
Perforation phasing and direction. The optimum be more susceptible to tortuosity than those that are
perforation strategy in vertical wells is to shoot 180- soft and ductile. This is because it is easier to create
phasing perforations, oriented in the direction of and propagate fractures in hard, brittle materials.
maximum horizontal stress. The fracture will also tend Contrast between maximum and minimum
to propagate in this direction (see Section 4-3.2.4). horizontal stresses. The effect of the contrast between
However, the only way to do this is through the use the minimum horizontal in-situ stress (h,min) and
of oriented perforating guns and the knowledge of the the maximum horizontal in-situ stress (h,max) (see
fracture azimuth a priori. Perforations that are pointed Section 4-3.2) is illustrated in Fig. 4-6. If the far-field
away from the direction of maximum horizontal stresses are in a significantly different direction from
stress will add to the complexity of the near-wellbore the orientation of the initial fracture propagation,
situation. In addition, perforation strategies that the fracture will have to make a radical change in
produce spirals of holes around the wellbore, rather direction at some point. When the horizontal stresses
than vertical lines, will also add to the complexity. are very similar, this will be a gradual change in
Shot diameter and depth of penetration. Recent work direction, accompanied by plenty of fracture width.
by Pongratz et al. (2007) indicates that the fracture However, when there is a large contrast between these
will initiate from the base of the perforation, at the stresses, the change in direction will be a lot more
cement-to-rock interface rather than at the tip or abrupt and will introduce a dramatic reduction in
side of the perforation tunnel. In such cases, small- width, leading to restricted flow. Large contrasts in
100
horizontal stresses are typically found in formations means that as the fracturing fluid is being pumped into
that have experienced significant tectonic and other the formation, a certain proportion of this fluid is being
geological activity. Therefore, it is reasonable to lost into the formation as fluid leakoff. Understanding
assume that formations in geologically active areas, the rate of fluid leakoff is essential to understanding
or with a significant history of faulting, will have an the fracture geometry. To start with, material balance
increased tendency towards tortuosity. tells us that the volume of the fracture, Vf, will be equal
to the total volume pumped (or injected), Vi, less the
volume of fluid leaked off, VL:
V f = Vi VL . (4-6)
Carter (1957) was the first to try to quantify fluid

leakoff and defined the leakoff volume in terms of an overall
leakoff coefficient, CL (referred to as Ceff elsewhere). The
velocity of fluid flow into the formation, perpendicular to
the fracture face at a specified point is given by
Figure 4-6 Illustration of the effects of horizontal stress
contrast on tortuosity (after Wright et al., 1996) CL
vL = ,
t texp (4-7)
Premature Screenouts. Tortuosity has a far more sinister
effect than simply adding to the BHTP. Unfortunately, where t is the time since the start of pumping and texp
because tortuosity is caused by restricted flow paths is the time at which the specified point of the fracture
between the perforations and the main body of the fracture, face was exposed to fluid loss. Solving for the entire
it can often be difficult to pump even moderate proppant fracture gives the following result:
concentrations into the fracture. Proppant bridging in
t A
this region is all too common. Often, the geometry of (4-8)
VL = 2
0 0
uL dAdt.
the near-wellbore flow paths is such that lower proppant
concentrations can pass through without any serious The double integral is the summation of leakoff
problems, but as soon as a critical proppant concentration over the entire fracture for the entire treatment period.
is passed, the grains bridge over the flow channel and Remember that each individual segment of the fracture
cause a premature screenout. There are numerous cases is leaking off at a different rate in proportion to the
of treatments screening out just as increased proppant square root of its exposure time. For a constant fracture
concentration reaches the formation. area, such as during a post-treatment pressure decline
Techniques for curing tortuosity rely on deliberately (provided pnet < pext), Eq. 4-8 can be approximated as:
screening out the narrowest of the flow channels and
then using increased rate and pressure to force the
VL CL Af t , (4-9)
remaining flow channels sufficiently wide (Cleary et
al., 1993). Of course the best plan is to incorporate in field units, where Af is measured in ft2 (and includes
drilling and perforating practices, along with appropriate both sides of the fracture), t in minutes and CL in ft
measurements of stresses and stress orientations, in a min-. Note that this fluid leakoff is independent of
holistic approach with hydraulic fracturing. changes in pressure, filtrate viscosity and formation
permeability to the filtrate (which can change due to
4-2.1.5 Fluid Leakoff and Slurry Efficiency relative permeability effects).
Often, the leakoff coefficient can be assumed to
Hydraulic fracture treatments are pumped into permeable be constant; there is usually not enough data available
formations there is little point in carrying out the to do anything else. This is what is usually calculated
process in a formation with zero permeability. This from a minifrac pressure decline (see Section 4-2.1.9).
101
Modern Fracturing
However, the coefficient can also be approximated by C A

w f
calculating three fluid leakoff components and then VL ,C =
t + VS , (4-10)
0.0164

combining these to form the overall leakoff coefficient,
as described by Howard and Fast (1970). The three where VS is the spurt loss volume, which is a function
components are the viscosity controlled coefficient of leakoff area. To account for this, it is usual to define a
(Cv), the compressibility controlled coefficient (Cc) spurt loss coefficient, Sp:
and the wall-building controlled coefficient (Cw)
often referred to simply as the wall-building coefficient Vs = S p A, (4-11)
(Howard and Fast, 1957). These three components
can be combined as dynamic fluid loss (Williams where A is the area over which the spurt loss occurs.
et al., 1979) or harmonic fluid loss (Smith, 1965),
although there are some problems associated with Fluid Efficiency. Fluid efficiency is a concept used in
this, including the fact that most of the required input many fracturing applications and is relatively simple. At
data is very rarely available and has to be estimated. any given time, the fluid efficiency is given by:
This approach is employed in some of the commonly
f V V V VL
used fracture models. Other models employ a grid-
= = i L
= 1 , (4-12)
Vi Vi Vi
based leakoff approach, as discussed by McGowen et
al. (1999). A more complex approach was proposed where Vi is the total volume of fluid injected into the
by Mayerhofer et al. (1993), which accounts for the fracture and Vf is the volume of the fracture itself. Thus,
filter cake and reservoir response, allowing for the the greater the fluid efficiency, the greater the volume
superposition of injection history, filter-cake deposition of the fracture and the less fluid leak-off. As a result,
and associated rate convolution. A full discussion of when a fluid system is described as being efficient, it
these fluid loss models can be found in Economides has low fluid-loss characteristics. Efficiency depends on
and Nolte (1987 and later editions). the fracture size and treating rate, and usually it refers to
the value at the end of fluid injection or pumping, p.
However, efficiency can be defined at any point when the
fracture is open (i.e., pnet > 0).
Care should be taken when using fluid efficiency as a
comparison between fluids and/or treatments. Efficiency
Slope = m is highly variable and not dependent solely upon fluid and
Total Filtrate Volume, ml
formation characteristics. Being fluid loss-based, it is also

dependent upon the fracture area, the pressure differential,
the pumping time and several other variables. This means
that for two treatments pumped into identical formations,
significantly different fluid efficiencies can be observed by
changing only the pumping rate or injected volume.
Spurt Volume
4-2.1.6 Dimensionless Fracture Conductivity
Time, mins
Figure 4-7 Illustration of the effects of horizontal stress Dimensionless fracture conductivity, CfD, is a measure
contrast on tortuosity (after Wright et al., 1996)
of how conductive the fracture is compared to the
Spurt loss. Spurt loss is the extra fluid loss that occurs formation. The consequences of this will be discussed
before the fracturing fluid builds up a complete filter in detail in Section 4-5. To start with, the dimensioned
cake on the fracture face. As illustrated by Fig. 4-7, fracture conductivity is:
the total fluid loss due to filtration-based fluid loss
can be expressed as: C f = wave k f ,
(4-13)
102
where wave is the average propped width (ft) and kf is These upper and lower boundaries are the result
the permeability of the fracture ( in-situ proppant of solving a polynomial equation. This means that
permeability) in md. Think of Cf as the flow capacity of for practical values of n, the lower boundary of e
the fracture. The dimensionless fracture conductivity is will be between 0.25 and 0.125, while the upper
calculated from the fracture conductivity: boundary will be between 0.3333 and 0.2. So any
straight line on a Nolte-Smith plot with gradient
Cf wave k f
C fD = = . (4-14) between 0.3333 and 0.125 probably indicates that
xf k xf k
there is very good height containment (provided
Equation 4-14 compares the ability of the fracture to the fluids bottomhole rheological properties are not
transport fluids to the wellbore (the numerator) with changing significantly).
the ability of the formation to transport fluids to the Nolte and Smith went on to define the pressure
fracture (the denominator). CfD is a major variable used response of other behaviors during fracturing. The results
in fracture design and has a massive influence on post- of this are illustrated in Fig. 4-8 and Table 4-2.
treatment production. It will be used extensively in
numerous subsequent sections of this book.
b
4-2.1.7 Nolte-Smith Analysis Predicting
a
Fracture Geometry from Pressure Trends log pnet II III
I
Nolte and Smith (1981) introduced a method for IV
analyzing the pressure response of a formation during

pumping, in order to interpret the fracture geometry
being produced. Based on the 2-D Perkins and Kern
Nordgren (PKN) fracture geometry (Perkins and Kern, log t
1961, and Nordgren, 1972), Khristianovich and Zheltov Figure 4-8 Nolte-Smith analysis pressure response
(1955) and Geertsma and de-Klerk (1969) (KGD), and
radial models (see Section 4-5.6), the method analyzes Table 4-2 Nolte-Smith Analysis Pressure Response
the expected pressure response from the formation during Modes (with reference to Fig. 4-8)
fracture propagation, and then predicts the pressure
Mode Behavior
response when certain types of behavior take place.
PKN fracture geometry assumes constant height, Propagation with PKN fracture geometry. Gradient
I is equal to e (see Eq. 4-16) for constant frac fluid
with length considerably longer than height, and also rheology.
that net pressure is a function of time such that pnet(t) Constant gradient = 0. Represents height growth
in addition to length growth, or increased fluid loss,
te, where 1/8 < e < 1/5 for a Newtonian fluid (see Section II or both. Can also be explained by a change in the
4-4.1.1). Taking logs of this relationship gives: relationship between pnet and wf (see Eq. 4-5 this
implies a change in rock mechanical characteristics).
Unit slope. This means that pnet is now directly
log pnet = e log t + constant.
(4-15) proportional to time (and also to rate, as this is
IIIa usually constant with respect to time). This behavior
is usually associated with additional growth in wf
This means that on a plot of log pnet against log t, such as during a tip screenout (see Section 4-7.3.2).
fractures exhibiting PKN fracture geometry will have IIIb
Slope > 2. Screenout, usually a near-wellbore event
with a very rapid rise in pressure.
a straight line of gradient e. This represents Mode I
Negative slope. Rapid height growth. Potentially
on the Nolte-Smith plot (see Fig. 4-8). For power law IV
KGD or radial fracture geometry.
fluids (see Section 4-4.1.3) the gradient e is defined
with upper and lower boundaries as: The introduction of Nolte-Smith analysis in the early
1980s coincided with the first practical use of computers
1
< e < 1
. (4-16) on fracturing locations. Consequently, fracturing service

4n '+ 4 2n '+ 3
103
Modern Fracturing
companies were able to provide the first real-time fracture steps before the fracture opens, so that a pressure trend can
diagnostics using Nolte-Smith plots that were continually be seen. At each rate step, it is important not to proceed
updated during the treatment. For the first time in to the next step until the pressure has stabilized, so the
fracturing history, it became important to stay inside the pump operator should get to approximately the right rate
control van and watch the computer screens. and then leave the pumps alone. It is more important to
Nolte-Smith analysis remains a very powerful tool stabilize the rate and pressure for each stage than it is to
for predicting fracture geometry, despite the fact that it is achieve exactly the right rate. Record the stabilized rate
based on explaining deviations from ideal 2-D behavior. and bottomhole treating pressure for each stage. As the
rate increases, bigger steps can be taken: Typical pumping
4-2.1.8 Step Rate Tests rates for a step-up test could be 0.7, 1.0, 1.2, 1.5, 1.8,
2.1, 2.5, 3.0, 4.0, 6.0, 8.0, 10.0 and 15.0 bpm. After the
Step rate tests, along with minifracs (see Section 4- test, plot stabilized BHTP against slurry rate to find the
2.1.9), are collectively referred to as calibration tests. fracture extension pressure, as illustrated in Fig. 4-9.
These are fluid injections before a fracture treatment,
4500
designed to collect important information to help
calibrate the fracture simulator and hence provide a 4300
more accurate prediction of fracture geometry. They
Gauge BHTP, psi

can also provide some important information about 4100
near-wellbore restricted flow. Step rate tests come in
3900
two varieties, step up and step down. pext= +/- 3920
When performing step rate tests, it is necessary

3700
to obtain reliable bottomhole pressure data. BHTP
calculated from STP does not provide accurate enough 3500
data for treatment-critical analysis such as this. Often, 0 1 2 3 4 5 6 7 8 9 10
Slurry Rate, bpm
BH memory gauges can be run on slickline and held Figure 4-9 Typical pressure-rate cross-plot from a step-
in place during the calibration tests, to be retrieved up test
afterwards. The data can be merged with surface data
to provide an accurate basis for the analysis. This can be The following points should be remembered when
done quickly and cheaply. conducting a step-up test:
Step-Up Tests. Step-up tests are performed with 1. Always start with the fracture closed.
the fracture initially closed. The objective of the test 2. Use any fluid. However, if BHTP is being
is to obtain the fracture extension pressure, pext, by calculated from surface data, the wellbore contents
injecting into the formation in a series of increasing must be understood.
rate steps and then by analyzing the data on a pressure 3. Note that the fracture extension pressure, pext, is
against rate cross-plot. not constant. It will vary throughout a fracture
To pump a step-up test, start with the fracture closed treatment, usually decreasing. As the fracture
(often, the step-up test is the first operation performed grows (per unit volume of fluid injected), it
on the well). Start at the lowest rate possible with the frac becomes easier to create fracture length and
pumps this is usually in the region of 0.5 to 0.75 bpm. height, and harder to create width.
Then step up the rate by as little as possible; an experienced 4. Although the step-up test is the least useful of the
pump operator will probably be able to get the next rate calibration tests, it does provide a useful upper
step at 1 bpm. Keep stepping up in small increments, boundary for finding pc from a pressure decline
a difficult task with modern, high-horsepower fracturing (which can sometimes be quite difficult). A pext
pumps, requiring good control from the pump operator. generated through a step-rate test is usually 100
However, it is important to get at least three good rate to 200 psi greater than pc.
104
Step-Down Tests. Step-down tests are usually pumped Perforation-dominated near-wellbore friction
in order to differentiate between perforation-dominated is easy to quantify. From Eq. 4-2 we can see that
and tortuosity-dominated near-wellbore friction, as ppf q2 and so a rapid increase in pressure with
documented by Cleary et al. (1993) (see Section 6- increasing rate is observed. However, if the number
6.2). The test starts with the fracture open. This means of open perforations taking fluid changes (opening of
that significant pumping has to be performed before blocked perforations) or if the perforation diameter
the step-down test can commence. Often, the step- increases (through erosion of blocking material),
down test is done as the second half of a combined the pressure will rise more slowly.
step-up and step-down test. Sometimes, a step-down For tortuosity-dominated near-wellbore friction,
test is performed at the end of a minifrac (this is not the relationship between pressure and rate is more
recommended, however, because the step-down can complex. Often this relationship is approximated
obscure both the pressure decline and any assessment to ptort q, although in reality the relationship
of the magnitude of pnear-wellbore calculated from the is quite a bit more complex, as there are several
difference between piw and pISI). possible underlying causes of tortuosity (indeed,
Performing a step-down test is radically different Wright et al., 1996, suggested that ptort q).
from performing a step-up test. It is important that Whatever the cause, it is clear that ptort will
the test is performed quickly, because the fracture must rise more slowly than q. This is because as the
remain open throughout the test. Start at fracturing rates pump rate increases, so does the fluid pressure
(ideally 15 to 20 bpm) with a significant volume already due to viscous friction effects within restricted
pumped into the fracture (at least 5 minutes pumping flow channels. However, the geometry of these
at fracturing rates). Step down rapidly in four or five channels is pressure dependent the higher
approximately equal rate steps, spending no longer than the pressure, the wider they get (consequently,
15 seconds at each step. Then plot BHTP against slurry one of the accepted approaches to dealing
rate in a similar fashion to the step-up test. Figure 4- with tortuosity is simply to pump as fast as
10 shows the results of two step-down tests performed possible). Therefore, friction effects, and hence
on the same well. The first test illustrates perforation- the pressure upstream of the tortuosity, increase
dominated NWBF, with the characteristic concave curve. more slowly than the slurry rate.
As a result of the test, the operator decided to re-perforate Things to remember when conducting a step-
the interval and perform a second step-down test. The down test:
results of the second test display the characteristic convex Make sure the fracture is open throughout the
shape of tortuosity-dominated NWBF. However, note test and is of significant size. The test is trying to
the dramatic decrease in BHTP as a result of the re- measure the pressure response of the tortuosity. If
perforating this clearly illustrates that the decision to the fracture is small, significant pressure changes
re-perforate was a good one. will be generated simply by the varying fracture
geometry; the larger the open fracture is, the less
10,200
significant these become.
Step-Down Test 1 If the results from a step-down test are to be
9800
used in the near-wellbore friction pressure tables
Gauge BHTP, psi
(Perforation Dominated)
9400
included in many modern fracture simulators,
then the step-down test has to be conducted
Step-Down Test 2
9000 (Tortuosity Dominated)
with the same fracturing fluid that is planned for
the main treatment. Otherwise the magnitude
8600 of the near-wellbore pressure response which
1 2 3 4 5 6 7 8 9 10 11
Slurry Rate, bpm
is viscosity-dependent will be meaningless.
If these tables are not to be used, then any
Figure 4-10 Results of two step-down tests performed
on the same interval, before and after re-perforating known fluid can be pumped.
105
Modern Fracturing
Step down as fast as possible, as described above. It is in terms of length, height and possibly width. However,
important that the geometry of the fracture behind this analysis is somewhat qualitative. To further increase
the tortuosity not change significantly during the the accuracy of the fracture geometry it is necessary to
step-down. If the steps take too long, the fracture will perform a pressure match (see Section 4-2.2.2). This
start to get smaller and may even close. process involves the input of selected treatment data (rate
Monitor the pressure decline after the last step and the and proppant concentration) into the fracture simulator
pumps have been shut down. If the test is performed model and comparing the predicted net pressure response
correctly, the fracture will still be open at the end of with the actual net pressure response. Key variables in
the step-down, and it may be possible to identify the simulator (usually stress, Youngs modulus, fracture
fracture closure from the decline curve. toughness and fluid leakoff) are adjusted until the
Some simulators are capable of importing the data predicted pressure response matches the actual pressure
and producing a plot of p against q (i.e., the response. The problems associated with this process are
change in BHTP against the change in slurry rate). discussed in detail in Section 4-2.2.2. However, the
Consequently, this is a direct plot of NWBF against pressure match does enable a much more quantitative
rate. Usually, such plots can back out the theoretical assessment of the fracture geometry.
Rate, bpm, and Proppant Concentration, ppa

perforation friction and produce a calculated ptort.
6000 50
The closer this calculated curve is to the actual Gauge BHTP Near-Wellbore
Friction
curve, the more the MWBF is dominated by 5000 40
tortuosity and vice versa. Of course, this calculation Pressure, psi Frature
4000 Geometry 30
is not to be trusted if the diameter and/or number Fluid Leakoff, Closure
Pressure, Reservoir
of perforations are unreliable. 3000 Slurry Rate
Pressure and Permeability 20
2000 10
4-2.1.9 Minifracs Tortuosity Proppant
Concentration
1000 0
0 5 10 15 20 25 30 35 40 45 50
The purpose of the minifrac is to provide the best
Elapsed Time, mins
possible information about the formation, prior to Figure 4-11 Example minifrac job plot, illustrating the
pumping the actual treatment. significant parameters (italics) that can be derived from its
The minifrac is designed to be as close as possible to analysis
the actual treatment, without pumping any significant Figure 4-11 also illustrates a proppant slug. This
volumes of proppant. The minifrac should be pumped is a test to see how aggressive any tortuosity is; simply
using the anticipated treatment fluid, at the anticipated measuring the quantity of the pressure drop is not
rate. It should also be of sufficient volume to contact all sufficient. The near-wellbore region needs to be tested
the formations that the estimated main treatment design for its ability to transmit proppant. The proppant slug
is anticipated to contact. A well planned and executed is designed to do this. It is a very short stage of proppant
minifrac can provide data on fracture geometry, rock usually less than 1000 lbs. As the slug passes into the
mechanical properties and fluid leakoff information fracture, the response of the BHTP is observed. Ideally,
that is vital to the success of the main treatment. Below no pressure rise is seen, indicating that the near-wellbore
is a summary of the process involved in minifrac analysis flow channels have adequate width. However, if tortuosity
see Martin (2005) for a more detailed discussion. is significant, a pressure rise will be observed. In the
worst cases, the well will screen out with the proppant
Minifrac Data During Pumping. As previously slug. While this may seem to be a disaster, it has in fact
discussed (see Section 4-2.1.1), the shape and gradient prevented a premature screenout during the treatment,
of the pressure-against-time plot tells us a great deal after pumping a wellbore full of proppant. Note also that
about the fracture itself. This is illustrated in Fig. 4-11. it is common practice to pump a series of proppant slugs
Nolte-Smith analysis (see Section 4-2.1.7) can be use to at increasing concentrations, in order to further test the
obtain a general idea of how the fracture is propagating response of the near-wellbore region.
106
Minifrac Pressure Decline\Up to Fracture fracture area and closure time will yield the leakoff
Closure. The analysis of the pressure decline after the rate and, hence, the leakoff coefficient. Note that
end of pumping is an important part of the minifrac this is an iterative process because leakoff also plays
analysis process. Analyzing the data up to fracture a significant role in the shape of the pressure plot
closure will provide the following information, as during pumping. The pressure match also needs
illustrated in Fig.4-12: to match the before-closure pressure decline and
the pressure response during pumping. Often, it is
5200 20
Bottomhole Treating Pressure (BHTP) easiest to match the pressure decline first and then
Total Near-Wellbore Friction
5100 16 adjust to match the pressure while pumping.
Closure Pressure
Pressure, psi
Rate, bpm
5000 12
Instantaneous Shut-In One method to determine fracture closure is to
4900
Pressure (ISIP)
Gauge BHTP 8 produce a plot of BHTP against the square root of time.
Because fluid loss through the fracture face is dependent
4800 Slurry Rate 4
upon the square root of time, in theory when the pressure
4700 0 data is plotted against t, there should be a straight line
24 25 26 27 28 29 30 31 32 33 34
while the fracture is open and a curve after it has closed.
Elapsed Time, mins
This works well for pressure-independent fluid loss.
Figure 4-12 Example minifrac pressure decline, However, when the fluid loss is dependent upon pressure,
illustrating the parameters (italics) that can be derived
from its analysis
the fluid loss rate will decrease as the pressure declines,
making the relationship much more complex.
1. Quantitative assessment of near-wellbore friction, Because finding fracture closure is all about finding
from the difference between pinj (BHTP) and pISI the point on the decline curve where the gradient changes,
(ISIP). Note that this may not be constant, as it makes sense to plot the gradient (or derivative) itself at
perforations can increase in diameter (erosion) and the same time. Often, it is easier to spot the change in
number (opening of blocked perforations), while gradient from the derivative plot than from the decline
tortuosity is a dynamic phenomenon, changing itself. Figure 4-13 shows the pressure decline from Fig.
with pressure, rate, time and viscosity. 4-12 on a square root time plot, complete with derivative
2. ISIP. At shutdown, when all friction goes to and identification of closure pressure.
zero, pISI = pf the pressure of the fluid inside
5100 80
the fracture. This is needed for the calculation Fracture Closure
of net pressure.
Derivative dp/dt0.5, psi/mins0.5

5050 Pressure Decline 40
3. Fracture closure. Closure pressure is the pressure
on the decline plot at the point when the fracture
Gauge BHTP, psi
5000 0
closes. This is usually marked by a change in
Derivative
gradient, indicating a change from Darcy linear 4950 -40
flow through the fracture faces to Darcy radial

flow from the wellbore. Sometimes this change in 4900 -80
gradient can be hard to spot, and there is a wide

4850
variety of different plots and methods available to 0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3
help with this process (see below). Square Root of Time Since ISIP, mins0.5
4. Net pressure. This is the difference between pISI Figure 4-13 Example minifrac pressure decline square
and pc. root time plot, showing derivative and fracture closure
5. Fluid leakoff. After the closure pressure has been
identified, the time taken for the fracture to close Nolte G-Function Analysis. Nolte (1979, 1986a)
can be measured. Analysis during pumping will developed a method for analyzing minifrac pressure
give the fracture geometry and, hence, area. The declines that has direct application to the three 2-D
107
Modern Fracturing
frac models (KGD, PKN and radial see Section 4- The actual value of used for the extrapolation is
6.2.2). This analysis is based on the use the G function dependent upon the fluid efficiency and n. Values
to help identify fracture closure and, from this, the tend to be almost always in the region of 0.5 to 0.7;
fluid leakoff and fracture geometry. in practice 0.6 is often used. Also, given that n is
Nolte derived the following relationships for the often variable, a quicker method is just to take the
decline curve: average of the upper and lower expressions for g(tD).
3/ 2
4 / 3(1 + t D ) t D3 / 2 As shown in Fig. 4-14, as tD increases, the difference
, (4-17)
g (t D ) = 1/ 2
between the upper and lower boundaries becomes
(1 + t D ) sin (1 + t D ) + t D
1 1 / 2
smaller and eventually becomes negligible compared
with to the accuracy of the rest of the system.
t t Nolte G time is then a function of tD such that:
i
t D =
, (4-18)
ti
G (t D ) = g (t D ) g (t D = 0).
(4-20)
where t is the time since he start of fluid
injection and ti is the injection time. The upper Note that for = 1 and = 0.5, g(tD = 0) is equal to 4/3
part of Eq. 4-17 represents the upper boundary, and /2 respectively.
and the lower part represents the lower boundary. A typical plot of a pressure decline against Nolte G
In practice, to find the actual value of g(tD), time is shown below in Fig. 4-15.
both values are calculated, and an extrapolation
is made based on the power law exponent of the
fracturing fluid (n) and the fracture geometry Additional Fracture Extension
(radial, PKN or KGD). (Note: When calculating
from the lower boundary, the trigonometric function
BHTP
Fracture Closure
is in radians, not degrees.) Ideal ISIP
5
Lower (B 0.5, g (%t D ) (1 %t D ) sin -1 (1 %t D )-1/ 2 %t D1/ 2 )
Gc
4
0
G (tD)
3
g(tD)
Figure 4-15 Idealized Nolte G time pressure decline plot.

The match pressure is the gradient of the straight line
2 g (%t D 0) Q / 2
section in the middle of the decline before closure.
1 g (%t D 0) 4 / 3
Upper (B 1, g (%t D ) 4 / 3[(1 %t D )3 / 2 - %t D 3 / 2 ]) Figure 4-15 illustrates three important points. First,
0
0.01 0.1 1 10
the ISIP recorded using field data may be artificially
tD
high, due to the effects of fracture storage and fluid
Figure 4-14 Graph showing the variation of g(tD) with tD friction. Second, there is a period of constant gradient
for Nolte analysis before the fracture closes, which is often referred to as the
match pressure (pm) and has pressure units (as G time is
The extrapolation is performed between two dimensionless). This is an important parameter in Noltes
values of the variable . At the lower boundary, = minifrac pressure decline analysis. Finally, closure occurs
0.5 and at the upper boundary = 1. The actual value when the decline pressure deviates from this constant
for is given as: gradient. At this point G(tD) = Gc.
It should be noted that if the closure time equals the
(2n '+ 2) / (2n '+ 3) : PKN
(4-19) pump time, then Gc = 1.
= (n '+ 1) / (n '+ 2) : KGD
From the g(tD) time at closure [ = g(tcD)], the fluid
(4n '+ 4) / (3n '+ 6) : Radiaal.
efficiency can be determined as follows:
108
g (tcD ) g (t D = 0) 1 v prop Finally, Nolte G time can be used to find the

= , (4-21) fracture dimensions:
g (tcD ) 1 v prop /
(1 )Vi
where vprop is the fraction of the total fracture volume Af =
, (4-26)
2 g (t D = 0) CL rp ti
occupied by proppant. For a minifrac, vprop will be equal
to zero. Therefore: where Af is the area of one fracture wing. Given that for
g t g t = 0 the 2-D models:
( cD ) ( D )
=
. (4-22)
g (tcD )
(4-27)
This can be simplified to:
G
c
,
(4-23) then the fracture length or fracture radius
2 + Gc
can be easily found. Average fracture width can also
which is a quick and easy method for determining fluid be obtained:
efficiency. Most modern real-time data monitoring
systems can plot G function in real time, so if the ( D ) Lp i
2 g t = 0 C r t
(4-28)
wave = .
closure pressure can be determined, the fluid efficiency (1 )
can be easily calculated from Eq. 4-23. When using Nolte analysis, remember
The fluid loss coefficient (constant, pressure that it is based on 2-D fracture geometry. An
independent) can be calculated as: example of Nolte G function analysis can be found
in Martin (2005).
p
CL = m s X , (4-24)
rp ti E '
4-2.2 The Role of Advanced Technology
where pm is the match pressure (see Fig. 4-15), s a in Design, Execution and Evaluation
geometry-dependent factor (see Eq. 4-25, below), rp is
the ratio of fracture area in permeable formation over Hydraulic fracturing is a technology-driven process.
total fracture area (i.e., net to gross area ratio for the Fluid systems, proppants, equipment, simulators and
fracture), E is the plane strain modulus (see Section fracturing theory are all in state of constant update,
4-3.1.3) and X is a factor dependent upon which and staying up to date with the latest and greatest
geometry model is being used (see Section 4-6.2.2), can often become a time-consuming process. The
such that for KGD, X = 2xf, for PKN, X = hf and for following list illustrates some of the major advances
radial, X = (32Rf /32), where Rf is the fracture radius in fracturing over the last 15 years.
for the radial fracture model.
(2n '+ 2) / (2n '+ 3 + a ) PKN 4-2.2.1 Recent Advances and Breakthroughs
s 0.9 KGD
(4-25) Advances in Fluids Tecchnology. Major
2
(3 / 32) Radial
advances in fracturing fluid technology have been
where n is the power law exponent for the fluid and made over the past 15 years or so. This has been
a is a variable describing how constant the viscosity of driven by a growing recognition of the following
the frac fluid is in the fracture, such that for a constant fluid issues: proppant transport; the potential for
viscosity, a = 1 and for a falling viscosity a < 1. Usually, damage to the proppant pack from fluid residues;
a is assumed to be 1. and reservoir compatibility, for high-permeability
Thus, not only is Nolte G time a useful tool for reservoirs where fracture-face damage is important
finding the ideal ISIP and the closure pressure, it can and for low-permeability reservoirs with respect to
also be used to find fluid efficiency and fluid leakoff the prevention of fluid retention issues. Chapter 7
(assuming a 2-D fracture geometry). covers fluid technology in detail.
109
Modern Fracturing
Proppant transport has become increasingly Resin-coated proppants have become increasingly
recognized as a major function of the fracturing fluid popular for a significant number of treatments (see
(see Section 7-2.2), thanks to the pioneering work of the Section 8-4). Resin coatings come in two types: those
Stim-Lab Rheology and Proppant Transport Consortium. designed to increase proppant (or more usually frac
The phenomena of convection and settling have been sand) conductivity by reducing point loading and the
identified, quantified and built into fracturing simulators production of fines; and those designed to prevent
(a good summary of this work was presented by Clark proppant flowback by physically adhering the grains
in 2006). In addition, the important role played by the together. Often, both of these effects are combined. As
fluids proppant transport characteristics in mitigating technology progresses, the performance of these coatings
the effects of tortuosity has also been recognized. Fluid continues to improve. Additionally, in many cases the
systems and treatments are now designed to account for problems associated with resin-contaminated frac fluids
these factors, and treatment results have reflected this. have been completely resolved.
The pioneering work performed by another of The use of low- and neutral-density proppants
the Stim-Lab consortia, the Proppant Conductivity (see Section 8-11.1) has become increasingly common
Consortium, led to a general recognition that the damage (Rickards et al., 2003). Currently, the full potential
caused by fracturing fluids to proppant packs was far of these products has yet to be realized in the field.
worse than had been previously realized (see Section Ultimately, they have the capability to completely
7-6.2). This in turn led to the advent of low-polymer- revolutionize the way treatments are performed, as
loading fluids and to the increasing use of polymer- it will no longer be necessary to use highly viscous
free fracturing fluid technology (such as viscoelastic fracturing fluids, nor to use complex blending
surfactant-based fluids and brine systems incorporating equipment to add proppant on the fly. Instead,
neutral-density proppants). Today, understanding the proppant will be pre-mixed into brine carrier fluids
regained permeability of the proppant pack is a key part of with relatively simple blending equipment.
the fracture design process. Proppant flowback is another area that has received
A more widespread understanding of issues relating considerable attention, led once again by another
to the effects of fluid imbibition, capilliary pressures of Stim-Labs consortia, the Proppant Flowback
and immobile fluid blocks, as presented by Bennion Consortium (see Sections 8-10 and 8-11.2). Various
et al. (1996), has significantly improved the ability of techniques have been used in the field, including mesh-
engineers to stimulate low-permeability formations. creating micro-fibers (Card et al., 1995), deformable
The use of low-surface-tension surfactants, methanol particles (Rickards et al., 1998) and surface coatings
and micro-emulsion additives have proved crucial added on the fly (Nguyen et al., 1998). In addition, Ely
in developing many tight gas reservoirs. In addition, et al. (1990) demonstrated that controlled flowback of
CO2-based systems (both as foams and as 100% CO2 fracturing fluid immediately after the treatment could
systems) have become increasingly popular to both help to stabilize the near-wellbore region of the fracture
reduce liquid content and increase fluid recovery. See and prevent proppant flowback.
Section 7-4.4 for more details.
Understanding Fracture Tip Effects. It had long been
Advances in Proppant Technology. Advances in recognized that the process of physically tearing the rock
understanding how proppant characteristics can affect apart at the fracture tip was consuming more energy
post-treatment production have taken a major step than was predicted by classical linear elastic fracture
forward (see Chapter 8). In a seminal paper, Vincent et mechanics (Griffith, 1921), as detailed in Section 4-
al. (1999) detailed how the effects of non-Darcy flow 3.4. Four different and not mutually exclusive theories
(see Sections 2-4.1 and 8-7.3) and multi-phase flow have been put forward to explain this phenomenon.
(see Section 8-7.4) could significantly affect the effective The first theory concerns the fluid lag effect, in which
permeability of the proppant, further reducing the viscous and surface tension effects prevent the fracturing
conductivity of the fracture. fluid (and hence the net pressure) from penetrating all
110
the way into the very narrow fracture tip (Jeffrey, 1989). by the propagation of the fracture or fractures, which
This acts to reduce the width at the fracture tip and has are assumed to occur around the fracture tip. If these
the effect of increasing the apparent fracture toughness. seismometers are placed effectively in 3-D space around
The second theory, proposed by Johnson and Cleary in the well to be fractured, the position of each microseismic
1991, involves non-linear elastic deformation of the rock event can be mapped, allowing real-time measurement
around the fracture tip, producing a pinching effect. This of fracture geometry. Tiltmeters are placed on the surface
is commonly referred to as dilatancy. The third theory and downhole to measure minute deflections in the
involves the production of a damaged zone (also referred earth, caused by the fracture. Tiltmeters are used mainly
to as the process zone) ahead of the fracture, in which for obtaining fracture azimuth. Section 6-8 has a more
micro-fractures form, absorbing additional energy (Yew detailed discussion of these subjects.
and Liu, 1993, and Valk and Economides, 1993). The
final theory concerns the plastic deformation of the Perforating for Fracturing. Of all the things that can
formation in a region around the fracture tip. Although be controlled by the engineer, the perforations probably
rocks are traditionally considered brittle materials, under have the most significant single influence on the success of
conditions of tri-axial loading they can deform and flow the treatment (Behrman and Nolte, 1999, and Pongratz
plastically, absorbing significant quantities of energy et al., 2007). When used effectively for fracturing, they
(Martin, 2000, and van Dam et al., 2000). can be used to control the point of fracture initiation.
Each of these theories is based on established Indeed, as most propped fractures will connect to the
principles, often taken from more general theories wellbore through fewer than 20 perforations, in low-
developed outside the oil industry. There is no doubt permeability formations there is little need to perforate
that these effects occur. Possibly the real answer to the rest of the net height (other than for reservoir
understanding tip effects lies in understanding how evaluation purposes). However, in high-permeability
significant each effect is in any given situation. gas formations, the fracture may become choked due to
turbulence effects, and it may be necessary to perforate
Unified Fracture Design. As previous discussed, low- longer intervals, if these will connect to the fracture
permeability formations require long, thin fractures, during proppant placement.
while high-permeability formations require short, wide
fractures. Consequently there has always been a problem Fracturing Equipment. To most people involved in
deciding the right combination of length and width for the fracturing process, the most obvious advances in
any fracture in a formation that falls between these two technology have been to the equipment used to perform
extremes. This conundrum was solved by Economides fracturing operations. Since the early 1990s, the fracturing
et al. in 2002, in introducing the concept of Proppant industry has fully embraced the rapid advances made in
Number and Unified Fracture Design. This theory states computerized control, measurement and display systems
that for a ratio of formation and proppant permeability, (see Section 9-5). These have produced significant
there is a specific ratio of propped length and propped operational and safety advantages. Fracturing pumps are
width that will produce the maximum possible no longer controlled by individual operators positioned on
production increase. This theory is dealt with in more or by the pumps. Instead, a single remote panel is placed
detail in Section 4-5. inside the control cabin or vehicle, allowing the pumps to
be run by a single operator, under the direct supervision
Microseismic and Tiltmeters. A significant step forward of the frac supervisor. Not only does this significantly
in understanding fracture geometry and azimuth has improve communication and reduce complexity, it also
been made by the use of microseismic and tiltmeter has a significant safety advantage, dramatically reducing
measurements, allowing the actual geometry of the the number of personnel exposed to the high-pressure
fracture to be mapped (Cipolla and Wright, 2000). lines. In addition, the electronic pump controls allow
Microseismic involves the use of surface and downhole more efficient operation of the equipment, prevent
seismometers to measure small seismic events produced damage to the equipment during equipment failure (e.g.,
111
Modern Fracturing
automatic engine shut-downs when oil pressure is lost) works. In spite of this need for an understanding of the
and allow the use of automatic, adjustable, over-pressure physics behind the fracturing process and the fracture
quick-to-neutral shut-downs. simulation, there is still an art to pressure matching.
In general, fracturing pumps have become larger, Some have a feeling for this process, and some do not.
as it became apparent that efficiency was gained by Pressure matching is a very powerful tool that allows
using a smaller number of larger pumps. Because the engineer to tune the fracture simulator to the
equipment reliability has increased as well, the formation. The idea is that once the simulator has been
effects of losing a pump during the treatment have tuned, further fracture simulations can be performed
not been significantly affected. with a high degree of accuracy.
Blending equipment has also seen a dramatic
advance. Today, almost all frac blenders are fully process- The Process of Pressure Matching. Pressure matching
controlled. This means that all mixing and blending is all about making the simulator predict the same
operations are run by computer-controlled positive pressure response as the reaction actually produced by
feedback systems. These adjust additive rates (liquids, the formation (see Fig. 4-16).
powders and proppant) automatically, significantly
Before After
reducing the complexity of the treatment for the blender
operator. The job is pre-programmed into the equipment,
Net Pressure
Net Pressure
and under most circumstances the blender will operate
automatically during the treatment. Manual overrides
usually allow the rate adjustments if necessary.
Job Time Job Time
Remote Data Transmission. With the advent of modern
Actual Net Pressure Calculated Net Pressure
communications technology, it is no longer necessary
Figure 4-16 Pressure matching: variables in the simulator
for the engineer to be on location during the treatment. are adjusted to make the calculated net pressure match
Treatment data can be transmitted in real time from the actual net pressure
any location to any location, often accompanied by a
voice link. The latest version of this technology uses the In Fig. 4-16, before the pressure match (left) the
Internet. The treatment is broadcast real-time on the net pressure predicted by the fracture simulator does
Web, with each control van or cabin having its own not match the actual net pressure in any way. After the
Web address. Anyone with the job-specific password pressure match has been performed (right) the computers
can log on and view the treatment. This means the pressure response prediction is very similar to that of
treatments can be monitored using standard PCs, with the actual treatment. Now, according to the theory,
no specialized software, from anywhere in the world the simulator has been tuned to the formation. This
with a high-speed internet connection. allows the frac engineer to input any desired treatment
Remote data transmission has its greatest use in remote schedule, and the simulator will be able to predict the
locations, especially offshore. It is an enabling technology fracture geometry with a reasonable degree of precision.
because it makes it easier for engineers to treat wells. There is no doubt that the advent of pressure
matching has greatly improved the success rate and
4-2.2.2 Pressure Matching effectiveness of hydraulic fracturing. Modern fracture
simulators equipped with this facility have gradually
Pressure matching is part science and part art. In order made the process increasingly user-friendly, helping
to perform a quick and efficient pressure match, it is to reduce the black art associated with fracture
essential to have a good knowledge of the fracturing engineering, as more and more engineers feel capable
process, an understanding of the various rock mechanical of designing a fracture treatment.
properties, an understanding of fracture mechanics and, However, there are some definite limitations to this
ideally, a reasonable idea of how the fracture simulator process:
112
Garbage In = Garbage Out. The computer model in each formation affected by the fracture. These variables
of the formation generated by this process is only as are Youngs modulus, stress, fracture toughness and fluid
good as the data used to create it. Poor data on items leakoff.So even for the most basic formation lithology,
such as permeability, net height, fluid properties (both the engineer will have to be able to track of a minimum
formation and fracturing fluids) and perforations can of 12 variables (four each in the zone of interest and the
make an otherwise perfect pressure match almost formations above and below).
irrelevant. Another major source of errors is the use of Of course, each fracture simulator comes complete
surface pressure data to calculate BHTP. In order to with a plethora of variables the user can adjust. In
calculate BHTP, the model first needs to calculate the fact there are so many that an engineer could vary
fluid friction pressure, something that is notoriously several hundred parameters for a complex reservoir
difficult to do for a crosslinking fluid. Variations in with several rock strata. This is for fracture simulator
fracturing fluid properties (such as those caused by and rock mechanical experts only. Unless there is a
problems with liquid additive systems or varying gel really good reason, the engineer is advised to stick to
properties) can also be very difficult to account for. the four main variables.
Therefore, the engineer should do everything possible
to acquire reliable bottomhole pressure data, such as Youngs Modulus, E. In order for the fracture to
that from a gauge or dead string. propagate it must obtain width, to a greater or lesser
extent. In order to do this, the rock on either side of the
No Unique Solution. The process of pressure fracture has to be displaced. As discussed in Section 4-3.4,
matching involves adjusting four major variables Youngs modulus defines how much energy is required to
(Youngs modulus, stress, fracture toughness and accomplish this displacement, which is a classic linear
leakoff) and many other minor variables for each elastic fracture mechanics concept. Rocks with a large
rock strata affected by the fracture, until the pressure Youngs modulus will require a lot of energy (i.e., net
response predicted by the model matches the actual pressure) to displace. In these formations, fractures tend
pressure response of the formation. This means that to be relatively narrow, and the rock is referred to as
the engineer may have 30 or 40 variables available hard. Similarly, rocks with a small Youngs modulus
for adjustment. It is therefore quite possible for require relatively little energy to produce width. In these
two engineers to get good pressure matches with formations, fractures tend to be relatively wide, and the
significantly different sets of variables. rock is referred to as soft.
A large Youngs modulus makes it harder for
The Fracture Model. The results of the pressure match the fracturing fluid to produce width. This will
are only as good as the fracture model itself. Modern make the fracture thinner, higher and longer. A
fracture simulators are tremendously advanced small Youngs modulus has the opposite effects.
the product of more than 20 years of innovation, Increasing E only in the perforated interval will have
experimentation and inspiration. However, different the effect of forcing the fracture out of the zone of
fracture simulators will still predict different fracture interest i.e., increasing fracture height. A decrease
geometries for the same input data. Which one in E has the opposite effect.
is right? This is difficult to say and the subject of
considerable debate in the fracturing industry. The In-Situ Stress, h. In-situ stress (often referred
popular conception is that one fracture simulator is to as confining stress and, in the vast majority of
good for a certain type of formation, while another is cases, horizontal stress) is the stress induced in the
good for a different type. The debate continues. formation by the overburden and any tectonic activity.
Put simply, it is pre-loading on the formation, the
The Four Main Variables. There are four main variables stress that has to be overcome (or pressure that has
that the engineer should be adjusting in order to achieve to be applied) in order to actually start pushing the
the pressure match that is to say, four main variables formation apart. The actual bottomhole fracturing
113
Modern Fracturing
pressure is the pressure required to overcome these in- Fluid leakoff is a loss of energy from the fracturing
situ stresses, plus the pressure required to propagate fluid: The total energy available for propagating the
the fracture (as a consequence of fracture toughness) fracture is equal to the net pressure multiplied by
and the pressure required to produce width. the fracture volume. High leakoff means low fracture
In the pressure matching process, increasing h volume, and vice versa. Therefore, an increase in fluid
reduces net pressure (for a fixed BHTP). This means that leakoff will tend to decrease width, height and length.
the fracturing fluid has less energy available to fracture The opposite applies for a decrease in leakoff.
the formation, and so the width, height and length of the The basic effect of each of these variables
fracture decrease. This in turn means that the volume of when applied to a fracture in a single formation is
the fracture has decreased. However, the same volume of summarized in Table 4-3.
fluid has been pumped into the formation, so an increase
in h also has the effect of increasing leakoff rate and Table 4-3 effects of increasing each of the Four Main
decreasing fracture efficiency. The opposite effect applies Pressure Matching Variables. (Note that these are the
overall effects when the change is taken in isolation (i.e.,
for a decrease in in-situ stresses. no other changes take place). It also assumes that the
fracture is unaffected by boundary layers above and
Fracture Toughness, K1c. Strictly speaking, K1c is the below.)
critical stress intensity factor for failure mode 1 (see Effect of an Increase in Selected Variable
Variable
Section 4-3.4 and more specifically fig. 4-30). However, Height Length Width
Net
Pressure
it is commonly referred to as the fracture toughness and
Fracture
is a measure of how much energy it takes to propagate Toughness, K1c
Decrease Decrease Increase Increase
a fracture tip through a given material. In hydraulic Youngs
Increase Increase Decrease Increase
fracturing, where the energy needed to propagate the Modulus, E
fracture comes in the form of fluid pressure, fracture In-Situ
Decrease Decrease Decrease Decrease
toughness is that portion of the available energy required Stress,
to physically split the rock apart at the fracture tip. Fluid Leakoff
Decrease Decrease Decrease Decrease
Rate, qL
Note that some fracture models have moved away
from the concept of fracture toughness and instead
model nonlinear elastic effects at the fracture tip as being Table 4-3 should be used with caution because
more significant. In such models, variations in Youngs it applies only when the fracture is confined within
modulus and in-situ stresses are far more significant than a single formation. If the fracture propagates
variations in fracture toughness. into separate formations above and/or below the
productive interval, then an increase in (for instance)
Fluid Leakoff Rate, qL. The fluid leakoff rate can be fracture toughness will make it harder for the fracture
controlled by altering a number of variables, depending to propagate through the main formation, forcing
upon the fracture simulator being used, and the fluid the fracture up and down. So, in this instance, an
leakoff model being employed: isolated increase in a property in just one formation
can actually increase the fracture height.
Pressure differential (fracturing fluid pressure minus
pore pressure) The Effect of Poissons Ratio. The Poissons ratio is
Formation permeability at the same time important and largely irrelevant to
Formation porosity pressure matching. It is important because it has a
Formation compressibility major effect on defining the horizontal stresses in a
Formation fluid viscosity formation. However, in most cases, the engineer will
Fracturing fluid filtrate viscosity be determining these stresses from pressure data, not
Fracturing fluid wall-building coefficient from the Poissons ratio data. In most fracture models,
Spurt loss the Poissons ratio is used in the form (1 - 2) to modify
114
the Youngs modulus (i.e., E/(1 - 2) the plane strain user enters the relevant formation data and treatment
modulus, see Section 4-3.1.3). This means that a large schedule, which can be loaded from a previously created
change in the Poissons ratio, say from 0.25 to 0.35, data file. The treatment starts, and the computer starts
will only change (1 - 2) from 0.9375 to 0.8775 (i.e., to collect the data. As the treatment progresses, the
a 40% increase in produces only a 6.4% decrease in simulator models the created fracture. The model will take
(1 - 2). Therefore, it is not worth varying the Poissons fluid, proppant, formation and wellbore characteristics
ratio during a pressure match. from the input model, and the pump rate, pressure and
proppant concentration from the real-time data. Using
4-2.2.3 Getting Closer this data, the simulator will model the fracture that has
to Understanding Fracture Geometry already been created, constantly updating as more data is
collected. This enables two operations:
Exactly how well do we understand fracture geometry? The engineer can perform a pressure match with the
Todays simulators are the products of 30 years of data that has already been collected by the simulator,
innovation, invention and discovery, from some of the until the net pressure predicted by the computer
most able minds in the industry. However, we must ask matches up with the actual net pressure.
ourselves just how reliable these models are. The engineer can instruct the simulator to run the
The first thing we must consider is the model job until completion, predicting the characteristics
itself. Currently there are four main models in use of the fracture based on the ongoing pressure match.
three simulators described as lumped-parameter 3- For the treatment schedule, the simulator will use
D models and one fully 3-D grid-based model. Feed the actual treatment data as far as possible, and then
the same input data into each model, and they will project forward until the end of the job using the
produce dramatically different results. So which one remaining input treatment schedule. This allows
is correct? This is an area of considerable debate and the engineer to predict the fracture characteristics
controversy in the industry. based on the most accurate data possible. This
Fortunately, two methods that are independent of process can be taken one step further: The engineer
the fracture simulators can be used to remove a lot of the can alter the remaining treatment schedule and
uncertainty associated with modeling. The first method predict the fracture characteristics based on this
uses microseismic measurements to produce an estimate revised schedule. Thus the engineer can redesign
of fracture height and length (as outlined in Section 4- the treatment schedule on-the-fly.
2.2.1 and discussed in detail in Section 6-8). The second
method is to use post-treatment pressure transient Limitations of Real-Time Modeling. The ability to
analysis (see Section 3-7) to evaluate fracture dimensions redesign on the fly, while usually not very popular
(Agarwal, 1980; Economides and Nolte, 1987; Arihara et with the frac crew, is a very powerful tool, with the
al., 1996; and Cipolla and Meyerhofer, 2001). following caveats:
Do not over-react to short-term trends. All fracture
4-2.2.4 Real-Time Analysis simulators treat formations as homogenous
materials with uniform rock mechanical properties
Some fracture simulators have a facility that enables throughout. In reality this is usually not the case.
fracture modeling in real-time. This is a very The fracture is constantly propagating through rock
powerful tool that under the right conditions can with varying properties, producing unpredictable
enable the engineer to redesign the main treatment variations in the net pressure plot. The engineer
on-the-fly, as it is pumped. needs to find an average value for each of these
The modeling computer is set up to receive data properties, such that the simulators predicted net
from either the data processing computer or the pressure curve follows the trend (an average value)
engineers computer. The user then runs the fracture of the job plot, but does not necessarily follow every
model, selecting the real-time data input option. The minute rise and fall in pressure.
115
Modern Fracturing
The engineer must be able to react quickly when Note that this is very similar to the formula for
a short-term trend has become a long-term trend. calculating pressure. Stress and pressure have the
When this happens, its time to start adjusting same units and are essentially the same thing
formation properties. stored energy. The main difference between the
Real-time modeling is only effective on long two is that in liquids and gases, the material will
treatments, where the engineer has time to spot flow away from an applied force until the force
long-term trends, adjust the model and still be and stress (or pressure) is the same in all directions
able to make change the treatment schedule (i.e. equilibrium has been reached). However,
in time for the changes to have some effect. If solids cannot deform in such a manner, so these
the job is too short, the crew can be pumping materials will always have a plane across which
the displacement before the engineer has the stresses are at a maximum. They will also have
finished the pressure match. a plane perpendicular to this, across which the
The problem above is exacerbated if the wellbore stresses are at a minimum.
volume represents a significant part of the treatment. Properties such as mass and volume
If this is the case, the treatment can be close to are said to be scalars they require only a
displacement before the proppant has even passed magnitude to define them. Quantities such as force
into the fracture. In such cases, there is little point and velocity are vectors they require not only a
in modeling the fracture real-time. magnitude, but also a direction in which they are
acting in order to be fully defined. Stress takes this
4-2.3 From Fracturing a Single Vertical Well one step further as a tensor property it can only
to Complex Well-Fracture Architecture be fully defined by a magnitude and an area across
which it is acting.
As it was discussed in Section 2-5.3, the industry
moves away from the single vertical hole paradigm Strain. Strain is measure of how much the material
and develops towards complex well-fracture has been deformed when a stress is applied to it.
configurations. It is now common practice to place As the force, F, is applied in the x-direction, the
multiple fractures into one wellbore (vertical or original height of the block of material, x, changes
horizontal) or even to combine the draining capacity by x (so that the new height is x - x). The strain
of multiple holes and multiple fractures. A draining in the x-direction, x, is given by:
system can be still characterized by its overall x
productivity, and that is how it should be analyzed x = .
(4-30)
x
and optimized. In Section 4-6 we will present a
significant tool for such analysis. Note that the strain is defined in the same
direction as the applied force, F, and perpendicular
4-3 Rock Mechanical Characteristics to the plane across which the stress acts. Strain is
important because this is the way we measure stress
4-3.1 Basic Definitions by observing the deformation of a known piece of
material. Strain is dimensionless.
4-3.1.1 Stress and Strain
4-3.1.2 The Poissons Ratio
Stress. If a force, F, is acting on a body with cross-
sectional area, A, perpendicular to the direction of The Poissons ratio, , is a measure of how much a
action of the force, then the stress, , induced in this material will deform in a direction perpendicular
body is equal to the force divided by the area: to the direction of the applied force, parallel to the
F plane on which the stress induced by the strain is
= . (4-29) acting (see Fig. 4-17).
A
116
F
y

E= .
(4-33)

x Because strain is dimensionless, E has the same units

as stress. Youngs modulus is a measure of how much
a material will elastically deform under a load. This is
another term for hardness.
On a more fundamental level, if stress and
pressure are closely related (applying a pressure to a
surface induces a stress), then in fracturing we can
think of Youngs modulus as a measure of how much
a material (i.e. rock) will elastically deform when a
pressure is applied to it. Because pressure is stored
y + y energy, E is also a measure of how much energy it
takes to make the rock deform.
Figure 4-17 Application of force F in the x-direction will Materials with a high Youngs modulus (e.g., glass,
also produce a deformation in the y-direction
tungsten carbide, diamond and granite) tend to be
very hard and brittle (susceptible to brittle fracture).
The strain in the x direction, x, is given by Eq. Conversely, materials with a low E (e.g., rubber,
4-30 (see Section 4-3.1.1). Similarly, the strain in the polystyrene foam and wax) tend to be soft and ductile
y direction is given by: (resistant to brittle fracture).
y
y = .
(4-31) Elastic vs. Plastic. Elastic deformation is reversible: If
y
the force (or pressure or stress) is removed, the material
Note that this value is negative, a result of the way returns to its original size and shape. If so much force
we define the forces and the direction the forces act. is applied to a material that it passes beyond its elastic
Compressive strain is positive; tensile strain is negative. limit, then the material will start to plastically deform.
This is permanent. A good illustration of this is the small
The Poissons ratio, , is defined by: spring from a ball point pen. When the spring is lightly
stretched, it will return to its original shape. However,
y
= .
(4-32) if the spring is stretched too far, it will be permanently,
x
or plastically, deformed. Youngs modulus only applies
The Poissons ratio is an important factor in determining to elastic deformation. As a group of materials, rocks
the stress gradient of the formation but is less important tend not to plastically deform very much. Instead they
in defining fracture dimensions, although it does have will elastically deform and then fracture if the stress
some effect. By definition, the Poissons ratio is always gets too high. Notable exceptions to this are salt beds,
less than 0.5 (otherwise a uniaxial compressive stress soft carbonates (e.g., chalk) and young coals.
would result in an increase in volume) and typical
values for rocks are between 0.2 and 0.35. The Poissons Static Youngs Modulus. This is the standard
ratio is, of course, dimensionless. measure of E and is applicable to hydraulic fracturing.
The material is being deformed slowly and in
4-3.1.3 Youngs Modulus only one direction.
Youngs modulus, E, (also known as modulus of Dynamic Youngs Modulus. This is the rock
elasticity or elastic modulus) is defined as the ratio of property measured by special sonic logging tools.
stress over strain: The material is no longer static; instead, it is
117
Modern Fracturing
being continually stretched and then compressed This property is important in hydraulic fracturing
rapidly. There is often a significant variation between because this stress level has to be overcome in order
static and dynamic values for E due to a process to split the rock. Usually, the frac gradient (the
known as hysteresis. Hysteresis is a retardation of pressure i.e., stress needed to fracture the rock)
the effects of forces, when the forces acting upon has two components: the stresses induced by the
a body are changed (as if from viscosity or internal overburden, and the tensile strength of the rock.
friction). In this situation, it represents the history See Section 4-3.2 for a more detailed explanation
dependence of the physical systems. In a perfectly of in-situ stresses.
elastic material, elastic stress and strain are infinitely Materials also have a compressive strength,
repeatable. In a system exhibiting hysteresis, the which is the compression load beyond
strain produced by a force is dependent upon the which a material will fail. Failure mechanisms
magnitude of the force and the previous strain history are more complex because the material is
(see Section 4-3.5.3). often compressed in several directions at once.
Generally, rocks are much stronger in compression
Plane Stain Modulus. In hydraulic fracturing, the than in tension, a fact we take advantage
strain in the direction perpendicular to the fracture of during fracturing.
plane (i.e., the direction in which fracture width is
produced) is effectively zero. This is because in this Shear Modulus. The shear modulus is similar to
situation the denominator in Eq. 4-30 (the x) is so Youngs modulus, except that it refers to the material
large that the strain is effectively zero, even though being in shear rather than in compression or tension.
there has been measurable material deformation. This It defines how much energy is required to elastically
is known as plane strain, which implies that strain deform a material in shear (see Fig. 4-18).
only exists in a direction perpendicular to the direction
x
in which strain is zero. To account for this anomaly,
fracture models use the plane strain modulus, E, to F
calculate fracture width:
E
E'=
. (4-34)
2
(
1 )
h
In fracturing, Youngs modulus will typically
have values ranging from as low as 50,000 psi
(for a shallow, very soft chalk or weak sandstone)

to as high as 6,000,000 psi for deep, tight,
b
shaley sandstone. It should be noted that Youngs a
modulus may not be constant in weak or Figure 4-18 Force F applied to produce a shear stress
unconsolidated formations.
The shear stress, , is given by:
4-3.1.4 Other Rock Mechanical Characteristics
F
= , (4-35)
A
Tensile Strength. The tensile strength of a material is
the level of tensile stress that is required to make the where A is the area of the block of material parallel to
material fail. Usually as stress is applied, the material the line of action of the force F (this is the plane along
will elastically deform (reversible), plastically deform which the shear stress acts) and is equal to a b.
and then fail. In most rocks plastic deformation The shear strain, , is defined as:
is negligible and the material will, for all practical
x
purposes, elastically deform and then fail. = = tan .
(4-36)
h
118
Therefore, the shear modulus, G, is equal to the shear Relationships Between the Four Elastic Constants.
stress divided by the shear strain: The four main elastic constants Youngs modulus,
shear modulus, bulk modulus and Poissons ratio are
Fh
G= =
. (4-37) all related. If two of these material properties are known,
Ax
the other two can be deduced:
Bulk Modulus. This is another elastic constant,
which defines how much energy is required to E = 3K (1 2 ) ,
(4-40)
deform a material by the application of external E
K= ,
pressure. This is a special form of compressive 3 6 (4-41)
stress, in which the applied compressive stress is
E
equal in all directions. Imagine a block of material, G= , and (4-42)
2 + 2
which originally has a pressure p1, applied to it
and has a volume V1. This pressure is increased to = 3K E . (4-43)
p2, which causes the volume to decrease to V2, as 6K
illustrated in Fig. 4-19. The increase in bulk stress Therefore, if the Youngs modulus and the Poissons ratio
is the same as the increase in pressure, p2 p1. The are known, the shear modulus and the bulk modulus
bulk strain is equal to the change in volume, V2 V1 can be deduced. Thus, fracture simulators only require
divided by the original volume, V1. Thus, the bulk the input of E and .
modulus, K, is given by:
4-3.1.5 Hookes Law
( p2 p1 ) V ( p p1 )
K = = 1 2 and (4-38) Stresses under the ground do not act only on a single
(V2 V1 ) V1 (V2 V1 )
plane. There is a complex three-dimensional stress
dp
regime. To simplify things, stresses are usually resolved
K = V
. (4-39) into three mutually perpendicular stress components in
dV
the x, y and z directions.
Additionally, because the stresses are three-
dimensional, so are the strains. The elastic
relationship between these stresses and strains in
the mutually perpendicular directions, x, y and z, is
V1
governed by Hookes law:
V2
1
x = x ( y + z ) .
(4-44)

E
Equation 4-44 means that the strain in any direction

can be found in a three-dimensional stress regime,
provided the stress in that direction and the stresses
in two other mutually perpendicular directions are
Figure 4-19 Volume changes from V1 to V2 as pressure also known. This will have implications for the in-situ
increases from p1 to p2 stresses as discussed in Section 4-3.2.
The minus sign is introduced into the equation 4-3.1.6 Failure Criteria and Yielding
because the term V2 V1 will always be of the opposite
sign to the term p2 p1. The bulk modulus is therefore Under conditions of uniaxial (one-dimensional) loading,
a measure of how much energy it takes to compress a material failure is simple: The material elastically deforms
material using externally applied pressure. until a yield stress occurs, and then plastically deforms
119
Modern Fracturing
until the material fails. The point at which the strain The three-dimensional state of stress at any given
changes from elastic to plastic is referred to as the yield point is described by the following matrix:
point or yield stress. The maximum stress the material x xy zx

can withstand under plastic loading is often called the
[ ] = .
xy y yz
ultimate tensile (or compressive) stress or strength. For yz z (4-45)
zx
brittle materials such as most rocks, plastic deformation
under uniaxial loading is almost nonexistent. The material Using standard three-dimensional transformation
will, for all practical purposes, deform elastically and fail techniques (see Budynas, 1999, for example),
(i.e., the yield stress ultimate tensile stress). these stresses can be resolved to a coordinate
However, under conditions of triaxial loading, such system such that:
as those experienced by subterranean rock formations, ' x 0 0

the situation is much more complex. To start with, the
[ ] = 0 ' y 0 .
three principal mutually perpendicular stresses (here 0 0 'z (4-46)

referred to as x, y and z), will induce shear stresses
on planes between the directions x, y and z. This means These are the three principal stresses, 1, 2 and
there are six principal shear stresses (xy, xz, yx, yz, zx 3, defined so that 1 > 2 > 3 (note that tension is
and zy) and therefore nine principal stresses overall (see positive and compression is negative, so a small tensile
Fig. 4-20). It must also be recognized that under triaxial load is greater than a large compressive load). To put
compressive loading, failure occurs by shearing, not by this another way, the direction of these three principal
conventional tension or compression. This means that stresses has been defined such that they are in a direction
under triaxial compressive loading, if a suitable balance that produces no perpendicular shear stresses. In all
is maintained among the three principal stresses, the other directions, shear stresses will exist.
shear stress can be kept below the level required for
failure, and the material can essentially withstand Triaxial Failure Criteria. Several methods have been
infinite stress. Such a situation occurs when a solid developed for deciding when a material will fail under
block of material is subjected to hydrostatic pressure. triaxial loading. The main ones used today are Tresca
Finally, it must also be recognized that under triaxial (maximum shear stress), von Mises (maximum-energy
compressive loading, even materials traditionally of distortion) and Mohr-Coulomb. Both Tresca and von
considered brittle (such as most rocks) can display Mises apply to ductile materials. The Mohr-Coulomb
substantial plastic deformation before failure. failure criterion, which is more applicable to brittle
materials, will be dealt with in more detail below.
Ty
Mohr-Coulomb Failure Criterion. This failure

Uyz Uyx criterion, as illustrated in Fig. 4-21, relies on the
Uxy use of two key factors that must usually be obtained
Uyz empirically: SUC (the ultimate strength under uniaxial
Tz
Uzx compression) and SUT (the ultimate strength under
Uxz
Tx Tx uniaxial tension). For rocks, as for most brittle materials,
Uzx Uxz
Tz SUC is usually many times greater than SUT.
Uzy
Uxy Figure 4-21 looks complex but can be easily
Uyz explained. To construct this diagram, first plot two
Uyx y
Mohrs circles based on the values of SUT and SUC.
These are the two dotted circles in Fig. 4-21. Then
z
Ty x draw tangents to these lines as illustrated; these are
the failure lines. Finally, construct a Mohrs circle
Figure 4-20 General state of stress (after Budynas, 1999) based on the magnitude and difference between 1
120
and 3, as illustrated by the solid circle in Fig. 4- H

21, remembering that 1 > 3. If the solid circle is v = n g hn ,
(4-47)
0
small enough to remain inside the tangents, then the
material will not fail. However, if the difference and where n is the density of rock layer n, g is the acceleration
magnitude of 1 and 3 are such that the solid circle due to gravity and hn is the vertical height of zone n,
goes outside the tangents and into the region marked such that h1 + h2 + ..... + hn = H.
Failure in Fig. 4-21, then the material will fail. In This is often expressed more simply in terms of an
Fig. 4-21 the stresses are such that the material is just overburden gradient, gob:
at the point of failure, with the principle stress Mohrs
circle falling right on the tangents. v = g ob H .
(4-48)
U
4-3.2.2 Horizontal Stresses
Failure
As previously discussed, a complex three-dimensional
Stability
stress regime exists in most subterranean rock
sUC T3 T1 sUT formations. To simplify things, stresses are usually
T
resolved into three mutually perpendicular stress
components: the vertical stress, v, and two horizontal
Failure
stresses, h,min and h,max.
Additionally, because the stresses are three-
dimensional, so are the strains. The elastic relationship
between these stresses and strains in three mutually
Figure 4-21 Mohr-Coulomb failure criterion for brittle
materials perpendicular directions, x, y and z, is governed
by Hookes law (Eq. 4-44 and Section 4-3.1.5).
For the case of elastic deformation with no outside
4-3.2 In-Situ Stress and Fracture Orientation influences (such as tectonics) in a homogenous
and isotropic formation, there are two important
In-situ stresses are the stresses within the formation, things to note. First, h,min = h,max because the
which act as a load (usually compressive) on the stresses will be symmetrical on the horizontal plane.
formation. They come mainly from the overburden, Second, as each individual unit of rock is pushing
and these stresses are relatively easy to predict. However, against another identical unit of rock with the same
factors such as tectonics, volcanism and plastic flow horizontal force, h,min = h,max = 0 (i.e., no deformation
in adjacent formations can significantly affect the in- on the horizontal plane).
situ stresses these factors are much harder to predict. Therefore, from Hookes law, setting x = h = 0,
In addition, the act of creating a localized anomaly x = y = h and z = v, then:
such as an oil well can also significantly affect the

stresses in a specific area.
h = v . (4-49)
1
4-3.2.1 Overburden Stress Equation 4-49 indicates that the Poissons ratio can
have a considerable influence on the horizontal in-
The stresses due to the overburden (also called the situ stresses. Additionally, given that < 0.5 and that
vertical stress) are simply the sum of all the pressures it is usually in the region of 0.25 for rocks, Eq. 4-
induced by all the different rock layers. Therefore, if 49 also shows us that unless there are some extreme
there has been no external influences such as tectonics outside influences, the horizontal stress will always be
and the rocks are behaving elastically, the vertical less than the vertical stress. This has implications for
stress, v, at any given depth, H is given by: fracture orientation (see Section 4-3.2.4).
121
Modern Fracturing
4-3.2.3 The Effect of Pore Pressure 4-3.2.4 Fracture Orientation

and Fracture Gradient
Fractures will always propagate along the path of least
After the work of Biot (1956) the vertical stress (Eq. 4- resistance. In a three-dimensional stress regime, a
48) is usually modified to allow for the effects of pore (or fracture will propagate so as to avoid the greatest stress
reservoir) pressure, such that: and will create width in a direction that requires the least
force. This means that a fracture will propagate parallel
v = g ob H pr ,
(4-50) to the greatest principal stress and perpendicular to the
plane of the least principal stress. This is a fundamental
where gob is the overburden pressure gradient (usually principle; therefore, the key to understanding fracture
between 1.0 and 1.1 psi/ft) and is Biots poroelastic orientation is to understand the stress regime.
constant, which is measure of how effectively the
v
fluid transmits the pore pressure to the rock grains. h, max
depends upon variables such as the uniformity and
sphericity of the rock grains. By definition is always
h, min
between 0 and 1, usually assumed to be between 0.7
and 1.0 for petroleum reservoirs. We can see from Eq.
4-50 that the vertical stress is reduced by the action of
the reservoir pressure, as if the reservoir fluids partially
lift the overburden. This value is often referred to as
the effective stress. It should also be noted that as the
reservoir pressure declines (i.e., during depletion), the
effective vertical stress can increase.
The horizontal stress is also modified to allow for
the effects of pore pressure. As a result of work by
Handin et al., 1963 (based on previous investigations Figure 4-22 Fracture propagation perpendicular to the
minimum horizontal stress
by van Terzaghi, 1923; and on Biot, 1956), Eq. 4-
49 is generally modified to allow for the effects Propagation perpendicular to the least principal
of the pore pressure: stress (usually h,min) means that the fracture will almost
always propagate on a vertical plane (see Fig. 4-22).

h = (v pr )
+ pr . (4-51) However, there are some exceptions.
1
Equations 4-50 and 4-51 define the magnitude
Under most circumstances, h will decrease as pore of horizontal and vertical stresses in undisturbed
pressure declines. formations. The horizontal stresses are induced by
the vertical stresses. There is evidence to suggest
Frac Gradient. Equation 4-51 gives the magnitude of that these horizontal stresses get locked into place
the horizontal stresses in the formation, provided the (Economides and Nolte, 1987) and remain relatively
two horizontal stresses are equal and the formation is constant regardless of what later happens to the
uniform. In order to determine the stress at which the vertical stress. Figure 4-23 illustrates what happens
formation will fracture, it is often necessary to add a when the vertical stress is reduced.
component (T) to account for the tensile strength of If formation is lost due to erosion or glaciation,
the rock (often minimal due to brittle fracture effects the overburden stresses are reduced. However, because
see Section 4-3.4) and the effects of tectonics, etc. the horizontal stresses are locked in, they have not
The fracture gradient, gf, therefore becomes: been reduced. Therefore, there is a region, close to the
new surface, where the horizontal stresses are greater
1 (v 2 pr )
g f =
+ pr + T . (4-52) than the vertical stresses. This means that the fracture
H 1
122
will propagate horizontally a pancake frac. Thus, wellbore, it will eventually reach a point at which the
in shallow formations in areas with a history of surface normal stress regime becomes more significant than the
erosion, horizontal fractures are not only possible they near-wellbore stress regime. At this point, the fracture
are in fact, likely. This does not apply to formations that will change orientation. Sometimes this re-orientation
are very weak or unconsolidated because stresses cannot can be quite sudden, resulting in sharp corners in the
be locked in if the rock strata have no strength. fracture, which can cause premature screenouts.
Magnitude of In-Situ Stress Magnitude of In-Situ Stress

4-3.2.5 Stress Around a Wellbore
Formation Lost
Due to Erosion and Breakdown Pressure
Depth
Depth
v v
A wellbore is essentially a pressure vessel with a very
thick wall. Consequently, the same theories that apply
h h to thick walled pressure vessels can also be applied
to wellbores, providing that the in-situ stresses and
Original Stress Regime Stress Regime after Loss
of Height by Erosion reservoir pressure are accounted for. Figure 4-24
illustrates how the stresses at any given point near the
Figure 4-23 Changes in stress regime due to erosion or
glaciation
wellbore can be resolved into three principal stresses:
vertical, radial and tangential stresses. Once again,
Another consequence of this phenomenon these are perpendicular to each other.
is that in formations where the v and the h are
v
approximately equal, it can be very hard to predict
fracture orientation.
The action of outside forces, such as tectonics t
and volcanism, can also significantly affect fracture

r r
orientation. The extra stresses imposed by the movement

of the Earths crust do not usually alter the overburden t
stress but can significantly alter the horizontal stresses.
In addition, formations can sometimes be bent and
v
buckled. For example, in Barbados, there is a formation
that has experienced so much tectonic stress that it now
Figure 4-24 Three-dimensional stresses around a
runs vertically. Its stresses have been locked into place,
wellbore
so now the original vertical stress is horizontal, and vice
versa. So the fractures propagate horizontally. The vertical stress, v , is as defined in Sections 4-
3.2.1 and 4-3.2.3 (Eqs. 4-47 and 4-50). From Deily
Influence of Wellbore Orientation. Drilling a well can and Owens (1969) we can get expressions for the
significantly alter the stress regime in an area around radial and tangential stresses induced by a pressure in
the well. How far this effect reaches from the wellbore the wellbore, pw, at a radius, R, from the center of the
depends upon the Youngs modulus of the formation. well (wellbore radius r w):
Hard formations (high E) tend to transmit stress
more easily than soft formations (which will deform r2 r2
r = pw ( pr + pw pR ) w2 + 1 + w2 v
to reduce the stress). Therefore, hard formations are R R (4-53)
affected more than soft formations.
In the area around the wellbore the area affected by and
the new stress regime fractures may propagate parallel r 2 2
to the wellbore, even if the wellbore is highly deviated t = ( pw pr ) w2 + 1 rw ( p p ) , (4-54)

R 1 R 2 ob r
or horizontal. As the fracture propagates away from the
123
Modern Fracturing
where pR is the pressure at a radius R from the 80 and 90% of the energy in the fracturing
center of the well, is Biots poroelastic constant fluid will be used simply to overcome the stresses
(Biot, 1956), pr is the reservoir or pore pressure in the formation.
and pob is the overburden pressure ( vertical In order to produce a fracture in the formation,
stress without the effects of pore pressure, see two forces have to be overcome. The first force is
Deily and Owens, 1961). the in-situ stress, which is defined in Eqs. 4-49
At the wellbore face, the stresses due to wellbore and 4-51 when there are no external influences
pressure will be at a maximum. Also, this is by such as tectonics, etc. The second force to be
definition the point at which the fracture initiates. overcome is the tensile strength of the rock,
Therefore, these are the stresses that interest us most. which is usually in the region of 100 to 500 psi.
At the wellbore R r w and pR pw so that: Roegiers (1987) defined the breakdown pressure in a
different manner to Eq. 4-57:
2
t =
1 ( ob
g H pr ) ( pw pr ) and (4-55)
pif , upper = 3h ,min h ,max pr + T and (4-58)
r = pw pr ,
(4-56) 3h ,min h ,max 2 pr + T
pif , lower = , (4-59)
2 (1 )
where gob is the overburden pressure gradient and H is
the vertical height, such that pob = gob H. where pif, upper is the breakdown pressure assuming no
Furthermore, Barree et al. (1996) went on to show fluid invasion into the formation (i.e., the maximum
that provided the rock does not have any significant possible theoretical breakdown pressure), pif, lower is the
tensile strength or plastic deformation, failure of the lower boundary for breakdown pressure, assuming
rock (i.e., breakdown) occurs when the tangential stress significant alteration of the near-wellbore pore pressure
is reduced to zero. Therefore, from Eq. 4-55 with t = due to fluid, and is a parameter defined by the Poissons
0 and pw equal to the breakdown pressure (or initial ratio and Biots constant, as follows:
fracture pressure), pif, rearranging gives:
(1 2 )
= .
(4-60)
2 2 (1 )
pif =
1 ( ob
g H pr ) + pr . (4-57)
If there are no significant external influences on
The breakdown pressure is the pressure the stress regime, the two horizontal stresses are equal
required to initiate a fracture from the wellbore. and Eqs. 4-58 and 4-59 can be simplified to:
Due to the effects of the stresses induced by the
presence of the wellbore, the breakdown pressure pif , upper = 2h pr + T and (4-61)
is usually greater than the fracture gradient,
which is a measure of the pressure required to 2h 2 pr + T
propagate the fracture through the formation, pif , lower = . (4-62)
2 (1 )
away from the influence of wellbore effects.
Both are usually expressed as pressure gradients The breakdown gradient is simply the breakdown
(i.e., in psi/ft or kPa/m) so that similar formations pressure, pb, divided by the TVD.
in different wells at different depths can The fracture gradient is the pressure required to
be more easily compared. The fracture gradient is a make the fracture propagate outside of the influences
very important quantity in fracturing because it is of the wellbore (the region referred to as far-field). As
the most significant contributor to the bottomhole stated above, this is often significantly lower than the
treating pressure, which in turn helps to define breakdown pressure, depending upon the viscosity of the
the surface treating pressure, the loading on the frac fluid, the reservoir pressure and the contrast between
completion and the proppant selection. Between maximum and minimum horizontal stresses.
124
4-3.3 Fracture Shape
4-3.3.1 Two-Dimensional (2-D) Fracture Geometry

wf (xf )
Until the early 1990s, fracture simulation was limited

to two-dimensional modeling. Although these hf
models had been extensively developed and refined

xf
over a period of 25 years or so, they represented a
simple, albeit elegant, approximation to simplified
fracture geometry. Three main models existed; radial,
KGD (Kritianovitch and Zheltov, Geertsma and Figure 4-26 KGD fracture geometry
DeKlerk, further refined by Daneshy) and PKN
(Perkins and Kern, Nordgren). A more detailed PKN. This fracture model was originally conceived
explanation of these models will be provided in by Sneddon (1946) and later developed by Perkins
Section 4-6.2.2. However, below is a brief, qualitative and Kern (1961), with further work by Nordgren
description of the models. (1972), Advanti et al. (1985) and Nolte (1986b).
In this model, the fracture height is again assumed
Radial. Various radial models have been developed, to be constant. However, this time there is no
but in each one the height is assumed to be directly slippage between the formation boundaries, and
related to the fracture length, such that hf = 2xf (= the width is proportional to fracture height,
2rf, the radius of the fracture). This is illustrated as illustrated in Fig. 4-27.
in Fig. 4-25. In this model, fracture width is
proportional to fracture radius.
wf (hf )
hf
rf
hf (= 2rf ) wf (rf )
xf
Figure 4-27 PKN fracture geometry
4-3.3.2 Elliptical Fracture Geometry

Figure 4-25 Radial fracture geometry
It quickly became apparent that the simple two-

KGD. This model was originally developed by dimensional fracture geometry, with either a
Khristianovich and Zheltov in 1955 and was later fixed fracture height or a radial shape, was not
modified by Geertsma and de Klerk (1969), Le adequate for many fractures. Consequently, attempts
Tirant and Dupuy (1967) and finally by Daneshy were made to extend the 2-D models to simulate
(1973). This model is illustrated in Fig. 4-26. In the fractures as ellipses (e.g., Martins and Harper,
this model, fracture height is fixed and width is 1985). Other attempts were based on extending the
proportional to fracture length. This model also radial model so that hf < 2xf (instead of hf = 2xf),
assumes constant width against height and slippage although the user still had to specify the
at the formation boundaries. ratio of length to height.
125
Modern Fracturing
Ultimately, attempts at 2-D elliptical sometimes fracture height is contained; the evidence,
fracture modeling did not find widespread use, usually based on microseismic analysis or post-treatment
as they were quickly superceded by lumped- pressure transient analysis, is hard to contradict. So what
parameter 3-D models. can cause fracture height containment?
Today, most fracture modeling is performed using
lumped-parameter 3-D simulators. These models are Stress Contrasts. The classic and often-abused formation
considerably more sophisticated than the 2-D models characteristic of stress contrasts is often used as evidence
outlined in Section 4-3.3.1 but are not fully 3-D. for fracture height containment. Without question, a
Instead, these models relate all fracture parameters significant increase in the minimum principal stress
back to a characteristic dimension, which is found (usually caused by a change in the Poissons ratio) as
using proprietary methods. These simulators model the fracture propagates up or down into a different
the fracture as two semi-ellipses, which meet on formation, can make it very difficult for the fracture
a horizontal line level with the point of fracture to extend. The fracturing fluid does not have sufficient
initiation, as illustrated in Fig. 4-28. pressure to separate the rock and without a significant
increase in pnet, will extend only laterally, rather than
grow into the high-stress region.
However, how often do these stress contrasts exist?
It is not appropriate to assume that an over- or under-
lying stratum will automatically have higher in-situ
stress than the reservoir sandstone. In fact, neighboring
layers are just as likely to have reduced stresses (in which
Upper
Semi Ellipse case we will get preferential fracture height growth) as
increased stresses. Unless there is some independent
Lower evidence to corroborate stress-based fracture height
Semi Ellipse
containment, it can not be relied upon.
Point of Fracture Initiation
Fracture Toughness and/or Tip Effects. In order for
the fracture to propagate, the rock must be physically
Figure 4-28 Lumped-parameter 3-D modeling, showing split apart at the fracture tip. In formations where this
two semi-ellipses
is easy (i.e., relatively little energy is used to do this),
Special sets of consistency equations are used to the formation is said to be brittle and has a low
ensure that the two semi-ellipses have the same length apparent fracture toughness. The opposite effect, in
and width characteristics where they meet. The main which relatively large quantities of energy are used to
differences between the models are how they find the split the rock apart, is referred to as ductility (high
characteristic length, how they ensure consistency apparent fracture toughness).
between the two halves of the fracture, and how they Although there is still considerable debate as to what
model propagation at the fracture tip. exactly is happening at the fracture tip, there is no doubt
that a significant portion of the net pressure (the energy
4-3.3.3 Limitations to Fracture Height Growth available in the fracturing fluid for making the fracture
grow) is used up at the fracture tip. How much energy is
Engineers often refer to boundary formations above used is controlled by ductile the formation is.
and below a zone of interest. Sometimes this is based If the fracture propagates from a formation that
on wishful thinking a desire to achieve preferential is brittle into a formation that is ductile, extra energy
fracture extension and minimal height growth. Often (i.e. net pressure) will be required to keep the fracture
there is very little evidence to back up these claims of propagating in that direction. If this energy is not available
fracture containment. However, it is also true that or if it is easier for the fracture to grow in an alternative
126
direction then the fracture will not significantly Simulators assume symmetry on either side of
penetrate the ductile formation. Consequently, the wellbore, but this is highly unlikely given the
contrasts in apparent fracture toughness can form the variations in formation properties and perforation
most reliable barriers to height growth, especially for characteristics that almost certainly exist.
shales (as many shales exhibit plastic properties). Simulators generally assume with notable
It is possible that fracture height containment exceptions an elliptical or semi-elliptical
often attributed to stress contrasts is in fact due to fracture geometry.
apparent fracture toughness contrasts. Simulators assume again with notable exceptions
horizontal isotropy in the formations. Even
Youngs Modulus. Contrasts in Youngs modulus are in simulators that allow variations in formation
not very good at preventing fracture height growth. characteristics on the horizontal plane (such as the
There is an inverse relationship between fracture grid-base simulators), there is usually very little
toughness and Youngs modulus; formations tend to data available to justify doing this.
be either soft and ductile or hard and brittle. This Simulators generally assume a single fracture on
means that a rapid increase in E can also coincide each side of the wellbore. In reality, engineers
with a rapid decrease in apparent fracture toughness, hope that a single fracture will predominate
making it easier for the fracture to propagate. and force any others closed, but there is no
However, as illustrated in Eq. 4-5, fracture width is guarantee that this will happen. Every perforation
inversely proportional to E for any given net pressure. is a potential source of fracture initiation, and
Therefore, a rapid increase in E will result in a rapid it is highly likely that more than one fracture
decrease in fracture width, possibly even to the point will propagate on either side of the wellbore
where it is too narrow for proppant placement. especially if the wellbore is deviated.
However, one over-riding fact determines
Fluid Viscosity. In spite of a common industry that relatively simplistic representations of fractures
perception, reduced fluid viscosity cannot be relied will persist. There is little point in making the
upon to reduce fracture height growth, although a models any more complex, as we cannot acquire
contrast between pad fluid and proppant-laden fluid the data necessary to do justice for even these
viscosity can under the right circumstances result simple models. Indeed, most inputs into the
in proppant-laden fluid entering only the central part current generation of fracture simulators require a
of the fracture. In fact there is considerable evidence considerable degree of educated guesswork, refined
from the Gas Research Institute (GRI) Staged Field by experience. Consequently, there is little point
Experiments (1988, 1989 and 1990), as discussed by in making the models even more complex until we
Wright et al. (1993), that formations will produce the can find significantly improved ways to determine
same pressure response (indicating similar fracture the required input data.
geometry) regardless of fluid viscosity. There is also
very little evidence from microseismic analysis that 4-3.4 Fracture Propagation,
thin fracturing fluids produce less height than highly Toughness and Tip Effects
viscous fluids (Cipolla, 2006).
4-3.4.1 Linear Elastic Fracture Mechanics
4-3.3.4 Complex Fracture Geometry
Linear Elastic Fracture Mechanics (LEFM) is all
In reality, even with the most advanced models available, about predicting how much stress (i.e. energy) is
we are using simplifications to model the hydraulic required to propagate a fracture. LEFM assumes
fracture. In general, what we try to do is make a regular, linear elastic deformation (constant Youngs
symmetrical, simulated fracture behave in a similar modulus) followed by brittle fracture it is implicit
fashion to a far more complex reality: that no significant energy is absorbed by non-linear
127
Modern Fracturing
or non-elastic effects. That is to say, energy stored In order to reach this relationship, Griffith makes a
as stress in the material is transferred directly to significant assumption that there is no energy lost
fracturing the material, and no energy is lost to at the fracture tip, so all energy is used to propagate
plastic deformation or other effects. the fracture, either to elastically deform or to rupture
the material. Therefore, there can be no plastic
The Griffith Crack. The first person to adopt a deformation at the tip and the Griffith model is only
meaningful analytical approach to studying the applicable to materials liable to elastic deformation
mechanics of fracture propagation was Griffith followed by brittle fracture.
(1921). Fig. 4-29 illustrates the concept of the
Griffith crack, which can be expressed as: Griffith Failure Criterion. Given that for a uniform
material with constant geometry U/a is a constant,
U 2 2 a

= , (4-63) there is a critical value of stress, c, at which the material
a E
will experience catastrophic failure, i.e., the fracture
where U is the elastic energy (i.e., the energy used propagates at high velocity across the material causing
to produce elastic stress on the material), a is the failure. This critical stress is defined as:
characteristic fracture length, is the far-field stress
EG
(i.e., the bulk stress away from the influence of c = 1c
. (4-65)
a
the fracture) and E is Youngs modulus. Therefore,
Eq. 4-63 defines the amount of additional energy The critical energy release rate, Glc, is determined
(U) that is required to make the fracture grow from experimentally and is a material characteristic,
length a to length a + a. although it will vary with both temperature and
the overall geometry of the test specimen. Equation
4-65 also defines, for a given stress, a critical
fracture length. If the fracture is less than this
critical length, the material will not fail. However,
if the fracture grows beyond this critical length,
the material will fail.
The subscript 1 refers to the failure mode, as
illustrated in Fig. 4-30. Failure mode 1 is the opening
mode, mode 2 the sliding mode and mode 3 the
2a tearing mode. In hydraulic fracturing, we are usually
concerned only with failure mode 1.
Mode 1: Mode 2: Mode 3:

Opening Sliding Tearing
Figure 4-29 The Griffith crack
Usually, U/a is replaced by 2G. G is referred

to as the elastic energy release rate or the crack
driving force, such that:
2 a Figure 4-30 Failure modes for linear elastic fracture
G =
. (4-64)
E mechanics
128
Stress Intensity Factor. With reference to Fig. 4-31, the characteristic, it is also a variable, depending upon the
stresses in the principal directions, at some point away gross geometry of the fracture and its surroundings,
from the fracture tip, can be expressed as: as well as temperature.
Assuming a constant temperature in any given
K1 3
xx = cos 1 sin sin and (4-66) instance, relationships linking K, a and for most
2 r 2 2 2
situations have been solved, either analytically or
numerically. At material failure, c can be described in
K1 3
yy = cos 1 + sin sin , (4-67) terms of a critical stress intensity factor, K1c, which is
2r 2 2 2
more commonly referred to as the fracture toughness:
where K1 is the stress intensity factor (failure mode 1) at
K1c
the polar coordinates (r, ) away from the fracture tip. c = .
(4-69)
a

This is a fundamental equation of Linear Elastic Fracture
Mechanics, where is a geometrical factor and is equal
y
to 0.4 for a radial fracture. For elliptical fractures, can
be found as follows (from Broek, 1986):
r 1.12
2
.
3 x f
Fracture + 2 (4-70)
x ) 8 8 ( 12 h
a f
K1c is related to G1c by:

K12c
G1c = (1 )

2
. (4-71)
Figure 4-31 Coordinate system for the stress intensity E
factor
For a given gross geometry, fracture toughness
Considering the plane strain situation (i.e., zz = 0, is a material characteristic. Equation 4-71 shows
an object with a thickness large enough to make strain that it represents the amount of mechanical energy a
on the z-axis negligible), and the case that a r, then material can absorb before it fails by brittle fracture.
the stress in the y-direction across the line of the Put simply, a material with a low K1c is brittle, and a
fracture (i.e., = 0) can be expressed as: material with a high K1c is tough.
K
yy =
1
. (4-68) 4-3.4.2 Significance of Fracture Toughness
2r
Obviously, from Eq. 4-68, as r 0, yy . This In hydraulic fracturing, fracture toughness represents
represents a fundamental difficulty in this approach the amount of energy required to split the rock apart
to modeling fractures because the stresses can become at the fracture tip. The relative values of fracture
infinite at the fracture tip. In reality, some kind of toughness and Youngs modulus determine how
plastic deformation is likely as the stress rises above energy is used to create fracture width and how much
the yield point. In brittle materials, the region around it is used to create fracture height and length.
the fracture tip in which plastic deformation occurs At the start of fracture propagation, the fracture
is very small and the amount of energy consumed half-length (equivalent to a in Eq. 4-69) will be very
by this process is not significant. This is not the small. While the stress (and hence energy) required
case for ductile materials. to create length is proportional to 1/xf, the energy
Using this approach, K is the only factor required to create width is directly proportional to
that affects the magnitude of the stress at a given the half-length. This means that the energy used
distance from the fracture tip. While K is a material as a result of the fracture toughness is much more
129
Modern Fracturing
significant for smaller fractures (or at the start of Jeffrey defines the effective fracture toughness
a treatment) than it is for larger fractures (or at produced by the fluid lag effect as:
the end of a treatment).
K1effc = K1c + K1lag
c and (4-72)
4-3.4.3 Complexity at the Fracture Tip
R R

K1lag
c = 2 ptip
arcsin , (4-73)
2 R
From the extensive work done in this field, it is clear
that LEFM alone does not adequately account for the where ptip is the net pressure in the non-wetted part of
pressure needed to make the fracture grow. There is a the fracture tip, R is the radius of the fracture at any
tip over-pressure effect, which means that more pressure given point, K1ceff is the effective fracture toughness
(energy) is required than is predicted by LEFM. A few caused by the fluid lag effect, K1clag is the additioanl
of the possible and not necessarily mutually exclusive fracture toughness caused by the fluid lag effect, and
theories for this are described below. is the length of the non-wetted zone (parallel to the
direction in which xf is measured). Note that R is the
Fluid Lag Effect. The fluid lag effect was first identified distance from the point of fracture initiation and does
by Khristianovitvh and Zheltov in their original 1955 not imply radial fracture geometry.
paper that is better known for the description of the Johnson and Cleary (1991) showed that the
frac model that would become known as the KZD length of the non-wetted part of the fracture could be
model. Jeffrey (1989) quantified the effect by defining approximated by:
an apparent fluid lag due to the crack tip pinching effect 2
caused by fluid lag as illustrated by Fig. 4-34. R pnet (4-74)
,
2 pnet + pc
Process Zone
if the condition set out in Eq. 4-75 is met.
pnet Crack Tip Dilatancy. The theory of crack tip dilatency

ptip was first put forward by Johnson and Cleary (1991)
and has been used extensively by them in one of
the major fracture models. They used the concept
pnet
of dilatancy (first observed by Reynolds, 1885) in
which granular materials are observed to deform
2xf
in a non-linear elastic fashion. During hydraulic
fracturing, dilatancy causes a volumetric expansion
of the material in the process zone, which requires
Figure 4-32 Crack tip pinching effect caused by the extra energy (i.e. net pressure). This approach almost
inability of the fracturing fluid to penetrate into the fracture entirely eliminates the concept of fracture toughness.
tip (from Jeffrey, 1989)
Instead, the theory states that deep underground,
Because the frac fluid cannot penetrate into the the effect of the confining stress is much more
far tip of the fracture, the rock at the tip experiences significant than the effect of the fracture toughness.
less pressure than the rock exposed to the fracturing Thus K1c can be ignored if:
fluid in the main body of the fracture. Because the
rock is subjected to less pressure, the fracture width is
pnet R K1c , (4-75)
reduced. This acts to decrease the apparent length of
the fracture, which has the same effect as increasing where R is the radius of the fracture (as defined above)
the fracture toughness of the rock. Therefore the main and is analogous to the LEFM characteristic of fracture
effect of the fluid lag is to increase the net pressure length. Johnson and Cleary combined the effects of
required to propagate the fracture. dilatancy and fluid lag, as illustrated in Fig. 4-33.
130
Dilation Contribution where C is the Kachanov (1971) parameter and is

Fluid Lag Pinching Effect considered a material characteristic.
Crack Tip Plasticity. The phenomenon of plastic flow

at the fracture tip, even in materials believed to be
pnet
w elastic-brittle, was first described in relation to hydraulic
fracturing by Yew and Liu (1993). van Dam, et al.
(2000) presented experimental evidence of the existence
of the plastic zone, while Martin (2000) described the
R shape of the plastic zone and quantified the amount of
Figure 4-33 Illustration of the Effects of Fluid Lag and energy lost to plastic deformation.
Dilatancy (from Johnson and Cleary, 1991)
Referring to Eq. 4-68, we can see that as we near the
Process or Damage Zone. Yew and Liu (1993) fracture tip (i.e., as r approaches zero) the stresses can rise
showed that a damaged zone exists beyond the fracture dramatically and will tend to infinity as r actually reaches
tip, in which extra energy is absorbed to create the zero. Obviously, this is not possible; as the material passes
damage, ahead of the propagation of the fracture. This its yield stress, y, it will plastically deform to reduce the
produced an increase in fracture toughness: stresses and absorb energy. This has the effect of producing
E Et a blunt fracture tip with a diameter, d, defined as:
72 2
K1c =
K , (4-76)
2 1
K12 (1 2 )
2
( )
1 + 3 (1 + ) Et
d= ,
(4-80)
E y
where Et is the Youngs modulus of the damaged
zone, K1 is the stress intensity factor (= pnet(xf )), remembering that K1 = pnet(xf). The plastic region
and is found from: around the fracture tip has a radius rp, as defined by:
3 2
y rp 1 3
tan =
2
,
112
(4-77)
K 2
1
=
4 2 ( )
sin 2 + (1 2 ) (1 + cos ) .
2

(4-81)
where is the internal friction angle of the material. Figure 4-34 shows Eq. 4-81 plotted in polar coordinates,
From Eq. 4-76 we can see that the change in fracture to give the size and shape of the plastic zone around the
toughness is driven by a change in Youngs modulus in fracture tip.
the affected area. If the Youngs modulus is unaffected
(i.e. Et = E in Eq. 4-76), then K1c is zero. 0.5
Valk and Economides (1993) used the
Continuum Damage Mechanics concept to modify y2rp
traditional LEFM to account for the existence of O = 0.25 K12
microfractures in a damaged zone normal to the
fracture path. They showed that Eq. 4-68 could be
modified to define the stress at a distance r from the R
fracture tip, to allow for the presence of microfractures -0.5 0.5
of average length, :
K1
(r ) = 1 / r h ,min . (4-78)
2 ( + x f )
They also showed that the fracture propagation rate, u,

could be determined by: -0.5
2 Figure 4-34 Polar coordinate plot showing the size and
C 2 K1
, shape of the crack tip plastic zone, for = 0.25 (from
u= (4-79)
h ,min + x f Martin, 2000)
131
Modern Fracturing
4-3.5 Measuring Rock and Cook, 1995, Obert et al., 1946). For example,
Mechanical Characteristics the dynamic Youngs modulus of sandstone is often
greater than the static value by a factor of 2 to 10.
4-3.5.1 Introduction The predominant view in the industry is that static
This section reviews testing methods that are generally values of elastic constants are a better representation
used in acquiring rock mechanical properties necessary of the values needed in hydraulic fracture models.
to run current hydraulic fracture models. In cases where Therefore, elastic constants derived by dynamic
core is not available, alternative methods of estimating methods often have to be corrected or transformed
mechanical properties are reviewed. into static equivalents before use in hydraulic fracture
Many hydraulic fracture models require some or models. Below, we discuss how these constants
all three of the following elastic constants for each layer may be obtained and how to transform dynamic
used in the simulation: values into static values.
Youngs modulus
Poissons Ratio Dynamic Methods. It can be shown (Auld, 1990) that
Biot-Willis poroelastic constant for linear elastic, isotropic, homogeneous solids, the
relationships among the dynamic Youngs modulus
In addition, proppants are used to maintain fracture (Edyn) and Poissons Ratio (dyn) and the longitudinal
width and enhance conductivity. If the formation is too and shear wave velocities are as follows:
soft, excessive proppant embedment may detrimentally 1 (u u )2 1
dyn = 2
p s
reduce conductivity. This is especially true for low 2
and (4-82)
proppant loadings. Formation hardness is a simple test (u p us ) 1
that can be used to assess the degree of embedment that
one can expect at net closure stress. Edyn = 2bus2 (1 + dyn ) , (4-83)
4-3.5.2 Methods of Measurement where up is the longitudinal mode of wave propagation,

us is the transverse mode of wave propagation and
The measurement of the elastic properties of the b is the bulk density, which can be calculated
formation have been divided into two general from porosity (), rock matrix density (ma) and
categories; static and dynamic. Dynamic elastic fluid density(f) by:
properties derive their name from the oscillatory
nature of the applied loads used in the measurement, b = ma (1 ) + f .
(4-84)
while the term static implies measurements in
which the loads are applied at such low rates that the Eqs. 4-82 and 4-83 form the basis for calculating
axial and confining stresses are uniformly distributed elastic constants from wireline conventional sonic and
over the entire sample. Static elastic constants are dipole sonic density logs. They are used throughout the
measured using conventional triaxial test equipment industry and usually provide a reasonable estimation
that measures the deformation of core samples as of the dynamic elastic constants. A more convenient
a function of applied stress. The rates of stress and form of Eqs. 4-82 and 4-83 uses the transit times or
strain in static measurements are orders of magnitude inverse wave velocities:
lower than those used in dynamic testing. The stress/
strain amplitudes applied in static measurements are 1 (t t )2 1
dyn = 2
s p
also order of magnitude greater than those used in
(ts t p ) 1 and (4-85)
2
dynamic testing (typically). Because of differences

in stress/strain rates and amplitudes between the two
Edyn = 2.694104 b (1 ts ) (1 + dyn ).
2
general test methods, the calculated values of the (4-86)
elastic constants usually turn out to be different (Plona
132
Shear (Gdyn) and bulk moduli (Kdyn) can also be pulsed through-transmission method is most common.
calculated from the transit times: The ultrasonic transducers are designed to generate
longitudinal, transverse or torsional waves. The
2
Gdyn = 1.347 104 b (1 ts ) and (4-87) transverse- and torsional-mode transducers generate
ultrasonic waves that travel at the shear velocity. The
2 2
K dyn = 1.347104 b (1 t p ) 4 3 (1 ts ) , (4-88) transverse wave measurement has the advantage over the

torsional wave measurement in that the transverse wave
where ts = 1/us and tp = 1/up. The coefficient in Eq. 4-86 is polarized and therefore suitable for measuring elastic
is the conversion factor allowing Youngs modulus to anisotropy. However, torsional-mode transducers tend
be expressed in units of millions of pounds per square to generate more energy; therefore, the signal-to-noise
inch (Mpsi), while the bulk density is expressed in units ratio is greater than that produced by polarized shear
of g/cc and the shear-wave transit time in sec/ft. Eqs. transducers. This can be important when measuring
4-85 and 4-86 are convenient because sonic data is shear velocities in highly attenuating rock, such as
typically expressed in units of sec/ft. unconsolidated, saturated sands.
The Biot-Willis poroelastic constant, , can also
be calculated from wave velocity data (Klimentos Static Methods. The most reliable method for
et al., 1995): determining the elastic constants for use in hydraulic
fracture models is by testing cores under triaxial
= 1 K dry K s ,
(4-89) conditions of stress. The most common triaxial method
tests cylindrical core plugs. Loads are imparted to the
where Kdry is the dynamic bulk modulus of the rock sample by a pressurized fluid (usually hydraulic oil)
tested under drained conditions and Ks is the average applied to the outer circumference of a sleeved core
bulk modulus of the mineral components comprising plug and an independent axial stress applied to each
the rock. Drained measurements are performed end of the sample by a hydraulic press. Typical stress
by allowing the pore fluids to freely drain from the paths involve an initial hydrostatic ramp to some
pore space, thus maintaining constant pore pressure predefined value, followed by an additional increase
during the test. The subscript dry is used in Eq. 4-89 in axial stress while maintaining a constant confining
because testing dry core is the easiest way to ensure this pressure. In some instances, the axial stress path may
condition is met. Like other dynamic measurements be cycled one or two times to remove stress hysteresis
of elastic properties in porous rock, obtained from effects from the strain measurements. The cores may or
wave velocity measurements may be different from may not be saturated with a fluid. For gas-producing
obtained using static testing methods. wells, testing the cores in a room-dried state usually will
Wireline logging companies provide options for not affect the elastic properties to a large extent.
obtaining the wave velocities of different modes of wave The axial and radial displacements are usually
propagation. For example the compressional, shear, and measured using either strain gauges or LVDTs
Stonely modes of propagation are routinely collected. (Linear Variable Differential Transformers). The
Conventional shear velocities are derived from mode- axial strain is calculated from the fractional change
converted compressional wave signals that fall incident in length along the deviatoric axial stress cycle. The
on the borehole wall. Dipole-sonic logs induce torsional radial strain is calculated from the fractional change
and flexural modes of propagation but travel essentially in diameter or circumference along the same portion
at the shear velocity. For elastic constants, data from the of the stress cycle. Youngs modulus is calculated
compressional and shear wave transit times are used in from the secant slope of the change in axial stress
Eqs. 4-85 to 4-88. to axial strain. Poissons Ratio is calculated from
The longitudinal and transverse wave velocities are the secant slope of the change in radial strain to
also measured in the laboratory on cores using a set axial strain. In some instance, the volume strain
of transmitting/receiving ultrasonic transducers. The is monitored during the stress cycle(s) to ensure
133
Modern Fracturing
that volume strain of the core is not in a state of Typically, BH values less than 5 to 6 kg/mm2
dilatation. Dilatation implies the volume is increasing indicate soft formations. Care should be taken to
with additional stress/strain and is associated with ensure that proppant loadings are high enough to
permanent deformation and the onset of plastic compensate for proppant embedment.
deformation (Reynolds, 1885). Sections of stress/
strain curves where dilatation is observed should be Formation Porosity-Permeability. Routine
avoided when calculating elastic constants. measurements of porosity and single-phase
permeability (either liquid or air) should be performed
Brinell Hardness. Formation hardness is easily if core materials are available. These tests are relatively
measured on core materials using a single-ball easy to perform and provide important data needed
penetrometer, which is used to determine the Brinell to calculate production rates, hydrocarbon volumes,
hardness (BH) of the material. This measurement is completion strategies, etc. This data is often needed
often performed to estimate the amount of proppant as input into hydraulic fracture models. Porosity/
embedment that can be expected in a hydraulic permeability cross-plots from core measurements
fracture at maximum anticipated closure stress. can often provide useful transforms for estimating
Brinell hardness is also measured to determine the permeability from wireline log data.
softening effect that fracturing fluids have on fluid-
sensitive formations such as shale and coal. Mineralogy. Mineral content as determined from
Brinell hardness, BH, is defined by ASTM E50 powdered X-ray diffraction (XRD) techniques
and ISO 6506 standards as the ratio of the applied provides clues to mechanical and petrophysical
force, F, of the steel ball to the contact area, Ac, of the properties. For example, XRD is routinely performed
indentation created by the steel ball: on cores to identify hydrophilic minerals such as
expandable illite/smectite clays that can be the cause
F
BH = ,
(4-90) of much formation damage from water-based muds
Ac
and fracturing fluids.
where the contact area is given by:
4-3.5.3 Core Selection/Sample

Ac = 2rh = 2r r

(r 2 r 2 ) , (4-91) Preparation Considerations
and where definitions of the penetration depth, h, Core material can be acquired by various means.
radius of indentation, r, and proppant radius, r are Conventional whole cores are extracted during the drilling
illustrated in Fig. 4-35. The test can be designed to operations and have to be planned with the drilling
measure either h or r. program. Typical coring barrels extract 4-in.-diameter
cores. Core barrel lengths typically come in 30-ft sections.
Although some analytical laboratories have the means
Proppant Grain to test whole core for mechanical, electrical and poro-
of Radius r
permeability properties, most testing is performed on core
plugs extracted from whole core sections. Cylindrical core
r plugs are usually drilled parallel and/or perpendicular to
ra the whole core axis. The size of the core plugs is variable,
h but typically, 1- and 1.5-in.-diameter plugs are taken.
The length of cores cut parallel to the whole core axis can
vary depending on the test objectives. Core plugs drilled
h = Embedment Depth
perpendicular to the whole core axis, are limited to the
ra = Radius of Indentation
whole core diameter. For mechanical properties, a length-
Figure 4-35 Diagram illustrating the Brinell hardness test to-diameter ratio of 2:1 is recommended.
134
Rotary and percussion sidewall plugs are a second 3. If available, use a LithoLog to select core.
source of core material. Rotary plugs tend to have less Choose enough core from each rock type so
mechanical damage than percussion sidewall plugs and that there will be enough core data generated
are favored over percussion sidewall for testing mechanical to establish empirical relations among the
properties, routine porosity and permeability. By different rock types.
definition, sidewall cores are drilled perpendicular to the 4. Find out if there is a core-log depth shift. Note the
axis of the borehole. In vertical wells, rotary sidewall cores shift when selecting core plug locations.
are correctly oriented for determining elastic properties in 5. For hydraulic fracturing, drill plugs in the
hydraulic fracture applications. Percussion sidewall cores direction of minimum principal stress. In a
are useful for assessing mineral content and grain size and vertical well with anticipated vertical fracture, drill
measuring formation hardness. The diameter of sidewall plugs perpendicular to the core axis. In horizontal
cores is slightly less than 1 in. wells with anticipated vertical fracture, drill plugs
Drill cuttings are a third source of core material. parallel to bedding planes (if visible). If horizontal
By definition, cuttings are fragmented rock that is fractures are anticipated, cores should be drilled
transported by the drilling mud to the surface. The in the vertical direction.
cuttings acquired at the surface are a mixture of 6. Drill plugs with length-to-diameter ratio
cuttings from different depths, and the depths from of 2:1.
which they are collected are usually not accurately 7. If the core was recently taken, try to preserve
known. The use of cuttings to determine mechanical formation fluids. Wrap core plugs in plastic
properties is therefore severely compromised. However, wrap and seal before shipping to laboratory.
advances in the use of cuttings to infer mechanical Tests should be performed with reservoir fluids
properties have been reported (Ringstad et al., 1998, in place if possible. In gas wells, testing as
and Santarelli et al., 1996). received should be adequate.
Outcrop is a fourth source of reservoir core material.
Caution should be exercised when interpreting data 4-3.5.4 Deducing Elastic Properties without Core
acquired from outcrop material. Outcrop is weathered
rock whose mechanical properties have most likely There may be occasions in planning hydraulic
been altered from initial in-situ values through drying, treatments when core material is not available. It will
freeze-thaw cycling and stress relaxation. In addition, then be necessary to use other resources to estimate
the location of outcrop is often miles from the well the static elastic constants. The following resources
site. Considerable effort is often required to relate the should be considered:
outcrop test results to reservoir properties.
1. Rock properties database
The following items should be considered when 2. Reference literature studies
selecting core for testing:
A database of core measurements (in-house) of
1. Try to obtain a representative sampling static and dynamic elastic properties can be used as
both the reservoir and the bounding rock a look-up tool to determine whether measurements
layers, if available. on cores from the same formation, location and/
2. Use available gamma ray (GR), density, and or rock type have been made. If so, this data may
compressive wave transit time (DTC), shear wave be correlated to available wireline log data to help
transit time (DTS) logs to determine the location generate a synthetic mechanical property log.
and number of samples needed. Select samples If not, then the use of empirical correlations
that span the range of density, porosity and GR reported in the technical literature between quantities
values. This will increase the chance of selecting a routinely measured by wireline logs and static/dynamic
wide range of rock properties. elastic constants on similar rock types may suffice. The
135
Modern Fracturing
most common wireline log data that often correlates 3. Han et al. (1986) provided empirical relations
well to elastic properties are porosity, density, gamma for shaly sandstone, relating the p-wave
ray (total clay) and sonic transit times. However, it and s-wave velocities to porosity and clay
is important that reported empirical correlations be content. The study was conducted on well-
applied only to the rock types similar to those in consolidated Gulf Coast sandstones with
which the study was conducted. porosity ranging from 3% to 30% and clay
One basic strategy for estimating mechanical content from 0% to 60%. These relations
properties from logs is to first determine which do not extrapolate well beyond the range
of the following information is known for a set of data tested.
of rock types: 4. Eberhart and Phillips (1989) related the p-wave
and s-wave velocities to porosity, clay content
1. Bulk density and effective pressure in water-saturated cores
2. Porosity from Hans data set.
3. Clay volume 5. Castagna et al. (1985) related p-wave and s-
4. P-wave transit time wave velocities to porosity and clay content for
clastic silicates composed of mostly clay and
For example, if bulk density, porosity and siltstone. The relations should be used only for
clay volume are known from wireline logs, then rock types of similar properties.
an appropriate correlation found in the literature
relating these two quantities to the p-wave (and/or s- If p-wave data is available, but not s-wave, the
wave) transit times can be used to generate a synthetic following sources may be of use:
p-wave (and s-wave) log. If s-wave log data is not
available, an appropriate correlation can then be used 6. Castagna et al. (1993) and Pickett (1963)
to estimate s-wave transit times from p-wave transit derived empirical correlations between p-
times. Once the p-wave and s-wave transit times have wave and s-wave velocities in water-saturated
been estimated, the dynamic elastic constants can be limestone and dolomite.
calculated from the above equations. The static elastic 7. Castagna (1985, 1993) and Han (1986) derived
constants can then be estimated from appropriate empirical correlations between the p-wave and
static-to-dynamic transforms. s-wave velocities in water-saturated sandstone,
Below is a partial list of references that may be shale and shaly sands.
useful in constructing synthetic logs from reported 8. Greenberg and Castagna (1992) gave empirical
empirical trends: relations of s-wave to p-wave velocities of the
mineral constituents making up water-saturated
1. Wyllie et al. (1956, 1958, 1963) provided time rock. Correlations were developed for sandstone,
average equation relates the p-wave velocity in limestone, dolomite and shale.
rock to the porosity and p-wave velocities in the
pore fluid and mineral constituents of the rock. If bulk density is the only known quantity, then
The correlation is limited to moderately to well- the following sources may be of use:
cemented sandstone with primary porosity where
all minerals have the same velocity. 9. Gardner et al. (1974) derived an empirical
2. Raymer et al. (1980) also related the p-wave relation between the p-wave velocity and bulk
velocity to porosity and the p-wave velocities density for many rock types.
of the pore fluid and mineral content. The 10. Castagna et al. (1993) improved Gardners
correlation is limited to moderately to well- relations by applying power-law fits to individual
cemented sandstone where all minerals sets of data representing shale, sandstone,
have the same velocity. limestone, dolomite and anhydrite.
136
Additional sources that may be of use include: is only very rarely a constant value, as it can change
dramatically with temperature, applied shear stress and
11. Brevik (1995) and Urmos and Wilkens (1993) fluid composition. Viscosity is defined as the relationship
presented relations between p-wave and s- between shear stress and shear rate (see below).
wave velocities in gas- and brine-saturated
chalk from sonic-density logs. Included 4-4.1.1 Shear Rate, Shear Stress and Viscosity
are relationships between p-wave, s-wave
velocities and porosity. Shear Rate, . In fluid mechanics, shear rate is a
12. Geertsma (1961) and Yale and Jamieson (1994) measure of how fast a fluid is flowing past a fixed
presented data for deriving correlations between surface. Shear rate can be thought of as a measure of
p-wave, s-wave velocities and porosity in dry how much agitation a fluid is receiving.
dolomite and limestone.
13. Blangy (1992) presented data for deriving Causes of Shear Rate:
correlations between the p-wave, s-wave Spinning centrifugal pump
and porosity in poorly consolidated, Flow through a pipe
saturated sandstone. Model 35 viscometer test
Jet mixer
After obtaining an estimate of the dynamic elastic Tank agitators
constants, the following sources may be useful in
estimating the static elastic constants. Most studies Shear Stress, . Shear stress is the resistance the fluid
focused on the relationship between the dynamic and produces to an applied shear rate. For instance, it
static Youngs modulus. To this authors knowledge, requires more force (or pressure) to pump water at
strong empirical correlations between the static and 20 bpm than at 10 bpm.
dynamic Poissons Ratio have not been published.
Viscosity, . The fluid property that defines how
14. Lacy (1997) derived empirical correlations much shear stress is produced by a shear rate, is called
between the static and dynamicYoungs modulus viscosity. The greater the viscosity, the greater the
for sandstone, shale and the combined data set resistance of a fluid to shear agitation.
of sandstone, shale and carbonate rock.
15. Yale and Jamieson (1994) presented data for 4-4.1.2 Measurement of Viscosity
deriving empirical correlations between static
and dynamic Youngs modulus, static and The standard method for measuring the viscosity of
dynamic Poissons ratio, static Youngs modulus a fluid is to agitate it at a known shear rate and then
and porosity, and static Poissons ratio and see how much force is produced on a fixed surface,
porosity in different types of carbonates. positioned close to the source of agitation, with a thin
16. Morales (1993) derived an empirical correlation layer of the test fluid between them. For a fixed rate
between static and dynamic Youngs modulus of shear, the greater the force on the fixed surface, the
for sandstone from various fields. greater the viscosity of the fluid.
Figure 4-36 shows the components of a typical
4-4 Fluid Rheological Characteristics viscosity-measuring device. The device consists of a
fixed solid cylinder (or bob) surrounded by a hollow
4-4.1 Viscosity cylinder, which is positioned concentrically to the bob.
The cylinder (also referred to as the rotor) spins around
Viscosity is a measure of how much a fluid resists the bob such that a fluid positioned between the rotor and
deformation as a result of an applied force or pressure. the bob will produce a drag effect on the bob. The greater
It is a measure of how thick the fluid is. Viscosity the viscosity of the fluid, the greater the drag force on the
137
Modern Fracturing
bob. The bob is connected, via a shaft, to a torsion spring

and a measuring device. As the fluid produces drag on = .
(4-92)
the bob/shaft assembly, it is allowed to deflect against the
torsion spring, so that the greater the drag force, the more For oilfield units, with in cp, in lbf/ft2 and in
the shaft and bob assembly will deflect. The deflection is sec-1, Eq. 4-92 is expressed as:
measured and displayed as viscosity. Because some fluids

have viscosity that is not constant (c.f.) and will vary with
= 47, 479 . (4-93)

shear rate, most viscometers allow the rotational speed of
the rotor (and hence the shear rate) to be varied. Newton was the first to realize the relationship in
fluids between an applied force and the resistance to
Torsion Spring
that force. His experiments were carried out on simple
fluids such as water and brine, and not on more
complex fluids, such as those used in stimulation
activities. A fluid that exhibits a linear relationship
Bob Shaft between shear rate and shear stress is referred to as
Rotor
being Newtonian. However, there are a wide variety
Fluid
Bob of fluids that do not exhibit this behavior and are
referred to as non-Newtonian (see below).
Cross-Section Through Rotor & Bob
Examples of Newtonian fluids include fresh
water, seawater, most brines, non-gelled acid,
Figure 4-36 Diagrammatic illustration of the rotor and
bob configuration used to measure viscosity diesel, alcohols and gases (at constant density
and temperature).
Viscometers based on this rotor and bob method
are available in various configurations, including fully 4-4.2.2 Non-Newtonian Fluids
computer-controlled versions capable of testing fluids
at high temperatures and high pressures. Non-Newtonian fluids do not exhibit a linear
relationship between shear rate and shear stress,
4-4.2 Fluid Behavior except in very specialized circumstances. They
can be divided into three basic types, as
4-4.2.1 Newtonian Fluids illustrated in Fig. 4-38.
Newtons law of fluids states that there is a linear

relationship between shear rate and shear stress, as
illustrated in Fig. 4-37. The gradient of this straight
line is the viscosity:
Shear Stress,
d
c
b
Shear Stress, t
Gradient =
0
0 Shear Rate,
0 Figure 4-38 The relationship between shear rate

0 Shear Rate, and shear stress for a) Newtonian fluid; b) power law
Figure 4-37 Relationship between shear rate and shear fluid (shear thinning); c) Bingham plastic fluid and d)
stress for a newtonian fluid Herschel-Buckley fluid
138
Bingham Plastic Fluids Shear-thickening fluids. These fluids have an n greater

This type of fluid requires an initial shear stress to be than one, so they exhibit an increase in apparent
induced before they will deform. Put another way, they viscosity as shear rate increases. Extreme examples of
have a yield point or gel strength that must be broken these fluids can behave as if they were solids when
before the fluid can move (although some fluids have a gel exposed to even moderate shear forces.
strength that has nothing to do with yielding). This type of
fluid is not Newtonian, although it usually has a constant Herschel-Buckley Fluids
viscosity once the initial gel strength has been overcome: Another example of a power law fluid is the Herschel-
Buckley fluid, which is often used to model the flow
= yp + p ,
(4-94) behavior of foams:
n
where yp is the yield point and has oilfield units of = 0 + K ,
(4-96)
lbf/100 ft2 (note that in laboratory measurements, yp
has the units lbf/ft2, so the value for has to be converted where o is the threshold shear stress, K is the
before it is used), and p is the plastic viscosity in cp. Herschel-Buckley consistency index and n the
Herschel-Buckley exponent.
Power Law Fluids Herschel-Buckley fluids are basically a combination
The next group of fluids is generally referred to as of Bingham plastic and power law fluids. An initial
power law fluids, although other names have been threshold shear stress has to be overcome before the fluid
used to describe them. In general, there is no linear will flow. Once this has happened, the viscosity is not
relationship between shear rate and shear stress, so constant and will vary according to the shear rate.
that apparent viscosity (the viscosity which the fluid
appears to have, at a specific shear rate) changes with 4-4.2.3 Apparent Viscosity
shear rate. Equation 4-95 describes the behavior
of the power law fluid: The apparent viscosity of a fluid is the viscosity of
the fluid at a specific shear rate. For a Newtonian
n
= K . (4-95) fluid, the apparent viscosity is the same as the actual
viscosity. For all other fluids, the apparent viscosity is
In order to find K (the power law consistency index) the slope of a line on a shear rate vs shear stress curve,
and n (the power law exponent), viscometer readings are from the origin to the line, at a specific shear rate, as
taken at a variety of different shear rates. The log of shear shown in Fig. 4-39.
rate is plotted against the log of shear stress, so that K is
the intercept of the log axis and n is the gradient. K
has the rather awkward units of lbf.secn/ft2, in order to be
consistent, while n is dimensionless.
Power law fluids can be divided into three
Shear Stress,
major categories:
1
t=
Shear-thinning fluids. In these fluids, n is less than 1, so en
dia 2
Gr nt =
the fluids experience a decrease in apparent viscosity as the ad
ie
G r
shear rate increases. Most of the fluids used for fracturing
fall within this category. 1 2
0
0 Shear Rate,
Newtonian fluids. Newtonian fluids are a special case
of power law fluids in which n is equal to one, i.e., the Figure 4-39 Change in apparent viscosity for a power
viscosity is constant and equal to K. law fluid at two shear rates
139
Modern Fracturing
As can been seen in Fig. 4-39, for a shear-thinning into turbulent flow. This is characterized by a series
power law fluid, the apparent viscosity of the fluid of small-scale eddies and whirls, all moving in the
(the slope of the two lines) decreases as the shear rate same overall direction.
increases. At shear rate 1, the slope of line 1, 1, (and The friction pressure produced by the fluid
hence the apparent viscosity) is greater than the slope flow is highly dependent upon the flow regime.
of line 2 at the greater shear rate 2. Hence the fluid Therefore, it is important to be able to determine
is said to be shear thinning. the flow regime.
In practice, it is the apparent viscosity that
is usually measured. The model 35 viscometer is 4-4.3.2 Reynolds Number
set up so that at 300 rpm (with an R1 rotor,
B1 bob and spring factor = 1), the apparatus The flow regime is found by using the Reynolds number
reads apparent viscosity directly with no additional (NRe), as follows:
calculations required.
NRe< 100 Plug Flow
4-4.3 Flow Regimes 100 < NRe< 2000 Laminar Flow
NRe> 2000 Turbulent Flow
4-4.3.1 Plug, Laminar and Turbulent Flow
The Reynolds number for pipe flow can be
Figure 4-40 illustrates the three different flow regimes that found from:
a fluid can experience, with plug flow being at the lowest
dv
fluid velocity and turbulent flow being at the highest.
N Re = , (4-97)

where is the fluid density, d is the inside diameter of
the pipe, v is the bulk fluid velocity along the pipe
and is the viscosity. Equation 4-97 is for SI units,
and for field units:

fluid q
N Re =132, 624
, (4-98)
d
Plug Laminar Turbulent
where fluid is the specific gravity of the fluid, q is the
Figure 4-40 The three flow regimes flow rate in bpm, d is the inside diameter in inches and
is the viscosity in cp.
Plug Flow. At low flow rates, the fluid flows with an Obviously, Eqs. 4-97 and 4-98 only apply to
almost uniform velocity profile. The fluid moves with Newtonian fluids, i.e., fluids with a constant viscosity.
a uniform front across almost the entire flow area. As stated before, in fracturing engineers only rarely
deal with Newtonian fluids, so below is Eq. 4-98
Laminar Flow. As the flow rate increases, the velocity converted for power law fluids:
profile begins to change. Fluid close to the walls of fluid v 2n
the pipe (or duct, or fracture) flows slowest, while N Re = 15.49
n
, (4-99)
K (96 d )
fluid in the center of the pipe flows fastest. Fluid
velocity is a function of distance from the pipe wall. where v is the velocity in ft/sec. To make things easier,
This is also known as streamline flow. v can be easily found from the flow rate, q:

Turbulent Flow. As the flow rate continues to q (4-100)
v =17.157 2 ,
increase, the contrast in velocity across the flow area d
becomes unsustainable, and the fluid breaks down with q in bpm and d in inches.
140
4-4.4 Fluid Friction 4-4.4.2 Predicting Pressure Loss due to Friction
One of the ultimate objectives of fluid mechanics as Fannings method uses a friction factor, , to calculate
far as the fracturing engineer is concerned is to be the fluid friction:
able to predict the friction pressure (ppipe friction) of the
fluid Lv 2 f
fluids that are being pumped. This is often very difficult p pipe friction = 0.325
. (4-101)
d
because fluid composition and temperature is constantly
changing as the treatment progresses. In addition, Eq. 4-101 is in field units, with the length of the pipe,
two-phase (liquid and proppant) and even three-phase L, in ft, the velocity, v, in ft/sec and the pipe inside
(liquid, proppant and gas) flow is common. diameter, d, in inches.
Predicting fluid friction pressure is, therefore, an The friction factor is determined by using the
unreliable process; there really is no substitute for reliable Reynolds number. For plug and laminar flow:
bottomhole pressure data. Failing that, the next best option
16
is to use friction pressure tables, using data generated by f = ,
(4-102)
N Re
actually pumping the fluid around a flow loop (therefore
data is based on a situation similar to the actual treatment and for turbulent flow for smooth pipes:
process). Most modern fracture simulators incorporate
0.0303
data from these tests in their fluid models, so friction f 0.1612 .
(4-103)
N Re
pressures predicted by these are also reasonably reliable
(although not perfect, as the temperature of the wellbore is
constantly changing) unless there is proppant in the fluid. Effect of Proppant on Fluid Friction. Proppant
When the three methods outlined above are not concentration has a significant effect on the
possible, the friction pressure may be calculated from fluid friction pressure of the fluids. Generally, increasing
data, using one of several available methods. The method proppant concentration will increase the friction
outlined in Section 4-4.4.2, based on the use of Fanning pressure. Shah and Lee (1986) developed the
friction factors, is fairly reliable (i.e. it is just as good as the following correlation, based on the effect of
data used as inputs), but is not intended for use in narrow- varying proppant correlation and pipe size, for
diameter pipes at higher-than-normal flow rates (such as HPG borate fracturing fluids:
for coiled tubing treatments).
m 1m
p friction = (r ) (r )
, (4-104)
4-4.4.1 The Influence of Flow Regime
where pfriction is the ratio of the fluid friction
Flow regime has a huge influence on friction pressure pressure with and without solids, r is the ratio of
because the mechanisms that create the loss of energy the apparent viscosities of the slurry to the clean
vary massively in significance. In plug and turbulent flow, fluid and r is the ratio of the densities of the slurry
the main energy loss is due to friction effects between and the clean fluid. The exponent m is the gradient
the fluid and the wall of the flow channel (usually a of the log-log slope of friction plotted against
pipe, but also a duct or even a fracture). This is roughly Reynolds number; often; 0.2 is used.
proportional to the velocity of the fluid.
However, for turbulent flow, the situation is far more 4-5 Optimum Treatment Design
complex. Inertial and internal viscous forces become
much more significant, and the energy lost increases Hydraulic fracturing treatments are designed using
much more quickly than the velocity of the fluid. It is various software tools. There are two main groups of
therefore important to know the type of flow regime hydraulic fracturing software systems. Almost all in
being experienced by the fluid because different methods commercial use can be called fracture simulators. The
are used for calculating the friction. others are fracture design programs.
141
Modern Fracturing
A fracture simulator takes as an input a treatment into account several technical and economic constraints.
schedule. Using additional information relevant to the Fracture design software does not create a fracture
well, formation, fracturing fluids and proppants, it geometry first and then investigate its likely effect on well
simulates the growth of the fracture, the distribution productivity, but rather starts from petroleum engineering
of the proppant within it, and the pressure response. principles, determines the best possible placement of a
The primary use of fracture simulators is to analyze (or given amount of proppant that maximizes production
match) actual treatment data commonly observed and then back-calculates the necessary parameters of
pressures and compare those pressures to the model the treatment schedule that will realize the optimum
response to estimate the various parameters of the treatment within the given constraints. Over many years,
created fracture. Ultimately, a good match of observed fracture design software has become sophisticated and can
and simulated pressures provides confidence that the also provide additional functionality, such as calculated
model describes well the already executed treatment. pressure responses and model parameter matching.
Because of this goal, many options are built in to Nevertheless, the ultimate goal remains to create a
influence the calculated pressure response. treatment design, rather than to reconstruct an already
Over the past many years fracture simulators have executed treatment. More importantly, an optimum
evolved into complex software that can do related tasks, design software, although it can respect logistical and
such as design a fracturing treatment. Starting with practical constraints, may also help to overcome such
two-dimensional (2-D) models, they have evolved into constraints by pushing the limits of execution and
pseudo-three-dimensional (p-3-D) and, even, in some improving practices and products.
versions, three-dimensional (3-D) models. Design While the basic results for optimum fracture
using these models is done with a series of simulations, dimensions were already presented by Prats (1961),
basically feeding numerous treatment schedules into Economides at al., (2002) suggested the first practical
the simulator, calculating the created fracture geometry step towards such an optimum design methodology. In
and using the geometry to predict the improvement the book Unified Fracture Design (UFD), the original
in well productivity. The final step is selecting the methodology was developed using only the simplest
best injection schedule from the set of calculated 2-D fracture propagation models (PKN and KGD).
ones. However, finding the optimum design for well Since the publication of UFD many applications and
production is not the original objective of such software; improvements have been presented in the literature:
rather, it is a special application of the computational Meyer and Jacot (2005) and Daal and Economides (2006)
model developed originally for other purposes. In considered non-square drainage area; Lopez-Hernandez
fact, even with the most up-to-date versions of such et al. (2004) applied the methodology to compensate
software there is no guarantee that the calculated subset for non-Darcy effects in the fracture; Dietrich (2005)
of treatment schedules contains the optimum one in considered steady-state performance optimization for
any reasonable engineering sense. Approaching the water-flooding scenarios; Poe and Manrique (2005) and
optimum depends very much on user intuition and Manrique and Poe (2007) extended the optimization
on the number of calculated cases. methodology for cases where the transient flow regime
The fact that fracturing technology is often new in lasts extremely long. Field-scale applications are described,
a given geographical area implies that merely copying for instance, in Economides et al.,(2004), Demarchos
standard treatment practices and then making small et al. (2004, 2005), Diyashev and Economides (2005),
steps of gradual improvement is not a viable strategy. In Oberwinkler and Economides (2003, 2004), Rueda et
other words, in a new environment there is less use for al. (2004, 2005), Timonov et al. (2006), Nikurova et al.
golden rules derived from accumulating experience (2006), and Rozo et al. (2007).
and more use of basic design principles and well-defined Some key concepts of the methodology
design methods. The ultimate goal is to provide a Dimensionless Productivity Index (JD), Dimensionless
systematic method for creating the optimum fractured Fracture Conductivity (CfD), Penetration Ratio (Ix) and
well performance for a given well and formation, taking Proppant Number (Nprop) are summarized below.
142
4-5.1 Dimensionless Productivity Index Following McGuire and Sikora (1960) it is

and Dimensionless Fracture Conductivity customary to plot JD (or some similar performance
indicator) as a function of CfD (or some similar variable)
A well in a reservoir developed on a certain pattern with an additional curve parameter representing Ix (or
has a finite drainage area. During most of its lifetime, some similar variable). Such a plot, shown in Fig. 4-42,
it is producing in a stabilized flow regime called used to be popular to select treatment size and fracture
pseudo-steady state (or more precisely: boundary- dimensions simultaneously. Unfortunately, it is not
dominated state). During the stabilized flow regime obvious which curve to select and what point to select on
the productivity index of a well (PI), defined a given curve, because this type of presentation blurs the
by the production rate divided by the pressure cost of creating a propped fracture.
drawdown, is calculated as:
q
J =
(4-105)
p pwf ye=xe
If we consider a square drainage area (see Fig. 4-41)

and compare two possible fractures executed for 2xf
a vertical well in the center, the fracture providing
the larger PI is better. xe
The Dimensionless Productivity Index, JD, is defined
as w
B
JD = 1
J (4-106)
kh h
For an unstimulated well in a circular reservoir, JD is

given by the well-known formula (see also Eq. 2-21): 2xf
1
JD = , Figure 4-41 Basic notation, top and side view (assuming
r (4-107)
ln 0.472 e + s fracture height and formation thickness are the same)
rw
2 Ix= 1
Dimensionless Productivity index, JD
with the skin factor, s, representing deviation ye = x e 0.9

0.8
from the base case (without any near-wellbore 1.5
2xf
0.7
damage or stimulation). xe 0.6
For a fracture stimulated well, JD is affected 1

0.5
0.4
by the volume of proppant placed into the 0.3
pay layer, by the permeability ratio of the proppant 0.2
0.5
0.1
bed and the reservoir, and by the geometry
0.01
of the created fracture. All these factors can be
0
characterized by two dimensionless numbers 0.01 0.1 1 10 100 1000 10,000
the dimensionless fracture conductivity, CfD, and Dimensionless Fracture Conductivity, CfD
the penetration ratio, Ix: Figure 4-42 McGuire-Sikora-type representation of
kf w Dimensionless Productivity Index
C fD = and (4-108)
kx f
As it was first shown in Valk and Economides
2x f
Ix = , (4-109) (1998), certain aspects of the fractured well performance
xe
can be described by a single combination of these two
where the notation is obvious from Fig. 4-41. dimensionless numbers. This combination is called the
143
Modern Fracturing
Dimensionless Proppant Number, Nprop, and it is the 4-5.2 Optimum Dimensionless Conductivity
appropriate way to express the relative size of a given
treatment. Nprop is defined as the ratio of propped volume As seen from Fig. 4-43, Nprop already
of the fracture in the pay to the volume of the reservoir, determines the achievable maximum dimensionless
weighted by twice the permeability ratio: productivity index from the given amount of
resources. The maximum productivity index is
f p 2k V
N prop
= = I x2C fD , (4-110) realized at a well-defined value of the dimensionless
kVr
fracture conductivity. Since a given proppant number
where Vp is the propped fracture volume contained represents a fixed amount of proppant reaching the
within the payzone and Vr is the drained volume of pay, the best compromise between length and width
the pay (in short, the reservoir volume). Recently the is achieved at the dimensionless fracture conductivity
name Stimulation Index has been also suggested for located under the peaks of the individual curves.
Nprop, (Poe, 2006). Nprop represents in a dimensionless The corresponding CfD is called the optimum
form the amount of resources spent on the treatment. dimensionless conductivity.
Algorithms are available to calculate JD as a function of Although large Proppant Numbers lead to larger
CfD with Nprop as a parameter; see, for instance, Romero Dimensionless Productivity Indices, the absolute
et al. (2003), and Meyer and Jacot (2005). Typical maximum for JD is 6/ = 1.909 (the JD for a perfect
results are shown in Figs. 4-43 and 4-44. linear flow in a square reservoir; see for instance, El-
Banbi and Wattenbarger, 1998).
Dimensionless Productivity Index, JD
0.5
Ix= 1 a: Np = 0.0001
At low Proppant Numbers, the optimal CfD = 1.6.
xe = ye
b: Np = 0.0003
c: Np = 0.0006
At larger Proppant Numbers, the optimum CfD is larger.
ye
0.4 2Xf d: Np = 0.001
e: Np = 0.003
When the propped volume increases or the reservoir
Np = 0.1
xe permeability decreases, the optimal compromise
0.3 Np = 0.06 happens at larger dimensionless fracture conductivities
Np =0.03
Np = 0.01
(the penetration cannot exceed unity).
0.2
Np = 0.006
e It is also customary to present the actual
c d
a b
performance of the fractured well in terms of
10 10-3 10-2 10-1 100 101 102
effective fracture length; see, for instance, Barree
-4
Dimensionless Fracture Conductivity, CfD

et al. (2005).
Figure 4-43 Dimensionless productivity index as a
function of dimensionless fracture conductivity with 4-5.3 Optimum Length and Width
proppant number as a parameter, for Nprop <= 0.1
A reasonable optimization scheme for fracture
design can be readily established because once
2.0
xe = ye Ix=1 Np= 100 the optimum dimensionless fracture conductivity
ye Np= 30
1.5 2Xf Np= 10
Np= 60
is identified, the optimum fracture dimensions of
Np= 6
xe length and width are determined by two equations
Np= 3
1.0 (compare with Eq. 2-35, where the one-wing
Np= 1
Np= 0.6 propped volume was denoted by Vf):
Np= 0.3
0.5 0.5
Np=0.1
k f V p / 2
0 x fopt = and
0.1 1 10 100 1000 C fD ,opt kh
0.5
C fD ,opt kV p / 2
Figure 4-44 Dimensionless productivity index as a wopt = .
function of dimensionless fracture conductivity with
kf h
proppant number as a parameter, for Nprop > 0.1 (4-111)
144
Example 4-1: Designing for formation permeability the proppant volume and the retained permeability
of the propped pack. Other items on the bill are
In the center of an 80-acre drainage area, a vertical secondary in the sense that they just assist with placing
well of radius rw = 0.5 ft is to be fracture stimulated. the proppant. Increasing the proppant number costs
The fracturing engineer plans to place 2500 lbm money. On the other hand, optimum dimensions
proppant (retained permeability kf = 50,000 md, with respect to a given amount of resources already
specific gravity of the proppant material is 3.3, and follow from Nprop via CfD,opt.
porosity of the proppant pack under closure stress is Theoretically, it is quite easy to determine the
0.20). The engineer is uncertain about only one thing: optimum treatment, after we know the amount of
the formation permeability. Optimistic estimates go up proppant (Vp) reaching the net pay. In real life, however,
to 10 md, but the pessimistic value is 0.1 md. From a considerable fraction of the injected proppant will end
the above data and taking just 1 ft of the formation, up outside the net pay and hence will not contribute to
the engineer can calculate the propped volume (Vp). the proppant number. A key parameter is, therefore,
Depending on the assumed formation permeability, the fraction of injected proppant reaching the pay: the
however, the Nprop, the maximum achievable JD and proppant placement efficiency, prop. Most profoundly,
the best possible pseudoskin vary considerably. Using prop is affected by the created fracture geometry. For
UFD, we can calculate the optimum placement for instance, if there is a considerable height growth during
various assumed permeabilities: the treatment, prop will be much smaller than for a well-
contained fracture. Also, proppant settling may reduce
Table 4-4 Results for Example 4-1 prop significantly. Therefore, a thorough design process
k,md Nprop JD,opt CfD,opt sf,opt xf,opt,ft has to find the optimum dimensions with respect
0.1 4.35 1.39 6.60 -6.18 757
to the actual Nprop calculated with the actual prop. In
1 0.435 0.695 1.86 -5.46 451
10 0.0435 0.391 1.63 -4.34 152
UFD, this is done iteratively because any change in the
treatment parameters affects prop via the most up-to-
We can continue the previous example, investigating date calculated fracture geometry.
the effect of a wrong assumption about permeability. For
example, if the engineer designs for a 757-ft half-length 4-5.5 Taking Into Account Operational Constraints
fracture, but the real formation permeability is k = 10
md, she will not realize a JD = 1.39 nor even a JD = 0.391 There are many technical constraints to consider during
value (the maximum value for the optimum treatment fracture design. By far the most significant limitation is
and resulting geometry for a 10-md reservoir). The the maximum concentration of proppant in the slurry
actual realized JD will be only 0.359. On the other hand, feed that we can safely pump.
if she designs for a 152-ft half-length fracture, but the This limitation (often expressed in added proppant
real formation permeability is k = 0.1 md, she will not concentration, ppga) often makes it impossible to place a
realize the achievable JD = 1.39 value, but less than half given amount of proppant according to the requirements
of it: JD=0.631. All of these calculations can be easily of the theoretical optimum. Depending on the equipment
done with the publicly available FracPI.xls spreadsheet, and materials, the pumpable limit is around 0.35 to 0.45
available via http://www.pe.tamu.edu/valko volume fraction of solid, translating to 14 to 18 ppga
added proppant concentration.
4-5.4 Treatment Sizing and Proppant When the target length is reached, the actual
Placement Efficiency dynamically created fracture volume will be determined
by the dynamically created width, which, in turn,
One can consider the selection of a target Proppant is determined by the reached length, the injection
Number as the sizing of the treatment. The sizing rate, fluid rheology and formation elastic modulus.
determines the treatment cost. In fact the operator (Interestingly, the leakoff parameters significantly
mainly pays for the Proppant Number that is, for affect pumping time but do not really influence the
145
Modern Fracturing
dynamically created width at the time of reaching fracture width is inflated until the required
the target length.) The dynamically created fracture amount of proppant is placed. In the TSO phase,
volume multiplied by the maximum pumpable the inflating width causes a sharply increasing
concentration gives the mass of proppant we can net pressure; this is the main signature of an
place into the fracture. If this amount of proppant is actual TSO. From the point of optimization,
less than the amount we started the design with (for the TSO design means we can work around the
instance the 2500 lbm per 1 ft of pay in our previous technical limitation of the maximum pumpable
example), the design engineer should select among concentration and use smart tactics to realize the
various practical options: theoretically achievable maximum JD.
1. A viable option is to keep the target length Additional constraints. In actual fracture design,
but place only as much proppant as the limit we have to consider various additional technical
allows. In simple terms we can call this the constraints. For instance an approximate knowledge of
reduced proppant mass option. This will result the damage penetration distance may put a special
in a suboptimal placement and a reduction in requirement on the minimum fracture length. Net
realized JD. However, the cost of the treatment pressure limitation may also force us to depart from
is also reduced because we place less proppant the theoretical optimum. When designing for longer
than originally targeted. fractures, one should consider a minimum width and/or
2. More rigorous is to keep the original mass of minimum areal proppant concentration requirement.
proppant unchanged and stretch the fracture All these issues can be handled similarly: The key is to
as much as necessary until it can accept all the depart from the theoretical optimum only as much as
proppant (in our specific example, 2500 lbm per necessary to satisfy an additional constraint.
1 ft). Such a design will be suboptimal in the
sense that it will realize less than the theoretically 4-5.6 Using Fracture Propagation Models
achievable JD. The actual JD achieved will be
suboptimal with respect to the theoretical The above described design logic is basically
setting but it will be optimal in a practical independent of the method of predicting height
sense: It will provide maximum JD within the containment and fracture propagation.
technical limitation of the maximum pumpable
concentration. As we express this fact in UFD, we 4-5.6.1 Height containment
depart from the theoretical optimum as little as
possible to satisfy an additional constraint. Height is often calculated by subtracting the
3. A third option in sufficiently soft, unconsolidated vertical stress profile from fluid pressure and
formations is to use the procedure of Tip calculating stress intensity at the upper and lower
Screenout (TSO) (see Section 4-7.3.2). In a tips. One way to provide sufficient data for the
TSO design, we adjust the proppant schedule so data-hungry 3-D simulators is using rock tables
that at the time of reaching the target length (for (Leshchyshyn et al., 2004).
instance, 152 ft for the 10-md case in the previous Recently considerable effort has been spent on
example) we also reach a critical situation with fracture diagnostics methods (Cipolla et al., 2000, and
respect to proppant transport inside the fracture. Mayerhofer et al., 2006.) A common theme emerging
The critical situation can be postulated in various from such investigations is that fracture height is
ways and can be specified in various forms of controlled by far more complex parameters than just
TSO criteria. The most important thing is vertical stress profile (Weijers, 2005). Consequently,
that once this critical condition is reached, the the simple concept of aspect ratio (2xf /hf) seems to
fracture length does not increase any more. In be more predictable, at least in a certain well-defined
the subsequent part of the treatment only the geographical/lithological setting.
146
4-5.6.2 2-D models Table 4-6 Design Procedure Based on UFD Concepts
1. Select the type of proppant to be pumped.
The two most widespread 2-D design models are due to
2. For various treatment sizes (e.g. 25,000, 50,000, 100,000,
Perkins and Kern, 1961 (PKN), who used the vertical 150,00, 200,00, 300,000, 500,000 lbm)
plane strain assumption; and Kristianovich-Zheltov, 3. Assume a fracture height at end of pumping and calculate
1955, and Geertsma and deKlerk, 1969 (KGD), who corresponding proppant placement efficiency.
used the horizontal plane strain assumption for the 4. Estimate retained proppant pack permeability, taking into
consideration possible further reduction due to non-Darcy flow
crack opening. For simplicity, those early models (see Sections 2-5.2, 8-7.3 and 8-7.4).
assumed a rectangular fracture shape (side view). 5. Using the proppant placement efficiency and the effective
The basic outcome of those models was the width proppant pack permeability, calculate Nprop, the optimum JD and
the optimum target length, xf. (Here the knowledge of formation
equation relating length and width via additional permeability is crucial.)
parameters including injection rate, fluid rheology
6. Calculate dynamic fracture dimensions and net pressure at the
parameters and formation elasticity parameters. A time of reaching the target length (see Table 4-5 for the simplest
modified PKN model assuming elliptical side view is 2-D model equations or use a more sophisticated 3-D model).
7. Using leakoff parameters CL and Sp and an opening time
summarized in Table 4-5. distribution factor (Economides et al., 2002), = 1.5, solve the
material balance:
qi ,1wing
Table 4-5 Modified PKN Model for Elliptical Fracture t (2 C L ) t 2 S p wavg = 0,
Shape (consistent units), power law fluid h x
f f
for injection time, ti, or use a 3-D model to obtain the necessary
1.107 + 0.891n injection time.
w0 = exp
1 + n Vi ,1wing = qi ,1wing ti
n 8. Calculate injected slurry volume from:
1
1 + ( 1)n 2+ 2n
Maximum Width at K 2+ 2n h f x f wavg
Wellbore n 9. Calculate fluid efficiency from: =
1 Vi ,1wing
q n x 2+2n
i f 10. Determine pad injection time, tpad, and create a proppant
E h n-1
f schedule according to Nolte (1986):

t t pad 1
c = ce where t pad = ti and = and
ti t pad 1+
Average Width wavg = w0
5
m1wing
ce =
Vi ,1wing
Surface of one
Face of one Wing
A = x f hf
4 11. Convert concentration into proppant added to clean fluid:
c
cadded =
1 (c / p )
Volume of one
V = Awavg
Wing 12. Assume uniform concentration at end of pumping, ce, or
calculate proppant concentration at end of pumping from a more
detailed model taking into account various possible proppant
Net Pressure at 2E transport mechanisms and possible settling.
pnet = w0
Wellbore hf
13. Using the concept of fixed aspect ratio or net pressure to
stress contrast and fracture toughness, recalculate fracture height
at end of pumping.
Whether a simple 2-D width equation or a more 14. Knowing the fracture height (and possible detailed proppant
sophisticated fracture simulator is used, the goal of distribution), recalculate proppant placement efficiency.
the design should be to create optimum fracture 15. Repeat steps 4 through 13 until convergence.
dimensions with respect to the actual treatment 16. Having constructed the optimum treatment for each specified
treatment size and having calculated the corresponding
size (proppant volume used) and to select among dimensionless productivity indices realized, select the particular
possible treatment sizes. size satisfying a specific economic criterion (e.g., maximizing Net
Present Value).
147
Modern Fracturing
The reader is cautioned that several software We use the thickness of the perforated interval as
products claim to do net present value optimization, the fracture height at the wellbore. However, if the
but they do not ensure that the individual fracture fracture design would call for larger lengths, we
geometries compared are optimal with respect to adjust the fracture height not to exceed the aspect
their own treatment size. ratio, AR=2xf /hf=4. (The calculations are made with
an in-house non-commercial fracture design program
Example 4-2: Optimum Design Based on Aspect Ratio written in Mathematica.)
and Minimum Height
Table 4-8 Main Results for Example 2
As an example, we consider a certain geographical Case A (k = 0.25 md)
area where, according to some fracture diagnostics Proppant in Fracture 300,000 lbm
results, the aspect ratio of created fractures seems to Clean Liquid Needed 117,200 gal
be 4. Because this is hard rock country, extensive
Fraction of Proppant in Net Pay (Proppant
TSO (width inflation) is not a viable option. We are 0.514
Placement Efficiency)
going to design an optimum treatment for two cases: Proppant Number 0.979
In Case A, formation permeability is 0.25 md; in Case
Optimum (and Realized) Dimensionless
B, formation permeability is 2.5 md. The vertical plane Productivity Index
JD = 0.884
strain elliptical model is used (PKN width equation
Theoretical Optimum Dimensionless
combined with an elliptical shape of the fracture face). 2.35
Fracture Conductivity
Additional data are summarized in Table 4-7. Target half-length 426 ft
213 ft (aspect
Table 4-7 Input Data for Example 4-2, Optimum Design Fracture Height at Wellbore
ratio controlled)
Based on Aspect Ratio and Minimum Height
Fracture face Area of one Wing 71,230 ft2
Drainage area 40 acre
Net pay thickness 70 ft Areal Proppant Concentration 2.11 lbm/ft2
Gross pay thickness Slurry Efficiency 0.416
(also perforated interval
180 ft
and hence minimum Pad Time 42.0 min
height)
Pumping Time 102 min
Plane strain modulus 2 106 psi
Proppant Intermediate strength ceramics Needed Maximum Added Proppant
7.2 ppga
Concentration
Total proppant mass 300,000 lbm
Proppant retained Nolte Exponent and Pad Fraction for
0.412
permeability (after all 15,000 md Proppant Schedule
effects accounted for)
Specific gravity 3.1
8
Proppant pack porosity
0.2
under closure stress
Allowed maximum 6
added proppant 15 ppga
concentration in feed
cadded, ppga
Frac fluid HPG 40 (Borate-crosslinked) 4

Rheology flow behavior
0.45
index
Rheology consistency 2
0.6 lbf sn / ft2
index
Slurry rate 30 bpm
0
Leakoff coefficient in
0.003 ft/min0.5 0 20 40 60 80 100
net pay
t, min
Spurt loss neglect
Fluid loss multiplier out Figure 4-45 Continuous and staged proppant schedule
0.67 (bottomhole) for Case A
of net pay
148
100 14
12
50 10
cadded, ppga
8
z, f t
0 6
4
- 50
2
0
- 100 0 10 20 30 40 50
t, min
0 100 200 300 400 Figure 4-47 Continuous and staged proppant schedule
x, ft
for Case B (at bottomhole conditions)
Figure 4-46 Created frac profile for Case A. Dashed lines
show gross pay; the gray area represents net pay (but its On the other hand, in the higher-permeability case
exact distribution within the gross is not known) (k = 2.5 md), the optimum placement would call for
a short fracture (xf,opt = 175 ft), but at that length the
Table 4-9 Main Results for Example 2 dynamically created width is not enough to take all the
Case B (k = 2.5 md) proppant, even at the largest pumpable concentration.
Proppant in Fracture 300,000 lbm
We use option No. 2 of Section 4-5.5 and elongate the
Clean Liquid Needed 57,600 gal
fracture as much as necessary. After elongation, the target
Fraction of Proppant in Net Pay
(Proppant Placement Efficiency)
0.598 half-length becomes 335 ft, and the height at the wellbore
Proppant Number 0.114
remains at the minimum value of 180 ft because the
aspect ratio condition would require a smaller height (so
JD,opt = 0.482
Productivity Index the already perforated 180 ft has no reason to increase).
1.62
See Table 4-9 and Fig. 4-48.
Fracture Conductivity
Realized Dimensionless Productivity
JD = 0.441
Index (with elongated frac) 100
Realized Dimensionless Fracture
1.62 / 1.912=0.444
Conductivity
50
335 ft (optimum
Target half-length
elongated by 1.91)
z, ft
0
180 ft (minimum /
Fracture Height at Wellbore
perforation controlled)
- 50
Fracture face Area of one Wing 47,300 ft2
Areal Proppant Concentration 3.17 lbm/ft2 -100
Slurry Efficiency 0.457
0 50 100 150 200 250 300
Pad Time 20.5 min
x, ft
Pumping Time 54.9 min
Figure 4-48 Created frac profile for Case B, the gray
Needed maximum added Proppant area indicates net pay (its exact distribution within the
15 ppga (limit controlled)
Concentration gross is not known)
Nolte Exponent and Pad Fraction for
0.373
Proppant Schedule
4-5.6.3 3-D models

In the lower permeability case (k = 0.25 md), the After the issues of treatment size and corresponding
optimum fracture half-length of 426 ft induces a fracture optimum fracture geometry are settled, it is useful to
height of 213 ft (via the aspect ratio condition). See Table take into account more information (if available) and
4-8. The fracture length creates enough dynamic width to simulate the growth of the fracture. Figure 4-49 shows
place the proppant with a moderate final added proppant additional key data necessary for simulating the fracture
concentration, way below the technical limit. propagation for Example 4-2 Case A.
149
Modern Fracturing
Stress
Gradient Stress
Youngs
Modulus
4-6 Predicting Production Increase
15900
16000
16100 4-6.1 Pseudo-radial Concepts:
16200 Equivalent Wellbore Radius, Fracture Skin
TVD (ft)
16300
16400
16500 In the case of a propped fracture there are several
16600
ways to incorporate the stimulation effect into the
16700
0.6 0.7 0.8 0 dimensionless pseudo-steady-state productivity index
7
11 00
12 00
13 00
0
6
+0
00
+0
0
0
0
(psi/ft)
indirectly, via an additional parameter. One can use
10
0e
0e
1.
5.
(psi)
(psi)
Poissons
Toughness
the pseudo-skin concept:
Ratio
1
JD = ,
r 3 (4-112)
ln e + s f
rw 4
or Cinco-Ley and Sameniegos (1981) f-factor
concept:
0.1 0.2 0.3
1
00
10 990
JD = ,
.0
9.
(4-113)
00
(psi^1/2) re 3
99
ln + f
Figure 4-49 Rock properties and stress state for 3-D xf 4
design
or Prats (1961) equivalent wellbore radius concept:
A snapshot of the areal proppant concentration at
the end of pumping as obtained with a commercial 1
JD = .
re 3
3-D fracture simulator is shown in Fig. 4-50. ln (4-114)
r 'w 4
Concentration / Area (EOJ)
The form of these expressions may induce
additional thoughts in petroleum engineers who
may be tempted to talk about pseudo-radial flow.
In reality, these expressions are all equivalent (if used
correctly) and are not associated with any deeper
information regarding the nature of the flow field. The
actual parameters, sf, f or rw can be easily converted
0 200 400 600
into each other, but they bear completely formal
Stress
Width meaning: They are just different scales to express the
Profiles
16280
deviation from some pre-defined basis of comparison
16320
(the undamaged, unstimulated well). These concepts
16360
16400 might be handy to use in the context of some part of
TVD, ft
16440 the transient flow, but they might be very misleading

16480 in other periods of the life of a well.
16520
16560
4-6.2 Finite Reservoir Concepts, Folds of Increase
16600
-0.5 0 0.5
0
11000
12000
13000
00
Width, in.
10
Stress, psi
To make things even more blurred, fracturing engineers
like to speak about folds of increase, flow efficiency,
Figure 4-50 Areal proppant concentration obtained
from 3-D design using a commercial fracture
effective half-length, etc. Each of these concepts has
simulator (Case A) some advantages in conveying certain basic information
150
visually. There are, however, serious dangers associated such a complete presentation is possible only for
with using any of the concepts of Section 4-6.1 and well-defined flow history. Mathematically, the
2. For instance, although Eq. 4-112 suggests that the easiest method is to handle the constant-rate type
fracture can be represented by a pseudo-skin, it is not true of flow history. In such case, the late-time stabilized
that the same pseudo-skin value can be used in transient part is called pseudo-steady state. Other types of
regime as in pseudo-steady-state regime. Also, we already flow histories (such as the one implicitly defined
saw in Example 4-1 that the achievable pseudo-skin is by constant wellbore pressure) may lead to slightly
limited partly by the fracture and partly by the well/ different productivity indices at any moment of time;
reservoir system, including the location of its boundaries. even their stabilized values might differ in pseudo-
This is why we suggest using JD directly. Equation 4- steady state (Helmy and Wattenbarger, 1998).
114 might suggest that an equivalent wellbore radius Of course, it is possible to calculate a productivity
remains the same in various flow regimes (and this idea index curve for any specified rate history, but this
might be reinforced further if we speak about effective would be impractical in general. In reality, we do
fracture length) but in reality none of these indicators not know the production history that will happen
can be used in such a general way. Strictly speaking, the in the fractured well in the future. Fortunately, the
productivity is a common feature of the reservoir, the productivity index curve obtained with the constant-
fracture, the boundaries and the actual flow history. rate condition is generally a good average indicator
that provides a reasonably accurate forecast of any
4-6.3 Combining Productivity particular production history. Valk and Amini
Index and Material Balance (2007) developed a reliable and efficient procedure,
the method of Distributed Volumetric Sources (DVS),
In order to forecast production from a fractured well, to generate the combined JD curve that describes
we need to combine the Productivity Index with both transient and stabilized (pseudo-steady state)
material balance. production regimes.
4-6.3.1 Pseudo-steady state Example 4-3: Transient and Pseudo-steady State

Dimensionless Productivity Index
The pseudo-steady-state value of JD can be easily
obtained from Figs. 4-42 and 4-43, or from the We calculate the productivity index of a fractured
FracPi spreadsheet (see Section 4-5.3). Optimum vertical well with CfD=1.6 and Ix = 0.25 (pseudo-
values can be also calculated from correlations steady state optimum for Nprop = 0.1) using the DVS
(see Section 2-5.1). method (Valk and Amini, 2007). See Figure 4-51.
After we know the JD value, we can obtain
the Absolute Openflow Potential (expressed in
102
Calculation Parameters:
mscf/day) from: CfD = 1.6
Ix = wx/xe = .025
kx k y h 101
AOF = J D m ( p ), (4-115)
1424 T
100
as a function of the average reservoir pressure, p .

10-1
10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 101
4-6.3.2 Combined transient and stabilized flow
Dimensionless Time, tDA
To describe the production (often quite significant) Figure 4-51 Transient and pseudo-steady-state
during the transient period, we need a description of dimensionless productivity index (Nprop = 0.1) calculated
JD covering the whole time span. Strictly speaking, with the DVS method (Example 4-3)
151
Modern Fracturing
The productivity index stabilizes around The notation JD,tDA in step 5 means that
dimensionless time, tDA = 0.3 (with respect to the we should use the dimensionless productivity
drainage area). The stabilized value is JD = 0.470 index corresponding to the dimensionless
according to the DVS method. This compares well equivalent of the current time (elapsed from the
with the Unified Fracture Design value of JD = 0.467 start of the production.)
(and probably is even more accurate).
The transient flow regime can last for months or 4-6.4 Reservoir Simulation and Nodal Analysis
even years for tight gas wells. It is sometimes speculated
that a fracture designed for optimum with respect to In their everyday jobs, petroleum engineers rely
pseudo-steady state would somehow under-perform more and more on reservoir simulators and wellbore
if we also take into account the transient period. This simulators. One of the key trends in the industry is
is actually a misunderstanding. For short times, some to integrate these powerful tools (and provide them
(shorter and wider) fractures can outperform the with as much information as possible). One should
pseudo-steady state optimized fracture, but sooner or keep in mind, however, that the use of such an
later the latter will catch up. By the time the whole ultimate tool neither is nor ever will be easy. Fit-
drainage volume is affected (that is, at dimensionless time for-purpose simulators and semi-analytic methods
about 0.3), the pseudo-steady state optimal fracture is are readily available and are usually more practical
not only the best for the current productivity, but also for optimization. As a further illustration we
has produced the maximum cumulatively with regard to consider a non-trivial problem.
the given drawdown history, or the least final drawdown
with regard to the given production history. Example 4-4: Horizontal well with transverse fractures
A rather straightforward approach to in an anisotropic reservoir
forecasting the production from a fractured well is
depicted in Table 4-10. We are to place four transverse fractures
intersecting a horizontal well in an anisotropic
Table 4-10 Production Forecast Method (Field units) reservoir. The drainage volume and permeability
1. Prepare pseudopressure function (see Section 2-3.1) from
anisotropy are shown in Fig. 4-52.
p
p
The well will also contribute to the production
m( p ) = 2 dp. and is considered to be of infinite conductivity.
p0
( Z ) p
The four fractures are of finite conductivity
2. Specify initial pressure. (infinite-conductivity fractures are more a
mathematical construction than reality) and suffer
3. Specify wellbore flowing pressure.
from convergence skin (see Section 2-5.3). The
4. Take a time interval, t. total amount of proppant used is 100,000 lbm,
5. Calculate production rate and production in the time interval
and the retained permeability of the proppant pack
(see also Section 2-3.2) from: is kf=10,000 md. The specific gravity of the
proppant material is 2.65, and the porosity
V = Awavg
and of the proppant pack is 0.3.
The reader may have the opinion that the
G p = qsc t. ultimate method of calculating production from such
a complex well/fracture system is to use a reservoir
6. Apply material balance and calculate new average pressure
(Section 2-7.1) from simulator with a wellbore model. That is a possibility
pi Z G p because the simulator can handle multiple transverse
p= 1 .
Z i Gi fractures with finite conductivity. However, despite
all the progress made in reservoir simulation, such a
7. Repeat steps 1 6.
task might be challenging even today.
152
kx= 4 md 10

ky= 0.25 md
kz= 0.1 md 7
5
ze=100 ft
kx 3
kz
xe=1320 ft 2
ky ye=1320 ft 1.5
1
0.00001 0.0001 0.001 0.01 0.1 1
Dimensionless Time, tDA
Figure 4-52 Drainage volume in anisotropic reservoir
(for Example 4-4) Figure 4-54 Log-log plot of the transient and pseudo-
steady state productivity index for Example 4-4. The
stabilized value is JD = 1.8
4-7 Fracturing Under Specific

Circumstances
ze
There are various different types of hydraulic fractures,

which have evolved around the basic process of creating a
fracture and then propping it open. The type of treatment
kx
kz selected depends upon formation characteristics
ky
xe (permeability, skin damage, fluid sensitivity, formation
strength), treatment objectives (stimulation, sand control,
ye
skin bypass or a combination) and practical constraints
Figure 4-53 Horizontal well (rw = 0.3 ft) drilled in the (cost, logistics, equipment etc).
direction of minimum horizontal stress (and minimum
horizontal permeability). The horizontal well is 75%
4-7.1 Tight Gas
penetrating and is intersected by four transverse
fractures created from a total of 100,000 lbm proppant
(for Example 4-4) This type of fracturing is often carried out in tight
gas formations, found in areas such as the Rocky
Using the DVS method, one can find out that Mountains, Algeria, Western Germany, parts
for the configuration shown in Figs. 4-52 and 4-53, of Australia and many other places worldwide.
the optimum fracture half length is xf=78 ft, and the Permeabilities for such formations range from 1 md
optimum areal proppant concentration is 4.6 lbm/ down to 1 d and less.
ft2.(This quite large value is not surprising because it Tight gas formations are referred to as tight
should compensate the detrimental effects of converging because they have low permeability and are also usually
flow in the fractures.) The resulting transient and composed of so-called hard rocks. In the context of
pseudo-steady state JD is shown in Fig. 4-54. fracturing, hard means high Youngs modulus and
After the transient and stabilized productivity of low fracture toughness, which in turn means that any
the well/fracture configuration is known, it can be fracture produced in such a formation will tend to be
easily incorporated into a reservoir simulation model long and thin. This is one if the rare instances in the
and/or a nodal analysis model. However, in such a case production of oil and gas where the natural tendencies
the actual productivity of the complex well/fracture of the formation work in our favor because the ideal
architecture is not an output of the simulation tool, fracture geometry for low-permeability formations
but rather an input to it. (see below) is also long and thin.
153
Modern Fracturing
Tight gas fracturing forms by far the largest sector to be long. This means that the fracturing fluid has
of the fracturing industry. In 2006, over 36% of the to suspend the proppant for a relatively long period
fracture treatments performed by one of the leading of time at bottomhole temperature.
service companies were in formations of below 0.1 md Therefore, hydraulic fracture treatments in low-
(and this does not include treatments performed in permeability formations tend to have fairly large
coalbed methane or gas shale formations). Most of the fluid and proppant volumes, although the overall
industrys perceptions of hydraulic fracturing originate proppant concentration in the fluid is relatively
from this sector of the business. Many gas reservoirs low. Pad volumes are small. Treatment fluids are
can only be produced because of hydraulic fracturing. usually fairly robust, capable of maintaining viscosity
In many areas, there is such confidence in the process for extended periods of time.
that whole fields are developed without any hope of
ever being economic without fracturing. 4-7.1.2 Effective vs Actual Propped Length
4-7.1.1 The Importance of Inflow Area It is not uncommon to perform pressure transient
testing on fractured tight gas wells. Because of the
In order for hydrocarbons to flow down the fracture, reciprocal productivity index methods pioneered by
rather than through the adjacent formation, Crafton (1997) (see Section 3-7), operators can do this
the fracture must be more conductive than the with post-treatment production data, rather than having
formation. Given that the permeability for 20/40 to wait inordinately long periods of time for pressure
frac sand is 275 darcies (if closure pressure is build-up data. More often than not, this analysis shows
below 3,000 psi), we can see that even a very narrow that the effective propped fracture length is much less
fracture will have a much higher conductivity than than that predicted by the simulators. Obviously, some
the formation. This does not consider the effects of of this inconsistency is due to the inaccuracies of the
non-Darcy flow (see Section 2-5). fracture simulators. However, it is also apparent (e.g.,
Therefore, the most significant limiting factor Cramer, 2003) that fractures in tight gas formations
defining how far the reservoir production has are also susceptible to loss in effective fracture length
increased is not how conductive the fracture is (any for a variety of reasons. This is because fracture
realistic propped fracture will be significantly more width is very limited, especially towards the fracture
conductive than the formation), but rather how easily tip, and conductivity is relatively low. As a result, it
the formation can deliver the hydrocarbons to the does not take very much damage to the fracture to
fracture. Therefore, when treating low-permeability change sections of the propped length from infinite
reservoirs, fractures should be designed with a specific conductivity (i.e., no significant pressure drop) to
minimum fracture conductivity and a large surface area. finite conductivity (significant pressure drop). When
Because formations are usually limited in height, this a fracture has some sections of finite and some sections
means designing for maximum fracture half length, of infinite conductivity, it will behave as if it were a
xf. See Section 4-5.2 for a detailed discussion of how to shorter fracture of infinite conductivity.
determine the required fracture conductivity.
Because formation permeability is low, Causes of such losses of fracture conductivity include:
fluid leakoff also tends to be low. This has two
consequences. First, pad volumes tend to be very Loss of fracture width due to embedment.
low, relative to the rest of the job volumes. In some Loss of fracture conductivity due to proppant
cases, a pad is hardly needed at all; the proppant- crushing (i.e., poor-quality proppant and/or
laden fluid can be used to create the fracture. The higher than expected closure stress).
second consequence is that fracture closure time Loss of fracture conductivity due to damage from
the length of time taken for the fracture to close on fracturing fluid residues (poor fluid design and/or
the proppant after the treatment has finished tends treatment execution).
154
Loss of fracture conductivity and/or length The phenomenon of non-Darcy flow

due to immobile fluids (reservoir fluids or through proppant is covered in more detail
fracturing fluids). in Section 8-7.3.
Insufficient fracture conductivity due to poor
treatment design. 4-7.2.2 Wellbore Connectivity
Reduction in effective fracture permeability
due to unexpected multiphase and/or non- As discussed in detail in Section 2-5, turbulence
Darcy flow effects. can have significant effects on the production
from high rate gas wells, due to the constriction
It is important to realize that good treatment of flow at the connection between the propped
design, together with selecting suitable fluid and fracture and the wellbore. Often, a hydraulic
proppant systems, can go a long way to mitigating fracture will only connect to the wellbore via a
or even eliminating these effects. However, it is small number of propped perforations and this
also important to realize that fractures in tight can have a significant impact on post-treatment
gas wells are particularly susceptible to loss of production from high rate gas wells. Not only is
effective fracture length, and this must be allowed production reduced because of turbulence effects,
for during treatment design. but the extremely high flow velocities seen in
such areas can produce excessive drag forces on
4-7.2 High-Rate Gas Wells individual proppant grains. This can lead to
proppant flow back (see Section 8-10).
4-7.2.1 Non-Darcy Flow Efficient and effective treatment design, coupled
with an engineered perforation strategy, can do
Section 2-3.1 introduced the concept of non-Darcy a lot to mitigate these effects. However, they
flow within the reservoir and also introduced the can still be significant and must be allowed for
Forcheimer equation (Eq. 2-2). Non-Darcy flow can when designing a treatment. For instance, there is
also be a significant problem within the propped little point in pumping a 600,000 lbm treatment
fracture. For linear flow along a propped fracture, the with a length of 400 ft, when the near wellbore
Forcheimer equation is usually expressed as: turbulence effects make it behave like a 150,000
lbm, 200-ft-long treatment.
p v
= + v 2 .
(4-116) The issues surrounding the connectivity of
L k
the fracture to the wellbore will be discussed in
The left-hand side of Eq. 4-116 represents the total detail in Chapter 6.
energy lost per unit length. The expression ( / k)
represents the energy loss due to viscous friction effects 4-7.3 High-Permeability Wells
and the expression (v2) is a kinetic energy term that
represents energy loss due to non-Darcy inertial flow High-permeability fracturing is, not unexpectedly,
effects. The constant, , is specific to a given proppant the opposite of low-permeability fracturing. In
type and closure stress, and can usually be obtained from high-permeability formations, moving the fluid
tables provided by proppant manufacturers/suppliers. through the rock to the fracture is relatively easy.
In hydraulic fracturing, the inertial flow The difficulty lies in creating a fracture that has
effects manifest themselves as an effective drop in sufficient conductivity to increase productivity. This
proppant permeability; this needs to be allowed means that for this type of treatment, generating
for during treatment design. Data for this is widely sufficient fracture conductivity is more important
available (such as from Stim-Labs excellent PredictK than generating fracture length (remembering that the
software) if the engineer has some idea of the fracture needs to be of sufficient size to mitigate the
expected post-treatment flow rate. effects of turbulence in the formation).
155
Modern Fracturing
4-7.3.1 The Importance of Fracture Conductivity transmitted to the tip. However, as proppant builds
up in the tip, the pressure at the tip will fall until
With reference to Eq. 4-108, in order to maintain the it is no longer sufficient to make the fracture
ideal fracture conductivity of 1.6 (see Section 4-5.2) grow. This is the TSO.
as formation permeability increases, it is necessary
Proppant Fracture Tip
to increase width, increase proppant permeability
and/or reduce fracture half length. Given that for
pnet
a specific treatment, proppant permeability will
be fixed, then there will be an ideal relationship
between propped length and propped width. p
Maintaining this relationship for increasing formation
permeability means designing treatments that are Figure 4-55 Diagram illustrating the tip screenout
increasingly short and wide.
In such fracture/formation systems, the factor that However, fluid is still being pumped into the
controls post-treatment production is the pressure drop formation at a (usually) constant rate. Given that the
through the proppant pack. These fractures are often TSO will not significantly affect the fluid leakoff rate
referred to as finite-conductivity fractures. (at least initially), the fracture volume has to increase
at the same rate, even though the fracture is no
4-7.3.2 The Tip Screenout longer gaining length and height. This means that the
fracture width has to increase.
As previously discussed, the relationship between fracture As the tip screenout occurs, the net pressure starts
length, width and height is defined by the formation to rise because this is usually directly proportional to
rock mechanical characteristics and is generally outside fracture width. The rate at which the net pressure
the control of the frac engineer. Usually, all that can be starts to rise is controlled by the formations Youngs
controlled is the fracture volume and the final propped modulus: If the rock is too hard, the pressure will
width (because the initial, created fracture width will close rise too quickly and the treatment will soon be over.
onto the proppant and hence the proppant concentration Therefore, in order for a formation to be a candidate
per unit area will control the final width). for a TSO treatment, it must have sufficiently high
However, as permeability rises, it becomes fluid leakoff to allow proppant accumulation in the
increasingly difficult to produce sufficient width fracture tip and sufficiently low Youngs modulus so
without also generating excessive length. A point is that the pressure does not rise too quickly.
reached at which is it no longer possible to generate
sufficient width; without artificial intervention, 4-7.4 Unconsolidated Formations
fractures would not be effective. This permeability
is in the region of 50 to 100 md. Above this The majority of high-permeability fracture treatments are
permeability range, it is necessary to use a technique performed in weak or unconsolidated formations. Such
known as the Tip Screenout (TSO) in order to treatments are often referred to as frac and packs or
artificially generate extra width. simply frac-packs because they combine the effects of
With reference to Fig. 4-55, a TSO is achieved by fracture stimulation and gravel packing.
forcing proppant into the fracture tip at a relatively
early stage in the treatment. As the proppant collects 4-7.4.1 Re-Stressing the Formation
in the fracture tip, a pressure differential is created
by fluid trying to penetrate through the proppant to An undisturbed formation will normally exist in a state
reach the tip. In the main body of the fracture, pnet > of three-dimensional compressive stress. Drilling the
pext, which means the energy in the fluid is sufficient wellbore and applying a drawdown radically change
to make the fracture propagate, if this pressure is the stresses. Although the tangential and vertical
156
stresses will remain compressive, the radial stresses may to ensure the maximum possible re-stressing of the
become tensile. When this happens in a weak or poorly formation. Therefore, the final stage is designed with
consolidated formation, sand production occurs. extra slurry, so that the screen/casing annulus is filled
The major mechanism for controlling sand with proppant after the screenout.
production (sand control) is re-stressing the formation.
In this process, proppant is forced into any void spaces 4-7.5 Skin-Bypass Treatments
behind the casing (including perforation tunnels),
physically compressing the formation. As the slurry is Skin-bypass treatments are designed to do exactly
pumped into these areas, the voids fill up and a screenout what the name describes bypass skin damage.
will occur. The higher the pressure rise that results, the These treatments are not necessarily designed to
more the formation is compressed, or re-stressed. be the absolute optimum stimulation treatment for
Fracturing the formation offers an unparalleled the well. Instead, these treatments are designed to
opportunity to re-stress the formation because be small, cost-effective and easy to perform. Often
the fracture width will compress the formation these treatments are pumped in places where space
either side of the frac. or equipment weight is a limiting factor, such
as offshore. In many cases, if the engineer were
4-7.4.2 The Frac-Pack Treatment given a technical free hand to design the optimum
treatment, the job would be much larger. However,
The main elements of a frac-pack completion are given the restraints of cost and space that are often
illustrated in Fig. 4-56. Put simply, the treatment is placed upon engineers, the skin-bypass frac is an
pumped down the tubing and into the crossover tool. attempt (often highly successful) to produce effective
The crossover tool transfers the flow into the annulus stimulation. Rae et al. (1999) derived the following
between the casing and the blank pipe. The slurry then equation for predicting production increase from a
flows down between the screens and the casing, and then skin-bypass fracture treatment:
on into the perforations.
JF ln (re rw ) + s
=
, (4-117)
Jo ln 4 (C fD x fD )

Tubing
where s is the skin factor of the unfractured well,
Packer Elements
Jo is the initial PI of the well, JF is the PI of the
Slips fractured well and xfD is the dimensionless fracture
half length (i.e. xf/re).
Crossover Tool The skin-bypass frac can also be considered as a
more effective alternative to matrix acidizing when
Blank Pipe
Proppant
factors such as mineralogy, temperature, logistics and
cost prevent the use of acid.
Figure 4-57 shows the basic concept behind
the skin bypass frac. Although the formation has
considerable damage (dark-shaded area), this is
Screens
effectively bypassed by the more conductive path
Sump Packer created by the fracture. In order for the fracture to
produce a production increase, it does not have to
be more conductive than the formation. It merely
Figure 4-56 Typical components of a frac-pack completion
has to be more conductive than the damaged area.
The treatment is designed as a conventional high- Of course, normally the fracture conductivity is
permeability TSO treatment. The only exceptions are significantly higher than this. Given that skin-
that the treatment is designed to screenout at the end bypass fracs are normally carried out on marginal
157
Modern Fracturing
wells (wells that cannot justify the expense of form is the wellbore and the next best place for it
a major stimulation treatment), often economics to form is the fracture. The worst place for it to
dictate that significant production increase (over form is the formation.
and above that obtained from bypassing the To that effect, fractures are designed to minimize the
skin) must be obtained. pressure drop in the reservoir. This produces the same effect
on fracture geometry as having low permeability: The
inflow area (and hence xf) needs to be at a maximum.
4-7.7 Shale Gas and Coal Bed Methane
A more detailed discussion of the specific aspects of

fracturing these types of unconventional reservoir will
be the subject of Chapter 11.
4-7.7.1 Gas Shales

Permeability
Low High Fracturing in gas-bearing shales (such as the Barnett shale
in north-central Texas) has become a major focus of the
Figure 4-57 Skin bypass fracture penetrating through fracturing industry over the last few years, see Fig. 1-13
skin damage (Lancaster et al., 1992, Fisher et al., 2004, and Schein et
al., 2004). Although the shales are porous and contain
4-7.6 Condensate Dropout significant quantities of gas, they have almost no matrix
permeability. Production is through natural fractures.
4-7.6.1 Description of Phenomena Hydraulic fracturing concentrates on trying
to link up the natural fractures. Treatments are
The phenomenon of retrograde condensation was extreme; rates over 100 bpm are common, with
described in detail in Section 2-2 and illustrated proppant concentrations rarely exceeding 2 ppga. The
in Fig. 2-1. Put simply, for wet gas reservoirs, treatments try to link sets of natural fractures with
as the pressure falls, liquids can condense out of the absolute minimum fracture conductivity. Even
the gas mixture. This can result in a reduction of with such low proppant concentrations, the fracture
the relative permeability to gas due to changes in still acts with infinite conductivity because matrix
saturation and the development of immobile fluid permeability is so low. The proppant is carried in slick
banks. In the context of fractured wells, retrograde water, and so the high rates are needed to carry the
condensation can be a problem if the condensate proppant deep into the formation.
builds up in or around the fracture.
4-7.7.2 Coal Bed Methane
4-7.6.2 Mitigating the Effect of Dropout
Hydraulic fracturing is one of several completion
Generally speaking, condensate dropout will occur methods used successfully in coal beds. The choice
wherever the gas pressure and temperature changes of method depends upon the highly variable nature
enough to move the gas into the two-phase envelope of the coal seams. However, fracturing is probably
in Fig. 2-1. If this happens in the formation or the the most widely used method (Nimerick et al.,
fracture, conductivity will be reduced. 1991, and Cramer, 1992).
The process of designing fractures to mitigate The vast majority of the coal bed methane (CBM)
the effects of condensate dropout is based on the fracturing takes place in 9 or 10 major basins in the US,
principle that the best place for condensate to Australia, Canada and China. CBM fracturing also takes
158
place in the UK, the Middle East and Russia. In all of that as much rock volume as possible can be removed.
these places, each particular coal field or basin tends to be The acid system is also usually viscosified to reduce
dominated by a single operating company. wellbore friction, to suspend and transport any fines
Each of these basins has its own particular released by the acid, and to retard the acid reaction rate.
characteristics, in terms of the age and maturity of the Crosslinked acid systems are commonly used to retard
coal, reservoir pressure, fines mobility, water production the reaction rate even further. A combination of retarded
and mechanical characteristics of the coal seams and reaction rate and high pumping rate is intended to get
surrounding rock layers. As a result, each operating live acid as far from the wellbore as possible.
company has developed its own particular method for It is usual practice to pump acid fractures at high
producing the gas, and when this involves fracturing, rate, usually the maximum rate possible based on
each has developed its own method for this as well. equipment, completion and wellhead constraints. This
CBM fracturing remains to this day, very difficult allows the acid to penetrate as far from the wellbore as
to simulate on a computer. Conventional models possible, while minimizing the near-wellbore pinching
cannot be applied to the coal due to the extensive effect described by Roodhart et al. (1993).
cleat systems that exist in the seams, the extremely At the end of the treatment, the fracture is allowed to
plastic nature of the coal and the shear decoupling close and the spent acid flowed back as soon as possible.
that exists between the coal and the over- and under- However, sometimes a closed fracture acidizing treatment
lying rock strata. Without the aid of reliable fracture may be performed (Frederickson, 1986). This consists of
models, engineers have developed a number of rules pumping additional acid down the closed fracture, below
of thumb for CBM fracturing, most of which are fracturing pressure, to widen the etched width of the
specific to a particular basin. fracture and hence artificially increase conductivity.
Generally, acid fracturing is limited to harder rock
4-7.8 Acid Fracturing formations because soft formations will plastically
deform into the etched width created by the acid,
Acid fracturing is an alternative to proppant fracturing reducing and sometimes eliminating the effects of
in carbonate reservoirs (King, 1986, and Kalfayan, the treatment. In addition, acid fracture conductivity
2007). The process relies on acid etching of the tends to be significantly less than that which can be
fracture face, rather than the placing of proppant, to generated by proppant fracturing. These two facts
produce conductivity. Consequently, its intentional mean that fracture acidizing is generally performed
use has been limited to carbonate formations, which on lower-permeability, tight carbonate formations,
are usually > 95% soluble in acid systems. However, it rather than soft, more permeable rocks.
has also been suggested that with modern acid systems, The design of acid fracture treatments is generally
these techniques could be applied to sandstone much less scientific than the process of designing a
reservoirs (di Lullo and Rae, 1996). propped fracture treatment. Available simulators tend
to be very primitive and almost unusable. In addition,
4-7.8.1 Description of Process it is often very difficult to estimate etched width and
length. Consequently, treatments are usually refined on
The basic difference between acid fracturing and an empirical basis over a campaign.
proppant fracturing is that the fracture conductivity
is generated by the acid reacting with the formation 4-7.8.2 Estimating Fracture Conductivity
at the fracture faces, rather than by the placement of
proppant. To start, a non-reactive pad fluid is pumped The most reliable method for estimating acid fracture
into the formation to create a fracture with the desired conductivity was presented by Nierode and Kruk (1973).
half length. This is followed by the acid system, which Other methods, including laboratory testing, have been
is usually as concentrated as is practical (the most developed subsequently, but none have proved to be as
commonly used acid fracturing fluid is 28% HCl), so reliable. A summary of the method follows.
159
Modern Fracturing
The created fracture conductivity, Cf (md-ft), can treatment produces less fracture conductivity than
be calculated using etched fracture width, wetch (in.), proppant fracturing. Selected recommended diversion
rock embedment strength, SRE (psi), and closure techniques are listed below, presented in no particular
pressure, pc, (psi): order. Their use depends upon specific wellbore and
operational circumstances. It is usual practice when
C f = C1eC2 pc and (4-118) fracture acidizing to use separate diversion stages, so
that a treatment sequence could be pad, acid, overflush,
7 2.47
C1 = 1.47 10 wetch .
(4-119) diversion, and then repeating as often as required before
concluding with a final over-displacement.
The average etched fracture width, wetch, can be obtained
by calculating the volume of rock dissolved by the acid 1. Foam (acid and non-acid)
and dividing by the fracture area. The C2 term is related 2. Ball sealers (only in perforated wellbores)
to rock embedment strength, SRE (see also Economides et 3. Benzoic acid flakes and other soluble particulates
al., 1994, pp. 412-413) as follows: 4. Viscoelastic surfactant systems
5. Self-viscosifying acid (as the acid neutralizes, the
pH rises and a crosslinker becomes active, causing
C2 = 0.001(13.9 1.3 ln S RE )
a dramatic increase in viscosity).
for SRE < 20,000 psi or (4-120)
C2 = 0.001(3.8 0.28 ln S RE )
References
for SRE > 20,000 psi. (4-121)
Advanti, S.H., Khattib, H., and Lee, J.K.: Hydraulic
The rock embedment strength has to be determined Fracture Geometry Modeling, Prediction and
experimentally from core or outcrop samples. Extensive Comparisons, paper SPE 13863, 1985.
library data for SRE is available for many of the more Agarwal, R.G.: A New Method to Account for
commonly fractured carbonate formations. Producing Time Effects When Drawdown Type
The obtained acid fracture conductivity can Curves are Used to Analyze Pressure Buildup and
be used in the expression of dimensionless fracture Other Test Data, paper SPE 9289, September
conductivity (Eq. 4-108) instead of the product kfw. 1980.
Therefore, using the penetration ratio Ix, an equivalent Arihara, N., Abbaszadeh, M., Wright, C.A. and Hyodo,
proppant number can also be calculated using Eq. 4- M.: Integration of Fracturing Dynamics and
110; consequently Unified Fracture Design is readily Pressure Transient Analysis for Hydraulic Fracture
applicable for acid fracture design. Evaluation, paper SPE 36551, October 1996.
Auld, B.A.: Acoustic Fields and Waves in Solids, v. I,
4-7.8.3 Use of Diversion Techniques II, Robert E. Krieger Publ. Co., Florida, 856 pp,
1990.
One major difference between propped fracturing and Barree, R.D., Cox, S.A., Gilbert, J.V. and Dobson,
fracture acidizing is the use of diversion techniques. In M.: Closing the Gap: Fracture Half-Length From
propped fracturing, precise control over fluid and slurry Design, Buildup, and Production Analysis, SPE
placement along the wellbore is required to maintain Production and Facilities (2005), 20, 274-285.
control over fracture dimensions and to ensure that a Barree, R.D., Rogers, B.A., and Chu, W.C.: Use of
treatment is not over-displaced. This is not so critical Frac-Pack Pressure Data to Determine Breakdown
for acid fracturing. Consequently, acid fracturing can Conditions and Reservoir Properties, paper SPE
become the stimulation method of choice over long 36423, October 1996.
perforated intervals and multiple intervals, even if the Behrman, L.A. and Nolte, K.G, Perforating
160
Requirements for Fracture Stimulation, paper Cipolla, C.L. and Wright, C.A.: State-of-the-Art
SPE 59480, 1999 in Hydraulic Fracture Diagnostics, paper SPE
Bennion, D.B., Thomas, F.B. and Bietz, R.F.: 64434, October 2000.
Low Permeability Gas Reservoirs: Problems, Cipolla, C.L. and Wright, C.A.: Diagnostic Techniques
Opportunities and Solutions for Drilling, to Understand Hydraulic Fracturing: What? Why?
Completion, Stimulation and Production, paper and How? Hydraulic Fracture Diagnostics, paper
SPE 35577, May 1996. SPE 59735, Gas Technology Symposium, Calgary,
Biot, M.A.: General Solutions of the Equations of Canada, 2000.
Elasticity and Consolidation for a Porous Material, Cipolla, C.L.: Personal Communication, 2006.
Journal of Applied Mechanics (1956), 23, 91-96. Clark, P.E.: Transport of Proppant in Hydraulic
Blangy, J.P.: Integrated Seismic Lithology Fractures, paper SPE 103167, September 2006.
Interpretation: the Petrophysical Basis, Ph.D. Cleary M.P., Johnson, D.E., Kogsbll, H-H., Owens,
dissertation, Stanford U. (1992). K.A., Perry, K.F., de Pater, C.J., Stachel, A., Schmidt,
Brevik, I.: Chalk data presented at workshop on H. and Tambini, M.: Field Implementation of
effective media, Karlsruhe (1995). Proppant Slugs to Avoid Premature Screen-Out
Broek, D.: Elementary Engineering Fracture Mechanics, of Hydraulic Fractures with Adequate Proppant
Kluwer Academic Publishers, 4th Ed., Dordrecht, Concentration, paper SPE 25892, April 1993.
1986. Cleary, M.P., Wright, C.A. and Wright, T.B.:
Budynas, R.G.: Advanced Strength and Applied Stress Experimental and Modelling Evidence for Major
Analysis, 2nd Ed., McGraw-Hill, Singapore, 1999. Changes in Hydraulic Fracturing Design and Field
Card, R.J., Howard, P.R. and Feraud, J-P.: A Novel Procedures, paper SPE 21494, January 1991.
Technology to Control Proppant Backproduction, Coffer, H.F., Bray, B.G., Knutson, C.F. and Rawson,
SPE 31007, SPE Prod. & Fac. (November 1995), D.E.: Effects of Nuclear Explosions on Oil and
No. 4, 110, 271-276. Gas Reservoir Stimulation, JPT, May 1964, 473-
Carter, R.D.: Derivation of the General Equation for 480.
Estimating the Extent of Fracture Area, Appendix Crafton, J.W.: Oil and Gas Well Evaluation using the
1 of Optimum Fluid Characteristics for Fracture Reciprocal Productivity Index Method, paper SPE
Extension, Drilling and Production Practice, 37409, March 1997.
Howard, C.R. and Fast, C.R., API, New York Cramer, D.D.: Limited Entry MHF Part 1 Limited
(1957), 261-269. Entry Extended to Massive Hydraulic Fracturing,
Castagna, J.P., Batzle, M.L., and Eastwood, R.L.: Oil & Gas J., December 14, 1987.
Relationships between Compressional-Wave and Cramer, D.D.: Limited Entry MHF Part 2 Study
Shear-Wave Velocities in Clastic Silicate Rocks, Indicates Guidelines Improve Results, Oil & Gas
Geophysics, 50, 571-581, 1985. J., December 21, 1987.
Castagna, J.P., Batzle, M.L., and Kan, T.K.: Rock Cramer, D.D.: The Unique Aspects of Fracturing
Physics the Link between Rock Properties and Western U.S. Coal Beds, paper SPE 21592, JPT,
AVO Response, in Offset-Dependent Reflectivity pp 1126-1133, October 1992.
Theory and Practice of AVO Analysis, ed. J.P. Cramer, D.D.: Evaluating Well Performance and
Castagna and M.M. Backus, SEG Tulsa, 1993. Completion Effectiveness in Hydraulically
Cinco-Ley, H. and Samaniego-V.F.: Transient Pressure Fractured Low-Permeability Gas Wells, SPE
Analysis for Fractured Wells, Journal of Petroleum 84214, October 2003.
Technology, pp. 1749-1766, 1981. Daal, J.A., and Economides, M.J.: Optimization
Cipolla, C.L. and Mayerhofer, M.J.: Understanding of hydraulically fractured wells in irregularly
Fracture Performance by Integrating Well Testing shaped drainage areas, Proc. SPE International
and Fracture Modelling, SPE 74632, SPE Prod. & Symposium on Formation Damage Control, 2006,
Fac., November 2001. pp. 71-82.
161
Modern Fracturing
Daneshy, A.A.: On the Design of Vertical Hydraulic Flowback, paper SPE 20708, September 1990.
Fractures, JPT, 83-93, January 1973. Fisher, M.K., Heinze, J.R., Harris, C.D., Davidson,
Deily, F.H., and Owens, T.C.: Stress Around a B.M., Wright, C.A. and Dunn, K.P.: Optimizing
Wellbore, paper SPE 2557, October 1969. Horizontal Completion Techniques in the Barnett
Demarchos, A.S., Chomatas, A.S., Economides, M.J., Shale Using Microseismic Fracture Mapping,
Mach, J.M., Wolcott, D.S: Pushing the Limits paper SPE 90051, September 2004.
in Hydraulic Fracture Design, (2004) Proc. SPE Fredrickson, S.E.: Stimulating Carbonate Formations
International Symposium on Formation Damage Using a Closed Fracture Acidizing Technique,
Control, 239-248. paper SPE 14654, April 1996.
Demarchos, A.S., Economides, M.J., Diyashev, I., Gardner, G.H.F., Gardner, L.W., and Gregory, A.R.:
Svaykin, V.A.: Analysis of the performance of Formation Velocity and Density the Diagnostic
hydraulic fracturing treatments and quantum Basis for Stratigraphic Traps, Geophysics (1974),
design improvements, (2005) Proc. SPE European 39, 770-780.
Formation Damage Conference, 271-277. Geertsma, J.: Velocity-Log Interpretation: the Effect
di Lullo, G. and Rae, P.: A New Acid for True of Rock Bulk Compressibility, SPEJ, December,
Stimulation of Sandstone Reservoirs, paper SPE 1961.
37015, October 1996. Geerstma, J., and de Klerk, F.A.: Rapid Method of
Dietrich, J.K.: JD as a performance indicator for Predicting Width and Extent of Hydaulically
hydraulically fractured wells, (2005) Proc. SPE Induced Fractures, JPT (December 1969), 1571-
Western Regional Meeting, 201-216. 81.
Diyashev, I.R., and Economides, M.J.: A general Grebe, J.J.: Tools and Aims of Research, Chem. &
approach to well evaluation, (2005) Proc. SPE Eng. News, December 10, 1943.
- European Formation Damage Conference, 279- Grebe, J.J. and Stosser S.M.: Increasing Crude
289. Production 20,000,000 Barrels from Established
Eberhart-Phillips, D.M.: Investigation of Crustal Fields, World Petroleum (August 1935), 6, 8, 473.
Structure and Active Tectonic Processes in the Coast Greenberg, M.L., and Castagna, J.P.: Shear-Wave
Ranges, Central California, Ph.D. dissertation, Velocity Estimation in Porous Rocks: Theoretical
Stanford U., 1989. Formulation, Preliminary Verification and
Economides, M.J., and Nolte, K.G.: Reservoir Applications, Geophysical Prospecting (1992), 40,
Stimulation, Schlumberger Educational Services, 195-209.
1st Edition 1987, 2nd Edition, 1994, 3rd Edition, GRI Staged Field Experiment No. 1, Gas Research
1998. Institute, Chicago, 1988.
Economides, M.J., Demarchos, A.S., Mach, J.M., GRI Staged Field Experiment No. 2, Gas Research
Rueda, J., and Wolcott, D.S.: Pushing the limits Institute, Chicago, 1989.
of hydraulic fracturing in Russia (2004) Proc. GRI Staged Field Experiment No. 3, Gas Research
SPE Annual Technical Conference and Exhibition, Institute, Chicago, 1990.
2329-2333. Griffith, A.A.: The phenomena of rupture and flow in
Economides, M.J., Oligney, R.E. and Valk, P.P.: solids, Phil. Trans. Roy. Soc. of London (1921), A
Unified Fracture Design, Orsa Press, Alvin, Texas, 221, 163 167.
2002. Han, D.-H., Nur, A., and Morgan, D.: Effects of
El-Banbi, A.H. and Wattenbarger, R.A.: Analysis of Porosity and Clay Content on Wave Velocities in
Linear Flow in Gas Well Production, paper SPE Sandstones, Geophysics (1986), 51, 2093-2107.
39972, March 1998. Handin J., Hager, R.V. Jr, Friedman, M., and Feather,
Ely, J.W., Arnold III, W.T., and Holditch, S.A.: New J.N.: Experimental Deformation of Sedimentary
Techniques and Quality Control Find Success in Rocks Under Confining Pressure: Pore Pressure
Enhancing Productivity and Minimizing Proppant Tests, Bul. Amer. Asso. Pet. Geol. (1963), 47, 717-
162
755. of Oil Reservoirs, Rev. Inst. Franais du Ptrole, 44-

Helmy, M.W. and Wattenbarger, R.A.: New Shape 108, 1967 (in French).
Factors for Wells Produced at Constant Pressure, Leshchyshyn, T.H., Beadall, K.K. Meier, P.E., Hagel,
paper SPE 39970, March 1998. M.W., and Meyer, B.R.: Using Empirically
Holditch, S.A.: Tight Gas Sands, JPT (2006), 58, Developed Rock Tables to Predict and History
86-93. Match Fracture Stimulations, paper SPE 86989,
Howard, G.C., and Fast, C.R.: Optimum Fluid March 2004.
Characteristics for Fracture Extension, Drill. & Lopez-Hernandez, H.D., Valko, P.P., and Pham, T.T.:
Prod. Prac., API, 1957, 261. Optimum fracture treatment design minimizes
Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, the impact of non-Darcy flow effects, Proc. SPE
Monograph Series Vol. 2, SPE, Dallas, Texas, USA Annual Technical Conference and Exhibition
(1970). (2004), 1725-1740.
Jeffrey, R.G.: The Combined Effect of Fluid Lag Manrique, J.F. and Poe, B.D.: Evaluation and
and Fracture Toughness on Hydraulic Fracture Optimization of Low-Conductivity Fractures,
Propagation, paper SPE 18957, March 1989. paper SPE 106317, January 2007.
Johnson, E. and Cleary, M.P.: Implications of Recent Martin, A.N.: Crack Tip Plasticity: A Different
Laboratory Experimental Results for Hydraulic Approach to Modeling Fracture Propagation in
Fractures, paper SPE 21846, April 1991. Soft Formations, paper SPE 63171 (revised),
Kachanov, L.M.: Advances in Creep Design, Applied October 2000.
Science Publishers, London, 1971. Martin, A.N.: Hydraulic Fracturing Manual, BJ Services,
Kalfayan, L.J.: Fracture Acidizing: History, Present 2005.
State, and Future, paper SPE 106371, January Martins, J.P. and Harper, T.R.: Mini-Frac Pressure
2007. Decline Analysis for Fractures Evolving From Long
Khristianovitch, S.A. and Zheltov, Y.P.: Formation Perforated Intervals and Unaffected by Confining
of Vertical Fractures by Means of Highly Viscous Strata, paper SPE 13869, March 1985
Liquid, paper presented at the 4th World Petroleum Mayerhofer, M, Ehlig-Economides, C.A and
Congress, II, 579-586, 1955. Economides, M.J.: Pressure Transient Analysis
King, G.E.: Acidizing Concepts - Matrix vs. Fracture of Fracture Calibration Tests, paper SPE 26527,
Acidizing, paper SPE 15279, JPT, pp 507-508, 1993
May 1986. Mayerhofer M., Stutz, L. Davis, E. and Wolhart,
Klimentos, T., Harouaka, A., Mtawaa, B., and Saner, S., S.: Optimizing Fracture Stimulation Using
Experimental Determination of the Biot Elastic Treatment-Well Tiltmeters and Integrated Fracture
Constant: Applications in Formation Evaluation Modeling, SPE Production and Operations, pp
(Sonic Porosity, Rock Strength, Earth Stresses and 222-229, 2006.
Sanding Predictions), paper SPE 30593, Oct, McGowen, J.M., Barree, R.D. and Conway, M.W.:
1995. Incorporating Crossflow and Spurt-Loss Effects
Lacy, L.L.: Dynamic Rock Mechanics Testing for in Filtration Modeling Within a Fully 3D Fracture
Optimized Fracture Designs, paper SPE 38716, Growth Simulator, paper SPE 56597, October
October 1997. 1999.
Lancaster, D.E., McKetta, S.F., Guidry, F.K. and McGuire, W.J. and Sikora, V.J.: The Effect of Vertical
Jochen, J.E.: Reservoir Evaluation, Completion Fractures on Well Productivity, Trans., AIME
Techniques, and Recent Results From Barnett Shale (1960), 219, 401-403.
Development in the Fort Worth Basin, paper SPE Meyer, B.R., and Jacot, R.H.: Pseudosteady-state
24884, October 1992. analysis of finite-conductivity vertical fractures,
Le Tirant, P., and Dupuy, M.: Fracture Dimensions Proc. SPE Annual Technical Conference and
Obtained During Hydraulic Fracturing Treatments Exhibition (2005), 2007-2030.
163
Modern Fracturing
Morales, R.H.: Fracturing of High-Permeability Damage Control Symposium (2004).

Formations: Mechanical Properties Correlations, Palmer, I.D. and Veatch, R.W.: Abnormally High
paper SPE 26561, October 1993. Fracturing Pressures in Step Rate Tests, paper SPE
Nguyen, P.D., Weaver, J.D., Dewprashad, B.T., Parker, 16902, SPE Prod. Eng., August 1990.
M.A. and Terracina, J.M.: Enhancing Fracture Perkins, T.K. and Kern, L.R.: Widths of Hydraulic
Conductivity Through Surface Modification of Fractures, JPT, September 1961, 937-49; Trans.
Proppant, paper SPE 39428, February 1998. AIME 222, September 1961.
Nierode, D.E. and Kruk, K.F.: An Evaluation of Acid Pickett, G.R.: Acoustic Character Logs and Their
Fluid Loss, Additives, Retarded Acids and Acidized Applications in Formation Evaluation, JPT, 15,
Fracture Conductivity, paper SPE 4549, October 650-667, 1963.
1973. Plona, T.I. and Cook, J.M.: Effects of Stress Cycles on
Nikurova, L.F., Surtaev, V.N. and Yamilov, R.R.: Static and Dynamic Youngs Moduli in Castlegate
Enhancing Well Productivity After Hydraulic Sandstone, Rock Mechanics, Daemen & Schultz,
Fracturing in the Priobskoe Oilfield, paper SPE ed., Balkema, Rotterdam. ISBN 90 5410 552 6.,
102194, 2006. 1995.
Nimerick, K.H., Hinkel, J.J., England, K.W., Norton, Poe, B.D. and Manrique, J.F.: Production Performance
J.L. and Roy, M.: Design and Evaluation of Design Criteria for Hydraulic Fractures, paper SPE
Stimulation and Workover Treatments in Coal 101722, Proc. SPE Annual Technical Conference
Seam Reservoirs, paper SPE 23455, October and Exhibition (2005), 3687-3700.
1991. Poe, B.D.: Design Criteria for Improved Performance
Nolte, K.G.: Determination of Fracture Parameters of Fractured Wells, paper SPE 104015, Proc. 1st
from Fracturing Pressure Decline, paper SPE Int Oil Conference and Exhibition in Mexico,
8341, 1979. 2006.
Nolte, K.G., and Smith, M.B.: Interpretation of Pongratz, R., von Gijtenbeek, K., Kontarev, R. and
Fracturing Pressures, paper SPE 8297, JPT, 1767- McDaniel, B.W.: Perforating for Fracturing Best
75, September 1981. Practices and Case Histories, paper SPE 105064,
Nolte, K.G.: A General Analysis of Fracture Pressure January 2007.
Decline With Application to Three Models, SPE Prats, M.: Effect of Vertical Fractures on Reservoir
12941, SPEFE, 571-583, 1986(a) Behavior Incompressible Fluid Case, SPE
Nolte, K.G.: Determination of Proppant and Fluid Journal, 1, pp 105-118, 1961.
Schedules From Fracturing Pressure Decline, Rae, P, Martin, A.N. and Sinanan, B.: Skin Bypass
SPEPE, 255-265, July 1986(b). Fracs Proof the Size is Not Important, paper
Nordgren, R.P.: Propagation of a Vertical Hydraulic SPE 56473, October 1999.
Fracture, paper SPE 3009, SPEJ, 306-14; Trans. Raymer, L.L., Hunt, E.R., Gardner, J.S.: An Improved
AIME 253, August 1972. Sonic Transit Time-to-Porosity Transform,
Obert, L., Windes, S.L., Duvall, W.I.: Standardised SPWLA, 21st Ann, Log. Symp., July 1980.
Tests for Determining the Physical Properties of Reynolds, O.: On the Dilatancy of Media Composed
Mine Rock, U.S. Bur. Mines Rept. Invest., 3891, of Rigid Particles in Contact. With Experimental
1946 Illustrations, The London, Edinburgh & Dublin
Oberwinkler, C., and Economides, M.J.: The Phil. Mag. & J. of Sci., LVII, December 1885, 469-
Definitive Identification of Candidate Wells 481.
for Refracturing, Proc. SPE Annual Technical Rickards, A.R., Brannon, H.D., Wood, W.D.
Conference and Exhibition (2003), 1385-1401. and Stephenson, C.J.: High-Strength, Ultra-
Oberwinkler, C., Ruthammer, G., Zangl, G., and Lightweight Proppant Lends New Dimensions
Economides, M.J.: New Tools for Fracture Design to Hydraulic Fracturing Applications, paper SPE
Optimization, Proc. SPE International Formation 84308, October 2003.
164
Rickards, A.R., Lacy, L., Brannon, H.D., Stephenson, Shah, S.N. and Lee, Y.N.: Friction Pressures of
C.J and Bilden, D.: Need Stress Relief? A New Proppant-Laden Hydraulic Fracturing Fluids,
Approach to Reducing Stress Cycling Induced paper SPE 13836, SPEPE, 437-445, November
Proppant Pack Failure, paper SPE 49247, 1986.
September 1998. Smith, J.E.: Design of Hydraulic Fracture Treatments,
Ringstad, C., Lofthus, E., Snsteb, E.F., Fjaer, E., paper SPE 1286, Dallas, Texas, 1965.
Zausa, F., and Fuh, G.: Prediction of Rock Sneddon, I.N.: The Distribution of Stress in the
Parameters from Micro-Indentation Measurements: Neighbourhood of a Crack in an Elastic Solid,
The Effect of Sample Size, paper SPE/ISRM Proc. Royal Society of London, 187, 229, 1946.
47313, July, 1998. Timonov, A.V., Zagurenko, A.G., Hasanov, M.M.,
Roegiers, J-C.: Rock Mechanics, Ch. 3 of Reservoir Pasynkov, A.G. and Afanasiev, I.S.: System
Stimulation, ed. Economides, M.J. and Nolte, Approach to Hydraulic Fracturing Optimization
K.G., Schlumberger Educational Services, 1987. in Rosneft Oilfields, paper SPE 104355, 2006.
Romero, D.J., Valk, P.P. and Economides, M.J.: Torrey, P.D.: Progress in Squeeze Cementing
Optimization of The Productivity Index and Applications and Technique, Oil Weekly, July 29,
the Fracture Geometry of a Stimulated Well with 1940.
Fracture Face and Choke Skins, SPE Production Urmos, J., and Wilkens, R.H.: In-situ Velocities in
and Facilities, vol 17 pp 57-64, 2003. Pelagic Carbonates: New Insights from Ocean
Roodhart, L.P., Kamphuis, H. and Davies, D.R.: Drilling Program Leg 130, J. Geophys. Res., 98,
Improved Acid Fracturing Treatment Designs 7903-7920, 1993.
Based on In-Situ Temperature Calculations, paper Valk, P.P. and Amini, S.: The Method of Distributed
SPE 26185, June 1993. Volumetric Sources for Calculating the Transient
Rozo, R., Paez, J., Mendoza, A., Milne, A. and Soler, and Pseudosteady-State Productivity of Complex
D.: Combining Acid- and Hydraulic-Fracturing Well-Fracture Configurations, paper SPE 106279,
Technologies Is the Key to Successfully Stimulating January, 2007.
the Orito Formation, paper SPE 104610, January, Valk, P. and Economides, M.J.: A Continuum-
2007. Damage-Mechanics Model of Hydraulic
Rueda, J.I., Mach, J., and Wolcott, D.: Pushing Fracturing, paper SPE 25304, JPT, 198-205,
fracturing limits to maximize producibjlity March 1993. See also discussions on this paper
in turbidite formations in Russia, Proc. SPE contained in papers SPE 26282, 26303, 27734
International Petroleum Conference in Mexico and 27738.
(2004), 371-382. Valk, P.P. and Economides, M.J.: Heavy Crude
Rueda, J.I., Voronkov, A., and Mach, J.: Optimum Production from Shallow Formations: Long
fracture design under transient and pseudosteady Horizontal Wells Versus Horizontal Fractures,
conditions using constant fracture volume paper SPE 50421, November 1998.
concept, (2005) 67th European Association of van Dam, D.B, Papanastasiou, P. and de Pater, C.J.:
Geoscientists and Engineers, EAGE Conference Impact of Plasticity on Hydraulic Fracture
and Exhibition, incorporating SPE EUROPE2005 Propagation and Closure, paper SPE 63172,
- Extended Abstracts, pp. 1795-1805. October 2000.
Santarelli, F.J., Marsala, A.F., Brignoli, M., Rossi, van Everdingen, A.F. and Hurst, W.: The Application
E., and Bona, N.: Formation Evaluation from of the Laplace Transformation to Flow Problems in
Logging on Cuttings, paper SPE 36851, 1996. Reservoirs, 1949, Trans., AIME, 186, 305-324.
Schein, G.W., Carr, P.D., Canan, P.A. and Richey, van Everdingen, A.F.: The Skin Effect and Its
R.: Ultra Lightweight Proppants: Their Use and Impediment to Fluid Flow into a Wellbore, 1953,
Application in the Barnett Shale, paper SPE Trans., AIME, 198, 171-176.
90838, October 2004. van Terzaghi, K.: Die Berechnung der
165
Modern Fracturing
Durchlassigkeitsziffer des Tones aus dem Verlauf Formations in Water Flood Operations Part II,
der Hydrodynamischen Spannungserscheinungrn, Oil Weekly, March 19, 1945.
Sber. Akad. Wiss, Vienna, 1923, 123, p105 (in
German).
Vincent, M.C., Pearson, C.M. and Kullman, J.: Non-
Darcy and Multiphase Flow in Fractures: Case
Studies Illustrate the Dramatic Effect on Well
Productivity, paper SPE 54630, May 1999.
Weijers, L., Wright, C., Mayerhofer, M. and Cipolla C.:
Developing Calibrated Fracture Growth Models
for Various Formations and Regions Across the
United States, paper SPE 96080, October 2005.
Williams, B.B., Gidley, J.L. and Schechter, R.S.:
Acidizing Fundamentals, SPE Richardson, Texas,
1979.
Wright, C.A., Weijers, L., and Minner, W.A.: Advanced
Stimulation Technology Deployment Program, report
GRI-09/0075, Gas Research Institute, March
1996.
Wright, T.B., Aud, W.W., Cipolla, C., Perry, K.F., and
Cleary, M.P.: Identification and Comparison
of True Net Fracturing Pressures Generated by
Pumping Fluids with Different Rheology into the
Same Formations, paper SPE 26153, June 1993.
Wyllie, M.R.J., Gregory, A.R., and Gardner, L.W.:
Elastic Wave Velocities in Heterogeneous and
Porous Media, Geophysics, 21, 41-70, 1956.
Wyllie, M.R.J., Gregory, A.R., and Gardner, G.H.F.:
An Experimental Investigation of Factors
Affecting Elastic Wave Velocities in Porous Media,
Geophysics, 23, 459-493, 1958.
Wyllie, M.R.J., Gregory, A.R., and Gardner, G.H.F.,
Studies of Elastic Wave Attenuation in Porous
Media, Geophysics, 27, 569-589, 1963.
Yale, D.P., and Jamieson Jr., W.H.: Static and Dynamic
Rock Mechanical Properties in the Hugoton and
Panoma Fields, Kansas, paper SPE 27939, May,
1994.
Yew, C.H. and Liu, G.F.: Fracture Tip and Critical
Stress Intensity Factor of a Hydraulically Induced
Fracture, paper SPE 22875, SPE Prod. & Fac., no
3, 8, 171-177, August 1993.
Yuster, S.T. and Calhoun, J.C., Jr.: Pressure Parting of
Formations in Water Flood Operations Part I,
Oil Weekly, March 12, 1945.
Yuster, S.T. and Calhoun, J.C., Jr.: Pressure Parting of
166
David Ross is an associate director of InTuition Energy Associates Ltd., an organization that provides oil
& gas training in technology developments, challenges and applications. He has a BS degree in petroleum
engineering from the University of Texas. Immediately after graduating, he began his career with BJ Services
Company. During 24 years at BJ Services, Ross gained a broad range of experience in the engineering
disciplines of hydraulic fracturing, well cementing, acidizing, sand control, completion fluids and coiled
tubing services. He worked in a wide range of geographical environments including Alaska, Europe / North
Sea, South America, Southeast Asia, Russia, the Middle East, Canada as well as South & West Texas. Ross
has been active in the SPE and has authored or co-authored numerous technical papers about cement slurry
designs, completion techniques, sandstone acidizing systems, personnel training and QHSE management
systems. He was also a principal co-author of the Formation Damage and Clear Brine Completion Fluids
technical manuals. He has also been named a 2007-2008 SPE Distinguished Lecturer.
George King has worked 36 years with the BP organization since joining Amoco in 1971. He has been
involved with nearly all phases of oil and gas well completions, stimulations, workover processes and
operations applications and has held the title of distinguished advisor since 1991. Current activities include
mentoring, teaching, field reviews, innovation issues, problem well analysis, intervention designs and general
consulting in the area of production engineering. Degrees include a BS inchemistry from Oklahoma State
University, and a BS in chemical engineering and MS in petroleum engineering from the University of Tulsa.
Technical accomplishments include two patents, 52 technical papers, a book on completions and workovers,
the SPE Production and Operations Award in 2004, the Amoco Chairmans Award for technology from
Amoco in 1997, SPE Distinguished Lecturer in 1985-86, lecturer in the SPE Short Course series in 1999,
and Technical Chairman for the SPE Annual meeting in 1992.
Chapter 5 the final well. If well design or well locations are not
changeable, the value improvements possible from the
Well Completions learning curve are largely lost.
David Ross, InTuition Energy Associates and In many cases after initial drilling and completion,
George King, BP reservoir barriers and flow paths are finally recognized
and re-drilling, re-completion or unplanned stimulations
are needed to fully and efficiently exploit the reservoir.
5-1 Wellbore Construction For a good initial completion, it is important to collect
and accurately assess the data at the earliest possible
Optimum gas well completion design combines time, and to build flexibility into the design for later
continuing reservoir engineering assessment and operations. Batch completions, where logging or
knowledge of production rate and characteristics to production data are not fully considered until the end of
achieve the best reservoir access from the most efficient the operation, depend on certainties and understandings
wellbore design. The design often must allow access that are not usually available during initial drilling and
to reservoir sections from compact and often remote completion. Batch completions may save drilling cost,
entrance points such as platforms and pad locations. but they can sharply reduce recovery in the field or
Multiple flow regimes may exist in a single wellbore. sharply increase re-completion cost later.
The efficiency of well operations over the economic
life of the well is affected by the changes in fluid types 5-1.2 Fitting Well Design to the Reservoir Potential
and rates, reservoir stresses, pressure support, the
overall well design and operational decisions. Gas well completion design must fit the needs of the
producing well throughout its lifetime. The life of
5-1.1 Effects of Uncertainty a gas well may range from a few years in a deepwater,
in Reservoir Description high-permeability formation, to 40 to 60 years in an
onshore tight gas reservoir. The physical changes that
A good reservoir description can lead to better well a gas well undergoes during its life are remarkable and
completion, one that can be fitted to the reservoir and often challenge the design limits. A 20,000 ft (6097 m)
the changing well characteristics. Uncertainty in data, HPHT well in a stacked pay in south Louisiana (USA),
both initial and current, may result in serious flaws for example, went from nearly 16,000 psig (1088 barg)
in well design and sharp reduction in hydrocarbon BHP, at rates of 30 MMscf/D (0.87 MMsm3/d) and very
recovery. Often design-sensitive reservoir factors minor water, to BHP of 4000 psig (272 barg), rate of
such as permeability, porosity, saturations, pressure, < 1 MMscf/D (< 0.028 MMsm3/D), increased water and
barriers, drive systems and flow paths are only fully sand production, all in a two-year period.
known after most of the wells in the field have been Production of fluids from any reservoir to the
drilled, completed and produced. Thus, often for long surface is a system approach of managing pressure drops.
periods there is a requirement for the design to have Increasing the drawdown pressure from the reservoir to
flexibility for re-completion to achieve maximum the sales line is usually the goal. Optimizing this pressure
recovery. Unless the engineer regularly reviews the drop starts with the completion design and continues
available data and updates the flowing model, recovery every day of production through the economic life of the
opportunity will be wasted. A well or field depletion well. The flowing system is dynamic and requires constant
plan, supported by surveillance and modeling optimization. Tightly locking a completion design in
exercises, is necessary for achieving the best well place minimizes the options for optimizing flow and
completion. Maximizing the economics of a project production enhancement by moving proved undeveloped
requires that early cost-saving learning be recognized reserves to proved developed reserves, and ultimately
immediately and the well design be changeable during reduces recovery. Leaving flexibility in the design allows
the period between field discovery and the drilling of gas flow to be optimized from the turbulence restrictions
169
Modern Fracturing
of initial production from a high-pressure, high-rate gas 3. Inflow Performance: The flowing connection of
well to the removal of hydrostatic head and compression a reservoir into the well is the source of many of
needed in a low-pressure, low-rate gas well. the problems with high skin values and resultant
production loss from tortuosity, macro- and
5-1.3 Well Design micro-barriers, and flow-path connection.
(See Section 2-3 for inflow performance
The design of a gas well completion can be divided into calculations.) The effect of the damage on the
four major parts: average permeability of a zone is a function of
1. Drilling the Pay. The completion begins when the level of damage and zone thickness:
drill bit penetrates the pay. Every action - pressure
differential; mud, fluid, or cement loss; time of
contact; and fluid loss control operation adds
potential for near-well damage. A good point to (5-1)
remember is that invariably everything done during
drilling can result in some formation damage. where kavg is the average permeability through the
2. Mechanical Design. A well is a pressure vessel with reservoir including effect of the damaged layer
often a hundred or more threaded connections, (md), ks is the permeability through the damaged
a remotely placed sealing system (cement), and or stimulated layer (md), k is the permeability
a dependence on mechanical devices and fluids in the undamaged formation, re is the reservoir
to form barriers for control of pressure and fluid drainage radius of the well (ft), rs is the radius of
movement. Putting it simply, a well is designed the damaged or stimulated zone (ft) and rw is the
from the bottom to the top and from the inside out. radius of the wellbore (ft).
The required size of the bottom-most completion To match the flow path to the well, the
assembly (usually tubing size or pump size) sets the completion designer must have accurate knowledge
required casing size, the drilled-hole size and the of how fluids can move in the reservoir. Brute force
entire upper casing string/bit selection. If the casing corrections of perforating or fracturing alone will
design cannot be run as planned, well production not overcome a poor selection of well placement,
may be affected because smaller pipe will force a trajectory, deviation, sand control or the decision
curtailment of flow through smaller bottomhole to case and cement rather than opting for open-
equipment. Well design is a sequence of casing and hole completions. Flow channels through natural
tubing strings and mechanical isolation devices fractures and larger pores exceed the matrix
that provide usually two or more barriers when permeability, usually by two orders of magnitude,
placed on production. Tubular and joint thread dominating the flow paths that the fluids use
selections are based on burst, collapse, tension, to move through the reservoir. Connecting the
completion bending, and other loads and forces. wellbore with these flow paths in a manner that
These conditions change from the initial tubular allows the absolute maximum of flow with
running through the complex cycles of cementing, options for zone conformance is the heart of
stimulation and production; and maximum and inflow completion work.
minimum critical flow velocity limits. Other 4. Outflow Performance: Flow of liquids from the
issues include corrosion effects such as CO2, H2S, wellbore to the surface focuses on managing
increasing high chlorides, and water flow that pressure drops in the outflow system. The main
changes over the course of the wells life. Numerous points of interest are sand control and perforation
papers have been written on the treatment of these restrictions, management of hydrostatic head of
forces and factors in field applications (Pattillo both liquids and gas, balancing tubular friction and
and Huong, 1982; Li et al., 2003; and Pattillo deliquification needs, backpressures from chokes,
and Kristiansen, 2002.) gathering lines, and separators, and pipeline
170
Chapter 5 Well Completions
entrance pressures. Specifically for gas wells, the The ability to maintain the integrity of the pressure
major impact areas are turbulent or non-Darcy vessel status of the well is a requirement for the
effects around the perforations and sand control ability and even the legal right to continue operating
completions, removing liquid heads (especially in the well. Well integrity is one of the most difficult
low-productivity gas wells), and surface pressure objectives to achieve because the operating conditions
constraints around entering the pipeline. Each effect change and mechanical properties of equipment can
must be weighted against cost and recovery losses. degrade with time and production.
A complete nodal analysis of the entire system, Basic well integrity starts with a proper mechanical
updated regularly with production information design and a high-quality cement job. Those two elements
and improvements in reservoir understanding, is are the primary basis for success or failure in well-
critical (see Section 5-4.5). containment longevity. The best practices in well design
Wellbore construction, evaluation and for integrity include: bringing the cement top to the
optimization continue through the entire surface on the most outer casing; completely cementing
life of a gas well. off both fresh water and corrosive water zones; managing
thermal pressure rise in trapped annuli; using special
5-1.4 Other Well Equipment designs around salt, faults or other high-stress areas; and
selecting well construction materials that will meet the
In modern gas wells, the mechanical design is frequently challenges over the life of the well.
augmented by downhole temperature and pressure As wells age, the challenges of maintaining well
gauges, remotely actuated valves and other sensors and integrity often sharply increase. In gas wells, entry of
control points that allow better diagnostics and constant water, corrosion by acid gases, erosion, subsidence,
flow (Holstein and Berger, 1997; Oberwinkler and and potential for leaks increase with age. Each
Standner, 2005; Sinha and Al-Qattan, 2004; Baksh, of these problems can cause loss of isolation in a
2005; and Dolle et al., 2005). well. Progressive erosion, corrosion and subsidence
Some equipment, like the downhole gauges, are are serious problems, but their consequences are
particularly suited to monitoring and even improving measurable with established methods and regular
fracturing treatments. By removing friction and storage monitoring. Leaks, however, have broad consequences
effects from the measurements during a fracturing for the operator, ranging from pollution to trapped
job, more accurate analysis of pressure build-ups, annular pressure increases that can lead to severe leaks,
fall-offs and fracture closures is possible. The right explosions and loss of the well. In trapped annular
selection of sensors can also contribute information spaces, low-pressure leaks combined with large
on clean-up and production flow. Completion changes in flowing temperature can create pockets of
design, at minimum should allow for incorporation high pressure that can rupture or collapse even heavy-
of pressure and temperature sensors. Multiple zones, wall, new pipe (Fig. 5-1).
depending on thickness and fracture design intent,
Gas liquid closed Gas liquid closed
may use multiple sets of sensors. In wells with zones system (cold) system (hot)
that are commingled, downhole gauges help predict Gas@115F
Gas@45F
individual layer flow contribution. Increase in
Liquid Volume
How high is
this pressure ?
5-1.5 Well Integrity

Liquid@115F
Liquid@45F
One of the most important considerations for a
producing well is a risk assessment of the well design and
the operations plans regarding well integrity (Pattillo et
al., 1996; Pattillo and Kristiansen, 2002; Vargo et al., Figure 5-1 Production flow heating of a trapped annular
2002; and Tewari et al., 2006). volume (Courtesy of Dave Andrews, BP)
171
Modern Fracturing
A gas cap over an annular space can offset the hydrocarbon-producing well. This is to ensure the
effects of expansion caused by fluid heating if the gas optimum productivity from producing zones or, in
volume is of sufficient quantity (usually 7 to 12% the case of an injector well, the optimum injectivity
of total volume) and if the gas is at a low pressure into injection zones. The loss of zonal isolation can
at the start of the temperature increase (see Fig. 5- lead to cross-flow of fluids from one formation to
2). This technique is commonly used in subsea wells another, or the production of undesirable fluids such
where the annulus cannot be easily drawn down as water that can severely impact the economics
during production startup. Whether pressure will of a well. Furthermore, ineffective zonal isolation
rise in a trapped annulus with flowing temperature can severely compromise subsequent stimulation
increase will depend on whether the expansion of treatments, resulting in insufficient stimulation of
the liquid can be offset by compression of the gas the target producing intervals and ultimately reduced
volume. If the gas volume is too small or is initially well productivity. This is especially important for
pressurized, the pressure created when liquid gas wells because an overwhelming percentage will
expands can be much greater. require an effective hydraulic fracture stimulation
treatment to maximize the production and reservoir
10,000 drainage potential of the well.
Frequently, gas migration through and along
Full Temperature, psi
8000
When Heated to
Pressure Rise
6000
cementing sheaths can lead to sustained annular casing
pressure at the wellhead, a huge problem that continues
4000
to plague the oil and gas industry on a worldwide basis.
2000 Sustained casing pressure due to gas migration is causing
0 operators increasing concern and significant remedial
0 50 100 workover costs in countries where environmental
Percent Gas in Annulus regulations are becoming stricter and venting of gas
to the atmosphere has been banned. Furthermore,
Figure 5-2 Annular pressure in a trapped annulus contaminating fresh water aquifers with hydrocarbon
generated as flowing temperature increases. Pressure
Rise in a Trapped 2200 ft Annulus when heated from 45 fluids resulting from the failure to obtain or loss of
F (7 C) to 135F (57 C). Initial press is atmospheric. appropriate zonal isolation has created devastating
liabilities for operators around the world.
A gas well integrity plan should incorporate both Therefore, to optimize well productivity and
understanding of operations and standards of operating minimize potential liabilities, there should be no
that will protect the isolation potential of the flowing corners cut when setting out to achieve the ultimate
system from the reservoir through the sales point. goal of effective zonal isolation for the life of the well.
Inspection, standards of design and operation, integrity In order to achieve this goal, several primary objectives
surveillance programs and accountability are critical must be accomplished:
elements to safe, responsible operations. Design a cement slurry that has the
proper characteristics to:
5-2 Gas Well Cementing Place the slurry efficiently in the annulus
Prevent gas invasion during and after the
5-2.1 General Objectives for Gas hydration process
Well Cementing Operations Provide the required mechanical parameters
when set, to withstand the induced stresses
The primary cementing of any well is the most that will occur on the cement sheath
important operation during its life. Achieving a throughout the life of the well.
hydraulic seal between the productive formation Design a cement pumping program that
and casing is of fundamental importance for any will facilitate:
172
Adequate hole and mud conditioning

prior to cementing k = 5.4 x1010 w2 ,
(5-3)
Loss control prior to and during
the cement placement where the slot width, w in inches and k is in md.
Effective mud displacement using washes, Based on these equations it is interesting
spacers centralizers and pipe movement (if to note that a micro-annulus or stress fracture
possible) during the cementing operation with a width as small as 610-5 in. would exhibit
Hydrostatic control at all times during 200 md of permeability.
the pumping operation The significance of gas flow through or along a
Execute the cementing operation as per the permeable cement sheath can be shown in Fig. 5-
slurry design and pumping program. 3 which was generated by using various values of
Equivalent Cement Permeability (k*) and the data
5-2.2 Gas Well Zonal Isolation in Table 5-1 to calculate the corresponding values
of qcem expressed in Mscf/D.
Although obtaining effective zonal isolation is
critical for any well, it is especially critical for gas Table 5-1 Example Data for Gas Flowing Through or
wells. This is due to the relative ease for gas to Along a Cement Sheath
pass through permeable cement or micro-annuli rw = 0.354 ft (8--in. OD)
in comparison to liquids. For example as per rcas = 0.292 ft (7-in. OD)
Economides (1990), the ability for gas to flow pi = 3000 psi
through a cement sheath is: pwf = 1000 psi
m= 0.025 cp
k * (rw 2 rcas 2 )( pi 2 pwf 2 )

qcem = (5-2) Z= 0.95
1424ZT (L) T= 640 R
where qcem is the flow rate through or along a L = 20 ft
cement sheath (Mscf/D), k* is the equivalent cement
permeability (md), rw is the wellbore radius (ft), rcas is
the casing radius (ft), is the gas viscosity at reservoir 1.E+02
conditions (cp), Z is the gas deviation factor, T is
1.E+01
the reservoir temperature (R) and L is the length
q, Mscf/D
of the cement sheath (ft). 1.E+00

Equation 5-2 is simply Darcys linear flow
equation expressed in oilfield units for gas. The 1.E -01
equivalent cement permeability (k*) accounts for the

1.E -02
cement matrix permeability or permeability due to
slot flow through a micro-annulus or radial stress 1.E -03
fracture. In most cases set cement permeability is 0.00100 0.10000 10.00000
k*, md
quite low unless it has been invaded by gas during
Figure 5-3 Gas flow rate through or along a cement
the hydration process after being placed in the
sheath for a range of equivalent cement permeabilities
annulus. It is rare that even gas-cut cement would
exhibit permeability values in the 10 to 100 md To quantify the amount of zonal isolation that
range; however, it is entirely possible that the exists in a given circumstance, Economides (1990)
equivalent cement permeability could reach values presented the Index of Zonal Isolation (IZI) defined
this high if a micro-annulus or radial stress fractures by Eq. 5-4, which expresses the ratio of flow rate into a
were present. Equation 5-3, below, expresses slot flow well from the intended formation to the potential flow
permeability in oilfield units: through the cement sheath:
173
Modern Fracturing
Proper selection of preflushes and spacers that are

q khL
IZI = = ,
qcem r (5-4) compatible with the mud and cement slurries
k * (rw 2 rcas 2 ) ln e + s Proper selection of fluid volumes (contact times)
rw
Use of computer simulation to determine flow
where q is the flow rate through the formation (stb/D rates for optimum displacement efficiency while
for water or oil, Mscf/D for gas), qcem is the flow rate maintaining well control and preventing losses.
through or along cement sheath (stb/D for water or
oil, Mscf/D for gas), k is the formation permeability 5-2.4 Predictive Wellbore Stress Modeling
(md), h is the reservoir thickness (ft) and the remaining
variables are as in Eq. 5-2. As discussed in the previous section, effectively placing a
Equation 5-4 is suitable for either gas or liquid flow cement slurry across any zone of interest is fundamental to
because it is simply a ratio of Darcys radial flow equation achieving good zonal isolation. However, even successful
using the formation permeability, k, in the numerator and cement placement may not necessarily guarantee that
Darcys linear flow equation using the equivalent cement adequate zonal isolation will be achieved or maintained
permeability, k*, in the denominator. The equation can be throughout the entire life of the well. During the well
used to ascertain the flow rate through the cement sheath construction, completion and production phases
matrix or micro-annulus for a given k*, or whichever k* of the life of a well, the cement sheath is subjected to
value is expected to achieve a desired flow rate ratio. radial and tangential forces that can be compressive
or tensile in nature. If the subjected forces exceed the
5-2.3 Review of Fundamental mechanical properties of the cementing material, a loss
Cement Placement Practices of zonal isolation could occur due to debonding of
the cement interfaces or possibly the creation of radial
To obtain effective zonal isolation for the life of a well, stress fractures within the cement sheath itself. These
steps must be taken to insure that the wellbore is suitable forces are induced by temperature and pressure changes
to be cemented prior to and during the cement placement within the wellbore as a result of drilling, completion,
operation. It is imperative that the drilling mud in the stimulation and production operations. Forces can
hole at the time of cementing be conditioned to optimize also be induced on the cement sheath by changes in
the ability of the preflush and/or spacer trains to displace temperature and pressure in the adjacent formation as a
it from the annulus completely before cement is placed. consequence of pressure depletion, injection operations
Inefficient mud displacement is a major cause of failing and/or far field stress variations.
to achieve adequate zonal isolation (Sauer, 1987). Recently, a considerable amount of research has
Although the best practices for achieving high been performed to understand better the forces imposed
displacement efficiencies have been known for years it is on a cement sheath during the life of a well and the
not always possible to execute all of these practices due to mechanical properties required from the cementing
overriding circumstances. Unfortunately, very frequently materials to withstand these forces (Thiercelin et al.,
many of the practices are ignored in an effort to speed 1998; di Lullo and Rae, 2000; Mueller et al., 2004; and
the drilling operation with complete disregard for the Gray et al., 2007). As a result of this work, engineering
critical need of achieving the zonal isolation required for tools, cementing materials and methodologies have
the productive life of the well. A detailed discussion of been developed to allow engineers to design fit-for-
these practices does not fall within the scope of this book. purpose cementing that can indeed provide long-term
However the list of fundamental cement placement zonal isolation if executed correctly.
practices, as outlined by Carter et al. (1973), includes: A predictive analysis of the wellbore stresses that
Proper mud conditioning prior to cementing will occur during the life of a well should be considered
Casing centralization an essential part of any engineered cement design. This
Reciprocation and/or rotation of the casing during is especially true for gas wells that will be hydraulically
mud circulation and possibly cementing fractured during the completion phase of the well.
174
Cement Stress Calculations Input Data Shading Test 2 Results @ 12015 ft
Low WBP Chg WBP Chg Res P Chg
Sample Case
psi F psi
Well Geometry 6500 6500 0
Hole Dia, Csg O D, I D (in) Radial Radial Tangential Negative

8.500 5.500 4.778
Overlimit Distance Stress Stress is
Additional Parameters in psi psi compression
Casing Cement Formation
Inner Radius, in 2.750 -1771 704

2.389 2.750 4.250
Youngs Modulus, psi 2.910 -1639 571
2.9e + 007 1.8e + 006 4.2e + 006
Poissons Ratio: 3.070 -1527 459
0.27 0.12 0.27
3.230 -1431 363
Comp. Ten Strength, psi: 3.390 -1348 281
3400 495
Overburden Grad, psi/ft: 3.550 -1276 209
1.000
3.710 -1214 146
Time to 50 psi under 12 hrs 3.870 -1158 91
use estimated Tensil Youngs & Poissons in the Calcs 4.030 -1110 43
4.190 -1067 -1
4.250 -1052 -16
Radial Stress Radial/Tangential Stress Field Tangential Stress
-1800 750
600
-1600
Radial Stress, psi
400
-1400
200
-1200
0
4 3.5 3 2.7 3
Dist from Borehole Axis, in Dist from Borehole Axis, in
Figure 5-4 Output screen from analytical predictive stress calculator (Courtesy of BJ Services)
An understanding of the forces that will be induced of stress. These can lead to erroneous results if used
on a cement sheath as a result of any anticipated well outside the scope of their limitations. These analytical
operation (e.g., hydraulic fracturing) will allow a cement calculations also require accurate values for the
design engineer to do one of two things: 1) optimize formation, cement and casing mechanical properties
the mechanical properties of the cementing material to generate accurate results.
to withstand the forces or 2) if the first option is not Figure 5-4 shows an example output for one of
possible, place restraints on the planned well operations these calculations, indicating the radial and tangential
in order to stay below the maximum forces the cement stress profiles for a given wellbore situation. In this
material can withstand. example, the stresses associated with increasing the
There are two modeling methods currently internal wellbore pressure by 6,500 psi were calculated.
available for performing predictive wellbore stress The radial dimensions of the casing, cement sheath
analysis. The first involves the use of fairly simplistic and formation interface were input in addition to
analytical calculators that are based on the solution the mechanical properties for each. In this particular
by Lam for cylinders under pressure. Thiercelin et case the cement analyzed exhibited an unconfined
al., 1998; di Lullo and Rae, 2000; and Mueller et compressive strength (UCS) of 3,400 psi and a tensile
al., 2004, have proposed the use of calculations to strength of 495 psi. The tangential and radial stresses
approximate the radial and tangential stresses that were calculated at various radial distances starting
will be imposed on a cement sheath subjected to with the casing/cement interface. Based on the
temperature and/or pressure change within a wellbore output, the maximum radial load from this scenario
or adjacent formation. These calculations are based was compressive and equal to 1,074 psi, considerably
on certain simplifying assumptions such as linear below the UCS of the cement. Unfortunately, the
elasticity, axisymmetric geometries and the initial state maximum tangential load was calculated to be 704
175
Modern Fracturing
psi and tensile in nature, which is 204 psi greater production strings. A detailed description of slurry
than the tensile strength of the cement. As a result the characteristics that aid in efficient placement is outside
calculations predicted a failure at the cement/casing the scope of this book. Nonetheless, a quick review of
interface, which can be seen graphically on the polar these parameters is warranted:
and Cartesian tangential stress plots.
More often the modeling approach to stress analysis Rheology. Slurry rheological properties are an important
incorporates numerical 3-D finite element analysis design criterion for any primary cementing operation.
models. This type of model permits the use of solids These properties define the slurrys mixability and
mechanics theories including elasticity and plasticity. pumpability in addition to the frictional forces that will
Although these models can provide more comprehensive be incurred during pumping. In fact it is essential to
results, they require time, specialized expertise, characterize the rheological properties and densities of all
experimental data and sizeable amounts of processing fluids that will be in the wellbore during the placement
power to run. On the cutting edge of stress analysis, operation. This allows calculation of an equivalent
Gray et al. (2007) have presented a staged finite element circulating density (ECD), which must be maintained
approach, which allows the tracking of stress and strain between the formation pore pressure and fracture
development throughout the life of a well. gradient at all times during the placement operation.
The rheological properties of the wellbore fluids will
5-2.5 Cement Slurry Criteria also dictate the flow regime the fluids will experience
for Hydraulically Fractured Gas Wells as a function of displacement rate. Understanding flow
regimes and rheological properties provides the basis
Cement slurry design is a critical component of any for calculating mud displacement efficiency, which
well cementing operation. The type of operation needs to approach 100% in order to achieve adequate
to be performed (whether primary or remedial) cement placement. Although a design engineer will
and the type of casing string to be cemented (in have some liberty to tailor the rheological properties of
primary cementing) will dictate the requirements a slurry, many times the ultimate goals can end up being
of the slurry. This section will deal specifically compromised due to the need to meet other slurry
with the slurry requirements for production criteria such as free water, fluid loss control etc.
casing strings in gas wells that will subsequently
be hydraulically fractured, which is the majority of Thickening Time. Thickening time is the length of
new wells drilled around the world. As mentioned time a cement slurry remains in a pumpable state under
in Section 5-2.1, a successful slurry design for any wellbore temperature and pressure conditions. This time is
primary cementing operation will: affected by many factors including temperature, pressure
Place the slurry efficiently in the annulus and the obvious chemical retardation induced by the use
Prevent gas invasion during and after hydration of cement retarders and other additives. Thickening time
Provide the required mechanical parameters also depends on several mechanical factors, including the
when set, to withstand the induced stresses that mixing energy the slurry is exposed to during the mixing
will occur on the cement sheath throughout the and pumping operation, and the dynamic and static fluid
life of the well. loss the slurry experiences during placement. Simply put,
the thickening time of a slurry should be designed to equal
5-2.5.1 Slurry Criteria for Optimized Placement the time required to mix and displace the slurry based on
the desired pump schedule, plus a safety factor. A safety
The ability to effectively place uncontaminated factor is needed because the standard API test does not
cement slurry in the casing/open-hole annulus is the take into account the dynamic or static fluid loss that
first requirement for obtaining zonal isolation. This can greatly affect thickening time. The interdependence
requirement is fundamental to all primary cementing between slurry thickening time and anti-gas migration
operations but especially true for those designed for properties will be discussed later in this section.
176
Fluid Loss Control. The inability to control slurry

dehydration during and after a pumping operation
can greatly compromise effective zonal isolation. The
dynamic fluid loss of cement slurry is affected by various
parameters such as formation permeability, mud filter
cake thickness and permeability, the amount and type of
cement fluid loss additive, and the differential pressure
the slurry is exposed to. In severe cases, slurry dehydration
can form bridges in the annulus that can lead to the
premature termination of a cementing operation. Rapid Set Prevents Gas Migration
During Transition from Liquid to Solid
5-2.5.2 Slurry Criteria for Anti-Gas Migration
Annular gas migration is a potential problem for any well

that requires cementation across a gas-bearing zone. The
causes of gas migration and methods proposed to prevent Folded Chart- Approximately
it have been well documented over the years. The range 37 hours not shown
of methods for preventing gas migration covers a wide
spectrum, and although many have shown promising
results, the reality is that none have been totally successful
on their own. A review of the literature indicates that
most authors concur on the following fundamental
requirements for preventing gas migration: Relatively Thin Slurry Allows
Compliance with the fundamental cement Easy Placement of Slurry
placement practices
Slurries must exhibit zero free water breakout and
minimal particle settling tendencies
Slurries must exhibit API fluid loss control values of
50 cc/30 min or less
Slurries must be designed to minimize cement matrix
volume reduction
Slurries must exhibit low permeability throughout
the liquid, transition and set states.
Although somewhat more controversial, another

methodology for preventing gas migration incorporates
the use of "right-angle set" (RAS) slurries. RAS slurries
as presented by Parcevaux et al. (1990) are slurries that
exhibit no progressive gelation tendencies but set very
rapidly due to hydration kinetics. The RAS phenomenon
can be seen on the standard API thickening time chart
(Fig. 5-5), which demonstrates the ability of a slurry to
go from a relatively low consistency value to over 100
Bc (Bearden units of consistency) in just a few minutes.
Due to the short transition phase from liquid to semi- Figure 5-5 Consistometer chart demonstrating RAS
solid, RAS slurries maintain the ability to transmit full phenomenon
177
Modern Fracturing
hydrostatic pressure up to the start of set. Furthermore, Rogers et al. (2004) pointed out that the
it is believed that the RAS phenomenon differs from the SGS transition time defined by Sabins et al.
rapid gel strength development presented by Sabins et al. (1982) is frequently confused with the transition
(1982) in that a true set occurs involving the deposition time associated with a standard thickening time test,
of mineral hydrates which aid in reducing permeability which measures the dynamic gelation profile. These
and the prevention of gas migration. authors also concluded that although beneficial,
Purvis et al. (1993) observed that the gelation neither RAS characteristics nor short SGS transitions
development profile of a slurry under dynamic conditions times (< 40 minutes) alone will provide a 100%
and static conditions can vary greatly. In other words guarantee that a slurry will be able to prevent gas
although the RAS methodology has merits for preventing migration. Their findings confirmed that designing
gas migration, the gelation profile occurs during a a slurry based on the fundamental requirements
dynamic state and may not occur in the same manner listed above is a good start to providing a fit-for-
while in a static state. The significance of this is that most purpose slurry; however, the only truly valid way to
slurries are designed with thickening times greater than determine if a slurry will provide gas-tight properties
the time required to actually place them. In this situation is to physically test it in a gas flow simulator, such
if the gelation profile of an RAS slurry differed greatly in as the one shown in Figure 5-6
the static state, the benefit of the RAS design would be Although the process of designing a fit-for-
negated. Therefore, the speed at which a slurry develops purpose gas-tight slurry can be somewhat laborious,
static gel strength (SGS) is also a very important property the benefits of doing it correctly far outweigh the cost
for preventing gas and fluid invasion. implications of not doing it correctly.
The transition time as defined by Sabins et al. (1982)
is the time in which a slurry goes from being a fully 5-2.5.3 Slurry Criteria for Long-Term Zonal Isolation
hydraulic fluid to a highly viscous mass that demonstrates
solid properties. Transition time is measured from the The third component of designing a slurry for
time the slurry develops an SGS value equivalent to long-term zonal isolation in a gas well pertains to
100 lbf/100 ft2 and ends when the SGS reaches 500 optimizing the mechanical properties of the set
lbf/100 ft2. It is widely believed that when the slurry cement. To begin, a predictive stress analysis must be
SGS value reaches 100 lbf/ 100 ft2 it is prone to gas performed for the numerous scenarios during the life
or fluid invasion because this is when the hydrostatic of the well (such as a hydraulic fracturing treatment)
pressure transmission begins to be restricted (Sabins and that will subject loads to the cement sheath (see
Sutton, 1986). As the SGS value increases to 500 lbf/100 Section 5-2.4). It is necessary to understand the
ft2, the slurry has generated sufficient gel strength to magnitude and direction of these forces in order to
resist the invasion of fluid or gas. establish the design criteria for the slurry. Although
the principal mechanical design criteria for cement in
Gas Flow Model
the past has been based on unconfined compressive
1000 psi
Oil Pressure strength, there are many other mechanical properties
Hydrostatic
Pressure
Pore Pressure
Filtrate
Piston
Gas Flow
Flowmeter of the cement such as tensile strength, Youngs

High Pressure
Zone
500 psi
Transducer
modulus and Poissons ratio that could be much more
important than compressive strength. In reality, the
Nitrogen
Supply
Cement
majority of the time the maximum stress loads placed

Pore
Pressure
Screen Control Panel
Cement
Slurry
Data Storage
Graphical Display on a cement sheath are tensile in nature rather than
Multi-tasking
Screen
Transducer compressive. Based on this premise it is imperative
300 psi
Low Pressure
Zone
300 psi
Back
Pressure Direct to Excel
that old paradigms such as the harder the cement,
Nitrogen
Supply
Flowmeter
Spreadsheet
the better be phased out for new paradigms that
Filtrate
focus on a more encompassing understanding of the
Figure 5-6 Gas Flow Simulator (from Rogers et al., 2004) resilient properties of cement.
178
During the past decade there has been a substantial 5-3 Identifying Gas Pays,
amount of research into cement additives that can enhance Permeability and Channels
the resiliency of oil and gas well cement to withstand
loads. For example, calcium silicate-based minerals, 5-3.1 Pay and Water Zone Logging Methods
polymeric blends and volumetric-equivalent sand
systems have proven to considerably enhance the tensile Logging, whether to locate gas pays, to assess
and flexural strength and Youngs modulus of slurries contribution over the zone, to estimate turbulence
(Heinold et al., 2002; and Mueller, 2003). Furthermore, problems, or to identify water entry points for repair,
unless absolutely required for well control purposes, the are all part of a larger surveillance effort necessary to
density of a slurry can be reduced from the traditional better define the formation and the flow behavior
densities associated with the API water requirements. (Julian et al., 2007; and Fox et al., 1999). The
Reduced-density slurries with a proper additive package surveillance package in a gas well, encompassing
can many times satisfy all three of the fundamental slurry open-hole and cased-hole or production logging,
design requirements for placement, anti-gas migration, is integral to a functional depletion plan to
and resilient mechanical properties, more easily than maximize reserve recovery.
higher-density designs (di Lullo and Rae, 2000). Logging to identify gas pays most commonly
In conclusion, it is essential to understand the defines the presence of gas, porosity, water saturation,
force loads that a given cement sheath will be exposed mineralogy and clay identity and presence. Open-
to and to design a slurry accordingly to provide the hole logging includes the initial pay identification
mechanical properties under downhole conditions to and characterization; cased-hole logging usually
withstand the subjected loads. focuses on production logging, conformance issues,
missed pay identification or other reserve addition
5-2.6 Fracturing Constraints Required opportunities. In general, logging determines if
to Maintain Long-Term Zonal Isolation there are commercial quantities of gas, if the zone
is sufficiently permeable to produce the reserves and
Under certain circumstances it may not be possible if the water saturation is mobile and manageable
to optimize the mechanical properties of the cement during production. The minimum log suite
slurry to withstand the forces predicted from a necessary to answer these questions includes gamma
proposed hydraulic fracturing job. In this scenario a ray, resistivity/conductivity and multiple porosity
predictive force analysis should be used to mandate the readings. Other logs are frequently used for special
maximum pressure and temperature changes allowed applications and confirmation of findings. Special
during the treatment to prevent compromising the plotting and cross-plotting of the log data, when
cement sheath and zonal isolation. combined with experience in an area, are powerful
Typically it will be difficult to reduce the bottomhole tools in locating and assessing a pay zone.
treating pressure associated with a fracturing treatment Gamma ray logs, which measure the natural
by reducing the pump rate because the wellbore pressure resistivity of the zone, help estimate lithology,
adjacent the perforations will be primarily controlled correlate depths across a field and specifically identify
by the mechanical rock properties. Nonetheless, the clean sands that often have higher permeability than
change in wellbore temperature during the job maybe the rest of the zone. Several different porosity logs
the overriding factor associated with the subjected stress (density, neutron, sonic and magnetic resonance)
loads. In this scenario it may be required to heat the are used to generate an estimate of permeability. The
fluid at surface prior to the job to minimize the change differences in measurements, when cross-plotted
in temperature the wellbore will experience. Although against other data, can frequently be used to identify
heating the fracturing fluid will add logistical complexities a gas-bearing zone. Alone, each has limits, but when
and operational cost to the treatment, the loss of zonal multiple porosity log types are processed, the porosity
isolation would be far more costly. estimate is very good. The neutron-density combo
179
Modern Fracturing
log is a standard tool for gas identification. Gas in During the well construction phase, logs are also used
the pores causes the density log to read high and the to monitor the quality and the top of cement (temperature
neutron log to read low. Normally, the density log and cement bond logs) and to estimate deformation
tracks just lower than the neutron log; however, when or other damage in the casing (caliper and ultrasonic
gas is encountered, the neutron log crosses over. A low inspection tools). After the pay is producing, logging is
gamma ray signal indicating a clean (low shale content) useful to assess how much of the zone is contributing to
formation and resistivity-calculated water saturation flow. Formation permeability often varies widely, even in
of less than 30 to 50%, depending on formation clay a single pay, due to depositional environment variances,
content, are secondary indicators. sediment reworking, natural chemical modification
A few of the logs and surveillance control methods and tectonic effects. The permeability variation, often
for gas wells and a very basic assessment of the two or more orders of magnitude, is a strong influence
information they offer are given in Table 5-2. on location of flow channels in the rock. Production
logging identifies points of fluid entry (and exit) in the
Table 5-2 Logging Tools for Open-Hole Sections wellbore. Special constraints of wellbore fluid type, well
Logs or Surveillance
Potential Information Available
deviation, tool conveyance, temperature effects, flow
Method
behavior and flow rate are important considerations
Gamma-ray Lithology and correlation
in the application of logging tools. Knowledge about
Water saturation,
Resistivity
potential hydrocarbon reservoir pressures and permeabilities can assist in
Capacitance Fluid type creating a design that can control cross flows.
Neutron, Density Porosity estimation, gas presence Optimizing flow and reserve recovery may require
Dipole Shear Sonic Formation strength balancing the gas production rate against water coning
Pulsed Neutron Water saturation potential or timing of gas recovery against recovery of oil
Temperature Cement top, inflow, channels reserves from a thin oil zone overlain by a large gas cap.
Noise Log Leaks and fluid entry These types of decisions require a surveillance plan, with
Cement Bond Log Cement bonding quality regular review, to achieve the reserve recovery goal.
Calipers Hole size and shape Reservoir boundaries and barriers are rarely
Sweeps w/markers
Swept hole volume adequately described through seismic or logging
prior to cementing
methods. Continuing production surveillance
Channel flow, water
Oxygen Activation
differentiation from hydrocarbons offers the possibility of more accurately defining
the reservoir and often identifying missed pay and
After the pay is identified, an assessment of re-completion opportunities.
water contact is made using the resistivity log.
Knowledge of how active the water movement may 5-3.2 Effect of Formation Clays and Micro-porosity
be and an estimate of the vertical permeability of the
reservoir are valuable in estimating effects of water Attention to the mineralogy and the overall formation
drive over time for both gas recovery and water clay content and location is needed, particularly
encroachment into the wellbore. This information when gas is indicated and high water saturation
is also required to optimize any fracture stimulation. makes the pay unattractive.
Estimates of formation vertical permeability - In a few cases, some clays such as smectite, kaolinite
which can range from < 1/100 of the horizontal and illite can create high micro-porosity in the pores,
permeability to equal values - is usually generated trapping water that increases the water saturation
from core studies, micro-logs, depositional reading but does not move. A gas well in the Nile delta,
environment and offset fluid movements. The for instance, indicated pay zone water saturation over
formation vertical permeability is a dominant control 50% but produced near water-free gas. The reason was
in horizontal well production but is often inadequately extensive clay in a form with space between the clay
known over the reservoir. platelets that trapped a significant amount of water.
180
Conversely, an indication of very low water Centralizing the logging tool, restricting flow around
saturation (< 10%) may be a dry or under-saturated the tool, tool movement while logging and routing all
zone that will adsorb water from any injected fluid, flow past the measuring surfaces were at least partly
resulting in a lower relative permeability to gas after successful in assisting the logging reading. However, in
the water contact (Bennion et al., 2000a, 2000b). This deviated wells with an un-isolated slotted or perforated
effect is rarely seen and may be the result of either too liner, the logging tool will very likely show all fluid entry
little connate water during cementing, or more likely, at the heel or the last point where the fluid flowing on
the water was flushed out by excessive dry gas flow over the outside of the liner can enter the wellbore. Although
geologic time. Whatever the cause, the effect will be the work was initially done to assist the interpretation of
sharply lower permeability to gas after any operation production logs, the implication for flow, especially in gas
that floods the matrix with a liquid. Recovery to wells, cannot be ignored. Deviated wells have excellent
initial state will be extremely slow. application benefits in many producing environments,
but the wellbore construction must take into account the
5-3.3 Wellbore Deviation and manner in which fluids flow.
Resultant Logging and Flow Problems
5-3.4 Completion Considerations
With more wells being drilled from pads and platforms, for Naturally Fractured Reservoirs
the effects of fluid movements along deviated wellbores
must be considered. Production of a single-phase Naturally fractured reservoirs (NFRs) are widespread
fluid in a horizontal well is rarely an issue, but as and are encountered much more frequently than
other phases are added, the flow behavior becomes one would expect. The percentage of global reserves
increasingly complex. In production logging, fluid contained in NFRs is currently unknown. Nonetheless
density segregation in the wellbore creates zones of the contribution from NFRs is significant and will
stratified flow and even areas of refluxing liquid that can probably increase as conventional reservoirs without
make log readings questionable at best and may create natural fracture networks are depleted. Although
significant flowing backpressures that can sharply affect typically associated with carbonate formations, natural
production in lower pressure gas wells (Scrimgeor et al., fractures are commonly encountered in sandstone and
1983; and Bamforth et al., 1996). shale formations as well. Naturally fractured reservoirs
The multi-phase flow effect was noted first in the field are of increasing importance in North America, where
as engineers recognized that log interpretations in deviated the exploitation of tight gas sands and more recently gas
wells were yielding inconsistent and confusing results. A shales (see Sections 11-3.2 and 11-5) is on the rise. In
series of flow loop tests in inclined test fixtures helped these reservoirs it is not uncommon for the rock matrix
identify the problem as phase hold-up, where the lighter permeability to be in the very low microdarcy or even
fluid segregates to the upper part of the flowing pipe cross- nanodarcy range. It is generally accepted that these
section, occupying a correspondingly smaller volume of tight gas plays would not be commercially viable in the
the flowing cross-section and leaving the higher-density absence of natural fracture systems.
fluids moving much more slowly along the bottom side. As pointed out in Section 5-3.1, a variety of electric
Several distinct flowing regions were recognized as the logging techniques and cut-off criteria are available for
deviation changed from full vertical through the Boycott determining productive pay in conventional reservoirs
settling range of 30o to 60o, to near horizontal flow. Even that produce via matrix permeability. However, accurately
changes of as little as 2o from horizontal affected the type identifying pay and characterizing reservoirs that contain
and location of the rapidly moving layer. Beyond 90o, in naturally occurring fractures caused by external or internal
the over-horizontal deviations, there is a reversal of the stresses within the formation is much more complex.
location of the fast moving layer, with the higher-density Characterizing a natural fracture system plays
fluids creating a faster flow along the bottom of the pipe a fundamental role in field development decisions
and occupying a smaller portion of the pipe. pertaining to well placement, wellbore azimuth,
181
Modern Fracturing
wellbore construction, completion sizing and associated with the measured matrix permeability,
stimulation design. Natural fractures can help or the variation is a strong indication that the borehole
hinder production. Depending on the degree of is in communication with a natural fracture system
the natural fracturing present, a reservoir could extending deep into the reservoir.
become very compartmentalized and drain less In reality, it may require a combination of
acreage than expected. Therefore, it is vital to observations and indicators to provide sufficient
identify the presence, type and density of natural evidence to classify a reservoir as naturally fractured.
fractures within the reservoir. In general, naturally fractured reservoirs require more
Understanding the orientation and magnitude of up front study and thought than simpler and more
present day horizontal stresses and the orientation of the straightforward conventional reservoirs.
natural fracture system will facilitate proper planning of
a well designed to optimize the drainage of the reservoir. 5-3.5 Formation Characterization
Avoiding geohazards is also critical. In the case of for Well Completions
deviated or horizontal well designs, the wellbore azimuth
may be planned so that the wellbore, or subsequently Formation characterization provides rock and fluid
induced hydraulic fractures, intersect the natural fracture property information for use in predicting reservoir
network for optimum inflow area or, conversely, may hydrocarbon volumes and well production rates and,
be planned to avoid intersection of natural fractures consequently, the appropriate well construction to
that connect undesirably to an aquifer. accomplish the task. The identification of problems
There are two categories of methods available that involve damage or the propensity to damage,
for determining the presence and degree of natural such as sand production, scales, paraffin or asphaltene
fractures in a given reservoir: direct observations deposition, etc., is very important.
and indirect indicators. Measurements come from a variety of tools:
Direct observations are typically the least Well logs, which measure resistivity, nuclear,
expensive method and can include close examination acoustic, and magnetic formation properties
of drill cuttings, analysis of conventional or rotary along the well length
sidewall cores and the interpretation of borehole Well tests, which measure formation pressures, obtain
image logs. Under low magnification, the presence reservoir fluid samples and sometimes measure
of mineralization on fracture surfaces observed in fluid flow rates from selected formations.
either drill cuttings or cores demonstrates that the Mud logs, which provide analyses of rock
fractures were naturally created and not induced and hydrocarbons circulated to the surface
during the drilling process. Additionally, borehole within the drilling mud from the formation
image logs provide a full 360o view of the wellbore that was drilled
and can provide useful qualitative information about Cores, which provide samples of formation rock
the, position, density, orientation and extent of for direct physical measurements.
natural fractures in the reservoir.
Indirect indicators while drilling include an Well logs are the predominant formation
increased rate of penetration in conjunction with characterization source, because they provide the
mud losses, bit chatter and multi-arm caliper most complete set of information for the lowest cost.
logs indicating eccentric or out-of-gauge holes. Of particular importance to well completions are
Substantial mud losses can indicate the presence thickness, porosity, lithology and heterogeneity. Well
of a significant natural fracture network. Indirect testing, described in Chapter 3, provides the zone
production indicators can also provide insight permeability, average reservoir pressure and skin. It is
into the presence and extent of naturally occurring the only method described in this section that can look
fractures. For example, if the actual production deep into the formation. For more detailed description
from a zone greatly exceeds the expected production within a presumed zone, production logging delineates
182
vertical heterogeneity. This is particularly important 2000). Open-hole completions represent the simplest
for stimulation because it provides answers for the type of completions and may have significant benefits
appropriateness of diversion and staging. in high-permeability oil reservoirs. Gas reservoirs,
Interpretation of all available well data is essential however, significantly benefit from hydraulic
for an accurate formation characterization. Well data are fracturing to reduce the effects of near-wellbore
applied using a petrophysical model that describes the drilling and completion damage, and from the
formation rocks and fluids. reduction of non-Darcy flow effects (see Section 2-4).
Formation rock and fluid properties that are usually Therefore, in most gas reservoirs cased and perforated
of most interest are: completions are still preferred, although the casing
Zone depth and thickness area opened by perforating must be maximized.
Lithology (e.g., sandstone, limestone) Open-hole completions may also present challenges
Pore types (e.g., between grains, fractures) in formation and water influx control, as well as
Porosity (percentage of reservoir volume that contains being significantly more restricted by near-wellbore
liquids and gas) damage from completion fluids.
Permeability (how well fluids flow through Open-hole completions have a large area of contact
rock pores) between the wellbore and the formation. Although
Damage, type of damage and potential for it the hydraulic fractures preferred on nearly all gas
Pore fluid saturations (percentage of pore volume will provide orders of magnitude greater contact area
occupied by each fluid) compared with an open hole, some restriction of flow
Pore fluid properties (e.g., density, water salinity) may be seen due to the small area of the perforations
Formation pressure at the wellbore entry points. Additionally, open-
Formation characterization is often hole completions avoid formation damage caused
expanded across an entire field or geologic basin. by cementing and damaging convergent flow effects
This effort requires corelating the formation caused by perforation restrictions. However, cased-
and its properties among multiple wells and hole wells can achieve open-hole productivity if
interpolating formation properties between sufficient number of perforations of sufficient size
wells, using well data already mentioned and the and penetration are used. Although it is difficult
following data acquired over time: to perforate through formation damage due to the
Seismic data resulting from the measurement of the crushed zone that exists around each perforation
path/time of surface-originating sound waves and tunnel (Fig. 5-7), it is possible to significantly reduce
their reflections from formation boundaries the effects of this zone by use of the appropriate
Well performance data including flow rates, formation perforating technique (see Section 6-4.3).
pressures and produced fluid properties.
Computers are used at every step of formation
Damaged Undamaged
characterization from data acquisition to well Casing Formation Formation
performance history matching, well performance
Cement
prediction, hydrocarbon volume determination and

the display of formation characterization results.
Perforation Tunnel
5-4 Sizing the Completion
Crushed Zone Around
Perforation
5-4.1 Initial Design Considerations
The first completion design decision on casing the

pay zone is not its size or weight but whether to run Figure 5-7 The crushed zone surrounding the perforation
casing at all (Bennett et al., 2000; and Parlar et al., tunnel
183
Modern Fracturing
Nonetheless, as discussed in Section 2-5.2 in the Casing strings are run to make continued
case of high-permeability, high-rate gas wells nothing drilling possible. The drilling fluid density operating
can compete with hydraulically fractured completions window formed between fracture breakdown
for gas wells. This is due to the significant reduction pressure as the upper limit and pore pressure and/or
in the non-Darcy flow effects, which greatly enhances hole stability (sloughing) control as the lower limit,
the production response beyond the benefits of a will narrow with tectonic, deposition, age, charging,
negative skin. Fracturing should be considered for all basin variance and other factors. Basic well design
gas wells. In low-permeability formations, fracturing uses a conductor pipe to keep out loose soil; surface
improves the flow towards the wellbore, and in higher- or protection strings to isolate water or flowing
permeability formations, very high-conductivity sediments; and production casing and liners to isolate
fractures can reduce non-Darcy flow effects. zones of productive interest from fluids, pressures or
Selecting the actual size of the flowing pathways in formations (see Fig. 5-8). In completions that require
the well is usually done with a nodal analysis package hydraulic fracturing, it should be noted that cased,
with the intent of optimizing flow, not only at the start cemented and perforated production strings provide
of production, but also repeated frequently during the best chance of maintaining zonal isolation and
wells life if possible. The selection of casing is dictated optimized fracture placement.
by the largest downhole production equipment needed Whereas the wells casing design is almost
for an application and the number of casing strings and completely concerned with the loads and forces
liners necessary for pressure and fluid containment to of completion and operation, tubing selection
deliver the casing size to that depth. These strings are must also optimize the ability to flow fluids to
set at the initial completion and generally cannot be surface. In order to optimize flow over the life of
changed, once set, without the expense of drilling and the well, the tubing may have to be changed. The
re-completing the well. size of the tubing is driven by the expected gas
Swab Valve production rate in conjunction with the ability
Flow Cross to optimize natural lift of any liquids by the
velocity of the expanding gas.
Choke
5-4.2 Flow Factors for Tubing Design
Wing Valves Wing Valves
The ability of a well to flow is addressed by the inflow

Upper and Lower
Master Valves performance relationship, IPR. In an example gas well
IPR (Fig. 5-9), the curve illustrates the relationship
Inner Annulus Access of productivity of the formation at drawdown to
Outer Annulus Access production rate of the well (Brown, 1982) The
Conductor Pipe IPR curve is essentially a snapshot in time of well
performance potential. As the well depletes in pressure,
Surface or Production
Casing String the IPR curve shifts to the left, indicating the effect
Cement
of pressure removal on the production rate. There is a
wellbore influence impressed on the IPR curve by any
Production Casing flowing restriction in the near-wellbore area.
Tubing String Figure 5-10 is an interesting comparison between
Packer cased-hole and open-hole gas production in a high-
PAY
permeability gas reservoir. In this data, a new well with
ZONE an open-hole completion replaced a cased and perforated
Perforations gas well. Although separated by only 20 meters, the
Figure 5-8 Typical wellbore design difference in production from the two wells was striking
184
- with an IPR indicated gain of over 30% (although counting the flow area of the perforation tunnel itself.
there could be reasons for this besides the change from The restriction caused by small or insufficient numbers of
cased to open hole). In gas wells, both fluid viscosity perforations presents a challenge. Fracturing is beneficial
and compressibility are pressure dependent. The IPR on almost any gas wells (problems with adjacent
modeling is also complicated by high velocities around water or adjacent depleted formations excepted), but any
the wellbore that produce turbulent flow. Most models limited entry into the casing poses a restriction. Large,
assume Darcy flow (laminar) and may not be accurate open perforations are a necessity in high-rate wells.
for the pressure drops produced by turbulence. The production effects of these limits of perforating
are rarely significant in low-permeability formations
IPR - Natural Lift, Tubing Performance (less than about 100 md) that are hydraulically
fractured, but the effect of the inflow constraints are
Small Tubing
increasingly felt as permeability rises.
Large Tubing
IPR Curve
Bottom Hole Flowing Pressure
5-4.3 Tubing Selection
} Drawdown
Tubing selection is based on both physical forces and
several minimum and maximum sizes that determine
production efficiency:
1. Tension, collapse and burst forces over the
entire cycle of completion, workover and
production operations. This is best done with
a stress analysis program. The design accuracy
is totally dependent on understanding the
Flow Rate
loads and forces and the accuracy of the data
Figure 5-9 Inflow performance relationship (IPR) curve
with overlaid tubing performance curves (TPC) for
used in the calculations.
tubing selection 2. Maximum flow velocity of produced fluids that
will not create fluid-based erosive damage of the
4000
pipe (erosion or erosion-corrosion). If solids are
Bottom Hole Flowing Pressure, psi
3500
produced, an entirely different approach will be
3000
needed; the maximum velocity for solids-laden
2500
fluids is a fraction of the acceptable velocities for
2000
1500
single-phase, low-density fluids.
Open Hole
1000
3. Minimum flow velocity needed to lift liquids in the
Cased & Perf
500 well. The Turner and Coleman equations (Turner
7" TPC
0 et al., 1969; and Coleman et al., 1991) are the basis
0 100 200 300 of design; however, well deviation makes natural
Production, MMscf/d
flow lift much more difficult.
Figure 5-10 Comparison of open-hole and cased/ 4. Minimum flow rate needed to prevent surface
perforated high-permeability gas wells separated by
less than 20 meters
fouling of the pipe.
The major difference in open hole vs. non-fractured 5-4.4 Multi-Phase Flow and Natural Lift
cased-hole completions is in the wellbore-entry
restrictions created by casing the well. For example, Lift in natural flow uses the rise and expansion of
a well with 7 in. casing, perforated with 0.75 in. (1.9 bubbles of free and associated gas to counter the
cm) entrance hole perforations at 12 shots per foot (39 fall of liquids and to push the liquids out of the
spm) has only 2% of the pipe wall area open to flow, not well. This is lift via hindered settling: The rising gas
185
Modern Fracturing
bubbles interfere with the density-induced settling of 5-4.5 Multiphase Flow and
the liquids through the gas and, if the interference is Flow Correlation Options
sufficient, liquids will be expelled from the well. In a
vertical well, liquid lift by hindered settling is described The ability to predict the pressure drop caused by
by well-known equations (Turner et al., 1969). For a frictional and hydrostatic losses from the produced
deviated well, however, rapid density segregation of gas fluid stream in a well completion is essential.
to the high side of the pipe will decrease the efficiency Numerous factors will significantly affect the total
of lift at less than turbulent conditions, making gas lift pressure losses during production; these include
calculations much more complicated. Multiple flow the flow area; flow velocity; gas-to-liquid ratio
paths, changes of deviation, pipe size and fluid entry (GLR); fluid density, viscosity compressibility and
or exit all complicate the design. composition at the bottomhole and surface; flowing
The bubble rise velocity is generally proportional pressures and temperatures; and well deviation.
to the inside diameter and roughness of the tube. Analytical equations are available for accurately
When fluid is lifted, the upward velocity of the gas is calculating the hydrostatic and frictional losses for
often 10 to more than 100 times the upward velocity single phases of both incompressible and compressible
of the liquid, creating liquid hold-up. The intent of (gas) fluid flow (Economides et al., 1994).
tubing selection and optimization is to minimize the For multi-phase flow conditions, there are no
liquid hold-up and the back-pressure it creates on the analytical solutions for deriving pressure losses. Over the
formation. The larger the tube (up to several inches in years a vast amount of research has been done studying
diameter), the faster the bubbles rise. Because the gas the effects of multiphase flow (oil, gas and water),
bubble grows as it rises in a producing well, the friction resulting in several correlations based on empirical
is least in the deeper sections of the tubing string and laboratory and field data. These correlations can be used
greatest near the surface. Considering the bubble to estimate the frictional and hydrostatic losses associated
behavior, a tubing string of a single diameter would with a wide variety of multi-phase flow conditions.
lift with least efficiency at the bottom of the well and Because the experimental set-ups differed - i.e., vertical
would have the highest friction (also a back pressure) at vs. horizontal pipes, and the type of flow regimes
the top. Some wells can benefit from a tapered string - a studied - some correlations have more appropriate
tubing string with small ID tubing at the bottom and applications than others. The best known correlations
larger ID tubing at the top. Difficulties with this type for calculating multi-phase pressure gradients are those
of string are the larger casing needed at the surface for by Hagedorn and Brown (1965), Brown (1977), Beggs
clearance of the larger upper-section of the tubing and and Brill (1973) and Brill and Beggs (1978).
the effect of the changes in fluid rates and types over Today, computer algorithms (nodal
even a relatively short production period. analysis) are used, and in general they follow the
The standard approach to a single-diameter tubing following procedure:
selection involves overlaying the tubing performance The well is segmented, starting either from the
curves for specific well conditions on the IPR curve for bottomhole or tubing head flowing pressures. Within
the well, as seen in Fig. 5-9. The part of the TPC curve each interval, the average temperature is either the mean
to the right of the minimum flowing pressure values are temperature based on some geothermal gradient or the
steady-state flow regimes if the right-hand part of the log-mean temperature (Bradley, 1987.) The pressure at
curve intersects the IPR curve. Flow rates to the left of the end of the interval is assumed and used with the
the intersection are unstable and associated with slugging starting pressure to determine phase compositions and
behavior. When the TPC curve does not intersect the to calculate physical properties
IPR curve, the well will not flow without energy from The first step in using a multi-phase correlation is
a lift system. As reserves and pressure are depleted in a to employ a flow-regime map to identify whether
formation, the tubing size selected by this method will the flow is bubble, slug, churn or annular. Flow regime
eventually become too large to enable liquid lift. identification as suggested by Duns and Ros (1963)
186
and Taitel et al. (1980) is essential in selecting the most being used, or to investigate the accuracy of the input
appropriate correlation and determining the hold-up data if the correlations being used have been proven
(the extra liquid present in each segment), a variable that effective under similar flow conditions.
is at the heart of all multi-phase correlations.
Then, the pressure gradient is calculated and 5-4.6 Critical Lift Factors
the outlet pressure determined and compared
with the assumed. The process is repeated until a Gas will quickly separate from liquids in any flow rate
convergence tolerance (e.g., 1 psi) is achieved. Then, less than turbulent flow. Loading up of gas wells by
the outlet pressure of the segment becomes the inlet accumulating either static water columns in the bottom
pressure of the next segment and the procedure is of the well or by refluxing water layers in the midst of the
repeated all the way to the top or the bottom of well are both symptoms of gas flow velocity that is too
the pipe. Table 5-3 gives recommended usage for slow to adequately lift the water. A number of studies,
various single- and multiphase flow correlations, starting with that of Turner et al. (1969), and continuing
according to wellbore conditions. with refinements to the present, have proposed methods
of calculating the minimum or critical rate to lift liquids
Table 5-3 Recommended Single- and Multi-Phase in droplet form from vertical wells (Fig. 5-11) (Coleman
Correlations and Calculations et al., 1991; Yamamoto and Christiansin, 1999; Lea
Type of Flow Recommended Method
Single-Phase Incompressible Analytical equations, constant
and Tighe, 1983; Lea and Nickens, 2004; Guo et al.,
Fluid (water or oil above bubble density and constant viscosity 2005; and Lea and Nickens, 2004). There is considerable
point pressure) in each segment
scatter in the predicted critical gas flow rates to achieve
Analytical equations, using
real gas law and pressure- liquid lift and the actual gas flow rates that effectively
Single-Phase Compressible
dependent properties. Trial unload the wells. The entrained drop or hindered
(Gas)
and error convergence is
necessary for each segment settling lift used as the basis of these models is affected
Multi-Phase Flow in Vertical
Bubble Flow: Griffith (1961) by a number of factors, including actual gas velocity over
Correlation
or Slightly Deviated Pipes
All other flow: Original
the entire wellbore from bottom to top, pipe roughness,
(Modified Hagedorn and Brown,
as in Brown, 1977)
Hagedorn and Brown (1965) liquid density, well deviation, gas pressure/density and
Correlation
how steady-state the inflow of each component really is.
Multi-Phase Flow in Highly Beggs and Brill (1973)
Deviated and Horizontal Pipes Correlation Variations in one of more of the conditions can make the
well load up at certain times of operation and unload at
It is fundamentally important that when a others. Although the aforementioned references provide
completion engineer is selecting the proper tubing a starting place for measurement, surveillance is required
size for a completion, he should investigate various to make sure the well unloads.
flow scenarios based on projected flow conditions
3000
throughout the life of the well. Fortunately, nodal 4.5" (3.958" ID)
analysis software can automatically select and 2500 3.5" (2.992" ID)
2.875" (2.441" ID)
switch between multi-phase flow correlations based 2.375" (1.995" ID)
Gas Rate, Mscf/D
2000
2.0675" (1.751" ID)
on calculated velocities, GLRs, wellbore trajectory
1500
and predicted flow regimes. Software such as this
1000
can facilitate performing sensitivity scenarios
with ease, without the need to manually switch 500
between correlations that may or may not be fully 0

0 100 200 300 400 500
understood by the person using them.
Flowing Pressure, psi
Nodal analysis solutions should be frequently
compared with actual well production. Significant
Figure 5-11 Turner unloading rates for water at
differences between the calculated and actual results pressures greater than 1000 psi (after J.F. Lea, personal
can indicate the need to modify the correlations communication with authors)
187
Modern Fracturing
Table 5-4 Common Lift Systems for Deliquifying Gas Wells

Lift System Comments
Beam Lift / Rod Pump Cannot tolerate solids, must vent gas up annulus. 50% efficiency. Removes all liquid head.
Handles large amounts of water, requires instrumentation for optimum lift and long life. Must vent
ESP (electrical submersible pump)
gas up annulus. 40% efficient. Removes most liquid head.
Simple system, wireline valve changes, can tolerate solids and any gas content. Reduces liquid
Gas Lift - continuous
head to about 4 to 5 lb/gal. 18% efficient.
Single-point lift, usually thru CT or down annulus with no packer/punched hole. Mostly for well
Gas Lift - intermittent
kick-off, 12% efficient.
A free-traveling plunger, pushing water above the plunger with gas pressure beneath. Requires
Plungers
wellbore setup but can operate automatically with instrumentation.
A water foamer applied by stick or liquid drop. Works well for small water volumes but requires
Soap (foam)
monitoring and surface break of foam. Unstable with hydrocarbon condensates.
Compression Reduces the back pressure of gas column, increases gas velocity in the string.
A short hang-off string from the tail-pipe to extend end-of-tubing to the liquid level in large casing
Siphon String
below tubing.
A tubing size reducing path such as 1 in CT inside the regular tubing. Usually flows up the
Velocity String
CT/tubing annulus. High friction for flow up the CT.
5-4.7 Liquid Hold-up and Back Pressure 5-4.8 Lift Options for Gas Wells
The effect of liquid loading in a low-pressure flowing Lift systems for gas wells are mostly based around liquid
gas well has been seen to curtail production by recovery as the gas flow rate falls below that needed for
as much as 80% and can easily kill the well if not liquid lift. In high-gas-content wells, gas lift is usually the
removed. Liquids may hold a back-pressure on gas- method of choice for larger liquid volumes and solids
producing formations in several ways, all of which can production, particularly offshore or where electrical
be minimized by well design or redesign. power is not available to the well site. Any lift system can
The first issue is in minimizing liquid hold-up along be used in a high-gas-content well, but some systems such
the wellbore by selecting the tubulars (or the lift system as beam-lift and electrical submersible pumps (ESPs) can
in some cases) to operate above the critical lift velocity in become gas locked if the well does not have adequate gas
all sections of the well. This can be difficult as the well diffusers around the pump (Lea and Nickens, 2004).
becomes more deviated because the minimum velocity In the lowest-liquid-volume gas wells, plungers are
lift equations are for vertical flow and, even then, are only used to unload small amounts of liquids. Additional lift
approximations. Reports of liquid loading in gas wells methods for these wells include foaming surfactants,
operated above the minimum critical rate are common. compression, jet pumps, and hybrid systems.
Surveillance is required to optimize flow. Table 5-4 is a general guide of the lift systems
The second issue is well path along the horizontal available for gas wells that require assistance to lift
sections. Wellbores that rise and fall create pockets of liquids to surface.
liquids that act as traps, holding back pressures and
contributing to slug development in flowing gas wells. 5-5 Completion Design
For any gas well that flows at less than full turbulence for Flow Assurance
along the wellbore, a single area of liquid collection for
lift system intake is needed. In most cases, the preferential 5-5.1 Completion Design for
location for this collection point is at the toe of the the Prevention of Gas Hydrates
well. This toe location helps prevent slug development
in deviated wells but requires consideration of how to Gas hydrates in the production stream of a gas well can
lift the liquid. Velocity and siphon strings, lift systems, pose a significant threat to flow assurance if measures to
foaming surfactants and operational adjustments can be prevent hydrate formation are not adequately addressed
effective if matched to well needs. during the completion phase of the well. Gas hydrates are
188
crystals formed with water molecules around a host gas and glycol can prevent the formation of hydrates and
molecule. The potential for gas hydrate formation exists remove hydrates that have already been formed. Large
anywhere gas is being produced in the presence of water. volumes of these types of inhibitors are often required,
However, wells that will be exposed to high pressures and which can lead to logistical difficulties, especially when
low temperatures (such as those drilled and completed in required offshore. Low-dose kinetic hydrate inhibitors
deep water or artic environments) have the highest risk (LDHIs) greatly reduce the volume of inhibitor required.
for hydrate formation. It is therefore critically important However, they are considerably more expensive and
to understand the steady state and transient, temperature cannot remove a hydrate plug if one forms. More
and pressure profiles within the well during production, recently, the use of hybrid hydrate inhibitors, a synergistic
shut-in and start up operations. combination of thermodynamic inhibitors and LDHIs,
Based on a known gas composition, a hydrate is becoming popular. They can greatly reduce the volume
equilibrium curve can be generated, which should of thermodynamic inhibitors required while providing
become the fundamental tool for a completion or flow superior hydrate prevention and the ability to remove a
assurance engineer to design a hydrate-free completion hydrate plug (Szymczak et al., 2005).
(for example, see Fig. 5-12.). Equilibrium curves show the The second method of hydrate prevention involves
thermodynamic stability of hydrates at various pressure maintaining the flowing well temperature at a high
and temperature conditions. If the flow conditions at a enough level to prevent the gas from entering the hydrate
given point in a system fall in the region to the left of formation area of an equilibrium curve. This can be
the curve, the system is at risk for hydrate formation. achieved by installing vacuum-insulated production
However, if the operating conditions fall in the region tubing or by using specially designed thermal insulating
to the right of the curve, hydrates will be unstable and packer fluids to minimize the heat loss to the annulus or
therefore unlikely to form. The term subcooling is surrounding environment. Although vacuum-insulated
often used as a measure for the potential for hydrates tubing can be very effective at minimizing heat loss, it
to form and is by definition the temperature difference is extremely expensive and steps must be taken to ensure
between the operating conditions and the equilibrium excessive heat loss will not occur at the connections.
curve. As the value for subcooling increases, so does the Insulated tubing can also restrict annular clearance and
potential for hydrate formation. lead to overly complicated completion string designs.
Due to the expense and other complicating factors
7000
pertaining to the use of vacuum insulated tubing, fit-
Stable Hydrate
6000
Region
for-purpose thermal insulating packer fluids have been
5000 gaining favor. These versatile fluids can be used in any
number of annular spaces including risers and concentric
Pressure, psia
4000 Shut-In
Conditions casing annuli. Commercially available insulating fluids
@ Wellhead 34 F Subcooling
3000
Hydrate Free are available in both aqueous and oil-base forms, although
Region
2000 increasing environmental regulations and concern about
1000
long-term compatibility with elastomers is limiting the
applications for oil-base fluids. Aqueous-based insulating
0
30 40 50 60 70 80 90 fluids developed over the past several years have proven
Temperature, F to reduce heat loss due to conduction and convection by
Figure 5-12 Hydrate Equilibrium Curve more than 90% compared to conventional packer fluids.
Extensive research has been performed with these fluids
There are two methods of preventing gas to better understand the critical relationships between
hydrate formation. The first is to chemically treat the the fluid viscosity and intrinsic thermal conductivity,
production stream with thermodynamic and/or kinetic which is essential to properly design the correct fluid
inhibitors that hinder the formation of hydrate crystals. for a given application (Javora et al., 2006a, 2006b; and
Thermodynamic inhibitors such as methanol, ethanol Wang et al., 2006a and 2006b).
189
Modern Fracturing
It is important to note that in addition to providing Production damage may comprise one or
insulating properties, these fluids must be able to pass all more of many factors including the hydrates
the other criteria required from a completion fluid such as covered in Section 5-5.1. Scale development from
an adequate true crystallization temperature (TCT) and connate waters, clay damage from encroaching
pressure crystallization temperature (PCT), compatibility waters and water blocking of pores are also common
with all formation fluids, compatibility with formation but are difficult to control in the completion
mineralogy, prevention of tubular corrosion including design, aside from limiting the commingling of
stress corrosion cracking, and compatibility with control zones with incompatible waters or controlling
line fluids.. A full description of all the compatibility injection conformance or water leaks. The type
tests required is not within the scope of this text. Instead, of damage is often predictable but its effect on
the reader is referred to Javora et al. (2006a and 2006b). production can vary widely from well to well. For a
Nonetheless, a full suite of compatibility tests should be damage mechanism to be a problem, it must be in
performed to eliminate the potential for catastrophic a form and a location that will create a dominant
completion problems during the life of a well. restriction to flow. Even significant damage to
the permeability of the walls of a fracture, for
5-5.2 Formation Damage in Gas Wells, example, is usually not a restriction due to the
Completion Damage and Scales enormous area of the fracture wall. However, even
a small amount of damage in a frac-pack or in the
Damage problems in gas wells may directly or indirectly near-wellbore inflow area in an unfractured well or
relate to one or more factors from the design of the well, in the outflow path of the tubing can create a back
the production of fluids from the reservoir, or both. The pressure that significantly decreases production.
damage may be in the formation, but often the most Some damage mechanics can produce skin, but the
severe damage for gas wells is in the perforations or well may be restricted by other factors such as tubing
the tubing with flow path restrictions at choke points flow limits, facility limits, etc.
having the greatest effect. Many formation damage and
flow assurance issues have their roots in a completion 5-5.3 Organic Deposits and Condensate Banking
design that did not consider how to effectively produce
the well through its life. Many initial drilling and Deposits such as paraffin and asphaltenes, common
completion damage mechanisms are substantially the to oil wells, are much less common in gas wells
same as in oil wells: mud and cement filtrate damage, unless there is associated oil or condensate. Dry
whole mud losses and poor perforations (Tiffin et asphaltene accumulations in the choke of gas wells
al., 2001; and Sanchez et al., 2004). These initial and paraffin deposits stranded by condensate flash-
damage mechanisms are often bypassed by either deep off are occasionally seen. More frequent forms of
perforations or fracturing stimulations. In addition to production damage in a gas well are often related
simply bypassing existing formation damage, a properly to relative permeability effects, in addition to the
placed fracturing treatment will reduce the near-wellbore water blocks described earlier. Reservoirs that
pressure drop during production (Rae et al., 1999) due are under-saturated with liquids may experience
to the associated, bilinear or pseudo-radial flow pattern development of a liquid phase in previously dry
as opposed to a radial flow pattern (see Section 4-7.5). pore space as pressure passes through the dew point.
The reduction in near-wellbore pressure drop can in After the liquid precipitates, there is less room for
itself prevent many types of damage such as inorganic the gas to flow and a lower permeability, sometimes
scale, organic deposition and fines migration. In some less than 20% of initial when the pores are small
cases formation sand control can be achieved. Reducing and permeability is low. Fracturing to add effective
or eliminating rate-dependent skin caused by non-Darcy wellbore-to-reservoir contact area is one of the few
flow effects in high-permeability, high-rate gas wells is workable approaches to controlling the condensate
covered in detail in Sections 2-4 and 2-5. precipitation damage effect.
190
5-5.4 Effects of H2S and CO2 on Corrosion The general corrosion caused by the presence of
CO2 in a wet gas is due to the dissolution of CO2 in the
Corrosion control is a fundamental responsibility formation brine, which forms carbonic acid (H2CO3)
of a completion or flow assurance engineer. The with an intrinsic pH value of around 3.5. Although in
knowledge of the fluid composition to be produced comparison to acids used in stimulation treatments the
from a given field is of paramount importance and will pH of carbonic acid is quite mild, it can increase the
dictate the material selection and required corrosion corrosion rate of carbon steel up to several mm/year
control program for the wellbore completions, without corrosion control measures.
flowlines and production facilities. The corrosive The presence of H2S in a wet gas will also accelerate
nature of a fluid will depend on factors such as the general corrosion rate of carbon steel and certain
temperature, salinity, pH and of most importance, corrosion-resistant alloys (CRAs), but more importantly
the presence of carbon dioxide (CO2) and hydrogen the presence of uncontrolled H2S can lead to sulfide
sulfide (H2S). The presence of CO2 and/or H2S in stress corrosion cracking (SCC). SCC is considered one
a wet gas stream can cause catastrophic corrosion of the worst forms of corrosion within the industry and
problems if not addressed early in field development can result in the splitting of tubulars and catastrophic
(Figs. 5-13a and 5-13b). failures of completion components within short periods
of time. Financial losses within the oil and gas industry
related to unplanned re-completions as a result of SCC
are in the hundreds of millions of dollars per year;
therefore, every effort to prevent the occurrence should
be made (Kermani and Harrop, 1995).
In general two basic types of corrosion prevention
methods are suitable for oil and gas wells. The first
is the use of corrosion inhibitors specially formulated
to protect carbon steel from corrosives. The second
is to replace carbon steel tubulars and components
with CRAs compatible with the anticipated fluid
composition. On occasion a combination of both
has been used: CRA components with carbon steel
components plus an inhibitor program. When
Figure 5-13a CO2 corrosion on the outside of a tubing
string suitable, the use of carbon steel materials and a fit-
for-purpose inhibitor program will provide a more
economic solution for corrosion prevention than
the utilization of extremely costly CRA materials.
Nonetheless, extremely corrosive fluid compositions
or problematic logistical issues may prevent the
frequent or continuous application of inhibitors and
necessitate the use of the expensive CRA materials to
provide the required corrosion control.
It is important to note that presently there are a
large number of corrosion resistant alloys commercially
available within the industry. Chrome 13, Super
Chrome 13 and Chrome Duplex 22 are all commonly
used grades of materials used in the industry today.
Each alloy has characteristics suited for specific fluid
Figure 5-13b Crevice corrosion at a coupling compositions and strength requirements. However,
191
Modern Fracturing
not all materials with the same grade perform in the Gas well operations to minimize sand movement
same fashion (Stevens et al., 2004). For instance, focus on preventing high flux rates to control erosion,
Chrome 13 from one manufacturer may have a and limiting high instantaneous production rate loads
slightly different metallurgic composition and yield during startup to limit screen or gravel pack invasion by
strength when compared to Chrome 13 material from fines. The fluids in the rock pores are an active support
another manufacturer. As a result of the expansive of the overburden pressures created by the upper strata
number of materials entering the market, a vast (see Section 4-3.2.3). As the fluids are removed and
amount of corrosion research is occurring within the pressure is drawn down, the formation matrix must
service and operator sectors of the industry. In short, a support an increasing amount of the overburden load.
completion engineer should have a detailed knowledge Formations that were stable even at maximum rates
of the metallurgic composition, mechanical properties in initial production can fail as pressure enters the
and corrosion resistance suitability for any material final phase of depletion. Sand control is often added
selected for a given application. to a design to offset the loads that will appear only in
In the end economics will decide the appropriate the later stages of life. This may be a wise or a foolish
corrosion prevention method. Regardless of the method decision, depending on the operators expectancy of
selected, a comprehensive corrosion-monitoring program economic well life, the method of well operation, the
must accompany a corrosion-prevention program to ease of well entry to add sand control and the potential
identify the actual corrosion rates during production for significant production in late well life.
in comparison to what was predicted. Obviously large
variations in the actual corrosion rate as compared to the 5-6.2 Is Sand Flow All Bad?
predicted rate will dictate the need for an adjustment to
the corrosion prevention program. Sand control for gas wells is a critical concern that will
grow in importance as exploration is increasingly driven
5-6 Sand Control for Gas Wells into more complex reservoirs such as coals and shales,
and into younger sand reservoirs. The best approach to
5-6.1 Why is the Sand Flowing? sand control, from a production standpoint, is to achieve
the necessary sand control without adding mechanical
The first necessity in design of a completion in a weak obstructions such as gravel and screens. Although effective
or unconsolidated sand formation is the need to predict in preventing sand movement, gravel and screens create
the strength of the formation (Vaziri et al., 2006; and a near-wellbore pressure drop and a filter that will plug
Nouri et al., 2006). Sand grains can move due to many over time. Production skins in sand control completions
forces: coupled drag on the grains by the moving fluid, typically range from a low of 0 to 3 for open-hole gravel
relative permeability forces (a related form of drag) packs and frac-packs, to skins of 10 to 15+ for cased-
when the fluid phase changes, lubricated shear failure hole gravel packs with gravel interface outside the casing
(bed fluidization) and spalling of the grains due to free (King et al., 2000). In contrast, a completion that can
face loads. These are not independent causes and this allow some sand to produce in a formation with a low
list is not exhaustive. Fine particles (defined here as but stable matrix strength, can create a cavity completion
smaller than 44-micron or 325 mesh) may be released with accompanying skin of about 1 (Palmer et al., 2005).
after solvent or surfactant contact,, with disaggregation, The drawback to creating cavities is the need to manage
or with drag from flowing fluids. Independently, some sand erosion, collection and disposal. Cavities may also
clays, such as smectite, can disperse fines when the brine collapse with time as overburden stresses increase. When
character changes. In at least six phases of operation, considering sand control for a gas pay, it is worth the effort
pressures, loads or drag forces strongly influence to investigate the factors that cause sand movement in the
formation strength: drilling and completion; clean-up; specific pay and determine if the well can be produced
initial production at maximum rates; shut-down, start- with a sand management strategy (Balgobin, 2003). The
up; and production at pressure depletion. best candidates for such a strategy are usually on-shore
192
gas wells with little or no oil residuals and an unconfined considerably less than that from a well that is operated
compressive strength range between perhaps 1000 and with guidance updated by sand flow monitors on or in
1500 psi. The cavity created by such a sand management the flow lines (Stein et al., 2005).
strategy is not easy to maintain, and shock loads of sudden In general, sand movement prediction in gas
shut-downs and start-ups must be avoided. In any weak well completion design is used decide if and when
sand completion without active sand control (screen sand control will be needed. However, the gas well
and/or screen and gravel pack), there will be an incidence operation strongly affects whether or not sand will be
of sand production at any event (shut-in, start-up, choke produced. Effective wellbore design and consistent
change, etc.) that disturbs the flowing equilibrium. The operation cannot be separated if the maximum
amount of sand produced at the onset of such an event possible economic recovery is to be achieved. The most
will almost always diminish with time if a stable cavity can effective sand control for an application will depend
be created. If the sand is too weak to maintain a cavity, the as much upon the way the well will be operated as it
perforation tunnels will collapse or fill with sand and the does on formation characteristics.
resultant skins may be very high. Accuracy in predicting sand strength and thus the
sand-free rate over the life of the well is difficult. This is
5-6.3 Establishing and because of the many variables involved with formation
Monitoringa Sand-Free Rate sand movement in a gas well. Sand bonding strength
(specifically, that part of the formation matrix strength
The conventional method of predicting a sand-free that deals with grain-to-grain bonding) is affected by
rate (e.g., using increasing choke settings until sand cohesion (fluid-grain), grain-to-grain cementation,
is seen at the surface) is giving way to predictive sand moving and non-moving fluid phases in the pores, shut-
movement methods that offer better production rates down and start-up surges, production drawdown and
at less expense. The potential for sand production peak flow from high-permeability intervals. The accuracy
changes with time; it is only a constant in the strongest of predictions of whether and when the sand grains
and the very weakest formations. Where a sand free rate will move is becoming steadily better, but there is no
is needed, varying the choke size can usually establish totally accurate solution. Sand production monitoring is
a flow rate that is free of sand. However, this may be required for optimized production operations.
A B C D E
Figure 5-14 Sand control completions: A) Screen Only, B) External or Open-Hole Gravel Pack, C) Internal or Cased-Hole
Gravel Pack, D) High-Rate Gravel Pack, E) Frac-Pack
193
Modern Fracturing
5-6.4 Sand Control Methods for Gas Wells Table 5-5 Reliability of Sand Control Completions for
Gas Well (after King et al., 2000))
From a formation view, the best sand control for gas Sand
Control Sand Control
producers will depend on the flow paths through the Type of Production Production
Population
reservoir (Fig. 5-14). In a high-permeability, thick, Completion Failures, Failures,
percent of failures/well/yr
blanket sand reservoir with few vertical permeability attempts
barriers, a open-hole gravel pack will probably be the Screen Only,
best completion (Bennett et al., 2000; Claiborne et Cased and
44 18% 0.048
Perforated
al., 2002; and Parlar et al., 2000), particularly when Completion
fracturing is not feasible for other reasons. However, Screen Only,
in a laminated, lower-permeability sand (< 100 to Open Hole 206 13% 0.042
Completion
perhaps about 250 md, depending on damage and Cased Hole
213 5.40% 0.013
kh/), a tip screen-out frac-pack may be favored Gravel Pack
(Norman, 2003; and Mullen et al., 1994) (see Open Hole
388 4.8%* 0.016*
Gravel Pack
Sections 4-7.3.2 and 4-7.4). In smaller hydrocarbon
High Rate
deposits where the cost of a fracture treatment is Gravel Pack
208 2.70% 0.009
uneconomic or the formation character discourages Tip Screenout
845 2.00% 0.005
an open-hole completion, a screen-only completion Frac Pack
or a conventional gravel pack may be preferred. In *Results over the past 100 OHGP show a failure rate of less
than 0.010 failures/well/yr.
general, a screen-only completion should be used in
an open-hole where the formation sand can cave in Reservoir characteristics such as laminations,
around the screen for erosion protection. A cased- deviation, permeability and potential for subsidence
hole gravel pack should be reserved for lower-rate influence the sand control completion, often
completions where the gravel-packed perforations destructively. For longevity, frac-packs, high-rate
(with additional gravel-to-formation sand interface) water packs (at fracturing pressure) and open-hole
will not be a limiting inflow factor. gravel packs show excellent results. Several wells in the
Although horizontal wells are popular for oil database have recorded more than 30 million barrels of
formations and some gas plays, a fracture offers more fluids produced and many billions of cubic feet of gas
formation contact. Single and, in some cases, multiple production (King et al., 2000).
fractures should be considered in combination with
horizontal well designs to maximize reservoir contact 5-6.6 Repairing and Restoring
and minimize the non-Darcy flow effects. See Sections Productivity in Wells hat Flow Sand
2-5.3 and Chapter 10.
Repairing a gas well completion that has flowed sand
5-6.5 Reliability of Sand Control Completions may be very difficult, especially if equipment in the
well has been damaged by erosion. Erosion sharply
The historical reliability for sand control completions reduces the strength of the reservoir, well tubulars,
for gas wells are summarized in Table 5-5 (King et al., wellhead and valves. The causes of sand control
2000). The failures vary with application but several failures in gas wells fall into a few basic categories,
trends can be drawn from the data from the 2000 many of which are related (King et al., 2000;
wells included in the failure database represented by Hamid and Ali, 1997):
the paper. Complex sand control completions, such 1. Screen plugging and hot spot development
as frac-packs and open-hole gravel packs have more 2. Gravel pack plugging by invasion of formation
early-time failures but many fewer later-time failures sand
than the simplest completions of screen-only and 3. Screen erosion from excessive production rate
cased-hole gravel packs. flux loading
194
4. Straight erosion of screens from moving formation Flow in Inclined Pipes, JPT, pp607-617, May
sand or fluidized gravel where voids were not filled 1973.
during packing operations or where inadequate Bennett, C., Gilchrist, J.M., Pitoni, E., Petit, G.,
gravel reserve above the top perfs is a problem Burton, R.C., Hodge, R.M., Troncoso, J., Ali,
5. Mechanical failure from loads during production S.A., Dickeron, R., Price-Smith, C., and Palar, M.:
(common with some plugging) Design Methodology for Selection of Horizontal
6. Subsidence or earth shift (tectonic movements) Open-Hole Sand Control Completions Supported
The first five involve a breach or plugging- by Field Case Histories, SPE 65140, 2000.
induced failure of the screen. Repair methods Bennion, D.B., Thomas, F.B., and Ma, T.: Formation
in these cases are always difficult and rarely Damage Processes Reducing Productivity of Low
long-lived. In the short term, insert screens Permeability Reservoirs, SPE 60325, 2000a.
and vent screens are the preferred method. In a Bennion D.B., Thomas, F.B., Imer, D., and Ma, T.:
very few cases the lower well may be re-drilled Low Permeability Gas Reservoirs and Formation
or the completion may be pulled on very rare Damage-Tricks and Traps, SPE 59753, 2000b.
occasions. Historically, many re-completion Bradley, H.B., Ed.: Petroleum Engineering Handbook,
attempts last less than two to three years SPE, Richardson, Texas, 1987.
(King et al., 2000). Sand control failure due to Brill, J.P. and Beggs, H.D.: Two-Phase Flow in Pipes,
subsidence is a problem without a good solution University of Tulsa, 1978.
or prevention method. Subsidence forces are so Brown, K.E, Overview of Artificial Lift Systems, SPE
large that the ability to resist the load is nearly 9979, JPT, pp2384-2396, October 1983.
impossible in the severe instances of subsidence. Brown, K.E.: The Technology of Artificial Lift Methods,
Pressure maintenance in the reservoir is a 1, Pennwell Books, Tulsa, OK, 1977.
common prevention, but the success depends on Carter, L.G., Cook, C. and Snelson, L.: Cementing
both elimination of the created void spaces and Research in Directional Gas Well Completions,
movement of the pressure supporting material, SPE 4313, 1973.
usually water, to the right places without undue Coleman, S.B., Clay, H.B., McCurdy, D.G., Norris
cycling of injected water. III, L.H.: A New Look at Predicting Gas Well
New techniques for repair, involving new resins, Loading-Up,, Trans. AIME, 291, 329, JPT, March
creation of secondary matrix strength, and limited 1991.
withdrawal rates are being researched. Claiborne, E.B., Malone, B.P., Marshall, J.C.: Ceiba
Completion Optimization: A Fast-Track Approach
to Success, SPE 77434, 2002.
di Lullo, G., and Rae, P.: Cements for Long Term
References: Isolation Design Optimization by Computer
Modelling and Prediction, paper IADC/SPE
Baksh, K.: The Value of Permanent Downhole 62745, 2000.
Pressure Surveillance in the Amherstia/Parang Dolle, N., Singh, P., Turner, R., Woodward, M.,
Field, Offshore Trinidad, SPE 94805, 2005. and Paino, W.F.: Gas Management, Reservoir
Balgobin, C.J.: Sand Management of Ultra-High- Surveillance, and Smart Wells An Integrated
Rate Gas Wells, SPE 948946, 2003. Solution for the Bugan Field, SPE 96429, 2005.
Bamforth, S., Beeson, C., Stephenson, K., Whittaker, Duns, H., Jr. and Ros, N.C.J.: Vertical Flow of Gas
C., Brown, G., Catala, G., Rouault, G., Theron, B., and Liquid Mixtures in Wells, Proc. Sixth World
Conort, G., Lenn, C., and Roscoe, B.: Revitalizing Pet. Cong., Frankfurt, 2, Paper 22, 1963.
Production Logging, Oilfield Review, Winter Economides, M.J.: Ch 1 in Well Cementing, Nelson,
1996, p44-60. D. Ed., Schlumberger Educational Services,
Beggs, H.D. and Brill, J.P.: A Study of Two-Phase Houston, 1990
195
Modern Fracturing
Economides, MJ., Ehlig-Economides, C.A., and Hill, Corrosion on the Oil and Gas Industry, SPE
A.D.: Petroleum Production Systems, Prentice 29784, 1996.
Hall, Upper Saddle River, NJ, 1994. King, G.E., Wildt, P.J., and OConnell, E.: Sand
Fox, P.E., Adnyiana, G., and Setiadi, I.: Applications Control Completion Reliability and Failure
of Carbon/Oxygen Logging in Indonesian Rate Comparison With a Multi-Thousand Well
Reservoirs, SPE 54353, 1999. Database, SPE 84262, 2000.
Griffith, G.A. and Wallis, G.B.: Two-Phase Slug Lea, J.F., and Nickens, H.V.: Solving Gas-Well Liquid-
Flow, J. Heat Transfer, pp307-320, August 1961. Loading Problems, SPE 72092, JPT, pp30-36,
Guo, B., Ghalambor, A., and Xu, C.: A Systematic April 2004.
Approach to Predicting Liquid Loading in Gas Lea, J.F, and Tighe, R.E.: Gas Well Operation with
Wells, SPE 94091, 2005. Liquid Production, SPE 11583, 1983.
Gray, K.E., Podnos, E., and Becker, E.: Finite Li, X., Mitchum, F.L., Bruno, M., Patillo, P.D., and
Element Studies of Near-Wellbore Region During Willson, S.M.: Compaction, Subsidence, and
Cementing Operations: Part I, SPE 106998, Associated Casing Damage and Well Failure
2007. Assessment for Gulf of Mexico Shelf Matagorda
Hagedorn, A.R., and Brown, K.E.: Experimental Island 623 Field, SPESPE 84553, 2003.
Study of Pressure Gradients Occurring During Mullen, M.E., Stewart, B.R., and Norman, W.D.:
Continuous Two-Phase Flow in Small Diameter Soft Rock Frac-Pack Completions in the Gulf of
Vertical Conduits, JPT (April 1965) 475-484. Mexico, SPESPE 28532, 1994.
Hamid, S., and Ali, S.A.: Causes of Sand Control Mueller, D.T.: Producing Stress-Resistant High-
Screen Failures and Their Remedies, SPE 38190, Temperature / High-Pressure Cement Formulations
1997. Through Microstructural Optimization, SPESPE
Heinold, T., Dillenbeck, R.L., and Rogers, M.J.: The 84562, 2003.
Effect of Key Cement Additives on the Mechanical Mueller, D.T., GoBoncan, V., Dillenbeck, R.L., and
Properties of Normal Density Oil and Gas Well Heinold, T.: Characterization Casing-Cement-
Cement Systems, SPE 77867, 2002. Formation Interactions Under Stress Conditions:
Holstein, E.D., and Berger, A.R.: Measuring the Impact on Long-Term Zonal Isolation, SPESPE
Quality of a Reservoir Management Program, 90450, 2004.
JPT, pp 52, January 1997. Norman, D.,: The Frac-Pack Completion: Why Has It
Javora, P.H., Berry, S.L., Stevens, R.F., Carpenter, J.F., Become the Standard Strategy For Sand Control?
Isaac, D.D., Dalton, F.G., Augsburger, J., Guy- SPESPE 101511, SPE Distinguished Lecturer
Caffey, J.K., Tomlin, M., Jones, T.A., Malachosky, Presentation, 2003.
E., Prasek, B.A., Foxenberg, B., Freeman, M.A.,. Nouri, A., Vaziri, H., Belhaj, H., and Islam, R.: Sand
Howard, S., Benton, W., Son, A., Eichelberger, Production Prediction: A New Set of Criteria
P., and Stark, C.: A New Technical Standard for for Modeling Based on Large-Scale Transient
Testing of Heavy Brines, SPE 98398, 2006a. Experiments and Numerical Investigation, SPE
Javora, P.H., Stevens, R., Devine, C., Jeu, S., Simmons, 90273, SPE Journal, 11:2, June 2006.
M., Firmin, G., Poole, G., Franklin, B., and Qu, Oberwinkler, C., and Standner, M.: From Real-Time
Q.: Deepwater Completion Challenges Redefine Data to Production Optimization, SPEPF, p229,
Best Practices for Completion and Packer Fluid August 2005.
Selection, SPE 103209, 2006b. Palmer, I., Vorpahl, D.G., Glenn, J.M., Vaziri, H., and
Julian, J.Y., King G.E., Robertson, D.B., Johns, J.E., McLennan, J.: A Recent Gulf of Mexico Cavity
and Sack, J.K.: Detecting Ultra-Small Leaks With Completion, SPE 86462, SPEDC, 20, September
Ultrasonic Leak Detection Case Histories from 2005.
the North Slope, Alaska, SPE 108906, 2007. Parcevaux, P., Rae, P., and Drecq, P.: Well Cementing,
Kermani, M.B., and Harrop, D.: The Impact of Schlumberger Educational Services, 1990.
196
Parlar, M., Bennett, C., Gilchrist, J., Elliott, F., Sinha, S.P., and Al-Qattan, R.: A Novel Approach
Troncoso, J., Price-Smith, C., Brady, M., Tibbles, to Reservoir Surveillance Planning, SPE 88792,
R.J., Hoxha, B., and Foxenberg, W.E.: Emerging 2004.
Techniques in Gravel Packing High Performance Stein, M., Chitale, A.A., Asher, G., Vaziri, H., Sun, Y.,
Completions in High Performance Wells, SPE Colbert, J.R., and Gonzalez, F.A.: Integrated Sand
64412, 2000. and Erosion Alarming on NaKika, Deepwater Gulf
Pattillo, P.D., and Huong, N.C.: The Effect of Axial of Mexico, SPE 95516, 2005.
Load on Casing Collapse, JPT, pp159, January Stevens, R., Ke, M., Javora, P.H., Qu, Q.: Oilfield
1982. Environment-Induced Stress Corrosion Cracking
Pattillo, P.D., and Kristiansen, T.G.: Analysis of of CRAs in Completion Brines, SPE 90188,
Horizontal Casing Integrity in the Valhall Field, 2004.
SPE/ISRM 78204, 2002. Szymczak, S., Sanders, K., Pakulski, M., and Higgins,
Pattillo, P.D., Kristiansen, T.G., Sund, G.V., and T.: Chemical Compromise: A Thermodynamic
Kjelstadli, R.M.: Reservoir Compaction and and Low-Dose Hydrate-Inhibitor Solution for
Seaflow Subsidence at Valhall, SPE 47274, 1996. Hydrate Control in the Gulf of Mexico, SPE
Purvis, D.L., Mueller, D.T., Dawson, J.T. and Bray. 96418, 2005.
W.S.: Thickening Time Test Apparatus Provides Taitel, Y., Barnea, D. and Dukler, A.E.: Modeling Flow
Method of Simulating Actual Shear History of Oil Pattern Gradients Occurring During Continuous
Well Cements, SPE 26576, 1993. Two-Phase Flow in Small-Diameter Vertical
Rae. P., Martin. A.N. and Sinanan, B.: Skin Bypass Conduits, AIChE J., 26 (6) 345-354, May 1980.
Fracs: Proof that Size is Not Important, SPE Tewari, R.D., Abd Raub, M.R., Omar, M.I., Fenghan,
56473, 1999. B., Moris, M., Jelai, J., Ramachandran, S.,
Rogers, M.J., Dillenbeck, R.L., and Eid, R.N.: Fooks, A.L., Peden, J.M., and Montague, E.:
Transition Time of cement slurries, definitions The Importance of Well Construction and Well
and Misconceptions, related to Annular Fluid Integrity for Reservoir Management A Mature
Migration, SPE 90829, 2004. Field Experience in Sudan, SPE 100813, 2006.
Sabins, F.L., and Sutton, D.L.: The Relationship of Thiercelin, M.J., Dargaud, B., Baret, J.F., and
Thickening Time, Gel Strength, and Compressive Rodriquez, W.J.: Cement Design Based on
Strength of Oilwell Cements, SPEPE, pp143, Cement Mechanical Response, SPE 52890,
March 1986. SPEDC, pp 266, December 1998.
Sabins, F.L., Tinsley, J.M., and Sutton, D.L.: Transition Tiffin, D., Stevens, B., Park, E., Elliott, F., and Gilchrist,
Time of Cement Slurries Between the Fluid and Set J.: Evaluation of a Filter Cake Flowback in Sand
States, SPE 9285, SPE J., pp875-882, December Control Completions, SPE 68933, 2001.
1982. Turner, R.G., Hubbard, M.G., and Dukler, A.E.:
Sanchez, E., Audibert-Hayet, A. , and Rousseau, L.: Analysis and Prediction of Minimum Flow Rate
Influence of Drill-in Fluids Composition on for the Continuous Removal of Liquids from
Formation Damage, SPE Journal (December Gas Wells, JPT (Nov 1969), Trans. AIME, 246,
2004) p403. 1475.
Sauer, C.W.: Mud Displacement During Cementing Vargo, R.F., Payne, M., Faul, R., and LeBlanc, J.:
State of the Art, SPE 14197, JPT, pp1091-1101, Practical and Successful Prevention of Annular
September 1987. Pressure Buildup on the Marlin Project, SPE
Scrimgeor, J.A., King, P.A., and Kelman, J.F.: Field 77473, 2002.
Examples of Production-Logging Problems in Vaziri, H., Allam, R., Kidd, G., Bennett, C., Grose, T.,
Large Diameter Deviated Wells with BiPhasic Flow Robinson, P., and Malyn, J.: Sanding: A Rigorous
and Their Implications on Reservoir Management, Examination of the Interplay Between Drawdown,
SPE 11899, 1983. Depletion, Startup Frequency, and Water Cut,
197
Modern Fracturing
SPE 89895, SPE Prod. Ops., 21, Nov. 2006.

Wang, X., Qu, Q., Stevens, R., and Javora, P.: Factors
Controlling the Proper Design of Effective
Insulation Fluids for Oilfield Applications, SPE
103132, 2006a.
Wang, X., Qu, Q., Javora, P.H., and Pearcy, R.: New
Trend in Oilfield Flow Assurance Management:
A Review of Thermal Insulating Fluids, SPE
103829, 2006b.
Yamamoto, H., Christiansin, R.L.: Enhancing Liquid
Lift From Low Pressure Gas Reservoirs, SPE
55625, 1999.
198
Barry Hlidek is stimulation engineering manager for BJ Services Company Canada in Calgary, responsible
for fracturing and acidizing engineering services, including treatment design and computer simulations.
With more than 19 years of experience in the oil service industry, Hlidek's areas of interest include research
and engineering for fracturing, cementing, acidizing, sand control and coiled tubing services. Initially hired
by the Western Co. as a field engineer, Hlidek spent 11 years as a research scientist in corporate laboratory
technical services, involved in global technical support and chemical product formulation for cementing,
fracturing and acidizing. He holds BSc degrees in chemical and petroleum refining, and mineral economics,
from the Colorado School of Mines. Hlidek is an SPE author and holds two US patents on foaming agents.
At the local level, he is chairman of the SPE Canadian Section and held multiple positions on the board
of directors for the SPE Canadian section, as well as organizational and technical committee positions for
several local SPE technical conferences, including the GTS. On the SPE International level, he is chairman of
the Production and Operations Award Committee and is a member of the Distinguished Lecturer Selection
Committee. Barry is a registered Professional Engineer in Alberta.
Leen Weijers is the Rocky Mountain regional manager for Pinnacle Technologies after eight years of focusing
on development and support of the FracproPT hydraulic fracture growth simulator. He completed his
doctoral research at Delft University of Technology in the Netherlands by conducting model experiments
to study hydraulic fracture growth in horizontal and deviated wells. He has been involved in numerous
multi-stage propped fracture stimulation efforts in horizontal wells. He was awarded The Chevron Way
Award for his involvement in stimulating several horizontal wells in the Lost Hills Field, California. Weijers
has authored numerous industry courses and publications, has conducted real-time fracture analysis on
hundreds of treatments, and plays a key role in the calibration of fracture growth models with various
fracture diagnostics, such as Pinnacle's tiltmeter and microseismic fracture mapping technologies.
Chapter 6 However, in hydraulically fractured wells, near-
wellbore damage is normally of little consequence because
Fracture-to-Well Connectivity the propped fracture will bypass this damage, providing
Barry Hlidek, BJ Services Company and wellbore connectivity from the reservoir farther from
Leen Weijers, Pinnacle Technologies the well. In all propped hydraulic fracturing treatments,
we must bypass near-wellbore damage, or treatment
objectives will likely not meet expectations (see Section
6-1 Introduction 4-7.5 for a discussion on skin-bypass fracturing).
Well inclination, and perforation orientation
Fracture-to-well connectivity: This is defined and placement will significantly impact fracture
broadly as the physical connection of a wellbore to a operations. In general, initial fracture propagation
reservoir or formation or, specifically, to a hydraulic direction is controlled by the local stresses around
fracture. The latter allows the flow of fluids either the wellbore. At some point, and within a short
from the fracture to the wellbore, or from the wellbore distance from the wellbore, formation stresses
to the fracture. The obvious example of general will dominate, and the fracture will change direction
connectivity is the production of hydrocarbons from to propagate perpendicular to the minimum
a reservoir to the wellbore for ultimate production principal stress. As discussed in Section 6-4.1, a
through surface facilities. Flow restrictions will large difference between maximum and minimum
always exist, and well completions need to be horizontal stresses (referred to as a high stress bias)
designed to minimize pressure drops, especially at the can result in a rapid change in direction, possibly
wellbore-to-formation interface. causing severe near-wellbore tortuosity. In addition
Various completion options are available, and to narrow and twisted fracture paths (pinch points),
choices depend on well objectives and physical multiple fractures are possible, or the fracture may
or mechanical constraints. Completion methods even propagate through a micro-annulus prior to
include conventional cased and cemented wellbores, re-orientation. Any of these possible events will
uncemented liners, pre-perforated liners, slotted increase near-wellbore pressure drops and the chance
liners and open-hole producing sections, among of a premature screen-out or sand-off. In some cases,
others (see Chapter 5 for a detailed analysis of well proppant and/or gel slugs early in the treatment have
completions). In all cases, a robust connection been shown to be an effective method to minimize
to the wellbore is of primary importance. In the near-wellbore complexity (Cleary et al., 1993).
conventional case of a cemented casing (or liner), However, it must be emphasized that even if fracture
perforation size, type, density, orientation and execution is facilitated, the presence of misaligned
operational procedures will have a significant impact and badly connected fractures will adversely affect
on well connectivity. It is not uncommon that subsequent well performance.
perforation strategies for production or stimulation Control of formation fines and/or proppant
(fracturing) conflict with reservoir evaluation flowback is another important consideration. Solids
objectives. In order to perform pressure transient flowback can greatly affect production, may require
analysis and other flow or pressure build-up test well intervention, may damage surface production
procedures, it may be desirable to perforate an entire equipment and may present significant safety and
net pay interval, whereas for fracturing, a limited environmental hazards. A wide variety of solids
perforated interval may be more desirable. Often, a control options are available. For some situations,
compromise must be reached. proper perforation orientation may be adequate to
Historically, a primary loss of well-to- control solids production; in others, resin-coated
formation connectivity has been attributed to proppants, proppant flow-back control additives or
near-wellbore formation damage that occurs mechanical screens may be required (see Section 8-7
during drilling and completion operations. for a discussion on proppant flowback and Section
201
Modern Fracturing
5-6 for sand control in gas wells). Screens should 6-2 Completion Techniques and Their
be a solution of last resort and, if required, it could Impact on Well Connectivity
be an indication that the stimulation treatment and
the well construction, including its orientation, In many oilfield applications, a well penetrates
were not designed properly. multiple target zones; in some fields, the pay interval
In the case of hydraulic fracturing the industry that contains hydrocarbons can be several thousand
has taken two paths. The first involves accepting feet thick. Without an effective wellbore distribution
whichever well geometry is presented and executing strategy, hydraulic fractures will typically grow in
the treatment to work around problems, no matter the layers that provide the least resistance, i.e., the
how unfavorable or inefficient the completion may zones in which the fracture initiation pressure is
be for both execution and (especially) subsequent lowest. Consequently, a stimulation treatment
production. Thus, perforation schemes, fluids to that would simultaneously target all pay zones in
be injected and the sequence of injection stages a thick formation may actually cover only a very
(such as proppant slugs, etc.), are determined based small portion of the interval. Techniques have been
on their ability to facilitate execution. Some have developed to improve fluid diversion (also referred to
taken it so far as to consider it an accomplishment as lateral fracture diversion) and vertical penetration
to be able to fracture highly unfavourable well of the fracture over the entire thickness of the
geometries, such as highly inclined wells with long target zone (see Section 9-7).
perforated intervals, etc. In the face of obvious Until recently, the objective for gas well
problems this will cause to future production, such stimulation has been to create and extend a simple,
thinking should clearly be avoided. largely planar fracture within a reservoir. However, in
A far more desirable and, indeed, far more several newly developed reservoirs, such as the Barnett
effective approach is to integrate drilling, completion, Shale (see Section 11-2), the emphasis has been on
perforating and fracturing so that wells are creating a complex fracture network because this is the
constructed with all these operations in mind a priori. only viable way to achieve economic production from
The important issues for appropriate fracture-to-well certain ultra low-permeability (naturally fractured)
connectivity are addressed in this chapter. formations (Matthews et al., 2007).
Far-field reservoir connectivity: Far-field
complexity can also limit reservoir-to-wellbore 6-2.1 Cased-Well Isolation Techniques
connectivity. The reservoir itself may present barriers
to production. Multiple layers or producing lenses Cased and lined wells have a large-diameter pipe placed
may not have vertical permeability, and alternative and cemented (see Section 5-2), providing significant
fracture treatment designs may be required to advantages over open-hole completions: first by
effectively connect the producing intervals to the protecting the up-hole layers from fluids, pressures and
wellbore. Using shale as an example, the reservoir well stability problems;secondly by isolating fresh water
matrix has microdarcy permeability, and the dominant formations;and finally, isolating zones of lost returns or
production mechanism is through natural fissures formations with significantly varied pressure gradients.
(see Section 11-5). In the case of such ultra low- In particular, the casing allows the use of mechanical
permeability fissured reservoirs, although the hydraulic isolation, and perforations in cased wells are beneficial
fracture is largely parallel to well-defined natural for mechanical and chemical diversion (see Section
fractures, treatment designs must connect the fissure 9-7).This makes cased wells very attractive for most
system to the wellbore in order to provide maximum stimulation and diversion techniques.
possible inflow area, for efficient production. Finally, Various isolation techniques can be used to
fracture diagnostics (such as microseismic, tiltmeters, temporarily isolate target zones from each other
etc.) can be utilized to help provide an understanding and from the rest of the well (Economides and
of far-field fracture complexity. Nolte, 2000).These isolated zones can be stimulated
202
Chapter 6 Fracture-to-Well Connectivity
separately or simultaneously. The first requirement across these perforations (Elbel and Britt, 2000) (see
for proper isolation is achieving a good cement Section 9-7.3). This strategy is normally implemented
bond between perforated intervals (see Section 5- in thick target zones to ensure stimulation of the entire
2).After the wellbore has been constructed, various gross pay. The high frictional pressure drop (e.g., 400
types of casing plugs for zonal isolation are available psi perforation pressure drop) is designed to offset
for conventional bottom-to-top treatment stage the stress differences between target zones to enhance
stimulation.Typically, perforations are shot in the injection into all the perforated target zones (Lagrone
casing, a fracture stimulation treatment is conducted, and Rasmussen, 1963). However, a sufficient injection
and the stimulated zone is isolated from new rate must be available to maintain this high differential
perforations that will be treated up-hole. Mechanical pressure across the perforations, both initially and as
bridge plugs can be run on tubulars or wireline the perforations are eroded by proppant.
(Brown et al., 2000). Sand plugs can also be pumped Part B illustrates unrestricted entry perforating
(sometimes as part of the propped fracture treatment) using a long interval, somewhat similar to an open-hole
to cover a previously stimulated perforated interval. interval. There is little or no initial control over fracture
Bridge plugs are the most reliable and commonly treatment distribution; differences in breakdown (and
used method to provide isolation between perforated fracture propagation) pressures along the depth of the
zones. Frac baffles are concentric baffle rings that are well dictate where fracture growth will occur. Ball sealers
run as part of the casing string and placed between are sometimes used in conjunction with unrestricted
individual target zones (Brown et al., 2000;and entry treatments (see Section 9-7.2). Ball sealers are
Robert and Rossen, 2000).Isolation is achieved by small spheres that are added to the fracturing fluid to
dropping balls or running wireline plugs that seat in seal the perforations that accept the largest quantities
the baffle rings. Some of these techniques are limited of fluid (Robert and Rossen, 2000). However, their
to vertical or nearly vertical wells. sealing efficiency is unpredictable because of the erosion
There is always an economic trade-off between of the hole diameter by injected proppant, potential
the number of fracture treatments and the ability cross-flow between perforated intervals, and failure to
to achieve effective fluid diversion. In cased wells, seat at low pump rates. Part C illustrates point-source
several perforation schemes can help to create good perforations used to minimize creation of multiple
vertical coverage (Fig. 6-1). fractures. Point-source perforation distribution is also
used in deviated wellbores to obtain fluid flow through
A. Limited entry B. Unrestrictedentry C. Point source D. Multiple point
only a very few perforations in a single section of the
* Initially good * Potentially poor * Sensitive to * Initially good
* Extreme case * Created at stress contrast * Minimized at well (Underwood and Kerley, 1998). This tends to
1 frac/perf wellbore * Minimized at wellbore wellbore
produce very few fractures or only a single significant
fracture. While it does not suffer from adverse fluid
flow distributions and difficult proppant placement,
this method could stimulate only a very small target
zone due to limited height growth as a result of
fracture containment. Point-source perforations
may also pose the risk of limiting or choking
Vertical Coverage Multiple Fractures
post-frac production or injection.
Part D illustrates multiple point-source/limited entry
Figure 6-1 Perforation schemes (from Minner et al., 1996) perforating, a hybrid strategy that consists of limited
entry implemented in the form of, e.g., three point-
In Fig. 6-1, part A illustrates limited entry source perforation intervals, allowing vertical coverage
perforation distributions that are used to obtain of the formation by fracture height growth, rather than
simultaneous fluid flow through a large number of by fracturing each interval. This strategy would have
perforations, by creating a large frictional pressure drop the advantages of both limited entry and point-source
203
Modern Fracturing
perforations while initiating fewer fractures at the conventionally staged stimulations, although there
wellbore, preventing premature screenouts, achieving are significant limitations to the slurry injection rates
larger fracture dimensions with greater proppant and pressures due to the relatively small diameter
concentrations per fracture, and covering a larger interval of the coil. As many as 20 fracture treatments
using a single treatment. While the major advantage of can be completed during a day.
this technique is to discourage the creation of multiple A second coiled tubing fracture technique involves
fractures, it can also be utilized to isolate and effectively running a jetting tool on coiled tubing into an un-
treat small net pay intervals. perforated well (Hejl et al., 2006). Perforations are cut
into the casing by first pumping a low-concentration
Casing-Conveyed Perforating: Casing-conveyed proppant slurry through the jetting tool. The propped
perforating technology refers to the utilization of fracture treatment is then pumped down the casing/
perforating guns that are attached to the outside of coiled tubing annulus. After the fracture treatment
the casing string prior to running and cementing is completed, the fractured perforation interval is
(Fig. 6-2). External control lines provide the ability covered with a sand plug to isolate it from the next
to sequentially fire the perforating guns for each fracture stage performed up the hole. After all of the
stimulation treatment. Internal valves provide fracture treatments are completed, the coiled tubing
isolation for the previously stimulated zones. As many string is used to clean out the well. As many as eight
as 17 discrete intervals have been treated in a 24 hour fracture stages have been performed per day.
period utilizing the process (Rodgerson et al., 2004). Shallow wells, (for the purpose of this discussion
The advantage of casing conveyed perforating is not about 3000 ft measured depth or less) containing
only the ability to rapidly stimulate multiple zones, zones of similar reservoir pressure, are ideally suited
but these systems have also been shown to improve for an alternate zonal isolation technique utilizing
well performance, reduce environmental impacts, and opposing swab cups on the coiled tubing string (or
provide a safer work environment (Krawietz et al., jointed tubing) to isolate multiple perforated zones.
2004). See section 9-7.5 for a detailed discussion. Utilization of cup tools is normally limited to about
a 20-ft straddle interval. Greater straddle lengths
Coiled Tubing Applications: Several hydraulic require very long lubricators above the wellhead, or
fracturing technologies involve the use of coiled tubing deployment under pressure, both of which are time-
(see Sections 13-1.4 and 13-4.5). One technique consuming and expensive. Utilization depth depends
involves pumping the fracture treatments through a on casing size and tool configuration. For deeper
coiled tubing string. For this procedure, perforations applications, in order to move the tool, nitrogen
are shot along the entire well first. Then, a bridge- injection (to reduce hydrostatics) or internal circulation
plug packer assembly run on coiled tubing is used devices may be required. An additional disadvantage
to place proppant into a selective set of perforations of the technique is that a downhole tubing check valve
(Zemlak et al., 1999). The assembly is quickly pulled cannot be installed for well control, and in extreme
up after a fracture treatment, to conduct the fracture circumstancesspecifically high pressure kicks
treatment in the next set of perforations up the well. the implementation of shear rams at surface may be
This technique has the advantages that it isolates required. The major advantage of this type of tool,
the wellhead and tubulars from treating pressures especially on coiled tubing, is the ability to hydraulically
and reduces execution time when compared to fracture multiple zones in rapid sequence.
Figure 6-2 External casing perforating (from Rodgerson et al., 2005)
204
6-2.2 Open-Hole Completions 6-2.3 Open-Hole and Uncemented

Liner Fracture Treatment Diversion
Wells are sometimes completed as an open hole
(without casing) through the geological section Obtaining desired zonal coverage during stimulation
of interest. The open hole may have a length of treatments using diversion techniques is difficult
1000 ft or significantly more. The main limitations in open-hole and uncemented liner completions.
of open-hole wells are that the well is unsupported The best method for zonal isolation in uncemented
and may collapse, and selective fracture treatments completions is mechanical isolation (see Sections 9-7.1
or remedial work within the open-hole section and 9-7.2). In this section, we discuss a few techniques
are more difficult. Open-hole stimulation is often that are sometimes used to improve fracture height
made even more difficult by the installation of a coverage in thick target zones. In general, open-hole
slotted or pre-drilled liner. diversion techniques are primarily suited for matrix
Effective zonal isolation for propped treatments and acid fracturing. Typically, fracture
fracturing in open-hole wells is a challenge for fluid and proppant placement in any given location is
fracturing engineers, but packers for open-hole unknown; therefore fracture treatment designs cannot
wells are available (Freyer and Huse, 2002). The be optimized (see Section 4-5 and 4-6 for treatment
packers are configured with one large element optimization parameters) and run the risk of being
that deforms easily to contact the uneven over-displaced. However, some propped fracture
surface of the drilled hole, but retain sufficient stimulation treatments, especially in slotted liner
strength and integrity to withstand the anticipated completions, have been successful.
temperatures and differential pressures. Open- Chemical diversion can be conducted by utilizing
hole packers could be successful for matrix fracturing fluid additives that temporarily seal the
stimulation, but the use of these packers is perforations or fractures that accept most fluid flow.
generally not recommended for fracture treatments Chemical diversion includes bridging materials such as
because, ordinarily, they cannot withstand high rock salt, naphthalene, benzoic acid flakes, wax beads
pressure differentials. Fractures typically initiate and foams. Bridging materials are used to bridge casing
along the open-hole section and tend to follow slots and/or pre-existing fractures (Brown et al., 2000).
this open-hole section along and past the packer. They share many of the limitations of ball sealers and
In addition, the fracture could initiate at the sand packs in that they are difficult to accurately place
packer face due to the compressive force exerted and keep in place. In general, conventional open-hole
by the packer on the borehole wall, or due to diversion techniques do not provide any significant
natural formation heterogeneity. control over when or where treatment fluids are being
Another isolation option is the use of solid- placed. However, wax beads have been used successfully
body open-hole packers that are run on a for water fracturing open-hole wells in naturally
tubing string and generally considered a permanent fractured chalk formations (Bell et al., 1993).
installation. In most cases the packers are pre- When using uncemented liners, completion
set with fracture port windows between packers; techniques may enable at least some degree of selective
these windows can be individually opened with fracture treatment isolation or remedial work within
balls or darts. Starting at the far end of the open- the (e.g., horizontal) target zone, when the liner has
hole section, the fracturing windows can be multiple point-source limited entry perforations in the
sequentially opened and then isolated during open-hole (uncemented) section. Liner perforations
the treatment so that all intervals (up to eight are either mechanically installed on the liner before
with some systems) can be stimulated in the same running in the hole, or conventionally shot after the
day. The distance between packers and the liner is in place downhole. An unrestricted slotted
location of frac ports is variable and depends only liner does not provide good diversion behind the pipe
on completion objectives. into the open-hole annulus.
205
Modern Fracturing
With the point-source perforation strategy, the in Section 6-4.3. As discussed in subsequent
hope is that the higher local pressure at each perforation sections, deep-penetrating perforations are
cluster initiates a fracture system nearby. A lateral of unnecessary (and occasionally detrimental) for wells
significant length may be effectively stimulated using requiring fracture stimulation.
this method (Fisher et al., 2004, 2005; and Van Dyke
et al., 2005). However, the location of the created 6-4 Perforating for Fracturing
fractures is not strongly correlated to the location
of the clusters of perforations. That is, the fracture The primary objectives of perforating for fracturing
network grows where in- situ stress conditions along are to minimize near-wellbore pressure drop during
the horizontal lateral dictate, and not exactly where production and to allow proppant slurries to enter
the perforations are shot along the uncemented the fracture during the treatment with minimal risk
liner. Located in central California, the Rose Field is of proppant bridging. Of particular importance is
a pioneering area for using the uncemented slotted the pressure drop due to tortuosity, caused by the
liner concept. Surface tilt mapping showed transverse misalignment among perforation tunnels and the
fracture growth at the toe and heel of a well, with a fracture initiation and ultimate propagation planes.
large longitudinal component in the central portion This will be discussed later in this section.
of the lateral (Minner et al., 2003). Injection rates In general, the perforation process results in three
of 60 to 75 bbl/min were required to distribute primary damage mechanisms: perforation tunnel
fluids across the entire lateral. blockage due to debris, migration of formation fines
into the perforated region, and a crush zone at the face
6-3 Perforating in General of the perforation tunnel, caused by shock waves, that
reduces localized permeability. Perforation damage
Perforating for production is a widespread procedure is further discussed in Section 6-4.3. Techniques to
practiced in the petroleum industry. The main remedy perforation damage include pre-fracture acid
objective is to provide maximum access to the reservoir. washes, underbalanced perforation techniques, extreme
A shot density of four or five shots per foot (SPF), overbalanced perforating and the correct selection of shot
covering all producing horizons, is recommended. density, charge size and orientation.
Studies as early as Muskats (1946), show that well Perforating for fracturing should include
productivity is essentially independent of perforation considerations of treatment size, proppant concentration,
pattern and is determined mainly by perforation proppant size and treatment pump rate. In addition, well
density. Since then other studies (e.g., Karakas and reservoir data are important factors when planning
and Tariq, 1988) have quantified the effects of a perforation program. Deviated and horizontal well
perforations on well performance by calculating the configurations demand alternate perforation schemes for
resulting mechanical skin effect. One important treatment success. Failure to consider these factors may
conclusion is that, in order to bypass near-wellbore lead to excessive formation breakdown pressures and an
damage caused by drilling, completion and cementing unacceptable screen-out frequency.
operations, the perforation needs to reach undamaged
reservoir rockalthough it is widely recognized 6-4.1 Oriented Perforations
that the perforating process itself will result in some
localized damage. In general, deep penetrating In vertical and horizontal wells, perforation orientation
perforations are only recommended for reservoirs has a major impact on fracture-to-well connectivity.
that are not hydraulically fractured. In addition, perforation orientation can be used to
Overbalanced perforating can initiate small minimize breakdown pressures and maximize near-
fractures near the wellbore, which may assist wellbore fracture width (Abass et al., 1994). As shown
subsequent fracture propagation. Overbalanced in laboratory experiments, 180 phasing within 30 of
and underbalanced perforating are addressed the maximum horizontal stress generally provides good
206
connectivity with the fracture. 60 phasing will insure that link. So, for a situation where the borehole is aligned
some perforations will be within 15 of the maximum with the stress, non-oriented perforations pose little
horizontal stress plane, but more shots are required and fracture linkage concern.
may not provide focused fracture initiation.
0.2
Figure 6-3 shows a wellbore with a perforation
in the preferred plane and other perforations that are
0.1
outside the preferred plane. If the perforations are not
aligned with the fracture plane that is controlled by
y, m
the far-field stresses, the fracture initiation planes are 0
tilted, which may hamper link-up between fractures.

Eventually, the fractures will abruptly align with the -0.1
maximum stress. The tangential stress varies along the

borehole, but if vertical stress is the maximum local -0.2
-0.2 -0.1 0 0.1 0.2
stress component, the fractures will tend to propagate x, m
along the wellbore. This assumes that the wellbore and Figure 6-4 Fracture growth from perforations positioned
the maximum (vertical) stress are perfectly aligned, 15 from the preferred fracture plane (from van de Ketterij
and de Pater, 1997)
which is usually not the case. At some distance from the
borehole, the vertical stress component is equal to the Figure 6-5 shows the result when the fluid
far-field vertical stress. Fracture interaction simulations pressure is much higher (caused by a dramatic increase
consider the fracture propagation in two dimensions in fluid pumping rate compared to the example in
with the y-stress equal to the vertical stress and a varying Fig. 6-4), equal to the maximum stress. In that case,
x-stress that corresponds with the hoop stress around the it is more difficult for the fractures to link because
wellbore. From 3-D simulations (van de Ketterij and de the pressure inside the fractures dominates the stress
Pater, 1997) it is clear that most of the re-orientation field. A hydraulic fracture only completely follows
occurs in the radial direction. the direction of maximum stress when the fracture
pressure is much smaller than the stress difference.
Variation of tangential stress
T2
along wellbore wall The fracture pressure itself also contributes to the
stress field on the fracture tip. The result is that
Preferred fracture
fractures tend to propagate along a straight path
Fracture
in preferred
H plane at very high fluid pressures (which may be hard to
plane T3
achieve in the field), reducing the chance of link-up.
Perf tunnel Fractures from This could lead to more tortuosity.
perforations
T1 outside the
preferred plane
0.2
0.1
Figure 6-3 A wellbore with a perforation in the preferred
plane and other perforations that outside the preferred
y, m
0
plane (from van de Ketterij and de Pater, 1997)
Figure 6-4 shows the fracture growth path for two -0.1
initial fractures from perforations that are 15 degrees

from the preferred plane. The perforations are spaced -0.2
-0.2 -0.1 0 0.1 0.2
as 2 SPF (0.5 ft apart, 0.15 m), and the initial fracture x, m
is taken as 0.8 in. (0.02 m). Where the borehole is Figure 6-5 Fracture interaction for fractures initiating
from perforations that are 15 degrees from the preferred
perfectly aligned with the vertical stress, the fractures fracture plane with higher fluid pressure compared to Fig.
would quickly reorient towards the borehole axis and 6-4 (from van de Ketterij and de Pater, 1997)
207
Modern Fracturing
Figure 6-6 shows how gun selection in vertical wells shot density is even smaller. This increase in vertical
is affected by the horizontal stress bias, or the contrast distance between perforations of the same phasing
between the minimum and maximum horizontal stress. decreases the chances of link-up between fractures
In areas with high horizontal stress bias, utilizing many initiating from the perforations.
perforations with a small phasing (30, 45 or 60) generally
provides better wellbore-to-fracture communication 6-4.2 Deviated and Horizontal Well Perforating
than 180 phasing because fracture growth starts only in
a limited number of favorably oriented perforations. If Deviated and horizontal wells present special perforating
180 phasing were chosen in this case, high near-wellbore challenges. It is not likely that the wellbore and fracture
friction could result from fracture initiation through planes will coincide (Veeken et al., 1989) unless the well
micro-annuli if the perforations are not oriented toward is drilled deliberately along a particular direction after
the maximum horizontal stress. significant measurements and effort during drilling.
Because perforation orientation is not likely to match
Increasing stress bias
fracture orientation, perforation strategies need to be
Tmax
Tmin evaluated to promote fracture development with low
60 perf phasing
near-wellbore pressure and adequate connection to the
Poor communication Good communication main fracture. Orientation of the wellbore and associated
due to near-wellbore
multiple fractures perforation location will determine the connection
between the fracture and the wellbore.
Typically, deviated/horizontal wellbores are
180 perf phasing not good candidates for hydraulic fracturing
Potentially poor
treatments unless they are specifically designed for
Reasonable communication communication due to
with potential problems due fracture initiation fracturing. In order to ensure successful fracturing
to reduced near-wellbore width from annulus
treatments, wellbores should be drilled so that they
are perpendicular to, or so that they trace one of the
180 perf phasing principal stresses: e.g., vertical (for vertical wells),
oriented guns
or one of the two horizontal stresses (for deviated
Good communication
Good communication
wellbores) to accommodate longitudinal or transverse
Figure 6-6 Perforation phasing considerations for areas
fractures. Because the exact azimuth of principal
with low and high stress bias stresses is typically not known very well, correct
wellbore orientation is difficult to achieve.
It is not common, or easy, to measure the horizontal A common industry practice, especially for wells
stress bias, but it can be achieved by analyzing induced drilled from platforms or pads, is to configure the
and natural fractures, as well as borehole breakouts in wellbore as S-shaped so that the wellbore is vertical
caliper or image logs. Also, oriented core can be brought throughout the producing interval. Because of the very
to surface, where relaxation of the core can be measured definition of principal stresses, vertical wells aligned
using anelastic strain recovery (ASR) (Teufel, 1983). with the principal vertical stress, exhibit no shear
There are two disadvantages of perforations with stresses. This helps to reduce fracture initiation pressure
30, 45 or 60 phasing. First, there is a chance that a more and, with correct perforating practices, can reduce
complex system of fractures may initiate, especially tortuosity substantially, thereby enhancing connectivity.
if the horizontal stress bias is not high. Secondly, a Although the ideal perforation orientation is 180 and
smaller perforation phasing generally means that the parallel to the maximum principal stress, the stress
vertical distance from perforations at the same phasing orientation must be known. Fracturing strategies for
increases because perforation companies typically do horizontal wells are addressed in Chapter 10. For most
not provide guns with shot density in excess of 12 deviated wellbores, tactics to reduce tortuosity must be
shots per foot (SPF). For smaller guns, the maximum implemented, as discussed in Section 6-5.
208
Chen and Economides (1995) presented a and the perforation skin as

calculation to determine both the initiation plane of
2khLgravel
the fracture and the fracturing pressure for arbitrarily sp = (6-2)
k perf Aperf N perf
deviated wells. Figure 6-7 shows the geometric
variables and stresses required for the calculation. where Lgravel is the connection flow length (difference
If the angles a (the difference in azimuth between between drilled radius and screen radius), kperf is the
the plane of the wellbore axis and the main fracture perforation permeability, Aperf is the perforation cross-
plane) and b (the well deviation from the vertical) are sectional area, and Nperf is the number of perforations
known, if the three principal stresses are known, and connecting the fracture and the wellbore. Welling
if the angle q (between the initiating perforation and (1998) used the above formulation to back-calculate
some reference point on the circumference of the well), that the number of perforations for actual frac-packs
then the angle g between the well axis and the fracture was, on average, about 2.3 SPF, and he attributed
plane can be determined. Chen and Economides this to the fracture aligning only with perforations
(1995) showed that substantial tortuosity and large oriented in or very nearly in the fracture plane. By
fracture initiation pressures should be expected in basing the calculation on SPF, this implicitly assumed
highly anisotropic reservoirs. that the fracture plane was aligned with the axis of
the wellbore, but the paper indicates that all wells
za under discussion are inclined.
T1
z ya Instead, Ehlig-Economides et al. (2008)
developed a calculation to take into account the
difference between the angle of the wellbore axis
H r (arbitrarily governed by the drilling trajectory) and
C
xa the angle of the main fracture plane (normal to
R the minimum stress direction beyond one to two
y wellbore radii). Instead of counting a fraction of
the perforations per foot based on the perforation
T3
phasing, this calculation counts the number of
perforations within wellbore segment that is actually
T2 x intersected by the main fracture plane:
B rhalo
SPF
N perfs = min L, (6-3)
sin min (, )
Figure 6-7 Fracture and deviated wellbore geometry
(from Chen and Economides, 1995) where L is the drilled wellbore length through the
formation, rhalo is the halo radius, is the difference
6-4.2.1 Production Impairment from Inefficient Fracture- in azimuth between the plane of the wellbore axis and
to-Wellbore Contact the main fracture plane, and is the well deviation
angle. The connection flow length is assumed to
Welling (1998) calculated the skin for hydraulically be the difference between the drilled radius and
fractured wells as the sum of the fracture skin and the the screen radius. Because of the halo effect, all
skin calculated from the pressure drop through the perforations within the connection flow length are
perforations connecting the hydraulic fracture to the assumed to connect to the fracture.
well. Welling also assumed that the perforations had Although the work described here is for all
30o phasing and 12 SPF. hydraulic fractures, the problem is particularly severe
The fracture skin was computed as in high-permeability, loosely consolidated reservoirs.
rw (1 / C fD + 2) The halo effect (Upchurch, 2001) mentioned above
s f = ln (6-1) has been postulated as a configuration in loosely
xf
209
Modern Fracturing
consolidated sandstone reservoirs where proppant is In many operating companies, production

placed concentrically around the well as well as in the engineers claim that they must complete a well
fracture. Because of the halo effect, Upchurch (2001) regardless of drilled angle because the driller decides
suggested that the flow from the reservoir into the the well deviation and azimuth. To simplify their
fracture and into the well is far less cumbersome than task and to cut costs, drillers will typically maintain
the traditional vision of tortuosity. one well deviation and azimuth required to reach the
Ehlig-Economides et al. (2008) adopted the halo specified reservoir location from the drilling location.
effect as the best-case scenario. Their approach shows If the platform is near the center of the field, well
the skin penalty when the hydraulic fracture does not azimuths will vary 360o, and only 2 directions will be
align with the wellbore trajectory, a penalty that is aligned with the main fracture plane.
easily avoided by ensuring the wellbore trajectory is Alternatively, the well could be inclined through the
vertical through the productive interval whenever the productive interval by turning the trajectory azimuth
maximum stress is vertical. normal to the minimum stress direction through the
Figure 6-8 shows a sensitivity study for a range productive formation. In some cases this may require
of SPF values and the worst-case scenario that the less change in angle than turning the trajectory
azimuth angle difference between the plane of the downward. The additional stimulation from a vertical
wellbore axis and the main fracture plane is 90o. well to an inclined well aligned with the main fracture
Perforation permeability is assumed to be 200,000 plane is quite small, but the skin and, especially, the
md, like that of a high-quality proppant, and the screen velocity increase from misalignment of an
halo radius is assumed to be 1 ft. For example, for inclined well is quite severe.
a well with an inclination of 45o and 6 SPF, the The most commonly encountered misconception is
perforation skin would be 8, whereas for 2 SPF the to justify inclining the well on the basis that there will
skin would be larger than 25. For a fracture skin equal be more contact between the well and the reservoir. This
to 4, the composite skins would be +4 and +21, will be true if the well is gravel-packed, but if the well
respectively. In fact, looking at Fig. 6-8, only at well is successfully hydraulically fractured with sufficient
inclinations of 15o or less and larger than 4 SPF will fracture conductivity, either the flow through perforations
the composite skin be zero or negative. that are not connected to the hydraulic fracture will be
The skin for any combination of and can negligible, or there will be sand production.
be determined from Fig. 6-8 for the same input To see this, suppose the main fracture plane is
formation geometry and permeability parameters. For formed at an angle with the wellbore trajectory
< , read the skin value for the angle . Otherwise, and, therefore, with less than all of the perforations
read the skin value for the angle . connecting the fracture and the wellbore. When the
well is brought on to production, initial flow may
80
occur through all of the perforations. As the well
1 SPF
70
2 SPF flow rate is increased, when the flow velocity into
60 4 SPF perforations connected only by radial flow from the
Perforation Skin, sp
6 SPF
50
12 SPF
formation exceeds the flow velocity at which formation
40 fines become mobile, these perforations will either
30 become plugged or they will permit fines to flow. In
20 the former case, soon flow through these perforations
10 will be negligible, and all flow will occur through the
0
perforations connecting the fracture to the wellbore.
0 10 20 30 40 50 60 70 80 90 In the latter case, because frac-pack gravel grain size
Deviation Angle, degrees
distribution is not necessarily designed to filter fines,
Figure 6-8 Frac-pack perforation skin, assuming 90o sand production and subsequent screen erosion may
angle between wellbore and main fracture planes (from
Ehlig-Economides et al., 2008) occur via the perforations that do not connect the
210
fracture to the wellbore. Meanwhile, linear flow into 6-4.3 Underbalanced vs. Extreme
a properly designed frac-pack should never exceed the Overbalanced Perforating
rate at which fines become mobile.
Therefore, during fracture execution, gravel or Underbalanced perforating (UB) means that operations
proppant injected through perforations that are not are performed below reservoir pressure. Extreme
connecting the main fracture to the wellbore is at best overbalanced perforating (EOB) is done at pressures well
diverted from propagating or widening the fracture and above formation breakdown gradient. Both techniques
reduces the final fracture length and/or conductivity. have applications, depending on specific well/reservoir
At worst, several highly inclined fractures may form parameters and completion objectives.
instead of one main fracture, or, as described above,
perforations not connecting the fracture to the wellbore Underbalanced perforating: UB perforating is
will enable sand production and/or screen erosion when primarily intended to sweep perforation tunnel debris
the well is put on production. into the wellbore from the perforation tunnel. In
Without explaining why all their wells were all types of perforating, three types of perforation
inclined, Welling (1998) recommended high-rate tunnel damage occur. The first is known as the crush
water packs over gravel packs for high-permeability zone caused by crushing and compaction of the
reservoirs, and open-hole gravel pack completions for rock matrix near the perforation tunnel. As a result
highest target well rates. That paper claims to be able of crushed sand grains in the matrix, pore throat size
to estimate the number of flowing perforations and the is reduced, resulting in a loss of permeability in that
gravel pack or frac-pack permeability. However, very region. The second damage mechanism is residual
likely, the perforations are miscounted in that analysis. perforating debris, lodged in the perforation tunnel. It
If, instead, well inclination were used to estimate the is not expected that this type of residual debris would
number of perforations with Eq. 6-3, the number of impair production, However, it may affect attempts
perforations would be smaller, and the estimated gravel at subsequent injections, including fracturing,
permeability higher. Figure 10 in the Welling (1998) resulting in higher injection pressures. Lab testing
paper claims that the observed gravel pack permeability has indicated that particle debris filling at least 50%
is higher than that of the frac-pack. of the perforation tunnel can substantially reduce
In reality, gravel pack should never be recommended injectivity without affecting producibility (Behrmann,
for offshore wells, particularly if they have sub-sea 1995). The damage may be aggravated in vertically
wellheads. Over time, the gravel pack will accumulate downward-oriented perforations. An additional
fines, and, because matrix acidizing may not be an option, damage mechanism is the migration of formation
the well rate would have to be cut back to avoid excessive material, including clays, caused by mechanical
drawdown across the completion. A properly designed disruption of the formation matrix.
and executed frac-pack will never accumulate fines. Removal of crush zone damage has been studied
Instead of working with the driller to design by several investigators. The pressure drop required
a well completion that can safely flow at target to remove crush zone damage has been evaluated
production rates, many production engineers monitor in the laboratory using surface outcrop samples
the arbitrarily inclined frac-pack wells to avoid screen (Behrmann, 1995). Equation 6-4 is an example of
failure. Although this strategy may save wellbores, it such an empirical relationship using data from Alaska
fails to achieve design flow rates and thereby penalizes and North Sea Brent reservoirs:
the project economics. Well design should be a unified
2
endeavor, not just a series of avoidable compromises. pUB = D p0.3 [90.4 7.86 ln k ] ,
(6-4)
When the majority of fractures are properly designed
in vertical wellbores, the average skin for frac packs where pUB is optimum underbalance pressure, psi;
will drop below zero, and there will be very few screen Dp is the perforation tunnel diameter, in. and k is
failures, even for very high production rates. reservoir permeability, md.
211
Modern Fracturing
Removal of particle debris can easily be accomplished laboratory measurements indicate that the pressure at
during UB perforating operations. According to the perforated interval should be in the range of 1.4 to
Behrmann and McDonald (1996), few (if any) laboratory 2.0 psi/ft (Behrmann and McDonald, 1996). Handren
studies are available to confirm and quantify injectivity et al. (1993) state that as a rule of thumb, overbalance
impairment caused by particle debris, but the following gradients need to be about 0.4 psi/ft greater than the
factors probably aggravate debris removal: fracture treatment gradient. In any case, the overpressure
1. Low underbalance pressure differential. gradient needs to exceed the reservoir breakdown
2. Use of big-hole charges. gradient. A third method involves propellant-assisted
3. Weak (brittle or unconsolidated) rock. perforating, combining perforation breakdown with
4. Downward-oriented vertical perforations. propellant in a single tool and operation (Loman et
5. Single-phase oil or gas flow versus two-phase flow. al., 1996). The perforating assembly has a propellant
6. Perforation tunnel diameter larger than entry- sleeve over a conventional perforating gun assembly.
hole diameter, creating a flow restriction in all When the guns are detonated, the propellant sleeve is
but hard rocks. ignited, instantly producing a burst of high-pressure
gas. The gas enters the perforations, breaks through
UB perforating is preferred for both injectors and any damage around the perforation tunnel, and creates
producers that are deemed candidates for hydraulic micro-fractures near the wellbore.
fracturing treatments, except in weak rock. UB perforating A number of potential benefits can be achieved
may be of particular importance if there is reason to through EOB perforating (Handren et al., 1993). It is
believe that high breakdown or injection pressures may believed that during conventional UB perforating, the
be encountered. A successful UB perforation cleanout perforation tunnel may become unstable due to a rapid
requires sufficient permeability and reservoir pressure loss of pressure. Rapid changes in the near-perforation
to successfully surge the perforations. In many depleted stress field may result in the rapid closure of near-tunnel
reservoirs, UB perforating may not be feasible because fractures, causing collapse of the tunnel. During the
of insufficient reservoir pressure. For example, using Eq. EOB process the sustained application of high pressure
6-4, if the reservoir permeability is 100 md and the Dp can serve to stabilize the tunnel walls.
is 0.5 in. then the required pUB would be about 2500 EOB perforating also promotes and extends a
psi. But if the permeability is 0.1 md then the required fracture network beyond the perforation tunnel. During
pUB would be a highly unrealistic 9400 psi. Thus, the perforating process, micro-fractures are created
for under-pressured or depleted lower-permeability around and at the tip of the perforation tunnel. Debris
reservoirs (almost all formations in North America) generated during perforating will migrate into the
UB perforating is not recommended. micro-cracks because of the extreme pressure differential,
establishing the main connectivity improvement
Extreme overbalanced perforating: EOB perforating is provided by the EOB process.
intended to promote the formation of micro-fractures The rapid invasion of fluid into the perforations
that extend beyond the conventional perforation tunnel. during perforating should extend existing micro-
EOB perforating is accomplished in three ways. The fractures and potentially create additional micro-fractures
first procedure involves pressurizing the wellbore before extending from the perforation tunnels. With sufficient
perforating. The pressure in the wellbore is increased to velocity, the fluid should erode the micro-fracture face,
a value that exceeds the expected formation breakdown which benefits reservoir-to-wellbore connectivity.
pressure, and then the perforation guns are fired. A Reservoir permeability is expected to have an
second technique creates a pressure surge into the effect on the efficiency of UB perforating. Variations in
perforations after conventional perforating, normally permeability within the perforated zone can determine
pumping nitrogen gas in the tubing through a rupture the effectiveness of perforation surging after UB
disc. Pressure requirements for EOB perforating perforating. Large permeability variances could render
vary depending on application. Field experience and some perforations completely ineffective. On the other
212
hand, permeability variations should have little impact Perforating with Lasers: Since 1997 (Gahan et al.,
on the effectiveness of EOB perforation operations. 2004; and Parker et al., 2003), the Gas Research
Although fluid velocity and pressure drop are the Institute (now the Gas Technology Institute or GTI)
dominant variables in EOB perforating, fluid choice and its research partners have been investigating the
can also have an impact on successful perforating. Acids application of laser technology for use in both drilling
can be used in acid-soluble formations to etch the walls and perforating (rock removal). High-power laser energy
of the micro-fractures, thereby enhancing connectivity. conveyed via fiber optic cable technology, appears to
Acid wash or soak treatments can also aid in removing be able to perforate wellbores. Laboratory testing has
acid-soluble perforating charge debris. Sand slurries can focused on using concentrated laser energy to remove
be used to etch the fracture faces and aid in perforation rock, with the intention of improving near-perforation
erosion, lowering subsequent fracture treatment tunnel permeability (Fig. 6-10). Beyond improving
pressures. Other options include the injection of sand perforation tunnel connectivity to the wellbore, laser
consolidation resins designed to control fines flowback technology has advantages in rapid deployment,
from unconsolidated reservoirs. improved wellsite safety, and reduced environmental
Disadvantages of EOB include operational impact. However, there are still significant technical
considerations regarding safety and tubular/wellhead and operational challenges to be overcome before these
capabilities. Perforating under elevated pressure requires systems can be deployed in the field.
a careful evaluation of pressure limitations (and age) of
surface and downhole equipment, including wellhead,
casing, packers, gun system, isolation tools and lubricators.
Additional pumping equipment, product storage and
logistics add to safety concerns.
Performing a hydraulic fracture treatment after
EOB perforating raises additional concerns. While
enhanced fractures in the preferred fracture plane will be
advantageous to the propped fracturing process and could
aid in primary fracture initiation, enhanced fractures
away from this plane could lead to a much more complex Figure 6-10 Small block laser perforation test, showing a
near wellbore situation. (Ideally, 180 phasing oriented cutaway of a lased perforation tunnel in Berea sandstone
sample (courtesy of Sandia National Laboratories)
to the fracture plane is preferred.) Moreover, the EOB
process encourages the development of multiple fractures
(see Fig. 6-9). A multitude of fractures in various spatial 6-5 Near-Wellbore Fracture Complexity
orientations may lead to competing multiple fractures,
which will contribute to lower width, increased fluid Near-wellbore fracture complexity, often referred to
leak-off and an increased possibility of a screenout. as tortuosity, can be described as a series of width
restrictions in the area of the fracture that connects
Tmax Tmax the wellbore with the reservoir or a far-field planar
fracture. During a hydraulic fracture treatment, these
restrictions may lead to a pressure drop in the fracture
very close to the well. This near-wellbore friction
Tmax Tmax results in a difference between the bottomhole
pressure and the pressure in the body of the fracture;
this latter pressure controls fracture dimensions (see
A B Section 4-2.1.4). When proppant is added to the
Figure 6-9 Potential fracture orientation from
EOB perforating (A) and from UB perforating (B)
fluid, near-wellbore friction may increase dramatically
(from Handren et al., 1993) due to difficulties transporting proppant-laden slurry
213
Modern Fracturing
through the tortuous near-wellbore region; this principal stresses were applied to the model sample,
may cause proppant bridging. It is likely that many and thus far-field fractures would preferentially
premature screen-outs are caused by tortuosity, and grow perpendicular to the minimum principal stress
not by depletion of the pad or far-field fracture width (sh,min) direction. In the slab closest to the wellbore,
restrictions. Fracture complexity and restrictions it is clear that starter fractures initiate from every
in near-wellbore width may also impair the flow of perforation along the wellbore. Each of these fractures
hydrocarbons during production. turns toward the neighboring perforation as the fluid
Near-wellbore fracture conductivity is of utmost pressure within the perforation results in a tensile
importance for hydrocarbon production. Bad tangential stress around the perforation. This attracts
conductivity near the well leads to choked fractures the approaching fracture tip and makes it turn towards
(Cinco-Ley and Samaniego, 1981). This section the perforation, resulting in a complex network of
describes near-wellbore complexity and how it can be fractures. Farther away from the wellbore, however,
quantified and remediated. several of these starter fractures stop propagating and
some coalesce, making the fracture geometry less
6-5.1 Near-Wellbore Complexity complex further from the wellbore. Below and above
the perforated interval, only the fractures that start at
The connection between wellbore and fracture is the upper and lower perforation remain.
generally not a clean, zipper-like connection. Instead,
y = -0.075 m -0.05 m -0.025 m 0.025 m 0.05 m 0.075 m
there is often a rather complex and tortuous fracture
network linking the main body of the fracture with
the wellbore (Palmer and Veatch, 1990; Cleary et al.,
1991 and 1993). Although it is possible to pump
fluid past this near-wellbore area during a fracture
treatment, pumping proppant-laden fluid can be
problematic due to insufficient near-wellbore fracture
0.30 m Tv
width. Near-wellbore fracture tortuosity is the cause
for the majority of premature screen-outs. Near- Th, max
wellbore complexity varies widely between fields and Th, min
even wells within a field. As a general rule of thumb,
near-wellbore friction pressures of greater than 500 psi Figure 6-11 Near-wellbore fracture geometry as
(based on diagnostics described later in this section) observed in a model test (from Weijers et al.,1994)
should be interpreted as a cause of concern. Near-
wellbore friction in excess of 1000 psi will probably Summarizing the phenomena of these laboratory
require serious remediation, fracture treatment re- experiments, near-wellbore tortuosity can result from:
design or alternative completion strategies (e.g. re- gradual or sharp fracture curvature due to fracture
perforating or alternate zonal isolation techniques). re-orientation from the plane of initiation
Unfortunately, there is no catch-all universal method toward the preferred fracture plane (plane of
for remediating near-wellbore friction and its impact far-field fracture growth);
on subsequent fracture treatments. It is, however, of sharp fracture curvature when the hydraulic fracture
paramount importance to recognize its potential impact intersects natural or drilling-induced fractures;
and to be prepared to address the issue. multiple hydraulic fractures that compete for
Figure 6-11 shows an example of a complex near- fracture width; and,
wellbore geometry in a laboratory fracture growth fracture growth along the annulus between the
experiment from a modestly deviated well (Weijers et cement and the stiffer casing due to inadequate
al., 1994). The test block has been cut into different cementing. Fractures may initiate into the formation
slabs to reveal the fracture geometry. Three different away from the perforated interval.
214
High tortuosity during a fracture treatment greatly coverage achieved on a limited entry fracture treatment.
increases the risk of a premature screen-out due to Measuring entry friction pressure using a step-down
proppant bridging in the tortuous near-wellbore region. test helps to make the correct decision for remediating
The fracture treatment may be terminated due to proppant entry problems (Fig. 6-12).
unacceptably high surface pressure before all the proppant
is successfully placed in the fracture. In the worst possible High entry friction
case, the well may be lost for production. Also, the

maximum proppant concentration that can be pumped High perf friction Severe fracture tortuosity
when tortuosity is present may be lower than called for
Re-perforate
Use proppant slugs
in the original design of the treatment, resulting in a less Initiate with high viscosity fluid
Future wells may have
than optimum stimulation of the well. altered perforation strategy
increase gel loading
Increase rate
such as MORE perfs
Future wells may have altered
6-5.2 Diagnosing and Quantifying Spot acid
completion strategy such as
Near-Wellbore Complexity (Tortuosity) Perform ball-out FEWER perfs
Figure 6-12 Remedial measures to overcome

Currently, there are no methods to reliably predict high perforation or near-wellbore friction (from
Weijers et al., 2000a)
tortuosity, but there are simple diagnostic tests to
measure tortuosity and ineffective perforations before In the equations below, ptotal frict represents all the
pumping the propped fracture treatment. Typically, frictional losses during pumping from the surface to the
such tests are quick, inexpensive and require only small main body of the fracture, and pnear wellbore represents all
volumes of fluid and proppant. the frictional losses between the mid-perf depth of the
Although a simple ISIP shut-down will determine wellbore and the main body of the fracture. Just before
total friction, a step-down test can be done before the end of pumping, the terms ptotal frict and pnear wellbore
the main treatment to quantify and isolate the are equal to the difference between the pumping pressure
presence of perforation friction and near-wellbore and the instantaneous shut-in pressure (ISIP) at surface
friction (see Section 4-2.1.8). and downhole, respectively:
Before the step-down technique was developed,
p = p pISI , surf
all fracture entry friction (bottomhole pressure minus
total frict inj
(6-5)
pressure inside the main fracture body) observed = p pipe friction + p pf + ptort ,
at treatment shut-ins (ISIPs) was attributed to
perforation friction. This often led to expensive and pnear wellbore = piw pISI = p pf + ptort . (6-6)
unsuccessful (or even counter-productive) ball-outs,
re-perforating, acid spotting and/or increases in pad During shut-in, all frictional components in Eqs. 6-5
volume. However, it is now believed that most of these and 6-6 are zero. It is therefore relatively easy to calculate
premature screen-out problems are actually due to a observed net pressure during shut-in, and net pressure
complex wellbore-to-fracture connection that causes matches therefore often focus on shut-in times. During
high near-wellbore friction (i.e. tortuosity). pumping, the frictional components in Eqs. 6-5 and
The step-down test can be employed as a routine 6-6 can all change. However, their general behavior
procedure before a propped fracture treatment to with flow rate changes is governed by several relatively
evaluate potential proppant placement problems. Step- simple relationships that can be used to isolate the three
down tests can be employed to evaluate the potential individual frictional components.
risk of a near-wellbore screen-out and to evaluate the Wellbore Friction. Wellbore friction (see Section
efficacy of remediation measures. Step-down analysis can 4-4.4) depends on the type of fluid and flow path
also reveal the approximate number of perforations that diameter, but for most fracturing fluids in turbulent
are accepting fluid, which can provide essential insight flow it is a function of the flow rate to some
into perforation breakdown and the possible interval power w (Gas Research Institute, 1995a):
215
Modern Fracturing

p pipe friction = K wellf q with 1.2 < w < 1.8,
w
(6-7) There are two extremes for the behavior of near-
wellbore friction with flow rate. In the upper bound, it
where Kwellf is a proportionality constant for the is assumed that fluid flows through two parallel plates
wellbore friction pressure as function of the flow at a fixed distance from each other. For laminar flow
rate squared. When surface pressure is available, the between these plates, Darcys law applies, and thus the
estimated wellbore friction (from flow-loop tests or friction over a certain distance through these plates
empirical correlations) should be subtracted from the is directly proportional to the flow rate. In the lower
total friction to obtain entry friction. bound, the width between the parallel plates is directly
Perforation Friction. (See Section 4-2.1.2) As shown, proportional to the average pressure that is applied from
perforation friction is simply dissipation of the kinetic the inside of the plates. The pressure required to sustain
energy that is imparted on the fluid as it flows through a flow rate though the moving plates is proportional to
a small orifice at high velocity. Perforation friction, the flow rate to the power . Therefore:
therefore, is proportional to the flow rate squared times
a proportionality constant Kperf which is determined by ptort = K nwb q nwb , with nwb 1. (6-9)
the slurry density, s, perforation diameter, Dp, number
of perforations Nperf, and the discharge coefficient Cd As an average for real-time field use, near-wellbore
(Economides and Nolte, 2000) to modify Eq. 4-2: friction (tortuosity) can generally be approximated to
be proportional to the square root of the flow rate, and
q
2
thus nwb = 0.5 (Gas Research Institute, 1995a).
p pf = 0.2369s = K q2 ,
(6-8) As shown in Fig. 6-14, the step-down test
C N D2
perf
d perf p
results will be distinctly different for cases that are
Tortuosity. Near-wellbore fracture tortuosity is the result of dominated by either perforation friction or near-
the complex fracture geometry immediately surrounding wellbore friction (see Section 4-2.1.8, Fig 4-10 for
the wellbore. In a typical fracture treatment a multitude an example). For a perforation friction-dominated
of fractures can be initiated from numerous perforations regime, the largest changes in entry friction occur at
and micro-cracks. The orientation of these near-wellbore the high rates, whereas for a near-wellbore friction-
fractures is dictated by the orientation of the existing dominated regime the largest changes in entry friction
micro-cracks, perforation-induced or otherwise, and occur at the low flow rates.
not by the far-field stresses that dictate the orientation
Bottomhole pressure - ISIP
of the main fracture body. As fractures grow away from

Rate, Bottomhole pressure
p1
the wellbore, they eventually re-orient to align themselves p1 Total friction
Perforation friction f(q2)
with the far-field stresses as depicted in Fig. 6-13. p2
p2
q1 p3 Near-wellbore
Narrow pressure q2 p4 p3 friction f(q)
dependen opening High velocity, kinetic q3
p4
energy dissipation q4
Time q4 q3 q2 q1
Rate
Wide far-field fracture
Bottomhole pressure - ISIP
Rate, Bottomhole pressure
p1
p1 Total friction
p2
p2
p3
p3
q1 Near-wellbore
Cased borehole with perforations p4 friction f(q)
q2 p4
q3 Perforation
Figure 6-13 Tortuosity and perforation friction behave very q4 friction f(q2)
differently as a function of flow rate. Tortuosity occurs in a q4 q3 q2 q1
Time
pressure-dependent opening, and therefore is only weakly
Rate
(a power between 0.25 and 1) dependent on flow rate.
Perforation friction is caused by kinetic energy dissipation Figure 6-14 Step-down test results dominated by
after the fluid obtains a high velocity, and is a function of perforation friction (top) and near-wellbore friction
the flow rate squared (from Weijers et al., 2000a) (bottom) (from Weijers et al., 2000a)
216
6-5.3 Minimizing the Effects of Tortuosity of the multiple hydraulic fractures. The bottom
diagram illustrates a proppant slug that is too
There are three main methods available to minimize the aggressive and screens out all the fractures. The
effect of tortuosity during a fracture treatment: balance of proppant slug concentration and
Perforate a limited interval and use a limited volume is difficult to quantify and is often
number of perforations to reduce the number of determined based on field experience.
simultaneously propagating multiple fractures in the During fracture initiation, increase the flow rate
near-wellbore region (Stadulis, 1995). to increase the radius of near-wellbore fracture
Pump proppant slugs to erode the near-wellbore curvature and increase the fluid pressure in
fracture system and plug off less-conductive the near-wellbore region. The injection rate
multiple fractures (Cleary et al., 1993, McDaniel during initiation may be maximized by using
et al., 2001a, 2001b). Determining the correct overbalance perforating (Dees et al., 1993) or
proppant slug concentration is an uncertain by utilizing water jetting (East et al., 2004).
process. On one hand, a proppant concentration Also, use high-viscosity fluids (Aud et al., 1994)
that is too low will not plug some of the multiple during initiation to increase the radius of fracture
hydraulic fractures that are created near the well curvature and possibly decrease the number of
(Fig. 6-15, top). On the other hand, pumping a fractures initiated. This will also increase near-
proppant slug that is too aggressive (Fig. 6-15, wellbore fracture width and improve proppant
bottom) can plug off all fractures and cause a transport later in the treatment.
screen-out on the slug.
Example Illustrating the Use of Proppant Slugs
After proppant slug
to Reduce Tortuosity:
Before proppant slug Dominant fracture(s)
Not plugging any

multiple hydraulic fractures Figure 6-16 is a plot of treatment parameters versus
Too gentle
time for a treatment that showed severe near-wellbore
Dominant fracture(s) tortuosity in the naturally fractured Desert Creek
Just right Plugging off most formation at about 8,300 ft depth in New Mexico.
multiple hydraulic fractures
A step-down test was performed at the end of the
Too aggressive
second diagnostic injection, which showed that
Dominant fracture(s)
Plugging off all multiple

near-wellbore fracture tortuosity was extremely
hydraulic fractures high (1900 psi) at the designed pump rate, and
perforation friction was very low. This low perforation
Figure 6-15 Conceptual representation of the effectiveness friction was expected because of the large number of
of proppant slugs (from Weijers et al., 2002a)
perforations. The high tortuosity was probably caused
On the left side of Fig. 6-15, a complex by the simultaneous propagation of near-wellbore
near-wellbore network of multiple hydraulic multiple hydraulic fractures.
fractures is created when the naturally fractured The long perforated interval (70 ft, 4 SPF) and
formation is broken down. This complex network a relatively large pre-frac acid treatment most likely
can be treated with proppant slugs in an attempt provided multiple fracture initiation points in this
to simplify this network of multiple hydraulic dolomite formation (which would also have had
fractures. On the right side of Fig. 6-15, the top significant natural fracturing). To mitigate the near-
diagram illustrates the use of a proppant slug wellbore complexity problem, two proppant slugs (25
that is too gentle, causing continuous growth bbl at 1 and 2 ppg) were pumped very early during the
of multiple hydraulic fractures after the slug. pad stage. Just before the first proppant slug arrived at
The middle diagram illustrates an appropriate the perforations, one pump had to be taken off-line to
proppant slug concentration that plugs off some reduce the overall pump rate from 40 to about 32 bpm,
217
Modern Fracturing
20.00 20.00
6000 100.0
Surface pressure
limitation 6000 psi
S/D#1: 1900 psi 1300 psi reduction
16.00 (1st slug) Zero tortuosity at 16.00
4800 tortuosity; small 80.0
perf friction end of pumping
12.00 S/D#2: 300 12.00

3600 psi tortuosity 60.0
Increased max
proppant loading
from 4 to 6 ppg
8.00 8.00
2400 40.0
4.00 4.00
1200 20.0
0.00 0.00
0 0.0
0.0 30.0 60.0 90.0 120.0 150.0
Time (mins)
Proppant loading (ppg) Btm propant loading (ppg)
Casing pressure (psi) Slurry rate (bpm)
Figure 6-16 A propped fracture treatment with severe tortuosity was minimized by pumping two proppant slugs, without
which the treatment would probably have pressured-out during the pad
because the surface pressure was reaching the maximum Mid-field tortuosity (Fig. 6-17) manifests itself as an
pressure limitation of 6000 psi. When the first proppant initial concave pressure decline immediately after a shut-
slug arrived through the perforations, the surface treating in, indicating a choke of fluid flow farther away from
pressure dropped dramatically due to the reduction in the wellbore inside the fracture, possibly at a distance of
near-wellbore friction. This allowed for an increase in between about 10 and 50 ft.
rate back to 40 bpm with surface pressures well below
Well Near-wellbore region
5000 psi. The arrival of a second proppant slug at the Several feet into fracture
perforations reduced tortuosity by another 300 psi. After "Relatively" near-wellbore region
these proppant slugs (but still during the pad), a second Tens of feet into fracture
step-down test confirmed that the proppant slugs had Far field region
Hundreds of feet into fracture
indeed dramatically reduced tortuosity to an acceptable
300 psi at 40 bpm, providing confidence that the propped
fracture treatment could now be pumped successfully.
No Mid-field Tortuosity
Net fracture pressure
Bottomhole Pressure
6-6 Mid- and Far-Field Fracture
Complexity During pumping
End of pumping ISIP
Closure
Tip
In addition to near-wellbore tortuosity, we sometimes Rate
Distance from well Time
observe mid-field tortuosity (Weijers et al., 2002).
Mid-field tortuosity is significantly different from near-
wellbore tortuosity because it represents a choking High Mid-field Tortuosity
effect for fluid flow farther away in the fracture.
Net fracture pressure
Bottomhole Pressure
During pumping
Because near-wellbore friction occurs in the immediate End of pumping Apparent ISIP
vicinity of the wellbore, its effect instantly disappears Stabilized ISIP
when the pumps are shut down. Mid-field tortuosity Closure
Tip Rate
does not instantly disappear upon shut-down and can Distance from well Time
cause pressure equalization between the wellbore and
Figure 6-17 Conceptual picture of mid-field tortuosity
the fracture for tens of seconds or even up to several caused by a pressure choke beyond the near-wellbore
minutes after pump shut-down. area in the fracture (from Weijers et al., 2002)
218
Multiple hydraulic fractures are fractures that induced stress changes (Wright et al., 1994a,1994b, 1995,
grow simultaneously from a wellbore and penetrate far 2001) and interference from other fractures that remain
into the formation during a fracture treatment. It has open after they are created and thus leave a residual stress
long been known that multiple fractures are almost (Griffin et al., 2000), etc. In this chapter, however, we
always present near the wellbore, due to individual will limit our discussions to fracture complexities that
fracture initiation from many perforations in a cased- arise from the simultaneous propagation of multiple
hole perforated completion or from existing (natural sub-parallel hydraulic fractures.
or drilling induced) fractures along an open-hole Why does growth of multiple fractures occur at all?
interval. However, it was typically assumed that only From a minimum-energy standpoint, the propagation of
one fracture on either side of the wellbore propagated a single fracture would be favorable. This could maybe be
beyond this near-wellbore area; the other fractures were true in a perfect material without any local weaknesses
assumed to either coalesce in the near-wellbore area or and when fracture initiation occurs at a single point. As
cease growth due to the stress influence from the other shown in Figure 6-18, both conditions are violated in
nearby fractures. Evidence from core-through and hydraulic fracturing operations.
mine-back experiments, however, has indicated that
multiple hydraulic fractures may sometimes continue
to grow beyond the near-wellbore region. Fracture initiation
at Most of the
Complex fracture growth during the hydraulic Perforations
fracture treatment can therefore sometimes be more

Long Perforated Interval
problematic from a proppant placement perspective. Short

From a perspective of wellbore-to-fracture Point Source
Perforated
connectivity, it can also have major consequences. Interval
The redistribution of proppant from a single (wide)

fracture to several (narrower) multiple fractures
can result in increased damage from polymer
residue, reduced effective fracture width due to Completion-Induced Formation-Induced
proppant embedment and filter cake residue and Multiple Fractures Multiple Fractures
increased sensitivity to pressure losses from non-

Darcy and multiphase flow. Figure 6-18 Two main causes for the propagation of
multiple hydraulic fractures: At left, multiple fracture
initiation points from a long perforated interval with
6-6.1 An Introduction to Complex Fracture Growth numerous perforations; at right, bifurcation of hydraulic
fractures as they intersect natural fractures (from
Weijers et al., 2000) or as fractures grow through layer
Fracture diagnostic observations have shown that many interfaces (Barree, 1998).
hydraulic fractures may not be single and planar, aligned
with the wellbore and confined to the target zone. The A completion generally provides multiple
recent wealth of direct fracture diagnostic data has shown starting points for hydraulic fractures, as multiple
that in reality, there can be significant discrepancies perforations are shot. Hydraulic fractures tend to
between how fractures are expected (or designed) to initiate where there is a pre-existing crack from a natural
grow and how they actually do grow (Wright et al., 1999; fracture or a perforation-induced crack, because these
Warpinski and Branagan, 1989). Observations have are structurally the weakest points. Hydraulic fractures
included T-shaped fractures, horizontal fractures, fractures thus initiate from several perforations (or existing cracks)
that grow along multiple planes, fractures that bifurcate along a wellbore and may then coalesce into a fewer
at layer interfaces due to differences in mechanical number of dominant fracture(s) at some distance
properties (Wright et al., 1999), fracture reorientation due away from the wellbore (Weijers et al, 1994, 2000b).
to injection rate and viscosity changes (Weijers, 1999), Competing parallel fractures may also contend
fracture re-orientation due to production/injection- to squeeze each other out if they are close to each
219
Modern Fracturing
other, potentially leaving a single dominant fracture if tiltmeter fracture mapping, have indicated that fracture
perpendicular separation is minimal. The initiation of growth behavior may be significantly more complex than
multiple hydraulic fractures from the wellbore can be has generally been believed.
minimized, and the connection of the wellbore to the
fracture maximized, with proper wellbore completion 6-6.3 Consequences of Complex Fracture Growth
(e.g., perforating only a short interval, positioning the
well favorably with respect to the preferred fracture The simultaneous propagation of multiple hydraulic
plane and utilizing a high pressurization rate to improve fractures can have significant consequences for both
the coalescence of initial fractures). propped fracture treatment execution and the obtained
Growth of multiple hydraulic fractures is not propped fracture geometry, if more than one fracture
only determined by the controllable completion and receives a significant volume of treatment fluids:
stimulation practices. Bifurcation of a hydraulic fracture A significantly increased screen-out potential because
may occur wherever it intersects a natural fracture. In individual multiple fracture widths are smaller than
naturally fractured reservoirs (virtually every rock is the width of a single fracture.
naturally fractured to some degree), multiple hydraulic Shorter, narrower fractures because the fluid
fractures are potentially initiated and propagated, and proppant has to be shared by a number of
often in increasing numbers as more natural fractures multiple hydraulic fractures. The total width of all
are intersected by the growing hydraulic fractures. fractures added together, however, is larger than for
This process is self-reinforcing, as the propagation of a single fracture, resulting in higher near-wellbore
multiple hydraulic fractures tends to elevate the net proppant concentration. Despite the high proppant
fracturing pressure, which in turn makes it easier concentration, near-wellbore conductivity will
to initiate hydraulic fracture propagation at newly generally be lower than for a single fracture because
intersected natural fractures. narrower fractures are more prone to the effects of
proppant embedment and polymer residue damage.
6-6.2 Evidence of Complex Fracture Growth Higher net pressures. This can be partly because
fractures open against each other and compete for
In the early 1980s, the first indications for the fracture width, and partly because of tip effects
simultaneous growth of multiple hydraulic fractures caused by the process zone around a fracture
came from indirect fracture diagnostic techniques, such tip (see Section 4-3.4). In coals, these higher
as net pressure analysis and pressure transient tests. High net pressures could potentially damage the rock
net fracturing pressures were attributed to fracture tip permeability surrounding the fracture.
effects and the competition for fracture width from the 10 1000
simultaneous propagation of far-field multiple hydraulic
Fracture radius Cumulative fracture width
fractures (Shlyapobersky et al., 1988). Pressure transient
tests showed that fractures are sometimes much shorter
Fracture radius, ft
Fracture width, in
1 100
and of lower conductivity than initially expected based Individual fracture width
on theoretical model predictions.
Multiple far-field hydraulic fractures have been 0.1
Minimum width for
12/20 mesh sand 10
directly observed in the laboratory (Weijers et al., 1994) Minimum width for
20/40 mesh sand
and in the field during several core-through and mine-
back experiments. These fractures contained offsets at
0.01 1
geological discontinuities such as joints, confirming that 1 10 100
Number of simultaneous propagating fractures
multiple fractures can be caused by bifurcation at natural
fracture intersections. Fractures turned as they grew and Figure 6-19 Change in individual fracture width,
cumulative fracture width and fracture radius with
developed with en-echelon multiple branches. Recently a change in the equivalent number of simultaneous
developed direct fracture diagnostic techniques, such as propagating multiple fractures (from Weijers et al., 2000b)
220
Figure 6-19 shows an example of the way in which cause of this pressure rise depends on the fracture
the simultaneous growth of multiple competing geometry. For fractures that grow parallel to each
fractures changes the fracture dimensions of a radial other, the net pressure increases because the fractures
fracture, using the results from a commercial fracture compete for fracture width. Multiple fractures can
growth simulator that incorporates these effects in also lead to higher net fracturing pressures through
a simplified manner. The achievable fracture radius increased fluid leak-off due to increased fracture
decreases as the number of equivalent fractures area and the resulting rise in slurry proppant
increases because fluid and proppant has to be concentrations within the fractures.
shared by more fractures. The fracture width for The decrease in individual fracture width can
each individual fracture becomes smaller than the be problematic for wellbore-to-fracture connectivity
width of a single fracture, but the total width of because complex fracture growth may result in loss
all fractures combined increases with an increasing of effective fracture conductivity and an increased
number of multiple fractures. Production will suffer sensitivity to pressure losses from non-Darcy and
in such cases. The smaller individual fracture widths multiphase flow. These effects are predominantly
can (and often do) lead to problems regarding the important if the fracture system is conductivity-
placement of proppant in the fracture resulting limited (e.g. high-permeability formations). If the
in a bridging screen-out. dimensionless fracture conductivity is very high (e.g.
The minimum fracture widths in Fig. 6-19 very low-permeability formations), complex fracture
were calculated using an admittance criterion growth may actually be beneficial because it may act
of three proppant grain diameters. In this to increase fracture inflow area.
example, the fracture will not accept 20/40-mesh When multiple fracture growth is predominantly
proppant when there are more than 14 equivalent formation-induced, it is very hard to minimize the
simultaneously propagating multiple fractures, and it simultaneous propagation of multiple hydraulic
will not accept 12/20-mesh proppant when there are fractures. However, numerous techniques (Hainey
more than 6 equivalent simultaneously propagating et al., 1995) can help to minimize the number of
multiple fractures. The effect of an increase in simultaneous propagating fractures that initiate
the number of fractures, n, on the predicted net at the wellbore during the beginning of a propped
fracturing pressure, pnet, fracture radius, R, and fracture treatment (although the exact impact
fracture width, w, is as follows: of each of the completion changes listed below
2
may differ from case to case):
pnet ,n = pnet n 3 ,
(6-10) Limit the length of the perforated interval.
Utilize a high fracture fluid viscosity during fracture
initiation and the main fracture treatment.
2
Rn = Rn 9 ,
(6-11) Implement a cautious feeler gauge proppant slug
strategy with multiple proppant slugs to potentially
and plug off some of the multiple fractures.
Utilize finer sand for the first part or all of the
5

wn = wn , 9(6-12) propped fracture treatment.
Utilize the highest feasible injection rate to increase
where pnet,n, Rn and wn represent the actual net pressure, fluid efficiency.
fracture radius and fracture width (respectively) Utilize oriented perforations shot in the direction of
experienced with n multiple fractures instead the maximum horizontal stress.
of the predicted single fracture. Pick perforated intervals in areas with a good
Simultaneous propagation of hydraulic fractures cement bond and where natural fracture density
also results in higher net pressures and higher is relatively low.
surface pressures during the fracture treatment. The Minimize wellbore deviation whenever possible.
221
Modern Fracturing
References Jackson, R., and Fisher, K.: Successful Application

of Hydra-jet Hydraulic Fracturing on Horizontal
Abass, H.H., Meadows, D.L., Brumley, J.L., Hedayati, Wells Completed in a Thick Shale Reservoir,
S., Venditto, J.J.: Oriented Perforations A paper SPE 91435, 2004.
Rock Mechanics View, paper SPE 28555, 1994. Economides, M.J., and Nolte, K.G. (Eds.): Reservoir
Aud, W.W., Wright, T.B., Cipolla, C.L., Harkrider, Stimulation Third Edition, John Wiley and Sons,
J.D., and Hansen, J.T.: The Effect of Viscosity on New York, 2000.
Near-Well Tortuosity and Premature Screenouts, Ehlig-Economides, C.A., Tosic, S. and Economides
paper SPE 28492, 1994. M.J.: Foolproof Completions for High-Rate
Barree, R.D., and Winterfeld, P.H.: Effects of Shear Production, paper SPE 111455, 2008.
Planes and Interfacial Slippage on Fracture Growth Elbel, J., and Britt, L.: Fracture Treatment Design,
and Treating Pressures, paper SPE 48926, 1998. Chapter 10 in M.J. Economides and K.G. Nolte
Behrmann, L.A.: Underbalance Criteria for Minimum (Eds.) Reservoir Stimulation Third Edition, John
Perforation Damage, paper SPE 30081, 1995. Wiley and Sons, New York, 2000.
Behrmann, L.A., and McDonald, B.: Underbalance or Fisher, M.K., Wright, C.A., Davidson, B.M., Goodwin,
Extreme Overbalance, paper SPE 31083, 1996. A.K., Fielder, E.O., Buckler, W.S., and Steinsberger,
Bell, C.E., Holmes, B.W., and Rickards, A.R.: N.P.: Integrating Fracture-Mapping Technologies
Effective Diverting on Horizontal Wells in the To Improve Stimulations in the Barnett Shale,
Austin Chalk, paper SPE 26582, 1993. paper SPE 77441, 2005.
Brown, J.E., Thrasher, R.W., and Behrmann, L.A.: Fisher, M.K., Heinze, J.R., Harris, C.D., Davidson,
Fracturing Operations, Chapter 11 in M.J. B.M., Wright, C.A., and Dunn, K.P.: Optimizing
Economides and K.G. Nolte (Eds.) Reservoir Horizontal Completion Techniques in the Barnett
Stimulation Third Edition, John Wiley and Sons, Shale Using Micro-seismic Fracture Mapping,
New York, 2000. paper SPE 90051, 2004.
Chen, Z., and Economides, M.J.: Fracturing Pressures Freyer, R., and Huse, A.: Swelling Packer for Zonal
and Near-Well Fracture Geometry of Arbitrarily Isolation in Open Hole Screen Completions,
Oriented and Horizontal Wells, paper SPE 30531, paper SPE 78312, 2002.
1995. Gabriel, G.A., and Erbstoesser, S.R.: The Design of
Cinco-Ley, H., and Samaniego-V., F.: Transient Buoyant Ball Sealer Treatments, paper SPE 13085,
Pressure Analysis: Finite Conductivity Fracture 1984.
Case Versus Damage Fracture Case, paper SPE Gahan, B.C., Batarseh, S., Sharma, B. and Gowelly, S.:
10179, 1981. Analysis of Efficient High-Power Fiber Lasers for
Cleary, M.P., Wright, C.A. and Wright, T.B.: Well Perforation, paper SPE 90661, 2004.
Experimental and Modelling Evidence for Major Gas Research Institute, Technical Description GRI-
Changes in Hydraulic Fracturing Design and Field 95/0257 on Real-Data (Net Pressure) Fracture
Procedures, paper SPE 21494, 1991. Analysis (1995a).
Cleary, M.P., Johnson, D.E., Kogsbll, H-H., Owens, Griffin, L.G., Wright, C.A., Davis, E.J., Wolhart, S.L.,
K.A., Perry, K.F., de Pater, C.J., Stachel, A., Schmidt, and Moschovidis, Z.A.: Surface and Downhole
H., and M. Tambini: Field Implementation of Tiltmeter Mapping: An Effective Tool for
Proppant Slugs To Avoid Premature Screen-Out Monitoring Downhole Drill Cuttings Disposal.
of Hydraulic Fractures With Adequate Proppant paper SPE 63032, 2000.
Concentration, paper SPE 25892, 1993. Hainey, B.W., Weng, X., and Stoisits, R.F.: Mitigation
Dees, J.M. and Handren, P.J.: A New Method of of Multiple Fractures from Deviated wells. paper
Overbalance Perforating and Surging of Resin for SPE 30482, 1995.
Sand Control, paper SPE 26545, 1993. Handren, P.J., Jupp, T.B., and Dees, J.M.: Overbalance
East, L.E. Jr., Grieser, W., McDaniel, B.W., Johnson, B., Perforating and Stimulation Method for Wells,
222
paper SPE 26515, 1993. Removal, paper SPE 84353, 2003.

Hejl, K.A., Madding, A.M., Morea, M.F., Glatz, C.W., Robert, J.A. and Rossen, W.R.: Fluid Placement and
Luna, J., Minner, W.A., Singh, T., and Stanley, Pumping Strategy, Chapter 19 in M.J. Economides
R.R.: Extreme Multistage Fracturing Improves and K.G. Nolte (Eds.) Reservoir Stimulation Third
Vertical Coverage and Well Performance in the Edition, John Wiley and Sons, New York, 2000.
Lost Hills Field, paper SPE 101840, 2006. Rodgerson, J.L., Lopez, H., and Snider, O.: Unique
Karakas, M., and Tariq, S.: Semi-Analytical Production Multistage Process Allows Pinpoint Treatment of
Models for Perforated Completions, paper SPE Hard-to-Reach Pay, paper SPE 90052, 2004.
18427, 1988. Rodgerson, J.L., Ruegamer, M.L., and Snider, P.M.:
Lagrone, K.W., and Rasmussen, J.W.: A New External Casing Perforating Provides Optimal
Development in Completion Methods The Treatment Coverage in Horizontal Pay, paper SPE
Limited Entry Technique, JPT (July 1963) 15:7, 97175, 2005.
695702. Shlyapobersky, J., Wong, G.K., and Walhaug, W.W.:
Loman, G., Phillips, D.T., and Cicon, H.N.: Overpressure Calibrated Design of Hydraulic
Stimulation of Horizontal Wells Using the Fracture Stimulations, paper SPE 18194, 1988.
Dynamic Gas Pulse Loading Technique, paper Stadulis, J.M.: Development of a Completion Design
SPE 37097, 1996. to Control Screenouts Caused by Multiple Near-
Matthews, L.H., Schein, G. and Malone, M.: well Fractures, paper SPE 29549, 1995.
Stimulation of Gas Shales: Theyre All the Same Teufel, L.W.: Determination of In-Situ Stress From
Right? SPE 106070, 2007. Anelastic Strain Recovery Measurements of
McDaniel, B.W., McMechan, D.E., and Stegent, N.A.: Oriented Core, paper SPE 11649, 1983.
Proper Use of Proppant Slugs and Viscous Gel Underwood, P.J., and Kerley, L.A.: Evaluation of
Slugs Can Improve Proppant Placement During Selective vs. Point-source Perforating for Hydraulic
Hydraulic Fracturing Applications, paper SPE Fracturing, SPEDC (September 1998) 13:3, 151
71661, 2001a. 1564.
McDaniel, B.W., Stegent, N.A. and Ellis, R.: How Upchurch, E.R.: Near-Wellbore Halo Effect
Proppant Slugs and Viscous Gel Slugs Have Resulting form Tip-Screenout Fracturing: Direct
Influenced the Success of Hydraulic Fracturing Measurement and Implication for Sand Control,
Applications, paper SPE 71073, 2001b. SPEDC (March 2001), 16:1, 4347.
Minner, W.A., Du, J., Ganong, B.L., Lackey, C.B., Van de Ketterij, R.G. and de Pater, C.J.: Experimental
Demetrius, S.L. and Wright, C.A.: Rose Field: Study on the Impact of Perforations on Hydraulic
Surface Tilt Mapping Shows Complex Fracture Fracture Tortuosity paper SPE 38149, 1997
Growth in 2500 Laterals Completed with Van Dyke, P.E., Weijers, L., Fisher, M.K., and
Uncemented Liners, paper SPE 83503, 2003. Robertson-Tait,A.: Evaluation of Oil-industry
Minner, W.A., Wright, C.A. and Dobie, C.A.: Stimulation Practices for Enhanced Geothermal
Treatment Diagnostics and Net Pressure Analysis Systems: Lessons Learned from the Barnett Shale,
Assist with Fracture Strategy Evaluation in the Trans. GRC (2005), 29, 189195.
Belridge Diatomite, paper SPE 35696, 1996. Veeken, C.A.M., Davies, D.R., and Walters, J.V.:
Muskat M., The Flow of Homogeneous Fluids Through Limited Communication Between Hydraulic
Porous Media, 1946. Fracture and (Deviated) Wellbore, paper SPE
Palmer, I.D., and Veatch, R.W.: Abnormally High 18982, 1989.
Fracturing Pressures in Step Rate Tests, paper SPE Warpinski, N.R, and Branagan, P.T.: Altered-Stress
16902, SPEPE (August 1990). Fracturing, paper SPE 17533, JPT (September
Parker, R.A., Gahan, B.C., Graves, R.M., Batarseh, S., 1989), 41:9, 990997.
Xu, Z., and Reed, C.B.: Laser Drilling: Effects of Weijers, L., De Pater, C.J., Owens, K.A., and Kogsbll,
Beam Application Methods on Improving Rock H.H.: Geometry of Hydraulic Fractures Induced
223
Modern Fracturing
From Horizontal Wellbores, SPEPF, May 1994, Zemlak, W., Lemp, S., and McCollum, R.: Selective
8792. Hydraulic Fracturing of Multiple Perforated
Weijers, L., Wright, C.A., Demetrius, S.L., Wang, Intervals with a Coiled Tubing Conduit: A Case
G., Davis, E.J., Emanuele, M.A., Broussard, J.B., History of the Unique Process, Economic Impact
and Golich, G.M.: Fracture Growth and Re- and Related Production Improvements, paper
orientation in Steam Injection Wells, paper SPE SPE 54474, 1999.
54079, 1999.
Weijers, L., Wright, C.A., Minner, W.A. Molesworth,
G., Huckabee, P., and Roberts, G.: The rate step-
down test: A simple real-time procedure to diagnose
potential hydraulic fracture treatment problems,
paper SPE 62549, 2000a.
Weijers, L., Wright, C.A., Sugiyama, H., Sato, K.,
and Zhigang, L.: Simultaneous Propagation of
Multiple Hydraulic Fractures - Evidence, Impact
and Modeling Implications, paper SPE 64772,
2000b.
Weijers, L., Griffin, L.G., Sugiyama, H., Shimamoto,
T., Takada, S., Chong, K.K., Terracina, J.M.,
and Wright, C.A.: The First Successful Fracture
Treatment Campaign Conducted in Japan:
Stimulation Challenges in a Deep, Naturally
Fractured Volcanic Rock, paper SPE 77678,
2002.
Welling, R.W.F.: Conventional High Rate Well
Completions: Limitations of Frac&Pack, High
Rate Water Pack and Open Hole Gravel Pack
Completions, paper SPE 39475, 1998.
Wright, C.A., and Weijers, L.: Hydraulic Fracture Re-
orientation: Does It Occur? Does It Matter? The
Leading Edge (2001), 20, 11851189.
Wright, C.A., Weijers, L., Davis, E.J., and Mayerhofer,
M.: Understanding Hydraulic Fracture Growth:
Tricky but Not Hopeless, paper SPE 56724,
1999.
Wright, C.A., and Conant, R.A.: Hydraulic Fracture
Re-orientation in Primary and Secondary Recovery
from Low-permeability Reservoirs, paper SPE
30484, 1995.
Wright, C.A., Stewart, D.W., Emanuele, M.A.,
and Wright, W.W.: Re-orientation of Propped
Refracture Treatments in the Lost Hills Field,
paper SPE 27896, 1994a.
Wright, C.A., Conant, R.A., Stewart, D.W., and Byerly,
P.M.: Re-orientation of Propped Refracture
Treatments, paper SPE 28078, 1994b.
224
Dr. D.V. Satyanarayana ("Satya") Gupta is product line research leader for fracturing technology at BJ
Services Company, working out of the Corporate Technology Center in Tomball, Texas. He has over 31 years
of oil field chemical product development and applications experience. Previously, Gupta was laboratory
manager for Fracmaster Ltd. and manager of chemical technology and chief chemical engineer for the
Western Co. He also worked for Pennzoil Technology and Gulf South Research. In his present position, he is
responsible for technology development and applications in the area of fracturing fluids, breakers, additives
and proppants. His main personal research interests are green chemical technologies and unconventional
fracturing fluids. He has chaired over 10 symposia on green corrosion inhibitors and has presented talks on
unconventional fracturing fluids all over the world. He is a member of SPE, NACE, ACS, AIChE and the
Petroleum Society of Canada. He is on the SPE editorial board and was on the editorial board of JCPT from
1995 to 2002. He has published over 40 papers and is an inventor on over 70 international and US patents.
He has a DSc degree in chemical engineering from Washington University, St. Louis, Mo.
Dr. Peter P. Valk is a professor of Petroleum Engineering and holder of the L.F. Peterson Professorship
at Texas A&M University. A native Hungarian, Valk holds BS and MS equivalent degrees in chemical
engineering and technical mathematics from Veszprm University, Hungary. He received his PhD
("Candidate of Sciences") in chemical engineering from the Institute of Catalysis, Novosibirsk, Russia, in
1981. Before joining Texas A&M in 1993, he was an adjunct professor at the Mining University in Leoben,
Austria; worked for the Hungarian Oil Company (MOL); and was a faculty member at Etvs University,
Budapest, Hungary. A Steering Committee member of the 1999 and 2006 Forum Series and previous
member of the Editorial Board of SPE Journal, Valk is an active participant in the SPE. At Texas A&M,
he teaches advanced hydraulic fracturing, well completion and stimulation, petroleum numerical methods
and general engineering courses. His research interests include design and evaluation of hydraulic fracture
stimulation treatments, rheology of fracturing fluids, performance of advanced and stimulated wells and
the underlying numerical methods. He published several dozen research papers in well-known journals.
Chapter 7 oil. Common to early oil-based fracturing fluids was
sensitivity to variations in the components of the base
oil, which affected viscosity development. In general,
Fracturing Fluids and the early oil-based fracturing treatments were limited
Formation Damage in size due to the cost of the fluid.
D.V. Satyanarayana Gupta, BJ Services and Oil-based fracturing fluids were the mainstay
Peter Valk, Texas A&M University of the industry until the 1960s, when economic
and safety considerations moved the industry
toward water-based systems.
7-1 Introduction The development of natural gas as a resource also
necessitated the evaluation of alternative fluids. By 1962,
In 1949, J.B. Clark published a description the use of water-based fracturing fluids surpassed oil-based
of the Hydrafrac process for increasing well fracturing fluids (Hassebroek and Waters, 1964). Initial
productivity. Included in the requirements of this concerns of formation compatibility with water-based
innovative process was a listing of the necessary fluid fluids (i.e., clay swelling and dispersion) were overcome
properties, which were: by the addition of salts such as potassium chloride,
(1) sufficient viscosity to create a fracture and calcium chloride or ammonium chloride. A second, more
transport the proppant, subtle reason that perhaps was not well-understood at the
(2) compatibility of the fluid with the formation to time was that the main rationale for using hydrocarbon-
minimize formation damage, and base fracturing fluids in water sensitive formations was
(3) a reduction in fluid viscosity after the proppant is not so necessary. Fracture face (leakoff) damage has little
placed to maximize fracture conductivity. impact in low- to medium-permeability reservoirs.
To obtain the desired viscosity to transport
Other than a recognition that viscosity has a minor the proppant, a variety of water-soluble polymers
influence on fracture geometry, the basic requirements of were developed based on either polysaccharides
hydraulic fracturing fluids have not changed in the past (starch, cellulose and guar) or synthetic polymers
60 years. However, many advances in the chemistry and (polyacrylamide). Over time, the guar-based
composition of these fluids have been made to achieve polymers became the dominant products to
predictable and reliable performance under increasingly viscosify water-based fracturing fluids based on
challenging conditions. Ultimately, the chemical availability, cost and performance.
composition of the hydraulic fracturing fluid dictates Water-based fluids viscosified with guar or guar-
the fluid properties and performance. This chapter derivatives provided acceptable performance within
will discuss the common components of hydraulic a limited temperature range. However, as hydraulic
fracturing fluids and the interrelationship between fracturing was applied in reservoirs with higher
fluid chemistry and fluid performance. In particular, temperatures, thermal thinning of the linear polymer
fracturing fluid formulations suitable for fracturing solutions resulted in increased screen-out frequency. Early
natural gas wells will be emphasized. efforts to expand the temperature limit of the linear gels
Due to concerns about formation compatibility focused on using ultra-high concentrations of polymer
with water, the earliest fracturing fluids were greater than 100 pounds per 1000 gallons (Alderman,
hydrocarbon-based and utilized napalm to obtain the 1970). However, the ultra-high concentration of polymer
required viscosity (Clark, 1949, and Hendrickson et al., caused excessive friction pressure and conductivity
1956). Napalm comprises a hydrocarbon liquid such damage, due to the high concentration of polymer
as gasoline or kerosene viscosified with an aluminum fragments left within the proppant pack. To improve
salt of naphthenic and palmitic acid. The flammable fluid performance at lower polymer loading, crosslinkers
hydrocarbons were subsequently replaced with other were developed to increase fluid viscosity and increase
more viscous refined oils and later with gelled crude the temperature limit. Delaying the action of crosslinkers
227
Modern Fracturing
improved the friction pressures, and lowering the polymer Specialty fluids that include methanol in the
load also improved proppant pack permeability. A variety water, 100% methanol, and liquid CO2-based
of compounds were found to effectively crosslink guar- fluids were all developed for specialty applications,
based polymers under limited conditions of pH and particularly to stimulate gas wells in formations with
temperature. Refinement of crosslinker compounds has severe water-sensitivity issues (See Sections 7-4.5.3,
resulted in the current family of products that can be 7-4.5.4, 7-4.5.5, 7-4.5.6 and 7-4.5.7).
used over a wide pH range as well as extreme temperature From a simple beginning of pumping small,
conditions. Better-quality polymers also helped reduce batch-mixed sand slurries, hydraulic fracturing has
the polymer loading further. evolved into a science of creating complex fluids with
In addition to single-phase oil-based or water- customized properties for each well. This chapter
based fluids, foams and emulsions have been used will discuss the current products used to create
in hydraulic fracturing treatments. The most widely fracturing fluids that provide controlled, predictable
used foams are composed of common fracturing viscosity development and desired conductivity for
fluids containing either carbon dioxide or nitrogen effective reservoir stimulation.
and a foaming agent (surfactant) (see Section 7-4.4).
Foaming a fracturing fluid improves the rheological 7-2 Fracturing Fluid Function
properties of the fluid and provides increased energy
to enhance cleanup and flowback of the fracturing The main functions of a fracturing fluid are to create
fluid after the treatment. This has been particularly and extend the fracture, transport proppant through
useful in under-pressured gas reservoirs. the mixing and pumping equipment and into the
With the development of more stable fluids, the fracture, and place the proppant at the desired location
impact of gelling agent residue on proppant-pack in the fracture. Failure to adequately perform any one
permeability has become a greater concern. To promote of these functions may compromise the stimulation
controlled degradation of viscosified fluids, breakers benefit of the treatment. This section examines the
have been developed that can be added to the fracturing role of the fracturing fluid in creating the fracture
fluid prior to pumping into the well. In recent years, the and transporting the proppant. Typical methods of
development of delayed-release breakers and polymer- evaluating fluids to confirm that they meet these
specific enzyme breakers has allowed greater application requirements are also detailed.
of breakers (see Section 7-5.1.6). In addition to new
breakers, the concern with polymer damage to proppant 7-2.1 Fracture Initiation
pack permeability has led to the development of new
viscosifying agents, such as viscoelastic surfactants. To create and extend a fracture, fracturing fluids
Viscoelastic fluids, because of their elastic properties, transmit hydraulic energy from the surface
can transport proppant with lower viscosity. In tight pumping equipment to the target reservoir. A low-
gas applications, fluid leak-off was not a major concern, viscosity fluid such as a brine or oil can be used
which made these viscoelastic fluids very cost-effective. to create a fracture (see Section 4-1). However, in
The relatively low-molecular-weight surfactant molecules a formation with natural fractures or reasonable
form micelles in water that interact, increasing fluid permeability, most of the fluid may leak off into the
viscosity. The viscosity development can be reversed by formation, thus giving up the ability of the fluid to
exposure to certain salts or hydrocarbons (see Section effectively transmit this energy to the formation.
7-4.5.1). The stimulation of low-permeability natural These fluids may also have high frictional pressure
gas wells, where proppant transport is not of primary losses, a factor that contributes to the efficiency of
importance, can be served properly by low-viscosity, low- the fluid. Adding friction reducers can minimize
cost fracturing fluids. Very large volumes of these fluids the frictional losses, and increasing viscosity
with small proppant slurry concentrations injected at increases the efficiency of the fluid in transmitting
high rates can create thin, long fractures. energy to the fracture.
228
Chapter 7 Fracturing Fluids and Formation Damage
In tight gas applications, it is also necessary to increased fluid drag forces on the proppant particles.
be able to create a very long fracture. See Section 4- During the second phase, the bank grows only in
7.1 for the appropriate design for low-permeability height until it reaches equilibrium height over its full
reservoirs. Other than controlling leak-off into natural length. Finally, in the third phase, the bank grows only
fractures that may be present, high viscosity is generally in length, and the injected proppant saltates over the
considered unimportant to create these long, thin full length of the bank toward the banks front, where it
fractures. However, there may be other reasons for settles, increasing the length of the bank in the direction
relying on viscosity to create and extend a single fracture of flow. The analytical relations derived for each of
(Aud et al., 1994). Cleary et al. (1993) discussed these phases in bank build-up has been confirmed by
the types of near-wellbore tortuosity that may exist experiments in a transparent model (Schols and Visser,
and the mechanisms that may control the initiation 1974; and Blot and Medlin, 1985).
characteristics of fractures. Near-wellbore tortuosity Water fracs (using slickwater or water with friction
can be generally attributed to two mechanisms or a reducers instead of gel to transport proppant) have
combination of the two: Multiple fractures competing been a successful fracturing technique in some tight
for opening space in the same region; and complexity gas reservoirs. Proppant placement is an essential factor
of the fracture pathway leading from the wellbore to that determines the effectiveness of such hydraulic
the far-field fracture region (see Section 4-2.1.4). fracturing treatments. Liu and Sharma (2005)
investigated the impact of fracture width and fluid
7-2.2 Proppant Transport rheology on proppant transport, including particle
settling and horizontal transport. Their results show
An equally important function of fracturing fluids is that the settling velocity drops significantly as fracture
the transport of proppant into the fracture. Various width approaches proppant diameter. They presented
mechanisms can be responsible for the transport of data to show that proppant usually flows at a lower
the proppant: i) when the settling velocity of the horizontal velocity than the fluid, particularly in
proppant is negligible, the slurry behaves as a perfect narrow fractures. The proppant is either retarded or
suspension, and the solid moves effectively with the accelerated depending on the ratio of the proppant size
slurry fluid velocity; ii) when the settling velocity of the to the fracture width. It has been found that for a single
proppant is significant, a proppant bank is created, and particle, when this ratio is small, the proppant travels
its top is continuously sheared off by the high-velocity faster than the average fluid velocity at that location
slurry above the bank, so the solid moves towards the because the proppant tends to be confined to the center
fracture tip with a slower mean velocity (Novotny, of the flow channel where the fluid velocity is higher.
1977). The transition between the two mechanisms As the proppant size increases, the effect of the fracture
depends mainly on two factors: the apparent viscosity walls becomes more important, and the walls retard the
of the fluid (at the settling conditions) and the density proppant. The retardation of particle relative to fluid is
difference between the proppant material and the greater for larger particles and greater proximity to the
fluid (Aboud and Melo, 2007). fracture walls due to the hydrodynamic stress exerted
In low-viscosity fluids, proppant is transported by on the sphere by the walls in the narrow gap.
stationary bed saltation flow, which is characterized With viscous fracturing fluids, proppant transport
by the deposition of a bed of proppant followed by is a much more complex phenomenon. The industry has
saltation flow of the proppant slurry above the proppant considered viscosity a major requirement for proppant
bed. Laboratory evaluations indicate that building transport and has used rheometers to calculate power
a proppant bank occurs in three consecutive phases. law parameters and require a minimum viscosity at
During the first phase, the bank builds up gradually 40, 80, 100 or 170 sec1 shear rate, depending on
as a function of time until an equilibrium height is previous experience, as a requirement for proppant
reached near the wellbore. The bank stops growing transport. Several industry tests have concluded, based
at this point as a result of the erosion caused by the on laboratory proppant transport testing, that borate
229
Modern Fracturing
crosslinked fluids were perfect proppant transport themselves to understanding the settling behavior of
fluids (Harris, 1988). Several initial proppant transport single particles or a single slurry concentration under
tests were done with single particles. It was disclosed mostly static fluid conditions.
that the static single-particle tests might not reflect Goel and Shah (2001), in a large-scale proppant
real-life multiple-particle transport under dynamic evaluation facility, looked into the dynamic settling
conditions. The proppant transport capability of a behavior of borate crosslinked fluids of various viscosity
crosslinked fracturing fluid is considered to occur due and elastic parameters. Based on their testing, they
to entanglements formed in a non-crosslinked polymer concluded that crosslinked gels exhibiting satisfactory
or to crosslink bonds created in a crosslinked gel (Harris proppant transport behavior had dissimilar viscosity
and Heath, 1998). These bonds provide a strong network and viscous moduli. All the gels that exhibited good
structure to the fluid and significantly improve on the proppant transport behavior had similar elastic
base polymers ability to keep proppant in suspension. moduli at low frequencies. Thus, they concluded
Thus, the network structure formed in a fluid should that elastic modulus was a better property than
be correlated with its particle suspension behavior. The viscosity to correlate the rheological characteristics of
discrepancy in describing particle suspension with fluid the fluid to proppant transport through a fracture.
viscosity either in static or dynamic conditions shows The author has made similar observations with
that the viscosity does not correctly depict the network viscoelastic surfactant gel fracturing fluids.
structure formed in fracturing fluids.
Borate crosslinked guar exhibits, besides non-
Newtonian behavior, viscoelastic characteristics with 7-3 Fracturing Fluid Rheology
both elastic and viscous components (Constien,
1989). Viscoelasticity, also known as anelasticity, Fluid selection for hydraulic fracturing treatments is
describes materials that exhibit both viscous and based on a variety of performance criteria that impact
elastic characteristics when undergoing deformation. the design, execution and cost of the treatment, as
Viscous materials resist shear flow and strain linearly well as the eventual well productivity. The rheological
with time when a stress is applied. Elastic materials properties of the fracturing fluid play a crucial part
strain instantaneously when stretched and just as because they directly affect the performance of the
quickly return to their original state when the stress fluid with respect to almost all the fluid functions
is removed. Viscoelastic materials have elements of listed above. Rheological properties can be directly
both of these properties and, as such, exhibit time- used to calculate frictional pressure losses in the
dependent strain. The elastic properties are considered wellbore, perforations and fracture. From the
to better describe the network structure formed in engineering point of view, the frictional pressure losses
borate crosslinked fluids and provide an alternative can be described using steady-state approximations,
technique to better describe the suspension properties and so the rheological characterization of common
of the fracturing fluid (Kramer et al. 1987; and fracturing fluids is performed with a steady-shear
Knoll and Prudhomme, 1987). viscometer (see Section 4-4.1.2).
Acharya (1986, 1988) observed that most fracturing The ultimate goal of a laboratory measurement
fluids are not just non-Newtonian power law fluids (see series is to provide the steady-state flow behavior
Section 4-4.2.2) but are really viscoelastic in nature in the form of a shear stress versus shear-rate curve,
and theorized that the elastic nature of the fluid may or rather to represent the behavior by providing
be more important. However, because most fracturing the parameters of a simple rheological model
design models consider fracturing fluids as power law representing the same information.
fluids, the industry has not truly considered the elastic In general, the steady-state rheological model is in
effects of the fluids. Very few attempts were made to the form:
correlate fluid elasticity with proppant suspension
(Jin and Penny, 1995). Even these authors confined = f ( ) , (71)
230
where is the shear rate measured in s-1, and (The field unit for the consistency index is lbfsn /
is the shear stress measured in Pa (or lbf/ft2). The ft2. To convert the consistency index into SI, where
shear rate describes the intensity of the flow that the unit for viscosity is measured in Pas 1000
is, the change of velocity with location. The shear times larger than the traditional cp the number
stress is associated with internal frictional forces should be multiplied by 47.88. For instance, if the
emerging during the flow, and it is the ultimate K =0.1 lbfsn /ft2, its SI value will be K = 4.788 Pasn,
cause of frictional pressure loss. If we compare two whatever the actual value of n is.)
fluids in the same flow geometry and at the same Another model called Yield-Power Law or Herschel-
flow rate, the shear rate distribution is essentially the Buckley model, reads as:
same, but the shear stress at the perimeter will be
n
larger for the thicker fluid. Dividing both sides of = y + K ( ) (75)
Eq. 7-1 by , we obtain:
a version of which was presented already in Section 4-
f ( )
= = a , (72) 4.2.2 as Eq. 4-96. This attempts to describe the existence

of a minimum stress for flow and contains an additional
where a is the apparent viscosity. Many fluids parameter, the yield stress, y, in psi. The Yield-Power
manifest a simple linear relationship between Law is quite general. It reduces to the Power Law model
and . Such fluids are called Newtonian and are (y = 0) or to the Bingham plastic model and with
described by the simplest rheological constitutive n = 1 and y > 0, where it is customary to use plastic
equation with constant viscosity: viscosity, p, instead of K to the Newtonian model
(n = 1 and y = 0). Unfortunately, the Herschel-Buckley
= . (4-92) model makes the solution of the flow equations difficult;
moreover, actual fracturing fluids seem to follow another
As previously discussed, the various mechanisms threeparameter rheological model named after Ellis
creating an internal structure increase the flow (Guillot and Dunand, 1985):
resistance, but at larger shear rates more and 2
(0 )
more of this extra resistance is lost. This is why =0 n
, (76)
K ( ) + 0
fracturing fluids are typically shear-thinning. One
of the simplest models describing this behavior where 0 is the viscosity at zero shear rate. However,
is called the Power Law model: the actual flow behavior is even more complex and
the description can be improved by increasing the
n
= K ( ) (73) number of parameters.
100
60F
where n is the flow behavior index (dimensionless) 80F
100F
125F
and K the consistency index. There is a difference 150F
between Eq. 7-3 which is the generalized power
Viscosity, Pas
law and others that use primed values of K and 10-1

n such as Eq. 4-95 in Section 4-4.2.2. Primed Newtonian
Region
constants assume a specific geometry, and they are
different for pipes or fractures. Power Law
Region
For shear-thinning fluids, the flow behavior
10-2
index is between zero and one. At n =1 the model 10-2 10-1 100 101 102 103 104
Shear Rate, s
-1
becomes Newtonian. Clearly, for a power law fluid
the apparent viscosity is: Fig. 7-1 Apparent viscosity as a function of shear rate
for a water-based (40 lb/Mgal) HPG polymer solution.
Solid lines correspond to the best fit of Eq. 7-6 (after
n1
a = K ( ) . (74) Guillot and Dunand, 1985)
231
Modern Fracturing
For most engineering calculations, the power then substitute it in Eq. 7-4 and use the Hagen-
law region is of primary importance, and it is not Poissuelle solution for Newtonian fluid (provided in
surprising that the simplicity and flexibility of the the second column of Table 7-1).
Power Law model has become the standard way of More rigorous calculations require the exact
characterizing fracturing fluids. It is common to solution of the flow equation for the Power Law case.
provide the parameter pair (n and K) as a function A closed-form solution for the pressure loss gradient is
of temperature and time (elapsed from the first given in the third column of Table 7-1. Also provided
encounter of shear). Commercial fracture design in the table is the equivalent Newtonian viscosity, e,
software packages provide this information from defined as the viscosity one would use in the Newtonian
their built-in fluid databases. In fracture design, Hagen-Poissuelle law to obtain the correct pressure
one needs to calculate routinely pressure loss in the loss gradient. Notice that the apparent viscosity from
fracture, proppant settling velocity in the fracture, Eq. 7-7 and the more rigorous equivalent Newtonian
and pressure loss in the pumping path, and so a viscosity do not coincide (Savins, 1958).
knowledge of n and K is essential. In engineering calculations it is often more
straightforward to compute the pressure loss calculation
7-3.1 Pressure Loss Gradient in the Fracture using the friction factor and Reynolds number. In terms
of friction factor, the pressure loss is given by:
The flow geometry in the fracture is often idealized
as parallel plate geometry. In such a geometry it is p = f vavg
2
/ w. (78)
advantageous to use the average flow velocity (flow L
rate divided by cross-sectional area, vavg = q/Ac), to A version of Eq. 7-8 was given as Eq. 4-101, Section
describe the flow intensity and represent the pressure 4-4.4.2 in field units.
loss gradient, p/L, as an expression containing the For laminar flow, the relation between Reynolds
rheological parameters and the channel width, w, in number and friction factor is also included in
addition to vavg. In almost all calculations, laminar flow the table. For a first glance, this is only formally
regime is assumed in the fracture. different from the direct calculation discussed above,
For less rigorous calculations, one can use an and even it has the drawback of including another
approximate formula for wall shear rate: quantity, the density. The use of the Reynolds
number, however, has the definite advantage to
6vavg
w = , (77) provide an additional way to check the validity of the
w assumption regarding laminar flow.
Table 7-1. Pressure Loss Gradient for Laminar Flow Between Parallel Plates
Newtonian Power Law

n

p 12vavg p 1 + 2n
Pressure Drop = = 2n+1 n
K w( n+1) vavg
L w2 L n

Equivalent n
e = 2n1 1 + 2n n 1
Newtonian e = K w1n vavg
Viscosity 3 n
Fanning
Friction factor
N Re = 2wvavg / e
f = 24 / N Re
Equivalent
Reynolds Number N Re = 2wvavg / e =2wvavg / e
N Re
Note: The Fanning friction factor for pipe flow is 16/NRe for laminar flow, as given in Section 4-4.4.2, Eq. 4-102.
232
In some fracture models, the cross section is a) Low gel loading
approximated by an ellipse with large aspect ratio.

Then the geometry is characterized by the width
at the center, w0, and the corresponding entries in
Table 7-1 take a slightly modified form. For instance,
the equivalent Newtonian viscosity will be given by
(Valk and Economides, 1985): 600
500
Na, 400
n
1 + ( 1) n
n 1 400
= 2
cp
Kw01n vavg
n 1
. (79) 200

e
n
0 300
H, 1/s
50 200
Example 7-1 Viscosity and Pressure Loss Gradient in 100

time, min
100
the Fracture
(Note: The rheological equations in Chapter 7 and Table b) High gel loading
7-1 are given in consistent SI units, as is done in the vast

majority of the fluid flow and rheology literature. To use
them with field units, it is recommended that field units
be converted to SI units first, then used in the equations
and then the results translated back into field units. For
example, velocity in ft/s becomes m/s by multiplying it 600
10
by 0.305; calculated viscosity will be in Pa.s, which can Na, 400
cp 8
be converted to cp by multiplying by 1000, etc.) 200
A typical Borate crosslinked hydroxypropyl- 0 6
w, mm
guar (HPG) gel is considered at a relatively high 50 4
temperature. Table 7-2 shows the measured (n and
time, min 100
K) pairs at a low and a high gel loading, as a function 2
of the shearing time. Figure 7-2 Apparent viscosity as a function of shear rate
and shearing time
Table 7-2, Power Law parameters for a typical
crosslinked HPG at 105 oC In the figure we use a cut at 200 cp, and the
t, min n K n K lower gel loading goes below that value for larger
shearing times. Such visualization can be made even
loading: 0.42 % (35
loading: 0.60 % (50 lb/Mgal) more practical if we use fracture width and calculate
lb/Mgal)
apparent viscosity from Eq. 7-6.
0 0.540 3.92 0.495 6.40
To this end, the injection rate and the fracture
30 0.564 3.26 0.508 5.83 height are necessary. For an injection rate of 20 bpm
60 0.589 2.62 0.520 5.09 (0.053m3/s) and fracture height of 50 ft (15.2 m),
the average velocity will be only a function of the
90 0.607 2.09 0.527 4.49
fracture width (Fig. 7-3).
120 0.625 1.62 0.532 3.77 The apparent viscosities shown in Fig. 7-3 provide
satisfactory information about the conditions that can
In general we know the ranges of shearing provide at least 200 cp viscosity. For instance, for the
time and shear rates the fluid will experience in the lower gel loading at shearing time, t = 60 min and fracture
fracture, and hence we can visualize the apparent width, w = 0.3 in., the calculated apparent viscosity (from
viscosity as in Fig. 7-2. Eqs. 7-4 and 7-7) is a=259 cp.
233
Modern Fracturing
a) Low gel loading a) Low gel loading
600 600
10 10
Na, 400 Na, 400
cp 8 cp 8
200 200
0 6 0 6
w, mm w, mm
50 4 50 4
time, min 100 time, min 100

2 2
b) High gel loading b) High gel loading
600 600
10 10
Na, 400 Na, 400
cp 8 cp 8
200 200
0 6 0 6
w, mm w, mm
50 4 50 4
time, min 100 time, min 100

2 2
Figure 7-3 Apparent viscosity as a function fracture width Fig. 7-4 Apparent viscosity as a function fracture
and shearing time for a fixed injection rate width and shearing time (with doubled injection rate,
40 bpm (0.106 m3/s)
The rigorous calculation for these conditions using 7-3.2 Rheology in the Presence of Proppant
Table 7-1 results in e = 293 cp. The corresponding Material and its Relation to Settling
pressure loss gradient is 0.962 psi/ft (21.7 kPa/m).
If this pressure loss is calculated through the friction When proppant is added to the fracturing fluid, the
factor/Reynolds number approach, then NRe = 8.9, rheological behavior of the slurry will change and the
clearly indicating laminar flow. calculated pressure loss gradient will increase. However,
For comparison, the higher gel loading would this effect can be often neglected or taken into account
result in a pressure loss gradient of 1.29 psi/ft using relatively simple assumptions and correlations.
(29.1 kPa/m) and the corresponding Reynolds More complex is the phenomenon of proppant
number is NRe = 6.6. settling. The starting point for settling calculations is
The final part of this example is the effect of the the Power Law version of Stokes law for the terminal
injection rate. For instance, if the fluid injection rate velocity of a settling sphere:
is doubled, the lower gel loading is hardly sufficient
1/ n
even at low shearing times, i.e., the viscosity is less ( ) gD n+1
vt =
p f prop
than 200 cp (see Fig. 7-4). n +1 , (710)
23 K
234
where, in addition to the acceleration due to gravity Here, f is the (Fanning) friction factor and should be
and the density difference between proppant and used in conjunction with:
fluid, the particle diameter plays an important role.
p 4
For actual proppant transport several additional = f ( 12 vavg
2
), (7-12)
L D
questions should be raised. For instance, can the usual
n and K parameters be used in the shear rate region where D is the diameter of the pipe (or equivalent
representative for the falling particle? Roodhart et diameter of the flow path) and vavg is the average velocity.
al., (1985) found that a better description of the In fact, the base gel itself, and any additional fluid loss
settling process would need the Ellis model (Eq. 7-6). reducing agents are used to approach the MDRA and
Another issue is that in the settling process, unless the thereby reduce the friction factor as much as possible.
particle is exactly in the center of the flow channel In the original work by Virk, the Reynolds number
the flowrate-induced shear is superimposed on the was relatively easy to calculate because the small quantity
one imposed by particle settling. In addition, several of the added polymer did not change the viscosity of
investigators studied the effect of fracture walls, the the base fluid. For high polymer loads, calculating
interaction of individual proppant particles, etc. the Reynolds number is far from trivial. A reasonable
The interested reader is referred to recent reviews by suggestion is to use the wall Reynolds number:
Brannon et al., (2005) and Clark (2006).
1 + 3n Dvavg
N Re, w = , (713)
4n
7-3.3 Impact of Fluid Rheology on Fluid Loss e
where
Rheology is one of the main factors responsible for n
controlling the loss of fluid into the formation during e = 2n3 1 + 3n KD1n vavg
n1
. (714)
n
the fracturing treatment. However, the apparent
viscosity having the direct influence is not of the Example 7-2 Calculation of Pressure Loss Gradient
fracturing fluid but of the filtrate penetrating the rock in Tubing
matrix (Williams, 1970).
What is the frictional pressure loss gradient, if
7-3.4 Calculation of Pressure Loss in the Wellbore 40 bpm (0.106 m3/s) of fluid with specific gravity
Using Rheological Parameters and the Virk 1, power law parameters n = 0.5 and K= 0.044 lbf.
Maximum Drag Reduction Asymptote sn /ft2 (2.1 Pa.sn ) is pumped through a 2.76-in.
(7 cm) inner diameter tubing?
The Power Law parameters are also used to carry out First the intermediate results are calculated: vavg
engineering calculations for pressure loss gradient = 90.35 ft/s (27.54 m/s), = 1000 kg/m3, e = 41.8
in the wellbore. Because the flow regime in the cp (0.0418 Pa.s) and NRe,w = 57,600. Then Eq. 7-11
wellbore is usually non-laminar, one has to use the yields the friction factor f = 0.00111. Substituting
concept of Reynolds number and friction factor into Eq. 712, the pressure loss gradient is obtained
(or should revert to direct empirical correlations as 24.1 kPa/m (1.06 psi/ft).
interpolating actual measurements). In general, the For comparison, if the injection rate is reduced
non-laminar flow of thick fracturing fluids can be by a factor of two, the pressure loss gradient decreases
considered similar to turbulent flow of Newtonian to 9.86kPa/m (0.436psi/ft).
fluids with some degree of suppression of the
turbulence. The basic engineering result is due to 7-3.5 Advanced Rheology
Virk (1975), who determined the maximum drag
reduction asymptote (MDRA): Apparent viscosity dependence on shear rate is only
one of the many phenomena related to flow of
f -1/2 = 19.0 log10 ( N 'Re, w f 1/ 2 ) 32.4. (711) fracturing fluids. Viscoelasticity is present to a certain
235
Modern Fracturing
degree in all major fracturing fluid families (not only in

the commercially named VES fluids) and is usually f = 16 , (7-16)
N Re,VE
inferred through dynamic oscillatory measurements
(Prudhomme, 1988). Such measurements are and for non-laminar flow it can be obtained from Eq.
quite reproducible, and the results can be used to 7-11, where again,
characterize the fluid and even to obtain the steady-
1 + 3n
state viscosity dependence on shear. N Re, w = N , (717)
4n Re,VE
7-3.6 Foam Rheology to convert into wall Reynolds number.
Foams are used extensively for various reasons, such as Example 7-3 Calculation of Pressure Loss Gradient
minimizing water-induced damage in water sensitive for Foam Flow in Tubing
formations, their positive effect on clean-up, favorable
energy transfer in the wellbore, good proppant carrying For simplicity, the calculations of Example
capability, etc. Their flow behavior is strongly affected 7-2 are used, assuming that the injected fluid
by foam quality, defined as the in situ ratio of dispersed is 70% quality N2 foam.
gas volume to fluid volume. The gas is N2, CO2 or a As previously, volumetric flow rate q = 40
mixture of the two. Typical foam quality in the fracture bpm (0.106m3/s), D=2.76in. (7 cm) and
is between 0.65 and 0.85. The rheological Eqs.7-3 to 7- base liquid density is .
5 are still applicable, but the two or three parameters (n, For illustration purposes, the volume-equalized
K and y) involved are generally considered a function of flow behavior index n=0.43 and the consistency
quality. Statistical methods are used to provide empirical index is KVE = 0.044 lbf.sn /ft2 (2.1 Pa.sn ). In general,
correlation for the quality dependence. Influential the volume-equalized n and KVE parameters are not
works of this type are for instance Reidenbach et al. very different from those of the base fluid.
(1986) and Sudhakar and Shah (2004). Because somewhere in the pipe the quality is 70%,
A departure from the foam quality based at that point the superficial liquid velocity is 30% of
correlations was suggested by Valk et al. (1992), the velocity calculated in Example 7-2; therefore vl =
Winkler et al. (1994), and Enzendorfer et al., (1995). 0.3 (27.54) = 8.26 m/s.
Their volume-equalized power law model automatically Now the volume-equalized Reynolds number can
accounts for quality changes and requires many fewer be obtained from Eq. 7-15:
parameters : the power law exponent, n, and the
23n D nl vl 2n
volume-equalized consistency index, KVE (compared N Re,VE = n
= 12.1103.
1 + 3n
with other models that require quality, etc.) The KVE
n
advantage of the volume-equalized approach is that in
steady-state isotherm pipe flow, the Reynolds number NRe,VE is then converted into the wall Reynolds
remains constant along the pipe and can be calculated number using Eq. 7-17:
using only liquid properties: 1 + 3n
N Re, w = N = 16.1103.
2 3n n
D l vl 2n 4n Re,VE
N Re,VE = n
,
1 + 3n
KVE (7-15) Thus, the flow is laminar and the corresponding
n
friction factor from Eq. 7-11 is f = 0.00206.
where l is the liquid density and vl is the superficial To use Eq. 7-12, the average density of the foam
velocity of liquid, defined as liquid flow rate and the overall foam velocity, v, are needed. For instance,
divided by total cross-sectional area. The resulting assuming 3000 psi (20 MPa) and normal temperature,
friction factor is also constant along the pipe; for the gas density will be g = 3.6 lb/ft3 (225 kg/m3) and the
laminar flow it is given by: foam density is calculated by:
236
102
1 3 3
= = 4.7lb/ft (293 kg/m ).
1 1 175 Nm
0.3 + 0.7 Smaller
f g Particles
N, Pa.s
100
750 Nm
This also means that the overall velocity is
1000/293 folds larger than the superficial liquid
10-2
velocity and hence: 10-1 100 102 104 106
H, s-1
1000
v= vl = 28.2 m/s. Figure 7-5 Effect of particle size on viscosity (from Hill
293 and Carrington, 2006)
Substitution of the friction factor, foam density and The fluid is shear thinning, i.e., viscosity decreases
foam velocity into Eq. 7-12 yields: at higher shear rates. The viscosity tends to be higher
p 4 with smaller particles. The changes in particle size
= f 1
foam vavg , foam are associated with an increase in particle number if
L D 2
all other system variables are kept constant. Smaller

particles, therefore, result in an increase in the number
The obtained friction factor can be used for of particle-to-particle interactions and an increase in
subsequent calculations at various locations resistance to flow. As shear rate increases, this effect
along the pipe, but the foam density and total becomes less marked. This suggests that any particle-to-
foam velocity should be recalculated because of particle interactions are relatively weak and are broken
the varying pressure. down at high shear rates. Figure 7-6 shows a Newtonian
system in the absence of the particles. Adding coarse
7-3.7 Effect of Proppant on Rheology particles (large-size proppant) results in an increase in
viscosity, but the system remains Newtonian.
Proppant size, proppant size distribution, and the
102
volume fraction of proppants present influence 28.6% fine talc
the rheology of slurry. Zeta potential, a measure (D50 = 5 Nm)
of the charge on particles in the system, also has

N, Pa.s
a marked effect; however, for proppants it is not 100 28.6% coarse talc
(D50 = 19 Nm)
an issue. The relationships between rheology,
especially viscosity, and particle size parameters
0% talc
are explained below.
For slurries, the volume fraction and the maximum 10-2
10-2 100 102
volume fraction can also influence viscosity. Maximum H, s -1
volume fraction (the highest volume of particles that can Figure 7-6 The effect of particle size on viscosity for a
be added to a fluid) can be thought of as the amount of Newtonian fluid (from Hill and Carrington, 2006)
free space the particles have in which to move around. Its
impact on viscosity is discussed below. Adding finer particles (smaller-size proppant)
results in a further, more significant increase in
1. Particle size. If the total mass of proppant viscosity, particularly at low shear rates. Colloidal
particles in slurry is kept constant but the particle repulsion among a relatively large number of particles
size is reduced, then the net effect is an increase in gives structure to the fluid, increasing resistance to
the number of proppant particles in the system. The flow. As in the previous example, this relatively weak
impact on viscosity of such a change, across a range of structure is broken down at high shear rates: The fluid
shear rates, is shown in Fig. 7-5. has become shear-thinning.
237
Modern Fracturing
2. Volume fraction. The effects of volume fraction and In this system, at the lowest volume fraction, the
maximum volume fraction on viscosity are described system is almost Newtonian. As volume fraction increases,
using the Krieger-Dougherty equation (Hackley and shear-thinning behavior becomes evident. Increasing the
Ferraris, 2001): volume fraction results in a higher degree of particle-to-

[ ] f s ,max particle interaction, and the resistance to flow increases.
slurry f
= 1 s . (7-18) The forces between particles are, however, broken down
fluid f s ,max
at high shear rates. A further transition in flow behavior
where slurry is the viscosity of the slurry, fluid is the occurs as volume fraction is increased to values above
viscosity of the base fluid, fs is the volume fraction of about 50% of maximum volume fraction. At these solids
proppants in the slurry, fs,max is the maximum volume loadings, the free movement of particles is significantly
fraction of proppants in the slurry and [] is the hindered as collisions between particles increase and the
intrinsic viscosity of the particles, which is 2.5 for system effectively becomes more congested. As shear rate
spherical proppant. Non-spherical particles will have is increased, the particles are trying to move more rapidly
higher values for []. The value of [] is also affected and thus the effect becomes more pronounced. Viscosity
by the particle size distribution. therefore increases with shear rate; the system is shear-
Equation 7-18 indicates that viscosity increases thickening at very high shear rates.
with increasing volume fraction. As the volume
fraction of solids in the system increases, the particles 3. Particle size distribution. Particle size distribution
become more closely packed, it becomes more influences the way particles pack together. A polydisperse
difficult for them to move freely, particle-to-particle particle population with broad size distribution will pack
interactions increase, and resistance to flow (viscosity) more closely than a monodisperse sample with a narrow
increases. As the volume fraction nears the maximum size spread. The effects on viscosity can be explained
that the sample can contain, viscosity rises very with reference to the Crier-Doughty equation (Eq. 7-
steeply. In addition to influencing the absolute value 18). For a monodisperse sample, the maximum volume
of viscosity, volume fraction also affects the nature of fraction is around 62%. With a polydisperse sample, the
the relationship between shear rate and viscosity for smaller particles can fill the gaps between larger particles
the system, i.e., flow behavior. The parameter fs,max is a and the maximum volume fraction is greater, more like
function of particle shape, particle size distribution and 74%. Therefore, increasing the particle size distribution
shear rate. Slurries with relatively low volume fraction for any given volume fraction of solids will result in
tend to behave as Newtonian fluids; i.e., viscosity reduced viscosity. This can be viewed as lubrication of the
is independent of shear rate. As volume fraction is movement of larger particles by smaller particles. Particle
increased, shear-thinning behavior is observed. This size distribution can be an especially valuable tool for
transition is illustrated in Fig. 7-7. manipulating viscosity when the volume fraction of the
system is fixed. Viscosity as a function of fraction of large
102
or small particles in solution is shown in Fig. 7-8.
61% Solids
Increase All Maximum All

Volume 750 Nm Volume Fraction 175 Nm
Fraction
N, Pa.s
100
55% Solids
Viscosity
45% Solids
10-2
10-1 100 102 104
0% Increasing amount of 175 Nm 100%
H, s -1
100% Increasing amount of 750 Nm 0%

Fig. 7-7 Effect of particle volume fraction on viscosity
(from Hill and Carrington, 2006) Figure 7-8 Effect of particle size distribution on viscosity
238
In this example, a synergistic effect can be 2. Dynamic Measurements: Most fracturing fluids
observed when particles of both sizes are present show elastic as well as viscous properties, as explained
at a certain concentration. The resulting viscosity before. The measurement of sample viscoelasticity can
is lower than that achieved using a monodisperse provide vital information not given by viscometry
sample of either sized particles. alone. With a rheometer, tiny movements (small
Keck et al. (1992), based on field-sized yard tests, strain oscillations) can be used to measure viscoelastic
came up with correlations that were valid for linear and properties without destroying the sample structure.
delayed-crosslinked HPG gels for predicting the effect of Oscillatory testing generates a mechanical spectrum
proppant on viscosity and friction pressures. for the material, and this provides a unique behavior
fingerprint. Typically, the viscometer can measure in
7-3.8 Laboratory Rheology Measurements the range of about 0.1 to 5000 s-1 while a rheometer
extends the measurement range from 10-6 to 105 s-1.
In the laboratory, the shear stress and shear rate values Clark (1979) explained the use of dynamic rheological
can be obtained with a variety of instruments. These evaluations of fracturing fluids in detail.
can be viscometers or rheometers. Most viscometers
operate by rotating a spindle in the sample. Viscosity 3. Laboratory Instruments: The first viscometer
is determined by measuring resistance to this commonly used in the industry was a Marsh funnel,
rotational force. Viscometers, in comparison to a conical-shaped funnel fitted with a small-bore tube on
rheometers, are usually relatively simple instruments. the bottom end through which fluid (originally drilling
Their simplicity of design and operation can mud) flows under gravity (Fig. 7-9). A screen over the
offer advantages for operator ease of use. Spindle top removes large particles that might plug the tube.
movement in a viscometer is in one direction, which
allows the measurement of viscosity. Rheometers
can apply oscillatory and rapid step changes in stress
and strain, and can therefore determine viscoelastic
properties (providing information on the structural
properties of the sample) as well as flow properties.
Viscometers employ a mechanical bearing that limits
the speed and torque capabilities of the instrument,
whereas rheometers generally use a low-friction air
bearing. The residual friction from the mechanical
bearing can make the measurement of low-viscosity
materials difficult. Some of the latest viscometers use
a variety of measurement geometries, enabling them
Figure 7-9 Marsh funnel
to cover broader range of viscosity. Rheometers, while
generally more expensive than viscometers, are more In the test standardized by API for evaluating
versatile and have a much wider dynamic range for water-base and oil-base muds, the funnel viscosity
control and measurement parameters. measurement is the time (in seconds) required for one
quart of mud to flow out of the Marsh funnel into
1. Measuring Viscosity: Rheometers function across a a graduated mud cup. Funnel viscosity is reported
very wide range of shear rates enabling the simulation in seconds (for a quart). Water exits the funnel in
of real processes that occur over vastly different time about 26 seconds. This test was one of the earliest
scales. Shearing occurs whenever fluids flow through mud measurements for field use. Simple, quick
pipes, perforations and fractures. In any of these and foolproof, it still serves as a useful indicator
configurations, the velocity is zero right at the wall of change in mud by comparing mud-in and
surface and maximum at the center. mud-out sample funnel viscosity.
239
Modern Fracturing
The next simple viscometer used in the industry is Gdanski et al., (1991) detailed the use of this
a simple couette viscometer (Fig. 7-10). instrument for measuring the rheology of titanate
crosslinked HPG fluids. Even though limited rheology
can be gleaned from measurements from this instrument
at a limited range of shear rates, it cannot evaluate the
effect of shear history of the fluids and the effect of
proppant on the viscosity. Craigie (1983) used a shear
history simulator for pre-conditioning the fluid prior to
testing with Fann 50 instrument. The capillary viscometer
consists of a fluid pump, a series of tubes and an additive
pump for injecting crosslinker. The length of the tubing
and the pump rate can be adjusted to simulate shear
histories in the tubulars during fracturing treatments.
As the fluid exits the tubular simulator, it flows into a
concentric-cylinder viscometer (i.e., Fann 50), where
selected fracture shear history and temperature conditions
are simulated. In this part of the apparatus, the long-term
degradation rates of the crosslinked fracturing fluids are
determined. Figure 7-12 displays a capillary reciprocating
viscometer, where the fluid is reciprocated in a capillary
and the pressure drop is measured and viscosity is
calculated from the pressure drop.
Figure 7-10 Fann 35 couette viscometer
Without means of heating it and keeping fluids

from evaporation, it is mainly used as a quality control
tool for base gels. The next-generation instrument
overcame this by having heating chamber with oil
and pressurized test chamber to keep solutions from
evaporating (Fig. 7-11).
Figure 7-12 Reciprocating capillary viscometer
Kraynik et al. (1984) developed a helical-screw

rheometer (HSR). Lord (1988) and Lord and Shackleford
(1989) described the application of the HSR for
evaluating stimulation fluids. This HSR is a single-screw
Figure 7-11 HTHP viscometers (Fann 50 and Chandler
5500) extruder composed of a single-flight metering screw with
240
a constant helix and a small, uniform channel depth. This

screw rotates at a constant speed with minimal clearance
inside an outer cylinder. Pressure-drop measurements
across the barrel are made with a closed-discharge valve.
Fluid rheology is determined from the differential
pressure and rotation rate. The HSR can measure
fluids containing particulate because fluid circulation
in the HSR provides a homogeneous sample, which
is an advantage over both conventional rotational and Figure 7-13 Foam flow loop setup
pipe viscometers. Although straightforward shear stress

and shear rate cannot be measured, a semi-empirical
approach extracted power law parameters from the
pressure-drop/rotational speed data.
De Kruijf et al., (1994) described another type of
HSV. This HSV has a helical screw impeller rotating in
a cylindrical glass draught tube mounted coaxially in a
cylindrical vessel. In this apparatus, the screw circulates Figure 7-14 Mixing tank, flow meters and additive pumps
approximately 1300 mL of test fluid around the draught
tube. The HSV measures rotation rate and torque. The
screw and draught tube were modeled as a system of
two coaxial cylinders to allow power-law calculations.
This HSV was reported to effectively measure the
rheological properties of proppant-laden fracturing fluids
because (1) its measuring gap is large in comparison
to the proppant-particle diameter, and (2) the fluid is
continuously circulated so that proppant particles cannot Figure 7-15 Flow tubes of the foam flow loop
settle to the bottom of the rheometer.

Despite these developments, the Fann 50
viscometer and its equivalent rheometers are still
widely used for measuring the rheology of fracturing
fluids. For this reason a method was sought for
measuring rheology when proppant was present in
the sample. Thesing (2000) developed a helical screw
bob that can be used with the Fann 50 to prevent the
settling of proppant in the viscometer cup.
Pipe rheometers or flow loops for measurement
of rheology use the principle of pressure drop across
various diameter pipes to correlate viscosity and power
law parameters. In addition to evaluating conventional
fracturing fluids, these have been useful to evaluate the
rheologies of foams and emulsions in single-pass or
recirculating mode. Figures 7-13 through 7-16 show
the foam flow loop to measure viscosity and foam
textures of foam fracturing fluids. Hutchins and Miller
(2003) described the development of a circulating foam
rheometer for measuring the rheologies of foams. Figure 7-16 View cell for evaluation of foam texture
241
Modern Fracturing
7-4 Types of Fracturing Fluids The most cost-effective solution is to fracture the
formation with the simplest of fluids. Low-viscosity
As has been described in Sections 7-2 and 7-3, the water or hydrocarbon with the fewest additives would
fracturing fluid is one of the important components for be the simplest fluids. However, these have very low
a successful hydraulic fracturing treatment. The required proppant transport properties, very little leak-off
properties and functions of the fracturing fluid, as detailed control and, if pumped at high rates, will result in
previously, can be accomplished by several means. A unacceptable friction. Friction can be controlled by
variety of conventional and unconventional fluids can be various additives, but if the formation has adverse
used for the purpose. These fluids can be water-based, saturation effects (see Section 7-6.2) even in tight gas
hydrocarbon- or oil-based, methanol-based or may not formations with very little leak off desired stimulation
contain any liquids at all. may not be achieved. Compatibility with clay containing
Why are there so many fluids? How does one select formations can be controlled by using salts in the
the right fluid for the right application, for example, gas fluids. Depending on pumping conditions, i.e., the
wells? This section will answer these questions. To start shear regime the fluid would experience, there may be
with, a variety of different formations with different need for shear-tolerant or shear-recoverable fluids. For
base lithologies need to be stimulated. Additional higher-temperature applications, these can be achieved
mineral components will make the compatibility of the by the use of organometallic or borate crosslinked
fluids with the formations unique. In addition, there water-based fluids and crosslinked oil-based fluids. If
may be swelling and/or migrating clays, which would the gas formations are under-pressured, the fluids can
require alternative fluids or additives. The formation be energized with N2 or CO2 or foamed with N2 or
permeability may determine the fluid loss of the fluid CO2 or a combination of the two. The foam fluids also
and the resultant efficiency of the fluid to create and provide good leak-off control. If compatibility with
propagate the fracture. The formation could contain water is an issue potentially due to wetting issues, the
different fluids in situ. Some situations will be dry gas use of viscoelastic surfactant fluids can be considered.
wells. Some will have condensate with the gas. Some They also do not damage the proppant pack and can
may have water produced alongside. The salinity of also be energized or foamed. If incompatibility is due
the water may play a role. All these permutations and to capillary and unloading issues, methanol-containing
combinations necessitate different fluids. fluid can be considered. If the incompatibility is severe,
In a tight gas application, long, thin fractures then crosslinked methanol based fluid or liquid CO2
are usually required (see Section 4-7-1). In a higher- based fluid may be the answer.
permeability formation, the need may be to create a Obviously, these various fluids have different
short, wide fracture, often with frac-pack applications rheological properties, which need to be considered
(see Sections 4-7.3 and 4-7.4). The fluid may before selection can be made. As will be explained later,
experience different pumping configurations. We may polymer loading may not have a bearing on the final
want to pump the fluid down casing or down tubing viscosity of the crosslinked fluid (see Section 7-4.1). It
or through the annulus or through coiled tubing. The is possible for a higher-loading polymer to yield lower
shear regime the fluid sees is different in these different crosslinked viscosity at temperature than a fluid with
scenarios. These considerations make the selection of lower loading of polymer.
the fluid very critical for optimum stimulation of a To pump a fracturing treatment through
particular well or a formation. coiled tubing, one would use a shear-resistant or a
Conventional fracturing fluids include water-based delayed crosslinking fluid to minimize friction and
and polymer-containing fluids; hydrocarbon-based fluids, provide good crosslinked fluid properties after the
energized fluids and foams. Unconventional fracturing fluid exits the coiled tubing.
fluids include non-polymer-containing fluids such as To stimulate wells in a coalbed methane formation,
viscoelastic surfactant fluids; methanol-containing fluids where production is by desorption, the fracture has
and liquid CO2-based fluids. to connect as many natural cleats as possible and
242
produce maximum possible inflow area (see Chapter 7-4.1.1 Low-Viscosity Fluids
11). However, if these natural fractures or cleats were
large, then fluid leak-off and the resultant damage to The friction-reduced water (slickwater) fracturing
the desorbing surface would be an issue. fluid introduced in the late 1950s and 1960s has come
back into vogue, particularly in shale fracturing. It is
7-4.1 Water-Based Fluids possible to create long fractures in shale reservoirs with
a large quantity of water pumped at very high rates
The water-based fluids can range from plain water with with minimal proppant and create partial proppant
a friction reducer to minimize horsepower needed for monolayer fractures (see Sections 8-1.3 and 8-1.4).
pumping to a complex crosslinked polymer fluid with Without friction reducers, at the high pump rates
a variety of additives. Water is relatively inexpensive employed, friction would be very high.
and widely available in most areas of the world; A partially hydrolyzed acrylamide polymer is
however, the quality of the water sometimes will not the most common friction-reducing agent used with
meet the minimum standards required for fracturing water for this purpose. The polyacrylamide polymers
(see Section 9-4.1 for details). Modern fluids can be are generally used as oil-external emulsions with
pumped in batch-mix or continuous-mix modes (see either a built in inverter or an additional inverter that
Section 9-3 for equipment requirements). Rheological flips the emulsion and hydrates the polymer, giving
properties (viscosity, for example) can be adjusted as some viscosity and reduction in friction properties.
desired very easily by adjusting polymer loading and Because viscosity is not needed for proppant transport
additive loading even during the job if required either under these turbulent conditions, the fracturing
in stages or continuously. treatment can be achieved at lower costs than would
In under-pressured gas wells, energized fluids are be possible to create similar conductive frac lengths
needed to help fluid recovery. Fluids energized with with crosslinked or foamed fluids.
CO2 will have low pH because of the solubility of Linear gels are another method of achieving low
CO2 and resultant carbonic acid in solution; therefore, viscosity. These use guar or guar derivatives as viscosifying
fluids that can be crosslinked and used at low pH agents. The guar derivatives include hydroxypropyl
are of interest (see also Section 7-4.3). The polymer guar (HPG), carboxymethyl guar (CMG) or
breaks down at low pH due to acid hydrolysis, and carboxymethylhydroxypropyl guar (CMHPG).
this helps in cleaning up the proppant pack.
In cold weather applications, e.g., Canada, Alaska 7-4.1.2 Crosslinked Fluids
and Russia, sometimes it is not possible to heat the
water in winter. Adding methanol can keep the base Crosslinking is the most cost-effective way of increasing
fluid from freezing. In certain formations, adding the viscosity of the fluid. It is an easy method of increasing
methanol to the fracturing fluid is necessitated to reduce the molecular weight of the polymer. Water-based
adverse capillary effects (see also Section 7-6). However, fluids can be crosslinked at high- or low-pH conditions.
un-derivatized guar is not soluble in methanol; it Low-pH crosslinked fluids are particularly suited for
can tolerate very little methanol even after hydration energizing or foaming with CO2. To achieve the same
(Ely, 1994). Some guar derivatives can be hydrated in viscosity at higher reservoir temperatures, one can use
aqueous alcohol solutions, and a special version of HPG an order of magnitude less polymer by crosslinking than
can be hydrated in non-aqueous methanol (see also with linear polymer in aqueous solutions. Historically,
Sections 7-4.5.3 and 7-4.5.5). these fluids typically required polymer loading from 40
Guar and guar derivatives also have low to 80 pounds per thousand gallons (pptg). However, with
residue, which is an additional benefit in reducing the higher-yielding guars and with better crosslinking
proppant pack damage. The latest un-derivatized technology, the loadings have dropped to as low as 12
guar has residue of less than 2%, and the derivatives pptg on the low end and as high as 35 pptg on the high
have as low as 0.5% residue. end for most applications.
243
Modern Fracturing
7-4.1.3 Borate Crosslinked Fluids fluids were limited in application to wells below
220 F. The newer high-temperature fluid can be
The most common water-based fluid is borate used to temperatures exceeding 300 F and can be
crosslinked (Fig. 7-17). formulated with high concentrations of KCl for clay
compatibility (Brannon and Ault, 1991; and Wang
CH2OH
et al., 2002). Figure 7-18 shows the new high-yield

CH2OH
guar-based borate fluid that gives similar viscosity at
HO O HO O 200 F at 24 pptg as previous guar-based borate fluid
HO
HO that needed 30 pptg. Figure 7-18 also shows that
O
O
acceptable viscosity for fracturing can be achieved at
HO
HO polymer loading as low as 16 pptg. Conventional guar
CH2
H2C
CH2
O O O O cannot be crosslinked at these low polymer loadings
H
H
O
H
O
H (to understand why, see Section 7-5.1.2).
H H H H H H H H
HO OH O O HO OH
O O 1000
B B 900
30 ppt Polymer A
800 16 ppt Polymer B
Viscosity, cp @ 100 sec-1

O O HO OH O O HO OH 18 ppt Polymer B
700 20 ppt Polymer B
24 ppt Polymer B
H H H H H H H H
600
O H O H O HO H
O O O O 500
CH2 CH2 CH2 400
300
Figure 7-17 Chemical structure of borate crosslinked guar
200
100
Borate fluids are shear-recoverable; i.e., the 0
fluid will shear degrade but will recover its viscosity 0 30 60 90 120 150 180 210 240
Time, min
when shear is removed. This is possible because
the crosslink mechanism is based on hydrogen Figure 7-18 Viscosity of borate crosslinked fluids with
bonding. In industry testing, it was found that a standard guar and high-yield guar at 200 F
properly designed borate fluid provides nearly perfect
proppant transport because these fluids have very 7-4.1.4 Metallic Ion Crosslinked Fluids
high low- or zero-shear viscosity (Asadi et al., 2002).
Borate crosslink is reversible if the pH drops below Metallic ion (organometallic) crosslinked fluids have
8, and so if there is water production in the well and also been used both at high pH and low pH. Prior
if the produced water has neutral or low pH, borate to 1995, organometallic crosslinkers included the
crosslinked fluids will uncrosslink. The breaker zirconium-based crosslinkers used today (Fig. 7-19)
designed to break the crosslinked gel will break plus aluminum, titanium, chromium and others. Due
the uncrosslinked gel much more efficiently, and to their lack of high-temperature stability, shear stability
the fluid will hence clean up extremely well. Being and environmental unfriendliness, all of them except
high-pH fluids, they are not compatible with CO2 zirconium-based ones have disappeared from the scene.
but can be energized with N2. Prior to the development of carboxylated derivatives
As detailed in the next section, there are of guar (e.g., CMG or CMHPG), the industry used
various ways of delaying the crosslink system with guar and HPG to crosslink with the metal crosslinkers.
borate. Until the development of organoborates CMG or CMHPG proved to be more versatile in
(as opposed to mono-borate salts) and borate ester applications with metallic crosslinked systems and
crosslinkers (where the borate ion is released by the are the common polymers used with them now.
hydrolysis of the borate ester polymer), the borate Originally, metallic crosslinked systems were developed
244
for high-temperature applications because mono- Low-pH fluids are also usually based on either
borate (molecules containing one borate ion) fluids CMG or CMHPG. At low pH, guar and HPG degrade
did not function well above 160 F. The zirconium by acid hydrolysis very easily, whereas the carboxylated
crosslinked fluids are generally shear-tolerant; however, polymers are relatively resistant to acid hydrolysis.
if excessive shear is experienced, the gels will shear The most common crosslinker used is zirconium.
degrade and will not shear recover. Typical high-pH The crosslinkers are typically chelated (reversibly
metal crosslinked fluids are either based on CMG bonded to a complexing agent) and can be either
or CMHPG and typically use low polymer loading, water- or alcohol-based. These types of fluids have
from 15 to 35 pptg. Higher loadings can be used for been successfully used for temperatures approaching
temperatures of 350 F and above. 450 F (Carman, 2007). The carboxylated polymers,
e.g., CMG, typically do not perform very well in the
O
Zr presence of KCl, and their best performance is limited
to use with KCl substitutes.
O
O
7-4.1.5 Delayed-Crosslink Systems
Zr
H+
There are a couple of distinct schools of thought
O
on delayed-crosslink fluids. Time or temperature
Zr OH
can delay the crosslink time. The main purpose
of delaying the crosslink time is to minimize the
friction avoid having to pump a highly viscous
O
HOR
fluid because of the resultant high horsepower
Zr requirements. Also, delaying crosslinking through the
high-shear wellbore environment will minimize shear
O O Zr R=GUAR degradation and loss of fluid viscosity.
Zr
O
With time-delay mechanisms, delay additives
can adjust the crosslink time. However, if the fluid
Figure 7-19 Zirconium colloidal structure and possible
crosslink mechanism with guar (after Kramer and
reaches a temperature between 120 to 140 F, the
Prudhomme, 1987) fluid will start crosslinking almost immediately even
when designed to crosslink at longer delay times
Figure 7-20 shows comparative viscosity of typical high at room temperature. These types of systems are
temperature borate crosslinked with high temperature typically used in the field with crosslinking designed
zirconium crosslinked fluids. to occur when the fluid is two-thirds to three-fourths
of the distance down the tubular.
2000
50 ppt Borate Crosslinked - 325F The temperature-delayed systems are chelated
1800
after Wang, et.al., 2002 so that they will crosslink instantaneously when the
1600 50 ppt Borate Crosslinked - 335F
1400 after Wang, et.al., 2002 fluid reaches a certain temperature. Dual-crosslinked
40 ppt Zirconate Crosslinked - 320F
1200 50 ppt Zirconate Crosslinked - 360F systems are typically used with these types of
1000 crosslinkers, where an initial weak crosslink occurs
800
almost instantaneously at room temperature and
600
the second crosslink occurs at a certain temperature.
400
200
The initial crosslink helps in transporting the
0 proppant through surface equipment and down the
0 30 60 90 120 150 180
Time, min wellbore tubulars; the second crosslink is used to
Figure 7-20 Viscosity of borate and zirconate crosslinked
transport and suspend the proppant in the fracture at
fluids at high temperatures bottomhole temperature conditions.
245
Modern Fracturing
Figure 7-21 Progression of vortex closure in a blender jar with a delayed zirconium crosslinker added to a base gel
Figure 7-21 shows the progression of vortex 600

Fast crosslinker, low-shear mixing (A)
closure in a blender jar by the delayed crosslinker. Slow crosslinker, high-shear mixing (B)
Viscosity, cp at 170 sec-1

500 Fast crosslinker, high-shear mixing (C)
Figure 7-22 shows the early time viscosity build as
400 A
measured on a Fann 50 viscometer of the same fluid.
Figure 7-23 and Figure 7-24 show the effect of shear 300
on a borate crosslinked and zirconium crosslinked 200 B

fluid showing the ability of the instantaneous borate
100 C
crosslinked fluid to shear recover and instantaneous
0
zirconium crosslinked fluid losing viscosity on 0 2 4 6 8
shearing (Cawiezel, et al., 2004; Craigie, 1983). Time, hr at 250 F
Delaying the crosslink can minimize this effect in Figure 7-24 Effect of shear history on the viscosity of
metallic crosslinked systems. zirconium crosslinked fluid in Fann 50 (after Craigie, 1983)
1000
Immediate Crosslink 7-4.1.6 Function of Breakers in Water-Based Fluids
900
Delayed Crosslink
800 More Delayed Crosslink
700
There has been some speculation on the damage
600 caused by the adsorption of friction reducers in shale
500 formations and the partial monolayer proppant pack,
400 and the use of oxidizers and encapsulated oxidizers to
300
minimize the damage has been advanced (Carman and
200
Cawiezel, 2007). Although the friction reducer loading
100
0
used is very low, because of the large fluid volumes used,
0 30 60
Time, mins
90 120 this can be a significant problem, but one that can be
alleviated by using appropriate breakers.
Figure 7-22 Early-time viscosity build of a delayed The guar and guar-derivative polymers can be
zirconium crosslinked fluid in Fann 50 degraded with oxidizing and enzyme breakers, as is
1000 covered in detail in Section 7-5.1.6).
Batch Mix Sample
900 2 minutes at 7500 sec -1
Viscosity, cP @ 100 sec-1
4 minute at 7500 sec-1

800 4 minutes at 2600 sec-1 7-4.1.7 Water-Based Fluids in Gas Wells
700
600
Low-viscosity fluids have been successfully used in
500
400
fracturing shales and other tight gas formations where
300 economics are very important. The borate fluids tend to
200 have excellent fluid loss control and hence are applicable
100 in reservoirs with high fluid leak-off. They can be
0
0 30 60 90 energized or foamed with N2. These fluids have been
Time, min
successfully utilized globally in both oil and gas wells
Figure 7-23 Effect of shear history on the viscosity of and are particularly suited because of high viscosities
borate crosslinked fluid (after Cawiezel et al., 2004) that can be achieved where the wells have high tortuosity
246
and near-wellbore fracturing issues (Wright et al., 1993). 7-4.2 Oil-Based Fluids
Being perfect transport fluids, they are also suited for
proppant transport in deviated and horizontal wells. They In recent years, there has been an evolution of
are also used for coalbed fracturing in certain formations; technology for oil-base fracturing fluids. The technology
others react badly to borates (see Chapter 11). The low- has evolved from sodium-carboxylated associative
pH fluids can be energized or foamed with CO2 or N2 polymers to aluminum-carboxylated associative
and are particularly applicable to under-pressured gas polymers to aluminum-phosphate associative polymers
reservoirs and wells with severe clay issues. The fluids (Fig. 7-25). In the last decade, oil-based fluids have
compatible with methanol have been successfully used moved to iron-phosphate ester technology.
in gas wells with sub-irreducible water saturation issues The first oil-based fracturing fluids were prepared
discussed later in this chapter. The ability to adjust the using ingredients such as alkaline metal or aluminum
viscosity of water-based fluids either continuously or in carboxylates. Gelled materials of this type were described
stages may be of special interest in tight gas wells, where as napalm as explained in Section 7-1. These aluminum
high-viscosity fluid can be used in the pad to overcome carboxylate salts were dispersed in refined hydrocarbon
near-wellbore tortuosity issues and the viscosity materials such as gasoline, diesel and kerosene. In
minimized to be able to transport proppant (Ely et al., early years, aluminum carboxylate gellant chemistries
1993). Fracturing of high-temperature wells (over 300 demonstrated a very effective viscosity performance;
F) is restricted to water-based fluids. however, one disadvantage was the inability of aluminum
carboxylates to effectively gel various crude oils. It was
Sodium Carboxylate suspected that some crude oils did not contain the type
Association Polymer and concentration of aliphatic constituents necessary
R R
to form an effective gel. Generally this chemistry was
R
confined to the gelation of refined hydrocarbons due to
O C C C
O O O O O O
their instability in some crude oils.
Na+ Na+ Na+ Na+
The next gel systems were colloid dispersions,
Na+ Na+ Na+
O
based on soaps, which exhibited many preferred fluid
O O O O O O
C C C properties. Soap micelles were used instead of aluminum
R R R carboxylate gels because of their better stability in oils
containing asphaltenes. The micelles were aggregates of
Aluminum Carboxylate surfactant molecules dispersed in a liquid colloid. The
Association Polymer aluminum carboxylate gel fluids also exhibited extreme
R R
friction pressure. Oil gels utilizing soap colloid dispersion
C
O
O
H
O O
C
H
O
O technologies have provided a more consistent gelled
O
Al O Al O Al
oil system (Hendrickson et al., 1957; and Grantham
O O O O O
C C and McLaurine, 1986). Additionally, these fluids
demonstrated better sand transport properties without a
corresponding increase in viscosity.
Aluminum Phosphate
The most important characteristic of soap dispersions
Association
was that viscosity was lost at high shear, which is typical
R R
R
O
R R R of pseudoplastic or shear thinning fluids. Fluids of this
O O O
O P O O P O
type were prepared by delivering organic fatty acid
H H
Al O Al O Al materials in the hydrocarbon. A concentrated solution
O
P
O O O of an alkaline earth hydroxide (e.g., sodium hydroxide) is
P
O
H
O
R H
O O
R then dispersed in the hydrocarbon mixture to neutralize
Figure 7-25 Evolution of oil-based fracturing fluid
the organic acid in situ and cause the saponification
chemistry process. Henrickson et al. (1957) reported that in
247
Modern Fracturing
oil systems, micelles have the polar portions of the 7-4.3 Energized fluids
carboxylate molecule buried in a central core, rather
than distributed over the outer surface, as is the case Fluids are energized either with N2or CO2 to minimize
in aqueous systems. Polar terminations interact, either the amount of liquids introduced into the formation
by dipole attraction or by specific hydrogen bonding and to enhance recovery of the fluids. Fluids are
of certain atoms. Caustic solution also renders different typically considered energized if the volume percent
properties to soap micelle gels. Soap concentration and of the energizing medium to the total volume (defined
type, and hydrocarbon type have dramatic effects on the as quality) is less than 53%; they are considered
rheological performance of the fluid. as foams if the volume percent is greater (Mitchell,
Metal salts of alkyl acid orthophosphate esters were 1969). It is common in the industry to use 25 to 30
developed as an improvement over soap dispersion volume percent of an energizing medium in fluids.
hydrocarbon gel systems (Griffin, 1979; Poklacki, Also, adding the energizing medium reduces the base
1977; Caldwell and Sorrell, 1972; Monroe and Rooker, fluid (water or oil) pumped into the well. Energized
1970; Flanagan, 1966; and Pianfitti and Janey, 1962). fluids are of particular importance in fracturing under-
Aluminum alkyl acid orthophosphate systems proved to pressured gas wells. The majority of gas wells fractured
be the most effective system (Canterino, 1972). These in Western Canada use energized fluids. There are
aluminum salts included aluminum isopropoxide and certain differences between using N2 and CO2 as
hydrated aluminum. Alkyl esters are blends of mono-, energizing media, as explained below.
di-, and trialkyl esters. The dialkyl ester is the major In most places, N2 is easily available at lower cost
component, with monoalkyl and trialkyl esters being than CO2. N2 is chemically inert and is not miscible
present in lesser amounts in the reaction product. with formation fluids, which is an advantage because no
These systems were applied as gellants for fracturing chemical interference occurs. It is a gas under conditions
fluids by first dispersing the substituted aluminum alkyl of pumping and recovery. Because it has very low density
acid orthophosphate ester in hydrocarbon, followed by compared to the fluids it is pumped with, it will easily
the introduction of caustic solution or a strong amine separate from the fluid leaking off into the formation.
compound to activate the mixture. Phosphate esters may Thus it is imperative that fluids energized with N2 are
be monoesters, diesters, or triesters. The aluminum salts recovered as soon as possible after pumping ceases;
were later replaced with iron salts (Smith and Persinski, otherwise, the energy for fluid recovery might not
1995, 1996, 1997a, 1997b, 1997c). be present. Because N2 is not soluble in the fluids it is
The newer systems are very efficient in building pumped with and will not alter the pH, it can be used to
viscosity with lower additive loading. They can be energize most fluids. As a gas it, N2 undergoes significant
pumped either as batch-mix system or on the fly. They volume change when pressurized, and designing proper
build viscosity very rapidly and can tolerate moisture downhole quality may be difficult.
in the base oils. They are not as sensitive to the additive On the other hand, CO2 under pumping
concentrations and relative concentrations of the additives conditions is actually a liquid or a supercritical fluid. As
as the earlier aluminum based systems. This makes the a supercritical fluid, it is at a temperature and pressure
systems very easy to pump operationally. above its thermodynamic critical point of 88 F and
The gels are broken with slowly dissolving additives 1070 psi. As a supercritical fluid it has the unique ability
such as calcium oxide, magnesium oxide or carbonates. to diffuse through pores like a gas but with the density of
Some of these systems are also compatible with select the liquid. CO2 is chemically reactive. It is very soluble
crudes. However, any asphaltenic materials in the crude in both water and oil, and in water it creates carbonic
may react with the iron, and the gel will lose stability. In acid, which actually might work as a breaker to hydrolyze
certain formations in the Western Canadian sedimentary guar and guar derivatives. The liquid or supercritical
basin, hydrocarbon based gels with frac oils have been CO2 has density close to that of water, and hence the
successfully used to stimulate tight gas formations that energized or foamed fluid does not have lower density
are water sensitive (Gupta and Leshchyshyn, 2005a). than the base fluid, which helps in the hydrostatics. Until
248
the fluid is recovered by loss of pressure, it remains as a them foams. Foams are stabilized by adding a chemical
liquid or supercritical fluid whose density is close to that foaming agent and a base gel. For economic reasons,
of water, and so no density segregation occurs. The fluid the most common gel used is guar.
can be shut in for long time periods without losing the In addition to N2 and CO2 foams, binary foams were
energizing effect. Because it is soluble in fluids, the energy introduced to the industry in 1988 (Toney and Mack,
is stored for longer times. It also has the carbonated cola 1991). Binary foams contain both N2 and CO2 and are
effect: As more and more pressure is released, more and said to have advantages over either N2 or CO2 foams. The
more of the gas will be released from the fluid because the major advantage is the requirement that the recovered
solubility decreases with lowered pressure. This typically fluid have a low CO2 content so that the produced gas
results in better fluid recovery. can be put in the pipeline much sooner without flaring.
CO2, either as a liquid or gas, is considered non- Typically, N2 is lower cost than CO2; however, designing
damaging to the formation, particularly in a gas well. It jobs with two compressible fluidsone potentially
can precipitate asphaltene or paraffin in heavy oil, which involving phase changemay become complicated.
is usually not seen in gas wells. The high viscosity of the foams helps the fluids have
Because of the reactive nature of CO2, it cannot good fluid loss control. They are compressible and so
be used in all fluids. For example, borate fluids are have applications in gas wells with natural fractures and
not compatible with CO2, nor are certain viscoelastic higher permeability. They are also the fluids of choice for
fluids. In aqueous solutions, CO2 forms carbonic acid, many coalbed methane wells with good interconnected
which reduces the pH and uncrosslinks the borate cleats (see Section 11-4). They are suited for formations
fluid. In viscoelastic fluids neutralized with salts, with marginal water compatibility and where better
CO2 interferes with the salts, and micelle structure proppant-regained permeability is required. They
is lost along with viscosity. Also, not all oil gels are tend to have higher friction than conventional fluids,
compatible with CO2. Aluminum-phosphate ester particularly delayed cross-linked fluids. They also impose
systems are broken with low-pH carbonic acid, which a limit on the maximum amount of proppant that can be
functions as a breaker for the system. transported downhole because the proppant is added in
the tub whereas the N2 or CO2 is typically added at the
7-4.4 Foams and Emulsions wellhead. Foam viscosity can be adjusted by changing the
base gel viscosity and the quality of the fluid. Proppant
There is often a lot of confusion about what typically forms part of the internal phase, and so the
constitutes foams or emulsions. Generally, any fluid amount of N2 or CO2 added during the job needs to be
that has two phases and viscosity increase because of adjusted as the proppant concentration increases during
the presence of the second phase can be considered the job to maintain constant Mitchell or internal phase
a foam or emulsion. When the quality exceeds 53 quality (Mitchell, 1969). As mentioned before, the
(i.e., 53% by volume is gas), there is bubble-to- viscosity of the foam goes up exponentially as the quality
bubble interference, and system viscosity increases (volume fraction of internal phase) goes up. If the foam
exponentially until about 90 quality, when misting is pumped as a constant surface quality without regard to
starts to occur, dramatically reducing viscosity. The the added proppant, as the proppant volume in the fluid
typical foams or emulsions used in the industry increases, the viscosity will increase exponentially as the
are limited to about 70 to 80 quality to minimize proppant forms part of the internal phase of the foam.
frictional effects due to the high viscosity. When the This will result in unacceptable friction, particularly with
internal phase is a non-liquid, it is considered foam; higher proppant concentrations.
when it is a liquid, it is considered an emulsion. By The only damage to the formation and or the
this definition, all CO2 fluids with greater than 53 proppant pack from foams is from the gelling agent,
quality should be called emulsions. However, when which can be broken using enzymes or oxidizing
pumping ceases and the fluid is recovered, it comes agents. One of the potential problems using foams
out as foam, so it is common in the industry to call is that the recovered fluid can come back as foam,
249
Modern Fracturing
which might be a nuisance to surface equipment. The Air
use of splittable foamers that lose their ability to

function as foamers has been disclosed to overcome
this problem (Gupta and Hlidek, 2007).
Monomer L1-Phase L1-Phase M-Phase G-Phase

7-4.5 Unconventional Fluids Spherical Micelles Cylindrical Micelles Hexagonal LC Lamellar LC
Surfactant Concentration
CMC
As the industry moves to tighter and tighter
formations, particularly systems such as shales or Figure 7-26 Artists conception of the formation of
coalbeds where production is controlled by desorption ordered structures with VES surfactants
of the gas rather than matrix flow, fluids that are non-
damaging to the proppant pack and formation are As the concentration of surfactant increases in
becoming important. Wells with adverse capillary water, micelles start to form. Further increasing the
effects (see Section 7-6) due to sub-irreducible water concentration exceeds the critical micelle concentration
or hydrocarbon saturation also require different fluids (CMC) for the surfactant in water; these molecules
to minimize those effects or mitigate effects caused by start interacting with each other. These interactions are
drilling with the wrong fluid. Several unconventional based on ionic forces and can be amplified by adding
fluids have been developed and successfully used for electrolytes (salts) or other ionic surfactants. Depending
these unconventional formations in the last decade. on the ionic charges and the size and shapes of the
These fluids are described in this section. surfactants and these counter ions, ordered structures
start to form, which increases viscosity and elasticity.
7-4.5.1 Viscoelastic Surfactant Fluids The reverse mechanism is true for breaking these
systems. The structures can be disrupted by adding
Viscoelastic surfactant (VES) gel systems have other surfactants, ionic additives and hydrocarbons
been described in the patent literature for friction (from the formation or mutual solvents or other
reduction and as well treatment fluids (Teot, 1988). solvents) or can be diluted by additional formation
Its use in everyday life has been around for some water. The most common commercial systems use
time. Its use in fracturing fluids is relatively a new cationic surfactants with inorganic salts (Teot, 1988)
phenomenon, but the patent literature has exploded or with anionic surfactants (Zhang, 2002). Anionic
in this area in the last few years. surfactants with inorganic salts are also common
Principally, these fluids use surfactants in combination (di Lullo et al., 2002). Zwitterionic and amphoteric
with inorganic salts or other surfactants to create ordered surfactants in combination with inorganic salts have
structures, which result in not only increased viscosity been used (Dahanayake et al., 2004).
but also elasticity. These fluids have very high zero- The common VES fluids have a temperature limit
shear viscosity without undue increase in high-shear in the range of 160 to 200 F without foaming. High-
viscosity. Thus, they tend to be shear-degradable fluids. temperature stabilizers have been known to increase the
They also have high elasticity. As explained by Asadi et temperature limit to 250 F.
al. (2002), zero-shear viscosity has been found to be an Not all of these fluids are compatible with
essential parameter in evaluating proppant transport. CO2. They have been shown to be economic
Therefore, these fluids can, with lower loading, replacement for conventional borate fluids for tight
transport proppant without the comparable viscosity gas applications (Rieb, 2007).
requirements of conventional fluids. The technology of At least with one of these fluids, the flow-back
VES systems can be broken down into several categories water from these treatments can be recycled (Gupta
based on the structure of the ordered structures or and Tudor, 2005). This particular fluid uses a cationic
micelles. They can form worm-like micelles, lamellar surfactant neutralized with an anionic surfactant. The
structures or vesicles (Fig. 7-26). flow-back water, in gas wells, tends to return some of the
250
cationic surfactant and most of the anionic surfactant. and Leshchyshyn, 2005b). In formations with potential
The flow-back water is typically collected for 48 hours to form water blocks, these fluids are particularly suited
into a tank. Initially, the fluid was filtered to remove because the leak-off fluid still contains the surfactants,
any formation fines. Based on experience, it was found which lower surface tension in the matrix, overcoming
that allowing the fines to settle was sufficient to remove capillary forces and helping in recovery of the fluid.
the fines. After settling, the middle 75% of the flow- These fluids have been shown to be suited for fracturing
back water was transferred to a frac tank and the rest coalbed methane wells that contain water because
of the required water for the fracturing treatment the foams control leak-off into the cleats without the
was made up with fresh water. Using analytical or damage from polymer residue.
viscoelastic measurements, additional surfactants were
used to reconstitute the fluid. Russell (2001) reported 7-4.5.3 Emulsion of Carbon Dioxide
the procedure and well production results from using with Aqueous Methanol Base Fluid
the recycled fluid in field study in Canada showing no
effect of recycling on well production. Certain formations have potential to retain even limited
These VES fluids are operationally very simple water used in foams and VES foams of over 70 quality.
as only one or two additives are added on the fly These fluids may damage these sensitive formations
without any need for hydration of polymers. They do because of sub-irreducible water saturation and liquid
not require any biocides because they do not contain trapping (see Section 7-6.2.1). In these formations,
any biopolymers. In fact, some of the cationic replacing 40% of the water phase used in conventional
surfactants may have biocidal properties. They do not CO2 foams (emulsions) with methanol can minimize the
require additional flowback surfactants because they amount of water. Gupta et al. (1997) showed that a 40%
have inherently low surface and interfacial tension. methanol aqueous system yielded the highest viscosity
No additional clay control additives are needed: They of aqueous methanol mixtures, has a freeze point close
contain either salts or cationic surfactants, which have to 40 C (which is the lowest operating limit for
properties similar to KCl substitutes. The surfactants fracturing equipment in the field) and surface tension
have molecular weights of hundreds as opposed to the around 30 dynes/cm. These emulsions use surfactants,
guar polymer with millions. which are methanol-compatible foamers, in the place of
Viscosity is broken by altering the surfactant conventional foamers. Typical CO2 quality approaches
properties, by adding other hydrocarbons or by altering 85, which results in high regained permeability and rapid
the salinity or pH. The regain permeability with these clean-up and production results in several Canadian gas
types of systems approaches 100%. Because of the wetting formations (Gupta et al., 2007)
tendencies of the surfactants in some of the VES systems,
they are useful even in formations with sub-irreducible 7-4.5.4 Crosslinked Foams
water saturation and liquid-trapping issues even though
these fluids are aqueous-based (see Section 7-6). Crosslinked foams can be considered conventional
or unconventional fluids. There have been a variety of
7-4.5.2 Viscoelastic Surfactant Foams publications on the rheology and application of N2
foamed borate fluids since the early 1980s (Veatch,
A natural extension of VES fluid technology is the 1983; Harris, 1988; and Harris and Heath, 1996).
VES foams (Zhang et al., 2002). These foams can be The CO2-foamed metal crosslinked zirconium fluids
formed with N2 or CO2. As mentioned before, not all were used both with and without foamers (Phillips and
VES systems are compatible with CO2. No additional Mack, 1991). Most of these fluids use 70 or 75 quality
foamers are needed with these systems. The foam CO2 or N2 (or both, when binary fracturing).The
viscosity can be adjusted by adjusting foam quality and foams were crosslinked to increase the temperature of
the viscosity of the base VES system. They have been application of these fluids or to minimize the amount
successfully used in gas formations to 250 F (Gupta of liquids in the formation. In most of these systems,
251
Modern Fracturing
the viscosity increase is realized from the crosslink, The most recent development (Gupta, et al., 1997;
and viscosity dilution occurs at less than 53 quality. and Mzik, 1993 and 1994) describes a modified guar
The same principle can be used using a CMHPG- dissolved in anhydrous methanol crosslinked with a
based fluid containing 40% methanol crosslinked borate complexer and broken by an oxidizing breaker.
with zirconium crosslinkers. Performance similar to This system has been successfully used in the field.
40% methanol-containing CO2 emulsion fluid (see In under-pressured wells, it has been energized with
Section 7-4.6.3) described earlier can be achieved N2. There has been an interest in a CO2-energized
with these crosslinked systems containing only 30 methanol fluid for similar formations with severely
quality CO2 for a more cost-effective system. CO2 under-pressured wells. Hence a new polymer that is
energized oil gels prepared with frac oils has been used soluble in non-aqueous methanol and compatible with
in Canadian gas formations very successfully (Gupta carbon dioxide was identified. These non-aqueous base
and Leshchyshyn, 2005a). gels can be crosslinked with borate at pseudo-high pH
(non-aqueous fluids do not have pH) or with zirconium
7-4.5.5 Non-Aqueous Methanol Fluids crosslinker at pseudo-low pH for CO2 compatibility.
The special version of this HPG can hydrate in
In formations with severe liquid (aqueous and 100% methanol to give viscosity to the base gel and
hydrocarbon) trapping problems (see Section 7-6), is also compatible with CO2 without precipitation
non-aqueous methanol fluid may be a solution. Over (Gupta et al., 2003) (Fig. 7-27).
the years, several authors have identified the advantages
of alcohol-based fluids (McLeod and Coulter, 1966;
Smith, 1973; Tiner et al., 1974; Thompson et al., 1992;
Jales et al., 1988; and Hernandez, et al., 1994). These
advantages include, but are not limited to, low freezing
point, low surface tension, high water solubility, high
vapor pressure and formation compatibility. Methanol
is also the fluid of choice for formations with irreducible
water and/or hydrocarbon saturation (Bennion et al.,
1993a, 1993b, 1994, 1995, 1996a, 1996b). Three
concerns with methanol all relate to safety: low flash
point, high vapor density and flame invisibility. With
special precautions, as previous authors have identified
(Thompson et al., 1992; and Hernandez et al., 1994),
methanol can be safely used in the field.
Several approaches to increasing the viscosity
of methanol have been described in the literature Figure 7-27 CO2 compatibility of special highly
(Thompson et al., 1992; Jales et al., 1988; Boothe and derivatized HPG in non-aqueous methanol (left),
Martin, 1977; Crema and Alm, 1985; and Gupta et al., compared with regular highly derivatized HPG in non-
aqueous methanol (right), which precipitates when
1997). These range from foaming methanol to gelling exposed to CO2 (after Gupta et al., 2003)
with synthetic polymers (e.g., polyacrylamide and
polyethylene oxide) and modified guar. Attempts were These fluids can be completely broken with special
also made to crosslink the gelled methanol with metal breakers, resulting in very high regained permeability
crosslinkers. However, limitations as described by Ely in the proppant pack and in very sensitive formations.
(1994) restrict the use of gelled non-aqueous methanol, These fluids should be selectively used in gas
including solubility of these polymers in both aqueous formations with special safety considerations. These
and non-aqueous methanol, ability to crosslink, new generation fluids also do not require any water
breaking the polymer, and temperature limit. for hydration or for breaking.
252
7-4.5.6 Liquid CO2-Based Fluids Proppant transport can be divided into two areas:
(1) to the perforations, and (2) in the fracture. Moving
Fluids based on liquid CO2 are the real unconventional proppant through surface equipment and wellbore
fluids. The concept and applications of these fluids tubulars is accomplished relatively easily by the high
require outside-the-box thinking. These fluids have pump rates typically utilized in liquid CO2 or CO2/
been very successfully used in tight gas applications in N2 fracturing. The high turbulence caused by the high
Canada and several US formations. Their chemistry and Reynolds number (because of the low viscosity and
physics have been extensively published, as summarized high velocity of the fluid) is sufficient to adequately
in a paper by Gupta and Bobier (1998) and described transport the proppant to the perforations. The fluid
in brief here. The family of these fluids consists of pure density (~ 1.1) also helps transport proppants. Both
liquid CO2 and a binary fluid consisting of a mixture of the viscosity and density values drop when N2 is
liquid CO2 and N2 to reduce costs. added to liquid CO2 at a 50:50 ratio.
Conventional fracturing fluids rely on viscoelastic Proppant transport in the fracture for turbulent,
properties to inhibit leak off from the fracture into the low-viscosity fluids has been explored previously.
reservoir. Filter cake deposition from long-chain polymers Of particular significance is the effect of turbulence
or high filtrate viscosity in oil gel systems provides fluid on reducing sand settling rates. Frictional contact
loss control to establish adequate fracture width. Sufficient between proppant and fracture walls is increased,
fracture width is required to allow proppant placement in thereby slowing sand settling rates. Uneven, tortuous
the fracture. Liquid CO2, has very low viscosity (Gupta fracture walls combined with narrow fracture widths
and Bobier, 1998) and thus does not have the viscosity also serve to hinder proppant from settling within
or filter cake properties to establish fracture widths when the fracture. Settari et al. (1986) also described the
pumped at typical rates (i.e., 20 bpm). transport in the fracture as suspended transport
Low-viscosity, low-temperature fluids have or bed load transport depending on the ratio of
higher leak-off rates than conventional fluids, but this terminal velocity to horizontal velocity.
characteristic has an upper limit that is dependent Surface operational techniques for fracturing with
on reservoir parameters. The use of a gas or liquefied liquid CO2 and CO2/N2 can be divided into three areas:
gas makes the fluid compressible. The fluid pumped product delivery, slurry blending and high-pressure
at high pressures and low temperature volumetrically pumping. A liquid CO2/N2 fracture treatment is a
expands when the fluid is exposed to lower pressure liquid CO2 fracture treatment with up to 50% N2 (by
and higher temperature of the formation. These fluids volume) added at the wellhead. The N2 is pumped with
are not in steady state, and thus positive transient effects conventional N2 pumping equipment and does not
occur. This thermal expansion effect inhibits leak-off require any special considerations.
near the fracture face and promotes the development The liquid CO2 is stored in bulk storage units on
of fracture width. A combination of thermal expansion, location at 300 psi and 30 F. Each unit has a capacity
relative permeability effects and possible turbulence of 300 bbl of liquid. The liquid CO2 is moved from the
through small reservoir pore throats curtails leak-off of storage unit through the CO2 blender and to the high-
liquid CO2 from the fracture into the reservoir. pressure pumpers by gaseous nitrogen. A nitrogen tube
Because these liquid CO2-based fracturing fluids trailer is used to perform this function.
are compressible by nature, their hydrostatic pressure is The CO2 blender is a pressure vessel and manifold
dependent on both temperature and pressure. The density system that blends the proppant into a liquid CO2
of liquid CO2 remains relatively constant during high- stream prior to the high-pressure pumps (see Section 13-
pressure fracturing pressures (Gupta and Bobier, 1998). 1.4, Fig. 13-5). Proppant is loaded into the CO2 blender;
Adding gaseous nitrogen to make a CO2/N2 mixture the unit is sealed and then filled with CO2. During the
requires the use of computer algorithms to calculate treatment, proppant is drawn into the liquid CO2 stream
values. The density and compressibility of nitrogen can be by hydraulically driven augers. CO2 blenders have been
calculated if both temperature and pressure are known. built in 44,000- and 88,000-lb versions.
253
Modern Fracturing
Radioactive densitometers are vital in the of this fluid is the additional viscosity gained by the
successful operation of a liquid CO2 fracture foam over liquid CO2. The use of 75 volume percent
treatment. Two densitometers, one on the blender of N2 also makes the fluid very cost-effective and
discharge and one on the high-pressure line, are used applicable to project frac applications where multiple
to ensure the treatment design is followed. jobs can be performed in a single day. The fluid has
Liquid CO2 is pumped with conventional also found niche application in coalbed fracturing in
pumping equipment. Care must be taken to ensure Canada on dry coalbeds where any water introduced
that the liquid CO2 is kept above the equilibrium into the formation damages the cleats.
pressure. Specialized techniques for the high-
pressure pumpers reduce the chance of vapor locks 7-4.6 Acid Fracturing Fluid
or mechanical failure due to low temperatures. The
wellhead treating temperature of liquid CO2 can be as Conventionally, acid fracturing has been used for
low as 15 F. With these low wellhead temperatures, carbonate formations. Acid fracturing is a simulation
considerations for the thermal effect on well tubulars technique in which acid is injected at pressures above
must be addressed prior to the treatment. the parting pressure of the formation so that a hydraulic
Several papers have described the unique nature of fracture is created. Usually, a viscous pad fluid is injected
liquid CO2 and liquid CO2/N2 mixtures as fracturing ahead of the acid to initiate the fracture, and then plain
fluids (Lillies, 1982; Tudor et al., 1994; Mazza, 1997; acid, gelled acid, foamed acid, or an emulsion containing
and Gupta and Bobier, 1998). In these systems, the acid is injected. Fracture conductivity is created by the
proppant is placed in the formation without causing acid differentially etching the walls of the fracture; i.e.,
damage of any kind, and without adding any other carrier the acid reacts non-uniformly with the fracture walls so
fluid, viscosifier or other chemicals. As was described that after closure, the fracture props itself open, with the
previously, the use of a reservoir friendly substance like relatively undissolved regions acting as pillars that leave
liquid CO2 (and inert N2) offers unique advantages more dissolved regions as open channels. Thus acid
through the elimination of capillary fluid retention and fracturing is an alternative to the use of proppants to
clay swelling (Mazza, 1997). create fracture conductivity after closure.
These low-viscosity fluids are not an obvious choice Recently Kalfayan (2007) introduced the concept of
of fracturing fluid. However, a large number of jobs sandstone acid fracturing. Acid fracturing fluids have been
have been successfully performed with them (Gupta used a means of extending the ability of acid to penetrate
and Bobier, 1998). One of the major limitations of this deep into the formation, creating higher-conductivity
technology has been their high treatment cost. Although fractures. The viscosity increase can be obtained with
stimulation treatments using the low-viscosity liquid surfactants, resulting in viscoelastic systems or with
CO2 system have been successful, the high rates required polyacrylamide polymers as emulsions or microgels
to place these jobs and the associated frictional losses crosslinked with zirconium. These systems can be used
raised horsepower requirements. to 300 F and can break in a controlled fashion by using
encapsulated breakers (Boles et al., 1996).
7-4.5.7 Liquid CO2-Based Foam Fluid
7-5 Fracturing Fluid Additives
Several attempts have been made to increase the
viscosity of CO2-based fluids while trying to maintain 7-5.1 Additives for Water-Based Fluids
the conductivity and formation compatibility of these
fluids with very little operational success (Bullen et al., 7-5.1.1 Friction Reducers
1987). The liquid CO2-based foam fluid consists of
a foam of N2 gas in liquid CO2 as the external phase The rapid rekindling of interest in non-gelled water as
stabilized by a special foamer soluble in liquid or fracturing fluid or slickwater fracturing, particularly
supercritical CO2 (Gupta, 2003). The main advantage for shale natural gas reservoirs (see Section 11-
254
5), has necessitated the use of friction reducers. are dehusked, milled and screened to obtain the guar
Friction reducers not only reduce the friction gum. It is typically produced as a free-flowing, pale
and associated horsepower requirement for the off-white colored, coarse- to fine-ground powder
pumping operation, but also protect the equipment (Fig. 7-28). Manufacturers define different grades
from wear and tear due to the high rates of these and qualities of guar gum by the particle size, the
jobs. The most obvious choice of friction reducers viscosity that is generated with a given concentration,
would have been surfactants. However, they do not and the rate at which that viscosity develops. Coarse-
function as friction reducers in the highly turbulent mesh guar gums will typically but not always
regime of high-injection-rate jobs. develop viscosity more slowly. They may achieve
Partially hydrolyzed polyacrylamide has been a reasonably high viscosity but will take longer to
used as a friction reducer in the oil patch from the achieve it. On the other hand, they will disperse
1960s. Non-damaging viscoelastic surfactant based better than fine-mesh gums, all conditions being
systems have earned attention, for example by Teot equal. A finer mesh, like a 200-mesh, requires more
et al. (1981), who added an organic electrolyte to effort to dissolve (Hazen, 2004).
associate with the surfactant, and Schwartz and
Smith (2007), who added a low-molecular-weight
polyethylene oxide to a viscoelastic surfactant system
to lower friction in turbulent flow. However, the
lower cost of the polyacrylamide system has won
the battle. Oxidizing breakers can minimize the
perceived damage of these polymers, as shown by
Carman and Cawiezel (2007). The polyacrylamide
may be added as emulsions or microemulsions in a
hydrocarbon carrier or prehydrated and supplied in
brine solutions or as powders. Depending on the
temperature of the carrier water and salt content, they
Figure 7-28 Guar pods, seeds, splits and powder
can hydrate almost instantaneously, especially with
the proper amount of shear.
Guar gum (Fig. 7-29) is extracted from the guar
7-5.1.2 Gelling Agents bean. The guar bean is principally grown in India
and Pakistan, with smaller crops grown in the United
To viscosify water, the industry has typically used States, Australia and Africa. Possessing almost eight
natural polymers such as guar, starches and cellulose times the water thickening potency as cornstarch, guar
derivatives. These natural polymers hydrate or swell gum is economical because only a very small quantity
in water to provide the viscosity of the base gel. In is needed for producing sufficient viscosity. In water,
many cases, the base gel provides adequate viscosity underivatized guar is non-ionic and hydrocolloidal. It
for limited proppant transport and fluid loss control. is not affected by ionic strength or pH but will degrade
These polymers can be crosslinked to enhance the at pH extremes at elevated temperature. It remains
performance and thermal stability and reduce the stable in solution over pH from 5 to 7. Strong acids
cost by using lower loading of the polymers. cause hydrolysis and loss of viscosity, and alkalis in
Guar is readily available, cost-effective and can strong concentration also tend to reduce viscosity. It
be easily derivatized and so is the most common is insoluble in most hydrocarbon solvents. Guar gum
natural polymer used as a gelling agent for fracturing shows high low-shear viscosity, which is important for
applications. Guar gum, also called guaran, is proppant transport, but it is strongly shear-thinning.
primarily the ground endosperm of the seeds from It is very thixotropic above a concentration of 1%,
Cyamopsis tetragonolobus (L.) Taub. The guar seeds but below 0.3% the thixotropy is slight.
255
Modern Fracturing
C 1-6 D-galactosyl introduced to the oilfield; however, its use is minimal

OHOH in fracturing applications. The cleaner and higher-
H O yield non-derivatized guar covers the majority of use.
H For higher-temperature and low-pH applications,
HO H
H OH
OH
either carboxymethyl guar (CMG) or carboxymethyl
H O H
HO O hydroxypropyl guar (CMHPG) are used (Fig. 7-
H
HO O 30.) Typically, guar is used for borate and zirconate
O H
HO H crosslinked systems (detailed later in this section);
H
O H
HO H
H
H CMG and CMHPG are used both at high pH and low
H n pH with zirconium crosslinkers. HPG can be used to
backbone poly (C 1-4 D-mannosyl) crosslink with both borate and zirconate crosslinkers.
Ratio of Mannose to Galactose ~3:2 to 2:1 Prior to the use of CMG or CMHPG, cellulose
Figure 7-29 Chemical structure of guar (courtesy of derivatives, particularly hydroxypropyl cellulose (HPC)
Rhodia) and carboxymethylhydroxyethylcellulose (CMHEC),
were used, especially in low-pH systems.
Guar is a galactomannan consisting of a (1 4)-
OX
linked -D-mannopyranose backbone with branchpoints
XO
from their 6-positions linked to -D-galactose (i.e., 1
6-linked --D-galactopyranose). There are between 1.5 O
- 2 mannose residues for every galactose residue. Guar XO
gum is made up of non-ionic polydisperse rod-shaped OX

O
polymers consisting of molecules made up of over OX
10,000 repeat units. Higher galactose substitution also *

OX XO
increases the stiffness (i.e., decreases the flexibility) but O O
O O
reduces the overall extensibility and radius of gyration of XO *
XO
the isolated chains (Petkowitz et al., 1998). The galactose n
residues prevent strong chain interactions because few
For Guar, X is HX = H or -R - C O O H
unsubstituted clear areas have the minimum number For Hydroxypropyl guar, X is H or CH2CH(OH)CH3
For Carboxymethyl guar, X is H or CH2COOH
(about 6) required for the formation of junction zones. For Carboxymethyl hydroxypropyl guar, X is H or
Of the different possible galactose substitution patterns, CH2CH(OH)CH3 or CH2COOH
the extremes of block substitution and alternating

substitution give rise to the stiffer, with greater radius of Figure 7-30 Chemical structures of guar derivatives
gyration, and most flexible conformations, respectively
(random substitution being intermediate). Since the early 1980s a majority of treatments
Natural guar contains a significant quantity performed, particularly in North America, have
of insoluble material or residue. It used to be in the been on-the-fly or continuous mix (see Section 9-
range of 5 to 10%; however, newer high-yield guar 4.2). For proper metering, the polymer is typically
varieties presently available have residue of less than slurried in a hydrocarbon solvent such as diesel or
2%. This insoluble material does not function as more environmentally friendly mineral oil or even
a viscosifier and can be damaging to the proppant glycol ethers for offshore use where the product
pack or form filter cake on the fracture face. Several does not create sheen on the surface of water. These
chemical modifications to guar have been performed to systems have residence volume to provide adequate
reduce the insoluble material and increase the stability time for nearly complete hydration, which can be
of the polymer at high temperatures and low pH to confirmed by realtime measurement of base gel
make it more CO2-compatible for energized systems. viscosity. In the near future, on-the-fly systems using
Hydroxypropyl guar (HPG) was the first derivative dry powders will become a reality.
256
Another development in the last decade has been Adding more polymer molecules increases the
the move toward low- and no-polymer systems. The low- interaction and minimizes the volume occupied by each
polymer systems were introduced in systems developed molecule; eventually the molecules run out of space, and
by Gupta and Franklin (1994). Dawson et.al., (2000) the only way they can occupy space is to interpenetrate.
further developed a theoretical basis for these systems. This point is marked by another slope change and
Lei and Clark (2004) discussed the concept of C* and labeled as C**. Crosslinking cannot occur below C*
C** and the minimum effective concentration for because there is not enough polymer-to-polymer
crosslinking to occur and the means for measuring it. interaction to promote intermolecular crosslinking.
When a water-soluble polymer is hydrated in water, the There can be intramolecular crosslinking, but this results
viscosity of the solution increases with increasing polymer in reducing viscosity, not increasing it. The effectiveness
concentration. This relationship is exponential for most of crosslinking increases between C* and C**. Between
polymers in their solvents. Earlier work showed the C* and C** is a critical crosslinking concentration (Ccc),
relationship of viscosity as a function of concentration can which is the minimum concentration of polymer needed
be divided into two regions separated by a concentration for a full three-dimensional structure to develop. The
of the polymer called, C* (Economides and Nolte, 1989; critical crosslinking concentration represents a lower limit
and Gidley et al., 1989). When the polymer is added to to the useful crosslinking concentration for fracturing
water, the molecule first swells as the solvent penetrates applications. Table 7-3 details the critical overlap (C*)
the polymer molecule, and the polymer molecule will and critical crosslinking concentrations for various
occupy some random conformation in space. For a polymers. GW-3 listed in their paper is a high-molecular-
random coil polymer this will be spherical. If a second weight, high-yield, and cleaner guar than conventional
polymer molecule is added, the two molecules will guar used in the industry. (Note: Concentrations have
not interact, and their effect on viscosity will be only been converted into oilfield units.)
hydrodynamic. As more and more molecules are added,
there will be a point when weak interaction starts to Table 7-3 Critical Overlap and Critical Crosslinking
occur, and slope of the viscosity/concentration curve Concentrations for Various Gelling Agents (after Lei and
changes. This can be seen when the log of the specific Clark, 2004)
viscosity is plotted against the log of concentration (Fig. Polymer C*,wt% C*, pptg
Ccc,
Ccc, pptg Crosslinker
wt%
7-31). Specific viscosity is given by:
GW-3 0.051 4.26 0.178 15 Borate
s
= 0 , CMG 0.054 4.51 0.193 16.1 Zr-chelate
sp s (7-21)
CMHPG 0.068 5.68 0.22 18.4 Zr-chelate
where o is the zero-shear viscosity and s is the viscosity Guar 0.078 6.51 0.26 21.7 Borate
of the solvent (water).
HPG 0.093 7.76 0.285 23.8 Borate
C**
The other major development involves the use of
Concentrated
surfactant or viscoelastic gels that contain no polymer at
log Nsp
C* all. This has been discussed in Section 7-4.5.1.
Semi-Dilute
7-5.1.3 Biocide
Dilute
The natural polymers used as gelling agents are a

log C good food source for bacteria. The bacteria degrade
Figure 7-31 Effect of polymer concentration on specific
the polymers, and the viscosity of the gel can be lost
viscosity in a few hours. In addition, certain bacteria can turn
257
Modern Fracturing
reservoir fluids sour. Biocide or bactericide is added

to the water in the tanks to prevent growth of the O
bacteria. Some biocides also inactivate the enzymes
produced by the bacteria. Biocides are regulated by N C
the Environmental Protection Agency (EPA) and has
to be registered with state agencies also for use in NH2
the particular state. The most common biocides are
either oxidizing type or cationic in nature.
Glutaraldehyde (Fig. 7-32) is the most common Br Br
and cost-effective biocide available for fracturing
fluids. It is a broad-spectrum biocide that functions Figure 7-33 Chemical structure of DBNPA (2,2-dibromo-
3-nitrilopropionamide)
in the pH range of 7 to 11. Biocides are typically used
to sterilize treatment water before adding the polymer
(i.e., to prevent the creation of a colony of bacteria)
Br
rather than to remove a colony that has already NO2
formedby that time, in most cases, it is almost H H
certainly too late to save the tank of gel.
O O HO HO
Figure 7-34 Chemical structure of BNPD
(bromonitropropane-1, 3-diol)
H H
THPS (tetrakishydroxymethyl phosphonium sulfate)
Figure 7-32 Chemical structure of glutaraldehyde (Fig. 7-35) is a broad-spectrum biocide with low toxicity.
It is compatible with most non-oxidizing biocides and
Glutaraldehyde is highly effective against sulfate surfactants. It is not compatible with oxidizing biocides
reducing bacteria (SRB). It has excellent bio-film and has limited stability at high pH.
penetrating properties and is generally compatible with
most fracturing fluid additives. Sometimes, it has an +
effect on the function of enzyme breakers. It will react CH2O
with amines and has an obnoxious odor. H +
P H
DBNPA (2,2 dibromo-3-nitrilopropionamide) (Fig.
HOH2C CH2O H SO42-
7-33) is the fastest-acting biocide used with fracturing H CH2OH
fluids. Because it is available as a powder, it is typically H H 2
available in water-soluble bags that can be dropped
into water tanks. It is a broad-spectrum biocide and Figure 7-35 Chemical structure of THPS
(tetrakishydroxymethyl phosphonium sulfate)
functions at very low concentration. DBNPA has low
environmental impact because it hydrolyzes very quickly
to leave no footprint. It has a very short half-life at high Isothiazolone (a mixture of 5-chloro-2-methyl-4-
pH, which may be a drawback. isothiazolin-3-one and 2-methyl-4-isothiazolin-3-one
Bronopol or BNPD (bromonitropropane-1, 3- in approx. 2.7:1 ratio) (Fig. 7-36) is a good broad-
diol) (Fig. 7-34) is typically used in conjunction with spectrum biocide used at very low levels (0.3 to 2 PPM).
isothiazolone (see below). It has low odor, is effective in It is a good bio-stat and a slow-acting biocide that can
a range of pH values. However, it releases formaldehyde be combined with quick-kill biocide for combined
on degradation. quick-kill and sustained activity.
258
are referred to as crosslinking agents or crosslinkers.

O O
Each crosslinking agent has its own strengths and
weaknesses. Each functions at a particular pH range
and temperature and will crosslink a particular type of
N H N
Cl polymer. As discussed earlier, guar and hydroxypropyl
CH3 CH3
S S
guar have hydroxy groups that can be crosslinked,
Figure 7-36 Chemical structure of isothiazolone carboxymethylhydroxypropyl guar has both hydroxy
and carboxy groups that can be crosslinked, and
It is always a good idea to rotate biocides so carboxymethyl guar has carboxy groups that can be
that the bacteria do not become resistant to one crosslinked (see discussion in Section 7-5.1.2).
type of biocide. The most common crosslinking performed on
these polymers are borate crosslinking, which occurs
7-5.1.4 Buffers at high pH, and metal or covalent crosslinking at
low or high pH. Borate crosslinking is by hydrogen
Buffers are used for various purposes in fracturing fluids. bonding and is shear-tolerant. The borate crosslinked
They are used for proper dispersion of the polymer fluid reheals on application of shear, and properly
particles in water and also in hydration of the polymers. designed borate fluid exhibits nearly perfect transport
The crosslinking of the polymer and its stability are properties. Because the crosslinking is very pH-
dependent on the pH. The pH is adjusted with buffers. dependent, when used in wells with nearly neutral-pH
In the industry, pH adjusters are sometimes mistakenly water, borate crosslinked fluid cleans up nicely giving
called buffers. pH adjusters shift pH of the solution but over 90% regained permeability in the proppant pack
do not possess buffering capacity. An acid or alkali can when used with enzyme breakers.
shift pH, but if the water contains natural buffers, it The most common borate crosslinked fluids use
may not maintain pH at a constant value. In addition, a form of mono-borate ion for the crosslinking. The
the pH drops when the solution is heated if it does mono-borate ion (BO3-) can be from borax, boric acid
not have sufficient buffer capacity. Buffer solutions are or similar species. Typically, these species crosslink at
solutions that resist change in hydrogen ion and the high pH (greater than 9.5) almost instantaneously.
hydroxide ion concentration (and consequently pH), They can be delayed by using a slow-dissolving
upon addition of small amounts of acid or base, or buffer (Cawiezel, 1990) or slowly soluble borate
upon dilution. Buffer solutions consist of a weak acid ores (Mondshine, 1986).
and its conjugate base (more common) (e.g., acetic Because of the potentially very high viscosity that
acid/sodium hydroxide, resulting in acetic acid/sodium can build with these borates, to minimize friction,
acetate) or a weak base and its conjugate acid (less it is desired to have delayed systems. Varying a
common) (e.g., ammonia/hydrochloric acid, resulting combination of pH and amount of borate ion can
in ammonia and ammonium chloride). Based on optimize the fluid. One of the problems of the mono-
the various combinations possible, different service borate systems, particularly at higher temperatures, is
companies use different combinations of acids, bases optimizing the system. Too much borate ion makes
and salts to arrive at buffers at various pH conditions the fluid experience synerisis, sometimes be referred
with various buffer capacities. to as over-crosslinking. The crosslinked fluid expels
excess water, becomes non-homogeneous and does
7-5.1.5 Crosslinkers not support proppant transport. Too little borate ion
make the gel weak and unstable at high temperatures.
Crosslinking is the most cost-effective means of Also, stabilizing the fluid at high temperatures with
increasing viscosity of fracturing fluids without mono-borate typically requires higher pH, which
increasing the amount of gelling agent used. The tends to delay the crosslink, or higher buffer capacity,
chemical compounds that perform this function which affects the performance of enzymes as breakers.
259
Modern Fracturing
The use of organo-borates such as borate esters (Gupta comprehensive examination of the proposed structures
et al., 2006) or polyborates (Dawson, 1992) make of titanium and zirconium alkoxides and aqueous-based
the borate crosslinked systems stable and usable to chelates, including a general discussion of the methods
temperatures in excess of 300 F. employed to manufacture these crosslinkers and factors
During the early days of fracturing, metallic present in Group 4 metal crosslinkers that may affect
crosslinkers meant titanate, aluminum, antimony their crosslinking response with polygalactomannans
or chromium compounds. All had limited high- (guar and their derivatives) (see also Section 7-5.1.2).
temperature stability (>250 F) and shear stability. There is also some discussion of other methods that
Some of the antimony and chromium compounds have been employed to control gelation rate. Finally,
were not considered environmentally friendly. there is some discussion about the effects of polymer
Titanates were typically used from pH 3.5 to 9.5. attributes on fracturing fluid performance, particularly
These metal crosslinked systems did not clean up well as it relates to crosslinker efficacy.
in laboratory tests and hence lost their usefulness.
Using low pH to clean up these fluids seemed like
a good idea. The use of aluminum and titanium for
this purpose looked good. However, the aluminum
systems were very shear-degrading, did not prove to
have good clean-up characteristics and had limited
thermal stability. The use of zirconium crosslinkers
for both high pH and low pH became prevalent and
is the metallic crosslinker used widely today.
Harry et al. (1997) and Moorhouse et al. (1996)
demonstrated that seemingly minor variations in
aqueous-based crosslinker chemistry can produce
significant differences in the rheological responses of
polygalactomannan-based fracturing fluid crosslinked
with these agents. While the work contained some
references to borate crosslinking and the principle
conclusion can be applied to both borate and
zirconium fracturing fluid systems (Fig. 7-37), the
emphasis was on zirconium crosslinked systems.
Not surprisingly, some of the observed differences in
performance could be attributed to variations in the
crosslinker chemistry. Crosslinker differences that
were isolated and evaluated included differences in the
ligand chemistry, the ligand-to-metal ratio, the nature
of the zirconium starting material, etc. (A ligand is
an ion, molecule or a functional group that generally Figure 7-37 Appearance of the base gel, borate
donates or shares one or more of its electrons through crosslinked gel and zirconium crosslinked gel
a coordinate covalent bond with a metallic ion).
Moorhouse et al. (1996) suggested several mechanisms 7-5.1.6 Breakers
by which these phenomena might be explained. The
authors stopped short, however, of making any attempt Breakers are chemicals that literally break the viscosity
to correlate the chemical structure of the crosslinker of the fracturing fluid, reduce the molecular weight of
with its rheological performance. Later, Harry et the polymer and help in cleaning the proppant pack and
al. (1999) advanced their studies to present a more the filter cake on the fracture face. The most common
260
types of breakers used with fracturing fluids are oxidizers, particularly polymer-specific enzymes, have been very
enzymes and acids. These products can be used as solids, successful in providing good clean-up of the proppant
slurries of solids in liquids, or liquid solutions. The most pack and filter cake (Brannon and Tjon-Joe-Pin, 1994).
common oxidizers are persulfates such as ammonium, Enzymes are catalysts and so can continue the degradation
sodium or potassium persulfates, peroxides such as process for a long time. They are not consumed in the
calcium, magnesium or organic peroxides or perborates. process like oxidizers are, and since they are catalysts,
All these oxidizers typically produce peroxygen, which very little goes a long way. Because they are very specific,
is a very reactive free radical species that attacks the they do not interact with other additives in the fracturing
polymer and degrades it. The reaction is stoichiometric; fluid or the formation and are very compatible with
i.e., one peroxygen radical reacts with one bond on the resin-coated proppants. With the use of biotechnology
polymer. In fact, the oxidizers are non-specific. They can and ever increasing interest in breakers, the application
attack not only polymer, but any species that is prone range of enzymes with respect to pH, temperature and
to oxidation, including the formation, tubulars and in other conditions is widening constantly.
particular resin-coated proppants. Figure 7-38 shows the difference in residue of
At low temperatures (<120 F), the oxidizer guar gel broken by an enzyme and an oxidizer. The gel
reaction rate is generally too slow, and catalysts or broken with oxidizer shows more residues than the one
accelerators such as amines or transition metal salts are broken by enzyme. Enzyme breakers may cost more on
used to increase the rate of reaction. Until the 1980s, a weight basis; however, as they are catalysts, they have
it was believed that no breaker was needed for high become cost-effective in use concentrations.
temperatures; however, diligent work by several authors
have shown that the clean-up or conductivity of the
proppant pack is better with breakers. Because the
half-life of the oxidizers is short and the reaction rate is
too fast at high temperatures, the use of encapsulation
technology became common in the late 1980s. Products
based on several mechanisms were developed; the chief
among them used osmotic burst effect (Walles et al.,
1988) and diffusion (Gupta and Cooney, 1992). The
use of slowly dissolving peroxides was an alternate
(Mondshine, 1993). The encapsulated or controlled-
release products not only increased the application
temperature of oxidizers, but could also be incorporated
into the filter cake for degradation.
The use of enzymes at low temperatures and nearly
neutral or low-pH conditions has been prevalent since
the 1960s. Gupta et al., (1992) disclosed the first high-
pH enzyme for guar borate system. The advent of
biotechnology resulted in the development of polymer-
specific enzymes, specifically for guar or guar derivatives,
which extended the application temperatures to near
350 F. Combination systems of enzymes with oxidizers
also have been common. The oxidizers not only degrade
the polymer but also reduce the pH of the fluid so the
enzymes become very active. Organic esters, which also
reduced the pH of the fluid, became very common for Figure 7-38 Differences in guar gel broken by an enzyme
use with enzymes (Dawson and Le, 2004). The enzymes, and an oxidizer
261
Modern Fracturing
Thermal effects and dissolved oxygen and other large quantities of fluid are trapped in the reservoir in
species, which affect the stability of the fluid, also degrade the near-wellbore area, in the area surrounding the
polymers. The effect of these natural breakers can be fracture and in the fracture itself. This trapped fluid has
minimized by the use of gel stabilizers such as oxygen detrimental effect on relative permeability, effective flow
scavengers (e.g., sodium thiosulfate or methanol) or free area and effective fracture lengths, and it impairs well
radical scavengers (Pakulski and Gupta, 1994). productivity (see Section 7-6 for details).
Surfactants can also be used to change the contact
7-5.1.7 Clay Stabilizers angle of the leak-off fluid into the pore space, which
changes the wettability of the fluid to the formation
As mentioned previously, the use of water-based fluids rock, again for easier recovery, especially in tight gas
became very popular with the addition of salts such as reservoirs. (Penny et al., 2005). Classes of surfactants,
potassium chloride or ammonium chloride to the fluids such as foamers, are used to generate and stabilize foams
to control clay swelling and make the water-based fluids for foam fracturing fluids. Certain biocides and clay
more compatible with clay-containing formations. control agents can also be surfactants. Surfactants can
However, several guar derivatives, particularly also be used to prevent or break emulsions formed with
CMHGP and CMG, lose some of their effectiveness any liquid hydrocarbons present in the formation.
(particularly the ability to prepare fracturing fluids
with very low loading and crosslinked with zirconium 7-6 Fluid Damage to Fractures and
crosslinkers) with these salts. Tetramethyl ammonium Sources of Productivity Impairment
chloride (TMAC) (Himes and Vinson, 1989) and
choline chloride (Hall and Szemenyei, 1992) have
been shown to function as temporary clay protection Cinco-Ley et al., (1978) and Cinco-Ley and
agents similar to potassium chloride. However, higher- Samaniego (1981) provided the understanding
molecular-weight amines and cationic polymers are of the factors affecting the performance of finite-
used for permanent clay stabilization. These materials conductivity fractures and identified the types of
are not washed in the presence of fresh water and damage impeding their performance:
also prevent clay migration and related permeability
damage in the near-wellbore area (McLaughlin et al., 1. Reduction of proppant pack permeability. This
1976). The permanent clay stabilizers are typically used kind of damage affects the proppant-pack inside
with temporary clay protecting agents like potassium the fracture and is a manifestation of proppant
chloride or KCl replacements. crushing and, especially, unbroken fracturing
fluid polymer. These phenomena have particularly
7-5.1.8 Surfactants detrimental impacts on the fracture conductivity
and should be avoided or minimized.
Surfactants provide several functions when used in The problem with proppant crushing can be
fracturing fluids. They are used to reduce surface tension, reduced considerably by selecting appropriate-
which makes fluid recovery easier after the fracturing strength proppants. This is addressed in Chapter 8.
operation. They can also reduce the interfacial tension To minimize polymer-related problems, extensive
between water and formation fluids, thus preventing the research has been conducted in the last several years
formation of emulsions, etc., which damage permeability. on breaker technology and the use and delivery of
In gas reservoirs, the use of water-based fluid can create appropriate chemical agents.
fluid retention problems (see also Section 7-6). Surfactants
are typically used to reduce the surface tension between 2. Choke damage. This refers to the near-wellbore
the treating fluid and the gas, emabling recovery of more damage inside the fracture (Fig. 7-39). It
fracturing fluid and restoring of relative permeability to can be depicted by a skin effect. This kind of
gas (see Section 7-6). When fluid recovery is not good, damage results either from fines migration
262
during production and their accumulation 4. Combined effects. Mathur et al., (1995) proposed
near the wellbore (within the fracture) or over- a means to account for composite damage that can
displacement at the end of the treatment (a be quantified by a skin effect expression:
disastrous error, if it happens) or inadequate
b2 k R (b1 b2 ) k R b1
perforations connected with the fracture. sd = + ,
2 b1k3 + ( x f b1 ) k2 b1k1 + ( x f b1 ) k R x f
It is possible to calculate the skin from
the choke-damaged fracture by assuming (7-24)
steady-state flow in the damaged zone (Cinco- where b1 is the radial damage in ft (equivalent
Ley and Samaniego, 1981): to rs); b2 is the width of the fracture face damage
in ft; k1, k2 and k3 are the permeabilities in the
xs k
s fc =
, (7-22) radial damage zone, the fracture face outside of
w fs k fs
the damage zone and the fracture face inside
where xs, wfs, kfs are the damaged fracture the radial damage zone, respectively; and
length, width and permeability, respectively. k is the reservoir permeability.
Figure 7-41 is a schematic diagram of the
Damage Zone Wellbore Fracture
composite damages accounted for in Eq. 7-24.
k Mathur et al., (1995) showed that positive skins
w kf kfs are possible from fractured wells if the penetration
and severity of fracture face damage are large. Such
xs xf
positive skin effect has caused some confusion
in view of the idea that fracturing is supposed
Figure 7-39 Vertical fracture with choke damage to result in a negative skin effect, bypassing
formation damage. The two concepts are not
3. Fracture Face Damage. This kind of damage, inconsistent. What the Mathur et al. (1995) work
caused conventionally by fracturing fluid suggests is that although fracturing eliminates pre-
leak-off, results in permeability impairment treatment radial skin effects, new skins resulting
outside the fracture, normal to the fracture from the fracture treatment itself can emerge.
face. Again Cinco-Ley and Samaniego (1981) They suggest better designs, better selection of
provided a means to account for this damage fracturing fluids and better execution.
with a skin effect, defined by: Radial Damage
Fracture Face Damage
Fracture
b k
s fs = s 1 , (7-23)
2 x f k s
where bs is the penetration of damage normal to the

fracture face and ks is the damaged permeability
inside this zone (Fig. 7-40).
rs
xf
Damage Zone
Wellbore
Figure 7-41 Fracture with composite damage
ks bs
kf wf All of these skins can be accounted for in

fractured well performance by calculating a new
dimensionless PI for the damaged fracture:
xf 1
Fracture J = .
D ,damaged 1 (7-25)
+s
Figure 7-40 Vertical fracture with fracture face damage J D
263
Modern Fracturing
An example of how fracturing fluid damage From Eq. 7-24, three things are important: the
affects production in quite different ways penetration of damage, the severity of damage and the
in low- and high-permeability reservoirs is fracture length. Aggour and Economides (1999) have
provided in Section 7-6.1. Several mechanisms studied experimentally and mathematically both the
of damage can be caused or accelerated by penetration of fracturing fluid leak-off perpendicular
hydraulic fracturing, and these are described, to the fracture face, and the severity of damage. In low-
starting with Section 7-6.2 permeability reservoirs up to 5 md, even linear gels will
not penetrate, and the leak-off zone would be of the
7-6.1 Example Calculation of Productivity order of fraction of an inch. Thus, from Eq. 7-24 and
Impairment from Fracture Damage using bs = 0.5 in., and even with k/ks = 10 for xf = 500
ft, then sfs = 0.001. From Eq. 7-25, with undamaged
Two wells are considered: one in a reservoir with JD = 0.49 the damaged JD = 0.489, an insignificant
permeability of 1 md and the second in a reservoir reduction. Fracture face damage in low-permeability
with 50 md permeability. All other variables are held reservoirs has little impact and should not be the over-
constant: net thickness of 50 ft and a fracture length riding concern in fluid design.
(tip-to-tip) to height aspect ratio equal to four. In both What would affect fracture performance is clearly
wells, 100,000 lbs of ceramic proppant is injected with the length. If the length were to be reduced from 500 ft
in situ proppant pack permeability (with zero polymer to, e.g., 150 ft without any damage, the JD would drop
induced-damage) of 220,000 md. to 0.38, a 22% reduction.
With the Unified Fracture Design (UFD) (see For high-permeability reservoirs the story is entirely
Section 4-5) for the 1-md reservoir, an undamaged different. The penetration of fracture-face damage can
fracture would lead to an optimum fracture half-length be as much as 1 ft (Aggour and Economides, 1999), and
of about 500 ft and a JD of 0.49. For the 50 md case, linear polymers can penetrate the reservoir itself. Aggour
the fracture half length would be about 140 ft with a JD and Economides found that cross-linked polymers
of 0.3. Assuming a drawdown of 1000 psi and B = 1 will not penetrate to 600 md permeability reservoirs.
(resbbl/stb)cp, the corresponding initial pseudosteady Even with mostly filtrate invasion, (bs = 1 ft, k/ks = 100
fractured well production rates would be about 175 and xf = 140 ft) the sfs = 1, resulting in JD,damage = 0.23,
and 5300 STB/d, respectively. a very serious 23% reduction from the undamaged
First, proppant pack damage is considered. Suppose case. This damage would be much more detrimental
that an unsuccessfully broken gel results in a proppant with smaller fracture length as is the practice in many
pack permeability one-tenth of the expected, i.e., high-permeability applications. A fracture length of 50
22,000 md. The JD for the 1-md reservoir (assuming ft would result in tripling the fracture face skin affect
the same length is executed) would drop to 0.27, a 45% and, with smaller undamaged JD in the first place, the
reduction. For the 50 md case, the JD would be 0.19, a productivity impairment could exceed 50%. Controlling
37% decline. (Note: If the proppant pack permeability leak-off and fracture face damage is thus very important
reduction were anticipated and the fractures designed in high-permeability reservoirs.
with that in mind, this would lead to a fracture half- Hunt et al. (1994) have suggested that the initial
length of 300 ft and a JD of 0.39 for the 1 md case productivity impairment because of fracture face
and a radial fracture of 50 ft radius and JD of 0.23 in damage may diminish over time for a suitably designed
the 50-md case. The production rates would drop to treatment. Aggour and Economides (1999) concluded
140 STB/d (20% decrease) and 4050 STB/d (23% that damage penetration normal to the fracture face
decrease)). With correct design, the decline percentage is more important than the degree of damage. If the
of the high-permeability reservoir is only slightly larger, penetration is minimized, even a 99% damage to
but the actual rate reduction is considerable, showing reservoir permeability should still result in a negative
the importance of controlling proppant pack damage skin. This is an important conclusion, suggesting that
in high-permeability cases. even though the performance of high-permeability
264
fractures is expected to improve with time as cleanup Capillary forces in the formation are the reason
of the fracture occurs, the appropriate choice and for fluid retention. Capillary pressure forces are
engineering of fracturing fluid can significantly reduce the difference in pressure between the wetting
or even eliminate the time of initial lost production. (typically water in gas reservoirs) and non-wetting
Aggour and Economides (1999) proposed the use of (gas) phases in the matrix. The imbibition effect
high polymer loads of cross-linked fracturing fluids has been observed as a particularly severe problem
to minimize damage penetration. To complete the in reservoirs where sub-irreducible water saturation
fracturing fluid design, there is a crucial need for exists. Sub-irreducible water saturation is believed
good breakers and filter cake building additives. This to have been created by a combination of factors,
type of fracturing fluid can minimize the spurt loss including dehydration, desiccation, compaction,
and leak-off. Thus, in high-permeability fracturing, mixed-wettability, significant height above the free
although creating a highly conductive fracture water level in oil reservoirs due to drainage, and
should be the primary concern, the treatment must diagenetic effects occurring during geologic time.
be engineered to prevent fracture face damage, a Laboratory capillary pressure measurements supply
potential highly detrimental event. good approximations of the irreducible water
saturation that would normally be expected, but
7-6.2 Formation Damage from Saturation Changes actual reservoir water saturation can be substantially
lower, i.e., a sub-irreducible level. The high capillarity
Adverse saturation in the formation can also contribute pressure associated with low-permeability microporous
to productivity impairment. Production has been reservoirs is illustrated in Fig. 7-42.
successfully achieved in formations with matrix
600
permeability as low as 10-3 md. However, adverse
0.001 md.
capillary forces, which result in high in situ saturation
of trapped water or liquid hydrocarbons even in 0.01 md.
500
very low-permeability formations, make economic 0.1 md
production difficult. Low-permeability formations 1.0 md.
are typically tolerant of only minimal saturation 400
damage due to the sensitivity to capillary retention
Capillary Pressure, psi
effects, rock-to-fluid and fluid-to-fluid compatibility

issues. In these wells, the damage from drilling and 300
completion can be overcome by a properly designed

frac treatment, which can penetrate beyond the zone
200
of induced invasion and damage.
7-6.2.1 Fluid Retention

100
The major cause of productivity impairment in gas

wells during drilling, completion and fracturing or 0
workover operations is fluid retention effects. These 0.2 0.4 0.6 0.8 1
can consist of the permanent retention of water Water Saturation, fraction
or hydrocarbon based fluids or the trapping of

Figure 7-42 Effect of capillarity on water saturation (after
hydrocarbon condensate fluids retrograded in the Holditch, 1978)
formation during the production of the gas. Bennion
and his collaborators have labeled this phenomena Measured capillary pressure values for four rocks
as aqueous or hydrocarbon phase trapping with permeability from 0.001 to 1.0 md are presented
(Bennion et al., 1994, 1995). to illustrate the greater imbibition effects of water in
265
Modern Fracturing
lower-permeability formations. The capillary pressure where APTi is the Aqueous Phase Trapping Index,
of the 0.001 md core at 40% water saturation is ka is the uncorrected formation air permeability
325 psi greater than that of the 0.01 md core at (md), and Swi is the initial (not irreducible)
initial saturation. This illustrates the higher capillary water saturation (fraction).
pressure available in tighter reservoirs to imbibe and APTi values may range from 0 to 1, but more
trap aqueous liquids due to capillary imbibition. typically vary from 0.3 to 1. Table 7-4 contains
Injecting water-based fracturing fluids into high- guidelines for interpreting APTi values.
capillarity reservoir results in the creation of a zone
of high water saturation in the near-wellbore or near- Table 7-4 APTi Values and Associated Reservoir
fracture face area. The relative permeability curves in Characteristics (after Bennion, 1996).
Fig. 7-43 show how increasing water saturation above APTi Range Reservoir Characteristics
the irreducible water saturation results in a dramatic formation unlikely to exhibit significant
decrease in gas relative permeability. APTi = 1.00 permanent sensitivity to aqueous phase
trapping
100
krw 0.80 < APTi < formation may exhibit sensitivity to aqueous
90 krg 1.00 phase trapping
80 formation will likely exhibit significant

APTi < 0.80
sensitivity to aqueous phase trapping
Relative Permeability, percent
70
Relative permeability
decreases rapidly as Water
60 Saturation increases.
A second diagnostic technique uses the
50 calculation of percent Bulk Volume Water (%BVW).
40 Bulk volume water is the percent of the total volume
(including rock) which is water. By comparing the
30
bulk volume water in a given formation versus
20 water production from various fields, a prediction of
10 water cut can be made in a given field (Davis and
Water
0
Imbibition Wood, 2004). Based on the formations log-derived
0 20 40 60 80 100 average porosity and initial water saturation values,
Water Saturation, Percent Pore Space
BVW can be calculated:
Figure 7-43 Effect of water imbibition on relative
% BWV =100 S w ,
(7-27)
permeability changes (after Keelan, 1975).
Producing gas results in the affected zone where %BVW is the Percent Bulk Volume Water,
reverting to the irreducible water saturation dictated is the porosity of the reservoir, fraction, and Sw
by the capillary effects of the system and not the sub- is the water saturation, fraction. Table 7-5 contains
irreducible saturation existing before. The net effect guidelines for interpretation of % BVW values.
is that the critical producing area of the well retains
the increased water saturation, a lowered relative Table 7-5 Percent BVW and its Effect on Reservoir
permeability to gas and therefore lower productivity. Characteristics (after Davis and Wood, 2004)
Bennion et al. (1996a and 1996b) developed an % BVW Range Reservoir Characteristics
empirical equation as a diagnostic tool to help evaluate formation unlikely to exhibit significant
% BVW 3.5 permanent sensitivity to aqueous phase
a reservoirs sensitivity to aqueous phase trapping. trapping
The equation is based on the average permeability 3.5 % BVW formation may exhibit sensitivity to aqueous
and initial water saturation values: 2.0 phase trapping
formation will likely exhibit significant
% BVW < 2.0
sensitivity to aqueous phase trapping
APTi = 0.25 log ka + 2.2 S wi , (7-26)
266
These are just guidelines, and exceptions abound, 7-6.2.3 Fluid/Fluid Interactions
particularly for over-pressured reservoirs where
the capillary imbibition effects can be overcome Fluid-to-fluid incompatibilities can result in formation
in a relatively short time frame (Bennion et al., of insoluble scales or precipitates due to adverse
1996a and 1996b). These correlations can be used chemical reaction of the invaded fracturing fluid
to estimate compatibility of the formation to filtrate with in situ formation water. The potential for
water-based fracturing fluids. stable emulsion formation also can exist, even with
water-based fluids in gas formations that may contain
7-6.2.2 Rock/Fluid Interactions an initial irreducible liquid hydrocarbon saturation.
The most common type of problematic emulsion is
1. Clays: The low permeability associated with many the water-internal emulsion in which small droplets
gas formations is caused generally by small grain of water are encapsulated in a continuous external oil
size or limited intercystalline porosity. Permeability phase. These types of emulsions can have very high
is also reduced by significant concentrations of clay. viscosity (up to 2 to 4 orders of magnitude above
A variety of different types of clays can be present. clean, non-emulsified oil) and hence may result in
Highly fresh water sensitive expandable clays such as the formation of permeability-inhibiting emulsion
smectite or mixed-layer clays can occur in shallower block. Poorly designed spent acids are a common
formations. When contacted by fresh or low-salinity offender in this area. Foamy oil would also fall into
water, these clays expand in size due to substitution of the category of a stabilized emulsion, where the oil
water into the clay lattice (Civan and Knapp, 1987). forms the eternal phase and small bubbles of trapped
Other types of clay, such as kaolinite, are susceptible gas form the internal phase. Usually associated with
to electrostatic deflocculation, where abrupt changes high-viscosity heavy oils, these fluids have been
in salinity and pH can cause the clay to disperse and documented to have viscosity substantially higher than
migrate to pore throats, where it may bridge, block non-foamy fluids (Bennion, et al., 1993b).
and cause reductions in permeability. The fluid lost to
the formation can extend a significant distance from 7-6.2.4 Wettability Alterations
the fracture face; this type of damage may be sufficient
to partially impair production. Many additives to fracturing fluidsparticularly
surfactants, defoamers, corrosion inhibitors and
2. Chemical Adsorption: Physical adsorption of high- some biocideshave polar adsorptive tendencies
molecular-weight polymers or oil-wetting surfactants that may cause them to establish an oil-wetting
can reduce permeability significantly in low-quality condition in the region of the reservoir they invade.
formations or preferentially elevate permeability Figure 7-44 illustrates the phenomenon of a near-
to water if free-water saturation is present in the wellbore wettability alteration. Water-wet rock,
matrix. In low-permeability reservoirs, the relative due to surface frictional drag effects associated
size of the polymer chain, even from a broken fluid, with the motion of the water phase, tends to have
when adsorbed on the surface of the porous media, fairly low final relative permeability.
is significant in comparison to the pore throats that Conversely if a rock is oil-wet, the water can
allow gas to flow from the matrix. move easily through the central portion of the pore
system, and effective final relative permeability
3. Fines Migration: Fines tend to move preferentially and water mobility are often much higher. If a
in the wetting phase; hence when only gas is formation is initially water-wet, transition to an oil-
flowing, the migration of particulate in the porous wet condition is akin to placing a semi-permeable
matrix is minimized. Problems can occur when the membrane around the wellbore, which tends to
fluid invasion occurs due to relatively high spurt hold oil back and preferentially let water through.
loss from the fracturing fluids. This may result in an undesirable increase in water-
267
Modern Fracturing
to-hydrocarbon production ratio if mobile water The precipitation of solids such as paraffin,
saturation is present in the matrix (Sharma and hydrates, diamondoids or elemental sulfur generally
Wunderlich, 1985; Ballard and Dawe, 1987; Cuiec, is initiated by reductions in temperature and also
1989; and Sanner and Azar, 1994). are sometimes a function of a reduction in pressure
(Bennion et al., 1995). Placing a hydraulic fracture
Water
Flow into a reservoir often acts to reduce the pressure and
Gas Flow temperature differential between the reservoir and the
wellbore, reducing the tendency for these phenomena
to significantly affect productivity.
Relative Permeability
Relative Permeability
7-7 Fracturing Fluid Selection
Wellbore
Fracturing fluid selection based on laboratory generated

Water Saturation Water Saturation data has been detailed by Devine (2005). A flowchart,
Wettability Altered Unaltered Fig. 7-45, is used to illustrate the overall procedure. The
procedure utilizes mineralogical evaluation using x-ray
Figure 7-44 Effect of near-wellbore wettability alteration diffraction analysis and scanning electron microscopy of
on relative permeability to hydrocarbon (after Bennion et the formation core to understand potential sensitivities
al., 1993a) of the formation material to fracturing fluids. The
mineralogical evaluation is designed to identify the
7-6.3 Formation Damage from Production appropriate fluid system to be tested under flow
conditions with the core. These tests require some time
Damage can occur during normal gas production and may follow immersion testing and/or capillary
operations due to phenomena including physical fines suction time testing. Immersion testing of rock chips and
migration, gas condensate banking, paraffin deposition capillary suction time testing are used to screen fluids.
from wax-rich gas wells, diamondoids and hydrate If sensitivity evaluation is required rapidly, immersion
plugging problems, and elemental sulfur precipitation testing and/or capillary suction time testing provide
from high H2S-containing gas. valuable results. Flow testing of core plugs is the final step
As explained before, fines generally tend to in determining if systems cause minimal damage to the
preferentially migrate in the wetting phase. For gas formation. Plugs must be of adequate size and adequate
wells, this would usually be water, and significant permeability for meaningful flow testing.
problems with fines migration do not occur unless
interstitial shear rate caused by very high gas flow rates Request for
Scanning fluid sensitivity k>
causes mobilization of connate water. Electron of rock 5 md
Microscopy
Condensate drop-out can be a serious issue in gas
formations. Similar to aqueous phase trapping, if a gas No Rush? Yes
Immersion
Testing
well does not contain pre-existing hydrocarbon saturation, Cappillary
X-Ray
sufficient hydrocarbon saturation to build a continuous Adequate Diffraction
Suction Time Adequate
Testing k and size
liquid film to allow the flow of the condensate to the k and size Analysis
wellbore must occur. This is also called hydrocarbon Any k

Any Size
phase trapping. This critical condensate saturation Fluid
and
Type
is dependent on reservoir lithology, wettability, and Sensitivity
Flow
condensate composition and drawdown pressure. Tight Testing
gas systems, due to adverse capillary effects, often tend to
exhibit higher critical condensate saturation and thus are Figure 7-45 General procedures for determining
sensitivity of formation rock to fracturing fluids (from
more susceptible to this mechanism of damage. Devine, 2005).
268
7-7.1 Mineralogical Evaluation The x-ray region of the electromagnetic spectrum

extends from 0.03 nm to 80 nm. XRD measures
7-7.1.1 X-Ray Diffraction (XRD) Analysis the physical structure of a material and is basically
a ruler, with coherent scattering of the x-ray
Quantitative assessment of reservoir rock mineralogy beam yielding information about the materials
is normally obtained from analysis of core (sidewall crystal lattice spacing (Klug and Alexander, 1974).
or conventional) samples or cuttings. The most Scattering is related to lattice d-spacing and the angle
commonly applied method for this purpose is x-ray of diffraction defined by Braggs Law:
diffraction (XRD) (Figs. 7-46 and 7-47), based on
diffraction of x-rays by the crystal planes of minerals
n = 2d sin , (7-28)
(Ruessink and Harville, 1992).
where n is an integer, is wavelength of the
radiation, d is lattice spacing (in angstroms), and
is the angle of diffraction.
The capability of XRD to identify and quantify
minerals is strongly dependent on a minerals crystal
structure. Decrease in crystalline size, degree of ordering,
and crystallinity have detrimental effects on the intensity
of the diffracted beam. XRD is a bulk analysis technique
and therefore does not assess mineral morphology and
texture. This is possible in scanning electron microscopy
(SEM) and other microscopic methods.
Rock samples can be analyzed in a variety of
ways. Analysis of the entire sample at one time is
called whole-rock or bulk analysis. Segregation (by
particle size) is a common method of analyzing
minerals that occur preferentially in a specific
particle size range. Techniques such as fluorescence
microscopy can identify very small portions of a
sample. Figure 7-48 shows an XRD scan of mixed-
layer clays in a bulk core sample.
10.0
Figure 7-46 X-ray diffraction apparatus
8.0
intensity, Counts
6.0
4.0
2.0
x103
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Two-Theta, deg
Figure 7-48 XRD scan of mixed-layer clays in a bulk

Figure 7-47 X-ray diffraction apparatus sample chamber core sample
269
Modern Fracturing
The black line is the scan of the air-dried sample from the electron source to the specimen chamber
vs. the red scan, which is the same sample in glycolated and a new secondary electron detection system. The
state. The change in the pattern here shows that the detection system can operate at high chamber pressures
clays present are swelling clays. (up to 1 psi at 25 C) under most gas environments
including water vapor, oxygen, methane, argon, carbon
7-7.1.2 Scanning Electron Microscopy (SEM) dioxide or organic solvent vapors.

During the past three decades, the conventional
high vacuum scanning electron microscope (SEM)
has become an indispensable research tool for many
petrologic endeavors. These devices can provide sub-
micron resolution and a depth of field roughly two
orders magnitude greater than that of an optical
microscope at any given magnification, The applications
of conventional SEM in petroleum technology are
numerous. Characterizing reservoir rocks, understanding
processes that control rock petrophysical properties and
revealing details and complexities of diagenesis beyond
the resolution limits of traditional petrographic methods
are noteworthy using this equipment (Wells, 1974;
Bentelspacher and Van Der Marel, 1968; Pittman and
Thomas, 1978; Callender and Dahl, 1984).
The conventional SEM, however, needs vacuum
of the order of 10-7 psi in the specimen chamber, and
it can be used only to examine dry and electrically
conducting solid samples. Wet or oily samples, or those
that undergo physical or chemical transformation
under the electron beam, cannot be examined unless
dried or specially prepared. However, dried samples
or those subjected to moderate/high vacuum during
coating may lose their original characteristics and
are likely to incorporate artifacts that could affect
interpretation of the images (e.g., collapsed clays,
over estimation of porosity). Because of these
limitations, the conventional SEM images, although
useful, cannot provide unambiguous information
about the rocks under wet conditions.
The environmental scanning electron microscope
(ESEM) allows imaging of fluids and virtually any sample
in its natural wet or dry state, In addition, the microscope
can be used to observe and record in real-time, dynamic
processes such as drying, melting, swelling, dissolution Figure 7-49 SEM of Rotlingen sandstone
and precipitation, and surface chemical reactions at
temperatures up to 1000 C (Mehta, 1991). This was Small chips of selected core are cleaned, using toluene
accomplished by incorporating a differential pumping vapor extraction, dried at 150 F overnight, affixed to a
system that creates a pressure gradient of 10-8 to 1 psi standard aluminum SEM stubs, and coated with gold,
270
using a sputter coater. Samples are then individually 7-7.1.4 Capillary Suction Time Testing
placed in the chamber of the SEM, and vacuum is
applied. The principal objective of this examination is Capillary suction time testing is utilized to determine
to observe the distribution of clay minerals and other the sensitivity of rock samples to potential base
authigenic cements in the pore system. Secondarily, the fluids for fracturing systems. The technique provides
overall morphology of the pore system, particularly with numbers for use in comparing fluids. Water sensitivity
regard to pore sizes and the degree of interconnection and clay swelling control are tested for water-based
of open pores are examined. If required, rock/fluid fluids. The test gives the time of movement of a
interactions may be directly observed with the ESEM. waterfront between two electrodes on a porous paper
Figure 7-49 shows the SEM of Rotlingen sandstone, a (Gale and Baskerville, 1967). The time of movement
framework of grains coated with illite. of the waterfront is related to the ability of the fluid
to flocculate or disperse clays in the sample. Figure 7-
7-7.1.3 Immersion Testing 51 shows a photograph of the unit on the left and the
results of the test with swelling clay like bentonite,
Immersion testing is utilized to screen potential base which does not release any water.
fluids for fracturing systems and to qualitatively
determine the effect of various fluids on rock chips,
in a minimal amount of time. Typically, a drilled
core plug is cut into wafers, approximately 0.5 cm
thick. Samples are cleaned by toluene and methanol
extraction prior to testing. The purpose of the
cleaning is to strip any residual fluids from the pore
system, which might prevent rock contact with the Figure 7-51 Capillary suction time apparatus (left) and
test fluid. Clean, dry wafers are then immersed in results with bentonite (right)
test fluids. If the sample is small, one wafer may
be cut into chips for testing. The test is limited For comparing different samples in the same
by sample permeability. Figure 7-50 shows the fluid, the longer the time of waterfront movement,
results of exposure of core sample to two fracturing the greater the water sensitivity of the sample (the
fluids. When there is detrimental interaction of greater the dispersion). For comparing the same
the fluid with the formation sample, the core starts sample in different fluids, the longer the time of
disintegrating, as shown on the left. The fluid on the liquid front movement, the poorer the clay control
right is compatible with this formation. by the fluid (the greater the dispersion).
7-7.1.5 Core Flow Analysis
The final elaborate test protocol that can be used

to evaluate the compatibility of a fracturing fluid
with the formation involves evaluating the regain
permeability of the core after exposure to the fluid.
The use of regain permeability to evaluate fluids is
well-established procedure (Klinkenberg, 1941; and
Levine, 1954). Figure 7-52 details the protocol for
evaluating the core for fracturing fluid compatibility.
Residual hydrocarbon and other contaminants are
Figure 7-50 Core compatibility testing with fracturing
fluids showing incompatibility (left) and compatibility
removed from newly drilled or submitted plugs, using
(right), in a core from a field in Kenai Peninsula, Alaska the solvent vapor extraction technique.
271
Gas Well
Temperature
Lower? Higher?
220oF
No Low Permeability? Yes No Low Permeability? Yes
Figure 7-53 Fracturing fluid selection chart for gas wells

Under Under
Pressured? pressured?
No Yes No Yes
Water Water Sensitive? Water Sensitive?
Sensitive? Water Sensitive?
No Yes Yes
Yes No
No
Energized Oil Gel;
Borate Crosslinked Oil Energized Crosslinked Methanol Borate Oil Gel or cross
Crosslinked Gel, Methanol linked Methanol
crosslinked; foams Emulsion; Energized Crosslinked
Gel Gel
crosslink Methanol
No Yes
Energized Oil Gel;
Energized Crosslinked
Crosslinked, Methanol Emulsion;
foams Energized cross
Under linked Methanol
Underpressured?
Pressured?
No Yes
No Yes
Water Sensitive?
Water Sensitive? Water Sensitive?
Water Sensitive? Yes
No
No Yes Yes
Yes No Yes VES foams, energized oil
No VES foams,
VES foams, foams, cross VES foams, oil gel, Gel, energized methanol
foams, energized
Modern Fracturing
link water based fluids crosslinked methanol, gel, unconventional fluids

Slick Water, VES VES Fluid, Oil Energized VES Energized VES, crosslinked Fluids
unconventional fluids
Fluid, low polymer Gel, methanol Fluid, Energized Energized cross
Borate crosslinked Gel Borate; CO2 linked Methanol:
energized Zr unconventional
crosslinked fluids
272
Full Diameter Core is the core face as opposed to into the core. Effluent
subsampled to provide
1.0" or 1.5" diameter fluids are collected for 36 minutes so that fluid loss
cylindrical core plugs
parameters (wall building coefficient and spurt)
Flow rate and differential
Core plug is ceaned of residual
hydrocarbons, salts and other
pressure are monitored. An
effective permeability to gas
can be calculated. Fracturing fluids are shut-in to
contaminants by solvent vapor
extraction. The sample is vented
at irreducible water saturation
(Kg,Swi) is calculated
break. A post flush of the pad fluid is used after the
and dried to a stable weight
fracturing fluid has broken. All test fluids are injected
Basic Properties are determined: Treatment regime - includes
Dry N2 Permeability (Darcy's Law) preover flushes, leak-off, etc. in a direction opposite to that direction in which the
Helium Porosity (Boyle's Law)
Irreducible water saturation is
initial permeability was established, except the pad
Plug is fully saturated with
brine - Actual or synthetic
re-established.
over-flush that is injected in the production direction
formation brine Laboratory
standard brine, ex. 2% KCl Flow rate and differential
pressure are monitored. An
to simulate the flow of formation fluids from the
effective permeability to gas
Irreducible water saturation at a secondary reducible formation into the well bore.
is established. water saturation
(Kg,Swi) is calculated Final fluids are displaced from the core with
humidified N2 gas. Flow is established in the
Figure 7-52 Core flow testing protocol production direction with humidified N2 gas to
steady state. The regained permeability is established.
In this technique, extraction is continued until Several fluids can be evaluated using the procedure
fluid in contact with cores is colorless. Fluids are further to identify the best suitable fracturing fluid
stripped from the pore system by methanol extraction. compatible with the formation.
After cleaning, plugs are dried at 150 F for 16 hours.
Suction is applied to plug surfaces to remove residual 7-8 Selection of Fracturing Fluids for
fines (generated by trimming plugs) from surface pores. Applications in Gas Wells
Porosity is determined using a dual-cell Boyles law
porosimeter and helium gas pycnometer, and baseline With all the permutations and combinations available for
permeability established relative to dry N2 gas. These fracturing fluids, it is not an easy task to develop a simple
tests are generally at ambient temperature and 250 selection criteria for identifying the right fracturing fluid
psi confining pressure. To better simulate reservoir for a particular reservoir. Holditch et al., (1993), Xiong
conditions, N2 permeability may be determined at and Holditch (1995), and Xiong et al., (1996) tried to
reservoir temperature and a high confining pressure. use expert systems and fuzzy logic to arrive at a selection
process. despite all the efforts, the selection of fracturing
Core flow Procedure Gas Wells: Plugs are seated in fluid for a particular application is still as much an art
rubber sleeves at 1500 psi confining pressure and are as it is science. We have attempted to come up with a
pressure- and flow-saturated with filtered formation simplified chart (Fig. 7-53) that may be used as a first
water. If formation water is not available, laboratory- order approximation in narrowing choices based on the
prepared formation brine is used. This procedure discussion in this chapter for gas wells.
ensures that all test cores used in comparative testing
are at the same (or very similar) initial water-wet state
prior to treatment with the test fluid.
Prior to flow, systems are heated to the reservoir References
temperature. Flow is established in an arbitrary
formation-to-wellbore (production) direction with Aboud, R.S. and Melo, R.C.B.: Past Technologies
humidified N2 gas to steady-state permeability. A gas Emerge Due to Lightweight Proppant Technology:
flow meter is placed in line, ahead of the humidifier. Case Histories Applied on Mature Fields, paper
Cores are then treated with the test fluid. SPE 107184 2007.
If a crosslinked fracturing fluid is tested, a Acharya, A.: Particle Transport in Viscous and
prepad is injected before the main fracturing fluid. Viscoelastic Fracturing Fluids, SPEPE, 104,
Crosslinked fracturing fluids are injected across March 1986.
273
Modern Fracturing
Acharya, A.R.: Viscoelasticity of Crosslinked Bentelspacher, H., and Van Der Marel, H.W.: Atlas of
Fracturing Fluids and Proppant Transport, SPEPE, Electron Microscopy of Clay Minerals and Their
483, November 1988. Admixtures, Elsevier Publishing Co., 1968.
Aggour, T.M, and Economides, M.J.: Impact of Fluid Blot, M.A., and Medlin, W.L.: Theory of Sand
Selection on High Permeability Facturing, paper Transport in Thin Fluids, paper SPE 14468,
SPE 54536, 1999. 1985.
Alderman, E.N.: Super Thick Fluids provide New Boles, J.L., Metcalf, A.S., and Dawson, J.: Coated
Answers to Old Fracturing Problems, paper SPE breaker for crosslinked acid, U.S. Patent No.
2852, 1970. 5,497,830, 1996.
Asadi, M., Conway, M.W., and Barree, R.D.: Zero Boothe, J.E., and Martin, F.D.: Thickened Alcohol
Shear Viscosity Determination of Fracturing Fluids: Well Treating Compositions, U.S. Patent No.
An Essential Parameter In Proppant Transport 4,012,327, 1977.
Characterizations, paper SPE 73755, 2002. Brannon, H.D., and Ault, M.G.: New, Delayed Borate-
Aud, W.W., Wright, T.B., Cipolla, C.L., Harkrider, Crosslinked Fluid Provides Improved Fracture
J.D., and Hansen, J.T.: The Effect of Viscosity Conductivity in High-Temperature Applications,
on Near-Wellbore Tortuosity and Premature paper SPE 22838, 1991.
Screenouts, paper SPE 28492, 1994. Brannon, H.D., and Tjon-Joe-Pin, R.M.:
Ballard, T.J., and Dawe, R.J.: Wettability Alteration Biotechnological Breakthrough Improves
Induced by Oil Based Drilling Fluids, paper SPE Performance of Moderate to High-Temperature
17160, 1987. Fracturing Applications, paper SPE 28513, 1994.
Bennion, D.B., Bietz, R.F., Thomas, F.B., Cimolai, Brannon, H., Wood, B., and Wheeler, R.: The Quest
M.P.: Reduction in the Productivity of Oil and for Improved Proppant Placement: Investigation
Gas Reservoirs Due to Aqueous Phase Trapping, of the Effects of Proppant Slurry Component
paper CIM 93-24, 1993a. Properties on Transport, paper SPE 95675 2005.
Bennion, D.B., Chan, M., Sarioglu, G., Courtnage, Bullen, R.S., Mzik, J., Richard, J.P.: Novel compositions
D., Wansleeben, J., and Hirata, T.: The In Situ suitable for Treating Deep Wells, U.S. Patent No.
Formation of Bitumen-water Stable Emulsions in 4,701,270, 1987.
Porous Media During Thermal Stimulation, paper Caldwell, J.E., and Sorrell, R.R.: Method of Refining
CIM 93-46, 1993b. Esters of Organophosphorus, U.S. Patent No.
Bennion, D.B., Cimolai, M.P., Bietz, F.R., Thomas, 3,706,822, 1972.
F.B.: Reductions in the productivity of oil and gas Callender, C.H., and Dahl, H.M.: Characterization of
reservoirs due to aqueous phase trapping, JCPT Petroleum Sandstone Reservoir Rocks by Scanning
(November 1994). Electron Microscopy, in Scanning Electron
Bennion, D. B., Thomas, F.B., and Bietz, R.F.: Microscopy, SEM Inc., Chicago, 1984.
Low Permeability Gas Reservoirs: Problems, Canterino, P.J.: Increasing Viscosity of Normally
Opportunities and Solutions for Drilling, Liquid Hydrocarbons and Gels Produced, U.S.
Completion, Stimulation and Production, paper Patent 3658491, 1972.
SPE 35577, 1996a. Carman, P.C., and Cawiezel, K.: Successful Breaker
Bennion, D.B., Thomas, F.B., Bietz, R.F., and Bennion, Optimization for Polyacrylamide Friction Reducers
D.W.: Water and Hydrocarbon Phase Trapping Used in Slickwater Fracturing, paper SPE 106162,
in porous Media, Diagnosis, Prevention and 2007.
Treatment, JCPT, (December 1996b). Carman, P.C.: Private Communication, 2007.
Bennion, D.B., Thomas, F.B., Bietz, R.F., and Bennion, Cawiezel, K.E.: Non-interfering breaker system
D.W.: Water and hydrocarbon phase trapping for delayed crosslinked fracturing fluids at low
in porous media Diagnosis, Prevention and temperature, U.S. Patent No. 4,969,526, 1990.
Treatment, paper CIM 95-69, 1995. Cawiezel, K.E., Wheeler, R.S., and Vaughn, D.R.:
274
Specific Fluid Requirements for Successful Coiled Return by Minimizing or Preventing Aqueous Phase
Tubing Fracturing Applications, paper SPE Trapping During Completion and Stimulation
86481, 2004. Operations, paper SPE 90170, 2004.
Cinco-Ley, H., and Samaniego-V., F.: Transient Dawson, J.C., Le, H.V., and Kesavan, S.: Polymer
Pressure Analysis: Finite Conductivity Fracture expansion for oil and gas recovery, U.S. Patent
Case Versus Damage Fracture Case, paper SPE No. 6,017,855, 2000.
10179, 1981. Dawson, J.C., Method for improving the high
Cinco-Ley, H., Samaniego-V., F., and Dominguez-A., temperature gel stability of borated galactomannans,
N.: Transient Pressure Behavior for a Well with a U.S. Patent No. 5,145,590, 1992.
Finite-Conductivity Vertical Fracture, SPEJ, (Aug. Dawson, J.C., and Le, H.V.: Fracturing using gel with
1978) 253-264. ester delayed breaking, U.S. Patent No. 6,793,018,
Civan, F., and Knapp, R.M.: Effect of Clay Swelling 2004.
and Fines Migration on Formation Permeability, De Kruijf, A., Davies, D.R., and Fokker, P.A.: Novel
paper SPE 16235, 1987. Rheological Equipment for the Characterization
Clark, J.B.: A Hydraulic Process for Increasing the of Hydraulic Fracturing Fluids, paper SPE 27398,
Productivity of Wells, paper SPE 949001, 1949. 1994.
Clark, P.E.: Stimulation Fluid Rheology A New Devine, C.S.: Approaches to Fracturing Fluid Selection
Approach, paper SPE 8300, 1979. Based on Laboratory Generated Data, paper CIM
Clark, P.E.: Transport of Proppant in Hydraulic 2005-228, 2005.
Fractures, paper SPE 103167, 2006. di Lullo Arias, G.F., Rae, P., and Ahmad, A.J.K.:
Cleary, M.P., Johnson, D.E., Kogsbll, H.H., Viscous fluid applicable for treating subterranean
Owens, K.A., Perry, K.F., de Pater, C.J., Stachel, formations, U.S. Patent No. 6,491,099, 2002.
A., Schmidt, H., and Tambini, M.: Field Economides, M.J., and Nolte, K.J., Reservoir
Implementation of Proppant Slugs To Avoid Stimulation, Prentice Hall, Englewood Cliffs, NJ,
Premature Screen-Out of Hydraulic Fractures 1989.
With Adequate Proppant Concentration, Ely, J.W., Hargrove, J.S., Wolter, B.C., and Kromer,
paper SPE 25892, 1993. J.C.: Pipelining: Viscous Fingering Prop Fracture
Constien, V.G.: Fracturing Fluid and Proppant Technique Finds Wide Success in Permian and
Characterization, Chapter 5, Reservoir Delaware Basins, paper SPE 26528, 1993.
Stimulation, 1st Ed., Economides, M.J., and Ely, J.W.: Stimulation Engineering Handbook,
Nolte, K.S., Eds., Prentice Hall, Englewood Cliffs, PennWell Books, Tulsa, OK, 1994.
NJ, 1989. Enzendorfer, C., Harris, R.A., Valk, P., Economides,
Craigie, L.J.: A New Method for Determining the M.J., Fokker, P.A., and Davies, D.D.: Pipe
Rheology of Crosslinked Fracturing Fluids using viscometry of foams, J. Rheology, 39:2, pp. 345-
Shear History Simulator, paper SPE 11635, 358, 1995.
1983. Flanagan, P.W.K.: Preparation of Gelled Liquids,
Crema, S.C., and Alm, R.R.: Foaming of Anhydrous U.S. Patent 3,243,270, 1966.
Methanol For Well Stimulation, paper SPE Gale R.S., and Baskerville R.C.: Capillary Suction
13565, 1985. Method For Determination Of The Filtration
Cuiec, L.: Effect of Drilling Fluids on Rock Surface Properties Of A Solid/Liquid Suspension, Chem.
Properties, paper SPE 15707, 1989. Inc., 9, 355, 1967.
Dahanayake, M.S., Yang, J., Niu, J.H.Y., Derian, P.J., Gdanski, R.D., Harris, P.C., Lord, D.L., Reidenbach,
Li, R., and Dino, D.: Viscoelastic surfactant B., and Shah, S.N.: Analysis of Fann Model 50
fluids and related methods of use, U.S. Patent No. for Fracturing Fluid Rheology, paper SPE 21647,
6,831,108, 2004. 1991.
Davis, B.J., and Wood, W.D.: Maximizing Economic Gidley, J.L., Holditch, S.A., Nierode, D.E., and Veatch,
275
Modern Fracturing
R.W. Jr.: Recent Advances in Hydraulic Fracturing, Aqueous Methanol Base Fluid: Chemistry and
SPE Monograph Volume 12, 1989. Applications, paper SPE 106304, 2007.
Goel, N., and Shah, S.: A Rheological Criterion for Gupta, D.V.S., Niechwiadowicz, G., and Jerat, A.C.:
Fracturing Fluids to Transport Proppant during a CO2 Compatible Non-Aqueous Methanol
Stimulation Treatment, paper SPE 71663, 2001. Fracturing Fluid, paper SPE 84579, 2003.
Grantham, C.K., and McLaurine, H.C.: Gupta, D.V.S., Pakulski, M., Prasek, B., and Franklin,
Thixotropy Without Viscosity: A new V.: High-pH-Tolerant Enzyme Breaker for Oilfield
Approach to Rheology Control of Oil Muds, Applications, paper SPE 23986, 1992.
paper SPE 15415, 1986. Gupta, D.V.S., Pierce, R., Litt, N.D., and Jerat, A.C.:
Griffin, T.J. Jr.: Method of Fracturing Organic Fracturing fluids containing borate esters as
Liquids, U.S. Patent No. 4,174,283, 1979. crosslinking agents and method of using same,
Guillot, D., and Dunand, A.: Rheological U.S. Patent No. 7,007,757, 2006.
Characterization of Fracturing Fluids by Using Gupta, D.V.S., Pierce, R.G., and Litt, N.D.: Non-
Laser Anemometry, paper SPE 12030, SPE Journal aqueous Gelled Alcohol Fracturing Fluid, paper
(February 1985), 39. SPE 37229 (1997).
Gupta, D.V.S.: Field Application of Unconventional Hackley, V.A., and Ferraris, C.F.: The Use of
Foam Technology: Extension of Liquid CO2 Nomenclature in Dispersion Science and
Technology, paper SPE 84119, 2003. Technology, NIST Recommended Practice Guide,
Gupta, D.V.S., and Bobier, D.M: The History and Special publication 960-3, US Department of
Success of Liquid CO2 and CO2/N2 Fracturing Commerce, 2001.
System, paper SPE 400016, 1998. Hall, B.E., and Szemenyei, C.A.: Fluid additive and
Gupta, D.V.S., and Cooney, A.: Encapsulations for method for treatment of subterranean formations,
treating subterranean formations and methods U.S. Patent No. 5,089,151, 1992.
for the use thereof, U.S. Patent No. 5,164,099, Harris, P.C.: Fracturing-Fluid Additives, JPT
1992. (October 1988) 40:10, 1277.
Gupta, D.V.S., and Franklin, V.L.: Method for Harris, P.C., and Heath, S.J.: Rheological Properties of
fracturing high temperature subterranean Low-Gel-Loading Borate Fracture Gels, SPEPF,
formations, U.S. Patent No. 5,305,832, 1994. (November 1998) 230-235.
Gupta, D.V.S., and Hlidek, B.T.: Method and Harris, P.C., and Heath, S.J.: Rheology of Borate
composition for treating a subterranean formation Crosslinked Foams, SPEPF (May 1996) 113.
with splittable foams, U.S. Patent No. 7,201,227, Harry, D.N., Putzig, D.E., Moorhouse, R., Del Presco,
2007. T., and Jernakoff, P.: Chemical Structures of Group
Gupta, D.V.S., and Leshchyshyn, T.T.: CO2 Energized 4 Metal Crosslinkers for Polygalactomannans,
Hydrocarbon Fracturing Fluid: History & Field paper SPE 50731, 1999.
Application in Tight Gas Wells in the Rock Creek Harry, D.N., Moorhouse, R., and Mathews, L.:
Gas Formation, paper SPE 95061, 2005a. Rheological Responses to Variations in aqueous-
Gupta, D.V.S., and Leshchyshyn, T.T.: Surfactant Gel Based Zirconium Crosslinker Chemistry, paper
Foam/Emulsion: History and Field Application in SPE 37280, 1997.
Western Canadian Sedimentary Basin, paper SPE Hassebroek, W.E., and Waters, A.B.: Advancements
97211, 2005b. through 15 years of Fracturing, paper SPE 801,
Gupta, D.V.S., and Tudor, E.H.: Method for 1964.
fracturing subterranean formations, U.S. Patent Hazen, C.: Hydrocolloid Handbook, Food Product
No. 6,875,728, 2005. Design (October 2004).
Gupta, D.V.S., Hlidek, B.T., Hill, E.S.W., and Dinsa, Hendrickson, A.R., Nesbitt, L.E., and Oakes, B.D.:
H.S.: Fracturing Fluid for Low-Permeability Gas Soap-Oil Systems for Formation Fracturing,
Reservoirs: Emulsion of Carbon Dioxide With paper SPE 685, 1956.
276
Hendrickson, A.R., Nesbitt, L.E., and Oakes, B.D.: Practices, API, 200, 1941.
Soap-Oil Systems for Formation Fracturing, The Klug, H.P., and Alexander, L.E.: X-ray Diffraction
Petroleum Engineer, May 1957. Procedures for Polycrystalline and Amorphous
Hernandez, J.M., Fernandez, C.T., and Sciana, N.M.: Materials, John Wiley & Sons, 1974.
Methanol as Fracture Fluid in Gas Wells, paper Knoll, S.K., and Prudhomme, R.K.: Interpretation
SPE 27007, 1994. of Dynamic Oscillatory Measurements for
Hill, A., and Carrington, S.: Understanding the Characterization of Well Completion Fluids,
link between Rheology and Particle Parameters, paper SPE 16283, 1987.
American Laboratory News, November/December Kramer, J., Prudhomme, R.K., Norman, L.R., and
2006. Sandy, J.M.: Characteristics of Metal-Polymer
Himes, R.E., and Vinson, E.F.: Stabilizing clayey Interactions in Fracturing Fluid Systems, paper
formations, U.S. Patent No. 4,828,726, 1989. SPE 16914, 1987.
Holditch, S.A.: Factors Affecting Water Blocking and Kraynik, A.M., Aubert, J.H., Chapman, R.N., and
Gas Flow from Hydraulically Fractured Gas Wells, Guyre, D.C.: The Helical Screw Rheometer: A
paper SPE 7561, 1978. New Concept in Rotational Rheometry, SPE
Holditch, S.A., Xiong, H., Rueda, J., Rahim, Z.: Using ANTEC, April 1984.
an Expert System To Select the Optimal Fracturing Lei, C and Clark, P.E.: Crosslinking of Guar and Guar
Fluid and Treatment Volume, paper SPE 26188, Derivatives, paper SPE 90840, 2004.
1993. Levine, J.S., Displacement Experiments in a
Hunt, J.L., Chen, C.C., Soliman, M.Y.: Performance Consolidated Porous System Trans. AIME, 1954.
of Hydraulic Fractures in High Permeability Lillies, A.T.: Sand Fracturing with Liquid Carbon
Formations, paper SPE 28530, 1994. Dioxide, paper CIM 82-33-23, 1982.
Jales, W., Grisdale, J., and Hossaini, M.: Fracturing Liu, Y., and Sharma, M.M.: Effect of Fracture
with Crosslinked Gelled Methanol: A New Width and Fluid Rheology on Proppant
Approach to Well Stimulation, paper CIM 88- Settling and Retardation: An Experimental
39-29, presented at the Petroleum Society CIM Study, paper SPE 96208, 2005.
Annual Technical Meeting, Calgary, 1988. Lord, D.L.: Helical Screw Rheometer: A New Tool for
Jin, L., and Penny, G.S.: Dimensionless Methods for Stimulation Fluid Evaluation, paper SPE 18213,
the Study of Particle Settling in Non-Newtonian 1988.
Fluids, JPT, (March 1995) 223. Lord, D.L., and Shackelford, D.: Real-Time Fracturing
Kalfayan, L.J.: Fracture Acidizing: History, Fluid Rheology Measurements With the Helical
Present State, and Future, paper SPE 106371, Screw Rheometer, paper SPE 19734, 1989.
2007. Mathur, A.K., Ning, X., Marcineau, R.B., Ehlig-
Keck, R.G., Nehmer, W.L., and Strumolo, G.S.: Economides, C.A., and Economides, M.J.:
A method for Predicting Friction Pressures Hydraulic Fracture Stimulation of Highly
and Rheology of Proppant laden Fracturing Permeable Formations: The Effect of Critical
Fluids, SPEPE (February, 1992). Fracture Parameters on Oilwell Production, paper
Keelan , D.K.: Rock Properties and their effect on Gas SPE 30652, 1995.
Flow and Recovery, presented at the 22nd Annual Mazza, R.L.: Liquid CO2 improves Fracturing, Harts
Southwestern Petroleum Short Course, Texas Tech Oil and Gas World (February 1997), 22.
Univeristy, Lubbock, Texas April 17-18, 1975. McLaughlin, H.C., Elphingstone, E.A., and Hall, B.E.:
Khade, S.D., and Shah, S.N.: New Rheological Aqueous Polymers for Treating Clays in Oil and
Correlations for Guar Foam Fluids, SPEPF (May Gas, paper SPE 6008, 1976.
2004) 19:2. McLeod, H.O., and Coulter, A.W.: The Use of Alcohol
Klinkenberg, L.J.: The Permeability of Porous Media in Gas Well Stimulation paper SPE 1033, 1966.
to Liquids and Gases, Drilling and Production Mehta, S.: Imaging of Wet Specimens in Their
277
Modern Fracturing
Natural State Using Environmental Scanning Poklacki, E.S.: Polyamine Salts of Aluminum Alkyl
Electron Microscope (ESEM): Some Examples of Orthophosphates, U.S. Patent No. 4,007,128,
Importance to Petroleum Technology, paper SPE 1977.
22864, 1991. Prudhomme, R.K., Ellis, S., Constien, V.G., and Knoll,
Mitchell, B.J.: Viscosity of Foam, Ph.D. Thesis, S.: Reproducible Rheological Measurements on
University of Oklahoma, 1969. Crosslinked Fracturing Fluids, paper SPE 18210,
Mondshine, T.C.: Crosslinked fracturing fluids, 1988.
U.S. Patent No. 4,619,776, 1986. Reidenbach, V.G, Harris, P.C., Lee, Y.N., and Lord,
Mondshine, T.C.: Process for decomposing D.L.: Rheological Study of Foam Fracturing
polysaccharides in alkaline aqueous systems, U.S. Fluids Using Nitrogen and Carbon Dioxide,
Patent No. 5,253,711, 1993. SPEPE (January 1986) 31.
Monroe, R.F., and Rooker, B.E.: Aluminum Salts Rieb, B.: Private Communication, 2007.
of Alkyl Orthophosphates, U.S. Patent No. Roodhart, L.P.: Proppant Settling in Non-Newtonian
3,494,949, 1970. Fracturing Fluids, paper SPE/DOE 13905, 1985.
Moorhouse, R., Harry, D.N., Mathews, L and Ruessink, B.H., and Harville, D.G.: Quantitative
Merchant, L.: Inter-relationships between Analysis of Bulk Mineralogy: The Applicability
Polymer/Crosslinker Chemistry and Performance and Performance of XRD and FTIR, paper SPE
in Fracturing Fluids, paper SPE 39531, 1996. 23828, 1992.
Mzik, J.: High Viscosity Crosslinked Gelled Alcohol, Russell, D.: Recycling of Water & Minimizing Fresh
Canadian Patent Application No. 2,108,194, Water Usage, presented at Shallow Gas Forum,
1993. PTAC, Calgary, Canada, June 2001.
Mzik, J.: High Viscosity Alcohol Gel For Shallow Sanner, D.O., and Azar, J.J.: Alteration of Reservoir
Wells, Canadian Patent Application No. Rock Wettability and its Flow Properties Caused
2,119,682, 1994. by Oil Based and Water Based Drilling Muds,
Novotny, E.J.: Proppant Transport, paper SPE 6813, paper SPE 27354, 1994.
1977. Savins, J.G.: Generalized Newtonian (Pseudoplastic)
Pakulski, M.K., and Gupta, D.V.S.: High temperature Flow and Stationary Pipes and Annuli, Trans.,
gel stabilizer for fracturing fluids, U.S. Patent No. AIME (1958) 213, 325.
5,362,408, 1994. Schols, R.S., and Visser, W.: Proppant Bank Buildup
Penny, G.S., Pursley, J.T., and Holcomb, D.: The in a Vertical Fracture Without Fluid Loss, paper
Application of Microemulsion Additives in SPE 4834, 1974.
Drilling and Stimulation Results in Enhanced Gas Schwartz, K.N., and Smith, K.W.: Friction reducing
Production, paper SPE 94274, 2005. composition and method, U.S. Patent No.
Petkowicz, C.L.O., Reicher, F., and Mazeau, K.: 7,205,262, 2007.
Conformational Analysis of Galactomannans: Settari, A., Bachman, R.C., Morrison, D.C.: Numerical
from Oligomeric Segments to Polymeric Chains, Simulation of Liquid CO2 Hydraulic Fracturing,
Carbohydr. Polym. 37, 25-39, 1998. paper CIM 86-37-67, 1986.
Phillips, A.M., and Mack, D.J.: Fracturing process Sharma, M.M., and Wunderlich, R.W.: The Alteration
using a viscosity stabilized energizing phase, U.S. of Rock Properties Due to Interactions with
Patent No. 5,002,125, 1991. Drilling Fluid Components, paper SPE 14302,
Pianfetti, J.A., and Janey, P.L.: Production of Aliphatic 1985.
Phosphate Esters, U.S. Patent No. 3,020,303, Smith, C.F.: Gas Well Fracturing using Gelled Non-
1962. Aqueous Fluids, paper SPE 4678, 1973.
Pittman, E.D., and Thomas, J. B.: Some Applications Smith, K.W., and Persinski, L.J.: Hydrocarbon Gels
of Scanning Electron Microscopy to the study of useful in Formation Fracturing, U.S. Patent No.
Reservoir Rock, paper SPE 7550, 1978. 5,471,287, 1995.
278
Smith, K.W., and Persinski, L.J.: Hydrocarbon Gels Fracturing Design and Treatment Technology,
useful in Formation Fracturing, U.S. Patent No. JPT (April 1983) 35, p853.
5,571,315, 1996. Virk, P.S.: Drag Reduction Fundamentals, AIChE J.,
Smith, K.W., and Persinski, L.J.: Hydrocarbon Gels 21, 625 (1975).
useful in Formation Fracturing, U.S. Patent No. Walles, W.E., Williamson, T.D., and Tomkinson, D.L.:
5,614,010, 1997a. Method for treating subterranean formations,
Smith, K.W., and Persinski, L.J.: Ferric alkyl U.S. Patent No. 4,741,401, 1988.
amine citrates and methods of making them, Wang, X., Qu, Q., McCarthy, S., Null, J., Bowen,
U.S. Patent No. 5,693,837, 1997c. K., and Neumann, L.: Successful Applications
Smith, K.W., and Persinski, L.J.: Preparation of of Borate Crosslinked Fracturing Fluids at High
hydrocarbon gels from ferric sources, polycarboxylic Temperature, paper SPE 73789, 2002.
acid compounds, and optional amines, in Wells, O.C.: Scanning Electron Microscopy, McGraw
combination with phosphate esters, U.S. Patent Hill, 1974.
No. 5,647,900, 1997c. Williams, B.B.: Fluid Loss from Hydraulically Induced
Sudhakar D.K., and Shah, S.N.: New Rheological Fractures, JPT (July 1970) 22:6, 882888.
Correlations for Guar Foam Fluids, SPEPF, (May Winkler W., Valk, P. and Economides, M.J.: Laminar
2004) 7785. and Drag Reduced Polymeric Foam Flow, Journal
Teot, A.S., Rose, G.D., and Stevens, G.A., Friction of Rheology, Vol. 38 (1) pp. 111-127, 1994.
reduction using a viscoelastic surfactant, U.S. Wright, T.B., Cipolla, C., Perry, K.F., and Cleary,
Patent No. 4,615,825, 1981. M.P.: Identification and Comparison of True
Teot, A.S., Ramaiah, M., and Coffey, M.D.: Aqueous Net Fracturing Pressures Generated by Pumping
Wellbore Service Fluids, U.S. Patent No. Fluids with Different Rheology into the Same
4,725,372, 1988. Formations, paper SPE 26153, 1993.
Thesing, A.: New Device for Rheology Measurements Xiong, H., and Holditch, S.A.: An Investigation Into
of Proppant-laden Fluids with the Fann 50 the Application of Fuzzy Logic to Well Stimulation
viscometer, paper SPE 58759, 2000. Treatment Design, paper SPE 27672, 1995.
Thompson, J.E. Sr., McBain, C., Gregory, G., and Xiong, H., Davidson, B., Saunders, B., and
Gerbrandt, D.: New Continuous-Mix Process Holditch, S.A.: A Comprehensive Approach
for Gelling Anhydrous Methanol Minimizes to Select Fracturing Fluids and Additives
Technology, JPT (July 1992) 832. for Fracture Treatments, paper SPE 36603,
Tiner, R.L., Stahl, E.J., and Malone, W.T.: 1996.
Developments in Fluids to Reduce Potential Zhang, K.: Fluids for fracturing subterranean
Damage from Fracturing Treatments, paper SPE formations, U.S. Patent No. 6,468,945,
4790, 1974. 2002.
Toney, F.L., and Mack, D.J.: The Next Generation of Zhang, K., Pierce, R., Litt, N.D., and Gupta,
Foam: A Field Study of Northwestern Oklahoma D.V.S.: Foam-fluid for fracturing subterranean
Foam Fracturing, paper SPE 21644, 1991. formations, U.S. Patent No. 6,410,489,
Tudor, R., Vozniak, C., Banks, M.L., and Peters W.: 2002.
Technical Advances of Liquid CO2 fracturing,
paper CIM 94-36, 1994.
Valk, P., and Economides M.J.: Hydraulic Fracture
Mechanics, Wiley, 1985.
Valk, P., and Economides, M.J.: Volume Equalized
Equations for Foamed Polymer Solutions, J.
Rheology, 36, pp. 1033-1055, 1992.
Veatch, R.W. Jr.: Overview of Current Hydraulic
279
Modern Fracturing
280
Dr. C. Mark Pearson is president and CEO of Golden Energy, LLC, and an adjunct professor of petroleum
engineering at the Colorado School of Mines. Pearson has BS and PhD degrees from the Camborne School
of Mines (UK) and graduated from the Advanced Management Program of the Harvard Business School.
He has over 30 years of experience in the oil and gas/energy industries, working in both the E&P and service
sectors. His E&P experience includes 12 years with Atlantic Richfield Company (ARCO), where he held a
variety of technical and management positions in engineering and operations. From 2001 to 2005, he was
president and CEO of NYSE-listed Carbo Ceramics Inc.
Harold D. Brannon is senior advisor for fracturing at BJ Services Company in the Corporate Technology
Center in Tomball, Texas. He has 27 years of industry experience in well completion, stimulation and
remediation technologies and has held positions in several arenas including R&D, product line management,
treatment engineering, technical support, chemical manufacturing, marketing and intellectual property.
Harold has a BS in chemistry from Baylor University. He is an active SPE member, having previously served
on various technical program committees including those for the Hydraulic Fracturing Conference, ATCE
Well Completions, ATCE Well Stimulation (chairman), and the Formation Damage Symposium. He is
currently serving as the Editorial Review Chair for the SPE Production & Operations Journal. Harold also
sits on several API and ISO Standard Practice Committees, has authored and co-authored over 60 technical
publications, and is listed as an inventor on over 35 US patents, issued and pending.
Chapter 8 Typically, the engineer has two primary questions to
answer: which proppant? and how much is needed?
This chapter reviews the criteria to answer these two
Proppants and Fracture questions. Although the optimum size of the fracture
Conductivity (the lateral and vertical extent of contact with the
Harold D. Brannon, BJ Services and reservoir) is briefly discussed, the primary focus is on
C. Mark Pearson, Golden Energy LLC understanding the effective conductivity of the propped
fracture and the characteristics of different proppant
8-1 Introduction types. The impacts of different reservoir conditions
on the proppant are reviewed with the laboratory
8-1.1 Overview procedures to measure proppant properties and how
they are used in the fracture treatment design.
To successfully enhance well productivity or injectivity,
a created fracture must have permeability (leading to 8-1.2 The Evolution of Proppants
conductivity) orders of magnitude greater than the
reservoir matrix permeability. When pumping has In the 1940s, the first experimental fracture treatments
ceased and the imparted hydraulic fluid pressure has in the industry did not use proppant. It quickly became
dropped below that required to keep the fracture open, apparent that an unpropped fracture healed, and a solid
the fracture may close, and in doing so, substantially material needed to be injected to facilitate production
eliminate the desired conductive pathway to the from the reservoir and to prevent the fracture walls from
wellbore. Proppants, or propping agents, are placed in closing (Howard and Fast, 1970). One of the earliest
the fracture to maintain the flow path after the treating proppants used in the 1950s was sand dredged from the
pressure is relieved. Ideally, the proppant will provide Arkansas River. Later, it became evident to engineers that
flow conductivity large enough to minimize pressure productivity could be improved by using screened and
losses in the fracture during production. In practice, processed sand. Monocrystalline sand was used from the
this may not always be achieved because the selection Saint Peter formation near Ottawa, Illinois. Referred to as
of a proppant involves many compromises imposed by white or Ottawa sand, the mined sand consists of grains
economic and practical considerations. that are single quartz crystals that offer superior strength
Thus, the proppant that is pumped as part of a properties compared to other sands. With the rapid
fracture stimulation treatment is an integral and vital increase in the number of fracturing treatments, increased
part of the completion system because it provides the proppant demand warranted additional supplies. In
connection for hydrocarbons to flow between the reservoir 1958, brown sand quarries were opened in the Hickory
and the producing wellbore. Of equal importance is sandstone formation near Brady, Texas. These sands are
optimization of the conductive or effective fracture area, polycrystalline, and each grain is composed of multiple
since stimulated well performance is directly dependent crystals bonded together. The existence of cleavage planes
upon the producibility of hydrocarbons from that area. within each grain results in greater proppant crush and
The conductive fracture area is defined by the propped reduced strength properties. The white Ottawa and
fracture height within the productive zone and the effective brown Texas sands continue to be the major sources of
fracture length. This metric is strongly influenced by the frac sand to the industry together with some production
ability to transport or place the proppant in the desired from the Jordan, Ironton, and Galesville sandstones
location. Flow within the propped fracture is concentrated in Minnesota and Wisconsin (Hoaberg and Koerner-
through a relatively small cross-sectional area. Therefore, Moore, 1978; Gidley et al., 1989). In international
great attention needs to be given to the design of this flow markets outside of North America, a number of sand
path in order to optimize the production and recovery sources are used locally, including sources in Australia,
from the reservoir. See Sections 2-5 and 4-5 for fractured Indonesia, Oman, Peoples Republic of China, Poland,
well performance and Section 4-6 for treatment design. Saudi Arabia, and the UK.
283
Modern Fracturing
Fracturing fluid technology was still in its infancy pellets are fired in a kiln to produce a strong crystalline
in the 1960s. A primary concern was the ability of the structure. These two products find wide application in
fluid to transport proppant down the fracture away from gas wells at greater than 10,000 ft depth.
the wellbore. Design proppant concentrations were kept The many problems associated with brittle failure
low and typically limited to placing less than lb/ft2 of of quartz spawned development of an improved sand-
proppant in the fracture. Glass beads were utilized with based product. Resin-coated sand (RCS) proppants
lighter, deformable materials such as walnut shells and were introduced, using a phenolic resin coating to
plastic beads. Engineering work focused on attempting encapsulate each sand grain. They are primarily used
to maximize conductivity by placing a partial monolayer in applications where regular sand does not quite give
of proppant such that flow capacity was obtained from sufficient permeability yet the added cost of a ceramic
the voids between proppant grains. Unfortunately, the proppant is not justified. The resin improves the
theoretical benefits of a partial monolayer were rarely properties of the sand by reducing the grain angularity
realized for several reasons: and the amount of crush by distributing the load
The proppant settled into a bed at the bottom of the more evenly and encapsulating the fines in the resin
fracture. coating (Graham et al., 1975; Johnson and Armbruster,
The proppant was brought into grain-to-grain 1984). Pre-cured resin-coated sand proppants are used
contact as the fracture closed due to fluid dehydration in shallow and moderate-depth fracture treatments of
after pumping. lower producing rate formations and occasionally have
Embedment effects were severe in softer formations, found application in frac-pack treatments.
resulting in loss of effective width. In 1985, the first lightweight ceramic (LWC)
Severe stress was applied to individual proppant proppant was introduced, with Al2O3 content around
grains because the closure stress over the entire 50% (Lunghofer, 1985). Although it is not as strong as
fracture face had to be supported by only a partial a bauxite or ISP proppant, it has similar density to sand
layer of proppant. with greater sphericity and improved strength. It found
application in shallower formations where production
Recognizing these problems, soluble proppant could be improved by increasing fracture conductivity
spacers were also mixed with the proppant material in beyond that of sand or resin-coated sand proppants.
a largely unsuccessful attempt to increase conductivity This product was used for the first extensive high-
after they were dissolved by produced hydrocarbons. permeability, hard-rock formation fracturing programs,
Another issue facing the industry was the which occurred in Alaska in the late-1980s (Pearson
exploitation of deep gas reservoirs that required hydraulic et al., 1992). Similarly, in the 1990s, LWC proppant
fracturing for successful commercialization. This was found wide application as fracturing technology
addressed in the mid-1970s when Exxon Production was applied to high-permeability, unconsolidated
Research invented the use of ceramic proppants by formations and is the most widely used proppant for
sintering pellets of bauxite an alumina-silicate clay frac-pack completions.
containing over 80% Al2O3 (Cooke et al., 1978). First Today there are over 50 different types of proppant
commercial production of bauxite proppant occurred in products coming in a variety of mesh sizes. However,
1979 for the exclusive use of fracturing deep gas wells. most can be categorized into one of the following
This was followed in 1982 by the introduction of what proppant types:
is commonly referred to as an Intermediate Strength Natural sand (white mono-crystalline, or brown
Proppant (ISP), which is manufactured from a raw multi-crystalline)
material with around 70% Al2O3 (Fitzgibbon, 1984). Resin-coated sand using a variety of phenolic and
Both of these products are manufactured from high- other coatings
alumina ores. They are first ground to a particle size less Ceramic proppants, ranging from lightweight
than 15 m and then formed into pellets using water ceramics with an Al2O3 of 45-50% to bauxites with
and binder in a high-intensity mixer. After drying, the an Al2O3 content in excess of 80%.
284
Chapter 8 Proppants and Fracture Conductivity
8-1.3 Fracture Conductivity STB). Therefore, in a fracture of w = 1 in., h = 100 ft,

with appropriate conversions v = 0.014 m/s (half of the
The fracture conductivity defines the conductive flow in each wing). Using Eq. 2-37 with = 800 kg/m3
path provided by the proppant material to enhance and = 1 cp (10-3 Pa.s) the Reynolds Number is 0.15,
deliverability and provide economic benefit when decidedly in the laminar flow regime.
the well is placed on production. Traditionally this is For natural gas flow, this is significantly different.
measured as the product of proppant permeability and For 1 MMscf/d and flowing pressure of 1500 psi the
propped fracture width (kfw). It is typically reported in velocity through the fracture would be 0.02 m/s, =
millidarcy-feet (md-ft) and is a key design parameter. 90 kg/m3 and = 0.015 cp (0.015 x 10-3 Pa.s). In this
Note that the units of permeability are length squared and case, the Reynolds Number is about 1.8, clearly in the
fracture width is unit length. Thus, fracture conductivity turbulent, non-Darcy regime.
can be thought of as the volumetric capacity of the At turbulent rates, the pressure gradients
fracture to transmit reservoir fluids. become proportional to the square of the velocity as
The above definition of fracture conductivity is, represented by Forcheimers equation (Forcheimer,
in fact, only a measure of the viscous drag effects of 1901) (see Section 4-7.2.1):
the producing hydrocarbon. Darcy (1856) showed
that the pressure drop through porous media is = v + v 2 .
p (4-116)
proportional to the fluid velocity: L k
Forcheimers equation states that the pressure
p v
= , (8-1) gradient is the sum of the viscous forces (v/k) and
L k
the inertial forces (v2). At low velocities, where
where p/L is the pressure drop per length of proppant inertial forces are small, Forcheimers equation
pack, is the fluid viscosity, v is the superficial fluid reduces to Darcys Law.
velocity and k is the permeability.
Equation 8-1 works for laminar flow, i.e., for 50
Pressure Drop per foot of frac length, psi/ft
Inertial Forces (Forcheimer Correction)

Reynolds Number less than 1 as measured in the porous Viscous Forces (Darcy's Law)
40
medium. The Reynolds Number is given by: Conditions:
50' Fracture Height
2#/ft2 ->0.15" after embedment
k f ,n (2-37)
30 2000 psi BHFP, 200 F
N Re = , 20/40 Ottawa Sand
2000 psi closure stress
20 Reference Permeability =
where kf,n is the nominal permeability (under Darcy 243 Darcy 55% gel damage
flow conditions) in m2, is the coefficient of 10
inertial flow resistance in m-1, v is the fluid velocity
at reservoir conditions in m/s, is the viscosity of 0
0 1 2 3 4 5
the fluid at reservoir conditions in Pa.s, and is the Well Flow rate, MMscf/d
density of the flowing fluid in kg/m3. The value of
is obtained from: Figure 8-1 Viscous and non-Darcy pressure drops in a
fractured gas well (after Vincent et al., 1999)
b
= (1108) , (2-38)
(k f ,n )a
For the fracturing of gas wells it is imperative to
where a and b are obtained from Cooke (1973). Some include the inertial flow effects and to define an effective
values for are given in Table 2-3. fracture conductivity. Using the above equations, the
For kf,n = 220,000 md, a = 1.54 and b = 110,470 pressure drop per foot of fracture length for a single-
(for 20/40 mesh proppant) is equal to 65,400. For phase fluid can be calculated as a function of velocity
example, at 1000 STB/d oil flow, the reservoir flow rate in the fracture or production rate through a given
would be 1200 resbbl per day (assuming Bo = 1.2 resbbl/ cross-sectional area of fracture. Figure 8-1 shows the
285
Modern Fracturing
pressure drop through a bi-wing fracture with a 50-ft Figure 8-3 provides an example of the necessity of
fracture height when propped with 2 lb/ft2 of 20/40 defining fracture conductivity under conditions of both
Jordan sand with 2000 psi of closure stress on the inertial and multiphase flow if they exist in the fracture.
proppant. It is apparent that Darcys Law is not valid Even with a fairly modest well (1 MMscf/d from a bi-
for calculation of the pressure drop within the fracture wing 50-ft fracture height), the inertial flow effects
at even moderate rates below 500 Mscf/d. Examination reduce the effective conductivity of the proppant pack
of Fig. 8-1 shows that the inertial pressure drops are by approximately 70%. Addition of 10 bbl/d of water
much more significant than the viscous losses and that to the 1 MMscf/d causes an incremental reduction of
the beta factor of the proppant is more important to approximately 25% of the reference conductivity.
ultimate productivity in a gas well than the laminar
4002
permeability (Vincent et al., 1999).
4000
In a gas well, there is typically a liquid phase
Jordan Sand
associated with condensate or water production. Resin Coated Sand
Several researchers have attempted to quantify the 3000 Light Weight Ceramic
Conductivity (md-ft)
effects of multiphase flow in fractures with laboratory
2032
work or by analyzing production data. While they 2000
1354
report differing results regarding the absolute effect
on conductivity, all conclude that the multiphase
1000
646
effects are substantial. The effect of multiphase and
467
246
non-Darcy flow can be reported in several ways. If
152
112
27
the pressure drop across the proppant pack increases 0
ow
Fl ase
ity
ten-fold, this is synonymous with reporting a ten-
ow
tiv
Fl
cy h
ar ltip
uc
cy
fold reduction in effective conductivity. Because Eq.
d
ar
-D u
ow n
on h M
-D
Fl Co
on
it
4-116 does not have a term for multiphase flow, the
ar e
W
N
in c
m ren
ith
N
La fe
W
incremental pressure drop is either accounted for by
Re
increasing the beta factor directly, or by determining Test Conditions:

20/40 Proppants at 2lb/sq ft, 6500 psi, 225F
a pressure drop multiplier referred to as gamma (). For a 50 ft. bi-wing frac height and 300 psi BHFP:
Reference Conductivity: 65 MSCFD dry gas
Figure 8-2 shows laboratory data of the impact of Non-Darcy velocity: 1 MMSCFD dry has
Multiphase: with 10 bwpd
mobile liquid saturation on total pressure drop in
the fracture where gamma () is defined as the ratio Figure 8-3 Impact of non-Darcy and multiphase flow on
of the actual pressure drop divided by the pressure measured fracture conductivity (after Penny and Jin, 1995)
drop for single phase flow.
8.2 Conductivity Impact on Fractured
100
Well Production Potential
A variety of methods are available to analyze formation
and fracture characteristics from production data. It
Gamma
10 is important to recognize that these techniques can

be used to both predict and evaluate production
performance if the reservoir and fracture characteristics
are specified as shown in Chapters 2 and 4.
1 The productivity improvement as a result of
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Increasing flow of mobile liquid phase
fracturing a well will depend on the initial condition
SVN
liquid SVN
gas + iquid of the formationi.e. whether it is damaged prior to
the treatment. The primary mechanism that improves
Figure 8-2 Example pressure drop multiplier for
multiphase flow (courtesy of Stim-Lab Inc.) production from a fractured well is the change in flow
286
pattern within the reservoir. In a natural completion permeability formations that are being frac-packed.
or matrix-acidized well, there is a radial flow pattern. The steady-state solution assumes a constant pressure
In flow from a fractured well there is a large portion boundary, while the pseudo steady-state solution
of production that will be channeled through the assumes a closed reservoir with no pressure support
fracture, particularly in the presence of any near- such that after the pressure drawdown has reached the
wellbore formation damage. reservoir boundary, the future rate of change of pressure
After a fractured gas well is placed on production, a is constant for a given production rate. In very low-
pressure drawdown moves down the fracture away from permeability formations (e.g. tight gas sands) the well
the well. Four different flow periods will result over time: may not reach pseudo steady-state conditions. In these
linear flow in the fracture, bi-linear flow in the fracture cases, the transient solution is used because it considers
and formation, linear flow in the formation and, finally, the transient response for all time. These issues are
pseudo-radial flow into the fracture. These different flow covered extensively in Chapter 2.
patterns are shown schematically in Fig. 8-4.
8-2.1 How a Propped Fracture
Well Fracture
Benefits Well Flow Rate
Fracture
If a well is damaged, a positive skin factor will be

a calculated from a pressure buildup test. In the case
Fracture Linear Flow c of high-permeability formations, positive skin factors
Formation Linear Flow in the range of +10 to +30 often result from gravel
pack completion practices. The production benefit of
Well bypassing this damage (s = 0) or from stimulating the
Fracture
well (s < 0) can be calculated readily using the equations
Fracture
given in Section 2-3.3. Post-fracture skin factors of 3
Well to 6 are common in hard rock completions.
The potential productivity benefit of the fracture
b is controlled by both the capacity of the formation to
d
Bilinear Flow Pseudo-Radial Flow deliver fluids to the fracture and the capacity of the
Figure 8-4 Flow periods for a fractured well (after Cinco fracture to produce the fluids into the wellbore. Many
et al., 1978) slightly different terms are used to define the relative
capacity. The most common term is referred to as the
Modeling of productivity improvement can take a dimensionless fracture conductivity (CfD) and was
variety of forms. If warranted, a fully compositional, defined in Section 4-5.1 as:
3-D reservoir simulator can be used to investigate the
k w
f
production benefit from fracturing. Typically, a reservoir C fD =
, (4-108)
kx f
simulator is used for field development planning, but it
is not practical for design of individual well completions. where k is the formation permeability in md, kf is the
Analytical solutions to the radial diffusivity equation fracture permeability in md, w is the propped fracture
are often used to estimate production benefits from width in ft, and xf is the fracture half-length in ft.
changes in completion practices. Based on the discussion in Section 8-1.3, it
Two approaches are used to calculate production is important to realize that the literature defines
rates: the steady- or pseudo steady-state condition, and fracture conductivity by the term (kfw). Since kf as
the transient flow solutions. These solutions are not measured under laboratory test conditions is a reference
alternatives, so much as different techniques applicable permeability, this should be restated in terms of effective
under different flow conditions. The steady- or pseudo fracture permeability including both non-Darcy and
steady-state solution is often more applicable to high- multiphase flow effects (kM).
287
Modern Fracturing
In an ideal completion, fracture length and 1.0
conductivity should be balanced to maximize

f
g xf
production as explained in Section 4-5. A re asin
k f wf Inc
design value of CfD will be determined in order r a 0.25
k ng
k fw
reasi
to provide a physical optimization of fracture 0.1
Inc
performance and an economically engineered balance

Steady-State Folds
between fracture flow capacity and potential reservoir of Increase
deliverability into the fracture. The optimum value

of CfD to physically maximize production for most
0.01
fracturing treatments is 1.6 (for medium- to high- 0.1 0.1 k f wf 10 100
CfD
permeability formations) but can be larger in lower- kxf
permeability reservoirs as shown explicitly in Section Figure 8-6 Producing rate folds of increase plot (after
4-5. The actual value for each completion will be Prats, 1961)
specific to the relative values of the formation and
effective proppant permeabilities. Figure 8-6 shows that for a low value of CfD
Tannich and Nierode (1986) presented the impacts < 1 the effective wellbore radius (rw) may be
of the relative capacity of the fracture and penetration approximated by:
for gas wells. As shown in Fig. 8-5 the productivity
k w
index ratio is provided as a function of the relative non- rw = 0.25 f . (8-3)
k
Darcy flow conductivity.
And for the case of infinite fracture conductivity (i.e.
10
CfD ), such that there is no significant pressure
9
drop in the fracture, the effective wellbore radius may
8 be approximated by:
7 0.8
rw = 0.5 x f .
(8-4)
Productivity index Ratio
6
0.7
5
The same concept is demonstrated where production
0.5
4 rates are governed by fracture conductivity at
3 0.3 low values of CfD and by reservoir penetration at
2
high values of CfD.
0.1
1
8-2.3 Transient Solutions
0
0.1 1.0 10 100 1000 10000
Models developed for pressure transient analysis can be
Relative capacity-RC = RCF
used to predict the production rate as a function of
Figure 8-5 Productivity increase plot (after Tannich and
time. Three basic models are used to analyze fractured
Nierode, 1986)
wells: the uniform flux, infinite conductivity, and
(the most appropriate) finite conductivity models.
8-2.2 Steady-State Solutions Figures 8-7 and 8-8 show example solutions for the
constant pressure and constant rate cases for the finite
Although not commonly applicable for gas wells, if there conductivity vertical fracture model (Cinco et al., 1978.)
is a case where there is pressure support for the reservoir, Note that the dimensionless fracture conductivity
a steady-state solution is more applicable. Prats (1961) term is defined a little differently:
defined the concept of an equivalent wellbore radius of k w
f f
a fractured well as shown in Fig. 8-6. Cr =
. (8-5)
kX f
288
10 a fracture designed for optimal late-time performance

may not be economically optimal, especially when the
1
economic discount factor is high or there is a need to
kh( pi p wf )
141.3qNB
.2 recoup development costs in a shorter pay-out time.

Cr = 0
1 The consequences are that treatment economics often

2
10-1 Type Curves require obtaining higher values of effective CfD.
qD
1
10 Finite Capacity Vertical

100 Fractures Constant
10-2
10
00 Pressure Case
8-3 Proppants
10-5 10-4 10-3 10-2 10-1 1
.000264kt
tD
GNCx f 2
The two main categories of proppants used today are
naturally occurring sands and man-made ceramic or
Figure 8-7 Constant-pressure, finite-capacity vertical
fracture type curve (after Cinco et al., 1978) bauxite proppants. Sands are used for applications where
10
the closure stress is less than 6,000 psi, most commonly
wk f
occurring at depths of less than 8,000 ft. Man-made
Cr
Q xf k ceramic proppants are used for deeper fracturing
1
applications in which the closure stress is greater than
kh( pi pwf )
141.3qN B
0.2
Cr = 6,000 psi. Man-made proppants may also be employed
1
in applications where high proppant pack permeability
pwfD
10-1 2 Type Curves

10 Finite Capacity Vertical is required for optimum fracture conductivity. The effect
Fractures Constant
100
00 Rate Case of closure stress on the fracture conductivity provided by
10-2 10
10-5 10 -4
10-3 10-2 10-1 1 various proppant types is shown in Fig. 8-9.
.000264kt
tD
GNCx f 2 Conductivity vs. Closure Stress for
Various 20/40 Proppant Types
Figure 8-8 Constant-rate, finite-capacity vertical fracture 100000
Ottawa sand LW Ceramic
Proppant Pack Conductivity, mD-ft
type curve (after Cinco et al., 1978)

Brady sand ISP
Pre-cured RC Sand Sintered Bauxite
At dimensionless time values above 102, the Cr 10000
= 10 curve approaches the 100 and 1000 cases

displaying infinite conductivity. 1000
An important contribution to the understanding

of fracture optimization was made using the concept 100
of the Proppant Number to determine the optimum

dimensionless conductivity for placement of a fixed 10
0 2000 4000 6000 8000 10000 12000 14000
volume of proppant (Economides et al., 1998. Also see Closure Stress, psi
Sections 2-5 and 4-5). In this case, where the fracture
volume is pre-determined, for small Proppant Numbers Figure 8-9 Effect of Closure Stress on Fracture
Conductivity for Various Proppant Types (from
the optimal value of CfD was found to be 1.6 to maximize Predict K, 2007)
production under pseudo-steady state conditions.
Economides et al. recognized that the indicated 8-3.1 Sands
optimal fracture dimensions may not be technically
or economically feasible. In practice, the engineer is Ottawa and Brady sands are by far the predominant
faced with additional costs of greater fluid volumes fracturing sands used in the oil and gas industry. These
and more out-of-zone growth when achieving greater are commonly called brown and white based upon
lateral extent of the fracture. The Proppant Number the sands primary color. However, color is not the
in that case is based on the proppant in the pay. In only difference in these proppants. Depending on
cases of very low-permeability reservoirs, with the well their overall balance of physical properties, sands
operating under transient conditions for a long time, can be subdivided into groups of excellent, good
289
Modern Fracturing
and substandard grades (API RP 56, 1983; and ISO Ottawa sands are considered to be the highest-
13503-2, 2006). The premium sands, which come quality fracturing sands and are the most widely
mainly from the north central United States, are employed proppants in the industry. They were formed
commonly known as Ottawa sand. Brady sand, from quartz grains freed by the erosion of granites in the
which is classified as a standard grade of sand, meets Canadian shield. The eroded sand grains accumulated
or exceeds the industry-accepted specifications for on the beaches of an ancient Cambrian Sea, where they
hydraulic fracturing proppant. These sands are used were repeatedly washed, sifted, sorted, and resorted by
widely in fracturing applications throughout the tides and winds over millions of years, before they were
world. Although less widely distributed, additional covered and protected from erosion and contamination.
sources of API-quality fracturing sands are located in This repeated washing and sorting yielded deposits that
England, Brazil, Saudi Arabia, and Russia. are particularly suited for use as fracturing sands.
Several other sands from geographical locations Ottawa sands are monocrystalline (have a single
such as Australia, Peoples Republic of China, Poland crystal phase), which results in superior individual grain
and Oman, are available in the marketplace. These strength properties compared to other sands (Fig. 8-
materials most typically fail to meet one or more of the 11). They are characterized by high purity, whiteness or
industry-accepted specifications and thus, are generally clarity, a high degree of roundness and sphericity, few
classified as being substandard. Nevertheless, such fines, and low solubility in acid. Ottawa sand formations
sands are being used as proppants in areas where they generally yield a higher percentage of smaller mesh sizes,
have been deemed fit-for-purpose with respect to the but are readily available in sizes from 12 to 70 mesh.
fracture stimulation of a particular reservoir. Both Ottawa and Brady sands exhibit particle specific
gravities of approximately 2.65.
8-3.1.1 Ottawa Sands
Most Ottawa sand is mined from geological formations

in the Midwest region of the United States. White,
Jordan, and Northern sands are other names
used to identify Ottawa sand. The sands are mined
primarily from three main sand deposits: the St. Peter,
the Jordan and the Wonewoc. Because the largest
concentration of mining is in the Ottawa District of
northern Illinois, the term Ottawa Sand is commonly
used to describe these sands. An Ottawa sand plant in
northwest Wisconsin is shown in Fig. 8-10.
Figure 8-11 Ottawa sand grains
8-3.1.2 Brady Sands
Brady sand is darker than Ottawa and consequently is

often referred to as brown sand (Fig. 8-12). Brady
sand is generally a less costly proppant and is widely
used, primarily in North America. Brady sand is mined
Figure 8-10 Ottawa sand manufacturing plant in
from the Hickory formation outcrops near Brady,
northwest Wisconsin (courtesy of Badger Mining Co.) Texas. These sands are polycrystalline: Each grain is
290
composed of multiple crystals bound together. Brady 8-3.2.1 Sintered Bauxite

sand is more angular and contains more impurities
(feldspars) than Ottawa sand but is thoroughly Sintered bauxite was the original ceramic-based high-
cleaned and processed to yield a high-quality product. strength propping agent (Fig. 8-13). Sintered bauxite
Impurities contribute to the brown color common proppants contain corundum, one of the hardest materials
to Brady sands. The existence of cleavage planes known, and offer the greatest proppant strength available
within each grain results in greater crush and reduced for deep wells with hostile, high-stress and -temperature
strength properties. The Hickory formation contains environments. The various commercial sintered bauxite
a variety of sand grain sizes, which are typically larger products have specific gravities of 3.4 or greater. Due to
than Ottawa sand proppants. Brady sand is commonly the manufacturing processes employed to make sintered
available in 8/12 to 20/40 mesh sizes. bauxite, the particles offer excellent roundness and
sphericity. Because of their relatively high cost, sintered
bauxite proppants are generally limited to wells with very
high closure stresses, typically greater than 10,000 psi.
Due to the high density of sintered bauxite proppants,
smaller grain sizes are typically employed to improve
transport capabilities. Sintered bauxite proppants are
available in sizes ranging from 12 to 70 mesh.
Figure 8-12 Brady sand grains
8-3.2 Ceramic Proppants
The exploitation of deep gas reservoirs beginning in

the 1970s necessitated the development of fracturing
proppants having greater strength. In response to this Figure 8-13 Sintered bauxite (courtesy of Sintex
need, Exxon Production Research introduced the Proppants)
first generation of ceramic proppants, which were
manufactured by sintering pellets of aluminosilicate 8-3.2.2 Intermediate Strength Ceramic Proppant
clay containing over 80% Al2O3, commonly known
as bauxite. First commercial production of bauxite Intermediate-strength proppants (ISP) are fused ceramic
proppant occurred in 1979 for the exclusive use of proppants which were first introduced in 1982. ISPs
fracturing deep gas wells. This was followed in 1982 have a specific gravity between 2.9 and 3.3. The variance
by the introduction of Intermediate Strength Proppant in specific gravity is due to the raw material sources
(ISP), which is manufactured from ore with 70% Al2O3. used by the various manufacturers of the proppant.
Both of these products are manufactured from high- Intermediate-strength proppants have lower strength
alumina ores, which are pelletized, dried and fired in a limits than sintered bauxites and are primarily used in
kiln to produce a strong crystaline structure. Ceramic applications having closure stress between 8,000 and
proppant products are most suitable for fracturing 12,000 psi. Like sintered bauxite, intermediate strength
applications at greater than 10,000 ft depth. proppants typically exhibit very good roundness and
291
Modern Fracturing
sphericity (Fig. 8-14). Intermediate strength proppants For conventional applications, LWC proppants
are commercially available for fracturing applications in are targeted for applications with closure stress from
sizes ranging from 16 to 70 mesh. 6,000 to 10,000 psi, but they eventually found
application in shallow, low closure stress formations
where production could be improved by increasing
fracture conductivity beyond that provided by sand
or resin-coated sand proppants. More recently,
LWC proppants have found application not only in
applications with stress environments beyond the
capabilities of sand-based proppants, but also in
applications where the highest possible conductivity
is needed, including fracturing high-permeability
formations (i.e. frac pack applications) and wells
exhibiting non-Darcy and/or multiphase flow
characteristics within the proppant pack.
Lightweight ceramic proppants are commercially
Figure 8-14 ISP proppant (courtesy of Saint Gobain) available for fracturing applications in sizes ranging
from 12 to 70 mesh.
8-3.2.3 Lightweight Ceramic Proppant
8-3.3 Resin-Coated Proppants
Lightweight ceramic (LWC) proppants have a large
mullite component, which is a hard alumina-silicate Problems associated with the brittle failure of sands
material. The first lightweight ceramic proppant was led to the development of resin-coated proppants
introduced in 1985. While not quite as strong as bauxite in which resin coatings are used to encapsulate each
or ISP proppant, the specific gravity of LWC proppant grain, primarily for purposes of improving grain
is approximately 2.72, thus more similar to sand than strength. The resin coating is usually at least partially
the bauxite and ISP proppants. The manufacturing cured during the manufacturing process to produce
process for LWC proppants yields particles with a non-melting, chemically inert surface film. Both
improved strength, greater sphericity, and more closely curable and fully pre-cured resin coatings are available
sized sieve distributions, all of which contribute to for most proppant types (sand to bauxite) and are
provide much higher conductivity than comparably- widely used in fracturing treatments.
sized high-quality sand (Fig. 8-15). The resin improves the proppant properties by
reducing the grain angularity. In addition, it improves
the amount of particle crush by distributing the stress
load over a larger area of the proppant grain, reducing the
point loading. Furthermore, when the proppant grains
are crushed, the resin coating serves to encapsulate the
fines and fragmented crushed portions, preventing the
small particles from migrating and plugging proppant
pack pore throats and flow channels.
Resin-coated proppants typically have higher
conductivity at elevated confining stresses than
uncoated proppant. Pre-cured proppants are those
in which the resin coating is fully cured by the
manufacturer. By contrast, the coatings of curable
Figure 8-15 LWC proppant (courtesy of Carbo
proppants are only partially cured when manufactured,
Ceramics Inc.)
292
with the intent of additional curing occurring after

downhole placement so that the coatings of discrete
proppant particles will bond together.
The primary application of pre-cured resin coated
proppant is to enhance the performance of the core
proppant at higher stress levels. The resin coating
also encapsulates the proppant grains, preventing the
migration of created fines during production. Pre-
cured resin-coatings (Fig. 8-16) have been shown
to be tolerant to formation brines and crude oil at
temperatures exceeding 300 oF.
Figure 8-17 Curable resin-coated proppant (courtesy of
Santrol)
Figure 8-16 Pre-cured resin-coated proppant (courtesy

of Santrol)
Curable resin-coated proppants are typically used

in producing zones that otherwise tend to flow back
proppant during production operations. The curable
coating bonds the proppant grains together after they
have been placed in the formation to form a consolidated
Figures 8-18 and 8-19 Curable resin-coated sand that
pack, which resists proppant flowback and the associated has been cured at elevated temperature and stress to
problems with productivity loss and damage to surface provide grain-to-grain bonding and consolidation of the
equipment. Rather than crush resistance, compressive proppant pack (courtesy of Santrol (8-18) and Hexion
Oilfield Proppants (8-19))
or tensile strength of the consolidated proppant pack is
often used as the critical property to describe resin-coated In recent years, several new types of resin-
proppant. Curable resin-coated proppants are typically coated proppants have been introduced to address
pumped at the end of a treatment, often referred to as a specific application needs beyond the capabilities
resin-coated tail-in stage (Fig. 8-17). After the treatment, of the previous generations. These include curable
the well is typically shut in for a time to allow the resin proppants providing greater tensile strength for
to bind proppant particles and cure into a consolidated, application in extreme stress environments and to
permeable pack (Figs. 8-18 and 8-19). better mitigate proppant flowback; proppants having
293
Modern Fracturing
improved chemical compatibility with fracturing

fluids; proppants tolerant of extreme temperature
environments; proppants that cure only in the
presence of confining stress to facilitate easier post-
frac wellbore cleanout; and proppants that cure at
very low reservoir temperatures.
8-3.4 Ultra-Lightweight Proppants
The commonly used sand and ceramic proppants

are each employed for their ability to withstand the
respective fracture closure stress environment. As the
relative particle strength of these various proppant Figure 8-20 Ultra-lightweight proppant (prop = 1.25) made
materials increase, so too does the respective particle from resin-impregnated and -coated nut hull (courtesy BJ
density or specific gravity. For example, sand has a Services Company)
specific gravity of 2.65, compared to a specific gravity
of up to 3.65 for sintered bauxite.
The settling rate of proppant in a fracturing fluid
is strongly influenced by the specific gravity of the
proppant. Bauxite, the strongest proppant available,
has the highest settling rate of common proppants,
even in a crosslinked fluid. Proppant settling rate is
typically assessed by comparing single-particle static
settling velocity in a column of water. Using this
metric, it can be determined that for equal-sized 20
mesh particles, bauxite falls at 23.2 ft/min and sand
at 16.6 ft/min. As the specific gravity of the proppant
approaches that of the fluid, a nearly neutrally Figure 8-21 Ultra-lightweight proppant (prop = 2.02) made
buoyant condition is approached and proppant from resin-coated porous ceramic particles (courtesy of
BJ Services Company)
settling velocity nears zero.
A new class of proppants commercially
introduced in 2004 has both low specific gravity and
the requisite mechanical properties to function as
a fracturing proppant at reservoir temperature and
stress conditions (Brannon et al., 2002; Rickards et
al., 2003; Wood et al., 2003). The new materials were
classified as ultra-lightweight proppants and defined
as exhibiting particle densities substantially lower
than sand-based proppants. The first generation of
ultra-lightweight (ULW) proppants, comprised of
resin-impregnated and coated walnut hulls, exhibit
a specific gravity of 1.25, which is less than half
that of sand (2.65). Subsequent development has
yielded additional ultra-lightweight proppants
Figure 8-22 Ultra-lightweight proppant (prop = 1.50)
with specific gravities (prop) of 2.02, 1.50, and made from plastic composite (courtesy of BJ Services
1.054 (Figs. 8-20 through 8-23). Company)
294
Long-standing fracture theory holds that placement

of partial monolayers of proppant offers high fracture
conductivity using a much reduced volume of proppant
than required for conventional multilayer proppant packs
(Darin and Huitt, 1959). Fractures containing proppant
partial monolayers possess similar fracture geometry but
with vacant areas around and between the proppant
particles, thereby increasing the relative conductivity of
the propped fracture. Designs using the ULW proppants
have been quite successfully employed to provide well
performance consistent with production from fractures
containing partial monolayers of proppant.
8-4 Proppant Properties, Testing

Protocols, and Performance
Considerations
8-4.1 Proppant Testing Procedure Standards

Figure 8-23 Ultra-lightweight proppant (prop = 1.50)
made from a thermoplastic nanocomposite (courtesy of
BJ Services Company) The American Petroleum Institutes Recommended
Practices for proppants used in hydraulic fracturing
The ultra-lightweight (prop = 1.25) proppant was and gravel packing lapsed (or expired) in the
shown to provide sufficient strength to be applied in late 1990s due to API policy that Standards or
reservoirs with closure pressures up to 5,000 psi and Recommended Practice documents must be reviewed,
BHSTs exceeding 200 F. The lower specific gravity revised, reaffirmed or withdrawn at least every five
of the ultra-lightweight (prop =1.25) proppant is also years. In 1998 the International Organization for
manifested as a much greater pack volume per unit Standardization (ISO), a network of national standards
of mass (lower bulk density), resulting in one pound institutes of 153 countries, formed a committee
of the proppant occupying an equivalent volume of to review and rewrite the APIs Recommended
greater than that of two pounds of similarly sized sand. Practices as ISO Standards. The committee elected
For this reason, comparisons of proppant performance to combine the three API Recommended Practices
should always be done on a volume basis. for proppant and gravel testing (API RP 56, 58, and
The proppant transportability provided by the 60) into a single standard. The ISO published ISO
ultra-lightweight proppants allows for fracturing 13503-2 Measurement of Properties of Proppants
fluids of lesser viscosity to be used to accomplish Used in Hydraulic Fracturing and Gravel-Packing
effective proppant placement. Ultra-lightweight Operations in 2006. (API subsequently adopted
proppants can be effectively placed using simple and published the ISO 13503-2 Standard as API
slickwater fluids (Brannon et al., 2005). The prop Recommended Practice 19C).
= 1.25 proppants have been employed widely with What does this new standard mean for the
brine-based slickwater fluids to yield a nearly neutrally industry? Laboratory data reported from these
buoyant proppant slurry, efficiently minimizing procedures shall reference these standards, and the
proppant settling within the created fracture. The industry shall recognize the procedure in which
minimized proppant settling afforded by the ultra- this data was obtained. The goal of such efforts is
lightweight proppants enables effective placement of to have all laboratories producing nearly identical
proppant partial monolayers. results for the same proppant.
295
Modern Fracturing
8-4.2 Proppant Sampling A box-type bulk sampling device is recommended for

sample collection (Fig. 8-24). The device has dimensions
Sampling and splitting of bulk material are critical steps of 8 6 4 in., with a in. slot, and is to be swept back and
to insure that any testing is carried out on a representative forth through the flowing proppant stream. Alternatively,
sample of the proppant. It is important to have a basic samples may be obtained by automatic sampling devices
understanding of segregation when sampling proppant. at the manufacturing plant or distribution center. These
Depending on the size, shape, distribution and devices conduct an intermittent but regular sweep of
mechanisms involved, there is usually a certain amount the proppant stream and deflect proppant into a sample
of variability involved in sampling due to segregation. bottle. When it is absolutely necessary to sample a bulk
Particles, such as proppants, will naturally find the container such as a bulk bag or truck, a sample probe
path of least resistance when moved or force is applied. should be pushed into the proppant approximately 3 ft
During transfer or movement, particles of differing size and then opened to obtain the sample.
and mass naturally separate or segregate. Fine particles Once a representative sample has been obtained
migrate and usually rest close to the area where they land. the sample must be split to insure accurate testing of
Heavier, coarser particles bounce or roll much farther, properties such as bulk density, specific gravity, crush
stratifying the material by size. The degree of segregation resistance and sieve sizing. Each of these properties can be
or stratification depends on the mechanisms involved in affected by variation in the sieve distribution. The sample
the proppant transfer. The sampling procedures defined should be split to obtain the required sample size for all
by ISO 13503-2 were developed to minimize the effects the tests that are to be conducted using a sample splitting
of segregation of particles by size. device (Fig. 8-25). For example, if it is only necessary to
The standard recommends that sampling be done conduct a sieve analysis, then the sample should be split
at the source-of-supply by sweeping the entire delivery to as close to 100 g as possible and adjusted to reach 100
stream rather than trying to sample from a bag, loaded silo g. If the material is sacked between 50 and 100 pounds,
or hopper. For example, during the transfer of proppant then the entire sack of material should be reduced using
at the well site, it is recommended to sample the proppant a sample reducer and then further split using the sample
stream every 20,000 pounds of material. The standards do splitter. These steps help ensure capture of representative
not recommend sampling from containers after they have samples so that erroneous results are not obtained by
been filled because proppant has a tendency to segregate, faulty sampling and splitting procedures.
and obtaining a representative sample is impossible.
(See also Section 9-5 for field procedures for proppant
quality control.)
Figure 8-24 Box-type bulk sampling device (courtesy of Figure 8-25 Proppant sample splitter (courtesy of Badger
Badger Mining Co.) Mining Co.)
296
8-4.3 Grain Size and Grain Size Distribution Fines invasion into the proppant pack can
result in partial plugging and rapid reduction in
Proppant mesh size has long been recognized permeability. Smaller proppants resist the invasion
as a key characteristic for describing and of fines, making them more suitable than larger
controlling the quality of a propping material. The ones in formations subject to fines production.
proppant size and its distribution directly affect Although the relatively smaller proppants offer less
the permeability of the proppant pack after its conductivity initially, the average conductivity over
placement in a hydraulic fracturing treatment. the life of the well will be higher and will more
Proppant grain size is characterized by the median then offset the initial high productivity provided by
diameter of the discrete grains, and grain size larger proppants, which is often followed by rapid
distribution relates to the breadth of the range production decline. This phenomenon is illustrated
of grain sizes exhibited within a given proppant. in Fig. 8-27 for a LW ceramic proppant, showing that
Proppants with larger grain sizes provide a more the larger sizes exhibit higher conductivity at lower
permeable pack. However, as grain size increases, stresses, but at higher stresses, the smaller 20/40
strength decreases. Thus larger grain sizes can be proppant provides greater conductivity.
more difficult to use in deeper wells because of greater
Conductivity vs. Stress
susceptibility to crushing. Figure 8-26 illustrates the Various LW Ceramic Proppant Sizes
effect of grain size on permeability of higher quality 100000
sand at increasing closure stresses. Conductivity, mD-ft
Proppant Pack
10000
Conductivity vs. Stress for Various
Ottawa Sand Proppant Sizes
100000 1000
Conductivity, mD-ft
10000
Proppant Pack
100
0 2000 4000 6000 8000 10000 12000 14000
1000
Closure Stress, psi
100 8/12 12/18 16/20 20/40 30/50
Figure 8-27 Effect of LWC proppant size on conductivity

10
0 2000 4000 6000 8000 10000 vs stress (from Predict K, 2007)
Closure Stress, psi
8/16 12/20 16/30 20/40 30/50 40/70 70/140 8-4.3.1 Proppant Size Testing
Figure 8-26 Effect of proppant size on the permeability of
fracturing sands (from Predict K, 2007) The purpose of this procedure is to ensure a consistent
methodology for sieve analysis and to provide a
Large proppant also requires a larger fracture width consistent procedure for sieve evaluation. Sieve testing
to enable transport without bridging out during for fracturing proppants is required to be carried out
the treatment. A bridge out condition occurs when on a Ro-Tap sieve shaker or equivalent (Fig. 8-
proppant is prevented from being transported further 28). This equipment differs from sieve stacks used
down the fracture due to the width limitation of the in soil mechanics or civil engineering applications
fracture. This is typically thought to occur when the in that it incorporates a hammer at the top of the
maximum proppant particle size is greater than one- stack. Often when proppant is analyzed on location,
third of the created fracture width (Gidley et al., sieve stacks lacking a hammer mechanism are used,
1989). As per Stokes Law, the settling rate of proppant which result in erroneously coarse distributions. The
increases with increasing size, proportionally to the repeated hammering action in the Ro-Tap stack during
square of the median diameter,; therefore, the use of sieving jars the stack and prevents proppant from
larger proppant sizes has a greater tendency to suffer shaking across individual screens. Standard US sieves
from proppant placement difficulties. are shown in Figs. 8-29 and 8-30.
297
Modern Fracturing
minutes. Table 8-1 shows the sieve sizes that should

be employed for testing various proppant sizes. The
new industry standard suggests using the following
limitations on sieve distribution for proppants
suitable for use in hydraulic fracturing:
- at least 90% of the material must fall between the
two mesh sizes
i.e. for 20/40, 90% must past though the 20
mesh and be retained on the 40 mesh
- no more than 10% of the material may be coarser
than the largest mesh size
i.e. for 20/40, < 10% can fall between the 16
Figure 8-28 Ro-Tap Sieve Shaker with sound enclosure
(courtesy of Badger Mining Co.) mesh and 20 mesh screens
- no more than 0.1% of the material may be coarser
than the next largest mesh size
i.e. for 20/40, no more than 0.1% can
exceed 16 mesh
- no more than 1% of material is permitted to fall
onto the pan
i.e. for 20/40, no more than 1% can be
smaller than 50 mesh.
8-4.4 Proppant Shape
Roundness and sphericity are proppant particle

properties that can affect performance by impacting
Figure 8-29 Standard US sieve stack (courtesy of Badger the proppant packing. Sphericity is a measure of how
Mining Co.)
close a proppant particle approaches the shape of a
sphere. Roundness is a measure of the relative sharpness
of corners or of curvature. Their importance depends
somewhat on the stress level at which the proppant is to
be used. Because the surface stresses are more uniform
on well-rounded, spherical particles, they are capable of
carrying higher loadings than a less-rounded particle.
Therefore, at increasing stress levels, a higher degree of
roundness and sphericity is preferred for better load
distribution, fewer crushed fines generated, and higher
proppant pack conductivity.
Grain robustness, or angularity, is a measure of
the relative sharpness of the grain corners, and the
Figure 8-30 US sieves (courtesy of Badger Mining Co.) particle sphericity is a measure of how closely the
shape approaches the shape of a sphere. Improved
The new ISO standard requires that the sieve roundness and sphericity will enable greater
stack comprise at least seven calibrated screens, that porosity and permeability than a pack comprised
a 100 g (20 g) split sample be used for testing, and of angular proppant after stress is applied and
that the sieve stack be shaken for a minimum of 10 grain rotation occurs.
298
8-4.4.1 Proppant Shape Testing 8-4.5 Proppant Bulk Density and

Apparent Specific Gravity
The standard method of determining roundness
and sphericity is the use of the Krumbien and Proppant density has a strong influence on proppant
Sloss chart (Krumbein and Schloss, 1963), as transport and placement. High-density proppants are
shown in Fig. 8-31. ISO 13503-2 requires the use more difficult to suspend in the fracturing fluid and
of a microscope to assess roundness and sphericity of to transport in the fracture. Placement of relatively
a minimum of 20 particles, and those values to be high-density proppants can be achieved by using high-
averaged. Roundness and sphericity measurements viscosity fluids, which carry the proppant for the entire
must be determined separately. The ISO13503- length of the fracture with minimal settling, and/or by
2 standard suggests that sand proppants have using a higher flow velocity.
roundness and sphericity of > 0.7, while ceramic The Bulk Density (BD) describes the mass of
proppants should be about 0.9. proppant that fills a unit volume, including both
the proppant and the void volume. This figure is
used to determine the mass of a proppant needed
0.9
to fill a fracture. The Apparent Specific Gravity
(ASG) represents the apparent density of a single
pellet. Although it does include any internal
0.7
porosity of the proppant particle, it excludes the
porosity between the grains. Generally, the ASG is
Sphericity
used to estimate proppant settling times, while the

BD is used to estimate transportation volumes and
0.5
actual volume of the fracture. Both properties are
measured without closure stress, so the bulk density
will increase substantially if the proppants crush
0.3
or if pack rearrangement results in loss of porosity.
In addition to being more expensive, stronger
0.1 0.3 0.5 0.7 0.9
conventional proppants are more dense. Table 8-2
Roundness
lists the apparent specific gravity and bulk density of
Figure 8-31 Sphericity and roundness chart (from
Krumbein and Schloss, 1963)
some commonly available proppants.
Table 8-1 Standard Sieve Stack Testing for Various Proppant Sizes (from ISO 13503-2)
(First and second primary sieves in bold)
Typical Proppant
6/12 8/16 12/20 16/20 16/30 20/40 30/50 40/60 40/70 70/140
Designations
Sieve
3350/ 2360/ 1700/ 1180/ 1180/ 850/ 600/
Opening Sizes, 425/ 250 425/ 212 212/ 106
1700 1180 850 850 600 425 300
m
4 6 8 12 12 16 20 30 30 50
6 8 12 16 16 20 30 40 40 70
8 10 14 18 18 25 35 45 45 80
Stack of 10 12 16 20 20 30 40 50 50 100
ASTM
Sieves 12 14 18 25 25 35 45 60 60 120
14 16 20 30 30 40 50 70 70 140
16 20 30 40 40 50 70 100 100 200
Pan Pan Pan Pan Pan Pan Pan Pan Pan Pan
299
Modern Fracturing
Table 8-2 Bulk Density and Apparent Specific Gravity for as invading fines: plugging pore throats within
Common Fracturing Proppants the pack and resulting in conductivity loss.
Apparent The percentage of impurities may also provide
Proppant Bulk Density Specific
Gravity an indication of the proppant strength because
Ottawa Sand 20/40 1.60 2.65 high concentrations of impurities are likely to
Ottawa Sand, 40/70 1.58 2.65 compromise proppant strength. Acid solubility is
Brady Sand, 8/12 1.60 2.65 generally used as an indication of proppant quality
Pre-Cure RC Sand, 20/40 1.59 2.45
in terms of the amount of carbonates, feldspar and
Curable RC Sand, 20/40 1.53 2.55
iron oxides present in the proppant.
LW Ceramic, 20/40 1.57 2.71
ISP, 20/40 1.88 3.27
Very fine particles in proppant can have health
Sintered Bauxite, 20/40 2.00 3.55 and safety implications, such as silicosis caused by
ULWP (prop = 1.25), 14/30 0.84 1.25 long-term inhalation of silicaceous dust. The presence
ULWP (prop = 1.05), 14/40 0.66 1.05 and relative concentrations of such finely divided
materials in a proppant can be determined using a
measurement of the turbidity of a fluid used to wash
Lighter proppants can provide clear benefits a proppant sample. A high turbidity measurement is
due to greater retained fracture widths. Although indicative of either low proppant quality or insufficient
proppant is typically purchased by mass, the benefit washing of the proppant by the supplier.
of a proppant is based on its volume. For example,
100,000 lbs of sand will occupy more volume than 8-4.6.1 Acid Solubility Testing
100,000 lbs of bauxite. Thus, for typical propped
fractures, the density of the proppant will significantly Acid solubility evaluation is used to determine the
impact the fracture width and, for a given proppant suitability of a proppant for use in applications
concentration in the fracture there will be a where proppant may come into contact with
proportionate decrease in propped fracture width for acids. This technique can also be useful to indicate
a higher-density proppant. the amount of soluble materials (i.e. carbonates,
feldspars, iron oxides, clays, etc.) present in the
8-4.5.1 Proppant Bulk Density and proppant. The results of the acid solubility are a
Specific Gravity Testing function of the proppant surface area, since smaller
proppant sizes have more surface area exposed for
Bulk density is measured by pouring the proppant acid reaction and thus will exhibit higher percent
sample through a funnel to fill a calibrated brass solubility than larger particles.
cylinder to a prescribed volume. The proppant The new ISO 13502-2 standard procedure for
added to fill the cylinder is then weighed and that measuring acid solubility is somewhat of a compromise
value divided by the known volume to arrive at of previous API RP procedures for proppant and gravel
the value for bulk density. Measurement of the pack sand. The new standard calls for testing proppant
specific gravity is conducted using a pycnometer, in 12:3 HCl:HF (i.e., 12% by mass of HCl and 3%
wherein a fluid or gas is used to fill the pore volume by mass of HF) over a period of 30 minutes at 150oF.
of a known mass of proppant. The remaining proppant is then separated and dried,
and the weight compared to the original weight prior
8-4.6 Proppant Quality to acid exposure. Sands sized from 6/10 to 30/50 mesh
should have less than 2% solubility.
Grain-size distribution and proppant quality (or In practice, acid solubility should be considered
mineralogical purity) are closely related. A higher when there is expectation of the need for remedial
percentage of smaller grains or impurities can have acid treatments typically used to remove formation
the same effect on proppant pack permeability fines damage or scale. This is one area where natural
300
sand has an advantage over synthetic proppants. reduced, these same stresses tend to close the fracture.
Although ceramic proppants are relatively unaffected If the proppant is not strong enough to withstand
by HCl or acetic acids, HCl/HF mud acid has the closure stress of the fracture, it will be crushed
been shown to attack all proppant types. Ceramic and the permeability of the propped fracture will be
proppants have been shown to lose more mass and drastically reduced. The smaller particles resulting
suffer greater loss of strength than sand proppants from crushing the proppant grains can significantly
(Cheung, 1985; and Roberts et al., 1990). reduce the permeability of the pack.
It is evident that the test conditions specified in The maximum effective stress on a proppant is
the procedures are not universally representative of commonly calculated using the difference between
downhole conditions; higher temperatures and/or the initial fracture gradient (minimum in situ stress)
longer exposure durations will increase dissolution and bottomhole flowing pressure (See Section 13-
of proppants. On the other hand, the tests are also 4.1 for the effects of depletion on effective closure
conducted on clean proppant samples, whereas in stress). The potential for maximum crushing can
an actual acid job, it would be expected that a large occur initially in the production of a well because the
volume of scale or formation fines would be present in fracture gradient is the maximum and decreases with
the proppant pack. These smaller contaminants with reservoir depletion. Usually, the bottomhole flowing
high ratios of surface area to volume are likely to be pressure is held constant and at a low value in order to
preferentially dissolved compared to the larger proppant maximize the production rate. However, if the well is
media. In essence, some percentage of the acid would be initially completed and produced at a high bottomhole
expected to spend on the targeted contaminants, either pressure and with nearly constant production rate, the
reducing the acid strength or the duration of contact maximum effective stress on the proppant is less and
with the proppant. Care should be taken designing any can increase with time; however, it should never exceed
acid treatments, but extra caution will be required if the calculated maximum effective stress.
acidizing a fracture containing ceramic proppant. Figure 8-32 shows a generalized comparison of
stress ranges for application of typical 20/40 mesh
8-4.6.2 Turbidity Testing proppants. Most typically, sand-based proppants are
used for fracturing formations with closure stresses
Turbidity tests measure an optical property of a less than 6,000 psi. In the range of 5,000 to 10,000
suspension that results from the scattering and psi, intermediate-strength proppants should be used.
absorption of light by the particulate matter suspended When the closure stress is in excess of 10,000 psi, the
in the wetting fluid. The higher the turbidity number, use of high-strength proppants is required.
the more suspended particles are present. In most
commercial turbidity meters the incident light beam Proppant Application Ranges
20/40, 2 lb/sqft-Minimum 500-md-ft
is normal to the detection path of the detector (the
Brown Sand
preferred method of measurement). The measured White Sand
percent of transmittance of a wash of the proppant Regular RC Sand
sample is compared to a calibrated formazin solution Premium RC Sand
curve. The results are expressed in Formazin Turbidity LW Ceramic

RC LW Ceramic
Units, or FTU. Standard procedures for turbidity
ISP
measurement are described in ISO 13503-2. Sintered Bauxite
RC Bauxite
8-4.7 Proppant Strength 0
0
0
00
00
00
00
00
00
00
00
00
00
20
40
60
80
10
12
14
16
18
20
Closure Pressure, psi

When a hydraulic fracture is created, the in situ
stresses must be overcome to open and propagate the Figure 8-32 Proppant application range vs. closure
fracture. Once the hydraulic pressure in the fracture is stress (from Predict K, 2007)
301
Modern Fracturing
8-4.7.1 Proppant Crush and Fines Generation It can be misleading to characterize proppants
and their potential application by using such tests to
At failure, most proppants crush to form particulates estimate the percentage crush under the given stress
of some smaller size (Fig. 8-33). The crushed particles conditions at which the proppant will be used. The
that are smaller than 100 mesh are generally considered crush test procedure identifies all material that falls
to be fines. The ability of fines to migrate through through the lower sieve size as crushed material. In the
the proppant pack depends on both the particle size case of a larger proppant material, the crush values are
and the pore structure of the pack. In a 20/40-mesh higher than for a smaller material. The fallacy of using
proppant pack, particles larger than 100 mesh do not the crush testing value as a measure of a proppants
migrate through the pack because they are too large to potential application is the implication that all crush
travel through the pore throats, and particles as small material is equally undesirable. Crush testing should
as 200-mesh range have only limited mobility. Fines be seen as a relatively quick test procedure that lends
productionrather than bulk material failureis the itself well to quality assurance testing. Conductivity
main reason why proppant pack permeability reduces testing is much more useful than crush testing to
with effective closure stress. Particles smaller than 200 compare the capability of a proppant.
mesh, however, have been observed in the laboratory Furthermore, the crush procedure employed
to travel throughout the cell and in practice may be is a single-cycle test, thus relating typically to the
produced into the wellbore. Table 8-3 provides published reservoir environment existing for only a very short
values for fines generated from crushing common sand period of time after the completion of the well.
proppants (from Stim-Lab Proppant Consortium Manyif not mostwells experience stress cycling
reports and supplier publications). during production operations, and proppant crush is
known to get progressively worse with repeated cycles.
Multiple crush cycle tests are useful to understand
the impact of stress cycling on fracture conductivity
because this will directly impact production, but they
are rarely run (Vreeburg, et al, 1994).
8-4.7.2 Crush Testing
Crush resistance tests are conducted on proppant

samples to determine the amount of proppant crushed
at a given stress. This test is useful for determining and
comparing the crush resistance of proppants. Standard
tests are conducted on samples that have been sieved
Figure 8-33 Ottawa sand after exposure to 6,000-psi so that all particles tested are within the specified size
closure stress (courtesy of Atlas Resin Proppants) range. The amount of proppant material crushed at each
stress level is quantified. Evaluation of test results should
Table 8-3 Crush Values for Common Proppants provide indications of the stress level where proppant
Proppant Mesh Size Stress, psi % Fines crushing is excessive and the maximum stress to which
12/20 3,000 5.4
Ottawa Sand 20/40 4,000 4.0 the proppant material should be subjected.
20/40 8,000 14.0 The first step of the procedure involves removing
12/20 3,000 15.5 material that is outside the upper and lower sieve ranges.
Brady Sand 20/40 4,000 11.4
20/40 6,000 15.1
Sieve size specifications allow up to 10% of material
LWC 20/40 8,000 6.0 outside the given range (i.e. 20/40). It is important to
ISP 20/40 10,000 6.0 remove this material before the sample is prepared for
Bauxite 20/40 12,500 6.8
testing in order to obtain repeatable results.
302
The standard testing procedures for measuring much higher than that of a full monolayer. Graphical
proppant crush involve loading a pre-set volume of depictions of full and partial proppant monolayers
proppant into a crush cell that has a floating piston of are shown in Figure 8-35.
2 in. diameter. (Previous standard procedures had not
defined the piston diameter). The loaded cell is then Conductivity vs. Stress for Various Concentrations
Proppant Pack Conductivity, mD-ft

of 20/40 Ottawa Sand
placed in a hydraulic press to apply a direct load onto 100000
0.25 lb/sqft
the proppant grains at a controlled rate increase of 2,000 2.0 lb/sqft
10000 0.5 lb/sqft
psi/min to the desired stress, after which the full load 4.0 lb/sqft
1.0 lb/sqft
is held for two minutes. The sample is then sieved to 1000 6.0 lb/sqft
determine the quantity of material crushed to a size
100
that falls below the lower mesh sieve, and the weight
percentage of the crushed material is measured. 10
0 2000 4000 6000 8000 10000
Closure Stress, psi
8-4.8 Proppant Concentration
Figure 8-34 Effect on sand concentration on the
conductivity of a packed fracture (from Predict K, 2007)
Proppant concentration refers to the mass of proppant
per unit area of fracture placed by the treatment. It is
usually expressed in lb/ft2 (or kg/m2) of the fracture.
The flow capacity of proppant packs increases
with fracture width and with increasing numbers
of proppant layers for a given proppant size. Most
standard laboratory testing is conducted between
hard sandstone core samples at a concentration of
2 lb/ft2. Laboratory testing has shown that 0.5 lb/
ft2 of proppant may be lost to embedment in the
fracture face of moderately hard formations (Stim-
Lab, 1997). Higher proppant concentrations within
a proppant pack are directly proportional to increased
fracture width and therefore relatively proportional Figure 8-35 Depictions of a full proppant monolayer and
to increasing fracture conductivity. However, fracture a partial proppant monolayer (from Edgeman et al., 2004)
conductivity is not directly proportional to proppant
concentration due to wall effects. As shown in Fig. Previous studies by van Poolen et al. (1958) had
8-34, from 0.5 lb/ft2 to 2.0 lb/ft2 the impact of shown the flow capacity of proppant packs increased
proppant concentration upon fracture conductivity is with fracture width or, for a given proppant size, with
fairly linear when tested between hard sandstone core increasing numbers of proppant layers. The prediction
platens exhibiting minor embedment. of a high degree of fracture conductivity provided by a
In 1959, Darin and Huitt reported the results of a partial monolayer of proppant was significant because it
study of fracture flow capacity based upon evaluation was thought to provide a means to achieve high fracture
of the permeability of propped fractures. They provided conductivity by placing a minimal amount of proppant.
a numerical description of the differences between the In the case of a full monolayer, the grains are envisioned
extremes of an open fissure and a packed fracture, in to be closely packed (grains touching neighbor grains).
which the space between the fracture faces is completely Fractures containing partial proppant monolayers are
filled with a single or multiple layers of propping envisioned to possess similar geometry but with vacant
agent. The study demonstrated that there is an areal area around and between the proppant particles, thereby
proppant concentration below that of a full proppant increasing the relative conductivity of the propped
monolayer in which the resulting conductivity is fracture. By definition, a proppant monolayer is created
303
Modern Fracturing
when a propped fracture has a width equal to one particle lb/ft2. It is therefore disturbing to note from Fig. 8-36
diameter with no remaining voids into which additional that the minimum conductivity for 20/40 sand occurs
particles could be placed. The areal concentration at a proppant concentration of between 0.2 and 1.25
required to create a monolayer with a particular proppant lb/ft2, well overlapping the commonly employed average
type and mesh size can be approximated by: concentration ranges used in typical job design.
The prospect of fracturing with proppant partial
Ca = 5.2 (1 p ,min ) prop D prop , (8-6) monolayers was met with a great deal of initial
excitement, which gradually waned as field attempts
where Ca is the minimum areal concentration to place partial proppant monolayers resulted almost
required to obtain a full monolayer (lbm/ft2), prop universally with disappointment and failure. Suggested
the minimum obtainable porosity of proppant pack reasons for the difficulties included the inability to obtain
(fraction), prop the proppant specific gravity and Dprop uniform and complete coverage of the fracture with
the average proppant diameter (in.). a partial monolayer, insufficient proppant strength to
The relationship of fracture conductivity to support the load, loss of fracture width due to proppant
concentration as defined by Darin and Huitt is shown embedment and, potentially deleterious non-Darcy
in Fig. 8-36 for 20/40 mesh sand at low closure stress. flow effects in the relatively narrow propped fracture.
Additional annotation is provided to illustrate the Additionally, contributing to the repeated failures in all
number of layers associated with the respective proppant likelihood was severe gel damage from the fluids bulked
concentration. The predicted fracture conductivity up with the dirty polymers of that era.
can be seen to peak near 9,000 md-ft at a proppant Recent development of ultra-lightweight proppants,
concentration of about 0.1 lb/ft2 which in this case is however, has rekindled interest and efforts to employ
approximately one half of a monolayer. A concentration partial proppant monolayers (Brannon et al, 2004). Initial
of over 4.0 lb/ft2 of 20/40 sand is necessary to attain a field success using low concentrations of ultra-lightweight
similar level of conductivity with a multi-layered pack. proppants led to refinements in placement techniques
and further reductions in proppant concentration
(less than 0.1 lb/ft2). The resulting stimulated well
Monolayer
productivities were reported far beyond historical

Full
Partial Monolayer Multilayer

105 expectation, ultimately leading to the conclusion that
highly conductive, proppant partial monolayers had been
placed over a much increased propped fractured area.
104 0.06 lb/ft 2
kfw (md-ft)
100000
2000 psi Stress
4000 psi Stress
103
2 lb/ft 2
6000 psi Stress
0.125 lb/ft 2
10000
Conductivity, mD-ft
102
0.01 0.10 1.0 10
Proppant Concentration,
lbm/ft2 of propped area 1000
Figure 8-36 Fracture conductivity vs. proppant
concentration for 20/40-mesh Ottawa sand at low
temperature and closure stress (from Darin and Huitt, 1959)
100
0.01 0.1 1 10
Some fracture engineers will design a packed ULWP (Hprop = 1.25) Concentration, lb/ft2
propped fracture regardless of the permeability of the

reservoir or extent (length) of the created propped Figure 8-37 Fracture conductivity vs. proppant
concentration for 12/20-mesh ULW proppant (prop
fracture. The practice for conventional proppants has = 1.25) at 100 F and 1,000 psi closure stress (from
been to strive for concentrations averaging at least 1 Brannon et al., 2004)
304
Figure 8-37 shows the laboratory measured slurry 1

r = = 2.5 n
,
conductivity of a 12/20-mesh ultra-lightweight (prop fluid (8-7)
1 f s
= 1.25) proppant at concentrations ranging from 0.1 f s ,max
to 1.0 lb/ft2. Note the signature of the curve is very where fs is the proppant volume fraction, fs,max is the
similar to that predicted by Darin and Huitt and that maximum fraction of the mobile slurry and n is the
a peak conductivity of greater than 100,000 md-ft was power law exponent.
achieved using a proppant concentration of 0.06 lb/ft2.
The substantial open flow areas provided by a partial 8-5.2 Convection
monolayer are illustrated in Fig. 8-38.
Density differences between fluids may result in the
lighter fluid overriding the denser fluid or, a denser
fluid flowing under the lighter fluid. This phenomenon
is known as convection and may be important in
situations such as a high-density slurry stage (higher
proppant concentration) under-riding a previously
pumped stage. Several authors (Barree and Conway,
1994; and Clark and Zhu, 2004) have shown that
convection is generally not of significant concern in
properly designed fracturing treatments. If excess pad
is pumped, convection could occur until the excess
Figure 8-38 Core slab showing an ultra-lightweight pad leaks off. Convection has an extreme sensitivity
proppant (prop = 1.054) partial monolayer, 0.02 lb/ft2 to fracture width, for example in low-modulus rock
(courtesy of Fractech Ltd.)
that is generally associated with high-permeability
formations. In such cases where the width is large,
8-5 Proppant Placement the occurrence of convection is more likely. If the
fracture width is not large, as is typical in fractures
The objective of hydraulic fracturing is to place proppant in low-permeability gas reservoirs, convection is
within a created fracture to provide a conductive pathway much less of a significant concern.
for production. The presence of proppant introduces
important considerations in the behavior of fracturing 8-5.3 Proppant Transport
fluids in creating the hydraulic fracture, including the
effect of proppant on fluid rheology, convection of the Proppant transport is an essential consideration in
slurry and proppant transport. treatment design to effectively place proppant in
the pay zone. The transport of proppant in fractures
8-5.1 Effects on Fluid Rheology during a hydraulic fracturing operation and the
suspension of proppant during the fracture and
At proppant concentrations greater than about 3 ppga, closure process can significantly impact the propped
the viscosity of proppant-laden slurry is higher than that area. Poor proppant transport can result in excessive
of the fracturing fluid alone (Novotny, 1977). Although proppant settling, often into the lower regions of
the relationship of proppant concentration to slurry the created fracture (and potentially below the
viscosity is well known for Newtonian fluids, it is less productive interval and/or in high-water-saturation
well established for substantially non-Newtonian fluids formations), yielding relatively short effective fracture
like those typically used for fracturing applications. lengths and insufficient coverage of the total height
A relationship for power law fluids was presented by of the productive zone. Additionally, proppant pack
Nolte (1988) (see also Section 4-4.4.2). The viscosity damage due to inadequate clean-up of the resultant
ratio, r , can be obtained from: propped fracture may cause significant reduction in
305
Modern Fracturing
the conductivity of the propped fracture area. The Law suggests that lighter particles fall or settle more
cumulative effects of these phenomena can result in slowly than heavier particles. Therefore, decreasing
a reduction of overall stimulation efficiency, yielding the specific gravity of the proppant to approach
steeper post-stimulation production declines than may that of the fracturing fluid is, theoretically, the most
be desired. Post-frac production analyses frequently practical method to improve proppant transport in
illustrate that the effective fracture area is less than low-viscosity fluids. It can be surmised from Eq. 8-8
that expected based upon the design, suggesting the that as the specific gravity of the proppant approaches
proppant was not placed effectively throughout the the specific gravity of the fluid, nearly neutral
designed fracture area or the existence of excessive buoyancy is approached and the proppant settling
proppant-pack damage (or that the fracture modeling velocity nears zero (Fig. 8-39).
was inaccurate). Optimizing effective fracture area is
often critical to economic exploitation of reservoir
assets, thus maximizing the propped fracture area is a
key parameter for generating desired stimulated well
performance. Efforts to improve effective fracture
area have historically focused on the proppant
transport capability of the fracturing fluid and the
fracture clean-up attributes.
The effects of convection on proppant transport
were discussed above. Consider a vessel of sand
mixed in water, which is shaken to disperse the
particles and then allowed to stand statically. The
sand will rapidly fall or settle to the bottom of the
vessel. Proppant settling rate is typically assessed by Figure 8-39 Ultra-lightweight proppant (prop = 1.25) in
10.2-ppg brine (from Brannon et al., 2005)
comparing single-particle static settling velocity in a
column of water using Stokes Law. As shown in Eq. Figures 8-40 and 8-41 illustrate proppant
8-8, the terminal velocity, vt, is a function of Dprop, distributions predicted by hydraulic fracture
the median proppant particle diameter, mfluid, the modeling software for comparing sand slurried in
Newtonian viscosity of the fluid, and the difference fresh water to ULW proppant (prop = 1.25) slurried
between the specific gravity of the proppant and the in 10-ppg brine. The benefits of the near-neutral
fluid, prop and fluid, respectively: buoyancy are clearly shown by the greatly increased
D prop
2 propped fracture area.
vt = 1.15103 ( prop fluid ). (8-8)
fluid
Increasing the fluid viscosity and/or decreasing the

proppant diameter have historically been used to
facilitate improved proppant transport. In each case, the
modification to positively impact proppant transport
has a negative effect on fracture conductivity. Increased
viscosity brings a greater likelihood of residual fluid
damage, and reduced proppant diameter directly
reduces proppant pack permeability.
The specific gravity of the proppant, prop , is
the single variable remaining for consideration as
the means to address the proppant transport issues. Figure 8-40 Software model of proppant distribution of
Assuming all other variables to be constant, Stokes sand in fresh water frac (from Edgeman et al., 2004)
306
treatment, if the fracture remains open for extended

periods. Fracturing fluids are typically viscoelastic in
nature, which has the effect of driving the proppant to the
center of the flow channel as a dense sheet surrounded
by clear fluid along the fracture walls. This has the effect
of accelerating particle settling, particularly for lower
proppant concentrations (Biot and Medlin, 1985; Goel
and Shah, 2001; and Clark and Zhu, 2004).
Stokes Law single particle setting rates in 2% KCl
for 20/40 mesh proppants are illustrated in Fig. 8-42.
The calculated static settling velocity of Ottawa sand in
Figure 8-41 Software model of proppant distribution of
ultra-lightweight (prop = 1.25) proppant in 10-ppg brine fresh water was 16.6 ft/min compared to 23.2 ft/min for
frac (from Edgeman et al., 2004) a sintered bauxite and 4.3 ft/min for an ultra-lightweight
proppant, less than 20% that of the high-density sintered
The basic equation for single-particle settling has bauxite with specific gravity of 3.4.
limited applicability in conventional hydraulic fracturing
because the fluids are typically non-Newtonian and the 23.2
25
proppant particles are concentrated and may be close 20.0
17.2 16.6
to the fracture faces. These phenomena can result in 20 15.9
hindered settling, where the particles can obstruct

Settling Velocity
t ft/min
15
each other, and clustered settling, where the particles
cluster and effectively increase the proppant diameter. 10
4.3
Economides et al., (2002) presented a correlation 5 1.4
for the particle velocity vsol in hindered settling in
terms of volume fraction of solids fs: 0
Bauxite Carbolite ULWP (Hprop = 1.25)
ISP Sand ULWP (Hprop = 1.05)
vsol = v F ( f s ) ,
(8-9) RCS
Figure 8-42 Static single particle proppant settling rates in

where water (from Brannon et al., 2005)

F ( f s ) = (1 f s ) ,
(8-10) Proppant transport studies in slickwater and
linear gel fluids were recently reported by Brannon
and the exponent ranges from 5.5 to 2 for low and high et al. (2007). Extensive testing in a large-scale slot
values of NRe, respectively. A generalized form of Stokes apparatus was conducted to evaluate the relative effects
law is used for power law fracturing fluids: of various component and treatment parameters on
1
the proppant transport capability of various slurry
( ) gD n '+1 n'
compositions. The acquired data were processed
prop sol
v =
fluid
,
3n '1
18 K ' (8-11) using the techniques published by Biot and Medlin
(1985) to determine the minimum horizontal slurry
where v is the particle settling velocity in a pseudoplastic velocities necessary for proppant transport using the
fluid, g is gravitational acceleration, Dsol is the solid respective slurry compositions.
particle diameter, n is the geometry dependent power The minimum horizontal velocity required to
law exponent and K is the consistency index. maintain proppant transport, a value called MHVST,
Some high-viscosity, crosslinked fracturing fluids may be defined using the Biot-Medlin process. At
may exhibit near-perfect transport, negating concern lesser velocities, proppant movement is limited to
about proppant settling until after the pumping sliding or rolling over a settled bed. A Slurry Properties
307
Modern Fracturing
Index, ISP, essentially the basic Stokes law equation, A simple linear fit of Fig. 8-43 leads to Eq. 8-13.
was used to characterize the physical properties of a
given proppant/slurry composition:
MHVST = CTRANS I SP , (8-13)
D prop
2
I SP = ( prop fluid ).
(8-12) where CTRANS , termed the transport coefficient, is
fluid
equal to 0.0117 from the linear curve fit and has the
Slurry Property Indexes and observed minimum units ft.sec-1.cp.mm-2.
horizontal flow velocities required for suspension Conventional proppants such as sand require flow
transport in several proppant slurries are velocities of an order of magnitude greater than those
shown in Table 8-4. required for ultra-lightweight proppants to be maintained
within suspension transport in a fracture. The ability to
Table 8-4 Slurry Factors and Calculated Minimum maintain suspension of ultra-lightweight proppants at
Horizontal Flow Velocity Required for Suspension very low minimum velocities illustrates the real potential
Transport, MHVST for Slot Flow Tests (from Brannon et
al., 2007) that ULW proppant partial monolayers can be placed to
Dprop2 ,
fluid fluid , cp ISP MHVST
provide very large effective propped fracture area.
prop
mm2
2.65 0.4032 8.34 7 119.3 1.400
2.65 0.4032 9.4 7 110.1 1.200
8-6 Fracture Conductivity
2.65 0.4032 10.1 5 104.0 1.188
2.65 2.070 8.34 60 78.56 1.018 The American Petroleum Institute (API) defined
2.65 0.4032 9.4 29 26.57 0.400 laboratory testing procedures to measure the proppant
2.65 0.4032 8.34 60 13.92 0.173 permeability and a reference fracture conductivity, which
1.25 0.4264 8.34 11 11.14 0.143 measures the viscous drag of the fluid upon the proppant.
1.25 0.4264 9.4 7 8.61 0.140 These procedures were recently incorporated in the
1.05 0.5810 8.34 5 6.43 0.102
publication of ISO Standard 13503-5. These procedures
1.25 0.4264 8.34 60 2.04 0.027
will be discussed first. However, they only measure the
1.25 0.4264 9.4 29 2.08 0.020
1.05 0.5810 8.34 29 1.11 0.017
effects of viscous drag on the proppant. As discussed in
Section 8-1.3, it is critical for gas well stimulation to
include the inertial effects of flow in the fracture due to
An empirical model was derived to relate the the relatively high velocity through the proppant pack.
minimum horizontal flow velocity required for suspension Multiphase effects can also be significant due to the
transport of a given slurry composition based upon its interaction of more than one fluid phase.
respective Slurry Properties Index (Fig. 8-43).
8-6.1 API Short-Term Testing Procedure
3.000
Minimum Horizontal Velocity for
2.500 A number of different linear and radial flow testing

Suspension Transport, MHVST
cells were used to measure proppant conductivity prior

2.000
to the mid-1980s (Gidley et al., 1989). In 1989, the
1.500 American Petroleum Institute published Recommended
1.000 Practice 61 for measuring proppant conductivity, which
involved flowing a single-phase liquid (water with 2%
0.500
KCl) through a 7 1 in. linear proppant pack cell
0.000 (Fig. 8-44) at flow rates of 1 to 10 ml/min (API, 1989).
0 50 100 150 200 250
Slurry Properties Index, ISP Measurement of the pressure drop across ports located
adjacent to the middle 5 in. of the proppant pack
Figure 8-43 Minimum horizontal flow velocity for slurry
transport as a function of the Slurry Properties Index, ISP
allows calculation of the permeability at a given stress
(from Brannon et al., 2007) condition according to Darcys Law.
308
8-6.2 ISO Long-Term Testing Procedure
A number of researchers investigated the effects

of longer periods of stress on the proppant. It was
found that the 15-minute steps in the short-term
API test procedure did not give adequate time for
the proppant pack to stabilize under the new loading
conditions. Recently the ISO testing standard 13503-
2 established a new procedure that requires each stress
level to be held for 50 hours.
The test apparatus includes a thermal housing
(Fig. 8-46). Because these tests are run at high
Figure 8-44 API fracture conductivity test cell (courtesy of
Carbo Ceramics Inc.)
temperature, they must use de-oxygenated, silica-
saturated liquids to replicate conditions in the
API RP-61 defines the measurement of what reservoir and minimize interaction induced by the test
is referred to as the short-term conductivity for a procedure between the fluids and the test apparatus,
proppant. The general procedure for testing is to load sandstone platens and proppant.
the cell with 2 lbs/ft2 of proppant and then increase the
load in 2000-psi increments, holding each stress level
for 15 minutes, during which distilled water is flowed
at rates from 2.5, 5 and 10 ml/min.
Figure 8-46 Long-term fracture conductivity test

apparatus loaded with three API linear flow cells
(courtesy of Carbo Ceramics Inc.)
In all short- and long-term tests, the flow rates

used are (per the API and ISO procedures) in the
range of 1 to 10 ml/minute. These rates correspond to
a range of 4 to 40 BOPD, or 13 to 130 Mscf/d of gas at
a flowing bottomhole pressure of 1000 psi, produced
from two fracture wings of a 30-ft-high fracture
(Vincent et al., 1999). The superficial velocity of the
water during the test is in the order of 0.2 to 2.0 in.
per minute. In actual fractures, the superficial velocity
can be several feet per second and the interstitial
velocity consequently even higher. These reference
conductivity values when used in fracture design need
Figure 8-45 API fracture conductivity test cell with two
cells loaded in a crush press, ready for testing (courtesy to be used with the appropriate adjustments for non-
of Carbo Ceramics Inc.) Darcy and multiphase flow effects.
309
Modern Fracturing
8-6.3 Non-Darcy Flow Testing

1 1
= +X. (8-15)
kF kD
Beta Factor, the inertial flow coefficient, is determined
by flowing more realistic velocities through the API/ The above value of kF can be substituted into equations
ISO conductivity cell after each step in the long- of flow as the effective value of permeability that
term conductivity procedure. The beta factor () is a includes both viscous (Darcy) and inertial (non-
proportionality coefficient that is determined by solving Darcy) flow effects.
Eq. 2-38 (see Chapter 2) to match the observed pressure Cooke (1973) ran laboratory tests with brine, gas
drops. To fully match the expected flow conditions in and oil, and showed that is a function of permeability
the proppant pack, it may be necessary to run tests at and independent of fluid type. Example plots are shown
rates of tens of liters per minute. However, in practice, in Fig. 8-48 for (a) 8/12 mesh, (b) 10/20 mesh and (c)
it is usually only necessary to achieve flow rates around 20/40 mesh sand proppants.
1500 ml/min of brine or 25,000 ml/min of nitrogen
gas in order to ensure fully developed inertial flow and 10-1
A Gas
accurate determination of . Brine
The Forcheimer equation, given as Eq. 4-116, may Oil
Beta, atm s2/gram

10-2
be linearized and rewritten as:
1
Y= +X,
(8-14) 10-3
kD
where Y = p/(L v) and X = (v/) and kD is the

10-4
reference permeability measured under laminar 10 100 1000 10000
(Darcy) flow conditions. A Forcheimer plot of Y vs. X Permeability - Darcies
has a gradient of and an intercept equal to (1/kD), 10-1
B Gas
as shown in Fig. 8-47. Brine
Oil
Beta, atm s2/gram
Gas Flow Data: 2 lb/sq ft 20/40 10-2

Carbo-Lite at 8,000 psi and 250 F
0.07
0.06 10-3
0.05
Y = 1/k
0.04
10-4
0.03 10 100 1000 10000
0.02 Permeability - Darcies
0.01 10 -1
C Gas
0.0
0 20 40 60 80 100 Brine
X = pv/N Oil
Beta, atm s2/gram
10-2
X < 20:ko=144 Darcies B = 0.00065 atm-s /g
2
X > 20:ko=121 Darcies B = 0.00063 atm-s2/g
Figure 8-47 Example Forcheimer plot (courtesy of Stim-

10-3
Lab Inc.)

A Forcheimer permeability (kF) may be defined 10-4
as the equivalent permeability of the proppant 10 100 1000 10000
Permeability - Darcies
incorporating inertial flow effects. From Eq. 8-14 or
Fig. 8-47, the value of Y for a given velocity (flow rate) Figure 8-48 Inertial coefficients for sand proppants with
will be equal to the reciprocal of kF. Thus, various fluids (after Cooke, 1973)
310
The Cooke (1973) relationship is of the form: Using the concept above and with X = (v/),
a plot can be made of the ratio of the fluid phase
= b a, (8-16) Xfluid = (fluidvfluid/fluid) to the total Xt = (Xfluid +
k
Xg) where Xg = (gvg/g). Data can then fit to a
where constants a and b describe the experimental power law model of the form:
lines from Fig. 8-48. Although the a and b terms b
X fluid

are used in some fracture design software programs, = 0 + a X , (8-17)
t
care needs to be taken because the values will be
impacted by additional proppant crush at a higher where 0 is the value for saturated gas in single-
stress and by any fracturing fluid damage that causes phase flow, and a and b are experimentally derived
proppant pore throat plugging. parameters. Figure 8-49 shows an example
Factors that influence the inertial flow effects for experimental plot of vs. (Xfluid/Xt) with matched data
single-phase flow include: for testing a 20/40 mesh sand proppant.
Initial proppant permeability Some models correlate liquid saturation and
Porosity of the proppant velocity in the fracture to actual liquid production
Curvature of streamlines (proppant angularity) in order to calculate multiphase flow effects.
Relative aperture of pore throat to pore space Unfortunately, many fracture design and production
Proppant size distribution simulators do not include non-Darcy and multiphase
Heterogeneities flow effects. If the production model is only utilizing
Surface roughness Darcys law, the engineer must calculate the total
pressure drop for the expected flow conditions and
8-6.4 Multiphase Flow Tests then convert the pressure drop into an effective
conductivity or permeability damage factor. The total
Multiphase flow testing follows the same procedure as pressure drop per unit length = viscous drag + inertial
the single-phase long-term conductivity test augmented effects + multiphase effects:
by introduction of the second fluid phase.
p v
Researchers have found it convenient to use a = + v 2 . (8-18)
L
k
term gamma (), which for a given flow condition
is defined as the ratio of the actual pressure drop The ratio of viscous drag forces to the total pressure
under multiphase conditions divided by the pressure drop allows calculation of an effective fracture
drop for single phase flow. conductivity, (kfw)e , compared to the reference or
nominal conductivity, kfw, where:
100
(v / k )
(k f w)e = k f w. (8-19)
(v / k + v 2 )
Alternatively, from Eq. 8-15, the multiphase flow

Gamma
10 effects will reduce the Forcheimer permeability by a

factor of gamma (). Thus, the equivalent multiphase
permeability (kM) is given by:
kF
kM = , (8-20)

1
0 0.02 0.04 0.06 0.08 0.1 0.12
or
(Xfluid /Xt)
Measured Predicted 1 1
Figure 8-49 Gamma vs. Xfluid/Xt for 3-lb/ft2 20/40 Jordan = + X . (8-21)
sand with 2000 psi closure and 100 psi internal pressure kM k D
at 75F (after Jin & Penny, 1998)
311
Modern Fracturing
Where kD is the permeability under Darcy flow It is important to also realize the impact of
conditions. As an approximation for non-Darcy and different proppant types on the cleanup of fracturing
multiphase flow effects, the ratio of single-phase to fluids. Figure 8-51 shows the cleanup effects of the
total pressure drops calculated above can be applied as same borate crosslinked guar fluid with three mesh
a damage factor to reference long-term conductivity distributions of proppant: 20/40 mesh lightweight
data that are typically included as lookup tables in ceramic (LWC) and 20/40 and 40/60 mesh Ottawa
fracture design models, where: Sand. There is a 250% difference in pressure required
k to initiate cleanup of the broken gel between the
Damage Factor = k = (1 + Xk D ) .
D
(8-22) narrowly sieved LWC proppant compared to the
M
more broadly sieved, angular 20/40 mesh sand. As
A variety of other factors can be investigated in the average proppant diameter and pore throat size are
laboratory using specific test conditions. reduced in going to a 40/60 mesh proppant there is
a similar increase in required differential pressure.
8-6.5 Gel Damage The consequence of this, as shown in Fig. 8-52, is
that although the retained permeability of the LWC
For practical reasons, most conductivity testing follows is close to 70% of the undamaged value, it is reduced
the API procedures of using a 2% potassium chloride to around 50% for the sand proppants.
(KCl) solution. In order to fully understand the impact
%p to Initiate Cleanup at 4.0 lb/sqft +
of the fracturing fluid on conductivity it is necessary Borate Crosslinked Guar + Breakers at 150F and
to replicate the treatment conditions of placing 2000 psi Closure Stress
0.06
proppant within the fracture under the actual reservoir
conditions of pressure and temperature. 0.05
Numerous studies have been conducted on
the impact of the gel filter cake and residue on 0.04
%p, psi/5 in.
conductivity (Stim-Lab, 1992, 1997, 2000; and

0.03
Brannon and Pulsinelli, 1990a and 1990b). For
example, Fig. 8-50 shows an SEM of a proppant pack 0.02
that was tested under simulated treatment conditions
with a fracturing fluid. The left side of the figure 0.01
shows gel damage concentrated at the interface with
0
the sandstone platens. The right side of the figure 20/40 LWC 20/40 White Sand 40/60 White Sand
shows gel residue within the proppant pack. Figure 8-51 Retained conductivities of various proppant
packs (courtesy of Stim-Lab Inc.)
Much of the early design of fluid systems

focused on rheology. Volumes of data were gathered
regarding the behavior of fluids as functions of shear,
temperature and time in order to develop fracturing
fluids capable of transporting proppant effectively
within any reservoir. The ability of a fluid to clean
up, or flow back out of the well after fracturing, was
traditionally related to the viscosity of the degraded
fluid. More recently, the secondary concerns of
emulsion creation, precipitation, clay swelling/
migration, polymer concentration and insoluble fluid
Figure 8-50 Fracture fluid damage in the proppant pack
(from Stim-Lab, 1992) residues have been receiving more attention.
312
Laboratory tests that attempt to replicate downhole 8-6.6 Other Factors

conditions and measure proppant conductivity
after damage have provided a better understanding Due to the high cost of long-term conductivity
of the impact of each component within modern testing, there are relatively few extensive databases
fluid systems. The range of results reported in the of tests. Other factors which can be investigated
literature has shown that each fracturing fluid should as warranted include:
be evaluated as a complete system. Attempts to draw Cycling: Operating conditions over the life of the
general conclusions regarding the efficacy of a specific well will result in cycling the effective stress on the
crosslinker or breaker may be invalid when other proppant as a result of shutting in and opening the
components of the fluid system change. well and the resultant changes in flowing pressure in
the fracture. It is possible to investigate this effect by
Conductivity and % Retained Permeability
vs. Proppant Types at 4.0 lb/sq ft
cycling the applied load to the API conductivity test
placed with YF130LG cell. Early studies showed progressive loss of fracture
12000 100
90
width and proppant permeability with successive
10000
80 stress cycles. This indicates that damage occurs in
% Retained Permeability
Conductivity, md-ft
8000
70 both the proppant and the fracture face. Ceramic
60 proppant exhibited better overall resistance to fatigue
6000 50
failure than natural sand, and softer formations
40
4000
30
showed increased sensitivity to repeated stress
20 cycles. The effect of cycling is believed to be more
2000
10 pronounced with lower proppant concentrations
0
20/40 20/40 40/60 20/40 20/40 40/60
0 (Ouabdesselam and Hudson, 1991).
LWC White
Sand
White
Sand
LWC White
Sand
White
Sand
Figure 8-53 shows results of testing a Lightweight
Figure 8-52 Differential pressure to initiate cleanup in the
Ceramic (LWC) and a Resin-Coated Sand (RCS)
proppant pack (from Stim-Lab Inc., 1997) proppant with a 25-cycle test phase between 8000
and 4000 psi to represent the effect of well operations
In the absence of specific gel damage tests relating over the life of the well. In this example the effects
to specific well and reservoir conditions, the practicing are 26% and 35% reductions in conductivity,
engineer has several alternatives for taking into account respectively. Similarly, for a 20/40 mesh sand
the effect of gel damage: proppant cycled between 4000 and 1000 psi there
1. Utilize a database of retained permeability data was a 30% loss in conductivity.
for similar fluid and reservoir conditions.
2. Utilize a design program from a proppant 10,000
testing laboratory which contains correlations

8,000
for calculating gel damage derived from an
Conductivity, md-ft
extensive test program under a variety of well

and Stress, psi
6,000
Ceramic loses 26%,
and reservoir conditions. RCS loses 35% due to 25 cycles
3. Estimate a damage factor to apply to the long- 4,000
term proppant conductivity data typically in

2,000
the order of 30 to 50%.
4. Conduct Pressure Transient Analyses (PTA) or
history matching of production data to estimate 0 100 200 300 400
Hours
the effective fracture conductivity of existing Stress (psi) LWC RCS
wells. The gel damage value can be back-

calculated and used as an estimate for future Figure 8-53 Multiple stress cycle test results for various
jobs with similar fluid systems. proppants (from Stim-Lab Inc., 2000)
313
Modern Fracturing
Long Term Exposure: Limited data are available simulated bottomhole stress and temperature. The
on the long-term stress exposure of proppant packs. circular indentations made by the proppants are
One study presented by Cobb and Farrell (1986) seen surrounded by fine-grained, crushed sandstone
tested a variety of proppants for three months of particles. The displacement of crushed fine formation
continuous loading. Results of testing a high-quality particles, a process called spalling, may further
LWC showed less than a 5% reduction in conductivity reduce conductivity as the particles create fines
over the period of testing. damage within the proppant pack.
Temperature: The effects of elevated temperature
on proppant permeability have proven difficult to 8-7 Proppant Flowback
measure consistently in the laboratory. Early literature
reported substantial degradation of all proppant types 8-7.1 Proppant Flowback Control
at high temperatures, but this was later found to be
an artifact of the test procedure; the KCl fluids were Proppant flowback after hydraulic fracturing
not fully silica-saturated, and the sandstone cores had treatments is a significant concern within the
started dissolving during testing. More recent work industry. Net effective stress in the early part of
confirms that ceramic proppants are virtually unaffected cleanup after a hydraulic fracturing treatment is
by high temperatures, which makes intuitive sense much lower than the closure stress on the well after
as they are sintered at approximately 2700 F during it is stabilized for production. Proppant flowback
manufacturing. Elevated temperatures do appear to can occur during that time. Similarly, when a well
have a detrimental effect on sand substrates and can is shut-in during a production cycle, the formation
affect the properties of resin coatings. pressure near the wellbore increases, and the fracture
experiences a simultaneous decrease in net closure
stress. Consequently, when production is re-initiated,
opportunity exists for proppant flowback.
Proppant flowback during cleanup and
subsequently over the lifetime of the well can have
detrimental effects to the success of a treatment.
Any significant flowback of proppant can be highly
detrimental because the fracture could close and act
as a choke to production. Also, abrasive damage can
occur to surface and production equipment such as
chokes, valves and separators. These issues can lead to
increased safety concerns, reduced well production and
increased costs. Factors identified as affecting proppant
flowback include formation type, fracture width,
closure stress, proppant size and type, and proppant
concentration. (See also Section 9-10.)
Figure 8-54 Proppant embedment and fines generation
Proppant flowback can be prevented by the
(from Stim-Lab, 1987)
appropriate use of perforating strategies (see Chapter
Embedment and Spalling: Embedment of 6), treatment design (see Chapter 4), treatment
proppant particles into the faces of the fracture execution (see Chapter 9) and well management
wall results in a reduction of the width of the (Vreeburg et al., 1994). However, there are many
fracture, directly resulting in decreased fracture treatments where these practices are not employed,
conductivity. The photographs shown in Fig. 8- for a variety of reasons. The next two sections detail
54 illustrate embedment of proppant particles into the main proppant-based techniques available for
a sandstone surface after 100 hours of exposure to preventing proppant flowback.
314
8-7.2 Curable Resin-Coated Proppant 8-7.3 Proppant Flowback Control Additives
Curable resin-coated sand was first introduced in 1975 Several on-the-fly additives were introduced in
as a means to control proppant flowback by locking the the late 1990s to control proppant flowback in
proppant in place in a conventional hard-rock propped applications where resin-coated proppants had proven
hydraulic fracture. The curable coating sets with ineffective. Among those were tacky resins, fibers
formation temperature to produce a bond between the and deformable particulates.
proppant particles. Resin-coated proppants are generally
effective for controlling proppant flowback. However, 8-7.3.1 Tackifiers
many wells treated with resin-coated proppants have,
after time, experienced proppant flowback, suggesting Tacky resins do not fully cure; therefore, the
that the inter-grain bonding of the resin-coated particles created inter-grain bonds are less subject to brittle
has failed. Figure 8-55 shows intact bonds between resin failure than the cured bonds among resin-coated
coated proppants; Fig. 8-56 shows bonds broken due to proppants (Nguyen, 1998). The tacky resins have
stress cycling fatigue. Bond failure can also occur due to been shown effective in fractures with high closure
the chemical effects of the fracturing fluid components stress but less beneficial in applications where high
(breakers, buffers, surfactants, crosslinkers, etc.), production rates or multiphase flow are the main
proppant size, or the physical effects of stress cycling, contributors to proppant pack failure.
closure stress, shear, and temperature.
8-7.3.2 Fibers
Fiber additives have small diameters and large length

relative to the proppant particles, providing for a
span across several proppant particles (Card, 1995).
When the fibers are mixed in a proppant pack,
they create a complex net that serves to impede
proppant particle movement. Fibers have been shown
to be effective for controlling proppant flowback,
and increasing the fiber concentration has been
shown to further improve control. However,
because they reside within the proppant pack
Figure 8-55 Bonding between resin-coated sand
particles (courtesy of Santrol) pore space, conductivity damage is directly
proportional to the amount of fiber added.
8-7.3.3 Deformable Particles
Deformable particle proppant flowback control

additives are sized similarly to the proppant and
intermixed within the proppant pack matrix
(Rickards et al., 1998; and Stephenson et al., 1999).
When stress is applied, the proppant particles create
indentations in the surface of the deformable particles
as the softer deformable particles absorb the energy.
The indentations create increased contact surface
Figure 8-56 Stress-induced fracture through resin-coated
area between the particles, providing for increased
grain and inter-particle bond (courtesy of Santrol) resistance to particle movement.
315
Modern Fracturing
Two forms of deformable particles are commercially simulated production rate of 101 bpd. The rate at failure
available. A substantially spherical deformable particle was increased to 782 bpd by adding 15% by weight of
is employed for applications with closure stress from deformable particles to the LWC proppant pack.
250 psi to 6,000 psi and reservoir temperatures up An ancillary benefit of the use of deformable
to 275 oF. Needle-shaped deformable particles are particles is the ability to reduce the amount of
used for higher-stress applications (Stephenson et al., proppant crushing experienced with a given proppant.
2002). Photographs of both particles after exposure to As shown in Fig. 8-59, 20/40 Ottawa sand exposed to
sufficient stress for deformation are shown in Fig. 8- 7,000 psi closure stress exhibits crush in the form of
57. The increased inter-particle contact area can be seen fines generation in the order of 25%. Adding 15% by
in the indentations on the surface of the deformable weight deformable particles to the Ottawa sand pack
particle. Laboratory testing indicates that a 3:1 ratio of decreased the amount of fines to 15%. Additional
proppant particles to substantially spherical deformable increase in the deformable particle concentration to
particles is the optimum for controlling proppant 25% by weight decreased the fines to 12%.
flowback and maintaining proppant permeability. A
Reduced Crushing of 20/40 Ottawa Sand with
ratio of 9:1 has been suggested as optimum for the Deformable Proppant Additive at 7000 psi Closure
higher-strength, needle-shaped deformable particles. 30
Experience has shown that fracture conductivity
25
reduction is minimized with the use of deformable
20
particles for proppant flowback control. Fines, %
15
10
0
0 5 15 25
Weight of Deformable Proppant Additive, %
Figure 8-59 Increasing concentration of deformable

particles reduces proppant crushing (from Stephenson et
al., 1999)
Figure 8-57 Deformable particles after exposure to stress
(from Rickards et al., 1998 and Stephenson et al., 2002).
8-8 Proppant Selection
Increased Flowback Resistance of 20/40 LW Ceramic
with 15% Deformable Proppant at 4000 psi and 250 F From all the chemicals and materials, and all the services
1200 NO FLOWBACK provided as a part of the hydraulic fracturing operation,
1000 only one thing will provide benefit to the productivity
bpd / Mscf/day
improvement of the well: Proppant! Selecting the

10-ft interval
800
600 1220 correct proppant to use, the amount to place, and the
400
propped fracture geometry is therefore at the heart of a
FLOWBACK 732
successful application of hydraulic fracturing. The only
200
101 14 way to evaluate the solution to these three parameters
0
w/o Deformable Incorporating 15% Deformable is through numerical simulation. However, even with
Proppant Additive Proppant Additive
that, the heterogeneity of the reservoir that is being
Figure 8-58 Deformable particle effect on proppant stimulated and our overall lack of knowledge about
flowback (from Stephenson et al., 1999)
the properties of the reservoir and bounding layers
Characterization of flowback of a 20/40 LWC may mean that our initial solutions do not result in
proppant with and without deformable particles, at selection of the best proppant type, proppant size or
250 oF and 4,000-psi closure is shown in Fig. 8-58. treatment volume/mass to pump. Actual production
The LWC proppant was observed to flow back at a results should always be benchmarked against expected
316
performance; armed with this experience, the engineer applied Damage Factor. The engineer should realize
should be prepared to make appropriate design that taking this approach, even for moderate-rate
and proppant selection changes. gas wells, will often result in calculating Damage
Fortunately, as a starting point for the practicing Factors in excess of 90% and even in excess of
engineer, a number of fracture design models have 95% for higher-rate gas wells.
extensive proppant databases built into them. Producing conditions over the life of the well,
However, without exception, all of these data are based including changes in both reservoir and bottomhole
on the API/ISO long-term conductivity standards for flowing pressures that may occur on account of
Darcy, low-rate, viscous flow conditions. Most of the production or well work activity.
models will include opportunities for the engineer
to evaluate non-Darcy, inertial flow effects that are 8-8.2 Flowback Control
critical for gas wells. In addition, the engineer can
select a damage factor to be applied to the data in The effective control of proppant and formation sand
order to account for the plethora of other issues production over the life of a well is critical to long-
which we have described in this chapter. term productivity. From a sand control perspective in
unconsolidated formations, the ideal solution is often
8-8.1 Productivity Potential to increase the fracture length until the matrix velocities
are low enough to prevent the movement of formation
Any proppant evaluation should start with identifying fines. However, estimating the energy required to
the production potential from different proppant liberate these fines, and the corresponding fracture
types and sizes. This will be conducted as a part of the length remains an imprecise science. Additionally, in
stimulation design optimization using the fracture design prolific wells, it is often difficult to provide enough
model. Typically, greater production benefit will be conductivity to effectively drain a long fracture
obtained with manufactured proppants compared with experiencing multiphase and non-Darcy effects.
natural sand proppants and with larger 16/20 or 12/18 In cases where fines migration cannot be eliminated,
mesh proppants. In this phase of the selection process, the size of the proppant should be determined using
the engineer should seek to understand for a given set the formation size distribution and appropriate sand
of reservoir conditions what potential benefit is obtained control criteria. This provides a bound for the smallest
from different proppant materials and sizes. proppant size to be selected.
Secondly, the above exercise should be conducted Applying resin-coating technology to either sand
with a range of proppant volumes/masses (see Sections or ceramic-based proppants has been used extensively
4-5 and 4-6). A small treatment that is constrained by to control both proppant and formation sand flowback.
fracture half-length penetration may not show much Over 70% of all resin-coated proppants are now used
productivity impact when changing proppant quality for applications to mitigate or minimize solids flowback.
and size. A larger treatment that might be expected to The resin chemistry is typically formulated for specific
produce at significantly higher rates could be extremely markets in order to provide the grain-to-grain bonding
sensitive to proppant quality or size. strength necessary for the pressure and temperature
As discussed in the text of this chapter, it is critical in conditions in a particular reservoir.
gas wells to address the following issues:
Non-Darcy and multiphase flow. If the fracture 8-8.3 Availability
design model does not include these effects, the
engineer should use a proppant characterization With over a hundred different proppant types and sizes
model to determine an appropriate reduction for the available to the industry, there is no fracturing market
API/ISO reference conductivity values incorporated in which all types and sizes of materials are readily
into the fracture design models as lookup tables. available. It is therefore necessary for the engineer to
This can then be manually incorporated into the understand what materials are typically available in a
317
Modern Fracturing
region/market and what are the availability and added measured proppant data, but be prepared to evaluate/
costs for other materials that may be considered in the correlate the treatment performance to expectations and
design/optimization phase. In offshore environments then, if necessary, to change the proppant type, size and
where frac boats are used, and in some international mass of material pumped in order to further improve
markets, there are often limitations on the available the performance of the hydraulic fracturing process
space for proppants both total volume and storage in future field completions.
of different types of proppant. The engineer should be
knowledgeable about what products are being used and
what products can be logistically handled. References
8-8.4 The Cost-Value Proposition American Petroleum Institute: Recommended

Practices for Evaluating Short Term Proppant Pack
Proppant cost will always be a significant component Conductivity, API RP 61, Oct 1989.
of the total treatment cost. In a small sand proppant American Petroleum Institute: Recommended Practices
treatment, the proppant may account for only 10 to for Measurement of Properties of Proppants Used
20% of the total treatment cost. In a large treatment in Hydraulic Fracturing and Gravel-Packing
utilizing man-made manufactured proppants, it is Operations, API RP 19C, 1st edition, 2007.
typical for the proppant cost to be in excess of 50% of American Petroleum Institute: Recommended
the total treatment ticket. Practices for Testing High Strength Proppants
Fracture design models produce a number of Used in Hydraulic Fracturing Operations, API
job size optimization plots in order to guide the RP 60, Dec 1995.
engineer toward the correct proppant selection and American Petroleum Institute: Recommended
schedule. Such plots which should be calculated Practices for Testing Sand Used in Gravel Packing
for each proppant type and mesh size being Operations, API RP 58, Dec. 1995.
considered include: American Petroleum Institute: Recommended
Treatment size (proppant mass) vs Practices for Testing Sand Used in Hydraulic
fracture half-length Fracturing Operations, API RP 56, Mar 1983.
Productivity vs. time for different treatment sizes Barree, R.D. and Conway, M.W.: Experimental and
Cumulative production (recovery) vs. time for Numerical Modeling of Convective Proppant
different treatment sizes Transport, paper SPE 28564, 1994.
Net Present Value vs. treatment size Biot, M.A., and Medlin, W.L.: Theory of Sand
Incremental Net Present Value vs. treatment size Transport in Thin Fluids paper SPE 14468,
Investment efficiency (NPV/Cost) vs. treatment size 1985.
Brannon, H.D., and Pulsinelli, R.J.: Evaluation of
The use of such plots, coupled with other Breaker Concentrations Required To Improve
appropriate financial metrics, should guide the engineer the Permeability of Proppant Packs Damaged by
to the correct proppant design. Hydraulic Fracturing Fluids, paper SPE 19402,
It should always be remembered that we are using 1990b.
laboratory data measured at specific conditions and Brannon H.D. and Pulsinelli, R.J.: Breaker
applying these data to a heterogenous environment in the Concentrations Required to Improve the
reservoir, which we do not completely understand. Many Permeability of Proppant Packs Damaged by
engineers have discovered significant financial benefit from Concentrated Linear and Borate Crosslinked
changing proppant types and treatment sizes from those Fracturing Fluids, paper SPE 20135, 1990a.
previously used in a particular fields development. We Brannon, H.D., Malone M.R., Rickards, A.R.,
encourage the practicing engineer to do likewise: to apply Wood, W.D., Edgeman, J.R., and Bryant, J.L.:
our best understanding of fluid mechanics and laboratory Maximizing Fracture Conductivity with Partial
318
Monolayers: Theoretical Curiosity or Highly Economides, M.J., Watters, L.T., and Dunn-Norman,
Productive Reality, paper SPE 90698, 2004. S.: Petroleum Well Construction, John Wiley and
Brannon, H.D., Rickards, A.R., and Stephenson, Sons, 1998, p 474.
C.J.: Lightweight methods for well treating, US Edgeman, R., Gertsner, M., Nelson, S., and Malone,
6,364,018 (2002). M.: Lightweight Proppant, A New Innovation
Brannon, H.D., Wood, W.D., and Wheeler, R.S.: in Hydraulic Fracturing, paper presented at the
A New Correlation for Relating the Physical 51st Annual Southwest Petroleum Short Course,
Properties of Fracturing Slurries to the Minimum Lubbock, Texas, 2004.
Flow Velocity Required for Transport, paper SPE Fitzgibbon, J.J.: Sintered Spherical Pellets Containing
106312, 2007. Clay as a Major Component Useful for Gas and
Brannon, H.D., Wood W.D., and Wheeler, R.S.: Oil Well Proppants, U.S. Patent No. 4,427,068
The Quest for Improved Proppant Transport: (1984).
Investigation of the Effects of Proppant Slurry Forcheimer, P.: Wasserbewegung durch Boden, ZVDI
Component Properties on Transport, paper SPE (1901) 45, 1781.
95675, 2005. Gidley, J.L., Holditch, S.A., Nierode, D.E., and
Card, R.J., Howard, P.R., and Feraud, J-P.: A Novel Veatch, R.W., Jr.: Recent Advances in Hydraulic
Technology to Control Proppant Back Production, Fracturing, Monograph Series, SPE, Richardson,
paper SPE 31007, SPEPF (November 1995), 10:4, TX (1989) 12, 109-130.
271-276. Goel, N., and Shah, S.: A Rheological Criterion for
Cheung, S.: Effect of Acids on Gravels and Proppants, Fracturing Fluids to transport Proppant During a
paper SPE 13842, 1985. Stimulation Treatment, paper SPE 71663, 2001.
Cinco, L.H., Samaniego, V.F., and Dominquez, A.N.: Graham, J.W., Muecke, T.W., and Cooke, C.E. Jr.:
Transient Pressure Behavior for a Well With a Method for Treating Subterranean Formations,
Finite Conductivity Vertical Fracture, SPEJ (Aug. U.S. Patent No. 3,929,191 (1975).
1978) 253-264. Hoaberg, R.K., and Koerner-Moore, J.: Silica-Sand
Clark, P.E., and Zhu, Q.: Fluid Flow in a Vertical Proppants Used in Well Stimulation, paper SME-
Fracture from a Point Source, paper SPE 28509, 80-345, 1978.
2004. Howard, G.C., and Fast, C.R.: Hydraulic Fracturing,
Cobb; S.L., and Farrell, J.J.: Evaluation of Long-Term Monograph Series, SPE, Richardson, TX (1970)
Proppant Stability, paper SPE 14133, 1986. 59-90.
Cooke, C.E.: Conductivity of Fracture Proppants in ISO 13503-2, Petroleum and Natural Gas Industries
Multiple Layers, paper SPE 4117, JPT (Sep 1973) - Completion Fluids and Materials: Part 2:
1101-1107; Trans. AIME, 255. Procedures for measuring properties of proppants
Cooke, C.E. Jr., Hedden, W.A., and Chard, W.C.: used in hydraulic fracturing and gravel packing
Hydraulic Fracturing Method Using Sintered operations, 1st edition, 2006.
Bauxite Proppant Agent, U.S. Patent No. Krumbein, W.C., and Schloss, L.L.: Stratigraphy and
4,068,718 (1978). Sedimentation, 2nd Ed., 1963, W.N. Freemen and
Darcy, H.: Au Filtrage DEau, published in Les Co., New York.
Fontaines Publiques de la Ville de Dijon, Paris, Jin, L., and G.S. Penny: A Study on Two Phase, Non-
1856. Darcy Gas Flow through Proppant Packs, paper
Darin, S.R., and Huitt, J.L.: Effect of a Partial SPE 49248, 1998.
Monolayer of Propping Agent on Fracture Flow Johnson, C.K., and Armbruster, D.R.: Particles
Capacity, paper SPE 1291, (1959); Trans., AIME Covered with a Cured Infusable Thermoset Film
(1960) 219, 31-37. and Process for Their Production, U.S. Patent No.
Economides, M.J., Oligney, R., and Valko, P.: Unified 4,439,489 (1984).
Fracture Design, Orsa Press, 2002. Lunghofer, E.P.: Hydraulic Fracturing Propping
319
Modern Fracturing
Agent, U.S. Patent No. 4,522,731 (1985). 1999.

Nolte, K.G.: Flow Considerations in Hydraulic Stim-Lab Proppant Consortium Report, 1992.
Fracturing, paper SPE 18537, 1988. Stim-Lab Proppant Consortium Notes, 1997.
Novotny, E.J.: Proppant Transport, paper SPE 6813, Stim-Lab Proppant Consortium Notes, 2000.
1977. Tannich, J.D., and Nierode, D.E.: The Effect of
Nguyen, P.D., Weaver, J.D., and Dewprashad, B.T.: Vertical Fractures on Gas Well Productivity, paper
Surface Modification System for Fracture SPE 15902, 1986.
Conductivity Enhancement, paper SPE 48897, van Poolen, H.K., Tinsley, J.M., and Saunders, C.D.:
1998. Hydraulic Fracturing: Fracture Flow Capacity vs.
Ouabdesselam, M. and Hudson, P.: An Investigation Well Productivity, Trans. AIME (1958) 213, 91-
of the Effect of Cyclic Loading on Fracture 95.
Conductivity, paper SPE 22850, 1991. Vincent, M.C., Pearson, C.M., and Kullman, J.: Non-
Pearson, C.M., Bond, A.J., Eck, M.E., and Lynch, Darcy and Multiphase Flow in Propped Fractures:
K.W.: Optimal Fracture Stimulation of a Case Studies Illustrate the Dramatic Effect on Well
Moderate Permeability Reservoir - Kuparuk River Productivity, paper SPE 54630, 1999.
Unit, Alaska, SPEPE (August 1992) 259-266. Vreeburg, R.-J., Roodhart, L.P., Davies, D.R., and
Penny, G.S., and Jin, L.: The Development of Penny, G.S.: Proppant Backproduction During
Laboratory Correlations Showing the Impact of Hydraulic Fracturing A New Failure Mechanism
Multiphase Flow, Fluid, and Proppant Selection for Resin-Coated Proppants, paper SPE 27382,
Upon Gas Well Productivity, paper SPE 30494, JPT (October 1994), 46:10, 884-889.
1995. Wood, W.D., Brannon, H.D., Rickards, A.R., and
Prats, M.: Effect of Vertical Fractures on Reservoir Stephenson, C.J: Ultra-Lightweight Proppant
Behavior - Incompressible Fluid Case, SPEJ (June Development Yields Exciting New Opportunities
1961) 105-18; Trans., AIME, 222. in Hydraulic Fracturing Design, paper SPE
Predict K, v.7.0, Stim-Lab, 2007. 84309, 2003.
Rickards, A., Lacy, L., Brannon, H., and Stephenson,
C.: Need Stress Relief ? A New Approach to
Reducing Stress Cycling Induced Proppant Pack
Failure, SPE Paper 49247, 1998.
Rickards, A.R., Brannon, H.D., Wood, W.D.,
and Stephenson, C.J.: High Strength, Ultra-
Lightweight Proppant Lends New Dimensions
to Hydraulic Fracturing Applications, paper SPE
84308, 2003.
Roberts, S.S., Binder, J.S., Lane, R.H.: Strength,
Volume and Weight Loss of Gravels and Proppants
Due to HF-Based Acids, paper SPE 30168,
1990.
Stephenson, C.J., Rickards, A.R., and Brannon, H.D.:
Exceptional Proppant Flowback Control for the
Most Extreme Well Environments: The Shape of
Things to Come, paper SPE 77681, 2002.
Stephenson, C.J., Rickards, A.R., and Brannon, H.D.:
Increased Resistance to Proppant Flowback by
Adding Deformable Particles to Proppant Packs
Tested in the Laboratory, paper SPE 56593,
320
Mark Malone is region technical manager for the Permian Basin at BJ Services Company. Malone has held
various engineering and sales positions during his 20 years in pressure pumping services, working in the
Denver-Julesburg Basin of Colorado, Raton Basin of Northern New Mexico and the Permian Basin of West
Texas and southeastern New Mexico. Malones most recent focus has been on the practical application of new
pumping services technologies. Malone is an active member of the Board of Directors of the Southwestern
Petroleum Short Course, an SPE member, author of multiple technical publications and patent holder. He
has a bachelor's degree in agricultural engineering from Texas Tech University.
John W. Ely is a principle in Ely & Associates, Inc., which he started in 1991 with three partners. He has a
bachelor's degree in chemistry from Oklahoma State University. He started his career with Halliburton Co.
in 1965, working as a technician for the analytical group while completing his college work and returning
for many roles from 1973 until 1980. Ely joined Nowsco Services in 1980 as engineering manager. In 1985,
he joined S.A. Holditch & Associates as vice president of stimulation technology. John holds several patents
and has numerous publications, including the Stimulation Engineering Handbook. He is also a contributing
author to the SPE monograph on hydraulic fracturing, writing the chapter on hydraulic fracturing fluids
and fracture fluids selection.
Chapter 9 allow for reasonable pumping pressures and still
maintain viscosities measured in 1000s of centipoise at
Execution of Hydraulic bottomhole temperatures greater than 350 F. Fracture
Fracturing Treatments design engineers have a tremendous number of options
Mark Malone, BJ Services and when designing fracture treatments. They can choose
from a very simple "waterfrac" fluid that may be only
John W. Ely, Ely & Associates
water and a friction reducer to an extremely complex
system comprising high-temperature stable guar
9-1 Introduction derivative, a metal ligand crosslinking agent, a buffering
agent, temperature stabilizer, breakers, surfactants, and
The industry has come a long way in its ability to a plethora of additional additives to fit the lithology of
properly execute hydraulic fracture treatments. As of the reservoir that is treated. (See Chapter 7 for more
the writing of this chapter hydraulic fracture treatments details about fracturing fluids.)
have been executed for more than 60 years. The very Although the first proppant used was sand from
first treatments were conducted using low-horsepower a river bed, today there are a vast number of choices
cementing units, and the mixing equipment consisted ranging from high-quality sand to resin-coated sand to
of a ground cement mixing tub and human beings ceramics and finally to the high-strength bauxite used
cutting sacks of sand into these tubs as the fluid was for closure pressures exceeding 10,000 psi. (See Chapter
pumped downhole. All personnel and any monitoring 8 for more details on proppants.)
equipment were outside, subjected to whatever the In addition to all the improvements in pumping
environment happened to be. It is very fortunate that equipment, metering capabilities and monitoring
the first well to be hydraulically fractured was highly software, one of the greatest improvements in
damaged; otherwise the onset of hydraulic fracturing execution was the introduction of pre-job diagnostics
would have been delayed considerably. on virtually every treatment. These calibration tests
From these humble beginnings the industry has (see Sections 4-2.1.8 and 4-2.1.9) allow the engineers
progressed to the point where we now have fracturing to know key issues such as number of perforations,
pumping units that yield in excess of 3000 hydraulic closure gradient, tortuosity and, most importantly,
horsepower. It should be noted that for many years fluid efficiency. By knowing fluid efficiency pre-job
fracturing pumps were controlled using a single individual the pad volume percentage can be adjusted to allow
sitting on the pump. Today we have multiple pumping for complete placement of proppant. By diagnosing
units controlled by one individual sitting next to the tortuosity or lack thereof the on-site personnel can
supervisor in the control van. Getting the personnel off take steps to remove near-wellbore problems before
of the units is not only a tremendous move in relation disrupting the placement of proppant.
to safety but also allows for much better control of the Obviously we have come a long way in having
fracturing treatment. For the first treatment, the fluids tools and technology to execute treatments. The
were mixed in the measuring tanks of the cement unit industry has also progressed tremendously in the area of
using a shovel. We now have computer-controlled blenders quality control, i.e., assuring both themselves and the
with microprocessor-controlled additive pumps. It is not customer that the job was pumped as designed. Today
uncommon for 8 to 10 additives to be run continuously there are very few treatments where the actual fluid
on a high-rate fracture treatment with all additives not to be pumped is not tested at reservoir conditions of
only controlled with microprocessors but also monitored temperature and shear prior to the treatment. Specific
through coriolis meters and the readout of volume and concentrations of gel breakers, etc., are selected based
rate shown in an air-conditioned van. on the actual performance of the fluids on the wellsite.
From the very first fluids, which consisted of In addition there is a very concerted effort by the service
World War II surplus napalm or simply ungelled companies to pump the job as designed, allowing for
hydrocarbons, we now have complex fluids that better post-job evaluation of performance.
323
Modern Fracturing
9-2 Function of Equipment went different directions in designing this type of

equipment. One service company made a commitment
The fracturing industry has come a long way in relation to utilize gas turbine units. These units were typically
to the pumping and monitoring equipment used for around 1000 HHP and had the advantage of low
hydraulic fracturing. From the time of the first fracturing weight to horsepower. This type of unit was used for
treatment with low-horsepower cementing trucks and more than 25 years before being phased out by more
when sand was added by sack cutters into a ground conventional equipment.
jet mixing tub (as seen in Fig. 9-1). Today, specialized
high-horsepower pumping and blending equipment is
available from a large number of suppliers. We also have
come from supervising fracturing treatments in the open
air with simple pressure and rate gauges to computerized
control vans where virtually every pressure, rate, and
chemical is monitored. In these vans it is typical that all of
the pumping and blending units are controlled, allowing
for a much safer and managed environment. It is quite
common for all of the data monitored in the treating
vans to be transmitted back to both the service company Figure 9-2 A pumping unit rated at 2700 HHP
office and the customers office or to any computer with
Internet access anywhere in the world. The major reason for the demise of the gas turbine
was their cost and the need for very specialized
maintenance. After World War II a very large number
of surplus Allison airplane engines were used for
fracturing equipment. Their use was primarily due to
their very low cost and availability. These units were
used until the surplus was expended.
Figure 9-1 Vintage ground jet mining tub in early

fracturing treatments
9-2.1 High-Pressure Pumping Equipment
As mentioned previously, the earliest pumping units

for hydraulic fracturing were typically cementing units
with less than 500 hydraulic horsepower. In addition to
the low horsepower, those units were typically limited Figure 9-3 Common fracturing unit from the 1950s
to pumping at pressure less than 5,000 psi.
Today the most common specialized pumping From the mid 1950s to the 1980s, the most common
unit is rated at 2,000 hydraulic horsepower and fracturing unit was a trailer or body load unit with
can work at pressures exceeding 15,000 psi. There two 500 HP pumps and twin 500 to 600-BHP drive
are specialized units with horsepower ratings engines as seen in Figure 9-3. This type of pumping unit
to 3000 HHP (see Figure 9-2). is still in use today in many areas due to its reliability
For a long time cementing pumps were adequate and functionality. As the industry was able to design
for fracturing treatments, but gradually the need for higher-horsepower pumps there was and continues to be
higher-horsepower, specialized pumping equipment a movement towards single-trailer-mounted fracturing
prevailed, and the various service companies trucks with a 2000 to 2500 BHP engine and a single
324
Chapter 9 Execution of Hydraulic Fracturing Treatments
pump with varying size fluid ends, allowing for high-rate To be able to conduct these treatments, all service
pumping at low pressures and low-rate pumping at high companies built long-stroke pressure multiplier pumps,
pressures. Table 9-1 is a typical rate horsepower table for which could pump at very high pressures for long periods
various fluid end sizes with a 2000 HHP pump. of time (see Fig. 9-4). Movement of the industry to casing
completions and to improvements in conventional
Table 9-1 Rate/Horsepower Table for 2000 HHP Pump equipment has led to these fleets of intensifiers being
Fluid End Size Maximum Rate Maximum Pressure phased out. There is a possibility that as we move to
Inches bpm psi
deeper and higher pressure reservoirs that we may have
4 12.6 14,147
to go back to this type of technology.
5 15.6 11,459
5 18.8 9,470
9-2.2 Blending Equipment
5 20.6 8,664
6 22.4 7,957
The blending of sand and chemicals in early fracturing
treatments was done typically by batching into tanks or
During the late 1960s and through the 1980s there adding materials manually. The very first blenders were
was a great deal of use of intensifier pumps for treating primarily a suction pump, discharge pump and a tub
at pressures between 10,000 and 20,000 psi. Because where sand and other materials were dumped (see Fig.
a large number of deeper wells were completed with 9-5). Modern blenders have microprocessor-controlled
small tubulars, there were a plethora of jobs requiring additive pumps, radioactive densiometers and precise
pumping pressures exceeding the ratings of conventional dry additive systems (see Fig. 9-6). These units are
pumps. In the late 1960s it was not uncommon for a typically controlled from the treating van, where the
fairly sizable fleet of conventional pumping equipment proppant and all additives are monitored. On larger
to be literally destroyed trying to pump for long periods treatments, a chemical additive truck or trailer, as
of time at pressures exceeding 12,000 psi. The state- shown in Fig. 9-7 is used. It contains chemical totes
of-the-art equipment at that time were very functional and metering pumps as well as precise flow meters that
if used at 60% of horsepower or pressure rating but are used to feed chemicals to the suction side of the
simply could not be used for any substantial period at discharge pump of the blender.
maximum pressure or horsepower.
Figure 9-5 Example of an early blender
Figure 9-4 Intensifier pump used on a treatment in Texas Figure 9-6 Modern fracture stimulation blender
325
Modern Fracturing
It should be noted that there are typically

check valves, plug valves or both between each
pumping unit and the manifold; there are also
check valves and plug valves isolating the manifold
from the wellhead.
Figure 9-7 Chemical additive unit
While equipment has improved, there also has

been improvement in the addition of powdered
products as slurries or suspensions. The industry
has come full circle in the addition of viscosifying
polymers using suspensions in hydrocarbons, but now Figure 9-8 High-pressure treating line truck to assist in
there is a trend to go back to dry materials and away rigging up
from hydrocarbon-suspending agents.
9-2.3 High-Pressure Treating Lines and Manifolds
In the early days of hydraulic fracturing, there was a large

emphasis on hiring very large and very strong personnel
to be able to rig up the treating lines on a location. Today
every service company has a high-pressure treating line
truck with a crane to assist in rig-up and typically has
an additional crane to assist in the rig-up of the treating
lines to the wellhead (see Fig. 9-8).
The major improvements have come in the area of
properly designed manifold trailers, which greatly assist
the personnel in rig-up (see Fig. 9-9). This manifold
Figure 9-9 Manifold trucks to help minimize lines and
which is parked between two lines of pumps minimizes discharge hoses
the lines to each truck and the number of discharge
hoses to each truck. The blending unit is usually 9-2.4 Nitrogen and Carbon Dioxide Pumping
connected to the manifold trailer by these discharge
hoses and the low-pressure manifold on the top of The industry was well into the 1960s before
the manifold trailer feeds all pumps. Conversely, the the introduction of pumping nitrogen (N2) or
high-pressure pumps are connected to the lower high- carbon dioxide (CO2). N2 and CO2 were initially
pressure treating lines at the bottom of the manifold, introduced as energizing media to help unload low-
and this manifold is connected to the wellhead via high- pressure wells, but in the late 1970s the use of both
pressure lines. As can be seen from Fig. 9-8 this allows foamed CO2 and N2 foam became commonplace
a minimum amount of treating lines, thus minimizing in the industry. There have been quantum leaps in
the chances for leaks and greatly assisting the personnel both the technologies in using foams and in the
in time and effort to rig up. equipment to pump these systems.
326
Early on, specialty companies furnished N2 inherent danger for potential solidification of the
pumping equipment and worked with the service CO2 as dry ice. Typically the suppliers of CO2 will
companies on foam fracs and energized treatments. provide storage vessels on location and additionally
Today virtually every major service company has its own (and most importantly) a booster pump to pressurize
N2 pumping equipment. Early N2 pumping equipment the fracturing pumps. Just as the N2 pumps have to
typically consisted of fairly low-rate equipment capable be cooled down to pump the N2 as a liquid, the CO2
of pumping only 2200 scf/minute. This required must be vented through the pumps such that the
sometimes 30 to 40 N2 trucks and numerous transports CO2 becomes liquefied and pumpable (see Fig. 9-12).
to conduct even moderate-sized foam frac as shown in Once the pumps are cooled down, signified by a large
Fig. 9-10. Today there are high rate N2 pumpers (Fig. layer of frost on the fluid ends and iron, the venting
9-11) have horsepower and pressure ratings similar to ceases and the CO2 is pumped down hole, mixed in
the high horsepower frac pumps. precise fashion with water or oil-based fluids.
Figure 9-10 Early N2 equipment on a moderate-size

treatment
Figure 9-12 CO2 pumps on location
Probably the most significant improvement in

energized foams, particularly in the case of N2, has
been the development of accurate means of measuring
gas flow rate. CO2 can be metered with a conventional
turbine meter, but N2 requires specialized temperature
corrected turbines or, typically, coriolis or target
meters. In the early days of foam fracs, the only means
of measuring rate was by using stroke counters, and
there was little real accuracy to measure downhole
foam quality. By utilizing computer control systems
with an accurate bottomhole pressure, the quality of
Figure 9-11 Modern N2 pump trucks foam fracs can be accurately controlled.
At approximately the same time as N2 was 9-2.5 Treatment Control Vans and Cabins
starting to be used as a foam, the use of CO2 as a
foaming agent was introduced. Pumping CO2, while Until well into the 1970s the standard practice in
not requiring specialized pumps like N2, does require the industry for hydraulic fracturing was to conduct
specialized high-pressure suctions on conventional the treatment with personnel out in the open air.
frac pumps and experienced operators due to the Typically the treater or the person in charge of the
327
Modern Fracturing
treatment would have a pressure gauge and some sort non-functional. Many operators request in their bids
of device to measure rate. As the industry progressed that they will require at least 50% standby pumping
electrical boxes were built with a strip chart with equipment. On location they find that there is barely
rate and pressure recorded. In the late 1970s service enough horsepower to do the treatment. Although the
companies began to introduce treatment vans (or equipment may be rated at a certain horsepower, if
cabins), which were typically of a square bread truck frequent maintenance has not been performed, the actual
configuration where the treater and engineer would deliverable horsepower may be much less.
sit with the company man and one could monitor The competent oil company representative should
pressure, rate and (sometimes) density of the fluid sit down with the service company supervisor and ask
while it was being recorded on a strip chart. not only what horsepower equipment is available but
With the advent of Nolte-Smith plots where also find out the pressure rating of the equipment and
net pressure data were monitored (see Section 4- the plunger sizes on the pump fluid ends. For instance
1.2.7) the industry introduced computer vans and if one has 10 pump trucks with small plungers capable
cabins that were typically on site in addition to the of only pumping 12 bbl per minute (bpm) per pump,
treatment vans. With the advent of quality control it is impossible to pump a treatment at 150 barrels per
(QC) efforts in the early 1980s requiring a great deal minute regardless of the pressure. Conversely if 20,000
of real-time monitoring of chemical additives, there hydraulic horsepower is on location but the trucks have
is now a single treatment van or cabin where many large fluid ends incapable of pumping at 10,000 psi, a
variables are monitored and plotted on flat screens. high-pressure treatment is doomed to failure.
Such treatment vans have not only made hydraulic All service companies will supply pressure and rate
fracturing more comfortable but have allowed for a curves for their equipment, and a diligent oil company
much better understanding of what is going on both supervisor should work closely with the service company
downhole and on the surface. supervisor to be sure sufficient operating equipment is
available for the job at hand. It is always a good idea to
9-3 Equipment Quality Control have all the available equipment on a treatment on-line
(primed) and pumping. The equipment used everyday
For the typical operator the question of quality by the service companies has a very long life if pumped
control of pumping and blending equipment is a at either 60% of pressure rating or 60% of maximum
moot issue. The normal response is that the treating horsepower. The vast majority of pumping equipment
lines, pumping equipment and blending equipment in the industry has a very short life if used at maximum
are the responsibility of the Service Company and pressure and maximum horsepower. A common analogy
therefore no real effort should be made in taking would be running the family car at 120 miles per hour
the time to look over and evaluate what is brought pulling a heavy trailer. The life of the car would be very
to location. In reality there are a huge number of short, as is the case with fracturing equipment.
reasons why the operator should not only notice One should also query the Service Company as to
the nature of pumping and blending equipment the pressure rating of the treating lines they are using.
but should become knowledgeable in various Typically the treating line is rated at 6000, 10,000,
aspects of the mechanical operation. 15,000, or in rare cases 20,000 psi. When dealing with a
major service company, this is normally not a problem,
9-3.1 How Much Horsepower but with small operators it is not uncommon to find
and What is the Pressure Rating? connections and fittings that were made in a machine
shop and have not been properly tested. The pressure
In boom times in this industry it is not uncommon rating is usually shown on the wing of the knock-up
for people and equipment to be pushed to their limit. connection. If concerned about the treating line integrity,
In some areas much of the pumping equipment never always test at 1000 psi above maximum pressure,
returns to the yard or maintenance shop until it is assuming the pressure rating of the iron is not exceeded;
328
after one successful pressure test have the service company Table 9-3 Suction discharge hose requirements
take the pressure back to maximum a second time. This Type Fluid Type Requirement
usually allows the detection of any fatigue or stress cracks 1 ea. 20 X 4 per
Thin Fluids
present in the treating iron. 10 bpm
Suction
1 ea. 20 X 4 per
Thick Fluids
5 bpm
9-3.2 How Many High-Pressure Lines All Fluids
1 ea. 20 X 4 per
10 bpm
and Suction Discharge Hoses to Use? Discharge
Use 3 hoses if rate
All Fluids
per pump < 5 bpm
Tables 9-2 and 9-3 show guidelines for the size and
number of high-pressure lines to be used based on the 9-3.3 Standby Pumping and Blending Equipment
treating rate. Insufficient suction hoses will result in
starving pumping equipment (cavitation) and can create It is always a good idea to have a discussion with the
a great deal of potentially hazardous vibration. supervisor for the fracturing company and find out
Exceeding critical velocity on treating iron is much just how much of the equipment is truly standby.
more dangerous. The erosion caused by pumping at Always have at least one spare pump or standby
excessive rates through treating iron can create well pump on location. All of the pumping equipment
control situations with abrasive material cutting out must have a pressure rating compatible with the
swings and chicksans, any place the proppant laden maximum pressure anticipated. It is always a good
fluid is turning a corner, or any area of turbulence. idea to have all of the pumps on location running
It should be noted that these guidelines do not at low rate and horsepower rather than having
hold for fluids without proppant or for pumping some of the pumps idling while other pumps are
through wellbore tubulars. High pressures from pushed to the limit. Having all pumps on line
turbulence will be noted, but excessive erosion is not and moving fluid not only eases the strain on the
normally an issue in straight pipe. equipment, but also eliminates potential problems
Additionally, one should consider the reverse with bringing pumps on-line while pumping
situation, where very low rates occur with heavy high proppant concentrations.
proppant concentrations such as seen on low-rate For jobs of any length, it is always a good idea to
treatments. Service companies can install smaller have a standby blender. Losing a blender during any
3-in. discharge hoses to create higher velocity and stage of a fracturing treatment can be catastrophic.
also not utilize manifold trailers or other areas The key to the use of a standby blender is that it
where proppant can settle out and potentially is truly on standby, i.e., it can be brought on line
damage a pump by entering the suction side in high without delay. Service companies can rig the blender
concentrations. Typically these situations occur at such that one has to simply change valves and switch
fluid rates less than 5 bpm per pump. monitoring connections without disrupting the
treatment. Another obvious point is that the standby
Table 9-2 High-pressure iron recommendations blender should be rigged so that it is ready to accept
MAXIMUM RECOMMENDED PUMP RATE THROUGH FIG. proppant and all fluid additives.
1502 HIGH PRESSURE DISCHARGE LINES
Line Working Pressure Pump Rate (bbls/
(in.)
ID (in.)
(psi) min)
9-3.4 Absolute Essentials for Every Job
2 1.775 15,000 8
3 2.750 15,000 20 Typically, large companies have rules stating that
4 3.750 6,000 37 there must be at least two barriers to restrict flow to
Note: This table does not give set values for the absolute surface, negating a well control situation. One should
maximum rates for the treating iron. Exceeding these rates, always have an isolating valve, a block or a check
utilizing high sand concentrations, yields a short life, with severe
abrasion, for the treating iron. When excessive rates are used valve between every pumping unit and the iron or
without abrasive materials, high friction pressures result. manifold. Additionally there should be a check valve
329
Modern Fracturing
installed as near the well as possible. When using a the first step in assuring an effective stimulation. After the
tree saver, the hydraulic valve or valves on the tree fluid has been selected for its specific properties, ensuring
saver can be used as the blocking valve. these properties are present during pumping is key to the
For additional safety and control of the treatment it success of the fracture design.
is always a good idea on high-pressure wells to have both The purpose of a fracturing fluid is to carry proppant
a check and blocking valve isolating each pump. To have into the induced hydraulic fracture as efficiently as
to shut down a frac treatment because of a small leak on possible while limiting the amount of damage associated
one truck can be devastating. with this invasion of the reservoir rock. A given fluids
ability to carry proppant is a function of its rheological
Table 9-4 Wellhead protection tool ratings (courtesy of properties. The objective of quality control (QC) is the
Stinger Wellhead Protection, Inc.) application of these rheological fluid properties and the
Stinger Isolation Tool Pressure Readings ability to reduce the same properties after the treatment.
Max
Mandrel
Rate
Max Tubing or
Comments
Fluid quality control can be summarized as the process
I.D. Size Pressure Casing Size
- BPM of evaluating the materials and equipment needed on
2.125 52 10000
3-4 location to ensure the fracture fluid performs as designed.
-5
Fluid systems should be designed, agreed upon and tested
1 12 15000 2
1.25 18 15000 2
in the laboratory for particular properties including
1.5 26 15000 2 viscosity profiles at bottomhole temperature. On location,
1.687 34 15000 2 High rate the QC process should ensure these same properties
1.75 36 15000 3 exist in the fluid and that the necessary equipment is
1.86 42 15000 3 High rate available for blending, metering and pumping the fluid
2 48 15000 4
system and associated additives.
4 - 5 and
2.5 72 15000
up
2.75 88 15000 5 and up 9-4.1 Quality Control of Water-Based
3 106 15000 5 and up Fracturing Fluids Before Arriving on Location
3.5 142 15000 5 and up
4 188 15000 7 and up The first step in the QC process is to gather samples of
5.063 298 15000 8 and up Stage Frac
all chemicals and water that are to be used in preparing
1.687 34 20000 3
the selected fracture fluid system. Samples from the water
2.25 52 20000 4 and up
source should be obtained and tested to ensure the water
Another very important and critical issue is to quality is acceptable before the water is transported to
establish the maximum rate allowed for the wellhead location. Each fluid system may have acceptable water
isolation tool (for example, see Table 9-4). If you exceed properties that vary from one another. Table 9-5 outlines
this rate the tool becomes a very good cutting device result levels for a standard water analysis, which should
and can sever the tubulars or casing. A common reason be acceptable for most crosslinked fluid systems.
for this is having the wrong mandrel in the tool, i.e.,
having a 2 in. mandrel in a 3 in. tool. Again always Table 9-5 Acceptable levels in water analysis
query the tool operator as to the particular tool to be Acceptable Levels in Water Analysis
utilized on your well site. pH 6-8
Iron <10 ppm
9-4 Quality Control for Oxidizing Agents None
Fracturing Fluids Reducing Agents None

Carbonates < 300 ppm
There are a multitude of fracturing fluids available to the Bicarbonates < 300 ppm
industry today. The selection of the appropriate fluid has Bacteria None
Cleanliness Reasonable
been discussed in Chapter 7, and this selection process is
330
Bacteria exist in all fluids, including waters that Water-based fracturing systems can be divided into
have been recently treated with biocides. Bacteria two basic fluid systems: linear and crosslinked. Mixing
levels in an acceptable water source should not reduce a polymer into water produces linear fluid systems,
base gel viscosity. However, a quick quantitative which are also the basic building block of crosslinked
analysis for bacteria in a fracturing fluid is often not fluids. Fracturing fluid polymers are usually produced
practical. A bacterial spectrometer (Keller and Segal, by the use of guar gelling agents or some derivative
1971), used to measure bacteria levels, may be present thereof, as described in Chapter 7. Fig. 9-13 indicates
in some laboratories, but probably will not be available the linear relation between fluid viscosity and the
for field use. Still, the presence of bacteria in a water addition of polymer in pounds per 1000 gallons of
sample must be determined. water at various temperatures.
A quick check for bacteria may be performed by
48
gelling (adding the polymer to the water and mixing
44
to ensure polymer distribution) a sample of water with 40 35 PPT
Viscosity at 511 sec-1, cp

the polymer to be used for viscosifing the water and 36
32
subsequently measuring the initial base gel viscosity 30 PPT
28
of the gelled fluid sample. The gelled fluid sample is 24 25 PPT
then allowed to stand for some predetermined time 20 23 PPT
interval, usually at ambient temperature. If the sample 16 20 PPT
12
loses viscosity during the test period, then the presence 8
15 PPT
of detrimental levels of bacteria is probable, since 4

50 55 60 65 70 75 80 85 90 95 100 105
most fracturing fluid polymers are organic in nature
and provide an excellent food source for bacteria. The Temperature, F
REMARKS:
bacteria present will consume the polymer that has been Fann 35 (R1Spring / B1 Bob) @ 300 rpm.
ADDITIVES: Water & Guar Polymer
added to the water, thereby reducing the viscosity of the WATER SOURCE: Tap Water
gelled fluid and producing a slight odor of hydrogen
sulfide (H2S). If bacteria are determined to be present Figure 9-13 Linear gel viscosity vs temperature for
in detrimental levels, then the water must be treated for various polymer loadings
bacteria or a new water source located.
After the water source has been tested and is Crosslinked fluid systems use metallic and non-
acceptable, samples of the chemical additives to metallic crosslinkers in acidic or basic pH environments
be used for the fracture fluid system should be to produce an ionic bond between polymer strands.
obtained and the chemicals selected for use should This bonding of polymer strands can create a fluid with
be isolated. In other words, the chemicals that have very high viscosity. The quality control of a crosslinked
been sampled should be placed aside from other fluid system begins with a polymer-enriched base fluid
chemicals in the service company warehouse or or linear fluid system. The linear fluid system is tested
facility and held for the specific treatment. This will to ensure the viscosity is in an acceptable range based
ensure that the exact chemicals that passed the QC on the amount of polymer added to the water and the
testing are actually used during the job for which the temperature of the water, given that the temperature
QC testing has been performed. This is important affects the viscosity as illustrated in Fig. 9-13.
to ensure that the performance of the fracture fluid The viscosity of fluid is indicative of the hydrated
chemicals is acceptable prior to arrival on location. polymer concentration and is fundamental in
Although the practice of isolating chemicals for a the development of the crosslinked fluid system.
given treatment is not used everywhere, it can serve Insufficient polymer yields a crosslinked fluid system
to eliminate one variable when troubleshooting fluid that can be unstable at reservoir temperature; too
problems on location and is therefore recommended much polymer may create such high viscosity that
on jobs of a critical nature. the subsequent crosslinked fluid system becomes
331
Modern Fracturing
difficult to break at low reservoir temperatures. temperature application range, the rheology testing
Excessive polymer loading may also cause excessive is performed to ensure that an adequate fluid break
friction pressure, large enough to cause a premature can be obtained in a reasonable time period upon
pressure-out of the treatment when pressure completion of the designed fracture treatment pump
limitations of the tubulars are met. time. Almost all fluid systems will eventually break
The approved water source sample and given enough time exposed to reservoir pressure
fracture fluid additives are then mixed as per the and temperature, although shorter break times are
specifications of the selected fracture fluid system preferred because they reduce the time and cost
(provided by the fracturing company) and tested necessary to recover the treatment fluids. However,
at simulated downhole conditions. This simulation initial production in low-temperature wells can be
can only be performed in a laboratory environment negatively effected if the fluid break is not adequate
with the use of specific fluid viscometers. The two and timely (Malone et al., 2000).
most widely used high-temperature/high-pressure In reservoirs with a BHST greater than 140 F,
(HTHP) viscometers in use today are the model maintaining adequate fluid viscosity and stability
Fann-50 and the Chandler model 5550 HTHP become the objective. A number of rules of thumb
viscometer (see Fig. 9-14). These viscometers are have been established for crosslink viscosity during
used to measure the viscosity of the fluid system in a fracture treatment. As an example, some suggest
centipoise under pressure and the bottomhole static that 60% of the original crosslinked viscosity
temperature (BHST) of the reservoir. should be maintained at the end of the pump time
of each stage. However, this is an arbitrary value.
Crosslinked viscosity is not equivalent for every fluid
system, and each fluid systems ultimate viscosity will
vary based on water quality, BHST, pH, chemistry
and a host of other factors. It is not necessary to
require a particular viscosity, but understand that
a practical viscosity range should be expected and
that a complete break of the final stage of the fluid
system should be achievable within 2 to 4 hours after
pumping the fracture treatment.
Figure 9-14 Fann-50 (left) and Chandler 5550 HTHP

viscometers (right)
There are two reasons for the reservoir simulation

of the fluid system: (1) To determine whether the
proposed crosslinked fluid system can generate and
maintain sufficient viscosity to effectively transport Figure 9-15 Fann-35 (left) and Ofite Model 900 (right)
viscometers
proppant into the reservoir for the duration of
the designed fracture stimulation treatment, and After the base water has been tested and accepted,
(2) To determine if the fluid systems viscosity the fluid is gelled to the appropriate base gel viscosity
can be reduced significantly after the fracture as measured on a linear viscometer. Usually, a Fann-
for the effective fluid recovery. 35 or Ofite Model 900 fitted with a B1 bob and an R1
Depending upon the application temperature, rotor is used to measure linear viscosity, in centipoise,
the objective of testing fluids can vary. At BHST at a rotational speed of 300 rpm or 511 sec-1 shear
less than approximately 140 F, fluid stability rate (see Fig. 9-15). These viscometers can also be
is relatively straightforward to obtain. In this used to measure crosslinked fluid viscosity by placing
332
a B2 bob (smaller and heavier bob) on the same Again, no definitive rules exist for crosslink
number one spring and measuring the dial reading at time and ultimate crosslinked viscosity, as
100 rpm. When used to measure crosslinked fluids, one must consider the dynamics of the fracture
the units of measure are not expressed in centipoise; treatment and the ultimate goal of the treatment,
instead, the dial reading of the viscometer is reported which should be to place proppant at a reasonable
and related to viscosity changes over time. and designed distance from the wellbore.
(The Fann-35 and Ofite Model 900 viscometers The optimum pH range and precise
are used in the field laboratory due to their portable concentration of crosslinker is determined by placing
size and rugged design. HTHP viscometers are usually the crosslinked fluid in an HTHP viscometer
used only in the laboratory due to their complexity with the proposed fluid additives (except the
and sensitivity in field environments.) breaker) mixed in the system. This sample without
The base gel viscosity should be within a 4 cp chemical breaker is referred to as the blank and
range of the required viscosity, which confirms the is used to optimize the fluid for stability and
polymer concentration as seen in Fig. 9-13. After the ensure the crosslinker and buffer concentrations are
base gel viscosity has been tested and is maintained for optimal for the desired or maximum crosslinked
approximately one hour to ensure no loss of viscosity fluid viscosity. After the fluid system additives have
(indicative of bacterial presence), then the buffer and been optimized, the breaker is then added to the
crosslinker are added to the sample to ensure the fluid system at varied concentrations and the tests
system will crosslink and do so in an acceptable time are repeated on the HPHT viscometer to ensure
period (i.e., the time required for the linear gel to a controlled fluid break can be achieved when
move to a crosslink state after the pH is adjusted and pumping the designed fracture treatment fluid
the crosslink ion is added to the system). volume and a complete break after the treatment.
Temperature and pH affect all crosslinkers and The HTHP viscosity testing should eventually
crosslink time. The effects of crosslink time and stability yield a fluid break profile that can be used in the
have been studied by Hodge and Baranet (1987) and field to provide the fluid technician on location
Malochee and Comeaux (2003). Crosslink time can be a series of breaker concentrations from which to
somewhat arbitrary depending upon the objective of select in the event that the treatment is redesigned
the fracture treatment, the rate at which the fracture or total treatment pump time changes.
treatment is pumped and the tubulars through which Figure 9-16 is an example of HTHP viscosity
the fracture treatment is pumped. testing which should be generated at the laboratory
The number of perforations, their diameter and with actual source water samples and isolated
shot density should also be considered in order to chemicals prior to arrival on location.
estimate the amount of shear at the perforations, which
5000 300
can be detrimental to the stability of some metallic
Temperature
crosslinked fluids as described by Baumgartner et
Apparent Viscosity at 40 sec-1, cp
4000 240
al. (1983) and further by Harris et al. (2005). If a
Temperature, F
metallic crosslinked fluid system is to be used and it 3000 180

is determined that the number of perforations is such
that the fluid will experience excessive shear, then a 2000 120
crosslink time, calculated to be past the perforations
Apparent Viscosity
or outside the pipe, is desired. However, if the treating 1000 60
rate is low and the ultimate proppant concentration

0 0
is designed to be relatively high (in excess of 4 ppg) 0 50 100 150 200 250 300 350 400 450 500
or if tortuosity is expected (see Section 4-2.1.4) then
Time, mins
obtaining a full crosslink before the fluid enters the
perforations is recommended. Figure 9-16 Typical HTHP viscometer test results.
333
Modern Fracturing
9-4.2 Fracture Fluid Blending Methods the method by which the fluid system is to be
blended (batch-mixed or on-the-fly) should also be
Before arriving on location it is important to know discussed and a consensus reached by the service and
whether the fluid system is to be batch mixed or operating company representatives before arriving
pumped on the fly. Both of these oilfield terms refer on location. Field QC of any fracture fluid system
to the method by which the fracture fluid system is primarily a repetition of the pilot testing that
additives are mixed or blended and more specifically was performed in the laboratory in preparation
the method by which the base gelling polymer (usually for the fracture treatment except the HTHP
guar or a guar derivative) is mixed. viscosity testing used to design the fluid system.
A batch-mixed treatment indicates the base Fracture fluid design work is performed in the
gelling agent is mixed and hydrated in holding tanks laboratory; field QC should only be used to verify
on location before pumping the fluid downhole and fluid and chemical performance.
subsequently into the formation with proppant. Batch On location, the fracture treatment proposal
mixing is usually performed on smaller treatments or or procedure should be verified to ensure all
treatments at high rates because this method allows parties involved are working from the same written
for full hydration of the base gel polymer before procedure. Then, an inventory of products should be
pumping. All fluid additives are mixed by circulating taken, including water and proppant.
a holding tank (usually a standard 500 bbl frac Sufficient water should be on location to
tank) with the downhole blender while adding the pump the designed fracture treatment and provide
required fluid additives. After the tanks are blended an adequate volume to flush or displace the final
or batch mixed, then only the crosslinker and breaker proppant stage to the top perforation in the wellbore.
are added while pumping the fluid downhole. There should also be enough excess water in each
When pumped on the fly, additives are being frac tank (tank bottoms) to prevent the pumping
mixed into the fluid as the job is being pumped into the equipment from drawing air with the blender
wellbore. This can only be done if the fracture treatment suction pump in the event the water in the tanks
rate is low enough to allow adequate hydration of the are drawn down too low. Drawing air can cause
liquid polymer before the fluid is crosslinked. This is high-pressure fracture pumps to cavitate, which
usually performed with the use of a hydration unit could result in reduced or terminated pumping
containing a holding tank or hydration tank. The of the fracture treatment at an inopportune time.
hydration unit draws fresh water from the frac tanks A general rule for fracture fluid tank bottoms is
and adds liquid polymer to this water. The hydration 40 bbl or 1500 gal per fracture tank (assuming
tank usually holds approximately 200 bbls of fluid standard 500 bbl tanks are used). This number
and allows enough time in the hydration tank for the could be more or less depending upon the
polymer to hydrate to form the base gel while moving designed downhole fracture rate.
through the unit. The hydrated base gel is then pumped After adequate water volume is confirmed,
from the hydration unit to the downhole blender, where a sample of water from each frac tank on location
the remainder of the fluid system additives are added, should be taken and tested using the seven-point
discharged and pumped downhole by high-pressure water test (see Table 9-5).
frac pumps downstream of the blender. Inventory of fracture fluid additives or
products should include some excess above that
9-4.3 Quality Control of Water-Based of the designed fracture treatment to allow for
Fracture Fluids on Location nominal chemical pump variation, the possibility
that a faulty or fouled fluid meter or chemical
The laboratory pilot testing of the fracture fluid pump is used during the fracture treatment,
system should be reviewed and approved before or in case of revised treatment design. This excess
stimulation crews arrive on location. In addition, chemical will help to ensure that the chemical
334
or product inventories are not exhausted before After the breaker tests and breaker concentrations
the end of the job. Most fracturing companies have been confirmed relative to the laboratory testing,
will usually transport 20% excess of all the fluid is ready to be pumped downhole.
chemical inventories to location. Finally, all chemical additives that are to be
After adequate inventories of water, proppant pumped or added on the fly during the fracture
and chemical product have been confirmed, treatment must be identified and the meters for
then samples of each product are collected and fluid pumps used to move these chemicals must be
a field pilot sample of the fracture fluid system is calibrated. This is performed with a stopwatch and
mixed. A standard Waring blender or equivalent a container of known volume. The pump is set to a
commercial blender of approximately 1-liter predetermined rate, usually in gallons per minute, and
capacity can be used to mix the water and gelling the time required to pump one gallon of chemical is
agent to produce a linear gel sample. The linear gel observed and the meter then calibrated to indicate the
sample is then checked for proper base viscosity measured rate. This calibration method is commonly
with a Fann-35 viscometer and subsequently referred to as a bucket test.
allowed to stand for one hour or more to ensure the Every chemical that is to be pumped on the fly should
viscosity of the sample does not decline, which may also have an alternative or a back-up pump available in
indicate the presence of bacteria. the event of a failure to the primary pump or primary
Buffer and crosslinker are then added to the pump meter. It is given that the back-up pump should
sample and the physical crosslinking of the fluid is also be calibrated prior to pumping.
observed and timed. Buffer should be added first, Rate meters should be monitored such that
pH checked to ensure the buffer is active and the pH a chemical rate is displayed in the treatment
is as per design, and then the crosslinker is added. control van during the fracture treatment for each
When mixing fluid in a Waring-type blender, the chemical that is pumped during the job on the fly.
fluid should be mixed at a high enough RPM setting These chemical rate displays indicate that each of
to create a vortex in the fluid. The crosslink time is the fluid chemical additives is being added at the
measured as the time from which the crosslinker is correct concentrations, which physical chemical
added to the fluid system until the fluid vortex is measurements during the job can confirm.
closed due to increased fluid viscosity.
Pilot testing of the chemical breaker system is 9-4.4 Quality Control of Other Fluid Systems
then performed by adding all fluid additives sampled
on location, including the designed chemical breaker, Quality control of other fluid systems (such as oil-
and once again crosslinking the fluid sample. This based systems, emulsions, gas assists and foams)
crosslinked sample containing the designed chemical are relatively different in process; however, the
breaker at the loadings determined by laboratory pilot goal is ultimately the same, which is to ensure (1)
testing is placed in an enclosed Fann-35 rotor, which Applicable laboratory pilot testing is completed before
is then placed on a Fann-35 viscometer equipped mobilizing any equipment or product to location;
with B2 bob, which is placed in a heat cup preheated (2) Sufficient volumes of all water, oil, gases,
to BHST (maximum of 190F). Readings are then proppants and chemical additives are on location
observed at 15 to 30 minute intervals and recorded. with a minimum of 20% excess; (3) All fluids and
These Fann-35 viscometer dial readings should decline chemical additives are pilot tested for performance on
sequentially to confirm a reduced fluid viscosity over location; (4) All chemical pumps are calibrated before
time. Laboratory pre-job tests should have determined pumping downhole and all chemical pumps have a
the impact of specific breaker concentrations on the calibrated back-up pump available; and (5) Fluid
dial readings over time at BHST; the dial readings in rates of all chemicals pumped on the fly are displayed
the field should follow the same trend, indicating the in the treatment control van or cabin. See Fig. 9-17
chemical breaker is functioning as designed. for a quality control checklist.
335
Modern Fracturing
Fracturing QC Checklist

Pre-Job Procedures:
Ensure the proper procedure is used for fluid, stage, proppant and fluid volumes.
Ensure tank straps indicate sufficient fluid volume for treatment.
Ensure manufacturers proppant weight tickets and sieve analysis (if provided) indicate the proper type and
amount of proppant was delivered.
Ensure Sieve Analysis was performed on each type of proppant and that they conform to specs.
Ensure Water QC Tests were performed on each tank and indicate proper fluid quality.
Ensure Gel Test was performed prior to gelling and viscosity was stable at ambient temperature. (If viscosity
declines, the tanks should not be gelled. May indicate bacteria or breaker. Repeat tests with fresh samples of
water to verify results.)
Ensure Gel QC was performed on gelled tanks or Hydration Unit and viscosity is within published viscosity
charts for the polymer loading (ppt).
Ensure Breaker Tests were performed with location chemicals and gelled tank water.
Ensure primary and backup chemical pumps have been calibrated.
Ensure chemicals have been measured (strapped) with hoses loaded prior to pumping.
MONITORING PROCEDURES
Ensure the measured viscosity (centipoise) is charted in real-time for all "On the Fly" jobs. (Not calculated
from gel pump rpm).
Ensure all "on the fly" chemicals are charted in gallons per thousand gallons fluid (gpt).
Ensure monitoring van operator will alert the Service Supervisor, Engineer and Fluid Tech immediately if "on
the fly" chemicals vary by more than 10% of plan during the job.
Ensure the proppant will be monitored and charted with a Radioactive Densiometer. (Not calculated from
sand screw rpm).
JOB EXECUTION
Be informed of job status during testing, gelling and pumping.
Ensure the fluid will be sampled and needed adjustments made by Engineer or Fluid Tech.
Ensure chemicals are strapped during the job, to confirm chemicals are on schedule.
Figure 9-17 Quality control checklist
9-5 Quality Control of Propping Agents Recently, these API RPs have been superceded
by a new standard from the International Standards
Within the industry there are relatively old Organization, ISO 13505-2 (see Section 8-4). Table 9-
recommended practices (RP) for propping 6 shows a comparison of two quarried proppants. Both
agents published by the American Petroleum of these proppants fit API specifications for sieve, but
Institute (API). These can be found in API RP- one of the proppants has effective permeability and
56 for fracturing sand, RP-58 for gravel pack sand conductivity 50% higher than the other.
and RP-60 for artificial proppants. These guidelines Based on the present API specifications virtually all
may be obtained by going to the American of the proppant can be on the smallest screen and yield
Petroleum Institutes website, clicking API Petroleum very low conductivity but still fulfill the API requirements
Standards and entering RP-56, -58 or -60. These concerning sieve analysis. With new quarries coming on
guidelines have been in place since the 1950s, line and with the movement to non-API man-made
and they are beneficial. However, the present proppants, there is a need for new and more precise
guidelines sieve analysis recommendations are too guidelines to relate sieve analysis to long-term downhole
broad and do not realistically represent the final conductivity at bottomhole temperature and closure (see
conductivity of the proppant. Section 8-7 for more details on conductivity testing).
336
Table 9-6 Comparative API sieve analysis for two quarried proppants
Proppant Type: Northern white = 1; Hickory = 2

Proppant Size: 20/40 = 1, 16/30 = 2
Variation 20/40 FP 20/40 FC 16/30 FP 16/30 FC
Proppant Type 1 1 1 1
Proppant Size 1 1 2 2
18 0.0 0.0 0.0 0.4

20 0.2 2.8 13.0 48.0
25 10.4 48.8 51.0 50.4
30 36.4 34.4 31.0 1.2
35 35.8 10.6 0.0 0.0
40 16.8 3.0 5.0 0.0
50 0.4 0.4 0.0 0.0
pan 0.0 0.0 0.0 0.0
Total % 100 100 100 100
Mean Phi 0.775 0.534 0.457 0.256

Mean D (mm) 0.584 0.691 0.728 0.837
0.228 0.215 0.221 0.133
Phi84-phi50 (Std Dev)
SG 2.65 2.65 2.65 2.65
Porosity (%) 42 42 43 43
Permeability
Calc 241 343 427 637
Actual 241 370 439 599
It is always a very good idea to gather contaminants are usually acquired from bulk handling
proppant samples from each compartment or equipment that has not been sufficiently cleaned prior
truck delivering proppant to location. In several to handling the proppant or frac sand.
cases in practice, this sampling has found not only On-site sieving and specific gravity testing are
different proppant than what was ordered but common, but there is virtually no acid solubility
mixed proppant or, more troubling, contaminants testing or measurements of fines released although
such as grain or cement. API has specifications on acid solubility and the
Each proppant sample should be sieved on amount of fines after crushing.
location. It is not a good idea to accept sieve Another area that needs to be addressed is the issue
analysis from the supplier because it is very easy of point stress and strength of particular proppants used
to have samples mixed up, and there is often no in water fracs. There is a well-based belief that due to
correlation between the sieve analysis from the very low concentrations of proppant pumped during
supplier and that conducted in the field. In addition water fracs, there is literally no time where a conventional
to the requirement for 90% of the proppant to multigrain proppant pack exists. Many proppants,
be between the uppermost and smallest screen, because of their small size, have extremely low
the API calls for a maximum percentage of conductivity in a pack mode, and it is felt that they
fines in the pan and no more than a certain function as a bridging/diverting material that holds the
percentage above the lowest screen. fracture open, allowing for large conductivity due to the
A very simple test to look for chemical existence of open fractures. Darin and Huitt (1960) first
contamination is to place the proppant in water described this phenomenon. Proppant partial monolayers
and measure the pH of the water to check for pH are described in Section 8-4.8.
contamination. This simple test has been very beneficial It is also important to do sieve analysis on artificial
in catching cement or flyash-contaminated proppant proppant to be assured that the proper mesh size
and should be regularly conducted on every job. These was received. There have been numerous occasions
337
Modern Fracturing
when different sizes of proppants - such as a few they can result in vast improvement in propping
bins containing 20/40 and others 16/30 - have been agent performance. The above recommendations
delivered. This may not result in a permeability issue, are followed by most of the major fracturing
but can be a serious execution issue if the perforations companies and operators.
are small or if the delivered proppant cannot achieve
the designed fracture width. 9-6 Quality Control and
Execution of Acid Fracturing
9-5.1 Quality Control Guideline
for Propping Agents The process of acid fracturing is much like that of
proppant fracturing (see Section 4-7.8). The difference
1. Obtain samples of proppant from every is the use of reactive fluid, usually hydrochloric acid
compartment, container, or truck delivered to (HCl) in formations that exhibit very high acid
location. If using big bags or super sacks, take solubility, such as carbonate reservoirs, and also the
one sample from every 20 bags, ensuring that each absence of a propping agent. Acid fracturing can be
batch number has at least one sample taken from performed with other organic or inorganic acids, but
it. Obtain an aliquot sample by digging down for the most part acid fracturing is done with HCl in
and catching a sample that does not contain a concentrations from 3% to 28%. To limit discussions
lot of dust. Also do not catch samples directly regarding the execution and quality control of acid
off the bottom, as this may cause errors due to fracturing, it will be assumed in this chapter that acid
segregation of larger particles. fracturing would include the use of HCl only.
2. Sieve the proppant on site or in a local field While the stimulation objective is the same
laboratory and make sure that the proppant fits in acid and proppant fracturing, i.e., to create
API guidelines. If samples are out of specification, a flow channel for hydrocarbons to move from
catch another sample to confirm it is not a the reservoir to the wellbore, the mechanism is
sampling problem. This procedure should be different. In acid fracturing, a conductive hydraulic
done long before the start of the job to eliminate fracture is created by etching the walls of a created
job delays if replacements are needed. fracture and essentially removing rock from
3. Place samples of the proppant from each the reservoir, thus removing it from the created
compartment in water and measure for pH hydraulic fracture as spent acid.
contamination. Proppant transport trucks The process of acid fracturing is performed by
often also transport cement and other high- pumping reactive HCl at pressures exceeding the
pH materials that can play havoc with fracture gradient of the reservoir, unlike matrix
stability and break characteristics of fracturing acidizing where the acid is pumped below fracture
fluids. It is always a good idea to physically pressure. Reservoir rocks treated by acid fracturing are
observe the proppant in each compartment normally limestones, chalks, dolomites or dolomitic
to be sure that mixing of proppants or limestones (Lund et al., 1973).
contamination has not occurred. The treatments are designed such that the HCl
4. If there is a question of the proppant being is pumped into an induced hydraulic fracture at rates
contaminated or out of specifications, higher than the formation leak-off rate of the treatment
always keep at least a gallon container from fluid or acid to promote fracture extension. The leak-
each compartment (or from 1 in 20 bags) off of the treating acid can be controlled by viscosifying
for post-job analysis. the acid, retarding the reaction rate, pumping a
Although the above guidelines do not fit the water-based pad ahead of the acid or placing a
rigorous stipulations for sample gathering given particulate leak-off control substance in the treating
by the API, nor do they cover such things as acid in an effort to seal naturally occurring fissures or
acid solubility, roundness and sphericity etc., high permeability in the reservoir. If the leak-off rate
338
is substantial, the acid will be at risk of treating only or emulsions (Bansbach, 1970; Moore et al., 1965;
the existing natural fissures in the reservoir and not and Thomas et al., 1993). Compatibilities should
establishing a dominant induced hydraulic fracture even be considered in gas condensate wells.
necessary for production enhancement. Samples of the reservoir fluids should be
The chemistry of the acid system is also modified obtained prior to mobilizing acid and pumping
in the case of high-temperature reservoirs (> 140 F) equipment. To ensure acid and reservoir fluid
in order to retard the reaction rate of the HCl. This compatibility, 50 cc of the reservoir fluid sample is
chemical retardation enables the acid to penetrate mixed with 50 cc of the proposed acid system and
deeper into the reservoir before the acid spends additives, and then blended with a predetermined
(reduced reaction rate). Additives that increase the volume of iron (usually on the order of 5000 to
viscosity of the HCl are also used in highly reactive 10,000 ppm in a 3:1 ratio of Fe2+:Fe3+). A breakout
reservoirs and pumped in sequential stages with non- of the fluids is observed according to API RP-42 to
viscosified acids or acids with different reaction rates to confirm the mixed fluids do not form an emulsion or
create a differentially etched conductive fracture that create an iron-induced sludge (API, 1990). Emulsions
will stay open after the treating pressure is relieved and and sludges can create severe and persistent damage
the treating fluid removed (see Fig. 9-18). in the reservoir if not considered during the selection
of acid additives. It is important to use fresh crude or
condensate samples for these tests because aging of the
sample (i.e., loss of the light ends) can significantly
affect the results of the testing.
On location, quality assurance involves
confirming the designed fluids and products
and their associated volumes to ensure the
Figure 9-18 Differentially etched rock (left) and rock treatment is pumped as designed.
with inadequate etching (right), taken from parallel plate
laboratory tests The HCl itself is of first priority, and it should
not be assumed that the HCl is mixed to the correct
There are several factors that affect the penetration concentration. The concentration of the HCl
distance or the created effective acid fracture length. can be easily determined by obtaining a sample of
These factors include, but are not limited to leak- the acid after the tanks have been rolled or properly
off of the acid into the reservoir, acid injection rate, mixed by the blender on location. This sample is
temperature, acid concentration, retardation of acid best caught via a submersible sampler (tank thief )
reaction rate, solubility of the reservoir and rock designed to retrieve the sample from the middle of
type as described by Williams et al. (1979). By the tank. In the absence of a tank thief a sample from
understanding the factors that affect acid penetration the top of the tank will suffice. Care should be taken
and effective stimulation, one may surmise which when a sample of any corrosive fluid is obtained,
factors are most important to a successful acid to ensure the safety of all personnel associated.
fracture treatment and thus develop focus areas The sample from the tank can then be titrated
to promote the treatment. to obtain the concentration.
Additionally, the specific gravity of the
9-6.1 Quality Control for Acid Fracturing HCl sample should be measured. The specific
gravity can be used to estimate the concentration of
The first step in the quality control of any acid fracture the HCl in the absence of titration or used to confirm
treatment is to ensure that the reactive acid system the titration is correct. It is also recommended
does not damage the formation. Damage mechanisms that the volumes of raw acid and mix water
include combining iron with reservoir hydrocarbons used to create the applicable HCl concentration
and their associated asphaltenes to create sludge and/ is known (see Table 9-7).
339
Modern Fracturing
Table 9-7 Specific Gravity and Mix Ratios for Standard reactions. Laboratory testing at designed reservoir
Oilfield HCl Concentrations temperatures is usually performed well in advance
Per Barrel Per 1000 gal (24 bbls) of the subject treatment or during the research and
Percent Specific
development of the acid system to achieve these viscosities
HCl Acid Gravity Raw Water Raw Acid and desired reactions. On location, simulation of these
Water
Acid per gal per
per gal results is not practical (other than breaker tests), and it
per gal (bbl) gal( bbl)
947 53 is therefore imperative to ensure the chemical additives

2 1.099 39.8 2.2
(22.5) (1.5) are added at the designed volumes, ratios or rates so
927 73 that the desired result is achieved in-situ.
3 1.015 38.6 3.4
(22.1) (1.9)
However, the base viscosity of any gelled acid system
902 98
4 1.021 37.9 4.1 should be measured on location prior to pumping
(21.5) (2.5)
5 1.025 36.8 5.2

887 123 downhole. The base gelled acid viscosity is then compared
(21.1) (2.9)
to the designed base viscosity at the ambient temperature
813 187
7.5 1.037 34.1 7.9
(19.4) (4.6) at the time of measurement to ensure the proper amount
747 253 of acid gellant has been used. The measured viscosity
10 1.050 31.4 10.6
(17.8) (6.2) of the gelled acid system is then compared to a quality
612 388
15 1.075 25.7 16.3
(14.6) (9.4)
control graph of viscosity vs. temperature for the gellant
470 530 used, as shown in Fig. 9-19.
20 1.100 19.7 22.3
(11.2) (12.8) The chemical additives for all HCl systems should be
230 770 separated by (1) those designed to be mixed in the acid
28 1.142 9.6 32.4
(5.5) (18.5)
tanks and (2) those additives designed to be injected with
HCl while pumping downhole or on-the-fly.
If a non-reactive pad fluid is included as part of The volume of all acid additives that are to be mixed
the treatment fluids design, the QC of this fluid is as in the acid tank should be measured prior to pumping
discussed in Section 9-4. them into the tank and then re-measured afterward to
A number of different acid systems are available ensure the proper volume of acid additives are added to
in the industry for use in acid fracturing. Some of each acid tank containing acid.
these systems include emulsified acids, surfactant
25
viscosified acids, gelled acids, cross-linked acids and 24
23
in-situ cross-linked acids. 22
21
An in-situ crosslinked acid is an acid system that will
Viscosity at 511 sec-1, cp
20
19
viscosify or cross-link upon spending to some specific pH 18
17
or pH range. After the acid has spent further and the pH 16
15
is subsequently increased, the acid will lose its enhanced 14
13
12
viscosity and return to a linear viscosity fluid. 11
10
Cross-linked acid systems may depend upon a 9
8
chemical reaction (or chemical breaker) to bring the 7
6
viscosity of the base acid system back to a linear viscosity. 5
If the chemistry of the cross-linked acid system depends
55
60
65
70
75
80
85
90
95
0
5
0
5
0
5
0
5
5
10
10
11
11
12
12
13
13
14
Temperature, F
upon a breaker for viscosity reduction, then breaker tests 15 gpt AG57L in 15% HCl 20 gpt AG57L in 15% HCl
of the cross-linked acid system should be performed on 15 gpt AG57L in 20% HCl 20 gpt AG57L in 20% HCl
location, as is done for any water-based cross-linked fluid REMARKS: ADDITIVES:

R1/B1 rotor/bob configuration with a 15% or 20% HCl,
system discussed in Section 9-4. spring factor of 1. Direct reading (field grade), 1 gpt
The performance of these complex acid systems viscosity @ 300 rpm (511 sec-1). Corrosion Inhibitor
WATER SOURCE: ACID SOURCE:
depends primarily on the volumes of the associated Tap Water Acid Tank.
additives required to produce these designed fluid Figure 9-19 Acid gellant viscosity vs. temperature
340
Volumes of the acid additives that are Table 9-8 Quality control on location - conducted on
designed to be pumped on-the-fly should be all acids
Proper Mixing Equipment Used for Blending Gelling
measured prior to starting the job to ensure that Agents
adequate volumes are available before starting Proper Amount & Type of Gelling Agent Used for
Design
the treatment and then compared to volume
Viscosity Measured & Corresponds with Proper
measurements after treatment. Viscosity
Calibration of the liquid additive pump to No Fish-eyes or Un-hydrated Masses, Precipitated
be used for pumping all on-the-fly chemicals Chemicals, Inhibitors on Surface of Tanks
should be performed prior to starting the treatment Batch Mix Chemical Schedule Report Prepared &
Verified
as described in Section 9-4. Additionally, real-time
metering of the chemical additives rates should On-the-Fly Schedules Prepared
be displayed in the monitoring van or cabin as On-the-Fly Chemical Pumps & Back-up Pumps
Calibrated
also discussed in Section 9-4.
Chemical Inventories: Pre-Job; Post Job; Chemical
To aid in the quality control of acid fracturing Variance Report
fluid systems, quality control check lists have been
included for base acid and several specialized acid Table 9-9 Quality check - Gelled & in-situ cross-linked
systems (see Tables 9-8, 9-9 and 9-10). acids
It should be noted that HCl in any concentration Proper Mixing Equipment Used for Blending Gelling
Agents
is corrosive and harmful to all personnel. Although the
safety aspects of handling corrosive and reactive fluids has Proper Amount & Type of Gelling Agent Used for Design
not been discussed, every effort should be made to limit Viscosity Measured & Corresponds with Proper Viscosity
exposure of all personnel on location to HCl during all
No Fish-eyes or Un-hydrated Masses, Precipitated
acid treatments. Personal Protective Equipment (PPE) Chemicals, Inhibitors on Surface of Tanks
should be required by all personnel while on location,
Batch Mix Chemical Schedule Report Prepared &
Verified
and care should be taken to eliminate all potential fluid
leaks during the treatment. Figure 9-20 is a photograph On-the-Fly Schedules Prepared
of several frac tanks containing acid on a location. Note On-the-Fly Chemical Pumps & Back-up Pumps
Calibrated
the PPE worn by all personnel, the climbing platform Chemical Inventories: Pre-Job; Post Job; Chemical
on each tank and the use of a large-diameter suction Variance Report
manifold to limit the number of hoses and associated
valves for transferring the acid while pumping and Table 9-10 Quality check Emulsified acids
limiting exposure to personnel.
Proper Mixing Equipment Used for Blending Emulsions
Pilot tests Conducted before Emulsiying to Ensure Proper

Emulsion
Emulsion Stable for Job Time
Viscosity Measured and Corresponds with Documented

Viscosity
Oil External Emulsions Tested for Oil Wet External Phase
No Fisheyes or Unhydrated Masses, Inhibitors on

Surface
Batch Mixing Schedules Prepared
On-the_Fly Schedules Prepared
On-the-Fly Chemical Pumps & Back-up Pumps

Calibrated
Chemical Inventories: Pre-Job; Post Job; Chemical
Figure 9-20 Acid tanks on location Variance Report
341
Modern Fracturing
Using the quality control checklists provided in is encountered, particulate materials can be used to
this chapter and understanding the goals and objectives bridge the face of the open-hole section or the initially
of the designed acid fracture treatment should enable induced fracture, thus diverting treatment fluids to
those on location to ensure that a proper and safe acid another section of the open hole by the subsequent
fracture treatment is executed. increase in hydraulic fracture pressure.
This method is not recommended as a completion
9-7 Multi-Stage Fracturing and method and should be used only when it is not
Isolation Methods possible or practical to case an existing open-hole
section for completion. However, if the only method
It has been established that hydraulic fracturing of completion is open-hole and the reservoir must
has a high probability of increasing production be fractured, then the use of granular-type bridging
when the designed treatment is properly placed in the material for diverting treatment fluids has had some
target reservoir. While the placement of the treatment is success (Howard and Fast, 1970).
almost certain to enhance the reservoir productivity for The most commonly used bridging material is
some duration, the physical placement of the treatment graded rock salt (NaCl) with water-based fracture
can be a challenge in wells containing multiple reservoirs fluids and/or oil-soluble naphthalene (Howard and
or massive potential pay intervals. Fast, 1970). The diverting material is usually placed
It is not uncommon for a reservoir to comprise in fluids at high concentrations between stages of
hundreds, and in the case of some horizontally drilled fracturing fluid and proppant. The expectation is
wells, thousands of feet of prospective producing that the diverting material will bridge the initially
reservoir. The placement of the hydraulic fracture induced hydraulic fracture and divert the next
treatment in these instances is often accomplished by pad fluid and subsequent proppant stages to a
pumping multiple fracture treatments in stages. The secondary producing interval.
perspective reservoir is divided into discrete sections Although increased pressure may be observed
and hydraulically fractured in sequential stages to at the surface after the bridging material impacts
place the desired treatment in the most productive the interval, it is not certain in each case that the
intervals and lessen the potential to fracture non- pressure increase is necessarily the result of diversion
productive reservoir rock. to another section of the open hole. The resulting
While the number of stages that should be pressure increase could be additional net pressure
performed and the interval that can be fractured are associated with the larger particulate material entering
sometimes in question and depend on a number of the first induced fracture. The increased pressure
reservoir and rock characteristics, the ability to fracture could also be a temporary diversion that occurs
in stages is not in question. Multi-stage fracturing has until such time that the abrasive proppant erodes
become a technically sound and economic method to the bridging material, either at the face of the open-
stimulate large pay intervals. This section discusses hole wellbore or some distance from the wellbore in
several methods used to stage treatments. the induced hydraulic fracture.
9-7.1 Diverting Agents 9-7.2 Ball Sealers
The use of particulate diverting agents during fracturing Ball sealers have been used successfully as a diverting
has been usually applied in open-hole or uncased agent during hydraulic fracturing in cased-hole
sections. The practice is not often recommended, applications for many years. When the interval to
the use of diverting agents in an open-hole section be stage-fractured is large enough that multiple
is considered when the section is large enough that perforations are scattered over a long producing
it would not be possible to mechanically isolate a interval, ball sealers could be considered for diversion
given section of the open hole. When this scenario during hydraulic fracturing operations.
342
Studies have been performed in areas where 9-7.3 Limited Entry

large pay sections were completed, comparing wells
with a single fracture treatment and wells stimulated Limited entry is a proven and technically sound method
with the use of ball sealers for diversion. The latter to divert fluids from one segment of an interval to
performed somewhat better when compared another. Limited entry is, as the name implies, a method
to a similar set of wells where no diversion was by which the number of perforations is limited within a
attempted (Baylocq et al, 1999). given wellbore interval or between two or more intervals
The use of ball sealers for stage fracturing is selected to be fracture stimulated.
an economically attractive method of fracture In order to treat a specific perforated interval in
stimulation when wellbore configuration or a well, the pressure must be increased such that it
downhole conditions prevent the use of plugs or exceeds the fracturing pressure of the reservoir to be
mechanical diversion techniques. stimulated. By limiting the number of perforations in
Although ball sealers have been used successfully the wellbore and their diameter, the friction pressure
at times, their wide use as a stage fracturing technique across the perforations varies with the rate in which
is not recommended. Ball sealers, like diverting agents, the fracture fluid is being pumped. When the fracture
do not allow the determination of the interval at which rate is increased, the perforation friction pressure
the diversion occurred nor whether the subsequently also increases, causing a restricted flow or choke
fractured interval was the desired one. The use of ball effect at the perforation and subsequently increasing
sealers is thus an unpredictable approach to diversion. the differential pressure between the wellbore and
In addition, ball sealers may provide a temporary reservoir. This increased differential pressure can cause
diversion that may or may not extend through the entire additional perforations or intervals to accept fluid or
treatment. Ball sealers can be eroded due to the erosive divert the fluid to another interval.
nature of proppants pumped at fracture rates. After Limited entry is a term for the practice of
this erosion occurs, the perforation initially plugged limiting the number of perforations in a completion
with the ball sealer may once again accept fracturing interval to promote the development of perforation
fluid. Perforations may also communicate behind friction pressure during a stimulation treatment. This
the casing during the fracture treatment, causing a process has been described by Lagrone and Rasmussen
loss of differential pressure between the reservoir and (1963) and Cramer (1987). The resulting choking
the wellbore and allowing the ball sealer to become effect creates excess pressure in the casing, allowing the
unseated and perforation isolation to be lost. simultaneous entry of fracturing fluid into multiple
If ball sealers must be used, several guidelines zones with varying in-situ stress.
have been established for their application. The ball The discharge coefficient Cd of a perforation can be
sealer should be larger than the perforation they defined as the ratio of the ideal rate of discharge, qideal, to
are to seal. As an example, if a perforation of 0.38 the actual rate of discharge, q:
in. diameter is to be shot, the diversion ball sealer q
should be a minimum of 0.5 in. in diameter. The ball C d =
. (9-1)
qideal
sealer should also be of adequate specific gravity that
it will not float in the selected fracture fluid, and it The ideal rate of discharge, qideal , can be derived
should fall to the bottom of the well after fracture from Bernoullis equation, combined with the discharge
stimulation operations are complete. The ball sealer coefficient multiplier to give Eq. 4-2:
should also be sufficiently durable to minimize
q 2 s
deformation during the fracture stimulation, thereby p
perf = 0.2369 2 4
, (4-2)
N perf D perf Cd2
preventing the ball sealer from partially or completely
entering the perforation and causing a permanent where pperf is pressure drop across perforations in
block of the perforation and potentially restricting psi, q is total flow rate in bbl/min, s is slurry density
production after the treatment. in ppg, Nperf is the number of open perforations, and
343
Modern Fracturing
Dperf is the perforation diameter in inches. When a bottomhole fracture pressure between intervals exists, the
fluid is entirely in turbulent flow, the effect of fluid number of perforations must be proportioned for each
viscosity plays a minimum role, and the pressure drop interval based on the pressure differential.
is primarily influenced by the degree of jet contraction As an example, the zone with the lowest bottomhole
while entering the orifice. pressure would receive the fewest perforations. Conversely
Crump and Conway (1988) evaluated the effect of a formation with a higher bottomhole fracture pressure
proppant slurries on perforation erosion. Their studies may receive more or larger-diameter perforations to
indicated that perforation entrances not only became proportion a specific volume of fluid into this interval.
rounded, but also increased in diameter. The effect of Limited entry can be used to reduce stimulation
an increasing Cd during a massive hydraulic fracture costs relative to stage fracturing. In instances
treatment would have a strong influence on the results where dual intervals are to be treated, it may be a
of limited entry fracturing. At the same time, increased preferred method for treatment.
proppant concentrations also resulted in an increase in Limited entry does have limitations. Like
the friction coefficient, effectively decreasing Cd . most diversion methods, certain unknowns
The design of limited entry fracs is often an iterative during a given treatment can negatively affect
process. Initially a number of perforations are selected to the outcome of the treatment and cause one or
adequately cover the zones of interest. Then, an injection more zones to be under-stimulated.
rate is chosen for the number of perforations selected with First, the effects of erosion by propping agents are
the aim of developing the maximum differential pressure difficult to quantify and can contribute greatly to varied
possible. If the rate exceeds the tubing or casing limitation distribution of fracture fluid as the treatment progresses
or creates insufficient differential pressure, then the rate is and the effects increase. The effects can not be measured
increased or the number of perforations or their diameter while pumping and can only be estimated in the design
changed in order to develop the maximum differential of a limited entry completion. Erosional effects cause
pressure possible with the given tubular configuration. diversion to vary as the treatment is pumped.
The calculations are then repeated until the maximum Second, the differences in fracture gradients among
differential pressure possible is achieved with the given various reservoir rocks are usually estimated and cannot
tubulars and zones to be treated. be measured until perforations are shot. If estimates of
To ensure the limited entry design is successful when reservoir properties (fracture gradients) are substantially
implemented, it is important to ensure the perforations different from those encountered while pumping, then
are open prior to the fracture stimulation treatment. the limited entry design can be grossly inaccurate and the
One successful technique is to isolate the intervals with calculated diversion ineffective.
a tubing packer and pump HCl acid at a pressure above
the bottomhole fracture pressure of all intervals selected, 9-7.4 Multi-Stage Fracturing
while dropping ball sealers in the treatment. Pumping with Mechanical Isolation
through the tubing and under a packer will ensure the
acid is being pumped into the perforations and result The objective of multi-stage fracturing (often referred
in total coverage of all perforations by ball sealers (ball- to as stage fracturing), regardless of the method of
out). A ball-out is indicated by an increased surface isolation employed, is to limit the interval of reservoir
treating pressure to the maximum pressure limit where rock to be treated, thereby allowing placement of the
pumping must be discontinued. stimulation more precisely in the zone or zones of interest.
Limited entry can be designed to distribute a specific To achieve a more effective stimulation, it is preferable to
volume of fracture fluid consistently across a selected stimulate smaller intervals, thus allowing for more precise
interval, or it can be used to hydraulically divide the placement of stimulation fluids.
fracture fluid evenly between two or more intervals. In Over-simplified pseudo-three-dimensional fracture
order to accomplish this, the bottomhole fracture pressure simulators sometimes predict extreme fracture height
of the intervals must be similar. When a variation of growth in large (several hundred feet) vertical reservoir
344
intervals. This has led to perforating these intervals set and pressure tested to verify that it will support the
with multiple perforations, assuming that enough vertical pressure necessary to fracture the ensuing stage. After the
fracture height will be generated and that the entire pay integrity of the plug is confirmed, a wireline-conveyed
section will be covered with one large fracture treatment. perforating gun is run into the wellbore and the casing
In actuality this technique can create multiple induced perforated for the second stage interval. This second
fractures along the thick pay section, each creating stage interval can then be broken down with the acid
excessive leak-off, depriving the proper growth of a fracture previously spotted in the flush of the first stage, and the
and even potentially leading to premature screen-outs. interval fracture stimulated.
It has long been recognized that a technique is needed This process is repeated until all the selected
to treat wells where more than one zone is present so that reservoir intervals within the well are fracture
each zone can be isolated and stimulated individually. stimulated. It is not uncommon to complete as many
Stage fracturing is the preferred method by which this is as six to seven stages using wireline-conveyed bridge
accomplished and has evolved over many years. plugs and perforating guns. It should be noted that
The early days of stage fracturing saw the use of plugs the spotting of acid is not always required but should
and packers to span or straddle the zone of interest. be considered in the case of formations where high
Tools were developed with dual packers arranged for breakdown pressures are expected.
variable spacing, which permitted the treatment of a Many kinds of bridge plugs are available for use
producing interval between the two packers. This method in staged fracturing. Most staged fracturing plugs
employed the use of tubing and would therefore limit (1) are classified by the material with which they are
the number of stages which could be effectively treated made. Bridge plugs can be retrievable for reuse after
due to the limited number of packers which could be the fracture stimulation is complete or expendable
simultaneously deployed and (2) the rate at which the (drillable), as shown in Figure 9-21.
stimulation treatments could be pumped due to high
friction associated with pumping fluids through small
tubulars, i.e., 2 in. tubing. Retrievable:
1. Packer-Type BP
The most prevalent method of stage fracturing 2. Wash Cup B
3. Wireline Retrievable BP
today uses bridge plugs set inside the casing. The
Drillable/Millable:
lowest or deepest interval (first stage) in the well is 4. Cast Iron BP (CIBP)
perforated first. The perforations in this zone may be 5. Wireline Set CIBP
6. Composite BP
pumped into and broken down with acid and tested for
injectivity prior to mobilizing fracture equipment to the
location. On location, the designed fracture treatment
is pumped into this same pre-perforated interval. Upon
completion of the first stage, the treatment is flushed
(displaced from the wellbore into the formation) to
the top perforation with acid and water or gelled
water. Acid is positioned in the flush such that it is
placed or spotted in the wellbore across the next
planned stage of perforations, thus allowing for acid
breakdown of the next perforated interval.
Meanwhile, a wireline lubricator (wellhead device
for the control of wellhead pressure) is installed on the
wellhead. A wireline-conveyed bridge plug is then run in
1 2 3 4 5 6
the wellbore and set just above the first stage perforations
and below the interval selected for the successive second Figure 9-21 Examples of retrievable and expendable
stage perforations. The wireline bridge plug is then bridge plugs
345
Modern Fracturing
Depending on the formation and its sensitivity to the opposite of the wellbore and into the formation
the fracturing fluids used, a bridge plug can be used (see Fig. 9-22). This process allows for the reduction of
that is removed hours or days after the treatment. near-wellbore tortuosity by aligning the perforations
Recent technology has incorporated the use of flow- with the fracture plane during the initiation of the
through bridge plugs, fitted with ports or flappers induced hydraulic fracture.
that allow flow in one direction, thus allowing the
simultaneous flow (fluid recovery) of all the staged Cross-Sectional View
fracture intervals after fracturing operations are
Formation
complete. These flow-through plugs allow the retrieval
90
or removal of the bridge plug to be delayed until the Cement
reservoir has been allowed to flow back and a large 22
percentage of the fracturing fluids are recovered. Casing

A A
Drillable bridge plugs can be made of soft metal alloys
or of composite material, enabling the drilling of the
plug in a matter of minutes. Firing
Cable/Line
9-7.5 New Multi-Stage Fracturing Technology

Figure 9-22 Casing-conveyed perforating shot direction
(from Rodgerson et al., 2004)
Recently, new techniques in stage fracturing have
been introduced that have allowed the completion of After the first stage interval has been fracture
more stages in less time than previously possible using stimulated, the second stage guns are fired, which
wireline-conveyed devices. causes a sliding sleeve above the previously fractured
One of these new staging technologies is casing- interval to be engaged. The sliding sleeve allows a flapper
conveyed perforating which has been applied valve to close in the casing and isolate the previously
successfully since 2001. Casing-conveyed perforating stimulated lower interval. The next stage or interval is
employs the use of perforating guns attached to the now ready to fracture stimulate. This method is repeated
outside of the casing along with isolation devices until each pre-selected interval has been fractured
located between each gun. Electric or hydraulic stimulated. As many as 23 stages have been successfully
control lines are attached to the casing and are stimulated over a 36-hour period with the use of this
extended to the surface during the running of the method. This advantages of this method are described in
casing for completion. The entire assembly, including detail by Rodgerson, et al., (2004).
multiple perforating guns, isolation devices and Another recent technique in stage fracturing is
electric or hydraulic lines, is then cemented into the just-in-time perforating (JITP). JITP perforating is
well along with the casing. a method of stimulation in which the zones of interest
The placement of these perforating guns and are fractured with wireline-conveyed perforating guns
isolation flapper valves requires planning and in the casing while the fracture treatment is being
coordination on location prior to the running of casing pumped; ball-sealers are used to divert to the next
into the newly drilled well. Zones of interest must be stage prior to perforating the next interval as described
identified and the guns and isolation devices must be by Lonnes et al. (2005). JIPT allows for treatment
placed at the precise depths for the stimulation of the with multiple stages or perforated intervals while
prospective pay zones (see Section 6-2.1). continuously pumping before it becomes necessary
After the well is prepared for stimulation, the to set mechanical plugs. Like the casing-conveyed
bottom gun is fired from surface by use of an electric perforating procedure, this method has been successfully
or hydraulic line. The guns are activated and six jet employed to treat multiple (in excess of 20) fracture
shots are detonated, of which three perforate through stages during continuous operations to reduce the cost
the casing and three fire outwardly from the casing to of multi-stage fracture stimulation.
346
9-7.6 Horizontal Well Multi-Stage Fracturing fracturing fluid efficiency (see Section 4-2.1.5) for
tight gas formations and should not be confused with
When fracturing horizontal wells with multiple treatments conventional minifracs (see Section 4-2.1.9) and step
the goal remains to separate discrete intervals within a rate tests (see Sections 4-2.1.7 and 4-2.1.8), which are
wellbore such that these intervals may be more efficiently together often referred to as calibration tests. In the past,
stimulated. See Chapter 10 for an analysis of horizontal there was a common belief that a minifrac on a well in
well fracturing for production enhancement. a field could allow the calculation of fluid efficiency and
The direction in which a wellbore is drilled controls leakoff coefficient for the entire field. Experience has
whether an induced hydraulic fracture is created shown that sometimes even when moving only a few
longitudinally along the horizontal lateral or whether acres away, fluid efficiency and leakoff characteristics
multiple induced fractures can be created transverse can change dramatically. It is now very common in to
to the drilled lateral (Edgeman and Walser, 2003; and conduct a FET just prior to the fracture treatment.
Soliman, et al., 2006). The planning process for hydraulic Conventional minifracs, where the actual fluid
fracturing of a horizontally drilled well must begin before to be pumped on the treatment is utilized, are still
the well is drilled. See also Section 13-1.4. run commonly on medium- to high-permeability
The lateral length should be chosen based on the wells and in reservoirs lacking the extensive fracturing
number of stages that can be practically completed history common to most US tight gas formations.
in a given horizontal wellbore. Many times excessive In addition, experience has shown that in many
horizontal sections (4000 ft or greater) are drilled to tight gas formations, because the very low reservoir
expose more reservoir, when in actuality the entire permeability (rather than the filter cake) dominates
horizontally exposed reservoir does not, or cannot, filtrate invasion, there is no significant difference
contribute to the productivity of the well due to the inability between fluid efficiencies and leakoff coefficients
(usually based on economics) to properly stimulate determined using crosslinked gels, linear gels and slick
enough of the horizontal reservoir. As with vertical water (Gas Research Institute (GRI), 1988, 1989-
wells, horizontal wells are more effectively completed 90 and 1990; Holditch et al., 1988; Robinson et al.,
through hydraulic fracture stimulation. 1991; Robinson et al., 1992; Northrop and Frohne,
Staging processes for horizontally drilled laterals 1990; Warpinski et al., 1994; and Cleary, 1994).
have evolved to optimize the hydraulic fracture Consequently, most FETs currently performed in tight
stimulation process but for the most part continue to gas formations use non-crosslinked fluids.
include variations of vertical techniques. Some of the Pre-job diagnostics should follow these guidelines:
horizontal well technologies include tractors to convey 1. For the data to be valid, there must be a bare
perforating guns and plugs through a horizontal minimum of entrained gas in the fluid in the
wellbore; pump-down guns and plugs; and sliding-sleeve wellbore. If the well has been flowing and consists
assemblies, which in some cases require pumping the of primarily gas then the entire wellbore must be
fracture treatment through the annulus of the casing or displaced prior to running the FET. This is many
liner and a tubing string which is subsequently needed times a fairly expensive process if the well is deep
to engage the sleeve assemblies and prepare the well for and 5- in. casing or larger is used. After the hole
the subsequent stage. Most of these methods have merit has been displaced, and over-displaced at least 10
but cannot be applied in all situations. bbl, a shutdown should occur; the FET should start
until after any fracture created during the wellbore
9-8 Pre-Fracture Diagnostics and displacement, has closed
Fracture Evaluation Tests 2. If the hole is full and no gas is present the fluid
in the wellbore can be utilized as the FET fluid.
One of the most beneficial practices for hydraulic fracture One should confer with the on-site personnel
stimulation is the running of pre-job Fracture Evaluation to be sure no gas is entrained due to flow nor
Tests (FETs). These are specialized tests used to determine are any significant increases in pressure noted
347
Modern Fracturing
at the wellhead. Another significant piece In medium- and high-permeability formations,

of data is that when pumping is initiated, where the filter cake dominates, and in formations
pressure starts building up. This is evidence that where the fluid leakoff is more complex than normal
the hole is full and little or no gas is present (e.g., natural fractures or pressure-dependent formation
in the fluid column. If a significant volume is characteristics), the relationship between pressure and
required to catch pressure, then it is probably time is often much more complex.
necessary to displace the hole totally and The engineer has an array of tools available for
then wait for closure from the gasified fluid analyzing the pressure decline (including t, 4t, log-
before initiating the FET. log plots, G-function analysis (see Section 4-2.1.9),
3. A typical volume for a FET is 200 bbl. This Horner plots all with or without various derivative
may be decreased for low-rate treatments down functions and many more increasingly sophisticated
tubing or increased for high-rate treatments and elaborate methods) all designed to make closure
down casing. One should plan on having a pressure (and hence fluid efficiency and leakoff
minimum of 4 to 5 minutes of pump time on coefficients) easier to find. What exists for pressure
the FET to be able to get past the water hammer decline analysis is a toolbox of methods, which should
effect and see a definitive closure. be applied when appropriate engineers should not
4. Many fracturing companies and operators limit themselves to any specific method.
recommend performing a step-down test (see The purpose of the FET is to provide a very
Section 4-2.1.8) at the end of the FET to analyze for quick and simple approximation for obtaining fluid
tortuosity. One must do this technique taking into efficiency and leakoff coefficients. It is most often
consideration that if the well is under-pressured or used in tight gas formations that have an extensive
has high leak-off, the closure pressure may be missed history of hydraulic fracturing (as in the United
while doing the step-down. Many disastrous screen- States) and for which the response to hydraulic
outs have been caused due to missing closure during fracturing operations is reasonably well understood.
a step-down analysis. Alternatively, a step rate test These low-permeability formations constitute
can be done as a separate pumping operation after approximately two-thirds of all treatments performed
closure has been seen from the FET. globally (see Section 1-6). (However, for higher-
5. It is the experience of the authors that the best permeability formations, conventional calibration
means of identifying closure pressure for tight gas tests are still recommended.)
formations is to use a combination of a square root In tight gas formations, methods that involve
vs. time plot combined with the first derivative. This time-consuming analysis and testing will defeat the
will often produce a straight line while the fracture original purpose of the test. The very simple equation
is open and will progress into a curve when the shown below allows an engineer to quickly insert
fracture has closed. However, this only occurs if the volumes pumped, pump-in rate, and leak-off rate
leakoff rate as the fracture is open is proportional to come up with an approximate fluid efficiency
to the square root of time and everything else is (Eqs. 9-2 and 9-3) which can be inserted into the
constant (see Section 4-2.1.5 and Eq. 4-9). The key extremely simple Kane equation (Eq. 9-4, Fritcher et
to seeing closure on a square root of time plot is al., 1995) to yield a pad percentage.
to minimize the pressure range so that deflections
in time are much more easily seen. A deflection in V
(9-2)
q = i ,
L
slope accompanied by a change in the derivative is a ti + t c
definitive method of seeing the change from linear
qL
to radial flow, i.e., the fracture closing. It is common = 1 , and (9-3)
qi
to see multiple closures during the pressure fall-off,
but only the first closure should be used to analyze
2
fluid efficiency and leakoff coefficient. Pad Volume % = (1 ) ,
(9-4)
348
where qL is the average leakoff rate (bbl/min), Vi is the fluid should be considered as a fluid loss additive. For
total fluid volume injected (bbls), ti is the total pumping even lower fluid efficiencies (10 to 20%), a particulate
time (mins), tc is the closure time (i.e., the time from fluid loss additive (such as starch) should be considered,
pump shut-down to fracture closure) (mins), is the although it is important to remember that these measures
fluid efficiency (see Section 4-2.1.5) (dimensionless), qi must be reversible after the treatment.
is the average injection rate (bbls/min) (= Vi /ti) and Pad Conversely, if high fluid efficiency is measured
Volume % is the percentage of the total slurry (fluid plus (i.e., above 50%), the pad volume can be decreased.
proppant) volume that should be used as pad. However, for normal and low-pressure reservoirs with
efficiencies above 75%, aborting the treatment due to
Discussion of Guidelines extremely low permeability should be considered. For
There is great controversy in the industry in relation geo-pressured reservoirs (i.e., reservoirs with pressure
to how to determine pad volume and which tools gradients above that of water), it is possible to have high
work best. The above guidelines come from literally efficiency with low fluid leakoff because the high reservoir
thousands of tight gas fracturing treatments, observed pressure dominates the leakoff. For normally pressured
over many years and the relative success achieved by reservoirs, fluid efficiency measured on each job is a good
following these guidelines. The applicability of using relative comparison of well-to-well permeability. This
the low-viscosity FETs rather than conventional relationship does not hold where depletion is present
minifracs is field-proven. or in over-pressured environments.
After obtaining the fluid efficiency from the above- Whenever possible, good bottomhole pressure
mentioned guidelines, the engineer has to decide whether data is essential. This is especially important for
to be conservative or aggressive when selecting the pad conventional minifrac analysis because analyzing
volume. If the well is normal or geo-pressured and there pressure during pumping is an important part of the
is a high degree of confidence in the closure pressure, process. Although the use of foamed or energized
it is recommended to add 10% to the fluid efficiency fluids during minifracs or FETs is rare, this is another
value to compensate for the increased viscosity and wall occasion when accurate bottomhole data must be
building when using crosslinked fluids. Some operators obtained. Using bottomhole pressure data eliminates
and consultants add 5% to the fluid efficiency in order to wellbore friction and hydrostatic pressure effects from
be less aggressive, while others simply use the measured the calculations, dramatically increasing the accuracy
fluid efficiency and then pump a larger pad volume. of the analysis. The two most common methods of
If the formation is naturally fractured, then the obtaining bottomhole pressure data are memory
FET fluid efficiency should be used. Some operators use gauges and a live annulus/dead string.
modified Kane equations (Eqs. 9-3 to 9-5) and others
have used the Shelly equation to use fluid efficiency Diagnostics in Soft Rock Environments
(Cipolla and Wright, 2000; Craig et al., 2000; and In general, soft rock environments tend to have higher
Gottschling and Myers, 2004). permeability than the formations described above,
When very low fluid efficiency is measured, this and a conventional minifrac is usually recommended.
may be an indication of high permeability, low reservoir Generally, these treatments involve the use of a tip
pressure or natural fractures (or a combination of the screenout (TSO) (see Section 4-7.3.2) to artificially
three). In such cases, the engineer should consider the increase fracture width, relative to fracture half-
use of a conventional minifrac to more accurately assess length. Consequently, tight control of the pad volume
the fluid loss characteristics of the formation. percentage is required, in order to generate the TSO at
To compensate for the increased leakoff rate, the the required point in the treatment. Typically the pad
options available are increased pad volume, increased volume percentages used in the soft rock reservoirs are
fluid rate and/or the use of fluid loss additives. In tight much smaller than conventional hard rock treatments
gas formations, for fluid efficiencies between 20 and 30%, and most soft rock design engineers do not consider a
the use of 3% diesel by volume, mixed into a water-based job successful unless a screen-out occurs.
349
Modern Fracturing
9-9 Real-Time Pressure Interpretation where very large intervals of similar lithology exist, actual
rock mechanical characteristics can vary little. This would
Often, during fracture execution, the well does not indicate that created vertical fracture height growth may
respond as the design has anticipated, and the treatment exceed the lateral fracture extension. In the Nolte-Smith
is modified or re-designed at the well site. Design analysis when presuming a PKN model (see Section 4-
modifications made on-site during the treatment are 3.3.1) the net pressure interpretation assumes that the
normally based on the pressure response of the fluid as it fluid pressure would increase as the fracture propagates.
is pumped into the reservoir. With a good understanding As the fracture height grows or if there is excessive
of the treatment objectives, perforation scheme, tubular leakoff the net pressure will decrease. The interpretation
configuration, formation reservoir, and fracturing fluid of fracture growth is explained by Nolte and Smith
systems being applied, much can be interpreted from the (1981), as modes or slopes of net pressure (as seen in
treating pressure generated during the hydraulic fracture Fig. 9-23). These slopes or slope modes can be used to
treatment. Tools are provided specifically for the analysis interpret fracture geometry and most importantly, the
of treating pressure with respect to fluid and proppant. geometry indicating critical pressure or the pressure
(See also Section 4-2.2.4) in which screen-out is probable.
Over-reliance on a net pressure plot during every
9-9.1 Nolte-Smith Plot (see also Section 4-2.1.7) fracturing situation can lead to ambiguous decisions
while pumping a hydraulic fracture treatment.
Most fracturing treatments now employ a Nolte- Although net pressure can be used to make basic
Smith plot (more commonly referred to as a net assumptions during the hydraulic fracture process, it
pressure plot). Although net pressure is a worthwhile is important to understand the limitations of a net
tool, it can be misinterpreted and fundamentally pressure plot and instances when other interpretive
misapplied during hydraulic fracture treatments. analysis should be used.
The classic Nolte-Smith (Nolte and Smith, The calculation of a net pressure plot requires
1981) analysis was designed to interpret net pressure that the reservoir closure pressure be known. In the
at a time when two-dimensional models were absence of a measured closure pressure, one is often
widely used for fracture design and most fractures estimated prior to the hydraulic fracture treatment.
were assumed to be vertically contained at some This would affect greatly the log-log net pressure plot
point during their propagation. and may make it unreliable during the treatment.
This suggests that the closure pressure should me
measured before the main treatment from either a
minifrac or other calibration test.
Other variables can affect the reliability of a net
pressure plot. One such variable is the use of small
tubulars, which magnify the effect of fluid and
proppant friction. As an example, one can expect
to encounter increased friction while pumping
through 2--in. (or smaller) tubing with 3 ppg or
greater proppant concentration (see Section 4-4.4.2).
Friction associated with the increased proppant
slurry will, in most cases, exceed the hydrostatic
Figure 9-23 Nolte-Smith analysis pressure response (see loss of increased density slurry, and an increase in
Section 4-2.1.7)
surface treating pressure can be expected. When
More recent understanding indicates that fracture relying exclusively on a net pressure plot, this may
height growth is largely controlled by in-situ rock stress be misinterpreted to be critical pressure and the
and other rock mechanical characteristics. In some cases treatment terminated prematurely.
350
The use of a dead string can eliminate this pumping via small tubulars and/or with foamed or
problem. A dead string is a string of tubing extended energized fluids. A net pressure plot is most accurate
into the casing and left open-ended (without a packer) and reliable when used with a dead string to measure
during fracturing operations. The fracture fluid and real time bottomhole pressure.
proppant are pumped via the casing/tubing annulus and
the pressure is monitored at surface on the static tubing 9-9.2 Surface Treating Pressure as a Tool
string. This static tubing string allows the accurate
measurement of the actual bottomhole treating pressure It is often necessary to rely on surface treating pressure
during the treatment. This measured bottomhole (STP), pinj , to interpret what is occurring in the
pressure excludes friction effects from the net pressure formation during fracturing operations. In order to rely
calculations while pumping and subsequently allows on pinj during the treatment it is necessary to understand
for the calculation of a reliable net pressure plot. This the relationship between the surface treating pressure
technique can also be applied to a live annulus, where and fracturing pressure, pfr , downhole. Eq. 9-5 shows
the fluid is pumped down the tubing and the BHTP is the relationship, assuming zero tortuosity:
monitored from the annulus.
Another occasion where net pressure plots may not pinj = p f phead + p pipe friction + p perf ,
(9-5)
be reliable is with the use of foamed or energized fluids
in casing or small tubulars. The use of small tubulars where pf is the bottomhole fracturing pressure
magnifies friction effects, and foamed or energized (i.e. fluid pressure inside the fracture), phead is the
fluid friction effects can vary dramatically during hydrostatic pressure of the fluid or slurry in the
pumping. When carbon dioxide (CO2) is used, it is wellbore, ppipe friction is the friction pressure of the fluid
cooled and pumped as a liquid, which is heated due to or slurry in the wellbore and pperf is the pressure loss
friction and increasing bottomhole temperature while due to perforation friction (see Eq. 4-2).
pumping. Depending upon the variation in reservoir
temperature, pressure and friction, the CO2 density Pressure/Rate/Density
6000 60
can vary during pumping and could undergo a phase
50
change, causing changes in friction and affecting the 5000
Rate & Density

calculation of net pressure. When nitrogen is used, it 4000 40
Pressure
is heated to a gas and pumped in fracturing fluids as a 3000 30
compressible gas. As such, it can exhibit unpredictable 2000 20

friction behavior that could negatively affect the 1000 10
reliability of a calculated net pressure.
0 0
The analysis of fracture pressure and the use of net 0 5 10 15 21 26 34 41 46 56 61 66
Time
pressure during fracturing can be used to estimate fracture
geometry and can describe the effect of fracture height Figure 9-24 Pressure, rate and density on a treating chart
with declining surface treating pressure
growth and fracture propagation. These interpretations of
fracture pressure analysis can be further used to optimize By understanding these contributing variables,
the design of subsequent treatments (Nolte, 1988 and one can use a process of elimination to surmise what
1991). In some cases where large infield stimulation is happening when pressure changes occur at surface
programs are performed, net pressure charts can be while pumping. The most common change in STP
used to compare well treatments relative to one another occurs when the density of the fluid changes while
and certain fracture geometries implied based on these adding proppants to the slurry. Adding greater proppant
comparisons coupled with other reservoir analysis tools. concentrations throughout the treatment creates an
In summary, net pressure plots should be used with increase in the hydrostatic head and subsequently causes
a measured closure pressure obtained from an injection a decline in STP throughout the treatment. Fig. 9-24
test. Net pressure plots should not be relied upon when displays a treating chart with declining STP associated
351
Modern Fracturing
with increasing hydrostatic head throughout the leak-off in low-permeability (< 0.5 md) reservoirs is
treatment. This fracture treatment chart is a standard rare and is usually only of concern in reservoirs with
rate-pressure-density chart commonly displayed during natural fractures or high permeability.
fracturing operations. This chart is recommended for There are also pressure increases associated with the
real-time use as it includes historical trends for pressure, interference from far-field fracture(s) that have been
rate and density. The decrease in STP due to hydrostatic executed earlier. The execution of multiple fractures
head will only change when the effects of friction are can cause large increases in STP followed by premature
greater than the decrease in hydrostatic head or the net screen-out of the treatment. In-situ stresses, reservoir rock
pressure in the formation is increased such that it is discontinuities and mechanical properties or variablilites
greater than the decrease in hydrostatic head. of the reservoir rock determine the pressure created
An increase in net pressure can be the result of a within the fracture. Pressure drops or variances within the
change in fracture fluid leak-off into the formation and fracture can be significant (Cramer, 1996). The creation
proppant dropping out the slurry. Screen-outs due to of multiple fractures near the wellbore can contribute
excessive leak-off (sometimes deliberate in the case of significantly to pressure increases while pumping in the
gravel packing of soft rock frac packs) normally take time initial stages of the fracture treatment.
to occur and are referred to as a tip screen-out (TSO). This phenomenon, known as tortuosity, acquires
This term describes the leading edge of the fracture fluid its name from the tortuous path the fracture fluids
or the tip of the induced fracture. It is at this point that must follow through multiple and twisting fractures
hydraulic fractures are initiated on a micro scale. When from the near-wellbore region into a dominant induced
fracture fluid at the tip of the fracture is lost such that hydraulic fracture (see Section 4-2.1.4). In some cases
only sand-laden fluid remains, a tip screen-out ultimately a single dominant fracture may not exist, and multiple
follows. For high-permeability fracturing, a TSO is a fractures created with tortuosity may not be sufficiently
designed occurrence (see Section 4-7.3). significant to place proppant. When unusually high
Involuntary TSOs can be overcome if recognized STP is encountered, terminating the treatment prior to
early, by increasing fracture rate (the rate at which pumping proppant and obtaining an ISDP (instantaneous
the fracture fluid is being pumped) in an effort to shutdown pressure, similar to the instantaneous shut-in
equal or exceed the leak-off rate during the treatment. pressure, ISIP, obtained in deliberate tests downhole for
This is usually an estimated increase governed by this purpose) can be useful in identifying tortuosity. This
the STP and how near the STP is to the maximum ISDP can be used to calculate a fracture gradient and
allowable STP or the pressure necessary to stay bottomhole fracture pressure. Additionally, the number
below completion or wellhead limitations. Increases of perforations that are open (receiving fluid) can be
in fracture rates are usually on the order of 5 bpm calculated and the induced pipe friction and hydrostatic
increments. When increasing the fracture rate, it is head pressure can be confirmed.
important to carefully study the STP for a minimum After the variables from Eq. 9-6 have been defined,
of one-half the pipe volume. This allows one to see any unaccounted additional pressure can be interpreted as
the effects of increased friction caused by the increase tortuosity or net pressure caused by multiple fractures in
in injection rate. After a rate change, the STP pressure the near-wellbore region or far-field fracture. If tortuosity
should be observed for a sufficient time to enable the is suspected or likely to occur in a given reservoir, it is
effects of friction to be noted. If the resulting STP prudent to obtain additional information about the
continues to increase at the same rate as prior to the reservoir through calibration tests (see Sections 4-2.1.8
rate change, then a second fracture rate increase should and 4-2.1.9). If the calibration test analysis indicates that
be considered. If, however, a change or stabilizing near-wellbore tortuosity is a potential problem, this can
(flattening) of the slope of the STP is observed, this at times be remedied by pumping short stages of low
is an indication that the rate of fracture fluid leak- concentrations of proppant (proppant slugs) in the pad
off has slowed or the increased fracture rate is now fluid. This may screen off some of the smaller multiple
greater than the rate of leak-off. Screen-out due to fractures and establish one predominant fracture with
352
a fracture width sufficient to accept greater proppant Understanding the perforation scheme during
concentrations (see Fig. 9-25). The use of proppant slugs pumping will also help one interpret the pressure
and viscous gel plugs (short stages of high concentrations responses observed during a given treatment. Two
of polymer-laden fluids) to control near-wellbore examples can be used to show how understanding
tortuosity has been used for many years and its success is the perforation design can effect real-time
well documented (Cleary et al., 1993; Aud et al., 1994; decisions during the treatment.
and McDaniel et al., 2001). The first example is the case of limited entry,
where the differential pressure from perforation
restriction is high and the loss of perforations
during pumping equates to large increases in STP.
Before Proppant Plug
In the case of limited entry it would be prudent
Dominant Fracture(s) (when possible) to shut down pumping during the
pad fluid and calculate the number of perforations
Plugging of open by rearranging Eq. 4-2:
After Proppant Plug Multiple
Hydraulic
Fractures q 2
N perf = 0.2369
s
(9-6)
Figure 9-25 Creating and controlling multiple near-wellbore p perf D perf Cd2
4
fractures with the use of proppant slugs (after Weijers et al.,

2002) Losing perforations while pumping or actually
initiating the job with a lesser number of perforations
By far the most effective, although far more involved, than designed can negatively affect stimulation,
way to control near-well fracture tortuosity is to drill wells resulting in treating the prospective pay or selected
that enter the formation vertically and perforate them pay intervals with less than the designed treatment.
with oriented perforating guns that create a path that Observing and understanding the STP while
would trace the expected fracture plane. This approach pumping can identify this limitation, if tortuosity
is indicated in a proper reservoir exploitation strategy is negligible. For limited entry treatments, if one
that assumes all wells will be hydraulically fractured. knows the number of perforations open compared
Any occurrence of tortuosity will always be problematic with number designed, it is possible to adjust the
during both treatment execution and - especially treating rate during the treatment to maintain the
- during the subsequent production. Production calculated perforation differential pressure required
problems from tortuous fracture paths are particularly to achieve the limited entry plan.
severe in higher-permeability reservoirs. Another example of perforations affecting,
or in this case, not affecting STP while pumping
9-9.3 The Effects of Perforations would be in the case a confined and homogeneous
on Surface Treating Pressure formation where the perforations are shot densely
(4 shots per foot or greater). In this instance,
The first step in fracture stimulation design should be pumping usually generates little perforation
the selection of the perforation scheme. The success restriction pressure. Understanding how little
of hydraulic fracture stimulation starts with the perforation restriction pressure is generated and
placement and design of the perforations. Hydraulic that communication between perforations is likely,
fracture stimulation, when unsuccessful, can usually it follows that any abrupt changes in STP most
be traced back to the perforation design. Perforating likely are not the result of perforation restriction
for hydraulic fracturing has been discussed in detail pressure but rather can be interpreted as a fluid
in Chapter 6. However the point is that hydraulic friction or formation response.
fracturing is most successful when the zone of interest When treating large gross reservoir intervals,
is perforated such that the fracture treatment is increases in STP may be associated with fracture fluid
pumped where needed to access the reservoir. leak-off due to the large amount of reservoir rock
353
Modern Fracturing
exposed to the fracture fluid. Inversely, decreases in Early in the fracture treatment, a logical response
STP may indicate extensive fracture height growth. to this type of pressure would be a rate increase to
In circumstances where a small confined reservoir determine whether this pressure increase is associated
interval is perforated, increases in STP are not likely with reservoir conditions. The goal of the rate increase
to be associated with fracture fluid leak-off, and would be to affect the dynamic conditions of the
unrestricted height growth is not likely. Increases in reservoir by overcoming the rate of fluid leak-off into
STP may not be affected by pumping rate increases the formation or to potentially increase the width of
and therefore may be associated with inadequate the created fracture. If the STP increase were associated
fracture width development, or the reservoir may be with one of these parameters, then an increase in
screening out or ballooning, meaning the induced pumping rate would either decrease the slope of the
fracture cannot accept additional fluid volume STP increase or eliminate it entirety.
and a screen-out may result. Although the effects of rate changes on STP are seen
Perforations, how they are placed and the size of immediately, other adjustments made at the surface are
the selected reservoir interval to be fractured should be not. Changes in fluid properties while pumping will
reviewed and their impact on the fracture treatment have gradual or delayed effects on STP. The full effect
studied prior to pumping the treatment as this aids in of a change in fluid viscosity will not be completely
understanding STP responses while fracturing. realized until the new fluid properties (i.e., viscosity
increases or decreases) have almost filled the treating
9-9.4 The Effects of Pipe Friction tubulars. As an example, it is common in fracturing
on Surface Treating Pressure low-permeability reservoirs to reduce fracture fluid
viscosity in subsequent proppant stages after a hydraulic
In order to determine if a change in the STP is caused fracture has been established and it is believed that
by pipe friction, one must be cognizant of the variables adequate fracture width has been established to accept
that effect pipe friction: rate, slurry composition greater concentrations of proppant. This decrease in
and proppant concentration. Rate changes during a fluid viscosity will reduce STP slightly, but the effect
hydraulic fracture treatment are common place and will not be seen until the lower-viscosity fluid fills the
necessary. It should be understood that small changes treating tubulars. This decrease may also be masked
in pumping rate can have a dramatic effect on pipe by changes in increased hydrostatic head caused
and perforation friction and that pipe friction changes by raising slurry density (i.e., increasing proppant
are not always immediate. Fig. 9-26 is an example concentration). However, if fluid viscosity is increased
of a hydraulic fracture treatment in which STP is (by design or otherwise) the converse would be true.
increasing throughout the job. This type of increase The STP would gradually increase as the higher-
can have multiple causes; a process of elimination can viscosity fluid begins to fill the treating tubulars, and
help to determine the fundamental cause. it might even be significant enough to overcome the
effects of an increased hydrostatic head associated with
Pressure/Rate/Density the increase in subsequent proppant concentrations.
6000 60
Thus, even small changes in fluid properties, rate
5000 50
adjustments and proppant concentration may
Rate & Density
4000 40 have significant effects on STP.

Pressure
3000 30 Friction changes during a hydraulic fracture

2000 20 treatment are the most misinterpreted and sometimes
misunderstood pressure events during a fracture
1000 10
treatment. Misinterpretation of friction changes can
0 0
0 5 10 15 21 26 34 41 46 56 61 66 lead to ambiguous decisions or even the premature
Time termination of the treatment. It would be prudent
Figure 9-26 Surface treating pressure increase to carry a fluid friction handbook on location
throughout a fracture treatment
354
while supervising a hydraulic fracture treatment. In general, it is best to flow the well back as soon as
Information on the friction properties of fracturing practical after the treatment, and flowing back should be
fluids for various sizes of treating tubulars at various thought of as an integral part of the treatment. However,
treating rates has been documented and published. it is also common practice to leave wells shut in for
Verified on location and in real time, fluid friction several days before recovering the fluids (such as when
properties may allow quick calculations or estimations fracturing multiple intervals), and modern fracturing
to determine the causes of STP changes. fluid systems do not appear to produce any significant
In summary, determining the cause of STP changes damage to many formations. As with many issues
involves understanding the variables that effect STP. in hydraulic fracturing, the effects of extended shut-
In addition it is important to also understand the in periods seem to be formation-specific; as such, the
reservoir. Knowledge of these interrelationships allows engineer should be prepared to be flexible. It also seems
interpretation of STP changes while pumping. It likely that practices that work well in low-permeability
should be noted, however, that hydraulic fracturing is gas formations may not be applicable to high-
a dynamic process and decisions must be made as the permeability formations and vice versa. The discussion
treatment proceeds. One may not have time to consider below investigates three of the major preconceptions
the definitive cause of a particular STP change, so regarding post-treatment flowback.
it is useful to understand that certain changes could 1. Is there a correlation between load recovery and
be detrimental to the objective of the treatment and well performance?
that, when encountered, some type of preplanned There is very little correlation between load
response should be implemented. recovery and well performance, especially in
The pumping rate is the most valuable tool available tight gas reservoirs, in which sometimes the best
to the supervising engineer, and it should be used performing wells are those which produce the
advantageously whenever possible. In some instances, smallest percentage of load before producing
increasing pumping rate is not logical. Examples would hydrocarbons. The literature (see above references)
be when the perforation design used is limited entry is full of contradictory case histories. In the case
and changes in pumping rate could negatively affect the of low-permeability reservoirs, good wells or wells
designed fluid distribution or when a treatment must with good permeability and flow capacity may often
studiously avoid zones known to be wet (containing start making significant hydrocarbons after very
formation water). Outside of these circumstances little production of the fracturing fluid. It seems
pumping rate increases can be used to affect reservoir counter-intuitive that in many formations low fluid
conditions and increase the probability of completing recovery equates to superior well performance;
the designed hydraulic fracture treatment. In many however, it is the authors experience that this is
cases during the process of hydraulic fracturing it may often the case in tight gas formations.
be safe to say, Rate is your friend. Fracturing companies have expended a great deal
of research and effort to create complex surfactants
9-10 Fracturing Fluid that help remove load water. Intuitively, these would
Recovery (Flowback) be valuable products to help increase production,
but that has not always proven to be the case.
One of the most discussed and controversial issues There is wide belief that improved fluid recovery
in hydraulic fracturing is the flowback of fracturing will aid in improved production, but in reality
fluids. Many papers have been written and many there has not been a direct correlation between
opinions have been formed. Often, these seem to load recovery and hydrocarbon production, as the
indicate contradictory practices (Penny et al., 1993; percentage of load recovery varies dramatically
Ely et al., 1990a; Sherman and Holditch, 1991; Ely, from reservoir to reservoir and is a function of
1994; Ely et al., 1995; Pope et al., 1995; Canon et many different reservoir parameters. Although the
al., 2000; and Cramer, 2000). affects of fluid recovery cannot be directly tied to
355
Modern Fracturing
productivity, it stands to reason that in the case of Experience with forced closure has disproved
low-permeability natural gas wells, recovering more this assertion in many US tight gas formations.
fracture fluid load in the early life of a producing Thousands of wells have been flowed back within
gas well would reduce the severity of liquid loading seconds of treatment shutdown, and studies have
later in the life of the well. shown that not only do the wells have significantly
Obviously for low-pressure wells there may be better near-wellbore conductivity due to hindered
benefit in products which are non-absorbing and settling of the proppant, but the wells also produce
allow for foaming and production of load to keep significantly less proppant in the long term (Ely,
the well from dying. However, non-absorbing 1996). The term forced closure was coined in
products can create foams/emulsions in reservoir 1986 to describe a process to speed up closure
pore throats and fractures. Any type of product on the proppant, in order to minimize near-
such as this should be only used with caution; wellbore proppant settling and to negate smearing
formation-specific testing (ideally with core of proppant due to ongoing fracture growth after
samples) is recommended. shutdown. In the early days of implementation of
An interesting trend that has gone on for the procedure, very low rates (0.25 bpm or lower)
some time, particularly in water fracs and in were used. As the process became better known,
some conventional crosslink gel jobs, is to totally rates were increased dramatically to 2 to 3 bpm for
eliminate surfactants from the fluid system. This tubing completions and 3 to 5 bpm with casing-
trend started out with cost savings in mind and only wells. The flowback rates were increased
progressed to purposely not running the products gradually because of fear of producing proppant.
to eliminate damage to reservoirs. The vast What has been learned is that the original
majority of surfactants used do indeed reduce purpose of the forced closure process (i.e., to speed
surface tension, but there is evidence to suggest up closure) was perhaps the least important benefit.
that in some low-permeability formations, they By close observation of fluid flow it was noted that
may plate out very quickly, leaving fluid with the process - combined with efficient fracturing-
high surface tension in hairline fractures and tiny oriented perforating techniques (see Chapter 6)
pore throats. Literally thousands of water fracs - created a means to bridge the proppant at the
have been conducted without any surfactants interior of the perforations and thereby assure
using the rationale that the surfactants allow high conductivity at the wellbore. Thus, a better
penetration of the fluid into fractures and tiny term for this process would perhaps be reverse
pore throats where it would have not penetrated gravel packing. In addition, the process exploits
without the surfactant. Again, formation-specific the supercharge effect from the fracture treatment
testing with core is recommended to determine if rather than watching the pressure bleed off prior
surfactants are beneficial, as well as what type and to opening the well. By utilizing forced closure,
loading should be used. Obviously one should slightly under-pressured wells can be treated
consider the use of non-emulsifiers where liquid without the need for N2 or CO2.
hydrocarbons exist in the formation and potential There has been much confusion about the
emulsion problems can occur. necessary time to open the well. For the process
There is also a strong feeling that fractures to work properly, the well should be flowed back
should be allowed to close before initiating as soon as possible, although most fracturing
flowback, to prevent proppant from flowing back companies prefer to rig down from the wellhead
out of the fracture and into the well. This may be a before doing this for safety and operational
moot point in medium- to high-permeability gas reasons. The well should not be left for too long, as
formations because fracture closure may only take significant delays will allow substantial proppant
a minute or less. However, in low-permeability settling, so the reverse gravel packing process will
gas formations, there is more flexibility. not work. Proppant production may ensue, due to
356
open fractures connecting with the perforations. Ideally, in tight gas formations, a great deal of
For the process to work optimally, short perforated care is taken to minimize pad, thereby speeding
sections (which allow for better bridging of the up closure time and also not drastically increasing
proppant) are preferred. Additionally, perfect the breaker when using crosslinked gels in the last
proppant transport fluids are very beneficial, stage of the treatment. However, it is important,
preventing rapid settling of the proppant. A especially for medium- and high-permeability gas
common mistake with fracturing treatments in formations, that sufficient pad volume be used
tight gas formations is over-breaking the fluid in to maintain the optimum relationship between
the near wellbore, causing near-wellbore proppant effective fracture length and effective average
settling and failure of the process. propped width (see Section 4-5).
If physically possible when the well is opened, It is also felt that forced closure is not
it is recommended that the well never be choked essential in tight gas water fracs or slickwater
back or shut in until hydrocarbons are being treatments, where very rapid settling occurs
produced, and then the well should be treated as and the mechanism of fracture conductivity
a normal gas producer. The worst mistake made is not from a conventional proppant pack
with forced closure flowback is to pinch back (see Section 8-4). Many wells performed well
flow in hopes of increasing pressure. During the without forced closure because of fluids banking
early stages of flowback, there is a fluid column, in the near-wellbore area. The real danger of
and the well should be allowed to flow at 2- to over-flushing and proppant smearing occurs
5-bpm rates until hydrocarbons are produced. when the fluids are perfect proppant transport
There is no danger of proppant crushing with a fluids and there is the real possibility of loss of
full column of fluid in the well. near-wellbore conductivity.
Forced closure is extremely beneficial when 3. Is leaving fluids in a formation for significant
using energized fluids. Obviously additional care periods of time detrimental to production?
should be taken in anchoring iron and assuring that Experience has shown that with properly
safe flowback procedures are followed. Not utilizing treated fluids, no damage is done to the
immediate flowback when utilizing energized fluids reservoir by leaving fluids in the formation.
is a serious mistake and in many occasions negates However, fracture fluid damage can be a
the benefit of the CO2 and N2. function of reservoir permeability and other
2. Should forced closure be used on every well? reservoir parameters. For very low reservoir
One of the worst reasons to skip forced closure permeability, the fracture face damage is not
is to allow a history match of falloff pressures that important (see Sections 7-6 and 7-6.1).
after a treatment. Jeopardizing the success of the But, in general it is agreed that fluids should be
treatment to gather data is a misplaced priority. recovered as soon as possible.
In addition, if a flowmeter is placed on the There has been and continues to be confusion
flowback line and the flowback rate and volume about regained permeability damage caused by
recorded, modern fracture simulators can take water-based fluids and damage caused by gel
the flowback into account when determining residue (see Section 7-6). Any so-called regained
fluid efficiency and leakoff coefficients. permeability damage is inconsequential when
There are, however, several situations where comparing the relative conductivity of the
forced closure is not applicable or should be avoided. propped fracture to that of the fracture face, for
The most obvious one is where the well goes on a low-permeability gas reservoirs. Field experience
vacuum. A second is where the operator is doing has shown that when modern breaker systems are
multiple stages and cannot afford to spend the time used correctly, there is no real loss of production
flowing back the well, with large numbers of people due to gel residue. The authors feel that success
and very expensive equipment on location. of the guar-based fluids is partially based
357
Modern Fracturing
upon its inherent enhanced fluid loss control, Aud, W.W., Wright, T.B., Cipolla, C.L., and Harkrider,
which helps to negate the damaging filter cake J.D.: The Effects of Viscosity on Near-Well Bore
created with less efficient fluids. Tortuosity and Premature Screen-outs, paper SPE
28492, 1994.
In summary, Bansbach, P.L.: The How and Why of Emulsions, The
1. In tight gas formations, shutting in wells to Oil & Gas Journal, September, 1970.
allow closure can be very detrimental to ultimate Baumgartner, S.A., Parker, C.D., Williams, D.A., and
production due to proppant settling in non- Woodroof Jr., R.A.: High Efficiency Fluids for
crosslinked fluids. Additionally, one loses the High Temperature Reservoirs, paper SPE 11566,
inherent supercharge from the treatment. The term 1983.
supercharge is used to describe the build-up of Baylocq, P., Fery, J.J., Para, L., and Derbez, E.: Ball
energy in the formation adjacent to the fracture, in Sealer Diversion When Fracturing Long and
the form of compressed minerals, liquids and gases. Multiple Triassic Sand Intervals on Alwyn Field,
This stored energy can be used to recover substantial North Sea, paper SPE 54740, 1999.
volumes of fluids even in under-pressured Canon, J.M., Romero, D.J., Pham, T.T., Valko, P.P.:
reservoirs if the well is flowed back immediately Avoiding Proppant Flowback in Tight-Gas
after the treatment. In medium- and high- Completions with Improved Fracture Design,
permeability formations, where crosslinked fluids paper SPE 84310, 2000.
are routinely used and proppant concentrations Cipolla, C.L., and Wright, C.A.: State of the Art in
within the fracture are considerably higher, wells Hydraulic Fracturing Diagnostics, paper SPE
should be shut in for enough time to allow the 64434-MS, 2000.
fracturing fluid to break, then flowed back. Cleary M.P., Johnson, D.E., Kogsbll, H-H., Owens,
2. Avoid using excessive amounts of breaker in the K.A., Perry, K.F., de Pater, C.J., Stachel, A., Schmidt,
later stages of tight gas treatments, for fear of H. and Tambini, M.: Field Implementation of
flowing back proppant due to lack of fluid viscosity. Proppant Slugs to Avoid Premature Screen-Out
Excessive amounts of breaker can cause severe near- of Hydraulic Fractures with Adequate Proppant
wellbore damage due to settling of proppant at the Concentration, paper SPE 25892, April 1993.
wellbore. It has been a rule used in the forced closure Cleary, M.P.: Discussion of Comparison Study of
process that if we dont see some viscosified fluid Hydraulic Fracturing Models Test Case: GRI
in the early flowback, we probably ran too much Staged Field Experiment No.3, paper SPE 28158,
breaker. One of the surprising things noted on SPEPF, p. 17, February 1994.
early forced closure treatments was that significant Craig, D., Odegard, C.E., Pearson, W.C., and Schroeder,
amounts of gel were produced without producing J.E.: Case History: Observations from Diagnostic
sand from the formation. Injection Tests in Multiple Pay Sands of the Mamm
3. After significant quantities of gas are observed when Creek Field, Piceance Basin, Colorado, paper SPE
flowing back a tight gas treatment, the procedure 60321, 2000.
should then revert back to the normal flowback Cramer, D.D.: Rewards & Pitfalls of Using Treating
procedures for oil and gas wells. Pressure Analysis for Evaluating Fracture Design,
paper SPE 36772, 1996.
Cramer, D.D.: The Application of Limited-Entry
References Techniques in Massive Hydraulic Fracturing
Treatments, paper SPE 16189, 1987.
API RP-42: Laboratory Evaluations of Surface Active Cramer, D.D.: Evaluating Well Performance and
Agents for Well Stimulation, RP-42 Second Completion Effectiveness in Hydraulically
Edition, January 1977, Reaffirmed February Fractured Low-Permeability Gas Wells, paper SPE
1990. 84214, 2000.
358
Crump, J., and Conway, M.: Effects of Perforation- paper SPE 16429, SPEFE, p. 519, September
Entry Friction on Bottomhole Treating Analysis, 1988.
paper SPE 15474, 1988. Howard, G.C., and Fast, C.R.: Hydraulic Fracturing,
Darin, S.R. and Huitt, J.L.: Effect of a Partial SPE Monograph Vol. 2, 1970.
Monolayer of Propping Agent on Fracture Flow LaGrone, K., and Rasmussen, J.: A New Development
Capacity. Trans., AIME (1960) 219, 31-37. in Completion Methods-The Limited Entry
Edgeman, J.R., and Walser, D.W.: Comparison of Two Technique, JPT, July, 1963, 693-702.
Low-Permeability Horizontal Devonian Projects in Lonnes, S.B., Nygaard, K.J., Sorem, W.A., Hall, T.J., and
the Permian Basin with Competing Completion Tolman, R.C.;: Advanced Multizone Stimulation
Techniques, paper SPE 84392, 2003. Technology, paper SPE 95778, 2005.
Ely, J.W.: Stimulation Engineering Handbook, Pennwell Lund, Kasper, Folger, H.S., and McCune, C.C.:
Publishing Co., 1994. Acidization I: The Dissolution of Dolomite in
Ely, J.W.: Experience Proves Forced Closure Works, Hydrochloric Acid, Chemical Engineer Science
World Oil, 1996. (1973) 28, 681-700.
Ely, J.W., Arnold, W.T., and Holditch, S.A.: New Malochee, S., and Comeaux, B.: Case Study: Analyzing
Techniques and Quality Control Find Success in Bottomhole Temperature Gauge Data in Gulf of
Enhancing Productivity and Minimizing Proppant Mexico Frac Packs to Optimize Fracture Fluid
Flowback, paper SPE 20708, 1990a. Crosslink, Stability and Break Times, paper SPE
Ely, J.W., Wolters, B.C., and Holditch, S.A.: Improved 84215, 2003.
Job Execution and Stimulation Using Intense Malone, M.R., Nelson, S.G., and Jackson, R.: Enzyme
Quality Control, 37th Annual Southwestern Breaker Technology Increases Production,
Petroleum Shortcourse, Lubbock, TX., 1990b. Grayburg-Jackson Field Southeast New Mexico: A
Ely, J.W., Brown, T.D. and Reed, S.D: Optimization Case History, paper SPE 59709, 2000.
of Hydraulic Fracture Treatments in the Williams McDaniel, B.W., McMechan, D.E., and Stegent, N.A.:
Fork Formation of the Mesaverde Group. SPE Propper Use of Proppant Slugs and Viscous Gel
29551, 1995 Plugs Can Improve Proppant Placement During
Fritcher, E., Byrd, A., and Stegent, N.: Optimized Hydraulic Fracturing Applications, paper SPE
Frac Pad and Gel Improve Well Productivity, Oil 71661, 2001
& Gas Journal, March 20, 1995. Moore, E.W., Crowe, C.W., and Hendrickson, A.R.:
Gottschling, J., and Myers, R.: Pre-Frac Treatment Formation, Effect and Prevention of Asphaltene
Analysis in the Hunterville Chert and Orinkany Sludges During Stimulation Treatments, paper
Sandstone, paper SPE 91419, 2004. SPE 1163, 1965
GRI Stage Field Experiment No.1, Gas Research Nolte, K.G.: Application of Fracture Design Based on
Institute, 1988. Pressure Analysis, paper SPE 13393, SPE Prod.
GRI Stage Field Experiment No.2, Gas Research Eng., 1988.
Institute, 1989-90. Nolte, K.G.: Fracturing-Pressure Analysis Nonideal
GRI Stage Field Experiment No.3, Gas Research Behavior, paper SPE 20704, JPT, 1991.
Institute, 1990. Nolte, K.G., and Smith, M.B.: Interpretation of
Harris, P.C., Morgan, R.G., and Heath, S.J.: Fracturing Pressures, paper SPE 8297, Society of
Measurement of Proppant Transport of Frac Petroleum Engineers of AIME, 1981.
Fluids, paper SPE 95287, 2005. Northrop, D.A. and Frohne, K-H.: The Multiwell
Hodge, R.M., and Baranet, S.E.: Evaluation of Field Experiment A Field Laboratory in Tight-Gas
Methods to Determine Crosslink times of Frac Sandstone Reservoirs, paper SPE 18286, JPT, p.
Fluids, paper SPE 16249, 1987. 772, June 1990.
Holditch, S.A., Robinson, B.M., Whitehead, W.S. and Penny, G.S., Soliman, M.Y. and Briscoe, J.E.:
Ely, J.W.: The GRI Staged Field Experiment, Enhanced Load Water-Recovery Technique
359
Modern Fracturing
Improves Stimulation Result, paper SPE 12149,

1983.
Pope, D., Britt, L., Constien, V., Anderson, A., and
Leung, L.: Field Study of Guar Removal from
Hydraulic Fractures, paper SPE 31094, 1995.
Robinson, B.M., Holditch, S.A., and Peterson, R.E.:
The GRIs Second Staged Field Experiment: A
Study of Hydraulic Fracturing, paper SPE 21495,
1991.
Robinson, B.M., Holditch, S.A., Whitehead, W.S. and
Peterson, R.E.: Hydraulic Fracturing Research in
East Texas: Third GRI Staged Field Experiment,
paper SPE 22878, JPT, p. 78, January 1992.
Rodgerson, J.L., Lopez, E., and Snider, P.: Unique
Multi-stage Process allows Pinpoint Treatment of
Hard-to-Reach Pay, paper SPE 90052, 2004.
Soliman. M.Y., Pongratz, R., Rylance, M., and
Prather, D.: Fracture Treatment Optimization
for Horizontal Well Completions, paper SPE
102616, 2006.
Sherman, J.B., and Holditch, S.A.: Effect of Injected
Fracture Fluids and Operating Procedures on
Ultimate Gas Recovery, paper SPE 21496, 1991.
Thomas, D.C., Becker, H.L., and Del Real Soria, R.A.:
Controlling Asphaltene Deposition in Oil Wells,
paper SPE 25483, 1993.
Warpinski, N.R., Moschovidis, Z.A., Parker, C.D., and
Abou-Sayed, I.S.: Comparison Study of Hydraulic
Fracturing Models Test Case: GRI Staged Field
Experiment No.3, paper SPE 25890, SPEPF, p. 7,
February 1994.
Weijers, L., Griffin, L.G., Sugiyama, H., Shimamoto,
T., Takada, S., Chong, K.K., Terracina, J.M.,
and Wright, C.A.: The First Successful Fracture
Treatment Campaign Conducted in Japan:
Stimulation Challenges in a Deep, Naturally
Fractured Volcanic Rock, paper SPE 77678,
2002.
Williams, B.B, Gidley, J.L., and Schechter, R.S.:
Acidizing Fundamentals, Society of Petroleum
Engineers AIME, 1979.
360
Steve Baumgartner is the region technical manager for BJ Services Company in Houston, managing an
engineering staff that provides technical support for oil and gas operators in the application of pressure
pumping services, primarily cementing and stimulation. He has held positions around the world in well
stimulation research and development, operations and technical sales throughout his 27-year career with BJ
Services. He has a BS in chemical engineering from Grove City College. Baumgartner is an active member
of the American Association of Drilling Engineers, American Chemical Society, American Institute of
Chemical Engineers, American Petroleum Institute, International Association of Drilling Contractors,
Intervention and Coiled Tubing Association and the SPE. He has co-authored SPE technical papers,
industry technical publications and internal technical publications on well stimulation chemistry and well
stimulation treatment design, execution and evaluation. He has prepared and presented numerous well
stimulation seminars, workshops and schools and served on the Well Completions, Production Operations,
Reservoir Monitoring, and the Production Monitoring and Control program committees for the SPE
Annual Technical Conference and Exhibition. Baumgartner is also a past chairman of the SPE Gulf Coast
Section Westside Study Group committee.
Chapter 10 laminated reservoirs, horizontal wells may have difficulty
competing with vertical wells and, certainly, they
Fracturing Horizontal Wells would not be attractive compared to fractured vertical
Michael J. Economides, University of Houston and wells (Economides et al., 1994).
Steve Baumgartner, BJ Services Diyashev and Economides (2006) presented
field case studies of more than 1000 wells in Siberia
and concluded that reservoirs whose permeability
10-1 Introduction is more than 10 md are attractive candidates for
undamaged horizontal wells and lower-permeability
Starting in the 1980s and, eventually, widely introduced reservoirs would perform far better with vertical wells
in the early 1990s, horizontal wells have proliferated and with fractures. The problem is that an undamaged
have become essential in oil and gas production. horizontal well is difficult to achieve in the field. They
We have already presented in Chapter 2 (Section 2- did not present fractured horizontal wells.
3.4) the inflow equations for non-fractured horizontal What are the unique issues associated with horizontal
well production. For, example, for steady state: wells for the two main themes of this book? First is
k H h( pe2 pwf2
) the performance of horizontal wells in gas reservoirs.
q= , (2-25) Second is the performance of hydraulically fractured

I ani h I ani h

1424ZT Aa + + Dq
ln horizontal wells and, in particular, transversely fractured
L rw ( I ani + 1)
wells with multiple treatments.
where One of the important problems in non-fractured
a + a 2 ( L / 2) 2 horizontal well performance is the vertical permeability

Aa = ln , of the reservoir. This becomes neutralized when the well
L/2
is hydraulically fractured. Even laminated reservoirs with
where kH is the horizontal permeability, h is the very poor vertical communication produce similarly to
reservoir thickness, and Z are the viscosity and reservoirs with much better vertical permeability if the
gas deviation factor, respectively, T is the reservoir horizontal permeabilities are similar. Thus, fracturing
temperature and Iani is a measurement of vertical-to- is the only way to remedy a situation that would
horizontal permeability anisotropy given by: recommend against the drilling of horizontal wells in
many reservoirs (Villegas et al., 1996).
k
I ani = H .
(2-27) Longitudinally fractured horizontal wells have
kV
been shown to perform very similarly to fractured
In Eq. 2-25, a is the large half-axis of the drainage vertical wells, and they do not warrant any special
ellipsoid formed by a horizontal well of length L. The consideration in this chapter (other than completion
expression for this ellipsoid is: issues, Villegas et al., 1996). This is not to say that
4 0.5
0.5
longitudinal fractures in horizontal wells should be
L
reH

a
= 0.5 + 0.25 + (2-28) dismissed. A long horizontal well drilled into the
2 L / 2
direction of maximum horizontal stress, if fractured,
L will be connected with longitudinal fractures. Done
for < 0.9reH ,
2 in stages, the executed fractures will approximate a
It is already established in many papers and books that hypothetical fracture that would be executed in a
horizontal wells are particularly attractive compared to single vertical well treatment but with a composite
vertical wells in reservoirs with good vertical permeability, length tip-to-tip, equal to the fractured horizontal
such as carbonates, and not as attractive in reservoirs with well length (Villegas et al., 1996).
bad vertical communication such as laminated structures. In horizontal wells, whether the hydraulic fracture
The thicker the formation, the more demanding is longitudinal or transverse depends on the a priori
this requirement becomes. In thick sandstones or knowledge of fracture azimuth and therefore the
363
Modern Fracturing
orientation of maximum and minimum stresses. Thus, Of unique interest in this chapter is the fracturing
it is the well that must be drilled appropriately to accept of horizontal wells transversely with multiple fracture
one or the other configuration. If the expected fracture treatments, properly spaced -- but all options are
azimuth is not known with certainty then it should subject to production engineering economics. The
be presumed that the fracture will not be perfectly selected configuration depends first on the physical
longitudinal and, at least for connectivity, the fracture performance and then on the economic optimization,
should be considered as transverse and all precautions e.g., maximizing the net present value (NPV). It must
to avoid tortuosity and multiple fracture initiation be emphasized here that economic decisions are very
should be taken as described by Abbas et al. (1993) and local and dependent on costs, logistics and, especially,
Brown and Economides (1992.) In general any well to the frequency of operations. Individual costs per
fracture misalignment of more than 17 degrees may fracturing treatment greatly depend on the number of
result in transverse type configuration (Behrmann and treatments over which costs are divided. In fracturing
Nolte, 1998.) The important perforation considerations horizontal wells this is particularly true.
for both longitudinal and transverse fracturing are Section 2-5.1 describes the sizing procedure
described in Section 10-5.2 for hydraulic fractures using the Unified Fracture
To establish an understanding of the important Design approach, and Section 2-5.2 shows the
reservoir issues for horizontal wells, an exercise is necessary adjustments in the nominal proppant pack
done with the results shown in Table 10-1 for three permeability because of turbulence effects. Section
permeability anisotropies and three thicknesses. In 4-5 presents the design procedure using the UFD
all cases, common characteristics are: kh = 1 md, g = approach. Turbulence effects are also addressed in
0.65, L = 2000 ft (horizontal well length), re = 1490 Section 8-9. One of the clear conclusions is that
ft, rw = 0.328 ft, pe = 4614 psi and pwf = 1500 psi. It is for natural gas wells hydraulic fracturing has a
clear from this example that for small thicknesses (e.g., double benefit: stimulation as for oil wells but also,
25 ft.) vertical-to-horizontal permeability anisotropy is and very importantly, the reduction of turbulence
not critical, and horizontal wells are far more attractive effects, which in higher-permeability gas wells
than vertical wells. In a very thick formation (e.g., 300 often dominate the production rate. Section 2-5.3
ft) permeability anisotropy is crucial to the point that a shows the effects of the choke skin, a configuration
horizontal well in a highly unfavorable setting such as Iani effect that adds further flow impediments and an
= 10 will under-perform a vertical well. enhancement to turbulence effects.
There are two ways to circumvent the problem
of thick reservoirs with poor vertical permeability. Table 10-1 Steady-State Gas Rates from Non-Fractured
One is to drill stacked multilateral wells, de-facto Horizontal Wells (where qv = production from fractured
vertical well)
partitioning the reservoir vertically. For example,
kv = 0.01 kv = 0.1 kv = 1
if three horizontal laterals were to be drilled h = 25 ft qv
(Iani = 10) (Iani = 3.3) (Iani = 1)
in the thick reservoir (i.e., effectively creating Production,
3.4 17.6 22.7 25.0
vertical no-flow boundaries and creating three MMscf/d
drainage zones), the production would be three

times that of the 100 ft-thick structure: 3 29.9
kv = 0.01 kv = 0.1 kv = 1
= 89.7 MMscf/d, a considerable improvement h = 100 ft qv
(Iani = 10) (Iani = 3.3) (Iani = 1)
over the single horizontal well. Production,
13.6 29.9 58.4 84.9
MMscf/d
The second way is to fracture the well longitudinally.
As mentioned earlier, Villegas et al. (1996) have shown
that vertical permeability anisotropy plays little role in kv = 0.01 kv = 0.1 kv = 1
h = 300 ft qv
the production from hydraulically fractured horizontal (Iani = 10) (Iani = 3.3) (Iani = 1)
wells, as long as the fracture is fully penetrating Production,
40.8 31.4 83.5 173.9
MMscf/d
the thick structure vertically.
364
Chapter 10 Fracturing Horizontal Wells
However, although fracturing generally reduces The Proppant Number of Eq. 2-32:
turbulence effects in gas wells, fracturing horizontal
wells transversely can actually induce new turbulence 4k f x f w 4k f x f whp
N prop = I x2C fD = =
effects that reduce the benefits of the fracture kxe2 kxe2 hp
stimulation. To further understand these issues the 2k f V p
next section in this chapter performs a series of = , (2-32)
kVr
calculations and examples taking into consideration
the effect of changing the shape and size of must be adjusted first by:
drainage, an important issue when fractures become
4k f whx f x f xe xe
tightly spaced as would be the case in fracturing N prop = = I x2C fD , (10-1)
kxe ye h x f xe ye
horizontal wells transversely.
where xe and ye are the dimensions of the drainage area,
10-2 Production from Transversely parallel and perpendicular to the fracture.
Fractured Gas Horizontal Wells
Table 10-2 Shape factors for equivalent Proppant
Transversely fracturing horizontal wells with multiple Numbers in irregular drainage areas (Daal and
treatments would result in a configuration as Economides, 2006)
shown in Fig. 10-1. Chapter 2 introduced the basic ye/xe CA
characteristics of flow from transverse fractures. Here, 0.1 0.025
a more appropriate focus is given for the effects of 0.2 2.36

0.25 5.38
multiple fractures dividing a given drainage area.
0.3 9
The first important issue is to recognize that multiple
0.4 16.17
fracturing treatments will no longer drain regular
0.5 21.84
drainage areas but quite irregular ones.
0.6 25.8
Tmin Tmax 0.7 28.36
0.8 29.89
0.9 30.66
1 30.88
In the tradition of shape factors, a comparison

of solutions between square and rectangular
rw drainage areas was obtained by Daal and
Economides (2006). This allows for calculating the
w JD of a rectangle in a comparable square drainage
area with the use of an equivalent Proppant
Number. The equivalent Proppant Number for
Figure 10-1 Multiple fractures penetrating a horizontal
Nprop < 0.1 is then defined as:
well transversely
1
if N prop 0.1
0.990 0.5 ln N prop
J D max ( N prop ) = (2-33)
6 0.423 0.311N prop 0.089( N prop ) 2
exp 2

if N prop > 0.1
1 + 0.667 N prop + 0.015( N prop )
365
Modern Fracturing
CA Table 10-3 F-function Constants

N prop ,e = N prop , (10-2)
30.88 ye/xe= 1 0.7 0.5 0.25 0.2 0.1
a 17.2 17.4 21.4 38.3 35 30.6

where CA is the Dietz shape factor (as given in Table 10-
b 54.4 55.5 54.3 46 59 89.6
2). The calculated Nprop,e can then be used to determine
c 52.5 53.3 56.3 71.1 70 70.2
the expected maximum dimensionless productivity index
d 16.9 16.9 16.9 15.84 16.3 17.8
from (see Eq. 2-33 at the bottom of the page).
The optimum CfD for Nprop < 0.1 is equal to 1.6.
For Proppant Numbers larger than 0.1, Table 10-4 F-function Prime Constants
Daal and Economides (2006) presented a semi- For All Shapes
analytical correlation: a 10
b 36
100 yeD C fD ,0.1
C fD,opt = ( N p 0.1) + C fD ,0.1 , (10-3) c 33
100
where 10-2.1 A Calculation for Transversely

ye Fractured Gas Horizontal Wells
yeD = , (10-4)
xe
To demonstrate the potential (and the pitfalls) of
and CfD,0.1 is the optimum dimensionless conductivity fracturing horizontal wells, a calculation is done,
at Nprop = 0.1 given by: using the concepts introduced in Chapter 2 and in
this chapter. Marongiu-Porcu (2007) performed a
1.6 If 1 yeD > 0.25 similar study on the subject.

C fD ,0.1 = (10-5) First, Tables 10-5 and 10-6 summarize

4.5yeD + 0.25 If 0.1 yeD 0.25. the important input variables for this
analysis. In Table 10-6 the reasonable variation of
Equation 2-33 is not applicable when the reservoir Youngs modulus with permeability is important as
aspect ratio changes from xe = ye to xe > ye, and a new will be demonstrated with the results.
function describing the optimum values has been The first calculations for comparison purposes with
introduced. Because JD,max depends on the CfD,opt, then the subsequent results are shown in Fig. 10-2. These are
(Daal and Economides, 2006): the folds of increase between fractured and non-fractured
vertical wells. Included in Fig. 10-2, in addition to the
1
J D ,max = , (10-6) gas well, are the folds of increase for an oil well. As is
0.63 0.5 ln( N p ) + Fopt
well known, as reservoir permeability increases, the
where folds of production increase over a non-fractured well
9.33 y 2 3. y .7 will decline. For example at 0.1 md permeability, the
eD eD
folds of increase may be over 10, whereas at 100 md, the
10 yeD
folds of increase may be closer to 2. (Note: This should
Fopt =
not confuse the issue of the desirability of fracturing.
a + buopt + cuopt2
+ duopt 3
2
If N p 0.1, (10-7) At a very high reservoir permeability, doubling the
a '+ b ' uopt + c ' uopt production would make practically any fracturing
and treatment very economically attractive.)
However, the most important point in Fig. 10-2 is
uopt = ln(C fD ,opt ), (10-8) the shape of the curve of the folds of increase for the gas
reservoirs. At low permeabilities, the gas trends are similar
and the constants for the F-function for Nprop 0.1 are to those in oil: Folds decline as permeability increases). But
presented in Tables 10-3 and 10-4. as permeability increases, the trends diverge: A fractured
366
gas well performs far better than a non-fractured high- Table 10-5 Common Input Data
permeability well because of the considerable reduction in for Calculation Case Studies
turbulence effects that adversely affect well performance INPUT DATA
and dominate radial flow. The obvious conclusions are Mass of Proppant, lbs 200,000
twofold. First, although fracturing is very attractive in oil Specific Gravity of Proppant 2.65
wells, its attractiveness may decline somewhat in higher Porosity of Proppant Pack 0.38
Proppant Pack Permeability, md 150,000
permeabilities (with non-fractured horizontal wells
Reservoir Thickness, ft 50
becoming a reasonable option). Second, in gas wells,
Fracture Height, ft 100
fracturing becomes imperative at any permeability. This
Drainage Area, acres 320
is a central theme of this book.
Well Radius, ft 0.4
16
14 GAS WELL ADDITIONAL INPUT DATA
FOI of JD (Frac/No-Frac)
12 Vertical Gas Well Reservoir Pressure, psi 4,000

10 Vertical Oil Well Bottom-Hole Flowing Pressure, psi 1,500
8 Reservoir Temperature, F 120
6 Gas Viscosity, cp 0.015
4 Gas Deviation Factor, Z 0.910
2
Gas Specific Gravity, g 0.71
0
0 0.1 1 10 100 Cookes Constant, a 1.54
Reservoir Permeability k, md Cookes Constant, b 110,470
Figure 10-2 Example calculation: folds of increase

JOB PLACEMENT INPUT DATA
between fractured and unfractured vertical wells. Fracture
width is as wide as determined from the optimum values Injection Rate, bpm 30
of JD and CfD (kp = 150,000 md) Rheology K, lbf/ft2. secn 0.018
Rheology n 0.45
There is a practical problem with the theoretical Youngs Modulus, psi 2 x 106 - 6 x 105
results of Fig. 10-2. Using the UFD design approach, Poisson Ratio 0.38
the fractures are allowed to become as wide as Leakoff Coefficient, ft/min0.5 0.004
dictated by the physical optimization. For natural gas,
following the procedure outlined in Section 2-5.2, as
Table 10-6 Permeability and Youngs Modulus Variation
the effective permeability of the proppant decreases
k, md E, psi
with increasing rate, the width of the fracture must
0.05 2 x 106
grow to highly unrealistic values.
0.1 2 x 106
16 1 2 x 106
Vertical Oil Well ADJUSTED
14 5 1 x 106
Vertical Oil Well not adusted

12 Vertical Oil Gas ADJUSTED 10 1 x 106
10 Vertical Oil Gas not adusted
30 7 x 105
8
50 7 x 105
6
4 100 6 x 105
2
0
0.01 0.1 1 10 100
If a maximum net-pressure constraint is imposed
Reservoir Permeability
(e.g., 1000 psi) then the results of Fig. 10-3 are obtained,
Figure 10-3 Example calculating folds of increase, net showing that in high-permeability gas reservoirs the
pressure constraint. Fracture width is now constrained
potential folds of increase will be significantly reduced.
in order to fulfill the condition of net pressure less than
1,000 psi (kp = 150,000 md) How can this be remedied?
367
Modern Fracturing
Looking at Eq. 2-35 it is clear that the indicated The choke effects will be a lot less pronounced
width of a fracture can be reduced significantly if the in lower-permeability reservoirs and much more
proppant pack permeability is increased: severe in high-permeability cases. Clearly, it can

0.5

0.5 be concluded from Eq. 2-39 that by increasing
k f Vf C fD ,opt kV f
x fopt = and wopt = (2-35) the proppant-pack permeability and fracture
C fD ,opt kh k f h
width, the choke effects will be reduced.
This can be done with the use of high-quality For gas wells the problem will be exacerbated
proppants. The reduced width also reduces the net because of enhanced turbulence effects. So, it should
pressure, so much larger treatments can be executed. be expected that the gains from fracturing, presented
Fig. 10-4 shows the shift in folds of increase earlier for vertical wells, especially for high-permeability
possible by using a much better quality proppant reservoirs, will be diminished considerably in
with pack permeability of 500,000 md. It must be fracturing horizontal wells transversely.
emphasized again that for gas wells, the effective For gas wells, the converging flow velocity inside the
permeability will be significantly lower because of fracture would be the volumetric flow rate divided by the
the turbulence effects inside the fracture (Section cross-sectional area of flow, 2rww:
2-5.2.) See also Chapter 8.
q
v=
(10-9)
16 2rw w
Vertical Oil Well ADJUSTED
14 Vertical Oil Well not adusted Eq. 10-9 provides for much larger velocity than would
Vertical Oil Gas ADJUSTED
12
Vertical Oil Gas not adusted be the case for flow in a vertical well fully connected
10 to a fracture, where the cross-sectional area of flow
8 would be 2whf (two wings). So, although the procedure
6 in Section 2-5.2 is repeated, now the turbulence effect
4 adjustment will be far more severe.
2 The second adjustment in the performance of
0 these fractures is the effect of boundaries by spacing
0.01 0.1 1 10 100
them near each other. This can be handled by following
Reservoir Permeability k, md
the procedure outlined in this chapter. In general, for
Figure 10-4 Example calculation with high-quality low-permeability reservoirs, huge increases in the
proppant. Fracture width is now constrained in order to
productivity index will be observed compared with
fulfill the condition of net pressure less than 1,000 psi (kp
= 150,000 md) vertical well fractures. This is because the boundaries
are brought closer by partitioning an original drainage
How would the production enhancement from area, enhancing the effects of drainage. (Note: this
fracturing horizontal wells transversely behave? How means that recovery will be accelerated with very
would it compare with vertical well fracturing? large early-time production rates but declining a lot
First, for each individual fracture the productivity more rapidly compared to vertical wells.)
index will be reduced as suggested by Eqs. 2-39 and 2- Fig. 10-5 shows the performance of transversely
40, providing for the choke skin effect from converging fractured horizontal oil wells for comparison to the gas
radial flow inside the fracture: case, presented below. The results for oil show that at low
reservoir permeabilities (e.g., 0.1 md), where the choke
kh h
sc = ln . (2-39) effects would be minimized, 10 transverse fractures,
k f w 2rw 2
tightly spaced, would deliver a composite JD equivalent
and to more than 25 fractured vertical wells. For very large
reservoir permeabilities (e.g., 100 md), 10 transverse
1
J DTH = (2-40) fractures would still result in the very attractive composite
1
equivalent of more than 5 vertical well fractures. These
J + sc

DV
368
results would be improved further if better-quality a net pressure constraint. Although for low reservoir
proppants are used because that would further reduce the permeabilities the results may be tolerable, they are
choke effects (see Chapter 8). totally unacceptable in higher peremabilities. Even at
The situation changes dramatically for gas wells. 1 md, five transverse fractures would lead to folds of
Turbulence begins to play a dominant role even at increase of slightly 1.5 times larger than that of a single
small permeabilities and becomes devastating at high vertical well fracture. Using better-quality proppants will
permeabilities. Thus, the enormous benefits from increase this value somewhat but not enough to make it
fracturing vertical wells, compared to non-fractured attractive, considering that vertical wells fractured with
wells, now become reversed. higher-quality proppants will also perform a lot better.
Clearly the attractiveness of fracturing horizontal
wells transversely in gas reservoirs would be
JD Multi-Frac / JD Fracced Vertical Well
27
sandwiched between the physical limit imposed by
24 1 Fracture
21 2 Fracture enhanced turbulence effects (likely for any well above
4 Fracture
18
5 Fracture
1 md) and by economic considerations for the lower
15
12
10 Fracture end of permeability values. Economic calculations are
9 in order but it is apparent that fracturing horizontal
6
3
wells transversely in gas reservoirs even in areas with
0.1 very low drilling and stimulation costs such as North
0.1 1 10 100
America would be difficult to justify in reservoirs of
moderate or high permeabilities (e.g., k > 1 md). For
Figure 10-5 Transversely fractured horizontal oil well: lower permeabilities the procedure can be attractive,
folds of increase over fractured vertical wells. Theorectical but incremental drilling and stimulation costs must
cases, fracture width is a s wide as determined from the be reduced dramatically and become commoditized.
optimum values of JD and CfD (kp = 150,000 md)
Otherwise fracturing vertical wells is indicated.
JD Multi-Frac / JD Fracced Vertical Well
5 10-3 Open-Hole
4 Horizontal Well Completions
1 Fracture
3
Open-hole completions are utilized in competent,
2
2 Fractures stable formations. Because horizontal wells often span
3 Fractures
considerable lengths, rock mechanical properties of
1 4 Fractures
5 Fractures
the reservoir must be fully understood for efficient and
0 effective open-hole completions. The primary collapsing
0.01 0.1 1 10 100
stress in a horizontal well is the vertical overburden
Solid Lines=Adjusted Dashed Lines=Unadjusted
component. Horizontal stresses reduce the potential
risk of vertical collapse. Wellbore instability problems are
Figure 10-6 Transversely fractured horizontal gas well:
folds of increase over fractured vertical wells
not normally encountered in low-permeability reservoirs
because the rocks are generally quite competent and
consolidated; these reservoirs are candidates for open-
Fig. 10-6 shows the resulting folds of composite hole completions (McDaniel et al., 2002).
JD, (i.e., the sum of the individual values) for multiple So, open-hole or barefoot completions are effective
fractures in horizontal wells in comparison to fractured in formations with low risk of wellbore collapse or sand
vertical wells. The dashed curves are for the cases where production, for example in dolomites, hard sandstones
the fracture width is allowed to increase to its theoretical or limestones and shale-free siltstones. An open-hole
value without regard to execution considerations such as completion involves running casing to the producing
net pressure. The solid curves represent the results with horizontal interval and leaving the end of the casing
369
Modern Fracturing
open. Barefoot completions are generally applied only

in formations with unconfined compressive strength
larger than 10,000 psi (Erwin and Ogbe, 2005).
For such completions it should be expected that for
long-term performance as the reservoir depletes, the
wellbore must remain open and solids-free. Fig. 10-7
shows an open-hole completion.
1000 to 5000 ft interval
Figure 10-9 Open-hole completion with perforated pipe
10-3.1 Perforating
Although it took some time in the early days of

horizontal wells to accept the practice, the creation
of perforations with a jetting tool results in very
Figure 10-7 Open-hole completion effective flow paths from the open hole into the
formation. The entire concept of creating perforations
in an open hole was at first difficult to accept
but it became a relatively common and effective
practice. The perforations essentially bypass the
damaged zone around the wellbore. A small jetting
tool is placed on the end of tubing or coiled tubing
and it is used to create small cavities or tunnels in
the formation. After each cavity is created, additional
fluid is pumped through the annulus to increase the
pressure in the cavity and initiate a small hydraulic
fracture. Jetted perforation tunnels have less damage
Figure 10-8 Open-hole completion with pre-perforated or than conventional shape-charge perforations. Jetted
slotted liner
perforations have small near-wellbore pressure
loss and tend to initiate simple hydraulic fractures
For less stable formations, an open-hole completion (McDaniel and Willet, 2002). Perforations can
with a slotted liner run into the horizontal section be located almost anywhere in the open-hole
helps prevent formation collapse near the wellbore, horizontal completion for effective stimulation
especially as the reservoir pressure depletes during and optimum well performance.
production. A disadvantage of the slotted liner
completion is the difficulty to effectively isolate and 10-3.2 Zonal Isolation
stimulate zones or sections in the wellbore (Navarro,
1996), as illustrated in Fig. 10-8. Another issue that took some time to be recognized
An open-hole completion with perforated pipe is that horizontal wells frequently need zonal
provides the most wellbore stability. Limited-entry isolation because, in spite of the old conventional
design is required to design the number and size of wisdom, reservoirs are not necessarily homogeneous
perforations to effectively initiate and place multiple along the horizontal well path. Zonal isolation in
hydraulic fractures in a single treatment (Austin et al., open hole completions, although at times difficult,
1988), as illustrated in Fig. 10-9. can be achieved with the installation of formation
370
packers at strategic locations along the slotted liner stimulation treatments along the horizontal wellbore
or perforated pipe in the wellbore, as illustrated in can be pumped in a single continuous operation,
Fig. 10-10. Horizontal well completion testing with which minimizes associated risks and allows optimum
formation packers often shows that they do not efficiency for personnel and equipment (Seale et
private complete isolation and they do not prevent al., 2006a). See Fig. 10-11. For such completions
flow in the annulus formed between the pipe and to work, the rock must be competent and wellbore
the reservoir. Formation packers should be utilized ovality must be minimal.
for zonal isolation or reservoir segmentation. The
formation packer needs to be energized to the 10-4 Open-Hole Fracturing
anticipated differential pressure or constructed to
maintain the ability to seal as it is energized by the Transverse fractures will be initiated and created
differential pressure (Henriksen et al., 2005). when the horizontal wellbore is oriented
perpendicular to the fracture plane (within 15-
degree tolerance). Longitudinal fractures will be
initiated and created when the horizontal wellbore
Inflatable packer
is oriented parallel to the fracture plane (within
less than 15 degrees) (McDaniel and Willet, 2002).
Transverse hydraulic fractures offer the greatest
potential to contact and drain a large section of the
reservoir. Longitudinal hydraulic fractures may allow
the pumping of higher proppant concentrations
with a lower risk of screenout.
Figure 10-10 Open-hole completion with external casing
Figure 10-12 shows properly spaced transverse
packers (Austin et al., 1988) hydraulic fractures; Fig. 10-13 shows improperly
spaced transverse fractures. Similarly, Fig. 10-14 shows
Zonal isolation in open-hole completions can appropriately placed longitudinal fractures, and Fig.
also be achieved with a series of mechanical open- 10-15 shows haphazardly fractured horizontal well in
hole packers deployed on a production liner with the longitudinal direction.
fracturing ports located between the packers to permit
stimulation of each interval discretely. The liner does
not need to be cemented in place and eliminates the
requirement to perforate, fracture, isolate, perforate,
etc. in a cemented horizontal wellbore. The liner
located in the open-hole section allows access to the
entire horizontal wellbore rather than encountering
problems associated with a barefoot completion. The
mechanical packers provide mechanical diversion
and isolation at high differential pressures. All of the Figure 10-12 Ideal transverse hydraulic fractures
Stage 12
Stage 11
Stage 10
Stage 9
Stage 8
Stage 7
Stage 6
Stage 5
Stage 4
Stage 3
Stage 2
Stage 1
Figure 10-11 Open-hole completion with mechanical diversion (from Seale et al., 2006)
371
Modern Fracturing
Post-treatment temperature and tracer logs indicate

that the heel and toe of the lateral are stimulated
with minimal or no stimulation in the middle
section of the well. Additional attempts to increase
acid coverage in the open-hole horizontal wellbore
may involve the use of an uncemented tubing string
with limited-entry perforated sections. After the
treatment the tubing string is pulled from the lateral
Figure 10-13 Non-ideal transverse hydraulic fractures and the well is produced as an open-hole completion.
Acid coverage throughout the lateral section is thus
improved (Willett et al., 2002)
10-4.2 Proppant Fracturing Execution
Obviously, one of the important considerations

is to place fracturing treatments appropriately
spaced and positioned along a horizontal well.
One method for selectively placing propped
Figure 10-14 Ideal longitudinal hydraulic fractures hydraulic fractures is to use coiled tubing to
create perforations in the wellbore and initiate the
fracture by pumping fluid through a jetting tool.
Conventional coiled tubing strings (1.5 to 2 in. OD)
are used for the perforating process, but they are
pumping-rate-limited for many hydraulic fracturing
applications. The hydraulic fracture treatment is
therefore pumped down the coiled tubing/casing
annulus to eliminate pump rate and treatment
size restrictions. Packers or bridge plugs are
Figure 10-15 Non-ideal longitudinal hydraulic fractures
not required to provide zonal isolation.
The creation of the perforation tunnels takes
10-4.1 Acid Fracturing Execution about 5 to 15 minutes depending on the perforation
design criteria. After the perforation tunnels are
Carbonate reservoirs are often acid-fractured (see created, the annulus is closed and pressure is allowed
Sections 4-7.8 and 9-3). The fracture geometry created to build up to initiate a fracture through the jetting
by acid fracturing depends on the acid volume and tool. Once the fracture is initiated the fracture
the injection rate. But live acid must be transported treatment is pumped down the coiled tubing/casing
as far away from the wellbore as possible to dissolve annulus. At the end of the treatment, a screenout can
and etch the carbonate in order to create conductive be induced or a high-concentration proppant plug
fractures. Pumping (bullheading) acid into an open- placed to provide isolation for the next treatment.
hole horizontal well will likely stimulate only the heel A common method is to underflush the treatment
portion of the well. An alternate method to improve and leave the perforation tunnel and wellbore full
stimulation effectiveness is to run tubing from the with a high concentration of proppant.
surface to the end of the lateral. Acid is pumped The only limitation on this technique is that
at maximum rates and pressures down the annulus the casing above the treatments cannot have any
with one set of fracturing equipment and down the perforations or mechanical integrity problems.
tubing with a second set of fracturing equipment. East et al. (2005) have developed a technique that
372
uses jetting first at the desired point, etching the 10-4.3 Cleanup
formation, and thus facilitating the execution of
each fracture without the problems and tortuosity of The first step after proppant fracture stimulation of an
normal fracture initiation. With appropriate zonal open-hole horizontal well is to clean out all the proppant
isolation a large number of transverse fractures can be left in the wellbore with coiled tubing and nitrogen, as
placed in open-hole horizontal wells. the fluid velocities generated by natural flowback will not
Another method for selective placement of be sufficient to entrain and carry any proppant left in
propped hydraulic fractures is the use of a mechanical the wellbore. After the proppant is cleaned out, the well
open-hole packer system. The system can withstand should flow back the fracturing fluid. If the well will not
the high differential pressures associated with flow enough to unload the broken fracturing fluid, then
hydraulic fracturing and contains fracturing ports nitrogen can be used to artificially lift the well as long
between the packers. The packer is designed to sustain as required. Cleaning the fracture after each perforated
differential pressures of 10,000 psi and temperatures interval is stimulated is important; leaving fracturing
of 425 oF, and to set in holes enlarged up to 50 fluid in the proppant pack may have a detrimental effect
percent for their nominal size. The fracturing ports on well production (Demarchos et al., 2006). This issue
are designed to provide a greater inflow area than the is covered extensively in Chapters 4 and 9. However, it is
cross-section of the liner, avoiding a pressure drop common practice to stimulate all intervals before flowing
through the completion. A series of packers and back. Therefore, it is important to be able to complete
fracturing ports are run in the open hole on a liner fracturing operations in as short a time as possible.
and set. The fracturing ports are opened with a ball
dropped from the surface and pumped down in the 10-5 Cased-Hole Completions
flush fluid of the previous treatment. The ball lands
in a specific seat to open the next fracturing port and 10-5.1 Cementing Horizontal Wells
provide isolation of previous treatments. Depending
on the open-hole diameter and liner size, up to 18 Cementing a well is a reliable method for controlling
treatments can be pumped in a single horizontal fracture placement in horizontal wells. Conventional
wellbore. The fracture treatments can be pumped in cement slurry formulations can have a negative
succession. After all the zones are stimulated, coiled impact on a completion. Such slurry systems have a
tubing is used to drill out the balls and fracturing low solubility in acid, which can cause difficulty in
port ball seats (Seale et al., 2006b). perforation breakdown, inhibit fracture initiation
A case history illustrates the efficiencies of the and cause excessive near-wellbore friction effects
mechanical open-hole packer system. Four offset wells during stimulation and production.
were drilled and completed with approximately the Compared with an open-hole completion, a
same horizontal length. Two wells completed with a horizontal well cased and cemented through the
cemented liner, perforated, stimulated and fixed with horizontal producing section of the well is generally
drillable bridge plugs for zonal isolation averaged 14 more desirable and less prone to failures for effective
days to pump five stages. Two wells completed with hydraulic fracturing. Of course, such a completion is
the mechanical open-hole packer system averaged more expensive, and it poses certain challenges on its
13 hours to pump six stages. A total of 13 days were own. A cased and cemented horizontal section allows
saved in well completion time, with no additional fracture initiation points to be controlled in order to
cost for coiled tubing or wellbore interventions to place multiple hydraulic fractures (Austin et al., 1988).
perforate or set and then drill out plugs. In addition The use of designated fracture initiation points along
to the time and cost savings for the completions, the horizontal section allows the appropriate spacing of
the wells completed with the open-hole mechanical either acid or proppant fracturing treatments. Figure
packer system had higher initial production than the 10-16 shows a cemented liner completion; Fig. 10-17
cemented-liner wells (Seale et al., 2006b). shows a cemented casing completion.
373
Modern Fracturing
which aids mud displacement and hole cleaning.

Slurry volumes are calculated to circulate cement to
the top of the liner plus 200 feet. After the cement
is displaced and the liner wiper plug lands, the liner
hanger packer is set and excess cement is circulated
out of the wellbore (see Section 5-2).
10-5.2 Perforating Cemented Completions

Figure 10-16 Cemented liner completion
Perforating for hydraulic fracturing is different

from perforating for production, and perforating for
hydraulic fracturing in horizontal wells varies from the
perforating of vertical well for fracturing (see Chapter
6). Gun size, perforation phasing, shot density, charge
type (entry hole and penetration), perforated interval
length, pressure conditions and fracture-to-wellbore
orientation are things to consider.
Figure 10-17 Cemented casing completion The perforated interval length can affect
the success of hydraulic fracturing in transverse
An acid-soluble cement system may be used to and longitudinal laterals.
cement the liner or casing. Such system rapidly dissolves Although for production considerations
when contacted by acid. The enlarged annular space there are distinct applications for longitudinal
created allows acid to contact the formation adjacent vs. transverse configurations (see Section 10-7),
to the perforations with minimal loss of energy and a perforated interval that is too long may lead to
effective zonal isolation of each perforated interval. the initiation and growth of multiple fractures
The acid-soluble cement system can be formulated near the wellbore. Multiple fractures can increase
to achieve the proper density, fluid loss control, bottomhole treating pressures and near-wellbore
compressive strength and free-water requirements for friction effects and will decrease fracture width
horizontal cementing. Gas migration can be controlled and fluid efficiency. Limiting the perforated
with unique cement additives or by foaming the acid- interval length for both longitudinal and transverse
soluble cement (Willett et al., 2002). configurations is always advisable.
Acid-soluble cement slurry densities can range For longitudinal fracturing, re-perforating the
from 13 to 15.8 lb/gal and can be formulated well after the fracture treatment is advisable. For
at lower densities with foam or extenders. The transverse fracturing, a limited set of perforations
removal of the acid-soluble cement adjacent to the extending up to four times the wellbore diameter or
perforations creates an enlarged area in the annulus less (usually 1 to 3 ft) reduces or eliminates multiple
while providing zonal isolation along the wellbore. near-wellbore fractures and especially the tortuosity
The dissolved acid-soluble cement around the associated with longitudinal fracture initiation
casing at the perforations minimizes tortuosity followed by transverse propagation.
and fracture entry pressure effects. Skin effects, For longitudinal fracturing, oriented perforating
reduced near-wellbore conductivity and perforation guns with 180-degree phasing, creating perforations
problems encountered with conventional that point up and down, are indicated. In certain cases,
cements are eliminated. 0-degree phasing is very desirable. To avoid penetrating
Cement slurries are usually batch mixed and the into a water zone, a well drilled in the bottom of the
densities are measured with pressurized mud scales reservoir can be perforated with 0 degree phasing,
before pumping. The slurry follows a weighted spacer with the perforations pointed up. This forces the
374
fracture to migrate upward, reducing the flow of water Multiple-trip, multiple-zone, tubing-conveyed
from the bottom. Abass et al. (1993) even suggested perforating may be used in a perforate/stimulate/
drilling horizontal wells in competent but non- isolate operation. The first zone is perforated with
producing formations, overlaying or underlying weak tubing-conveyed guns with tractor and stimulated. A
but producing formations, and then perforating and bridge plug is run and set with tractor. The second
fracturing, thus producing through the fracture with zone is perforated and the process is repeated until all
the well never entering the formation. the zones in the lateral are stimulated. The fracturing
If phased perforations are executed, deep- operation may be conducted continuously until
penetrating perforations are not required because the complete or scheduled for one or two zones per day
fracture initiation point is at the cement/formation due to operator or wellsite restrictions.
interface and not along the perforation tunnel. External casing perforating may also be
Perforation phasing affects hydraulic fracturing. If the used in horizontal wells. The perforating and zonal
perforations are within 30 degrees of the preferred isolation systems are attached to the outside of the casing
fracture direction, near-wellbore tortuosity can be (see Section 9-7.5). The external casing perforating
minimized. Perforation phasing of 60 degrees or less is system is cemented in place with conventional
preferred (Brown and Economides, 1992). cementing techniques. The perforating modules are
For horizontal well perforating, a typical procedure selectively fired to perforate and allow the zone to be
would involve first a liner cleanout trip where the stimulated (Rodgerson et al., 2005). See Fig. 10-18. One
liner, casing and tubing are pickled with a solvent and disadvantage of the system is a limit of cased wellbore
hydrochloric acid. Acetic acid is then spotted in the diameter, as the external casing perforating gun system
lateral, and tubing-conveyed perforating guns are picked maximum casing diameter is 3.5 inches.
up and run in the well. Pumping equipment activates As with open-hole completions, creating
the pressure-actuated tubing-conveyed perforating guns. perforations in cemented completions with an abrasive
More acetic acid is then pumped into the formation and jetting tool leads to large, clean holes within a very
breaks down each perforation interval. short interval. An abrasive jetting tool can be used to
Traditional explosive charge guns require 3 to 5 create six deep, penetrating holes on top of the liner or
ft to place the number of perforations for the fracture casing and six deep holes on the bottom in less than one
treatment. This length of perforated interval may foot. Another tool allows the creation of a radial slot in
contribute to tortuosity or multiple fracture initiations. the casing, which might appear to be an ideal option
Tortuosity may require extra fracturing pressure that may if not for the potential danger of cutting the casing
not be available due to pressure limitations of the casing entirely and shifting it. Both tools perform well in field
or liner, and the fracture may not initiate. In addition, a applications; the important issue is minimizing the
tortuous path near the wellbore may detrimentally affect open interval. This prevents the formation of a tortuous
production. This is explained in Chapter 6. path of longitudinally initiated fracture, turning into
Single-trip, multiple-zone, tubing-conveyed transverse (Demarchos et al., 2006).
perforating can be used to achieve high-quality Coiled tubing provides a method to deploy
perforations. For example, in a lateral 3,600 ft long, up a bottomhole jetting assembly. The bottom hole
to 12 zones can be shot on a single trip. Point-source assembly is moved to the first zone to be stimulated.
cluster perforating is designed to achieve a limited-entry Pumping down the coiled tubing string, as outlined
fluid distribution for an equal volume of treatment in Section 10-3.1 creates the perforations.
fluid in each perforated zone. Fewer perforations are
shot in the clusters near the heel of the lateral (highest 10-5.3 Zonal Isolation in Cased Completions
perforation pressure drop) and more perforations
are added to the clusters approaching the toe (lowest To isolate a zone for treatment, drillable composite bridge
perforation pressure drop). The primary application is plugs are conveyed with coiled tubing and set electrically
acid fracturing (Willett et al., 2002). or hydraulically. A hydraulically set bridge plug is the
375
Modern Fracturing
faster method. As soon as pumping stops and the well is The pressure differential across the perforations
under control, coiled tubing can be rigged up to run the is highest in the heel section and decreases to the
bridge plug to depth and set it. The advantage of setting lowest in the toe section. The heel sections are the
the bridge plug electrically is that the use of electric line first to encounter the proppant laden fluid and the
coiled tubing allows a plug to be set on depth with a collar erosion effects will enlarge the perforations in the heel
locator for more accuracy. (Demarchos et al., 2006). sections more rapidly than the toe sections. More
In the external casing perforating system, flapper treatment fluid will be placed in the heel section and
valves provide zonal isolation. The flapper valves are less in the toe section. Allowances can be made for the
protected from stimulation fluids during fracturing erosion effects by reducing the number of perforations
operations by a sliding sleeve. At the end of the first in the heel section and/or increasing the number of
treatment, the second perforation module is fired and perforations in the toe section.
the process allows the sliding sleeve to open and a
ceramic flapper valve closes to isolate the first zone. 10-6 Fracturing of Cased-Hole
The second zone can be immediately fractured. The Completions
process of perforating, isolating and fracturing each
zone individually can continue with less than 15 10-6.1 Acid Fracture Execution
minutes between treatments until all the zones are
stimulated (Rodgerson et al., 2005). As in open-hole fracturing, live acid must be
Another method of isolation is used in transported as far away from the wellbore as possible
some applications where coiled tubing deploys to dissolve and etch the carbonate in order to create
a jetting tool to cut perforations. In these cases, conductive fractures. But in cased holes, such
the fracture treatment is pumped down the casing/ treatments are a lot easier to accomplish than in open
coiled tubing annulus. A final proppant stage holes. The perforating methods discussed in Section
of non-crosslinked fluid with high proppant 10-3.1 and the zonal isolation techniques discussed in
concentration is pumped to induce a near-wellbore Section 10-3.2 can be applied to stimulate multiple
proppant pack that acts as a diversion for the next zones in a cased-hole horizontal completion.
treatment (see Section 13-1.4). In addition, diversion techniques can be far
Limited entry fracturing uses perforation friction more easily applied to acid fracturing, than to
pressure to control the stimulation fluid distribution into proppant fracturing (see Section 4-7.8.3). Therefore,
each perforated interval (see Section 9-7.3). Limited entry it is feasible to pump an acid fracture treatment
fracturing has been successfully used in vertical wells. In into a wellbore containing several open perforated
horizontal wells, limited entry fracturing incorporates intervals. Effectively applied diversion practices,
fluid friction pressure loss along the long lateral, little or generally consisting of a combination of ball
no change in hydrostatic pressure and perforation erosion sealers and viscosity-contrast techniques, allow the
effects caused by proppant laden fluid. Pipe friction fracturing fluid to be efficiently diverted away from
pressure in long horizontal wells may limit the maximum the weakest interval to the next weakest interval.
pumping rate. Perforation erosion effects, which may If several diversion stages are employed alternately
only have a minor effect on fluid distribution in vertical with several acid fracture stages, it is possible to
wells, may significantly affect the fluid distribution in a place acid fracture treatments effectively over
horizontal stimulation treatment. several perforated intervals.
Figure 10-18 External casing perforating (from Rodgerson et al., 2005)
376
10-6.2 Proppant Fracturing Execution ranged from 1,500 to 4,000 ft. The bottomhole static
temperature was 145 oF. Table 10-7 contains a summary
One method for selective placement of propped hydraulic of these fracture treatments.
fractures is to perforate, stimulate and isolate a zone, clean Five days were required to stimulate three
up the zone and perform a well test. The length of time zones completed with the perforate, stimulate
required to perform this method in a multi-fractured and isolate system. The external casing perforating system
horizontal well may not be economical. allowed from six to 13 zones to be completed in one
An alternate method is to minimize the total day of continuous operation. Treatment design called
completion time by performing the perforate, for placing 200 lbs of proppant and pumping 1000 gal
stimulate and isolate cycle as quickly as possible. of fluid per foot of lateral length. In Well A, individual
Perforating and zonal isolation methods outlined in chemical tracers were run in each frac stage. Figure 10-
Sections 10-5.2 and 10-5.3 are applicable. Cleanup of 19 shows that after six to eight hours of flowback, each
all the stimulated zones is performed at the same time of the six fractured zones was flowing. The initial and
after all zones are stimulated. long-term production from both completion methods
A case history describes successful stimulation of a was comparable (Rodgerson et al., 2005).
tight gas sand formation. A 2500 ft horizontal lateral
1200
was drilled and an external casing perforating with zonal
Stage 1 Stage 3 Stage 5
isolation system cemented in place. The total vertical
1000
depth of the well was 10,600 ft and the measured depth Stage 2 Stage 4 Stage 6
was 13,200 ft. The bottomhole static temperature was

Tracer Concentration, ppb
800
280 oF. Each proppant fracture stage was initiated with
a 3 ft perforation module. A total of eight stages were 600
completed in 9 hours. The average pump rate for each

treatment was 40 bbl/min with an average surface 400
treating pressure of 8000 psi, and a total of 1.3 million

200
lbs of proppant placed. Fracture gradients measured in
the eight stages ranged from 0.95 to 0.97 psi/ft. Eight 0
hours after completion, the well was flowing gas to
0.58
1.25
2.42
4.52
8.50
15.50
29.50
45.50
73.50
136.50
196.50
311.50
436.50
502.50
sales and recovered 50 percent of the fracturing fluid
in 5 days. One coiled tubing intervention removed the Elapsed Flowback Time, hr
isolation valves. The initial and long-term production

Figure 10-19 Well A chemical tracer response (from
of the horizontal well is significantly greater than offset Rodgerson et al., 2005)
fractured vertical wells (Rodgerson et al., 2005).
Table 10-7 External Casing Perforating System Well 10-7 Rationale and Conditions
Summary (from Rodgerson et al., 2005)
of Fracturing Horizontal Wells
Well A Well B Well C Well D
Number of Modules
in Gas Formation
6 13 13 10
/ Frac Stages
Lateral Length 1,543 ft 2,822 ft 4,160 ft 2,500 ft
True Vertical Depth 4,450 ft 4,150 ft 4,400 ft 4,250 ft The first issue is whether a horizontal well makes
Measured Depth 6,200 ft 7,800 ft 8,400 ft 7,000 ft any sense or not. Starting with the basic premise of
this book that all wells in natural gas reservoirs of
A second case history describes shale gas formation any permeability should be hydraulically fractured,
has been successfully completed with the perforate, then the work by Diyashev and Economides
stimulate and isolate system and the external casing (2006) is relevant. They found that for reservoirs
perforating system. Laterals in the four offset wells with permeability larger than 10 md, properly
377
Modern Fracturing
designed and undamaged horizontal wells would leading to unacceptable reduction in the rate
perform better than vertical wells with fractures. per fracture. This was discussed extensively
The problem is that an undamaged horizontal well is in Section 10-2.
difficult to achieve in the field. The last point leads to a sandwich between the
Bypassing damage is one of the important attributes economic and the physical. In general, transversely
of fracturing, and horizontal wells with fractures fractured horizontal wells in gas reservoirs have a
would benefit from this. For gas wells, reducing very narrow band of attractive reservoir permeability,
turbulence is of paramount concern, and fractures e.g., 0.1 md < k < 0.5 md.
are the most effective way to accomplish this.
Furthermore, stimulation would provide additional This discussion is summarized in Table 10-8
incentives, and all these benefits are available in
fractured horizontal wells. Table 10-8 Suitable Options for Fracturing Gas Wells
Although for horizontal oil wells there is Permeability
Best Technical Solution Comments
Range, md
clear delineation between longitudinal (for Horizontal Wellbore,
>5 In all cases
higher permeability) and multiple transverse Longitudinal Fractures
fractures (for lower permeability, see Brown Dependent upon
Horizontal Wellbore,
project economics and
and Economides, 1992), the delineation for gas Longitudinal Fractures
the relative costs of
0.5 to 5 OR
wells is not that clear. vertical and horizontal
Vertical Well with
wellbores and zonal
Longitudinal fracturing of horizontal gas wells will Fracture
isolation techniques
benefit any reservoir above 5 md. The performance of Above 0.5 md, the
choked connection
such wells would be as described by Villegas et al. (1998) between the fracture
0.1 to 0.5 Horizontal Wellbore,
and discussed in Section 10-1. Reducing turbulence is Transverse Fractures
and the wellbore
makes transverse
perhaps more important than the stimulation itself, as fractures relatively
expounded upon in Chapter 2 of this book. inefficient
Dependent upon
For transverse fracturing in gas reservoirs, Horizontal Wellbore,
project economics and
Transverse Fractures
the situation is more complex. Although for < 0.1 md OR
the relative costs of
vertical and horizontal
vertical wells in low-permeability (e.g., k < 0.1 Vertical Well with
wellbores and zonal
Fracture
md) formations, fracturing is the only way to isolation techniques
monetize them, multiple transverse fractures in
horizontal wells will provide major incremental
benefits over vertically fractured wells.
There are two reasons: 1) the choke effects (Section References
2-5.3) caused by the inefficient contact between
the well and transverse fractures in such low- Abass H.H., Wilson, J.M., Venditto, J.J. and Voss,
permeability reservoirs is not significant; and 2) R.E.: Stimulating Weak Formations Using
interference among multiple fractures is minimal. New Hydraulic Fracturing and Sand Control
However, these incremental benefits are subject Approaches, paper SPE 25494, 1993.
to economic considerations. Are they attractive Austin, C.E., Rose, R.E., and Schuh, F.J.: Simultaneous
compared to incremental costs of drilling horizontal Multiple Entry Hydraulic Fracture Treatments of
wells, multiple fracturing and zonal isolation? In Horizontally Drilled Wells, paper SPE 18263,
established areas such as the United States the 1988.
answer is yes; elsewhere it is not so clear. Bartholomew, P., Portman, L., Frost, R., Nunn. P.,
As reservoir permeability increases (e.g., k > Vause, P., Edmondson, D., Andrich, L. and
0.5 md) the choke formed by the inherently LaGrange, T.: Near a Kilometer of Perforating
inefficient contact between a transverse fracture Guns, in a 7-km, Extended Reach WellCoiled
and the wellbore adds additional turbulence, Tubing Shows its Mettle in New Zealand, paper
378
SPE 101065, 2006. External Casing Perforating Provides Optimal

Behrmann, L.A. and Nolte, K.G.: Perforating Treatment Coverage in Horizontal Pay, paper SPE
Requirements for Fracture Stimulations, paper 97175, 2005.
SPE 39453, 1998. Seale, R., Donaldson, J., and Athans, J.: Multistage
Brown, J.E. and Economides, M.J.: An Analysis of Fracturing System: Improving Operational
Horizontally Fractured Horizontal Wells, paper Efficiency and Production, paper SPE 104557,
SPE 24322, 1992. 2006a.
Daal, J.A., and Economides, M. J.: Optimization of Seale, R., Themig, D., and Athans, J.: Effective
Hydraulically Fractured Wells in Irregularly Shaped Stimulation of Horizontal Wells A New
Drainage Areas, paper SPE 98047, 2006. Completion Method, paper SPE 106357, 2006b.
Demarchos, A.S., Porcu, M.M., and Economides, M.J.: Villegas, M.E., Wattenbarger, R.A., Valk, P., and
Transversely Multi-Fractured Horizontal Wells: A Economides, M.J.: Performance of Longitudinally
Recipe for Success, paper SPE 102262, 2006. Fractured Horizontal Wells in High-Permeability
Diyashev, I., and Economides, M.J.: A General Anisotropic Formations, paper SPE 36453,
Approach to Well Evaluation, SPEPO (August 1996.
2006) 394-401. Willett, R.M., Borgen, K.L., and McDaniel, B.W.:
East, L., Rosato, J., Farabee, M., and McDaniel, Effective Stimulation Proved to be the Key
B.W.: Packerless Multistage Fracture-Stimulation to Economic Horizontal Completions in Low
Method Using CT Perforating and Annular Path Permeability Carbonate Reservoir, paper SPE
Pumping, paper SPE 96732, 2005. 76725, 2002.
Economides, M.J., Hill A.D., and Ehlig-Economides,
C.A..: Petroleum Production Systems, Prentice
Hall, NY, 1994.
Erwin, M.D., and Ogbe, D.O.: Predicting Openhole
Horizontal Completion Success on the North
Slope of Alaska, paper SPE 97121, 2005.
Henriksen, K.H., Augustine, J., and Woods, E.:
Integration of New Open Hole Zonal Isolation
Technology Contributes to Improved Reserve
Recovery and Revision in Industry Best Practices,
paper SPE 97614, 2005.
Marongiu-Porcu, M.: Physical and Economic
Optimization of Hydraulic Fracturing, MS
Thesis, University of Houston, 2007.
McDaniel, B.W., and Willet, R.M.: Stimulation
Techniques for Low-Permeability Reservoirs
with Horizontal Completions that Do Not Have
Cemented Casing, paper SPE 75688, 2002.
McDaniel, B.W., East, L., and Hazzard, V.: Overview
of Stimulation Technology for Horizontal
Completions without Cemented Casing in the
Lateral, paper SPE 77825, 2002.
Navarro, J.B.: Slotted-Liner Completions Used in
the first Horizontal Wells in Mexico, paper SPE
37110, 1996.
Rodgerson, J.R., Ruegamer, M.L., and Snider, P.M.:
379
Modern Fracturing
380
Gary Schein is the region technical manager for BJ Services Company in Dallas and is been responsible for
the treatment design and completions engineering for the East Region including northern and eastern Texas,
northern Louisiana and Mississippi. Schein has a bachelor's degree from Northern Arizona University. He
has worked in well completions and stimulation for over 29 years and has worked in several other areas
including R&D, technical support, marketing and field engineering positions. Most of his focus has been
on improving production from unconventional reservoirs such as shale gas and tight gas sands. He was a
2004-2005 SPE Distinguished Lecturer and has served on SPE technical paper selection committee for well
completions. He has also authored several SPE papers on well stimulation and completion.
David Mack is a senior staff completion engineer with Marathon Oil Company in Houston. He has a
petroleum engineering degree from Colorado School of Mines. For 27 years he held various research and
operational positions with pressure pumping companies in northern Texas, the Rocky Mountains, and the
Mid-Continent and Appalachian Basins. The past four years, he has worked with Marathon in the Reservoir
and Well Performance Group. He has authored some 20 SPE papers and trade journal articles and has served
on regional and national SPE technical conference committees and section leadership teams. He is listed as
inventor on four US patents.
Chapter 11 the porosity and cannot be connected to the wellbore
through natural permeability. Wells drilled in tight
Unconventional Gas gas reservoirs require large stimulation treatments to
Gary W. Schein, BJ Services and gain commercial production, as described in previous
David J. Mack, Marathon Oil Company chapters. The Energy Information Administration
(EIA) estimated that over 21 percent of the natural
gas reserves in the United States are from tight gas
11-1 Introduction reservoirs (NaturalGas.org, 2000).
Unconventional gas reservoirs offer unique sets of Geo-Pressured Zones

challenges that are not commonly found in traditional Geo-pressured zones are typically found in the US
reservoirs. The term unconventional gas typically refers Gulf Coast Region and were formed as a result of
to six types of reservoirs. These include tight gas sands, compaction of silt and clays, causing the natural gas to
geo-pressured zones, deep gas, methane hydrates, coalbed be forced under high pressure into bounding sandstone/
methane, and shale gas. The most common reference to siltstones. These geo-pressured zones are found well
unconventional gas production is associated with tight below 10,000 ft to the extreme depths of greater than
gas, coalbed methane and shale gas. These reservoirs have 30,000 ft. Because of the extreme conditions, the
provided the most interest over the past 10 to 20 years. The production of gas from these reservoirs is extremely
volatility of natural gas prices prevented the economical costly. However as a result of deepwater technology
commercialization of much of these gas reservoirs development, the commercial production of gas from
for many years. The emergence of new technology, geo-pressured zones is not far away.
government regulation/deregulation, increasing stability
of natural gas prices, and the demand for cleaner-burning Deep Gas
fuels provided the opportunity for energy companies to The NGSA defines deep gas as reserves found at depths
capitalize on producing and replenishing reserves. greater than 15,000 ft (NaturalGas.org, 2000). At one
Before we go any further, it is important to provide time, deep gas was considered uneconomical because
a brief description of each of these types of reservoirs to drilling to depths in excess of 15,000 ft were difficult, at
better gain an understanding of the primary focus of this best, using conventional drilling techniques. With the
unconventional gas chapter. Because the production improvements in drilling technology and geoscience
of unconventional gas resources is appropriate for a methods to find these reservoirs, deep gas is now
mature and yet technologically developed petroleum becoming more conventional and is often associated with
environment, such as the United States, this chapter is other tight gas sands. These deep reservoirs have been
intended to describe the completion and stimulation located in the Gulf of Mexico, Permian basin region of
of these resources and to serve as a technology West Texas, southern Louisiana, Mississippi, Alabama,
transfer for other places in the world where operators Rocky Mountains and Western Oklahoma in the USA.
are contemplating the exploitation of their own Deep gas reservoirs have also been identified of the coast
unconventional gas resources. of Nova Scotia as well as areas within the Middle East,
South China Sea and Siberia. As technology continues
11-2 Description of to improve, these deep gas reservoirs will account for
Unconventional Reservoirs significant reserve replacement.
Tight Gas Natural Gas Hydrates

Tight gas sands typically refer to gas reservoirs where There has recently been much discussion concerning
the permeability of the rock is less than 1 md. Most hydrates. The NSGA refers to hydrates as formations
reservoirs considered to be tight gas are sandstone or that are essentially frozen water and natural gas crystals
limestone formations in which the gas is trapped within found beneath the permafrost and under the ocean floor.
383
Modern Fracturing
Hydrates can be found in the Arctic coastal regions of reserves in Canada, India, Russia, China and Australia as
both the USA and Canada, along the East Coast of the well. Figure 11-1 shows the extent of many of the worlds
USA, Gulf of Mexico, offshore Oregon, and Gulf of coalbed methane resources from Scott, 2004. Additional
Mexico (Yougquist and Duncan, 1993). Some suggest data can be found at www.eia.doe.gov (2007). The CBM
that gas hydrates may contain many times more fossil reservoir and production mechanism will be further
energy than is found in conventional oil, gas, and coal described later in this chapter.
deposits (Kvenvolden, 1993). The potential hazards
and environmental concerns relating to production of Shale Gas
this type have prohibited any significant attempts at As a result of increasing natural gas prices and the
producing this resource. Although natural gas hydrates increasing need to replace reserves, production from
are considered unconventional gas, they will not be shale reservoirs has gained significant interest in the
discussed in this chapter because they have not provided recent years. Although it seems as though this is a
significant impact on natural gas production. recent activity, the earliest gas production from shale
dates back to approximately 1821, when it was used to
Coalbed Methane light street lamps in Baltimore, Maryland (NaturalGas.
Coalbed methane (CBM) is produced from gas that was org, 2000). Shale reservoirs are best described as fine-
trapped within the coal and was formed under similar grained sedimentary rock inter-bedded with siliceous
circumstance as conventional oil and gas. Traditionally, and carbonaceous material. The properties of the
coalbed methane presented a safety concern and created shale result in its ability to be easily cracked both
problems with mining and producing coal. With the vertically and well as laterally. The Devonian Shale
economic environment and issues associated with venting in the Appalachian region provided initial substantial
methane to the atmosphere, CBM has become a viable production. After significant production from this area,
and profitable resource. The EIA estimated in 2000 that shale reservoirs were discovered in Michigan (Antrim
recoverable reserves of greater than 60 Tcf exist within Shale), Illinois (New Albany Shale), the San Juan Basin
the lower 48 states of the US. There are significantly more of Colorado and New Mexico (Lewis and Mancos
Arctic Ocean Arctic Ocean

16,116
4,000 269
200 161
2,060 4,383
Western
Atlantic Europe 2,546
Pacific
Former
951
Soviet Union
508 55 0
275
27
North Africa Middle

America East
Ocean Indian
Ocean
3,360
248
32
Ocean
15
524 646 South

America
Natural Gas Proven Reserves (Tcf)
Asia Coal-Bed Methane Resources (Min-Max, Tcf)
Pacific Source: BP Statisical Review (year-end 2005), and Andrew R. Scott.
Figure 11-1 Worldwide coalbed methane resources (after www.eia.doe.gov, 2007)
384
Chapter 11 Unconventional Gas
Shale), and of course the prolific Barnett Shale of the in the cleat network (Fig. 11-2). The dewatering period
Ft. Worth Basin of North Texas. Additional aspects of the time from initial production to peak production
production for these and other shale reservoirs will be can be anywhere from weeks to years.
further discussed in this chapter.
11-3 Production Mechanisms

Peak Gas Rate
11-3.1 CBM (Coalbed Methane)
Gas Rate
In conventional gas reservoirs, the gas is stored within
the rock matrix pore space and/or naturally occurring
fracture networks (secondary porosity). The gas is then
Time to Peak Gas Rate
produced through the permeability connecting the pore
space and existing natural fracture networks. Time
Such is not the case for a CBM reservoir, which Figure 11-2 Typical CBM production profile (from Roadifer
is naturally fractured, low-pressured and normally et al., 2003)
water-saturated (Seidle and Aril, 1990). Here the
gas is stored as free gas in the natural fracture cleat Numerous methods are employed to dewater a
system (secondary porosity) and in the surface matrix CBM reservoir. Methods include hydraulic gas pump,
of the coal by adsorption. (See Section 11-4.2 for a plunger lift, rod pump, progressive cavity pump and
discussion of the cleat system.) Adsorption typically others. Simpson et al. (2000) and Bassett (2006)
accounts for more than 99% of the gas-in-place in a provide excellent overviews of dewatering methods,
CBM reservoir. (Roadifer et al., 2003). operations and considerations.
Production from a CBM reservoir is therefore
almost exclusively from the desorption of gas from 11-3.2 Shale Gas Reservoirs
the coal matrix by depressurizing the coals (Holditch,
1993). Because most CBM reservoirs are 100% water- Shale gas is stored as natural gas in the reservoir in
saturated in the cleat system, the water must be removed three ways:
in order to achieve any appreciable gas production 1. Free gas. This is gas that occupies the pore space/
(Seidle, 1993). The process is termed dewatering. matrix of the rock, and it is similar to more
Dewatering reduces the hydrostatic pressure, conventional as well as other non-conventional
allowing the gas to desorb from the coal matrix. In reservoirs such as tight gas sands.
conjunction, the water saturation is reduced, thus 2. Adsorbed gas. This is gas attached to the organic
increasing the relative permeability to gas, permitting matter within the shale. Because shale reservoirs are
the desorbed gas to flow through the cleat network to fine-grained mixtures of clays, silica material and
the wellbore. (McGinnis et al., 1982). sometimes carbonate material with an extremely
Desorption of the gas from the coal matrix changes low permeability rock matrix, the organic-rich
the fluid mobility in the cleat network. This creates nature of shale reflects the potential for adsorbed
another production feature unique to CBM reservoirs: gas. The adsorbed gas is directly related to the
an initial increase in the gas production rate (negative degree of thermal maturity of the shale system and
decline rate) (Roadifer et al., 2003). A conventional is typically determined as a measure of vitrinite
gas reservoir typically has peak production rates at the reflectance (VRo). Because vitronite (kerogen)
beginning of the production cycle followed by a rate changes both predictably and consistently upon
decline. Unlike a conventional reservoir, a CBM reservoir heating, the relative reflectance can be measured
will display an increasing rate as the coal is dewatered, gas utilizing a microscopic spectrophotometer. VRo
desorbs from the coal matrix and gas mobility increases is a method to identify the temperature history of
385
Modern Fracturing
sedimentary rocks. It is generally accepted that the TOC is measured in the laboratory using a
use of this measure can indicate the maturity of the process called pyrolysis. Pyrolysis is a form of
hydrocarbon source. It is also generally accepted that incineration that chemically decomposes organic
the onset of oil generation in low-maturity shale will materials by heat in the absence of oxygen. A core or
range from 0.5 < 1.0 % VRo. Although oil generation formation sample must be obtained to perform this
will usually terminate with reflectance ranging from measurement. In general, the TOC content measures
0.85 - 1.1 %, a VRo in the 1.0 - 1.4 % ratio will the amount of rock material that is available that
typically result in higher thermal maturity and will can convert to hydrocarbons. The higher the TOC
reflect higher gas flow rates (Jarvie, 2006). content, generally, the better. Lower TOC contents
3. Gas from Natural Fracture Systems. Because in can also be offset by a number of other factors such
shale reservoirs the rock matrix can be of extremely as VRo, presence of natural fractures, thickness of
low permeability (< 0.001 md), the existence and interval, porosity, water saturation, etc.
presence of natural fractures become a major factor
in the commercialization of the current major 11-3.3 Shale Gas Reserves
producing shale reservoirs such as the Devonian,
Lewis, and Barnett shales. The gas shale production that has occurred over a
A significant factor in analyzing a shale reservoir period of 15 years in the United States was the result of
is to determine if a natural fracture system exists government incentives and tax credit. However when the
and how it was created. A common misconception tax credit expired in 1992 many operators that had found
is that if natural fractures are found, they are success continued to expand drilling and completion
open and can produce without stimulation. These programs targeting gas shale reservoirs, indicating that
fractures typically occurred because of tectonic the financial incentives provided by the government have
stresses and are a result of the combination of been successful. Outside the USA, where conventional gas
silt/clay, siliceous content and carbonate content. reserves are more abundant (see Chapter 1), production
Many times if natural fractures are present, they are from gas shales is not significant.
small healed fractures that are not inter-connected There have been many attempts to accurately
or well-connected to the wellbore. estimate the reserves from these unconventional gas
An additional measure of the quality of shales; however, with the ever-expanding exploration of
these organic-rich shale reservoirs is to determine new gas shale plays the reserve estimates will certainly
the Total Organic Carbon (TOC) content. continue to rise. This will be a direct result of product
Table 11-1 Comparison of Producing Shale Basins in the U.S. (From Faraj, 2004 and Pickering Energy Partners, 2005)
Barnett Ohio Antrim New Albany Lewis Fayetteville
Depth, ft 6,500-8,500 2,000-5,000 600-2,200 500-2,000 3,000-6,000 1,500-6,500
Gross Thickness, ft 150-700 300-1,000 160 180 500-1,900 50-325
Net Thickness, ft 100-600 30-100 70-120 50-100 200-300 20-200
Bottomhole Temp, F 200 100 75 80-105 130-170
TOC, % 4.5 0.0-4.7 1-20 1-25 0.45-2.5 4.0-9.5
Total Porosity, % 4-5 4.7 9 10-14 3.0-5.5 2-8
Gas Filled Porosity, % 2.5 2 4 5 1-3.5
Water Filled Porosity, % 1.9 2.5-3.0 4 4-8 1-2
Flow Capacity - kh, md-ft 0.01-2 0.15-50 1-5,000 6-400
Gas Content, scf/ton 300-350 60-100 40-100 40-80 15-40 60-220
Adsorbed Gas, % 25 50 70 40-60 60-85 50-70
Reservoir Pressure, psi 3,000-4,000 500-2,000 400 300-600 1,000-1,500 600-2,000
Pressure Gradient, psi/ft 0.43 0.15-0.40 0.35 0.43 0.20-0.25
Water Production, Bwpd 0 5-500 5-500 0
Well Spacing, Acres 40-160 40-160 40-160 80 80-320
Recovery Factors, % 10-20 10-20 20-60 10-20 5-15
Gas-In-Place, Bcf/sq. mile 50-150 6-15 6-15 7-10 8-50 25-60
Reserves, MMcf 500-4,000 150-600 200-1,200 150-600 600-2,000
386
pricing and new technology to unlock these reservoirs. Coals are heterogeneous and anisotropic porous
Recent estimates have placed recoverable reserves from media. They are characterized by three distinct porosity
the current gas shale plays in the United States at systems: macropores, micropores and nanopores (Fig.
approximately 60 Tcf of gas. (NaturalGas.org, 2000). 11-3). Macropores are the natural fractures or cleats
To put this in perspective, the Gas Technology Institute common in coals. They contain free gas and water and
(GTI) and Pickering Energy Partners, Inc reported serve as the main conduit to the wellbore. Micropores
estimated production in 2005 from the Barnett shale of and nanopores are the coal matrix itself and contain
450 Bcf per year, which is approximately 58% of shale gas the vast majority of the gas.
production. The Appalachia/Ohio shale and the Antrim
20
shale accounted for approximately 18-21% each. The 18
Sorption, Vc (m3 gas/ m3 coal)

Lewis Shale of the San Juan Basin was much smaller at 16
14
+/- 3% of 2005 production. (Hayden and Pursell, 2005). 12
Table 1 provides a comparison of current producing gas 10
8
shale basins in the U.S. This demonstrates the variability 6
4
that is seen in productive shales and provides insight as
atmos atmos
below above
2
to the unique nature of these reservoirs. 0
2
4
11-4 CBM Reservoirs 0
2
4
6
8
0
0
atmos 0.
0.
0.
0.
1.
2.
3.
4.
Absolute Pressure, MPa
11-4.1 Coalbed Description Figure 11-4 Sorption isotherm at 33.5 C, gas mixture of
95% methane and 5% CO2 (from Gray, 1987)
Coal is a sedimentary rock that contains more than 50
weight percent and 70 volume percent organic material, Well Well Well
A Permeability B C
predominantly carbon, hydrogen and oxygen, along with Facies Change
inherent moisture. Hydrocarbon production from coal

is generally termed methane. However, the produced Channel
Sandstone Belt
gas is typically a mixture of methane with some ethane
Coal
and traces of propane, butane and heavier hydrocarbons Pinch
Out
along with nitrogen and carbon dioxide. Fault
Offset
Figure 11-5 Coalbed reservoir geometry (from

Nelson, 1999)
CBM reservoirs are different from conventional

gas reservoirs in three distinct characteristics (Gray,
1987). First, coal is both the reservoir and the source
rock. Second, gas is stored by adsorption in the coal
matrix instead of by compression in pore spaces as in
conventional sandstone or carbonate reservoirs. The
CBM pressure-volume relationship is defined by the
desorption (adsorption) isotherm, in place of the real gas
Figure 11-3 Schematic of coal structure showing a dual-
law. And third, the cleats are typically filled with water.
porosity system (plan view) (from King et al., 1986) If the gas content in the coal matrix is below equilibrium
387
Modern Fracturing
or the reservoir pressure is above a critical pressure (below (Gray, 1987). One is reservoir compaction, which will
which the gas will start to desorb), it will be necessary increase the effective horizontal stress and subsequently
to produce the water, known as de-watering, to reduce decrease cleat permeability. The other is desorption of gas
the pressure below the equilibrium pressure allowing the (primarily methane) from the coal matrix, causing the
gas to desorb. Figure 11-4 is a typical sorption isotherm. coal matrix to shrink, which reduces the horizontal stress
It shows the amount of gas that will be sorbed into the and allows cleat permeability to increase.
coal matrix as a function of pressure change at a constant Irving Langmuir (1881-1957), who was awarded
temperature. Data for these curves are obtained directly the 1932 Nobel Prize in chemistry (the first ever
from a volumetric standpoint. industrial chemist to receive the recognition) and is
A CBM reservoir typically contains multiple considered the father of surface chemistry, developed
coal layers or seams and can be complex in a remarkably simple expression relating the mass of
nature with many types of common geological gas adsorbed on a surface:
heterogeneities (Fig. 11-5) p
Ga = ,
1+ p (11-1)
11-4.2 CBM Fractured Systems
where Ga is the mass (or standard volume) of gas
Production of gas from a coal seam is primarily adsorbed, p is the pressure and is a constant,
through the natural fracture or cleat system. The cleats experimentally determined (Langmuir, 1960). The
are well-defined and almost uniformly distributed. constant is the Langmuir adsorption constant,
There are two subdivisions in the cleat system. The which increases with the strength of adsorption
face cleat is continuous throughout the reservoir and between a solid and a gas, and decreases with
generally normal to the bedding plane. The butt cleat is temperature. For a given temperature the relationship
discontinuous and terminates at the intersection with between Ga and p becomes a Langmuir isotherm.
the face cleat. Figure 11-3 is a schematic of the system. Variants of Eq. 11-1 appear in a great number of
Figure 11-6 depicts two coals, showing the general publications, including some that have different
orthogonal nature of the cleats. constants for different components, etc.
Carbon dioxide is frequently injected for enhanced
recovery/CO2 sequestration. The CO2 adsorbs on the
coal because it has a greater adsorption capacity than
methane, causing the coal matrix to swell. The swelling
will in turn cause an increase in the effective horizontal
stress and a decrease in the cleat permeability. The
reduction in permeability is estimated to be as much
as two orders of magnitude at the Allison pilot in the
Figure 11-6 Gibson coal (left) and Anderson coal (right),
San Juan Basin (Pekot and Reeves, 2003). It should
both 2.0 in. (5.08 cm) in diameter (from Robertson, 2005) be noted here that changes in cleat permeability and
fracture width are directly related.
These cleats provide the primary permeability
of a coal and are the most important parameter for Permeability Determination
commercial CBM production. The face and butt Determining the permeability and cleat width has
cleats are usually subvertical. As a result, changes in been an evolving study. Based on the orthogonality of
cleat permeability are dependant on the horizontal the cleats, Palmer and Mansoori (1998) used a cubic
stresses acting across the cleats. relationship between permeability and porosity to
As the reservoir pressure is depleted through evaluate changes in permeability:
dewatering and primary production, two distinct and
k 3
opposing effects occur related to cleat permeability = 3. (11-2)
ko o
388
The relationship between and o is described by: matrix volume resulting from the desorbing or
adsorbing of gas) (ft3/scf ), and b is the Langmuir
1 K p p constant (Langmuir, 1960) (psi-1).
o = ( pr po )1 l r o ,
M M pr + P po + P Robertson and Christiansen (2006) based their
work on the cubic nature of the cleat system as a
(11-3) function of cleat or fracture width, w, and the spacing
between the cleats or fractures, a, where w and a are
where k is the current formation permeability (md), ko is consistent units and k is millidarcies:
the original formation permeability (md), is the current
w3
formation porosity (fraction), o is the original formation
k= . (11-7)
12a
porosity (fraction), M is the constrained axial modulus
(psi) (see Eq. 11-4), pr is the current reservoir pressure However, Robertson and Christiansen (2006) cite
(psi), po is the original reservoir prerssure (psi), K is the three factors that influence changes in the cleat
bulk modulus (psi), P is the matrix-shrinkage Langmuir width as different stresses are applied to the coal:
curve-fitting parameter (psi) and l is the maximum pore volume compressibility, matrix compressibility
matrix shrinkage strain (dimensionless) at V = VL (where governed by mechanical elastic moduli and sorption-
V is the gas content in scf/ft3 and VL is the Langmuir induced strain. They derived equations to describe
volume (the gas volume at infinite pressure) (scf/ft3)). all three factors and present a relationship between
The constrained axial modulus, M, is related to permeability and initial permeability:
the Youngs modulus and the Poissons ratio (Palmer 1e( p po )

3 12 max
pL + pr
3co ( pr po ) ln
and Mansoori, 1998): k

+
o E

pL + po pL + po
(11-8)
= e

,
ko
1
M
= E . (11-4) where co is the original formation compressibility (psi),
(1 + )(1 2 )
is the rate of change in cleat or fracture compressibility
Mavor and Gunter (2004) subsequently extended (dimensionless), pr is current pore pressure (psi), po is
this equation to include enhanced CBM (ECBM) original pore pressure (psi), pL is the Langmuir pressure
recovery and CO2 storage. (the pressure at which sorption induced strain equals
Shi and Durucan (2005) developed a dynamic half of the maximum (Langmuir) volume) (psi) and
permeability model based on the model presented by max is the linear strain at infinite pore pressure on an
Seidle et al. (1992): unconstrained sample (dimensionless).
1.2

k = ko e
3c ( )
frac o
, (11-5)
where cfrac is fracture compressibility (psi-1), is the

Permeability ratio, k/k0
0.8
current formation hydrostatic stress (psi) and o is the
original formation hydrostatic stress (psi).
Their model defines the change in horizontal stress
0.4
as:
v E V bp bpo Shi-Durucan model
o = ( pr po ) + s L r , Palmer-Mansoori model
1 v 3(1 v) bpr + 1 bpo + 1 Robertson-Christiansen model
0
0 200 400 600 800 1000 1200
(11-6)
Average pore pressure, psia
where is the Poissons ratio (dimensionless), Figure 11-7 Model comparison of permeability changes
E is the Youngs modulus (psi), s is the matrix for nitrogen flowing through an average coal core
as pore pressure is lowered (from Robertson and
shrinkage/swelling coefficient (the change in Christiansen, 2006)
389
Modern Fracturing
Figures 11-7 and 11-8 compare the three to experimental binary and ternary data. However, a
models as presented by Robertson and Christiansen study by Clarkson (2003), showed that the variation
(2006), concluding that the effect of sorption- of the CO2 separation factor. With changes in pressure
induced strain on permeability is exaggerated in resulted in less than desirable adsorption predictions
the other two models. with the extended Langmuir equation.
Thermodynamic considerations were used to develop
5
Shi-Durucan model both the ideal and real adsorption solutions, with both
Palmer-Mansoori model
4 Robertson-Christiansen model
using the real gas equation of state (Manik et al., 2002;
Stevenson, 1997; and Hall et al., 1994). Stevenson
Permeability ratio, k/k0
3
compared these models to the measured adsorption of a
ternary gas (methane, carbon dioxide and nitrogen). He
2
found that the ideal adsorption solution gives acceptable
predictions in many coalbed reservoir conditions,
1
while neither model produces acceptable predictions
at high pressure conditions.
0
Clarkson and Bustin (1999) further compared
0 200 400 600 800 1000 1200 the ideal adsorption solution using the Dubinin-
Average pore pressure, psia Radushkevich, Rubinin-Astakhov and Langmuir
Figure 11-8 Model comparison of permeability changes isotherms and the extended Langmuir equations
for methane flowing through an average coal core
as pore pressure is lowered (from Robertson and
with experimental binary gas data (methane and
Christiansen, 2006) carbon dioxide) They found that the ideal adsorption
solution using the Dubinin-Astakhov isotherm
11-4.3 Adsorption/Desorption was the most accurate. However, Hall et al. (1994)
showed that the ideal adsorption solution with the
Adsorption is the collecting and holding of gas on Langmuir isotherm and the extended Langmuir
the surface of a solid. Conversely, desorption is the equations yielded a better match to component data
release of gas from the surface of a solid. Coals can than the other two methods.
adsorb a large quantity of gas due to their typically The diffusion process component of the gas
very large internal surface area, which can range from adsorption rate was originally described as a one-
30 to 300 m2/gm (Berkowitz, 1985). Other factors step process through the development of pseudo-
and properties of coal will affect adsorption; however, steady state and unsteady state relationships. A
surface area is by far the largest factor. good review of these models is presented by King
Considerable work has been done regarding and Ertekin (1989a and 1989b).
adsorption and desorption of gas mixtures as it relates to However it has been observed that diffusion
coal mine safety. (Joubert et al, 1973; Ruppel et al, 1974; through the coal matrix may be better described by
Greaves, 1993; Yang and Saunders, 1985; Chaback et al, a bidisperse diffusion model (Siemons et al., 2003;
1996 and De Gance et al , 1993) Mazunder et al., 2003; Ruckenstein et al., 1971; Cui
As pointed out by Wei et al. (2005), the gas et al., 2004; Shi and Durucan, 2003a and 2003b).
adsorption rate is controlled by the adsorption process This type of model uses a two-step approach to gas
and the diffusion process. Adsorption is a characteristic diffusion in a coal matrix. There is surface diffusion in
of the particular coal and diffusion is that of the gas the microporous system of the coal matrix and pore
through the coal matrix. diffusion in the mesopore and/or macropore system.
Several techniques have been suggested to describe The gas adsorption takes place in the micropores
gas adsorption. Ruthven (1984), and Yang (1987), while the mesopores and macropores are the storage
use an extended Langmuir isotherm. Arri et al. (1992) areas for free gas and provide the highly tortuous path
suggested this technique could provide a reasonable fit to the micropores and the cleat system.
390
The models use different controlling mechanisms Lamarre (2007) developed a new procedure to
for gas diffusion in the micropores and macropores. determine the critical desorption pressure and gas content
The surface diffusion and pore diffusion sorption rates downhole. This procedure uses a Raman spectroscopy-
are thus influenced by the respective model. Wei et based sensor capable of measuring trace amounts of
al. (2005) provide a comparison of the six referenced solution gas. The solution gas quantity is then related to
models in Table 1 of their paper. the partial pressure of methane via a solubility relationship
There is still work to be done with regard to the such as Henrys law, the relationship between gas
bi-disperse, surface diffusion in micropores and pore partial pressure and the concentration of gas in
diffusion in macropores, pore diffusion model. Wei et water at constant conditions.
al. (2005) list the areas as:
1. an improved numerical representation of the 11-4.4 Stimulation Techniques
multi-scale structure;
2. concentration-dependant diffusivity; Most CBM wells must be stimulated in order to be
3. the counter-diffusion effect observed during commercially viable. The types of completions can be
primary and secondary methane production; and broken down into two general categories: open-hole
4. the effect of water on gas diffusion. cavity and hydraulic fracturing.
A CBM well that is over-pressured with high coal
Desorption of gas from the coal matrix is described permeability appears to be a candidate for an open-
in a theory presented by Zhu et al. (2002a). They hole cavity completion. Here the wellbore is drilled
extended the convective and adsorptive exchange process through the coal with casing set at the top of the
from a chromatography point of view to include both seam. The coal itself may be under-reamed. As the
adsorption and desorption. The theory also incorporates well is allowed to produce, the coal tends to slough
gas properties as described by Peng and Robinson into the wellbore for some period of time. Eventually
(1976) and multiphase fluid flow in porous media the sloughing will cease.
as presented by Dindoruk (1992). Palmer (1992) points out that this process is similar
Gas desorption pressure and gas content have to sand flowback in unconsolidated formations, except
typically been determined using the industry- here the effect has a positive result. Unlike that in an
accepted practice of retrieving a core sample under unconsolidated sand completion, production from a
pressure, then desorbing the gas from the coal sample coal seam is improved. This is due to the removal of
and relating the resultant gas content to critical damage to the wellbore from drilling fluids, coal fines
desorption pressure and thus partial pressure via an and/or relative permeability effects, i.e., a positive
adsorption isotherm (see Fig. 11-9). This procedure is skin. In areas where sloughing will occur, it typically
very time-consuming and expensive. occurs naturally and has been shown to be the most
productive of all completion strategies.
Radii of 2 to 5 ft are typically estimated from the
Methane Storage Capacity, scf/ton
500
volume of coal fines recovered at the surface (Jones
Total Reservoir
400
Pressure
et al., 1984; and Logan et al., 1986). A collaborative
300 study (Mavor, 1991) used a sonar logging tool to
200
Desorbed map a cavitated wellbore. The team found an average
Gas
Critical
Desorption Content radius of 4.6 ft (Fig. 11-10).
100
Pressure In areas where sloughing will occur, it is typical that it is
0
0 500 1000 1500 2000
naturally occurring and shows to be the most productive of
Pressure, psia all completion strategies. Two coal-sloughing techniques
have been used in cased-hole completions. The cold
Figure 11-9 Standard technique for deriving critical
CO2 huff-n-puff technique (Puri et al., 1991a) uses
desorption pressure from gas content and adsorption
isotherm data (from Lamarre, 2007) thermal contraction and/or the swelling and shrinking
391
Modern Fracturing
of the coal caused by the adsorption and desorption of Water fracs, both with and without friction reducer,
CO2 to induce failure of the coal. The other technique are less damaging to the cleat permeability, although
is foamed water clean-outs (Kutas, 1990). Here, low- friction reducer will have some negative effect on the
rate injection of nitrogen-foamed water pressurizes the proppant pack permeability. This fluid has very low
coal. The well is then rapidly opened to the atmosphere, viscosity and poor proppant carrying capabilities and may
causing the coal to fail. not be able to prop all the coal seam that is hydraulically
fractured. However, it is still shown to be an effective
Gamma Ray I #2 Cavity Radius treating fluid in certain areas (Palmer et al., 1991).
0.0 GAPI 200 -8.0 Feet 8.0
3149 Penny and Conway (1996) extensively studied the
damage mechanisms of polymers and surfactants on
3150 coals. One of their conclusions was that all foaming
agents reduce the relative permeability of the cleat
3160 to gas. Their study did reveal a foaming agent that
minimized the relative permeability reduction.
Depth, ft
3170 Myers, Mack et al. (1999) presented laboratory

and field data that showed this foaming agent to be
3160 beneficial in a CBM reservoir.
Viscoelastic technology increases the carrier fluid
3190 viscosity and the proppant-carrying capability of a
water-based fluid. Unlike conventional polymer systems,
3200 a viscoelastic fluid is solids-free. A laboratory study by
Density
1.0 gm/cc 3.0 Olsen et al. (2003) reports the regained permeability of
the viscoelastic fluid in a coal-packed conductivity cell
Figure 11-10 Cavity dimensions inferred from sonar log in
the San Juan Basin, with coal density < 1.7 g/cc shaded in
was 230% greater than that of a borate crosslinked fluid
the center track (from Palmer, 1992) and a linear gel fluid. Chen et al. (2006) performed
additional laboratory tests and found that even with the
Most CBM wells are completed through casing addition of surface tension-reducing surfactants, polymer-
with perforations, although some operators, particularly based fluid does not achieve regained permeability to the
in shallow CBM areas, complete their wells open-hole. level of the viscoelastic fluid. Field application through a
Several techniques are used for zonal isolation, such re-fracturing program using viscoelastic-based nitrogen
as bridge plugs and sand plugs. Which techniques are foam fluid produced an estimated ultimate total reserves
implemented depend on operator preference; no one (EUTR) improvement of 25-30% as compared to the
technique has technical superiority over another. initial completions that used a conventional polymer-
As in a conventional sandstone completion, coals based fluid system (Fredd et al., 2004).
are affected by the composition of the fracturing fluid. Proppant pack conductivity reduction caused by
Puri et al. (1991b) demonstrated that adsorption of formation fines is a concern in all hydraulically fractured
water into the coal matrix swells the matrix, thus reservoirs. A major source of CBM fines comes from
reducing the permeability of the cleats. They showed the much harder proppant being pressed against the
that this reduction can be as much as 5- to 10-fold. relatively soft coal cleat face. These fines would tend
They also showed the cleat permeability is further to accumulate in the proppant pack, reducing its
reduced by whole gel that has leaked off into the cleats. conductivity and hence the well productivity. The
The mechanisms of gel damage were explained by introduction of a surface modification agent (SMA)
Conway (1992). This led many operators to use water has been shown to improve the productivity of a
fracturing (Palmer et al., 1989), no-polymer foams or CBM well (Lehman et al., 2003). The SMA coats the
linear polymers such as hydroxyethyl cellulose (HEC) proppant and filters the coal fines, preventing their
as the treating fluid (Spafford, 1991). intrusion into the proppant pack.
392
Most CBM wells are propped with natural frac often greater the horizontal permeability, thus the
sand. Due to the softness of the coal, proppant indirect fracture only needs to penetrate the coal to
embedment into cleat faces can be significant. It effectively drain it. The high leak-off caused by the
is not unlikely that one or more proppant grains cleat permeability will tend to draw the proppant
would be embedded into the coal surface on each into the cleat, creating a highly conductive path.
side of the cleat. To minimize this negative effect Also, the low fracturing gradient of the clastic rocks
on the proppant pack conductivity, two approaches will ensure an elastically coupled intersection into
are often taken. One is to place as many proppant the coal. Successful applications of IVFC have been
grain layers as possible into the cleat. This provides performed in the US in Utah, the central Rocky
sacrificial grains for embedment, leaving additional Mountains and the San Juan Basin.
grain to keep the cleat open. This is typically achieved Guidelines for IVFC applications are provided by
by adding very high-concentration proppant Olsen et al (2007):
slurries, 10 ppg or even higher. 1. In general, thick, high-fracture-pressure coal
A second technique is to use as large a proppant seams have inefficient complex fractures and are
grain diameter as is practical. The larger grain diameter probably not good candidates.
will distribute the cleat closing stress over a larger 2. An understanding of the mechanical
surface area, thus reducing the depth of embedment properties of the formations and the stress
of a proppant grain. Many CBM well are treated profile is required. A clastic layer with a low
using 12/20 mesh sand. The use of large-diameter, Poissons ratio and a high Youngs modulus
deformable proppant would further enhance the is the most efficient for the development of
stress distribution effect by creating an even larger induced hydraulic fractures.
surface area for stress distribution. 3. Areas that have high tectonic compression
or have highly laminated coals are
not good candidates.
4. Avoid excessive fracturing fluid leak-off. Besides
robbing fracture energy and increasing the
possibility of premature screenout, excessive
leak-off will substantially increase the minimum
horizontal stress to a point where it is nearly
equal to the vertical stress or the maximum
horizontal stress. Inefficient complex fracturing
will most likely be the result.
Figure 11-11 Optimized IVCF fracture geometry:

high fluid leakoff in the coals will assure proppant
11-4.5 Alternate Completions
connectivity to coals (from Olsen et al., 2003) and Enhanced Production Techniques
In many basins, a low-stress sandstone or siltstone Initial completions in CBM were in vertical
is adjacent to a coal seam or is sandwiched between wellbores. However, it was soon realized that CBM
coal seams. This clastic layer provides an efficient reservoirs exhibit the three characteristics that
means of propagating the hydraulic fracture away make horizontal completions effective: They are
from the wellbore while intersecting the adjacent coal relatively thin, naturally fractured and known to have
(Fig. 11-11). The technique, called Indirect Vertical anisotropic permeability.
Fracture Connectivity (IVFC), was proposed for The US Bureau of Mines (Duel and Cervik,
CBM applications by Olsen, Brenize and Frenzel 1977; Prosser et al., 1981) originally considered
(2003). They cite two reasons for the viability of this directionally drilled coal wells for degasification
technique. The vertical permeability of coal is most before the mining of the coal seam.
393
Modern Fracturing
A simulation study by Delmbacher et al. (1992) Coal strength expressed as its unconfined compressive
compared vertical CBM wells to wells drilled strength (UCS) is related to the coal rank, which is
horizontally in an azimuth perpendicular to the based on the percentage of volatile matter (VM)
maximum permeability (i.e., the face cleats). It in the particular coal (Bell and Jones, 1989; and
showed that horizontals can provide significantly Jones et al., 1988). Table 11-2 summarizes the likely
higher gas production rates. In addition, peak gas minimum UCS for each coal rank.
flow rate is generally attained sooner as the coal is Drilling, stimulation and production
dewatered more efficiently. guidelines with regard to coal stability have been
Before it is decided to exploit a coal reservoir suggested (Palmer et al., 2005; Moschovidis et
with horizontal wellbores, there are other conditions al., 2005). They include:
to be considered (Osisanya and Schaffitzel, 1996). 1. Wellbore stability during drilling depends
Several core samples should be tested to determine on the in situ stresses, coal strength
maximum and minimum stress directions and wellbore deviation.
and the strength and friability of the coal. 2. Well stability models should be adapted for coals
These will determine the well azimuth (normal to and used in drilling design.
the maximum stress) and if the coal is stable 3. Wellbore stability generally decreases with depth,
enough to support a long lateral section. The but individual cases should be studied.
core samples should also be tested to determine 4. The acceptable mud weight window will decrease
the appropriate type and weight of drilling with reservoir depletion.
fluid to minimize the pressure leak-off into the 5. Frac gradient greater than 1.0 psi/
coal; however, a breaker should be incorporated ft is not necessarily indicative of a
to remove the filter cake after drilling has been horizontal fracture as the fracture itself can
completed (Baltoiu et al., 2006). increase the vertical stress.
Osisanya and Schaffitzel (1996) also suggested 6. The extent of the failure zone from a hydraulic
implementing a medium-radius build, with the curve fracture is dependant on coal strength and fracture
section built before the coal is entered. This is due to pressure and can extend 50 ft or more.
the general weakness of the coal. 7. Shear failure extending from the fracture
face may also create:
Table 11-2 Likely Minimum Unconfined Compressive Microseismic emissions
Strength of Coals Based on Rank (from Palmer et al.,
2005)
Coal dilatancy and permeability increase
Coal Rank Likely min UCS (psi)
(production enhancement)
(decreasing VM-daf% Based on literature Fines creation or movement (production
order) data
detriment)
Anthracite Less than 14 1780
LV 14 to 22 490 8. As depth increases, the stability during production

MV 22 to 31 497 of an open-hole lateral decreases.
HVA 31 to 39 1050
9. Sanding methodology can be applied to stability
predictions of open-hole lateral coal wells.
HVB 39 to 42 4800
HVC 42 to 55 5930 In addition to single horizontal wells, multiple

Lignite Greater than 55 9400 laterals perpendicular to the maximum stress
VM-daf% = percent volatile material - dry, ash free can be beneficial in exploiting a CBM reservoir.
A simulation study by Maricic et al. (2005)
Wellbore stability knowledge is important because investigated options with as many as four laterals
it will impact the economics of drilling, hydraulic from a single vertical wellbore and a pinnate
fracturing and producing a horizontal CBM well. fishbone pattern (Fig. 11-12).
394
Enhanced recovery from a CBM reservoir (ECBM)

is an important aspect of this resource exploitation.
One ECBM technique is the injection of nitrogen,
which reduces the partial pressure of methane with
Single Lateral Dual Lateral
little sequestration of the nitrogen. Depending on the
coal, about one volume of methane will be desorbed for
every one-half to one volume of nitrogen adsorbed. A
pilot project in the San Juan Basin reported a methane
production increase of several fold; however, nitrogen
breakthrough was very rapid, necessitating the extra
Trilateral Quadlateral
production process of nitrogen removal (FETC, BP-
Amoco and IEA GHG Programme., 1999).
As discussed in Section 11-4.3, carbon dioxide
readily adsorbs to the coal matrix, forcing methane
to desorb at a faster rate than pressure depletion,
making it available for production as free gas.
Pinnate
Coal will desorb about one volume of methane
Figure 11-12 Potential lateral configurations (from for every two to threes volumes of CO2 adsorbed,
Maricic et al., 2005) depending on the coal itself.
The first successful CO2-enhanced recovery project
Parameters included in this study included was undertaken in the northern New Mexico portion
number of laterals, total lateral length drilled of the San Juan Basin, where gas production was
and the spacing between laterals. From the study, increased as much as 150% over convention pressure-
the authors concluded that: depletion (Stevens et al., 1998). This CBM reservoir did
Production, recovery and economics are all improved exhibit the key criteria for a potentially successful CO2
with horizontal and multilateral completions as injection project: laterally continuous and permeable
compared to a vertical completion. coal seams, concentrated seam geometry, and minimal
Longer horizontal lengths increase the recovery faulting and compartmentalization.
of a CBM reservoir; however, the associated There are two main sources of CO2 for injection:
drilling costs negate any economic advantage extraction from underground reservoirs and capture
beyond a certain length. from industrial flue gas. Although the former is typically
For their set of parameters, a well with four laterals the least expensive, the latter has the added benefit of
from a single vertical wellbore (a quadlateral) with preventing the addition of this greenhouse gas into the
lateral spacing of 680 ft and total horizontal length Earths atmosphere as the CO2 would be sequestered
of 3100 ft yielded the best economics. in the coal matrix (Wong et al., 2000). However, the
economics of capturing CO2 from flue gas must be in
Many successful horizontal projects have been place to take advantage of this process (Espie, 2005;
executed. One such project targeted the thin (as Sengul, 2006; and Imbus et al., 2006). The potential for
narrow as 4 ft) Hartshorne Coal in the Arkoma Basin CO2-ECBM/sequestration in the United States is 150
of Oklahoma, where production increases of five- to Tcf incremental methane production and 90 Gt (giga-
six-fold as compared to vertical completions were tonnes) of CO2 sequestered (Reeves, 2002-2003).
experienced (Mutalik and Magness, 2006). Laboratory studies and simulations have been
Successful side-track CBM drilling has been performed to investigate the potential for injecting
achieved in the San Juan Basin (Noynaert et al., 2007) a binary gas, CO2 and N2 (Tang et al., 2005; Zhu
and a horizontal project in southern Illinois was planned et al., 2002b; and Jahediesfanjani and Civan,
(Rodvelt and Oestreich, 2005). 2006). It was found that:
395
Modern Fracturing
CO2 tends to move in a more piston-like manner be impaired. Among their recommendations was to
with high efficiency. test charges with coal targets and the development of
N2 is more dispersed and will break through better penetration models for various coal types, such
more rapidly. as bituminous and anthracite.
A binary system high in N2 percentage yields a
higher initial methane production rate. 11-4.6 Fracture Modeling of CBM Wells
A CO2-rich system is more efficient and
longer lasting. Fracturing fluid options and additives are discussed
Water content in the coal reduces CO2 and N2 in Section 11-4.4. This section will look at fracturing
adsorption rates. treatment design and modeling.
A higher ratio of N2 to CO2 is required in wet Many CBM reservoirs comprised several
coals to enhance methane production. narrow coal seams found at various depths in a
wellbore. Individual fracturing treatments on each
A simulation using an extensive characterization seam can be very time-consuming, inefficient and
of the low-rank Wilcox Coal of East Texas looked more costly than one large job covering all or
at the overall economics of CO2 sequestration, groups of seams. A limited entry technique can
methane gas production and the effects of binary be employed to fracture stimulate multiple seams
gas mixtures (Reeves, 2002-2003). The study during a single treatment (see Section 9-7.3).
concluded that the Wilcox coals could sequester Limited entry is the practice of limiting the number
a large quantity of industrial CO2; however, the of perforations in each zone to be treated such
economics were unfavorable. The best economic that excess pressure is created in the casing, forcing
scenario was at an N2-to-CO2 injection ratio of nearly the treating fluid to simultaneously enter multiple
7:1, however, this ratio significantly decreased the zones with variable in-situ stresses.
CO2 sequestration volume. Successful limited entry programs have been
Additional information on CO2 sequestration achieved in the US San Juan Basin (Bazan et al., 2002;
can be found at the Coal-Seq Consortium and Bazan and Larkin, 2002). For these treatments the
website, www.coal-seq.com. authors determined the total number of perforations
Microbially enhanced hydrocarbon recovery required from:
has experienced a degree of success. A current study
q 2
suggests that the inclusion of nitrogen-fixing, carbon n perf =
s
4
, (11-9)
dioxide-fixing and methane-eliminating bacteria into 0.323Cd ( Dc D p ) p pf
a binary CO2/N2 gas injection could increase methane and
production by 8 to 10% (Jadhav, 2007).
55
One area of CBM well completions that has Dc =
, (11-10)
y
commonly been overlooked is the perforations
through the casing and cement into the coal. The where Nperf is the number of perforations, q is the total
perforation tunnels provide the connection between flow rate (bpm), s is the slurry density (lbm/gal), Cd is
the cleat system/ hydraulic fracture and the wellbore. the perforation discharge coefficient (see below), Dp is
Conventionally, correlations used for relatively the perforation diameter (inches), ppf is the perforation
weak, low-density sandstone were applied to CBM friction (psi) (see Section 4-2.1.2) and y is the yield
perforations. Snider et al. (2006) demonstrated strength of the pipe metal in 103 psi (such that, for
that the penetration and diameter of perforations example, for P-110 pipe, y = 110 kpsi).
from shaped charges were considerably less than Crump and Conway (1986) suggested a value for
predicted by the models. They also stated that a visual Cd of 0.6 for non-abrasive fluids ranging up to 0.9 for
inspection of damaged perforation tunnels in the coal proppant-laden fluids. For the US San Juan project, a
suggest the communication to the wellbore would value of 0.8 was found to be appropriate.
396
In order to achieve limited entry it is generally which relates shut-in time to total pumping time (see
felt that a minimum pressure drop across the Section 4-2.1.9 for a detailed explanation of G-Function
perforations, ppf, should be 400 psi. With this in analysis). A plot of the superposition function Gdp/dG
mind, typical limited entry designs employ pressure versus G will yield a characteristic hump indicative of
drops in the range of 600 to 1000 psi. pressure-dependent leak-off (Fig. 11-13). The pressure at
Although designing based on perforation pressure which the superposition curve returns to linearity is the
drop alone has been successful in the US San Juan fissure- or cleat-opening pressure.
Basin, Barba et al. (1999) warn that areas such as the Application of this technique is discussed by
US Black Warrior Basin, where mechanical properties Bazan et al (2002) and Ramurthy and Lyons (2007)
and in situ stresses vary significantly between seams, will using commercially available lumped 3-D or fully
require additional consideration. For limited entry to be functional, grid-oriented models.
successful under these conditions it will be necessary to Knowledge of PDL along with the geomechanical
rigorously calculate the number of perforations to shoot properties of the coal (Rahman and Khaksar, 2007)
in each seam such that the treating pressure in the casing can provide for a successful application of an effective
is the same at all perforation locations, allowing for the CBM fracturing treatment.
effects of hydrostatic pressure in the wellbore. CBM fracturing treatment design must also be
The hydraulic fractures created in a coal are very coupled with the economic impact of the treatment.
complex because they tend to follow the face and butt Having the fracture design simulator linked to a
cleats. Even with this complexity, 3-D fracture models production forecasting model is most beneficial.
can be used for treatment design, keeping in mind that Although a larger treatment may yield greater
the modeling is representative of the treating pressure production potential, the cost of the treatment
response but may not fully represent the complex and associated cleanup, disposal, etc., charges may
fracture geometry. However, care must be taken to actually reduce the overall economic benefit. Analysis
analyze minifrac and post-treatment pressures to better of the net present value, NPV, in association with
define in situ coal properties, in particular pressure- fracture treatment design has proven to provide for
dependent leak-off (PDL). PDL will occur when the the making of sound business decisions (Bazan et al.,
treating pressure within the main fracture increases to a 2002; and Valencia et al., 2005).
level where the butt cleats open; robbing fracturing fluid
and possibly proppant from the main fracture. 11-4.7 Fracturing Treatment
Evaluation of CBM Wells
1000 4700
900
4600 The Evaluation and refinement of CBM hydraulic
Bottom Hole Pressure, psi
800
Pressure Derivatives
700 4500 fracturing treatments should be a continuous process

600 pfo= Fissure opening pressure
4400 in field exploitation. The analysis of post-frac pressure
500
4300
decline data has been discussed in the previous section.
400
300
Such an analysis was performed in the Scotia Field in
dp/dG 4200
200 eastern Australia (Johnson et al., 2002). Through this
4100
100 Gdp/dG analysis Johnson et al. were able to determine bottom
0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
4000 hole treating pressure trends and delineate better areas
of productivity. The end result was a more predictive
Figure 11-13 Derivative plot for pressure-dependent
leakoff (from Barree, 1998) model to be used to develop the field.
In addition to pressure history matching, wellbore
The pressure at which the butt cleats will open can tiltmeters have been employed to map fracture height
be determined by analyzing the pressure decline after growth (Stutz et al., 2002); (Mayerhofer et al., 2003);
a pre-frac injection test using the Nolte dimensionless (Wolhart et al., 2004). In these studies the fracture-
time function, G (Castillo, 1987; and Barree, 1998), induced tilt signals are monitored and mapped providing
397
Modern Fracturing
fracture width and growth history. This information (Zuber and Semmelbeck, 1990; and Sawyer et al, 1987). As
is then used to further calibrate the fracture treatment can be expected, the quality of the data used in a simulator
design model and optimize future fracturing treatments. is critical. Zuber and Olszewski (1992) categorized
In a 16 well pilot program in Wyoming, the authors the effects of the various measured CBM properties
report a first year savings of $1,300,000 through this on production forecasting. They found that adsorbed
optimization process (Mayerhofer et al., 2003). gas content, desorption isotherms (more precisely
There are times when hydraulic fracturing of a coal the Langmuir volume) and the initial water saturation
seam does not produce favourable results. An eastern had significant impact on well production forecasts.
Australia CBM reservoir contains seams of high and low Parameters with a moderate degree of significance
permeability. Various hydraulic fracturing treatments are permeability, net pay thickness and porosity.
and open-hole completions were tested and production And parameters that are relatively insignificant are
trends closely monitored. Through this intense study ash content, desorption time and initial reservoir
it was determined that the high permeability seams pressure. They also reported that errors associated
are best completed open-hole with the potential with determining adsorbed gas content, permeability,
implementation of short radius horizontals. It was porosity, desorption isotherm and initial water saturation
also observed that the potential exists to optimize the were most significant in forecasting production for
hydraulic fracturing of cased-hole low permeability years 1 to 4. Beyond 4 years, errors in these parameters
seams (Johnson et al., 2006). become less significant.
An estimate of the gas-in-place can be made
11-4.8 Estimation of Reserves and with the knowledge of four physical reservoir
Production Data Analysis parameters (Nelson, 1999): reservoir or well
drainage area, gross reservoir rock thickness,
In some respects, CBM reservoirs are similar to average reservoir rock density and average in
conventional oil and gas reservoirs in that it is important situ gas content. With these four parameters,
for engineers to generate production forecasts, evaluate gas-in-place is determined from:
in-place reserves, provide information to design field
facilities and optimize reservoir development. This task Gi =1359.7 AhGc ,
(11-11)
is more difficult for CBM wells than for conventional
wells due to the more complex nature of the gas storage where Gi is the initial gas in place (scf), A is the reservoir
and flow mechanisms in a coal. area (acres), h is the average reservoir net height (ft) and
Malone et al. (1987) were among the first to Gc is the average in situ gas content (scf/ton).
perform gas-in-place analysis. One technique used for Nelson (1999) points out that errors in
CBM production forecasting and reserves estimates determining these reservoir properties will compromise
was decline curve analysis (Richardson et al., 1991; the accuracy of the calculation. Six major sources of
and Hanby, 1991). This, however, requires the error were identified:
establishment of a sustained production decline trend. Assumption that there is lateral continuity
Dewatering of the coal can result in two years or more of the reservoir thickness when determining
before the establishment of the appropriate decline rate. reservoir volume.
It has also been shown that the decline rate can vary Determining reservoir thickness with a maximum
significantly between wells in a field (Hanby, 1991). log density value that is too low.
These factors tend to make decline curve analysis Using rule-of-thumb coal density values to
generally unsuitable for CBM forecasting. determine in situ reservoir rock density.
Reservoir simulators were adapted for CBM Using ambient temperature conditions to
analysis. These tools correctly account for gas desorption, determine the gas desorption data used in
gas diffusion, relative permeability effects, operating calculating in situ gas content.
procedures and well-to-well interference simultaneously Using drill cuttings to obtain gas desorption data.
398
Assuming residual gas volume is negligible in volume factor was set at 1.0 resbbl/stb, water influx
determining in situ gas content. was neglected and water saturation remains constant
over time. Although constant water saturation is not
Table 11-3 provides an analysis protocol to avoid totally accurate, changes occur over a matter of years,
or minimize common errors in determining coalbed and the saturation change from initial to abandonment
reservoir properties used to calculate gas-in-place. conditions is often minimal. It is therefore necessary
Conventional gas well and reservoir to select a water saturation to be used throughout
analysis utilizes the mass balance equation all calculations. Seidle suggests that the initial water
with a p/Z vs. cumulative production plot (Lee saturation be used for CBM wells with free gas.
and Wattenbarger, 1996): Z* is now defined as:

p pi p Z
= i Gp , (11-12) Z = ,
Z Zi Z i Gi ZTpSCVmpc (1 a )bb)
1 c f ( pi p ) (1 S w ) +
Z SCTSC (1 + bp ) i
where Z is the gas correction factor, pi is the initial
reservoir pressure, Zi is the gas correction factor (11-13)
at initial reservoir conditions, Gi is the original
gas in place (OGIP) and Gp is the cumulative gas where cf is the formation compressibility (psi-1), Sw is
production from the reservoir. A plot of cumulative the water saturation (fraction), T is temperature (R),
gas production vs. p/Z is a straight line with intercept Vmpc is the pure coal Langmuir volume (constant, scf/
equal to the OGIP. The initial reservoir pressure can ton), a is the mass fraction of ash (fraction), b is bulk
be determined with a simple calculation. coalbed density gm/cc, pSC is the pressure at standard
conditions (usually 14.7 psi), TSC is the temperature at
Table 11-3 Gas-In-Place Analysis Protocol (from standard conditions (usually 519.67 R or 60 F), ZSC
Nelson, 1999) is the gas deviation factor under standard conditions
Perform gas desorption measurements at reservoir
1.
temperature on conventional core samples.
(dimensionless). b is the Langmuir pressure (constant,
Estimate the total gas content of each sample using the
psi 1) and i is the initial porosity (fraction).
2.
Direct Method lost gas content procedure. The material balance equation now becomes:
Relate the total gas content of multiple samples to sample
3.
composition. p pi pi
= G p . (11-14)
4. Relate the sample composition to density. Z
Zi
Z i Gi

Determine the in-situ moisture content from laboratory

5. For a volumetric coal deposit, a plot of cumulative
moisture content measurements.
Estimate the gross thickness and average in-situ density production vs. p/Z* will be a straight line with
6.
from open-hole density log data. an intercept of OGIP. The initial reservoir
Compute the in-situ gas content at the average reservoir pressure is then calculated.
7.
density and moisture content.
A generalized form of the material balance equation
8. Compute the gas-in-place volume.
is offered by Ahmed et al. (2006) as:
King (1990) adapted this form of the material
balance equation to CBM usage with the introduction G p = Gi + GF Ga GR ,
(11-15)
of a Z* variable, which considers adsorbed and free gas
as well as connate and influx water. However, when where Gp is the cumulative gas produced, Gi is gas
this method was applied to actual coal wells, less than originally adsorbed (assumed equal to OGIP), GF is
desirable results were observed. original free gas, Ga is gas currently adsorbed and GR is
This led to additional work by Seidle (1999), remaining free gas. The four variables to the right of Eq.
who redefined the Z* term. He did this by making 11-15 are determined from:
several simplifying assumptions: water and formation
compressibilities were neglected, water formation
Gi =1359.7 AhGc , (11-11)
399
Modern Fracturing
The modified dimensionless time is defined by

GF = 7758 Ah (1 S w ) Egi , (11-16) Delmbacher et al. (1992):

0.00634k
and, tD =
t , (11-21)
(g ct )i rw
2 ma
5.615Z T p pi

Egi = SC SC i
= 198.6 scf / bbl, (11-17) where g is the gas viscosity (cp) and ct is the total
pSC TZ i TZ i
reservoir compressibility (psi-1) both at initial reservoir
where pSC is the pressure at standard conditions (psi). conditions rw is the effective wellbore radius (ft) and
tma is the material balance pseudo-time:
Ga =1359.7 Ah BV p ,
(11-18)
(g ct )i t q dt
tma =

q 0
,
g ct
(11-22)
and
where and are the average gas viscosity (cp) and total
GR = 7758 Ah reservoir compressibility (psi-1) respectively, averaged
BwWP over the producing life of the reservoir until time = t.
+ (1 S wi ) ( pi pr )(c f + cw S wi )
7758 Ah The authors (Pinzon and Patterson, 2004; and
Eg ,
1 ( pi pr ) c f Delmbacher et al., 1992) also caution that flowing

material balance and type curves should be used in
(11-19) conjunction with other reservoir simulation methods.
where Bw is water formation volume factor (bbl/ 11-5 Shale Gas

stb), WP is cumulative water production (stb), Swi
is initial water saturation (fraction), cw, is the water 11-5.1 Shale Description
compressibility (psi-1), Vp is the total gas absorbed at
current reservoir pressure pr (scf/ton), and: Each of the gas shale systems are unique in their own right
and have variability within a producing area. Therefore
p

Eg =198.6 r scf / bbl. (11-20) careful planning and execution of a proper development
TZ
plan is required to obtain optimum and commercial
Substituting Eqs. 11-11, -16, -17, -18, -19 and production. The success of the Barnett Shale of the Ft.
-20 into Eq. 11-15 and rearranging, Ahmed et Worth Basin has received the most attention, and its
al. (2006) presented an equation in the form of explosive growth has made it one of the most prolific gas
y = mx + b. The slope and intercept are used to shale plays in the United States (Mathews et al., 2007).
calculate original gas-in-place. They further extend Figure 11-14 shows an overview of the gas shale
the calculations to determine the average reservoir basins and plays across the continental United States.
pressure and future production. A 13-step analytical A common mistake in developing and attempting
technique is presented for the prediction of CBM completions in many of these shale reservoirs is to
reservoir production forecasting. treat the properties of one like that of another. One
Type curves have also been presented to analyze of the many lessons learned is that not only will shales
CBM production and production forecasting. (Aminiam vary from basin to basin but also they will vary within
et al., 2004 and 2005; Mohaghegh and Ertekin, 1991; the same field. The productivity of a shale reservoir
and Pinzon and Patterson, 2004). Clarkson et al. (2003) will have significant differences in hydrocarbon
pointed out that these and other type curves can be used content, fracture barriers, mineralogy and structural
for single-phase coal gas production analysis if CBM environment. Variations in these four parameters
storage mechanisms, i.e., adsorption, are accounted for across and within each producing basin greatly impacts
by substituting material balance pseudo-time for time in not only the well plan but also specific engineered
the dimensionless time equation. designs required to optimize productivity.
400
Gammon Excello/Mulky
Niobrara
Bakken
New Albay
Antrim
Green
River
Monterey
Devonian/Ohio
McClure
Cane
Creek Floyd and
Lewis and Conasauga
Palo Duro
Mancos
Barnett and Caney and
Woodford Woodford
Barnett
Woodford
Fayetteville
Figure 11-14 Gas shale basins of the USA (from Brown, 2006b)
As was discussed earlier in this chapter, these shale The Antrim Shale is characterized by a low victrinite
reservoirs consist of fine-grained sedimentary rock reflectance (VRo) of between 0.5 and 0.75%; however,
inter-bedded with siliceous material and carbonaceous the Total Organic Carbon (TOC) content can be as high
material. Although some of these reservoirs could as 20% (Jarvie, 2006) Wells producing from the Antrim
produce relatively low volumes of gas naturally, most Shale will have relatively low production rates, e.g., 135
require some form of well stimulation. Mscf/d; however, they will produce for a long time, such
A more detailed shale description will be discussed as 20 years. In many cases large quantities of water are
later in this chapter where individual US basins are produced with the gas. Geochemical analysis has indicated
discussed with factors that effect stimulation designs that the water is usually fairly fresh. Although natural
and production in each one. fractures exist because of the low bottomhole pressure of
this shallow 500 to 2,500 ft formation, the contribution
11-5.2 Thermogenic and Biogenic Systems from these fractures may not be relevant. Much of this
early-generation system appears to be somewhat sensitive
Two types of productive shale gas systems must to water, and therefore water is minimized in the drilling
be understood in developing unconventional and completion process. Gas production is closely tied to
plays of this type. de-watering the system to gain economic production.
Biogenic Systems Thermogenic Systems

Relatively few biogenic gas systems are currently The majority of producing shale gas reservoirs in the US
producing economical gas within the United States. are thermogenic systems. They are characterized by a much
The Antrim Shale in Michigan is one of those systems. higher victrinite reflectance (VRo) of between 1.0 and
401
Modern Fracturing
1.8%; however the Total Organic Carbon (TOC) content The following sections will look at these shale
can range from 2.5 to 4.5% (Jarvie, 2006). Thermogenic reservoirs individually because their variability and
gas occurs as a result of a primary thermal cracking of the completion methods cannot be considered as a whole.
organic matter into a gaseous phase. Secondary thermal
cracking of remaining liquids also occurs. For this reason 11-5.3 Ft. Worth Basin Barnett Shale
production systems are variable. For example, sometimes
within a field an area with lower maturity may produce The Barnett Shale of North Texas is a Mississippian-
more liquid hydrocarbons. Thermal maturity of the age marine shelf deposit that exhibits variations in
system is one factor in determining productivity, but mineralogical and geochemical properties. It is described
gross reservoir interval is also important. The overall as black, organically rich, fine-grained shale. (Lancaster,
shale interval can range from a few hundred feet to 1992) It ranges from 3500 ft deep and 150 ft thick to
over 1000 ft in overall thickness. Wells completed in the South and West in Erath County to over 8000 ft
these intervals tend to be characterized by significant deep and 1000 ft thick to the North in Denton County.
and economical production rates. Production examples The Barnett has provided excellent data resources to
will be discussed later in this chapter. Unless fracture define the favorable geochemical attributes for gas-
barriers are absent, the gas produced in a thermogenic productive shale as well as the transition area into oil-
environment will be relatively dry. productive shale reservoirs that are not as thermally
Systems of this type include many of the prolific mature. (Mathews et al., 2007).
shale reservoirs currently being produced in the US. Early field development focused on an area that
These include the Barnett of the Ft. Worth Basin, Barnett was very easily developed, in that the shale was thick
and Woodford of the Delaware Basin, Fayetteville Shale and well-bounded by hard, dense limestone, both above
in Arkansas, Woodford and Caney Shale of the Arkoma and below, and the structural setting was relatively calm
Basin, and others such as the Floyd and Conasauga with very minor faults, if any, and no karsts. After its
Shale in the Black Warrior Basin. A common element productive capability was firmly established and the
of these systems is their location along the Ouachita limits of the core area were developed with vertical
Fold Belt that runs from far West Texas to Western wellbores, exploration reached into the more challenging
Alabama. The location of these shales is important not environments where fracture boundaries were weak,
only in understanding the production system but also and where faults and karsts communicated with a wet
in developing a plan for completion. Proper completion Ellenberger (Dolomite) below. With proven economic
planning can be critical as these reserves continue to be success in these challenging environments, horizontal
developed in urban areas (Fig. 11-15). completions have now become the standard throughout
the basin (see also Section 12-4.3).
The Barnett Shale is best described as naturally
fractured shale located in the Fort Worth Basin of North
Texas (Lancaster, 1992). The drilling and completion
practices have seen significant changes since one of the
first productive vertical Barnett Shale well, C.W. Slay
#1, was drilled and completed in 1981. Because this
is described as shale, it was considered to be water-
sensitive, and early fracture treatments comprised
a large volume of nitrogen as the only stimulation
fluid. Limited production response led to subsequent
treatments using CO2-foamed fluids; however, because
the matrix permeability is so low, it was felt that large-
volume treatments would be required to extend long
Figure 11-15 Fracturing treatment performed in the urban
environment within the Fort Worth, Texas, city limits fracture lengths into the reservoir.
402
Table 11-4 Barnett Shale Mineralogy

Barnett Shale
Formation Components Minerals 7165- 7265- 7365- 7465- 7535- 7565- 7595- 7635-
7175 7275 7375 7475 7545 7575 7685 7645
Quartz (Si02) 32 29 42 38 41 34 33 29
Framework Grains Plagioclase Feldspar 2 2 2 2 2 2 3 2
Potassium Feldspar 2 1 trace trace trace trace trace trace
Calcite (CaCO3) 14 9 11 9 7 4 10 15
Carbonates
Fe-Dolomite
5 24 2 4 4 4 3 4
({Ca,Fe}{CO3}2)
Sulfide Pyrite (FeS2) 2 3 4 3 4 4 4 6
Apatite (Ca5{PO4}3{F, Cl,
Phosphate 1 trace 1 1 trace 1 1 1
OH})
Illite + Mica 2 2 2 3 2 3 2 3
Kaolinite 2 trace trace -- -- -- --
Clays
Mixed-Layer Illite80/
37 29 35 39 39 47 43 39
Smectite20
Totals 100% 100% 100% 100% 100% 100% 100% 100%
Stimulation methods used to obtain and optimize treatment on a vertical well have been approximately
production from the Barnett Shale have evolved 2,000 to 2,400 gallons per foot of gross interval. An
over time. Conventional crosslinked fluids were used important factor in the success of using a slickwater
throughout the later 1980s through 1996. These initial fracturing technique is that the volume of fluid pumped
fracturing treatments were developed using state-of-the- has significant impact on well performance.
art technology available at that time. These common Several factors become evident in understanding
stimulation techniques were also used to develop tight why large-volume slickwater fracturing has been
gas sands. These initial Massive Hydraulic Fracture successful in a shale reservoir containing various mixtures
(MHF) treatments consisted of titanium- and zirconium- of clay, silicate, carbonate minerals, etc. A typical core
based crosslinkers with HPG and CMHPG and used analysis of the Barnett is shown in Table 11-4, which
approximately 750,000 gallons of fracturing fluid, often is representative of this shale formation throughout
including nitrogen to assist in fluid clean up (see Chapter the trend. (BJ Services, 2006-07). Typically the shale
7 for a more detailed description of fracturing fluid consists of approximate 30 to 39% clays, 29 to 38%
systems). The proppant placed ranged from 500,000 to quartz, and 25 to 30% other minerals such as calcite,
1.5 million lbs of 20/40 mesh white or northern type dolomite, feldspars and pyrites. The very nature of this
sand (see Chapter 8 for a more detailed description of type of mineralogy causes the Barnett Shale to be very
proppant systems). Continued frac fluid improvements brittle and therefore easily hydraulically fractured. Core
allowed the use of lower-polymer borate crosslinked analysis and improved logging techniques have also
fluids with tapered gel loadings from 25 lbs down to as shown the presence of natural fractures, although these
low as 17 lbs of polymer per 1000 gallons of fluid. fractures are essentially closed and they can also be filled
As a result of increasing success using slickwater with secondary deposits such as halite and calcite. The
fracturing in the Cotton Valley Sand in East Texas production of hydrocarbons from these intervals tends
through 1997, several operators began experimenting to be primarily dependent upon not only the fracture
with this technique in other tight gas sand reservoirs extension but also on interconnecting and re-opening the
such as the Barnett Shale (Walker et al., 1996). Many natural fractures that do exist. These fractures can occur
early treatment successes were met with difficulty in in fracture swarms and are considered to be very complex.
placing high concentrations of proppant, therefore Much of the success in the Barnett Shale is related to rock
limiting the proppant concentrations from 0.05 to 0.5 quality in terms of permeability, porosity, gross interval,
ppg average. The fluid volumes currently for a typical and the existence of this natural fracture network.
403
Modern Fracturing
Because of these factors and, in particular, the presence obtained. First, it will enable placing a higher proppant
of natural fractures, a hydraulic fracture created in the concentration and therefore larger volume of proppant
Barnett Shale cannot be considered a typical two-wing in this naturally fractured reservoir. Second, the smaller
planer fracture. Description of a Barnett Shale fracture particle size 40/70 mesh material could be suspended
treatment is considered to be more complex, and through better in slickwater and would have a settling velocity
the evidence of fracture-mapping technologies, a fracture less than the 20/40 mesh particles (Schein et al., 1994).
fairway has been shown with multiple orientations Most of the treatments using the 40/70 mesh sand also
(Fisher et al, 2002). Figure 11-16 shows an example of utilized 20/40 mesh sand as a tail-in proppant to ensure
the type of fracture network associated with stimulation improved conductivity near the wellbore.
of the Barnett Shale, where fracture mapping is used to After thousands of jobs, the smaller-mesh proppant
monitor microseismic events during the stimulation. In have proven to not only provide the opportunity to
this example, a vertical well is fractured with as many as place more proppant, but result in improved fluid
five offset wells being affected by the treatment. recovery and higher production rates. More recently
with the expansion of the Barnett Shale play to counties
1500
outside of what is considered the Newark East core
1000 area (Denton, Wise, and Tarrant counties) 100-mesh
sand has been used with good result.
500
Most recently the application of horizontal drilling
0 has changed the boundaries and limits of the vertical
wells. The first horizontal well was drilled in the Barnett
South-North, ft
-500
Shale in 1992 and was completed with results not much
-1000 better than the vertical wells. Ten years later, improved
-1500
drilling techniques along with better understanding of
optimum wellbore azimuth and completions has led to the
-2000 significant expansion of drilling successful wells outside
-2500
the core area. The horizontal wells have lateral lengths
ranging from 1800 to 4000 ft and are typically drilled in
-3000
-1000 -500 0 500 1000 1500 2000 2500 3000 a 120 to 140 degree or 300 to 320 degree azimuth. This
West-East, ft is generally transverse to the induced fractures; therefore,
Figure 11-16 Fracture network fairway from during fracture stimulation, multiple treatments allow
microseismic evaluation of a vertically fractured well (from
Fisher et al., 2002). for maximum exposure to the reservoir.
Large surface areas are being contacted through 11-5.3.1 Barnett Shale Slickwater
hydraulic and interconnected fractures that will ultimately Treatment Design Considerations
contribute to production (Fisher et al, 2002). Initially
almost all fracture treatments in the Barnett Shale once The primary considerations in the Barnett Shale are to
used 20/40 mesh white or northern sand. However, contact as much interval as possible to maximize the
because of the complex nature of the fracture system, area of the conductive flow path to the wellbore, and
there were limits to how much proppant could be placed. to maintain the fracture treatment within the interval.
The average proppant concentration was kept very low This is particularly important outside the core area
and rarely exceeded 0.3 ppg until the final stage, where such as Johnson, Hood, Erath, Parker, Hill, and Bosque
proppant concentration was increased or ramped up to counties. The use of horizontal drilling and slickwater
1 and 2 ppg. In 2000, trial treatments were performed fracturing has allowed expansion of activity with
utilizing small 40/70 mesh proppant in the early stages excellent results where the wet Ellenberger underlies
of the job, with the idea being that although the material the Barnett Shale directly without the presence of the
will have less conductivity, two benefits might be dense Viola Lime as a barrier.
404
Initially in these long cemented laterals, wells Because of the complex fracture network, there
were completed using multi-stage, (80 to 180 BPM) is a significant amount of communication between
treatments. It is now more common to perform five wellbores and offset wells thousands of feet away
to seven stages per well at lower rates (50 to 80 BPM) (Fisher et al., 2002). The theory of simultaneous
depending upon the area and potential for fracture fracture treatments is to minimize the communication
containment. Table 11-5 shows an example of a and subsequent killing of offset wellbores. Wells that
typical fracture in the Barnett Shale in an area where have undergone such treatments have significantly
the Viola Lime is present and potential for fracturing outperformed their offsets; however, one must also
into water is minimal. Table 11-6 is an example of a consider reservoir quality and completion methods
fracture in an area where the potential to fracture into as well. Figure 11-19 represents production offsets of
water is high. In this case the treatment consists of two pairs of wells fractured individually and two pairs
predominantly 100-mesh sand. of wells fractured simultaneously. In this example
A more recent trend in the Barnett Shale has been the fluid and proppant volume are very similar. The
to use a simultaneous fracture technique of offset wells azimuth and overall lateral length of all four of these
drilled parallel to one another (Figs. 11-17, -18a and - wells were almost identical. Significant production
18b). In the figures, two wells are seen on the same improvement using the simultaneous fracture
location; their horizontal wellbores are parallel. technique is evident based on these factors.
1,000,000
Simo Pair Wells

Mscf/ Month
100,000
Single Pair Wells
~120,000 Bbls slick water & 1.6 x 106 lbm sand per well
10,000
Figure 11-17 Simultaneous fracturing treatment 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
performed in Barnett Shale Months
Figure 11-19 Simultaneous vs. single fracture stimulation

in pairs of wells
1,000,000
Simo Pair Wells

Mscf/ Month
100,000
Single Pair Wells
Lateral lengths for each well pair 1800-2400 ft.

10,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Months
Figure 11-20 Simultaneous vs. single fracture stimulation

with different proppant volumes
An additional example of the simultaneous fracture

Figures 11-18a and 11-18b Fracturing equipment
rigged up to the two wellheads for simultaneous fracture technique and the significance of fluid and proppant
treatments volume are noted in Fig. 11-20. This compares two
405
Modern Fracturing
Table 11-5 Cemented Liner, Four- to Five-Stage Treatment in the Core Area
Procedure
Fluid Proppant
Stage Volume Conc. Cum
Type Type Stage (lbs)
(gal) (ppa) (lbs)
1 Slick Water 30,000 Diagnostics
2 Slick Water 90,000 Pad
3 Slick Water 20,000 0.20 100% Sand, White, 40/70 4000 4000
4 Slick Water 90,000 Pad 4000
5 Slick Water 15,000 0.20 100% Sand, White, 40/70 3000 7000
6 Slick Water 90,000 7000
7 Slick Water 70,000 0.25 100% Sand, White, 40/70 17,500 24,500
12 Slick Water 70,000 0.50 100% Sand, White, 40/70 35,000 164v,500
20 Slick Water 15,000 0.50 100% Sand, White, 20/40 7500 380,500
21 Slick Water 7000 0.75 100% Sand, White, 20/40 5250 385,750
24 Slick Water 9366 Flush
Total 1,270,366 401,500
simultaneously fractured wells with two individually fluid recoveries may be a strong indication of the fracture
fractured wells in which the lateral length for the treatment being pumped out of zone and contacting
simultaneously fractured wells is approximately 20% the lower wet Ellenberger formation. Experience has
shorter than the second set of wells. All the wells used taught that observed net pressure increases during the
similar fluid volume; however the simultaneously fracture treatment gives an indication that the treatment
fractured wells were completed using 25% more is contacting additional reservoir through its complex
proppant and were fractured simultaneously to maximize geology. A final key to success with regard to treatment
the contact of the shale to the wellbore. design in the Barnett has been to optimize both fluid
Successful completions in the Barnett can also and proppant volumes in order to achieve the maximum
be characterized by several factors observed during contact of the reservoir to the wellbore.
and upon the initial flowback of the well. The initial
fracture gradients will typically be less than 0.75 psi/ 11-5.4 Barnett and Woodford
ft; higher fracture gradients have been observed and Gas Shale, Delaware Basin
may be an indication of the lack of the natural fracture
network needed for good production. In addition, the The Barnett and Woodford shale formations located
initial load recovery of treatment fluid on economically within the Delaware basin of West Texas are, for the
productive wells is usually less than 55%. Higher initial most part, in Culberson, Reeves and Pecos Counties.
406
Table 11-6 Six- to Seven-Stage Treatment in the Non-Core Area

Stage Description Volume Fluid Proppant Concentration
1 Load Hole 500 Water Frac
2 Acid 2500 15% HCl
3 Spacer 10,000 Water Frac
4 Acid 2500 15% HCl
5 Pad/Stepdown 40,000 Water Frac
6 5000 Water Frac 0.10 ppg White 100 Mesh
7 Pad 30,000 Water Frac
9 Pad 30,000 Water Frac
12 12,000 Water Frac 0.30 ppg White 100 Mesh
22 12,000 Water Frac 0.40 ppg White 40/70
24 5000 Water Frac 0.80 ppg White 40/70
29 9000 Water Frac 1.00 ppg White 40/70
30 Flush 9000 Water Frac
Total 605,500 353700
The Woodford was deposited initially in the late treatments extend into this water-bearing interval, they
Devonian period. The geochemical properties of can result in operational and economic failure. Early
both shales are very comparable to one another based indications are that the West Texas version of the Barnett
on the data currently available, particularly in TOC appears to have significant gas content; however, the
values. However thermal maturity appears to vary production mechanism particularly as it relates to
considerably over this expansive area of West Texas. rock mechanics appears to be significantly different
The Barnett Shale in this area varies from 7,400 from the North Texas Barnett.
to over 17,000 ft deep. The earliest completions, Shale is a complex mixture of clays, silicates,
located in the shallower area of the field, did not carbonates and detrital materials, and most shales
produce economic hydrocarbons, which is believed follow this mixture of mineralogy with various
to be a result of the immature nature of the shale degrees and percentages that may be an indication
in this area because the wells had indications of of how the shale might fracture (Simons, 1963).
being low-productivity oil wells. Comparing the mineralogy reports of several North
Shale thickness ranges for both Barnett and Texas shale cores with that of the US Permian Basin
Woodford formations vary from less than 100 to Shale (Barnett and Woodford) indicates that almost
over 600 ft. The Devonian limestone lies beneath the all the base components are similar, with relatively
Woodford and behaves very similar to the Ellenberger small differences in the percentages of silicates
of North Texas, in that it contains water, and if fracture present (usually 20 to 40% and clays ranging from
407
Modern Fracturing
Table 11-7 Mineralogy of the Permian Basin Shale

Mineralogival Analysis of West Texas Shales
a b c d e
Mineral Phases Barnett Shale Miss. Woodford
12,250 12,340 12,430 12,570 12,650
Quartz (SiO)2 41 27 33 42 37
Plagioclase Feldspar 3 4 4 6 4
Potassium Feldspar ~ ~ ~ trace ~
Fe-Dolomite ([Ca,Fe][CO3]2) 3 4 5 14 4
Siderite (FeCO3) ~ trace trace ~ ~
Apatite [(Ca5(PO4,CO3)3(F,OH,Cl)] 1 1 1 trace trace
Pyrite (FeS2) 3 4 4 2 4
Chlorite 2 2 1 trace 2
Mixed-Layer Illite>90/Smectite<10 46 57 51 35 49
TOTALS 100% 100% 100% 100% 100%
25 to 45%). The clays are comprised mostly of The location of carbonate minerals in the shale
mixed-layer illite/smectite with varied percentages of may also indicate a presence of natural fractures in
kaolinite, illite and mica. Some percentage of sulfides the reservoir rock or, in the case of the Permian Basin
(usually in the form of pyrite) and phosphates are Shale, their absence. The carbonate-bearing natural
generally present in both shale reservoirs. The most fractures present in the North Texas Barnett Shale
significant mineralogical difference noted is the can make a significant contribution to the total area
complete absence of carbonate material in the Permian exposed to an induced hydraulic fracture. This is one
Basin shale. It was noted previously that the North of the reasons it is theorized that slickwater fracturing
Texas Barnett Shale can be generally categorized as works so well in the North Texas Barnett. In contrast,
being approximately 1/3 clay, 1/3 silicate and 1/3 the absence of carbonate material of any significance
carbonates, while the shale in West Texas is more on in the Woodford and Barnett in West Texas could
the order 50:50 quartz and clay (BJ Services, 2006- indicate a lack of natural fracturing in the reservoir
06). Table 11-7 represents a typical mineralogy core rock or perhaps natural fractures that do not fracture
report from a West Texas Shale (note the absence as efficiently, meaning a softer clay-lined fracture
of carbonate minerals relative to the North Texas more prone to embedment, fracture face softening or
Barnett shale from Table 11-4). healing without adequate propping.
Carbonate material is typically found in the rock These shales tend to exhibit higher fracture
pore matrix or occurred as a secondary deposition gradients as compared to North Texas shale, typically
within the natural fractures. These carbonates are greater than 0.8 psi/ft. An additional factor to these
usually present in the form of Calcite (CaCO3) and higher fracture gradients is evidence of higher fracture
can also appear in less abundance (< 5%) as iron-rich initiation pressures as well as difficulty in obtaining
Fe-Dolomite ((Ca, Fe) (CO3)). Carbonates are known reasonable injection pressure. These high surface
to be relatively brittle and less elastic compared to treating pressures are often associated with near wellbore
clays and, to some degree, quartz. The absence of tortuosity many times caused be several smaller multiple
carbonates in the Permian Basin Shale may indicate fractures.(Suarez-Rivera, 2006).
a more elastic and less brittle reservoir rock that does This might also indicate that hydraulic fractures
not fracture as readily and perhaps as efficiently as created in this type of shale will be more conventional
silica-rich and carbonate-bearing shale. The absence of in nature and that fracture treatments, instead of
carbonates may also lend itself to rock face softening creating a broad network, result in the traditional
in the presence of some fluids. two-wing, vertical geometry.
408
11-5.5 Fayetteville Shale in Arkansas pressured. This led to initial completions utilizing
nitrogen foam fractures. These wells produced gas
The eastern side of the Arkoma basin is home to the at commercial and economic rates sufficient to
recent surge in drilling in the Fayetteville Shale, which warrant the drilling and completion of additional
lies between 2000 and 6000 ft deep and ranges from wells through 2004. As a result of the success in
75 to over 300 ft in thickness. Both the depth and these vertical wells and the industry success in the
thickness increase from the west to the east portion of Barnett Shale in the Ft. Worth Basin, the Fayetteville
the basin. The geology is complex with a multitude of Shale play has developed and evolved much more
fractures and faults related to the Ouachita Thrust Belt. rapidly (Mathews et al., 2007).
The Fayetteville Shale is Late Mississippian Chesterian In 2005, more vertical wells were completed with
in age and described as black, fissile, concretionary, good results, leading to the initial drilling and testing
clay shale with dark gray, fine-grained limestone inter- of horizontal wells. The lateral lengths of many of the
bedded within the shale package. The geochemistry of early drilled wells were in the 1,800 to 2,300 ft range
the Fayetteville is higher in thermal maturity relative to (Williams, 2006). These wells were completed with
the North Texas Barnett and has been confirmed with multi-stage fracture treatments using nitrogen foam.
dry gas production in Faulkner, Cleburne, Van Buren, The completions consisted of three to four stages
Conway, White, Pope and Franklin counties. averaging 150,000 to 200,000 lbs of proppant and
Horizontal wells have been tested, but lateral 200,000 to 300,000 gallons of total foamed fluid per
lengths have been limited due to the small size of the stage. The horizontal wells initial production proved to
fault blocks drilled to date. The Fayetteville play is still be substantially better than the vertical wells by several
early in the technological evolution for maximizing folds. Although these horizontal wells continued to be
reserve recovery as well as defining the limits of successful there was concern about decline rates and
TOC and thermal maturity for commercial deposits. costs associated with foam frac treatments.
TOCcontent has been reported to range between 4 The next step in evolution of completions in
and 9.5% (Brown, 2006a). This compares favourably these horizontal wells was the evaluation of methods
to the well developed Barnett Shale, with an average to reduce cost, improve production and flatten
TOC content of approximately 4.5%. The higher TOC decline rates. More recently, slickwater fracturing
content is favourable to the Fayetteville Shale, as its technology and crosslinked fluid treatments have been
overall thickness is smaller than what is seen in the attempted. Treatments have been four to five stages
more developed areas of the Barnett Shale. using approximately 450,000 gallons of slickwater
It appears the most productive portion of the and 200,000 lbs of proppant and 250,000 gallons of
Fayetteville Shale is where the interval contains a crosslinked fluid and 200,000 lbs of proppant.
higher percentage of chert and siliceous material
inter-bedded within the shale (Brown, 2006b). 11-5.5.1 Treatment Design
The upper Fayetteville Shale contains much Considerations Fayettville Shale
higher concentrations of mica/illite, which would
indicate a section of extremely low permeability, The initial results of incorporating changes in the
porosity and lack of natural fractures. stimulations and increasing the density of fractures along
The first development of the Fayetteville shale the horizontal lateral has resulted in not only shallower
occurred in early 2004, with a few vertical wells to declines but also improved production. One must
test further potential. The Fayetteville was found to remember that in order to fully develop shale reservoirs,
have mechanical and physical rock properties similar a primary goal should be to maximize the amount or
to the Barnett Shale (Southwestern Energy, 2004). volume of rock exposed to the wellbore.
The reservoir pressure, however, trended toward being To compare the nitrogen foam and slickwater
more normal to somewhat underpressured, whereas stimulation techniques used in the Fayetteville Shale,
the Barnett Shale is considered to be somewhat geo- a production comparison of two offset wells located
409
Modern Fracturing
in Van Buren County, Arkansas that have very similar 300 ft thick. Much like the Fayetteville to the east,
lateral lengths and completion intervals, is included the geology is complex with the potential to contain
(Fig. 11-21). Two offsetting Fayetteville Shale wells were fractures and faults related to the Ouachita Thrust Belt.
stimulated: one using nitrogen foam and the other using The Woodford and Caney shales are Lower Mississippian
a large volume of slickwater. The foam-fractured well was Upper Devonian in age and described as black, fissile,
stimulated with about 1.2 million gallons of nitrogen concretionary, clay shale with dark-grey, fine-grained
foam and 750,000 lbs of proppant. The slickwater limestone inter-bedded within the shale package. The
treatment used about 2.8 million gallons of slickwater geochemistry of the Woodford and Caney, like the
and 940,000 lbs of proppant. The wells were completed Fayetteville, is higher in thermal maturity relative to
in the same month, with cumulative production for the North Texas Barnett.
the foam-fractured well at around 200,000 Mscf The current area of significant activity is centered
of gas and the slickwater-fractured well at 340,000 on the southeastern portion of Oklahoma in Coal,
Mscf of gas. Although the reason for this difference Hughes, and Pittsburgh counties. A recent gross
can be partly attributed to the volume of proppant production from the Woodford Shale has been reported
placed, a more important factor is that the volume of to approach 100 MMscf/d. The initial completions in
fluid should create opportunity for more shale to be these shale reservoirs involved drilling vertical wells,
fractured and exposed to the wellbore. The trend for the where the primary target zones included the Wapanuka,
area is to perform more slickwater and crosslinked gel Caney, Cromwell, Hunton, and Woodford. In over 100
treatments based on the well performance improvements wells drilled, intervals were completed individually and
(Southwestern Energy, October 2006). commingled for production.
Operators found the Woodford Shale contributed
100000
significant production to the wells and began
Slickwater
exploitation of the Woodford alone. Given the success
of the Barnett Shale to the south and implementation
Mscf/ Month
N2 Foam
10000 of new drilling and completion techniques, horizontal
wells have been drilled and tested with excellent
Slickwater 4 Stages 940 klbm Prop successes. This Woodford/Caney play is early in the
N2 Foam 4 Stages-750 klbm Prop technological evolution for maximizing reserve recovery
1000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 as well as defining the boundary limits. Although these
Months shale intervals may be thin in areas, the TOC (Total
Figure 11-21 Production after a nitrogen foam frac vs.
Organic Carbon) content can be in the 6 to 8% range
slickwater frac in the Fayetteville Shale
as compared to the 4.5% that was previously discussed
This rapidly growing Fayetteville Shale area is in the Barnett (Haines, 2006).
continuing to be defined as far as its extent and has An important factor to note is that the
become an extremely large logistical project because of composition of the Woodford and Caney shales
the activity planned. have some marked differences. The Woodford more
resembles the Barnett, with siliceous material inter-
11-5.6 Woodford/Caney Shale, Arkoma Basin bedded within the shale, whereas the Caney Shale
contains higher concentrations of clay minerals. This
On the western side of the Arkoma basin are two shale higher clay content is indicative of a reservoir that
reservoirs that, like the Barnett or Fayetteville shales, will not respond well to stimulation treatments, as
are not new discoveries but have been known for quite the West Texas shales have shown.
some time. The Woodford and Caney shales have been The Woodford however has provided renewed
productive in non-specific areas with inconsistent interest with the implementation of drilling horizontal
success. Woodford and Caney wells are generally 6,000 wells and multiple-stage fracture stimulations along the
to over 11,000 ft deep and pay ranges from 50 to over lateral. The lateral sections have ranged from 2,000 to
410
Table 11-8 Typical Pump Schedule in the Woodford Shale

Stage Volume Fluid Proppant Concentration Proppant
1 2500 Slick Water
2 2500 15% HCl
3 2500 Slick Water
4 2500 15% HCl
5 80,000 Slick Water
6 5000 Slick Water 0.50 ppg 100 Mesh Sand
7 9500 Slick Water
8 2500 Slick Water 1.00 ppg 100 Mesh Sand
10 9500 Slick Water 0.10 ppg White 30/50
11 9500 Slick Water
13 9500 Slick Water
15 9500 Slick Water
17 9500 Slick Water
19 9500 Slick Water
21 9500 Slick Water
23 9500 Slick Water
25 9500 Slick Water
27 9500 Slick Water
29 9500 Slick Water
31 9500 Slick Water
33 9500 Slick Water
35 9500 Slick Water
37 9500 Slick Water
39 9500 Slick Water
41 9500 Slick Water
Total 446,000 191,200 pounds
over 3,500 ft, with initial production rates reported as with continuously placing proppant into the reservoir.
high as 10 MMscf/d (Newfield Exploration Company Although the proppant size was reduced, additional
Press Release, November 2006). problems were encountered: either fracture width
Using the information learned from the Barnett was not sufficient near the wellbore or significant
Shale, slickwater fracturing became the most viable proppant fallout in the fracture caused premature
and economical stimulation method for these shales. screen-out. Therefore, pad sweeps were introduced
The early stimulation attempts experienced difficulty to the design, which allowed the proppant bridge or
411
Modern Fracturing
settled prop to be washed away from the near-wellbore 32% range and illite/smectite clays comprising 45
area. Including pad stages between proppant stages to 55% of the matrix. The remainder of the rock
also allows injection of an increased volume of fluid consists of various amounts of minerals such as
into the well. Increased fluid volume allows more siderite, pyrite, mica/illite, and significant quantities
opportunity to maximize zonal coverage. A typical of chlorite. The mineralogy of this rock is somewhat
treatment might consist of 500,000 to 700,000 variable and could change as different areas of the
gallons of fluid and about 200,000 lbs of 30/50 Floyd Shale are explored. This formation is in the
or 40/70 northern frac sand per stage. Table 11-8 earliest stages of development, with operators drilling
shows an example of a pump schedule containing the a few wells to sample the potential.
proppant stages and pad sweeps as an example. The The Conasauga Shale, by comparison, has a higher
trend has been to increase the density or number or thermal maturity and appears to be more similar to
frac stages in horizontal wells. the Fayetteville Shale in Arkansas.
Both the Woodford and Caney shales, although Completion attempts have mimicked both
well-known potential producers, are in their infancy the Fayetteville Shales large-volume nitrogen
as far as the development, particularly with the foam fracture treatments and the Barnett Shales
application of horizontal well technology and continued large-volume slickwater fracture treatments. In
improvements in stimulation techniques. addition there has been some use of and continued
interest in CO2 foam fluids as well as 100%
11-5.7 Floyd Shale/Conasauga Shale, CO2 fracture treatments.
Black Warrior Basin (Alabama) Several horizontal wells have also been drilled
and completed using nitrogen foam fluids as well
On the western side of Alabama in and around the as slickwater. By the time of this writing there has
Black Warrior Basin is a shale reservoir that has not been a commercial or economical completion
recently received much interest as another Barnett reported; however, it is important to consider
- the Floyd and Conasauga shales. The Floyd Shale that this interval is on the eastern edge of the
has had significant leasing activity but very few wells Mississippi Embayment, and higher-potential shale
completed by the time of this writing. This interval sections may yet be discovered.
occurs at depths from 4,000 to over 6,800 ft deep
and ranges between 50 ft and over 200 ft thick. The 11-5.8 Mancos and Lewis Shales
Conasauga Shale is thicker and occurs at depths of
approximately 8,000 ft. These shale reservoirs have The Mancos and Lewis shales are located within
been compared to both the Barnett Shale in North the San Juan Basin of northwestern New Mexico
Texas and the Fayetteville Shale in Arkansas. The and southwestern Colorado. The San Juan Basin is
rock is best described as a mixture of fine-grain clays, one of the more prolific producers of natural gas
silicates, plagioclase feldspars and other detrital in the Rocky Mountain Region. This area has been
materials. The intergranular porosity is poor with known for its prolific coalbed methane production
most of the microporosity associated with the clays. and production from several tight gas sands. This
There is little evidence of any natural fracturing. is a relatively thick shale ranging from 1000 to over
When compared to other productive shale reservoirs, 1500 ft. This helps compensate for a lower average
the Floyd Shale lacks significant amounts of inter- TOC content of 0.45 to 2.5% when compared to
bedded siliceous or calcareous material (Mathews et other shale reservoirs. The shale section consists of
al., 2007). Based on the history of productive shales, layers of shale, siltstone and sands, which can effect
this combination of material and the lack of natural the completion. The rock matrix is much like that
fractures would indicate a rock that is more elastic of other shale sections with porosity of less than 2%
and less brittle. In general, the Floyd Shale cores and permeability as low as 0.0001 md. The reservoir
examined to date have quartz content in the 28 to is typically separated into four sections: Ute, Navajo
412
City, First Bench of Otero and Second Bench of Mineralogy Evaluate the mineral content of the
Otero. (PTTC, 2001) The majority of interest in this rock to help determine fluid sensitivity, anisotropy
shale is associated with the Navajo City and Otero and potential for clay/fines mobilization
benches because most of the natural fractures tend to Water saturation To determine potential for
be located within these sections. These sections also water production and fluid clean-up issues
have more silica- and quartz-rich material and are Permeability Both vertical and horizontal
therefore more prone to fracturing. Porosity
Because the reservoir pressure is low (1,000 to Natural fractures
1,500 psi), this reservoir is typically fracture stimulated Rock mechanics Determine elastic properties,
using nitrogen foam fluids. The economics of this embedment potential, and the stress contrast for
reservoir work well because most of the Mancos and zonal isolation and or horizontal drilling potential
Lewis shale completions are commingled with other
tight gas intervals such as the Mesa Verde and Dakota Fracture identification logs have been important
(Shirley, 2001). However, it is difficult to evaluate the in showing the existence of natural fractures as well
incremental production from the shale. It has been as helping determine the orientation of natural and
estimated that the average production is in the range induced fracture direction. Most shale reservoirs
of 100 to 300 Mscf/d per well. depend upon some level of microfractures or natural
fractures for enhanced production.
11-6 Shale Treatment Because the general mineralogical properties of
Design and Evaluation shale reservoirs consist primarily of clay minerals
such as smectite, a perceived fluid sensitivity
11-6.1 Stimulation and Treatment led operators to minimize the fluid placed in these
Design for Shale Reservoirs reservoirs. In addition, the early development of shale
reservoirs occurred in the Michigan Basins Antrim
Essentially all shale reservoirs require hydraulic Shale and the Devonian shale section in the North
fracturing to be economically viable. In reviewing the East with bottomhole pressures of less than 1000
US shale reservoirs described above, it can be seen that psi, which roused concerns not about fluid clean
a variety of hydraulic fracturing techniques have been up, leading to treatment designs using N2 or CO2-
utilized. In order to determine the best stimulation based foam fluids. Increasing natural gas price and
method, all aspects of the reservoir must be evaluated. government tax credits resulted only in increased
Prior to the design of a fracturing treatment, several activity in many shale plays throughout the country
steps should be taken not only to quantify the reservoir but also experimentation with new stimulation and
properties but also to qualify the potential of the shale completion technology. Operators began to spend
to be productive and, more importantly, economic time to better evaluate these unconventional shale
to produce. This should initially include some form reservoirs and better understand the mechanisms that
of reservoir evaluation/assessment. A geophysical effect production and completion.
evaluation of the reservoir using 3-D and 4-D seismic Although slickwater fracturing has been used in
has become extremely important to properly identify gas reservoirs for over 30 years, experimentation in
the extent and potential of these unconventional unconventional reservoirs didnt start until Mayerhofer
shale intervals. Other steps prior to developing et al. (1997) presented a new understanding of
a completion plan must include well testing and fracturing tight gas sands. In this case the success
an understanding of the reservoir size. Significant of large-volume conventional fracture treatments in
information has been obtained though core analysis the Barnett Shale and an effort to reduce completion
and well logs. Core analysis is used to evaluate the costs led to the use of large slickwater treatments.
stimulation potential success and challenges and to Of course, the reason behind the slickwater success
determine shale properties such as: is that that in these very low-permeability reservoirs,
413
Modern Fracturing
very long fractures are indicated, with width as a Relatively Dry Gas-Saturated Rock. With high in
secondary priority. Thus, very large volumes can be situ water saturation, fluids could block formation
injected, extending the length greatly without the pore throats and adversely effect conductivity of
need for large propped widths. these low-conductivity reservoirs.
The subsequent success of large-volume slickwater Very Low Permeability. These reservoirs
treatments in the Barnett Shale has made this type of require large surface-area fractures and fracture
job the treatment of choice. Although there is a broad extension with low leak-off.
variation in the design of a slickwater treatment, the Natural Fracture System. Reservoirs with
common goal is to establish a conductive fracture microfractures or other natural fractures will
that exposes as much reservoir as possible to the contribute production.
wellbore. The success of slickwater fractures in Formation Barriers. Some level of barrier above
unconventional gas reservoirs is based on several and below the zone offers the possibility of keeping
factors (Schein, 2004-2005): the frac in zone and achieving frac containment
In very low-permeability reservoirs, the length of (see Section 4-3.3.3 for a discussion on barriers
the fracture is by far the most important feature to fracture height growth).
of the fracture. Width is far less important (see Low Formation Sensitivity to Water. Assessing
Sections 4-5.3 and 4-7.1). formation sensitivity to large volumes of
Fractures create rough surfaces, and some water may require core flow testing and
residual conductivity may exist in these mineralogical analysis to confirm.
fractures after closure. New Economic Considerations. Wells in which
Natural fractures contribute to production. massive and large-volume fracture treatments
Tight, low-pressure reservoirs that exhibit clean- were volumetrically successful but economically
up problems with conventional gels. marginal. (This is caused by a number of factors
The failure mechanism in the rock where including product price, depletion and poor
the potential for shear slippage or formation reservoir quality, among a few.)
deformation may occur Non-Depleted Reservoir Pressure. Wells must still
Slickwater fracture treatments may provide have sufficient reservoir pressure to assist in flow back
for increased fracture extension and potential and clean-up of the fracturing fluids.
height containment. Low Water Saturation. Too much water can tend to
over-saturate the reservoir, leading to poor fluid
It is important to note that for slickwater recovery and poor production response.
fracturing to be successful a reservoir
should exhibit a number of characteristics With the fracture mechanism of most shale reservoirs
that enable low-proppant-concentration fractures and the above criteria, it becomes evident why large-
to produce gas. There is confusion sometimes volume slickwater treatments are tending to become the
when talking or writing about these fractures. The preferred method of fracturing for most shale gas plays.
conductivity of these fractures is actually not The differentiating factor with regard to choosing the
low if you speak of dimensionless conductivity, correct fracturing fluid tends to be the reservoir pressure.
which is the ratio of fracture permeability times Low-pressure (< 1,000 psi) reservoirs tend to be in a
fracture width, divided by reservoir permeability more relaxed stress environment and may not clean
and fracture length. The slickwater treatments up well with large volumes of fluids. These reservoirs
typically result in long fractures with little propped such as the Antrim, Devonian, and Conasauga are
width. Consider the following as reservoir screening being fractured with N2 and CO2-based foam systems.
criteria for slickwater fracture treatments; wells with More normally or over-pressured reservoirs have
the following characteristics will tend to be more shown the best success utilizing large volumes of water
successful(Schein, 2004-2005): with low concentrations of proppant.
414
An excellent example of the evolution of fracturing minimize horsepower requirements. With this large
fluid and treatment type can be seen in the production volume of fluid one should also consider the use of
life of the discovery well in the Barnett Shale: the C. a bactericide, scale inhibitor, and iron scavenger. In
W. Slay #1 well drilled in 1981 (Fig. 11-22). Because addition, surfactants have been used with varying
of the concern with fluid sensitivity, the initial degrees of success. Microemulsion surfactant
stimulation consisted of about 1.5 MMscf of nitrogen technology has provided measurable success in
without proppant. Although the well responded with several areas of the Barnett Shale.
gas production, the flow rate was not economically Proppant type and size have continued to trend
sufficient. The well was re-fractured approximately toward the use of smaller-mesh proppants. As
a year later with 132,000 gallons of CO2 foam fluid discussed previously, the proppant initially used in
and approximately 188,000 lbs of 20/40 white sand. fracturing shale was typically 20/40 mesh white sand.
Production doubled, and this well continued to Seeing the increased need to place more proppant,
produce at marginally economic rates for the next 19 proppant bridging and premature screenout became
years. Based on the success of slickwater fracturing in concerns, particularly since water was being used as
other Barnett Shale wells, the well was re-fractured the primary transport mechanism. Therefore 40/70
again in mid-2000 using about 31,000 bbls of mesh proppant became available and provided several
slickwater and 95,000 lbs of 20/40 white sand. After benefits. Although its conductivity is approximately
almost 20 years of production, the well responded one-half that of 20/40 mesh proppant, the 40/70
with production rates peaking at 1.475 MMscf/d proppant average particle size diameter is also
and has had cumulative production of about 1.25 Bcf approximately one half that of 20/40 mesh. As a result
since the re-fracture with slickwater (see Section 13-5 3 to 4 times the volume of 40/70 proppant can be
for more discussions on re-fracturing). pumped, resulting in even longer propped fractures. In
addition, in several areas ultra-lightweight proppants
(ULWPs) in 20/40 and 45/65 mesh sizes are used to
Gas Production, Mscf/ Month
132,000 gals CO2 Foam

& 188,000 lbm Ottawa increase proppant-carrying capacity (see Section 8-
10000 31,000 Bbls Slick Water 3.4). Table 11-9 shows the settling velocity of various
95,000 lbm 20/40 Ottawa
size and specific gravity proppants in slickwater, and
Figures 11-23 and 11-24 compare conductivity of
ULWPs with natural sands.
1.5 MMscf N2 frac Table 11-9 Settling Velocity of Various Proppants in

100 Slickwater
1
16
31
46
61
76
91
6
1
6
1
6
1
6
1
6
1
6
1
10
12
13
15
16
18
19
21
22
24
25
27
Specific Settling Velocity

Months 20/40 Proppant
Gravity ft/minute
Figure 11-22 Stimulation and production evolution in Bauxite 3.65 23.2
the Barnett Shale discovery well, CW Slay #1 White Sand 2.65 16.6
The transition to slickwater fracturing treatments Resin Coated
2.55 15.9
Sand
in new shale play areas has occurred much more quickly
ULWP (prop = 125) 1.25 4.3
in recent years. This is a result of the relatively lower
cost of slickwater as a fracturing fluid and the ability
Specific Settling Velocity
to pump a large volume of fluid to contact/connect a 40/70 Proppant
Gravity ft/minute
large surface area of shale with the wellbore. Bauxite 3.65 6.0
The large volume of fluid being pumped and White Sand 2.65 4.4
the high rates necessary to create and extend a Resin Coated
2.55 4.2
Sand
fracture into the shale require several key additives
45/65 ULWP
to the fluid. Obviously, the use of friction reducer (prop = 125)
1.25 1.2
is needed to achieve the higher pump rates and
415
Modern Fracturing
5000
11-6.2 Fracture Modeling
4500
4000 Utilizing conventional fracture models proves to be
3500 a challenge when considering the complexity and
Conductivity md. ft
3000 variation in how shale reservoirs fracture. The original

20/40 ULW-125
2500 pseudo-3-D fracture models were developed to model
2000 conventional gelled or crosslinked fracture fluids,
1500 assuming typical two-wing fracture growth (see Section
1000
4-3.3.2). The hydraulic fracturing of shale reservoirs
500
20/40 White Sand and the use of slickwater in unconventional reservoirs
0
0 1000 2000 3000 4000 5000 6000
has radically changed how frac models can and should
Stress, psi
be used: Due to the complex nature of frac growth
20/40 ULWP (Hprop = 125) 20/40 White Sand
and broad extension of the fracture network in this
environment, the use of a two-wing fracture model
Figure 11-23 Conductivity comparison of ultra- is not realistic. There are, however, a few deep, high-
lightweight proppants with 20/40 sand
temperature, high-pressure (HTHP) shale reservoirs in
1000
45/65 ULW Conductivity
the early stages of exploitation, where the shale appears
to be more elastic in nature. In these cases, the hydraulic
fracture may be prone to more conventional bi-wing
Conductivity md. ft
100
40/70 White Sand fractures. Before using a fracture model, its important
45/65 ULW Permeability to understand the fracture geometry and the potential
for complex fracture systems.
10
11-6.3 Summary
1
0 1000 2000 3000 4000 5000 6000 It is evident that shale reservoirs are much more
Stress, psi complex than most other tight gas reservoirs being
Conductivity, ULWP 40/70 Permeability, White Sand
produced today, with the possible exception of coalbed
Conductivity, White Sand Permeability, ULWP
methane. Although the Barnett Shale has only been
Figure 11-24 Conductivity comparison of 45/65 ultra- actively produced since the late 1980s it has become
lightweight proppant (prop = 1.25) with 40/70 mesh
white sand the model for other shale reservoirs throughout the
world. Unconventional thinking and applications
Most recently the concept of smaller-mesh proppant of a diverse number of technologies have created a
has extended to the use of 100 mesh sand as the primary frenzy of activity over the past ten years. Not every
proppant. The small mesh size allows the proppant to be shale is just like the Barnett Shale; however, many of
more easily carried in the fracturing fluid and can also the tools used to evaluate shale reservoirs have been
enable placement in small fracture systems. One might developed in this time frame.
be concerned that 100 mesh proppant might act as a Several technological advancements have provided
sealant because 100 mesh sand has been used for years in a better understanding of how to complete and
conventional fracturing as a fluid loss additive. However, produce these complex formations. The application
in slickwater fracturing, with the low concentration of of horizontal drilling has had a significant impact on
proppant being placed in the fracture system, fracture production and potential reserve recovery. Fracture
conductivity may be a result of simply holding open stimulation technology and application improvements
a fracture, thus allowing maximum conductivity. have and will continue to impact decisions about
Additional experiential evaluation is needed to evaluate what treatment method will provide optimum
long-term production issues that may exist. economic production from shale reservoirs. Whether
416
the best method is slickwater, foam, crosslinked fluid zones. Lastly it is important to evaluate the mineralogy
or some form of hybrid stimulation treatment, each of the shale itself. Silica-rich shales containing calcite-
case must be evaluated individually because one filled natural fractures are easily stimulated and have
size does not and will not fit all. proven to be the best producers. By contrast, the
The ultimate goal of gas shale completions is to expose more clay-rich shales that lack interbedded silica
and interconnect the maximum surface area of shale to and have zero or low volumes of carbonate materials
the wellbore. Drainage penetration into the matrix of have been the most challenging not only to fracture
the shale is measured in inches due to the extremely low but also to produce commercially.
permeability. Therefore, economical completions must Additional unconventional thinking and
connect vast quantities of rock surface through fractures applications will be required to develop many of these
to generate sufficient production volumes. Due to limits gas shale plays in the future.
on bottomhole treating pressure, rate and fluid volume, Outside of the USA, very few shale gas reservoirs
knowledge of the bounding rock layers is required to have been fractured. Instead, the work performed has
design optimal completions. been in the nature of exploration and appraisal, rather
In the case of proximity to water-bearing zones, than development. This is largely due to the relative size,
horizontal wellbores have been utilized to minimize immaturity and productivity of gas reservoirs outside
height growth and improve aerial fracture complexity of North America. Put simply, the gas industry outside
and thereby expose the maximum rock area within of the US is not yet short enough of conventional
the gas shale. With better understanding of fracture gas to turn en masse to shale gas. Additionally, the
complexity generated from horizontal wellbores, massive fracturing infrastructure required to complete
infill drilling in well-bounded areas (e.g., Barnett shale gas wells effectively only exists in a few isolated
Shale Newark East core area) has intensified due locations outside of North America. However, many
to the improved production rates achieved with countries do not have their own independent sources
tighter spacing. Variations in horizontal completion of gas, and the governments of these countries are
technology abound as engineers experiment with continuously looking for secure sources of energy that
perforation cluster design, stage length, pump rate, they can control. Consequently, it is only a matter of
fluid type and volume, proppant type and proppant time before the world starts to look at the technologies
concentration, seeking to find the optimum and techniques developed in places like the Barnett
combination for the geologic environment of each and Fayetteville shales by the pioneering independent
area. The Newark East field has demonstrated how operators of the US gas industry.
the optimum treatment design can vary significantly
among wells within a field.
Micro-seismic fracture mapping now provides References
valuable information concerning the best direction for
a horizontal well or the extent and zonal coverage of a Ahmed, T., Centilmen, A., and Roux, B.: A Generalized
stimulation treatment. The use of 3-D and 4-D seismic Material Balance Equation for Coalbed Methane
information has significantly improved the potential for Reservoirs, paper SPE 102638, 2006.
success. Thorough understanding of the shale reservoir Aminian, K., Ameri, S., Bhavsar, A., and
(depth, thickness, thermal maturity, lateral extent, Lakshminarayanan, S.: Type Curves for
porosity, interbedded permeability, natural fracturing, Production Prediction and Evaluation of Coalbed
and mineralogy) is key to any successful completion. Methane Reservoirs, paper SPE 97957, 2005.
The extent of natural fracturing, as discussed in this Aminiam, K., Ameri, S., Bhavsar, A., Sanchez, M., and
chapter, is one of the significant factors in identifying Garcia, A.: Type Curves for Coalbed Methane
the most productive intervals and best wells; however, Production Prediction, paper SPE 91482, 2004.
these fractures may extend (or be extended through Arri, L.E., Yee, D., Morgam, W.D., and Jeansonmne,
stimulation) into water-bearing or non-productive M.W.: Modeling Coalbed Methane Production
417
Modern Fracturing
with Binary Gas Sorption, paper SPE Chen, Z., Khaja, N., Valencia, K.L., and Rahman, S.S.:
24363,1992. Formation Damage Induced by fracture fluids in
Baltoiu, L.V., Warren, B.K., and Natras, T.A.: State Coalbed Methane Reservoirs, paper SPE 101127,
of the Art in Coalbed Methane Drilling Fluids, 2006.
paper SPE/IADC 101231, 2006. Clarkson, C.R.: Application of a New Multi-
Barba, R.E., Dein, C.W., and Woodruff, R.A.: Component Gas Adsorption Model to Coal Gas
Application and Evaluation of Advanced Adsorption Systems, SPE Journal, 8 (3), 236-250
Completion Optimization Technology in the Black (2003).
Warrior Basin, paper SPE 36673, 1999. Clarkson, C.R., and Bustin, R.M.: The Effect of Pore
Bassett, L.: Guidelines to Successful Dewatering of Structure and Gas Pressure Upon the Transport
CBM Wells, paper SPE 104290, 2006. Properties of Coal: A Laboratory and Modeling
Barree, R.D.: Application of Pre-Frac Injection/Falloff Study 2. Adsorption Rate Modeling, Fuel, 78,
Tests in Fissured Reservoirs Field Examples, 1345-1362 (1999).
paper SPE 39932, 1998. Conway, M.W.: A Chemical Model for Coal-Fluid
Bazan, L.W., and Larkin, S.D.: Limited Entry Interactions, paper presented at the 1992
Techniques Prove Successful in Simultaneously International Gas Research Conference, Orlando,
Stimulating the Fruitland Coal and Pictured Cliffs Fla.
Formations in the San Juan Basin, paper SPE Coulter, G.R., Benton, E.G., Thomson, C.L.:Water
74353, 2002. Fracs and Sand Quantity: A Barnett Shale
Bazan, L.W., Larkin, S.D., Jacot, R.H., and Meyer, Example, paper SPE 90891, 2004.
B.R.: Modeling of Simultaneous Proppant Crump, J.B. and Conway, M.W.: Effects of Perforation
Fracture treatments in the Fruitland Coal and Entry Friction on Bottomhole Treating Analysis,
Pictured Cliffs Formations in the San Juan Basin, paper SPE 15474, 1986.
paper SPE 78694, 2002. Cui, X., Bustin, R.M., and Dipple, G.: Selective
Bell, G.J. and Jones, A.H.: Variation of Mechanical Transport of CO2 CH4 and N2 in Coals: Insights
Strength with Rank of Gassy Coals, Proc. 1989 from Modeling of Experimental Gas Adsorption
Coalbed Methane Symp., paper No. 8924, Data, Fuel, 83 (3), 293-303, (2004).
Tuscaloosa, AL, April 1989. DeGance, A.E., Morgan, W.D., and Yee, D.: High-
Berkowitz, N.: Coal Science and Technology 7-The Pressure Adsorption of Methane, Nitrogen and
Chemistry of Coal, Elsevier Science (1985) Table Carbon-Dioxide on Coal Substrates, Fluid Phase
3.16, p.88. Equalibria (February, 1993) 215.
BJ Services Company, Tomball Technology Center, Delmbacher, F.X., Economides, M.J., Helnemann,
Goemechanics & Petrologic Laboratories, Various Z.E., and Brown, J.E.: Comparison of Methane
shale core reports, 2004 to 2006. Production From Coalbed Using Vertical or
Brown, D.: Barnett May Have Arkansas Cousin, Horizontal Fractured Wells, JPT (August 1992),
AAPG Explorer (February 2006a) 8. 930-935.
Brown, D.: Shales Require Creative Approaches, Dindoruk, B.: Analytical Theory of Multiphase,
AAPG Explorer, (November, 2006b) 6. Multicomponent Flow in Porous Media, Ph.D.
Castillo, J.L.: Modified Fracture Pressure Decline dissertation, Stanford University, Palo Alto, Calif.
Analysis Including Pressure-Dependant Leakoff, (1992).
paper SPE 16417, 1987. Duel, M. and Cervik, J.: Methane Drainage in the
Chaback, J.J., Morgam, W.D., and Yee, D.: Sorption Pittsburgh Coal, Proc. 17th Intl. Conf. of Mining
of Nitrogen, Methane, Carbon-Dioxide and Safety Res., Varna, Bulgaria, October 1977, pp9-
Their Mixtures on Bituminous Coals at In-Situ 15
Conditions, Fluid Phase Equalibria (March, Espie, A.A.: CO2 Capture and Storage: Contributing
1996) 289. to Sustainable World Growth, paper IPTC 10936,
418
2005. Bennaceur, K., Gupta, N., Rigg, A., Hovorka,

Faraj, B., H. Willims, G. Addison, and B. McKinstry, S., Myer, L., and Benson, S.: Critical Issues on
2004, Gas potential of selected shale formations CO2 Capture and Storage: Findings of the SPE
in the Western Canadian Sedimentary Basin: Advanced Technology Workshop (ATW) on
Houston, Hart Publications, Gas TIPS, v. 10, no. Carbon Sequestration, paper SPE 102698, 2006.
1, p. 21-25 Jadhav, M.V.: Enhanced Coalbed Methane Recovery
FETC, BP-Amoco and IEA GHG Programme, CO2 Using Microorganisms, paper SPE 105117,
Capture and Geologic Sequestration: Progress 2007.
Through Partnership Workshop Summary Jahediesfanjani, H., and Civan, F.: Effect of Resident
Report, Houston, Texas, 29-30 September, 1999, Water on Enhanced Coal Gas Recovery by
p. A128. Simultaneous CO2/N2 Injection, paper SPE
Fisher, M.K., Wright, C.A., Davidson, B.M., Goodwin, 102634, 2006.
A.K., Fielder, E.O., Buckler, W.S., and Steinsberger, Jarvie, D.M.:Exploring for and Producing Shale
N.P.: Integrating Fracture Mapping Technologies Gas:Application of Detailed Geochemical Data
to Optimize Stimulations, paper SPE 77441, in Ascertaining Prospectivity and Producibility,
2002. WTGS Fall Symposium, 2006
Fredd, C.N., Olsen, T.N., Brenize, G., Quintero, B.W., Johnson, R.L., Flottman, T., and Campagna,
Bui, T., Glenn, S., and Boney, C.L.: Polymer-Free D.J.: Improving Results of Coalbed Methane
Fracturing Fluid Exhibits Improved Cleanup for Development Strategies by Integrating
Unconventional Natural Gas Well Applications, Geomechanical and Hydraulic Fracturing
paper SPE 91433, 2004. Technologies, paper SPE 77824, 2002
Gray, I.: Reservoir Engineering in Coal Seams: Part 1 Johnson, R.L., Scott, S., and Herrington, M.: Changes
The Physical Process of Gas Storage and Movement in Completion Strategy Unlocks Massive Jurassic
in Coal Seams, SPERE (February, 1987) 28. Coalbed Methane Resource Base in the Surat
Greaves, K.H. et al.: Multi-Component Gas Basin, Australia, paper SPE 101109 presented at
Adsorption-Desorption Behavior of Coal, Proc. the 2006 Asia Pacific Oil & Gas Conference and
1993 Intl Coalbed Methane Symp., Birmingham, Exhibition, Adelaide, Australia, 11-13 September.
AL. Jones, A.H., Ahmed, U., Bush, D.D., Holland, M.T.,
Haines, Leslie: Activity Builds In Woodford Shale, Kelkar, S.M., Rakop, K.C., Bowman, K.C., and
Shale Gas Supplement to Oil and Gas Investor, Bell, G.J.: Methane Production Characteristics
(January 2006) 17. for a Deeply Buried Coalbed Reservoir in the San
Hall, F.E., Zhou, C., Gasem, K.A.M., Robinson, Juan Basin, paper SPE/DOE/GRI 12876, Proc.
R.L., and Yee, D.: Adsorption of Pure Methane, Unconv. Gas Recov. Symp., p.417, Pittsburgh, Pa.,
Nitrogen and Carbon Dioxide and Their Binary May 1984.
Mixtures on Wet Fruitland Coal, paper SPE Jones, A.H. et al: A Review of the Physical and
29194, 1994. Mechanical Properties of Coal with Implications
Hayden, J. and Pursell, D., Pickering Energy Partners: for Coalbed Methane Well Completions and
The Barnett Shale: Visitors Guide to the Hottest Production, in Geology and Coalbed Methane
Gas Play in the U.S., October, 2005 Resources of Northern San Juan Basin (ed. J.E.
Hanby, K.P.: The Use of Production Profile for Fosset), Colorado and New Mexico Rocky Mtn.
Coalbed Methane Valuations, Proc. 1991 Assoc. of Geologists, Denver, 1988.
Coalbed Methane Symp., University of Alabama, Joubert, J.I., Grein, C.T., and Bienstock, D.: Sorption
Tuscaloosa, AL, 443-452. of Methane in Moist Coal, Fuel, (July, 1973)
Holditch, S.A.: Completion Methods in Coal-Seam 181.
Reservoirs, JPT (March 1993), p.270. King, G.R., and Ertekin, T.: A Survey of Mathematical
Imbus, S., Orr, F.M., Kuuskraa, V.A., Kheshgi, H., Models Related to Methane Production from Coal
419
Modern Fracturing
Seams, Part 1: Empirical and Equilibrium Sorption Methane Reserves with Insight into the Advantages
Models, 1989 Coalbed Methane Symposium. and Disadvantages of Each Method, paper 8716,
University of Alabama, Tuscaloosa, AL, (1989a). Proc. 1987 Coalbed Methane Symp., University of
King, G.R., and Ertekin, T.: A Survey of Mathematical Alabama, Tuscaloosa, AL, 73-79.
Models Related to Methane Production from Coal Manik, J., Ertekin, T., and Kohler, T.E.: Development
Seams, Part 2: Non-equilibrium Sorption Model, and Validation of a Compositional Coalbed
1989 Coalbed Methane Symposium. University of Simulator, JCPT, 41, No. 4, (2002).
Alabama, Tuscaloosa, AL, (1989b). Maricic, N., Mahaghegh, S.D., and Artun, E.: A
King, G.R., Ertekin, T., and Schwerer, F.C.: Numerical Parametric Study on the Benefits of Drilling
Simulation of the Transient Behavior of Coal Seam Horizontal and Multilateral Wells in Coalbed
Degasification Wells, SPEFE (April, 1986) 165; Methane Reservoirs, paper SPE 96018, 2005.
Trans., AIME, 281. Mathews, H.L., Schein, G., and Malone, M.:
King, G.R.: Material Balance Techniques for Coal Stimulation of Gas Shales, Theyre all the same...
Seam and Devonian Shale Gas Reservoirs, paper Right? paper SPE 106070, 2007.
SPE 20730, 1990. Mavor, M.J.: Western Cretaceous Coal Seam Project,
Kutas, G.M.: Remedial Treatment for Coal Degas July 1991 Progress Report, Letter to R.A. McBane,
Wells, U.S. Patent No. 4,913,237, April 3, 1990. Gas Research Institute, August 8, 1991.
Kvenvolden, K.A.: A Primer on Gas Hydrates: The Mavor, M.J., and Gunter, W.D.: Secondary Porosity
Future of Energy Gases, U.S. Geol. Survey Prof. and Permeability in Coal vs. Gas Composition and
Paper 1570, p.279-280, 1993. Pressure, paper SPE 90255, 2004.
Lamarre, R.A.: Downhole Geochemical Analysis of Mazunder, S., Hemert, P.V., Bruining, J., and Wolf,
Critical Desorption Pressure and Gas Content for K.H.A.A.: A Preliminary Numerical Model of
Carbonaceous Reservoirs, paper SPE 111091, CO2 Sequestration in Coal for Improved Coalbed
2007. Methane Production, unpublished, (2003).
Lancaster, D.E., McEtta, S.F., Hill R.E., Guidry, Mayerhofer, M.J., Richardson, M.F., Walker Jr., R.N.,
F.K., and Jochen, J.E.: Reservoir Evaluation, Meehan, D.N., Oehler, M.W., and Browning Jr.,
Completion Techniques, and Recent Results from R.R.: Proppants? We Dont Need No Proppants,
Barnett Shale Development in the Fort Worth paper SPE 38611, 1997.
Basin, paper SPE 24884, 1992. Mayerhofer, M., Stutz, L., Davis, E., and Wolhart,
Langmuir, I.: The Collected Works of Irving Langmuir S.: Optimizing Fracture Stimulation in a New
(Surface Phenomena), 9, Pergamon Press, New Coalbed Methane Reservoir in Wyoming Using
York, 1960. Tiltmeters and Integrated Fracture Modeling,
Lee, J., and Wattenbarger, R.A.: Gas Reservoir paper SPE 84490, 2003
Engineering, SPE, 1996. McGinnis, N., Scheppan, J., Popovich, G., and
Lehman, L.V., Shelly, B., Crumrine, T., Gusdorf, Rightmire, C.: Systems for Dewatering
M., and Tiffin, J.: Conductivity Enhancement: Unconventional Gas Wells, International Energy
Long-Term Results from the Use of Conductivity Conversion Engineering Conference paper IEEE
Enhancement material, paper SPE 82241, 2003. 829179, 1982.
Logan, T.L., Seccombe, J.C., and Jones, A.H.: Mohaghegh, S., and Ertekin, T.: Type-Curve Solution
Hydraulic Fracture Stimulation and Openhole for Coal Seam Degasification Wells Producing
Testing of a Deeply Buried Coal Seam in the Under Two-Phase Conditions, paper SPE 22673,
Piceance Basin, Colorado, paper SPE 15251, 1991.
Proc. Unconv. Gas Tech. Symp., p.501, Louisville, Moschovidis, Z.A., Cameron, J.R., and Palmer, I.D.:
Ky., May 1986. Methodology and Examples of Wellbore Stability
Malone, P.G., Briscoe, F.H., Camp, B.S., and Boyer, in Coalbed Methane Wells, paper 517, Intl.
C.M., II: Method of Calculating Coalbed Coalbed Methane Symp., Tuscaloosa, AL, 2005.
420
Mutalik, P.N., and Magness, W.D.: Production Data Tuscaloosa, AL.

Analysis of Horizontal CBM Wells in Arkoma Pekot, L.J., and Reeves, S.R.: Modeling the Effects
Basin, paper SPE 103206, 2006. of Matrix Shrinkage and Differential Swelling
Myers, R.R., Mack, D.J., Awny, J.A., and Creech, on Coalbed Methane Recovery and Carbon
G.: Field Application of an Alternative Foaming Sequestration, paper 0328, Proc. 2003 Coalbed
Agent for Coalbed Methane in the Nora Field in Methane Symposium, Tuscaloosa, AL.
Southwestern Virginia, paper SPE 57445, 1999. Penny, G.S., and Conway, M.W.: Coordinated Studies
NaturalGas.org, Overview Unconventional Natural in Support of Hydraulic Fracturing of Coalbed
Gas Resources, 2000 Methane, Gas Research Institute publication 95-
Nelson, C.R.: Effects of Coalbed Reservoir Property 0283, February, 1996.
Analysis on Gas-In-Place Estimates, paper SPE Peng, D.Y., and Robinson, D.B.: A New Two-Constant
57443, 1999. Equation of State, Ind. Eng. Chem. Fund. (1976),
Newfield Exploration Company, Press Release 15, 59.
(November 2006). Pinzon, C.L., and Patterson, J.: Production Analysis
Noynaert, S., Pumphrey, D., Pink, T., Eiden, T., of Coalbed Wells Using Analytical Transient
Hartensteiner, F., and Nelson, C.: Drilling for Solutions, paper SPE 91447, 2004.
Coalbed Methane in the San Juan Basin with Prosser, L.J., Finfinger, G.L., and Cervik, J.: Methane
Coiled Tubing: Results, Learnings and a World Drainage Study Using Underground Pipeline,
First, paper SPE/IADC 105874, 2007. Marianna Mine 58, US Bureau of Mines RI
Olsen, T.N., Brenize, G., and Frenzel, T.: Improvement #8577, 1981.
Process for Coalbed Natural Gas Completion and PTTC Workshop: The Lewis Shale, San Juan Basin:
Stimulation, paper SPE 84122, 2003. Approaches to Rocky Mountain Tight Shale Gas
Olsen, T.N., Bratton, T.R., Donald, A., Koepsell, R., Plays PTTCs Southwest Region, 2001
and Tanner, K.: Application of Indirect Fracturing Puri, R., Yee, D., Buxton, T.S., and Mahajan, O.:
for Efficient Stimulation of Coalbed Methane, Method of Increasing the Permeability of a Coal
paper SPE 107985, 2007. Seam, U.S. Patent No. 5,014,078, May 14,
Osisanya, S.O. and Schaffitzel, R.F.: A Review of 1991a.
Horizontal Drilling and Completion Techniques Puri, R., King, G.E., and Palmer, I.: Damage to Coal
for Recovery of Coalbed Methane, paper SPE Permeability During Hydraulic Fracturing, paper
37131, 1996. SPE 21813, 1991b.
Palmer, I., and Mansoori, J.: How Permeability Rahman, K., and Khaksar, A.: Implications of
Depends on Stress and Pore Pressure in Coalbeds: Geomechanical Analysis on the Success of
A New Model, SPEREE (December, 1998) 539. Hydraulic Fracturing: Lessons Learned from an
Palmer, I., Moschovidis, Z., and Cameron, J.: Coal Australian Coalbed Methane Gas Field, paper
Failure and Consequences for Coalbed Methane SPE 106276, 2007.
Wells, paper SPE 96872, 2005. Ramurthy, M., and Lyons, B.: Lessons Learned from
Palmer, I.D.: Review of Coalbed Methane Well Modeling Hydraulic Fracture Treatments in Coals
Stimulation, paper SPE 22395, 1992. Using a Fully Functional Three-Dimensional
Palmer, I.D., Davids, M.W., and Jeu, S.J.: Analysis Fracture Model in a San Juan Basin Project, paper
of Unconventional Behavior Observed During SPE 107972, 2007.
Coalbed Fracturing Treatments, paper SPE 8993, Reeves, S.: Enhanced Coalbed Methane Recovery,
1989. paper SPE 101466, SPE Distinguished Lecture
Palmer, I.D., Fryar, R.T., Tumino, K.A., and Puri, R.: Series, 2002-2003.
Comparison Between Gel-Fracture and Water- Richardson, J.S., Sparks, D.P., and Burkett, W.C.: The
fracture Stimulations in the Black Warrior Basin, TEAM Project Reserve Analysis: A Comprehensive
Proc. 1991 Coalbed Methane Symp., p.233, Evaluation to Predict Ultimate Recovery of Coalbed
421
Modern Fracturing
Methane, Proc. 1991 Coalbed Methane Symp., Reservoir Models for Coalbed Methane
University of Alabama, Tuscaloosa, AL, 293-306. Simulation, paper SPE 21599, 1990.
Roadifer, R.D., Moore, T.R., Raterman, K.T., Farnan, Seidle, J.P., Jeansonne, M.W., and Erickson, D.J.:
R.A., and Crabtree, B.J.: Coalbed Methane Application of Matchstick Geometry to Stress
Parametric Study: Whats Really Important to Dependant Permeability in Coals, paper SPE
Production and When? paper SPE 84425, 2003. 24361, 1992.
Robertson, E.P.: Measurement and Modeling of Sengul, M.: CO2 Sequestration A Safe Transition
Sorption-Induced Strain and Permeability Changes Technology, paper SPE 98617, 2006.
in Coal, Ph.D. Dissertation, Colorado School of Shi, J.-Q., and Durucan, S.: A Bidisperse Pore
Mines, Golden, Colorado (2005). Diffusion Model for Methane Displacement
Robertson, E.P., and Christiansen, R.L.: A Permeability Desorption in Coal by CO2 Injection, Fuel, 82
Model for Coal and Other Fractured, Sorptive- (10), 1219-1229, (2003a).
Elastic Media, paper SPE 104380, 2006. Shi, J.-Q., and Durucan, S.: A Model for Changes
Rodvelt, G.D., and Oestreich, R.G.: Case History: in Coalbed Permeability During Primary and
First Commercial Illinois Coalbed Methane Project Enhanced Methane Recovery, SPEREE (August,
Commences Through a Structured Resource 2005) 291.
Evaluation Plan, paper SPE 97720, 2005. Shi, J.-Q., and Durucan, S.: Gas Storage and Flow
Ruppel, T.C., Grein, C.T., and Bienstock, D.: in Coalbed Reservoirs: Implementation of a
Adsorption of Methane on Dry Coal at Elevated Bidisperse Pore Model for Gas Diffusion in Coal
Pressure, Fuel, (July, 1974) 152. Matrix, paper SPE 84342, 2003b.
Ruckenstein, E., Voidyanthan, A.S., and Youngquist, Shirley,K.: Lewis Not Overlooked Anymore, AAPG
G.R.: Sorption by Solid Bidisperse Pore Structure, Explorer (March 2001)
Chem. Eng. Sci., 26, 1305-1318, (1971). Siemons, N., Busch, A., Bruining, H., and Kroos,
Ruthven, D.M.: Principles of Adsorption ans B.: Assessing the Kinetics and Capacity of Gas
Adsorption Processes, John Wiley & Sons, Inc. Adsorption in Coals by a Combined Adsorption/
New York (1984). Diffusion Method, paper SPE 84340, 2003.
Sawyer, W.K., M.D. Zuber, V.A. Kuuskraa and D.M. Simpson, D.A., Lea, J.F., and Cox, J.C.: Coal Bed
Horner: Using Reservoir Simulation and Field Methane Production, paper SPE 80900, 2000.
Data to Define Mechanisms Controlling Coalbed Simons, L.H.:Elastic Properties of Shales, Drilling
Methane Well Production, paper no. 8763, and Rock Mechanics Symposium, paper 496-MS,
1987. 1963
Schein, G.W., Carr, P.D., Canan, P.A., and Richey, Snider, P.M., Walton, I.C., Skinner, T.K., Atwood,
R.: Ultra Lightweight Proppants: Their Use and D.C., Grove, B.M., and Graham, C.: First
Application in the Barnett Shale, paper SPE Laboratory Perforating Tests in Coal Show Lower-
90838, 2004. Than-Expected Penetration, paper SPE 102309,
Schein, G.W.: The Application and Tehnology of 2006.
Slickwater Fracturing, SPE Distinguished Lecture Southwestern Energy, Press Release (August 2004).
Series, 2004-2005. Southwestern Energy, Operational Update (October
Scott, R.A.: Preliminary Assessment of Worldwide 2006).
Coalbed Methane Resources, Proc. AAPG Annual Spafford, S.: Rock Creek Methane From Multiple
Meeting, 2004. Coal Seams Completion Project, 1991 Annual
Seidle, J.P.: A Modified P/Z Method for Coal Wells, Report, presented at the Gas Research Institute.
paper SPE 55605, 1999. Stevens, S.H., Spector, D., and Riemer, P.: Enhanced
Seidle, J.P.: Long-Term Gas Deliverability of a Coalbed Methane Recovery Using CO2 Injection:
Dewatered Coal, JPT, (June 1993) p.564. Worldwide Resource and CO2 Sequestration
Seidle, J.P., and Aril, L.E.: Use of Conventional Potential, paper SPE 48881, 1998.
422
Stevenson, M.D.: Multi-component Gas Adsorption Butterworth Publishers, Johannesburg, South

on Coal at In-Situ Conditions, Petroleum Africa (1987).
Engineering, University of New South Wales, Yang, R.T. and Saunders, J.T.: Adsorption of Gases
Sydney (1997). on Coal and Heat-Treated Coals at Elevated-
Stutz, H.L., Victor, D.J., Fisher, M.K., Griffin, L.G., Temperature and Pressure. 1. Adsorption from
and Weijers, L.: Calibrating Coal Bed Methane Hydrogen and Methane as Single Gases, Fuel
Fracturing Geometry in the Helper Utah Field (1985) 64, No. 5, 616.
Using Treatment Well Tiltmeters, paper SPE Zhu, J., Jessen, K., Kovscek, A.R., and Orr, F.M.:
77443, 2002 Analytical Theory of Coalbed Methane Recovery
Suarez-Rivera, R., Green, S.J., McLennan, J., and Bai, by Gas Injection, paper SPE/DOE 87336,
M.: Effect of Layered Heterogeneity on Fracture 2002a.
Initiation in Tight Gas Shales, paper SPE 103327, Zhu, J., Jessen, K., Kovscek, A.R., and Orr, Jr., F.M.:
2006. Recovery of Coalbed Methane by Gas Injection,
Tang, G.-Q., Jessen, K., and Kovscek, A.R.: Laboratory paper SPE 75255, 2002b.
and Simulation Investigation of Enhanced Coalbed Zuber, M.D., and Semmelbeck, M.E.: Simulators
Methane Recovery by Gas Injection, paper SPE Analyze Coalbed Methane, The American Oil and
95947, 2005. Gas Reporter (September, 1990), 40-44.
Valencia, K.J., Chen, Z., Hodge, M.O., and Rahman, Zuber, M.D., and Olszewski, A.J.: The Impact of
S.S.: Optimizing Stimulation of Coalbed Methane Errors in Measurements of Coalbed Methane
Reservoirs Using an Integrated Fracture Modeling Reservoir Properties on Well Production Forecasts,
Approach, paper SPE 93245, 2005. paper SPE 24908, 1992.
Walker, R.N. Jr., Hunter, J.L., Brake, A.C., Fagin, P.A.,
and Steinsberger, N.P.: Proppants, We still Dont
Need No Proppants A Perspective of Several
Operators, paper SPE 49106, 1998.
Wei, X.R., Wang, G.X., and Massarotto, P.: A Review
on Recent Advances in the Numerical Simulation
for Coalbed Methane Recovery Process, paper
SPE 93101, 2005.
Williams, Peggy: New Shale-Gas Play Unfolding,
Shale Gas, a supplement to Oil and Gas Investor,
(January 2006) 18.
Wolhart, S.L., Stutz, H.L., and Mayerhofer, M.:
Applying Advanced Fracture Diagnostics to
Optimize Fracture Stimulation in Coalbed
Methane Reservoirs: Case History of Two Field in
the Rocky Mountains, paper SPE 91376, 2004
Wong, S., Gunter, W.D., and Mavor, M.J.: Economics
of CO2 Sequestration in Coalbed Methane
Reservoirs, paper SPE 59785, 2000.
www.eia.doe.gov, 2007
Youngquist, W., and Duncan, R.C.: North American
Natural Gas: Data Show Supply Problems, Natural
Resources Research, 12:4 (December 1993), p229-
240.
Yang, R.T.: Gas Separation by Adsorption Processes,
423
Modern Fracturing
424
books, including The Color Of Oil, and more than 200 journal papers and articles. Economides does a
wide range of industrial consulting, including major retainers by national oil companies at the country level
and by Fortune 500 companies. He has had professional activities in over 70 countries. .
Randy LaFollette is the manager of geoscience technology for BJ Services Company in the Corporate
Technology Center in Tomball, Texas. LaFollette has a BS in geological science from Lehigh University. He
has 30 years of experience in the oil industry including field, region and research positions. His previous
employers have included the Western Co. and Reservoirs Inc. He is active in SPE, aiding with Advanced
Technology Workshop organization and presenting at some ATWs. Current job responsibilities include
coordinating laboratory and applied support projects; structuring and implementing geospatial studies
of stimulation effectiveness, linking reservoir quality and treatments performed to production results;
providing stimulation and remedial treatment recommendations; training personnel; and consulting with
operators and BJ field/region personnel on technical support matters.
Dr. Don Wolcott is president of E&P for Aurora Oil & Gas Co, an independent production company
headquartered in Moscow. Before working for Aurora, Wolcott worked for various companies including
YUKOS Oil Co., ARCO Oil & Gas Co., and Schlumberger Inc. While working at YUKOS, Wolcott played
a pivotal role in its becoming the top oil producing company in Russia, the lowest-cost producer and one
of the top reserve companies in the world. Wolcott has a PhD in petroleum engineering from the Colorado
School of Mines, and BS and MS degrees in petroleum engineering from the University of Wyoming.
Andronikos Demarchos is a senior engineer on the stimulation and completions team with the Hess
Corporation. Previously, he worked for eight years with Economides Consultants on projects around the
world, primarily in hydraulic fracture design, analysis and optimization, and for three years as a production/
operations engineer with Union Pacific Resources. He has authored several professional papers. Demarchos
holds an MS in petroleum engineering from Texas A&M University and a BS in electrical engineering from
the University of Houston.
Chapter 12 otherwise there would be radial flow. The duration of
linear formation flow will increase as the stimulation
Fracturing for level increases. At the limit, only linear flow is seen
Reservoir Development for a fully penetrating fracture in pseudosteady-state
Michael J. Economides, University of Houston, well drainage, with adequate fracture conductivity.
Randy F. LaFollette, BJ Services The well will have no radial flow period. In
Don Wolcott, Aurora Oil & Gas short, as the level of stimulation increases, so does
and Andronikos Demarchos, Hess the duration and magnitude of linear flow with a
corresponding decrease in radial flow.
Because well stimulation increases the length and
12-1 Introduction
duration of linear flow, optimal well placement during
development planning should look to accommodate
Clearly hydraulic fracturing has been established and enhance linear flow. The optimization criteria for a
already as perhaps the most compelling and attractive given development project is to deliver the maximum
single well production and injection enhancement production and recovery with the minimum number of
technique. As such, it should be difficult to decide wells. With longer fractures providing linear flow, the
against hydraulic fracturing in a well. However a optimization of rate and recovery leads the development
number of formation evaluation considerations to rectangular drainage shapes. There is an optimal
such as the location and distance of water and aspect ratio for the rectangular drainage shape, which is
hydrocarbon contacts, layering, barriers and others introduced later in this chapter.
must be accounted for and may, in fact, limit The shape of fractures forces the creation of
achievable production results. elongated drainage areas, and the question arises
Even more interesting are reservoir engineering regarding well spacing and, again, positioning.
issues, such as the deployment of fractured wells for Obviously, knowledge of the would-be fracture
reservoir exploitation strategies. What is the impact of azimuth is quite important and, indeed, crucial,
fractured wells in reservoir depletion and oil and gas in infield drilling for production and planning
recovery? How do they compare with unfractured wells? the positioning of wells.
There have been a number of recent important papers Important exercises in any reservoir strategy are
that address the issue of fracturing, optimizing the the evaluation of past performance and past jobs,
treatments and maximizing their impact on production the identification of gaps and the remediation of
enhancement (Romero et al., 2003; Economides et problems. A new, powerful approach is the use of data
al., 2004; Demarchos et al., 2004; Rueda et al., 2004; mining. Data mining is not new as an approach to
Demarchos et al., 2005; Rueda et al., 2005; Sabaev et al., identify patterns from large numbers of influencing
2006; Vakhitova, 2006; and Baranov et al., 2007). variables; only the types of questions it can answer
This chapter will focus on well positioning, have changed. In the 1960s, computing power was
configuration and spacing. How should fractured scarce and expensive; thus, data mining answered
wells be configured? Given a certain drainage area questions such as what was the total production by a
for the reservoir, how should it be divided among field for the last three years? In the 1980s the question
fractured wells? Unquestionably, it should not be the had evolved to what was the total production in the
same as for unfractured wells with radial flow. southeast area of the field in the second quarter of
Fracturing increases well potential and, in some 1983? In the 1990s, the question was what is the
cases, recovery when compared to an unfractured well. total production in the south east section of the field
The orientation of a fractured well in relation to other in the second quarter with water-cut above 50% and
wells plays a significant role in optimizing rate and drilled to the depth of well #15? Today with vast
recovery. When a well is hydraulically fractured, there computer power, geographical information systems
will be an introduction of linear formation flow, where (GIS), and artificial neural networks, data mining can
427
Modern Fracturing
answer more complex questions such as, why is the The orientation of a fracture in a well in relation
water cut in well 15 above 15%? Or, more relevant to those of other fractured wells plays a significant
to fracturing, what factors contribute to successful role in optimizing rate and recovery. When a well
fracturing treatments in a formation, a field or a is stimulated, there will be an introduction of linear
section of the field? or What went wrong, what went formation flow, where otherwise there would be radial
right and what things could be improved? flow for an un-stimulated well. The duration of linear
formation flow will increase as the stimulation level
12-2 Impact of Fracturing on increases. At the limit, only formation linear flow is seen
Reservoir- or Drainage-Wide for a fully penetrating fracture (i.e., one whose length,
Production tip-to-tip, extends to the drainage boundaries) with
adequate fracture conductivity under pseudosteady
In Chapter 2 of this book the performance of a state (closed boundaries). In such a case, the well will
hydraulically fractured well was introduced through the have no radial flow. This was shown in Chapter 2: The
use of the dimensionless productivity index, JD. Also, the theoretical limit of JD in a fully penetrating fracture
concept of the Unified Fracture Design was presented and fully developed linear flow in a closed square is
as a means to physically optimize well performance 6/. In summary, as the level of stimulation increases
(Economides et al., 2002a and 2002b). The design so does the duration and magnitude of linear flow
procedure was expounded upon in Sections 4-5 and with a corresponding decrease in radial flow.
4-6 using the Proppant Number as the catalyst in The shape of a hydraulic fracture lends itself
defining both the maximum potential JD and the to rectangular drainage area in a fully developed
optimum fracture conductivity, CfD,opt , whose value leads reservoir exploitation strategy. The first important
to the desired fracture dimensions. issue is to recognize that multiple fracturing
The maximum JD was presented in Chapter 2 both treatments will no longer drain square drainage
in graphical form and with convenient correlations: areas, but elongated ones, whose aspect ratio depends
where Nprop is defined as in Eq. 2-32, Section 2-5. (See on the number of infield wells.
Equation 2-33 bottom of page). The Proppant Number of Eq. 2-32:
Similarly, correlations were presented for
2
4k f x f w 4k f x f whp 2k f V p
the optimal dimensionless fracture conductivity N prop = I x C fD =
= = (2-32)
kxe2 kxe2 hp kVr
for the entire range of Proppant Numbers: (See
Equation 2-34 bottom of page). must be adjusted first by
The drainage area and drainage shape are of major
4k f whx f x f xe xe
importance in fractured well performance. The UFD N prop = = I x2C fD , (10-1)
kxe ye h x f xe ye
design procedure is based on pseudosteady state, and
the equations above are for such flow and also for where xe and ye are the dimensions of the drainage,
wells depleting square drainage areas. parallel and perpendicular to the fracture.
1
if N prop 0.1
0.990 0.5 ln N prop
J D max ( N prop ) = (2-33)
6 0.423 0.311N prop 0.089( N prop ) 2
exp 2

if N prop > 0.1
1 + 0 . 667 N prop + 0. 015( N prop )

1.6 if N prop < 0.1

0.583 + 1.48 ln N prop

C fD ,opt ( N prop ) 1.6 + exp
=
1 + 0.1142 ln N

if 0.1 N prop 10 (2-34)

prop

N prop

if Nprop >10
428
Chapter 12 Fracturing for Reservoir Development
Daal and Economides (2006) showed that for larger Proppant Numbers (i.e., as the treatments
small Proppant Numbers (Nprop < 0.1) an equivalent increase and their lengths increase). In these cases,
Proppant Number can be defined: an elongated drainage aspect ratio becomes quite
CA
attractive exactly because of the evolution of the
N prop ,e = N prop 30.88 , (10-2) far more desirable linear flow in the drainage. As
the treatments become larger, a square drainage
where CA is the Dietz shape factor (as given in Fig. shape is no longer desirable.
10-2) and the calculated Nprop,e can then be used
with the normal correlations for both maximum 12-2.1 Example Application of Infield
JD and optimum CfD. For Proppant Numbers Drilling and Fracturing of Gas Wells
larger than 0.1, Daal and Economides (2006)
presented a semi-analytical correlation, which Start with a large drainage area of 320 acres in a
is also shown in Section 10-2. reservoir where k = 1 md, with all other important
The benefit is an overall increase in well potential reservoir and fracture treatment data shown in Table
per unit area, with fewer wellbores. The total JD would 12-1. The sizing of the fracture follows the iterative
be the sum of individual JDs. procedure outlined in Sections 4-5 and 4-6, with the
proppant pack permeability reduced because of the
10
y e = xe 2ye = xe non-Darcy flow in the fracture (see Section 8-6). Table
4ye = xe 5ye = xe
10ye = xe 12-2 presents the results, first for one well drilled in
the entire drainage and fractured. The maximum JD for
one well is 0.393, leading to an initial pseudosteady-
state production rate of 15 MMscf/d.
JD, max
1
Table 12-1 Reservoir and Treatment Data for Single,
Two and Four Fractured Wells in a Given Drainage
INPUT DATA
Mass of proppant, lbs 200,000

0.1 Specific Gravity of proppant 2.65
0.1 1 10 100
Np Porosity of proppant Pack 0.38
Proppant Pack Permeability, md 150,000
Figure 12-1 Maximum JD as a function of the Proppant
Formation Permeability, md 1
Number and Drainage Aspect Ratio (xe is the drainage
side parallel to the fracture and ye is the drainage side Reservoir thickness, ft 50
perpendicular to the fracture) Total Drainage Area, acres 320
Well raidus, ft 0.4
Reservoir Pressure, psi 3,300
Figure 12-1 presents JD,max vs. Nprop for Bottom-Hole Flowing Pressure, psi 1,500
several aspect ratios of the drainage areas (Daal Gas Specific Gravity 0.71
and Economides, 2006). This graph allows
an interesting look at the behavior of fracture Table 12-2 Examples of Single, Two and Four Fractured
penetration vs. drainage aspect ratio. At small Wells in a Given Drainage
Proppant Numbers (i.e., small treatments relative Number
Total Rate q,
Total
of JD,max Rate q, ye/xe Nprop CfD,opt xf , ft
to the drainage area and, thus, small penetrations fractures
JD,max MMscf/d
MMscf/d
relative to the drainage dimensions) the
productivity index is reduced as the drainage becomes 1 0.393 0.393 15,058 15,058 1 0.0440 1.600 310
elongated because on either side of the fracture it

2 0.418 0.836 16,028 32,057 0.5 0.0847 1.600 304
becomes increasingly difficult to drain the reservoir.
Square drainage is the best for production in 4 0.375 1.499 14,356 57,425 0.25 0.1737 1.392 330
such cases. But an inverse behavior happens at
429
Modern Fracturing
Cutting the drainage in half (160 acres, where, tpss is the time until pseudosteady-state in
aspect ratio 2:1), means that now the two wells are hours. Equation 12-1 is for a non-fractured well
fractured in two adjoining and parallel drainage in the center of a square drainage area and should
areas, leading to individual well JDs equal to represent the upper limit of time to pseudosteady
0.418 for a composite of 0.836 and a total initial state. Fractured wells will enter pseudosteady state
pseudosteady-state production rate of 32 MMscf/d. more quickly than unfractured wells, especially if the
Finally, four wells, splitting the total drainage area in drainage shape is not square but instead rectangular,
four parallel quarters (80 acres, aspect ratio 4:1) lead with the boundaries parallel to the fracture coming
to composite initial pseudosteady-state production ever closer because of infield drilling.
rate equal to 57.4 MMscf/d. (Note: In this Using the variables for the study in Section
example the fractures are, of course, parallel to 12-2.1 and porosity equal to 20%, a drainage area
each other. Simply the total drainage is split from of 320 acres (re = 2100 ft) and a permeability of 1
top to bottom, creating first a 1:1, then two 2:1 md, the time to pseudosteady state would be a
and then four 4:1 drainages. In all cases the total maximum of 230 days. Halving the drainage area
area remains the same: 320 acres.) would result in roughly halving the time and, for
There are two important points that need to be the four fractured wells, halving the time again to
made here. First, as will be shown in Section 12-3, less than 2 months. Forecast of well performance
the goal of infield drilling is to accelerate recovery, beyond that point would then be done under
and the four-well configuration will lead to a much pseudosteady-state conditions.
faster rate of production decline than the two- These times would increase by a factor of 10 if the
well or single-well option. Second, a more subtle reservoir permeability is reduced by a factor of 10 to
observation is necessary. Table 12-2 shows a superficial 0.1 md, as can be surmised by Eq. 12-1.
abnormality. The individual JD of two fractures goes
Production Rate Under Transient Conditions
up compared with that of one fracture, and then 30
Transient Production Rate, MMscf/d
the JD of each of four fractures goes down. This is

not an error. The relationship between the distances 25
of the parallel and perpendicular boundaries vis a
vis the fracture length and the ultimate evolution 20
of full linear flow are complex. The issue was
addressed in detail in Daal and Economides
15
(2006) and is shown in Fig. 12-1.
10
12-2.2 Transient Flow of Fractured Gas Wells 0 1 2 3 4 5 6 7 8
Time, months
Before pseudo-steady conditions ensue, the well flow Figure 12-2 Transient rate decline for a single fractured
well in the example in Section 12-2
rate is under transient conditions. Chapter 3 of this
book presents all important flow regimes, including Figure 12-2 shows transient production rates for
bi-linear, linear and pseudo-radial. They are important the single fractured well in the example presented
not only in well-test analysis but also for forecasting in Section 12-2.1, showing a large rate decline from
well performance until boundary-dominated flow an average of 26 MMscf/d in the first month to 15
emerges. The onset of pseudosteady state can be MMscf/d during the eighth month, coinciding with
forecast through numerical simulation, but there are the onset of pseudosteady state. This results in the
also good analytical approximations, one of which is production of about 4 Bcf of gas.
from Economides et al., 1994: Using A = 320 acres, h = 50 ft, and Sw = 0.2 at
the initial reservoir pressure of 4000 psi and Z = 0.85,
ct re 2
t
pss =1200 , (12-1) the initial formation volume factor, Bg is 0.00349 (see
k
430
Section 2-3.3, Eq. 2-18) resulting in an initial gas-in- 0.85 and 0.78, respectively and from Eq. 2-
place of about 32 Bcf. At the end of transient conditions 49, Gp = 17.15 Bcf, leading to a recovery of
about 12.5% of original gas is produced, resulting in a about 54% at abandonment.
pressure decline to about 3300 psi. For each of the draining configurations in Fig.
In the next section, we will present a more 12-3 there will be two periods: a transient stage and a
appropriate material balance-based forecast under pseudosteady-state stage.
pseudosteady-state conditions.
12-3 Forecasting Natural Gas

Well Performance and Recovery
In Chapter 2, the natural gas material balance
equation was given as:
Bgi
G p = Gi G = Gi Gi , (2-48)
Bg
where Bgi and Bg are the corresponding formation

volume factors. Substitution of those expressions
results in:
p / Z
G p = Gi 1
. (2-49)
pi / Z i Figure 12-3 Drainage configurations for drainage and
recovery case study
With Eq. 2-49, for any new average reservoir pressure
the corresponding cumulative recovery, Gp, can be Figure 12-4 is a forecast of total production
readily calculated. rates for the four configurations showing the very
large early-time production from the four fractured
12-3.1 A Case Study for Reservoir Recovery wells. Eventually, as should be expected, the rates
Using Unfractured and Fractured Wells may flip-flop, but the true impact is shown in Fig.
12-5, where the cumulative recoveries are shown.
A case study is presented below using the expected Fractured wells provide production enhancement
productions from the reservoir described in Table 12-1 plus accelerated recovery, providing a huge boost to
and the configurations shown in Fig. 12-3, representing project economics and net present value. In this case
respectively, one and four unfractured wells and one study, the time to ultimate recovery from the four
and four fractured wells, partitioning the drainage fractured wells will be about 2.4 years, whereas the
accordingly. Total initial gas-in-place is given by same recovery would take more than 30 years from
a single unfractured well. (For the four unfractured
Ah(1 S w )
Gi = ,
(12-2) wells, the time would be 7 years and for the single
Bgi
fractured well would be about 10 years.)
and with A = 320 acres (43,560 ft2), h = 50 ft, = Tables 12-3 and 12-4 contain the actual rates
20%, Sw = 0.2 and Bgi = 0.00349 resft3/scf (at and cumulative recoveries for the four fractured
4,000 psi, calculated from property correlations and unfractured well configurations, respectively.
presented in Chapter 2), then the initial gas-in-place The results on these tables show the enormous
from Eq. 12-2, Gi = 32 Bcf. impact of fractured wells on accelerated production.
An abandonment pressure of 1700 psi is For example, after about three months, the four
assumed in this case study. The values of the gas fractured wells would deliver about 45 MMscf/d
deviation factor, Z, at 4000 psi and 1700 psi are whereas the four unfractured wells would produce
431
Modern Fracturing
about 20 MMscf/d. The four fractured wells would Table 12-4 Production Rate and Cumulative
decline to this rate about six months later (nine Production Forecast for Four Unfractured Wells in a
Quartered Squared Reservoir
months from the start.)
P, Gp, Gp, q, t, t,
psi Bscf Bscf MMscf days days
80
4000 0 0 26.75 0 0
70 1 Fractured Well
4 Fractured Wells 3900 0.43 0.43 25.28 17 17
60 1 Unfractured Well
4 Unfractured Wells
Transient to PSS 3700 * 1.68 1.27 22.42 72 55
q, MMscf/d
50
Transient to PSS
40 Transient to PSS 3500 2.61 0.91 20.36 116 44
Transient to PSS
30 3300 3.59 0.98 18.33 169 53
20 3100 4.97 1.38 16.18 250 81
10
3000 5.84 0.87 15.10 308 58
0
0 365 730 1,095 1,460 1,825 2,190 2,555 2,920 3,285 2800 7.26 1.42 12.98 414 106
Time, d
2600 8.72 1.46 10.90 544 130
Figure 12-4 Production rate forecast for drainage in case 2400 10.51 1.79 8.79 736 192
study
2200 12.30 1.79 6.70 985 249
2000 14.08 1.78 4.70 1332 347
18 1800 16.08 2.00 2.73 1957 625
15 1700 17.15 1.07 1.78 2560 603
12 *End of the transient period
Gp , Bscf
1 Fractured Well
9 4 Fractured Wells
1 Unfractured Well
4 Unfractured Wells
6 Transient to PSS
Transient to PSS
12-3.2 Field Development Strategy
3 Transient to PSS
Transient to PSS
0 Sabaev et al. (2006) have shown that a given Nprop has
0
0
60
90
20
50
80
10
40
70
00
30
60
90
0
0
an optimum drainage aspect ratio that can provide

73
,22
,95
1,4
2,1
2,9
3,6
4,3
5,1
5,8
6,5
7,3
8,0
8,7
9,4
10
10
time, d
the largest cumulative JD for a given development
Figure 12-5 Cumulative recovery forecast for drainage in area. In the same logic, for any given well drainage
case study
aspect ratio and Nprop , there is an optimal CfD and
Table 12-3 Production Rate and Cumulative Production
and a penetration ratio, Ix , at which JD is maximized.
Forecast for Four Fractured Wells in a Sliced, This means that that for the same proppant volume,
Squared Reservoir Vf, a rectangular drainage area gives a higher JD
P, Gp, Gp, q, t, t, than a square drainage area.
psi Bscf Bscf MMscf days days
As shown in Chapter 2 the JD limit for a well in
4000 0 0 76.44 0 0
a square drainage shape is 6/. Sabaev et al. (2006)
3800 0.87 0.87 70.09 12 12
3600 2.14 1.27 63.33 32 20
showed that the maximum JD in a rectangular drainage
3400 3.10 0.96 57.14 48 16
area under linear flow is approximated by:
3200 4.45 1.35 50.62 74 26

6 ye
3050 * 5.40 0.95 45.92 95 21
J Dmax = . (12-3)
xe
3000 5.84 0.44 44.30 105 10
2800 7.26 1.42 37.92 141 36 Because rectangular geometry provides larger
2600 8.72 1.46 31.75 185 44 JD, it is obvious to consider this geometry
2400 10.51 1.79 25.49 252 67 for field development. Sabaev et al. (2006)
2200 12.30 1.79 19.37 338 86 presented type curves showing the maximum
2000 14.08 1.78 13.53 458 120 possible JD and optimum CfD for a given
1800 16.08 2.00 7.83 676 218 rectangular aspect ratio (rectangle length to
1700 17.15 1.07 5.11 886 210 width ratio Ar=ye/xe). They further provided a
*End of the transient period methodology to determine the optimum aspect
432
6
Np=1000
3 500 1
Xf Xf JD, max=3*6/Q=5.73
ye 1 0.9
5 100
xe
50 0.8
2x f
4 Ix
xe
Penetration Ratio, Ix
kf w 0.7
3
CfD 20
kx f
N prop I x2* Cf D * Ar 10 0.6

2
5 0.5
3 0.4
1 0.3
1
0.5 0.2
0 0.1
0
0.1 1 10 100 1000
Figure 12-6 Pseudosteady-state fracture performance type curve (Ar = 3) (courtesy of Yukos)
ratio for a given set of reservoir conditions. They with fewer wells as shown by comparing Fig. 12-7 a
constructed type curves for various aspect ratios of and c. The development with Ar = 2 is not as clear a
2, 3, 5, and 10. The type curve for aspect ratio benefit, requiring economic justification because the
of 3 is shown in Fig. 12-6. cumulative JD is less by 30%, but the development
requires half as many wells (Fig. 12-7 b).
Ar= 1 Ar= 2 Ar= 3 Sabaev et al. (2006) presented a correlation
relating the drainage aspect ratio and the
required Nprop to maximize the drainage JD (see
Fig. 12-8). For each value of Nprop in Fig. 12-8
the curve cutting through the family of curves
represents both the maximum JD and the optimum
fracture conductivity.
The Proppant Number for rectangular drainage
Figure 12-7 Different development strategies for a given
reservoir area shapes is defined by:

2
Using rectangular-shaped patterns with aspect N

prop = C I
fD x rA . (12-4)
ratio greater than one, in field development provides
proportionally larger potential for a given JD. Due to From Fig. 12-8 for an aspect ratio of 3 it is clear
the larger well cumulative JD, for the same Nprop, a that the required Proppant Number should be
larger well spacing can be used. Larger well spacing about 10. The question is whether such Proppant
leads to a decrease in the total number of wells in Number can be placed, which goes back to
the field development and will further improve the concept of pushing the limits. For low-
investment efficiency. For the same development area permeability reservoirs, such Proppant Numbers are
the Ar = 3 development has a higher cumulative JD not only desirable but also feasible.
433
Modern Fracturing
Nprop=100
6
5
50
40
4
30
20
3
10
2
5
2
1
1
0.5
0.1
0
0 1 2 3 4 5 6 7 8 9 10
Aspect Ratio, Ar
Figure 12-8 Optimum aspect ratio for pseudosteady-state production (from Sabaev et al., 2006)
12-4 Impact of Fracture Of considerable interest in reservoir exploitation

Azimuth on Well Planning strategy are the fracture azimuth (because of the drainage
patterns), the positioning of wells and, especially, the
Hydraulic fractures generally have a well- partitioning of a given drainage either by the drilling of
defined orientation (vertical, comprising the infield wells in ever-closer well spacing or the fracturing
overwhelming fraction of fractures in petroleum of horizontal wells with multiple treatments. The
formations vs. horizontal) and azimuth. What performance of fractures in relation to their drainage
controls these features of the fracture is the size and shape was covered in Chapter 2.
state of stress (see Section 4-3.2). There are three issues related to stress and stress
In any geologic structure there are three identifiable magnitude that a fracturing engineer must be aware of,
principal stresses: one vertical and two horizontal. The especially in dealing with exploiting a field:
latter are distinguished as the minimum and maximum What are the magnitudes of the stresses in both the
horizontal stress. A fracture will propagate toward target and neighboring layers?
the path of least resistance, i.e., perpendicular to the What is the extent of stress anisotropy?
minimum stress. Because in the case of the vast majority What are the orientations of the stresses, and
of oil and gas formations the smallest of the three stresses consequently what would the direction and
is one of the two horizontal stresses, a hydraulic fracture azimuth of the fracture be?
is typically vertical. In shallow formations, a horizontal
fracture may be created, but this should be considered The answer to the first question is important for
the exception. There are also some aberrant formations a number of reasons. First, the fracture geometry,
where horizontal fractures may be created at depth, such as the width, greatly depends on the value of
or a T-shaped fracture may appear. A comprehensive the minimum stress. Also, the net pressure during
discussion on these subjects can be found in Chapter 3 fracturing and the fracture geometry are strongly
of Economides and Nolte (2000). Complex hydraulic dependent on these values. This subject has been
fractures, such as those documented in the Barnett covered extensively in Chapter 4, including the
Shale, are yet another fracture type. techniques for measuring minimum stress.
434
The second and third questions are what control More sophisticated techniques involve analysis
the fracture orientation and azimuth. The extent of of oriented cores. These techniques include Anelastic
stress anisotropy (i.e., how much smaller the minimum Strain Recovery, ASR (Teufel, 1983) and Differential
stress is compared to the intermediate stress) is critical Strain Curve Analysis, DSCA (Siegfried and Simmons,
in avoiding complex fracture geometry such as T-shaped 1978; and Strickland and Ren, 1980). Both of these
fractures. In fracturing, stress anisotropy is a desirable techniques provide the orientation of the stresses
thing. When the stresses approach each other, significant (minimum, maximum and intermediate) and their
complications can arise in fracture propagation. Thus, relative magnitudes. They provide a definitive verdict
measuring the stresses and especially stress anisotropy is on fracture azimuth. For a detailed description of all
important and will be covered below. techniques, see Economides and Nolte (2000).
12-4.1 Determination of Fracture Azimuth 12-4.2 Considerations Regarding Directional

Permeability in the Reservoir
Assuming that the vertical stress can be readily
calculated (see Eqs. 4-47 and 4-48, Section 4-3.2) The state of stress in a reservoir, in addition to
and that the fracture is expected to be vertical, to the impact that it has on hydraulic fracture direction
determine definitively the fracture azimuth, one of (vertical vs. horizontal) and azimuth, also greatly
the first steps is to obtain the value of the minimum affects reservoir permeability. Stress anisotropy
stress, at times known as the closure pressure. breeds permeability anisotropy, especially when
Techniques such as the pump-in/flowback test permeability is dominated by natural fissures or
(Nolte, 1979) or the step-rate test have been devised fractures (Buchsteiner et al., 1993).
to obtain the value of the closure pressure, which for In-situ stresses, being compressive in nature, tend
a single-layer formation is the value of the minimum to reduce the width of natural fissures. Therefore,
horizontal stress. Some attempts have been also made fissures perpendicular to the maximum horizontal
to estimate the value of the maximum horizontal stress stress are likely to be significantly narrower than fissures
from the fracture initiation (breakdown) pressure, perpendicular to the minimum stress. This would suggest
but these attempts have been hampered because the that the permeability in the direction of maximum stress
breakdown pressure depends on the elastic response would be larger than the permeability in the direction
of the formation and fluid leakoff into the formation of minimum (horizontal) stress. The larger the stress
before fracturing, and is also heavily influenced anisotropy, the more heterogeneous the distribution of
by the geometry and orientation of the wellbore natural fissures will be. Coalescing and jointed fissures
(Detournay and Cheng, 1992). form natural fractures, whose width is further enhanced
Still, knowledge of these two values does not by being normal to the maximum stress. The width of
provide their azimuth. these natural fractures is greater if stress anisotropy is
The most obvious techniques for determining this more pronounced (Elkins and Skov, 1960; Finley and
involve geological mapping. Large natural fractures Lorenz, 1989; and Lorenz et al., 1986).
and faults generally would be parallel to hydraulic Actual measurement of permeability anisotropy in
fractures except in those circumstances where stresses the field is a cumbersome and costly process involving
have rotated from the time of creation of the faults multi-well interference testing (see Earlougher, 1977).
and natural fractures to present day. Formation micro- A far more convenient technique has evolved since the
resistivity logs can provide the orientation of natural introduction of horizontal wells. Well testing of single
fractures, another indicator of fracture azimuth. An horizontal wells and the identification of early-time
oriented multi-arm caliper log, providing wellbore radial, middle-time linear and late-time pseudoradial
eccentricity, may also provide indications of stress flow regimes allows for the calculation of all three
orientation: The smaller axis of the wellbore ellipse principal permeabilities two horizontal and the
will be parallel to the maximum stress. vertical (see Economides et al., 1994).
435
Modern Fracturing
There are some direct implications for hydraulic 10:1 would result in a well-test-derived apparent
fracturing from this notion. First, vertical hydraulic length that would be less than one-third of the
fractures, normal to the minimum horizontal stress real length. Knowledge of permeability anisotropy
are also normal to the lowest horizontal permeability can reconcile this disparity.
in the reservoir. Second, the frequently repeated Walsh (1981) presented a model for the relationship
statement that hydraulic fractures are executed in hopes between initial (kfi) and some later value of fissure
of intersecting natural fractures is almost always false. permeability (kf) and the effective stress, :
Pronounced and conductive natural fractures are certain
3
kf 2h ' 1 b( ' *)
to be parallel to a hydraulic fracture. Of course, in a = 1 ln , (12-6)
k fi * 1 + b( ' *)
naturally fissured reservoir, even if there is a preferred
orientation, the fissures are in every direction, and where, is the fissure half-aperture, h is the root-mean-
during execution large leakoff is likely to be evident. square value of asperity (fissure) height distribution, b is
But during production, permeability anisotropy an experimentally determined constant and * is some
would be felt, and it would be adverse compared to reference, initial stress for the experiments.
expectations based on radial permeability. Based on experimental results, Walsh (1981)
In an areally anisotropic reservoir with directional suggests that the second (linear) term in his model
permeabilities kx and ky , the geometric average can be neglected except for very high effective
permeability would be equal to (kx ky) which is the stresses. This is in agreement with the Jones and
permeability that controls radial flow and the one Owens (1980) experimental relationship between
that would be obtained form a traditional well test permeability and effective stress for naturally
such as pressure buildup test. Two wells with radial fractured carbonate samples:
permeability of, e.g., 10 md would produce the same
3
under radial conditions, but one could be in a perfectly k
f = [ A + B ln '] , (12-7)
isotropic reservoir where the two permeabilities
are both 10 md while the other could mask severe where A and B are experimentally determined
anisotropy with, e.g., kx = 100 md and ky = 1 md. constants.
Such severe permeability anisotropy is possible in Comparison of the simplified Eq. 12-6 and 12-7
carbonate reservoirs, but even in sandstone reservoirs shows that
permeability anisotropy of 3:1 and greater can be
*
encountered (Buchsteiner et al., 1993.) A = B ln ,
(12-8)
When wells are hydraulically fractured in the two
reservoirs in the example above, the performance will and the model can be modified to
not be the same, and certainly not the performance 3
*
that can be forecast based on radial permeability.
k f = B ln . (12-9)
'
Nolte and Economides (1991) have shown that
the apparent (effective) length will be related This modified model has been used by
to the actual length by: Buchsteiner et al. (1993) with * denoted as
ky the closure quality of the fractures/fissures. The
4
x f ,eff = x f
, (12-5) relationship between permeability and effective
kx
stress (and thus stress-sensitivity) is determined
where kx is the permeability along the fracture and by *. Closure quality is affected by the type of
larger than ky which is the permeability perpendicular porosity, degree of mineralization, or tortuosity
to the fracture. Permeability anisotropy is one of and shape of the flow path.
the reasons why in post-treatment well tests, the In an anisotropic distribution of orthogonal
obtained fracture half-length is invariably smaller fractures/fissures, Eq. 12-9 can be written for the
than the one placed. Permeability anisotropy of e.g., geometric average:
436
* *y 2
3
White at the base of the section represents wet,
k = Bx By ln
x ln
p . (12-10) porous and permeable Ellenberger Dolomite. The

p

r
y r x
grey interval is the Barnett Shale, separated from
the Ellenberger by the Viola Limestone (lower blue
Equation 12-10 (where is the Biot poroelastic limestone) only to the northeast. Note also that the
constant) suggests that reservoir permeability will Barnett increases in depth and thickens to the northeast
decline with decreasing pressure, but it may also decline and is also divided into Lower and Upper intervals in that
heterogeneously if the closure qualities in the two direction by the Forrestburg Limestone (middle blue).
directions are different. Buchsteiner et al. (1993) suggested On this section, the Barnett is overlain throughout by
that permeability anisotropy may change direction with the Marble Falls Limestone (upper blue). The Marble
time, with maximum switching with minimum. They Falls, however, is lost as a thick limestone upper frac
proposed experimental determination of the constants Bx barrier when moving from the core area of Denton,
and By, the directional closure qualities, in reservoirs (such Wise, and Tarrant Counties to the southeast.
as carbonate rocks) susceptible to such phenomena.
12-4.3 Barnett Shale Case Study
The Barnett Shale in the Fort Worth Basin of Texas

is an enormous gas shale reservoir (see Section 11-5.3).
As of 2003, the USGS National Assessment of Oil
and Gas indicated that the Greater Newark East Frac-
Barrier Continuous Barnett Shale Gas and Extended
Continuous Barnett Shale Gas Assessment Units together
encompassed 4.1 million acres of land with mean
undiscovered continuous oil and gas reserves estimated
at 26.2 Tcf gas. It is a formation with a long history of
experimentation, success and failure. From drilling the
first well in the formation in 1981, through April 5,
Figure 12-9 Barnett Shale play well locations as of April
2007, over 6,500 vertical, deviated or horizontal Barnett 5, 2007
tests or producers have been drilled. Over 6,000 are listed
as active wells (Fig. 12-9) in the IHS Energy US Well Barnett production patterns analyzed using GIS
Database. During that time, vertical wells predominated (Geographical Information Systems) methods show
for 20 years or more, and were then supplanted by clear evidence of sweet and not-so-sweet spots (Fig.
horizontal well drilling technology (Fig. 12-10). 12-12). The illustration shows the best and worst 10%
Studying public data pertaining to the North Texas of Barnett vertical wells based on normalized 6 month
Barnett play indicates there is no simple recipe for success cumulative production. The color scale represents log10
in the Barnett. There are different reasons for Barnett of normalized 6-month cumulative gas production with
successes and failures in different parts of the formation. best producers in red and becoming progressively poorer
Technologies used successfully in one area may fail through purple. Note the generally good geographical
miserably in another. Barnett successes and failures separation from most to least production, indicating
result from the intersection of reservoir quality, well reservoir quality drivers. The purple dashed line indicates
architecture, geohazards, completion, and stimulation the approximate Viola frac barrier pinch-out. Viola is
processes. Reservoir quality and geohazards vary according absent southwest of the line. The implication is clear:
to geographic position (Fig. 12-11). Examination of The best vertical wells have Barnett overlying Viola
Fig. 12-11 shows gross stratigraphic changes occurring limestone, generally preventing fracs from contacting wet
regionally from southwest to northeast. Ellenberger dolomite below.
437
438
-7500
-7400
-7300
-7200
-7100
-7000
-6900
-6800
-6700
-6600
-6500
-6400
-6300
-6200
-6100
-6000
-5900
-5800
-5700
-5600
-5500
-5400
-5300
-5200
-5100
-5000
-4900
-4800
-4700
-4600
-4500
-4400
-4300
-4200
-4100
-4000
-3900
-3800
-3700
-3600
-3500
-3400
-3300
-3200
-3100
-3000
-2900
-2800
-2700
-2600
-2500
Depth
Subsea
201ELBG
403MBFLM
8800 8700 8600 8500 8400 8300 8200 8100 8000 7900 7800 7700 7600 7500 7400 7300 7200 7100 7000 6900 6800 6700 6600 6500 6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200 5100 5000 4900 4800 4700 4600 4500 4400 4300
354BRNT 4200 4100 4000 3900
Modern Fracturing
E RATH
8600 8500 8400 8300 8200 8100 8000 7900 7800 7700 7600 7500 7400 7300 7200 7100 7000 6900 6800 6700 6600 6500 6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200 5100 5000 4900 4800 4700 4600 4500 4400 4300 4200 4100 4000 3900 3800 3700
E RATH
8400 8300 8200 8100 8000 7900 7800 7700 7600 7500 7400 7300 7200 7100 7000 6900 6800 6700 6600 6500 6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200 5100 5000 4900 4800 4700 4600 4500 4400 4300 4200 4100 4000 3900 3800 3700 3600 3500
PARKE R
8100 8000 7900 7800 7700 7600 7500 7400 7300 7200 7100 7000 6900 6800 6700 6600 6500 6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200 5100 5000 4900 4800 4700 4600 4500 4400 4300 4200 4100 4000 3900 3800 3700 3600 3500 3400 3300
WIS E
Figure 12-11 Southwest-to-northeast cross-section across the Barnett Shale play

8400 8300 8200 8100 8000 7900 7800 7700 7600 7500 7400 7300 7200 7100 7000 6900 6800 6700 6600 6500 6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200 5100 5000 4900 4800 4700 4600 4500 4400 4300 4200 4100 4000 3900 3800 3700 3600 3500
WIS E
8200 8100 8000 7900 7800 7700 7600 7500 7400 7300 7200 7100 7000 6900 6800 6700 6600 6500 6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200 5100 5000 4900 4800 4700 4600 4500 4400 4300 4200 4100 4000 3900 3800 3700 3600 3500 3400
WIS E
201ELBG
354BRNT
403MBFLM
Depth
-7500
202VIOL -7400
-7300
-7200
-7100
-7000
-6900
-6800
-6700
-6600
-6500
-6400
-6300
-6200
-6100
-6000
-5900
-5800
-5700
-5600
-5500
-5400
-5300
-5200
-5100
-5000
-4900
-4800
-4700
-4600
-4500
-4400
-4300
-4200
-4100
-4000
-3900
-3800
-3700
-3600
-3500
-3400
-3300
-3200
-3100
-3000
-2900
-2800
-2700
-2600
-2500
Subsea
BARNETT 0407 UPDATE

5935 Samples for 6585 Wells
1000000
900000
6 Month Cumulative Gas Production, Mscf
800000
700000
3
600000
500000
400000
2
300000
200000
100000
0
81
83
86
/1 89
/1 91
/0 94
/0 97
/2 00
/2 02
/3 05
07
19
19
19
19
19
06 19
03 19
11 20
08 20
06 20
20
1/
8/
4/
0/
5/
0/
6/
2/
7/
3/
0/
/0
/2
/2
/2
01
09
06
03
12
09
Completion Date
Figure 12-10 Barnett Shale well completion timeline

(white = vertical wells, yellow = deviated wells, blue =
horizontal wells) Figure 12-13 Log of oil-to-gas ratio in Barnett Shale
vertical wells
Well architecture can be the difference between

completing a better or poorer Barnett well (Figs.
12-14 and 12-15). When Figs. 12-14 and 12-15 are
compared, the map clearly shows that hydraulically
fractured horizontal well technology radically extended
operators potential to make good wells, even in the
absence of the Viola frac barrier.
Figure 12-12 Best and worst 10% of Barnett vertical wells

based on normalized 6 month cumulative production
Hydrocarbon liquids and water have negative

effects on Barnett production (Fig. 12-13). The
trend for the Barnett to be oilier to the north and
west is well documented and is a function of thermal
maturity. Liquids in narrow pore throats reduce
relative permeability to gas and result in decreased gas Figure 12-14 Log of 6-month normalized
rate to the north and west. cumulative gas production for vertical Barnett wells
439
Modern Fracturing
are located within the central part of the area, away from
the major faults. They were fracture stimulated mainly
with 30,000 to 40,000 bbls of slickwater and 60,000 to
226,000 lbs of proppant. Average proppant quantity was
161,000 lbs. By comparison, the poorest 5% of wells
(Fig. 12-18) were fractured using approximately 9,000 to
44,000 bbls of slickwater, averaging 27,000 bbls. These
treatments carried 40,000 to 600,000 lbs of proppant,
averaging 210,000 lbs. Treatment ranges are generally
similar, although more proppant on average was pumped
into the poorer wells. The conclusion is that pumping
larger treatments into wells in an area of poorer reservoir
quality may not be helpful.
Fracture Structures
1500
1000
Figure 12-15 Log of 6-month normalized cumulative gas
production for horizontal Barnett wells 500
0 Observation
Different well completion/stimulation types Well
South - North, ft
have been attempted throughout the history of the --500

Barnett play (see Section 11-5.3.1). Over time, two
--1000
technologies have been critical to Barnett success;
slickwater fracturing and horizontal drilling. Early --1500
attempts at nitrogen fracs and foam fracs produced poor --2000

production results. Massive cross-linked, gelled water
--2500
fracs produced more gas than did nitrogen or foam
fracs, but they were expensive. Slickwater fracs yielded --3000
- 1000 --500 0 500 1000 1500 2000 2500 3000
essentially the same production results as cross-linked,
West-East, ft
West-
gelled water, but at significantly reduced cost. Proppants
have evolved from 20/40 mesh standard to 40/70 and Figure 12-16 Complex hydraulic fracturing in a
Barnett Shale well is evident from microseismic frac
100 mesh sizes, trading conductivity for increased monitoring, (Courtesy of Pinnacle Technologies, from
proppant transport and fracture surface area. Fisher et al., 2002).
In addition to the technology factors discussed
above, the Barnett is also a reservoir in which complex
hydraulic fracturing is critical to success (Fig. 12-16.)
Complex hydraulic fractures are beyond the current
capabilities of industry hydraulic fracture modeling,
and knowledge of the geometry and extent of these
fracture networks is derived almost entirely from
microseismic monitoring. Optimum job sizing is chiefly
accomplished through trial and error in different parts
of the play and is area specific.
Review of a specific part of the Barnett play immediately
north of the Mineral Wells fault (Fig. 12-17) shows that
Figure 12-17 In the Alligator Head area north of
within that geographic area, the best 5% of Barnett wells the Mineral Wells fault, the best 5% of wells fall
geographically within the red shaded area
440
Methods of computer mining large data sets were

originally developed for the retail industry when a
team of researchers at the IBM Almaden Research
Center developed a computer algorithm to efficiently
find rules relating sales of individual retail items
purchased together (Agarwal, 1993). Discovering
these previously unknown rules then allowed retailers
to maximize profit through such means as timing of
sales and item positioning on shelves. Since that time,
data mining has been advanced significantly, with
commercial database and software vendors developing
a number of different tools and methodologies
Figure 12-18 The poorest 5% of wells in the Alligator
Head area fall mostly outside the best wells area, closer with which to mine digital data.
to the bounding faults Data mining for the purpose of gas well enhancement
is a natural evolution from the formation studies that
Even with over 6,500 Barnett wells drilled, the have been performed in the well service industry for many
contribution of natural fractures to Barnett production years. Data sets have been computerized and have grown
is controversial, as natural fractures reported from core ever larger, but the goal has remained the same: identify
are mainly healed. What is critical is that the calcite- the fracturing fluids, proppants, additives and techniques
cemented natural fractures comprise closely spaced planes that yield the most incremental gas for the least cost in a
of weakness within the formation that break during the particular formation and field area.
hydraulic fracturing process. Whether open or healed, it
is generally agreed that the principal natural fracture trend 12-5.2 Data Sources
in the core area of the Barnett is WNW, approximately
296 degrees. Present-day h,max is approximately 40 to 45 Data are defined as numerical or other information
degrees, as shown in Fig. 12-16. Thus, present-day h,max represented in a form suitable for processing by
is approximately perpendicular to the principal natural computer (American Heritage Dictionary of the
fracture direction. The best core area horizontal wells are English Language, Fourth Ed., 2000. online). Data types
drilled on an azimuth in the range of 300 to 320 degrees include numbers, dates and text capable of being stored
or its mirror image, nearly parallel to the natural fractures, and processed. In the context of gas well enhancement,
and the main induced fracture fairway is transverse with data may be categorized as pertaining to reservoir quality,
respect to the wellbore. The optimum direction in which drilling, completion, stimulation, operations, workover
to drill a horizontal well is not constant throughout processes and production results. Production results
the Barnett productive area. begin with reservoir quality and are impacted positively
or negatively by drilling, casing and cementing, well
12-5 Data Mining Techniques completion and stimulation, operational practices, and
work-over processes. Gas-in-place,
12-5.1 Purpose of Data Mining
Ah(1 S w )
Gi = ,
(12-2)
Bgi
The purpose of data mining is to cost-effectively
generate useful knowledge from large data sets. The and the gas well producing rate,
process of data mining is a process of discovering
previously unknown qualitative or quantitative
kh( pe2 pwf
2
)
association rules among data items (Thearling, 2007). It q , (12-11)
[ ln(r r ) + s ]
e w
is desirable that the knowledge generated be predictive,
so that it may be used going forward. define reservoir quality in an engineering sense.
441
Modern Fracturing
Some of these data items are easily measurable, low- already operating in an existing database environment
cost, hard parameters while others are difficult or costly complete with data mining tools. When beginning a
to measure. Other data items not included in the above project, it is good practice to review the input data
equations, but with strong potential to affect production file types and formats that the project database and
results, are easily measurable, soft parameters, such the software application are designed to accept and
as clay type, occurrence and abundance. Although will read correctly. Attempting to input a date field in
these data are easily measurable, they are frequently mm/dd/yyyy format when the application is designed
not available because taking core is expensive. In gas to read dd/mm/yy is easily preventable. Particular
reservoir development, major reservoir quality data items care must be taken to remove all commas from data
(in particular permeability and drainage radius) are spreadsheets prior to converting to csv files required
frequently unknown because measurement is costly. for input into some software applications.
Resource ownership patterns and stimulation work being Data preparation includes quality checking and
split among multiple service companies also limit the formatting. Some of this work may be done automatically,
amount of data available for analysis. for example using a logical computation to identify
Data sources globally are mixed with respect to the and delete azimuth values from horizontal wells with
data items available, largely due to resource ownership. indicated azimuth outside the range of 0 to 360 degrees.
Data in the USA and Canada are available from Alternately, the user may elect to identify wells with out-
both public and proprietary sources. Government- of-range data and then undertake manual repairs. Data
mandated well and production records are available preparation may include reconciling multiple sources
through government agencies such as state oil and gas of data. For example, cross-plotting the public data for
commissions, and typically are also compiled into master pounds of proppant pumped for a series of fracs against
databases by commercial vendors. One such database in the same data item from the operator or service company
the USA contains thousands of data items from over 3 dataset will readily identify off-trend, mismatched
million wells. Additionally, proprietary data is held by proppant quantities. Certain data mining tools may
individual operators and service companies. In particular, require normalization of data to within a specific range,
most pumping service companies maintain electronic e.g., 0 to 1 for input into artificial neural networks.
databases of their well treatment reports and do not Suspect data may be deleted from the database,
have access to other service company data, for obvious either automatically, or based on the expert judgment
reasons. Operators maintain varying levels of databases, of the user. Another method of handling suspect data in
from enterprise-level Oracle or SAP databases to project- some software applications uses a quality field where
specific spreadsheets on a particular PCs hard drive. questionable data are identified, e.g., by a question
Throughout much of the rest of the world, oil mark or other identifier. Good software should allow
and gas resources are owned by governments. Data the operator the choice of whether to include any
collection, storage and availability are highly variable particular data item that has been marked as being
from one country to another. In many cases, this leads of questionable quality in any logical, computational,
to data collection at the field level, not well-by-well. It map, or well selection operation.
seems obvious that well production must be measured
on the formation or zone level, well-by-well, in order to 12-5.4 Selected Data Mining Tools
be useful. Lack of well-by-well allocated production data
does not permit useful data mining. Although data mining is frequently defined as an
automated or machine-learning process, much human
12-5.3 Data Preparation thought is still required for a successful outcome.
Many different tools, having greater or lesser degrees
After sourcing key reservoir quality, stimulation of automation, may be used to find solutions to the
and production data, the data must be prepared for issues surrounding gas well enhancement processes.
entry into the project database, unless the user is Pragmatically, any computer method that aids in
442
deconvolution of the contributions of reservoir Figure 12-19 shows the stimulation cost vs. total
quality and drilling/completion/stimulation/ proppant mass for the two proppants. Bubble size
operational processes to production results should represents cumulative production for each well after 90
be considered as data mining tools or techniques. days. Two very clear trends appear. First, as expected,
Artificial neural networks and their subset, Self the stimulation cost is much higher for the premium
Organizing Maps, decision trees, genetic algorithms, proppant than for the economy proppant. The second
nearest-neighbor methods, rule induction, fuzzy and more important trend is that fracture treatments are
logic, geographical information systems (GIS), and much more effective using economy proppant than using
the more standard statistical tools, i.e., histograms, premium, higher-quality proppant with higher proppant
linear and multiple linear regressions, etc., may all be pack permeability. This is contrary to what was promised
used in knowledge discovery. and expected. Why did it happen?
Figure 12-20 shows the 90-day cumulative
12-5.5 Data Mining Case History production vs. total proppant mass. The economy
proppant shows two trends. For both of these trends,
A South Texas operator was performing hydraulic cumulative production increases with increasing
fracture treatments on gas wells using an inexpensive, proppant mass. The premium proppant does not
angular, lower-permeability proppant. The operator show conclusive trends. In fact, some of the worst
had performed approximately 100 fracture treatments cumulative production is from wells stimulated with
before switching to a higher-permeability, more high volumes of premium proppant.
spherical and more expensive proppant. After about
400,000
100 treatments using the better proppant, the operator Economy Proppant
noticed that production from these wells was much Premium Proppant
worse than production from wells fractured with
the cheaper, lower-quality proppant. In some cases,
300,000
production was less than half.
Cum Prod after 90 days, Mscf
350,000
200,000
300,000
250,000
100,000
Stimulation Cost, $
200,000
0
150,000 100,000 200,000 300,000 400,000 500,000
Total Proppant Mass, lb
Economy proppant
Figure 12-20 Cumulative production after 90 days vs.
100,000
Premium proppant total proppant mass
Figure 12-21 shows the stimulation cost vs.

50,000
100,000 200,000 300,000 400,000 500,000 cumulative production after 90 days. Quadrant 1
Total Proppant Mass, lb represents high stimulation costs with low cumulative
production, and all data points in this quadrant
Figure 12-19 Stimulation cost vs. proppant mass
with bubble size representing cumulative production represent wells stimulated with the premium
after 90 days proppant. Quadrant 2 shows high stimulation
443
Modern Fracturing
cost and high cumulative production. Quadrant Figure 12-22 shows the average fracture
4 shows low stimulation cost and low production. conductivity vs. the cumulative production. These
Quadrant 3 is the best performer, showing low are tight gas reservoirs, and high conductivity
stimulation cost and high cumulative production. is not important. In fact, the results show that
Only two wells stimulated with premium most of the wells with high fracture conductivity
proppant fall into this category. have low cumulative production.
Figure 12-23 shows the fracture length vs.
350,000
cumulative production. This is a tight gas reservoir
1 2 and higher production is expected for longer fracture
lengths. This is not the case for many of the wells
300,000
treated with the premium proppant, but it is the
case for the majority of the wells treated with
the lesser-quality proppant.
Stimulation Cost, $
250,000
2400
Economy Proppant
4 3 Premium Proppant
200,000
2200
150,000
2000
Fracture Length, ft
100,000
0 50,000 100,000 150,000 200,000 250,000 300,000 1800
1600
Figure 12-21 Stimulation cost vs. 90-day cumulative
production
1400 1400
1200
1200 0 100,000 200,000 300,000 400,000 500,000
Average Fracture Conductivity, md.ft
Figure 12-23 Fracture half-length vs. 90-day cumulative

1000 production
Figures 12-19 through 12-23 represent a

typical analysis performed by many companies. The
800
highest number of variables that can be represented
using such an analysis is three. Typically, a plot
will have an x-axis, a y-axis and bubble size. This
600
is the Data Mining of yesterday.
Figure 12-24 represents Data Mining today. It
Economy Proppant
Premium Proppant was produced using a method called Self-Organizing
400
0 100,000 200,000 300,000 400,000 500,000
Maps (SOM), first introduced by Kohonen in 1997.
Cum Prod after 90 days, Mscf SOM is a form of neural networks. It organizes data
according to similarities and displays the results in
Figure 12-22 Average fracture conductivity vs. 90-day
cumulative production
a map as shown in Fig. 24. The small black circle
444
in the top left corner of the proppant type num in poor to fair net pay, with poor ultimate recovery
square represents the proppant type for well A. The and stimulated with low volumes of proppant mass.
remaining wells are arranged around well A according Poor candidates combined with less-than-optimum
to their similarities to well A and to each other, based stimulation yielded poor results.
on all the variables used for the analysis. The small The selected area in Fig. 12-27 shows expensive
circle in the top left corner of each of the remaining treatments with poor results. All wells in this area
squares represents the corresponding value for well were stimulated with premium proppant. They
A. The color gradation represents the magnitude are from all formations, show good net pay but
of the variable for each well, and the scale for each poor ultimate recovery and low- to mid-range
square is shown in the bottom. proppant mass. This group represents expensive
For example, in the proppant type num square, treatments that yielded poor results.
well A is dark blue, which means the proppant used The results from Data Mining are combined
for the treatment in well A was of type 4. In the with Unified Fracture Design and sample results are
formation num, the color for well A is red, meaning shown in Table 12-5. The first column shows what
that well A was completed in formation 6. The five a typical design should have been for the premium
squares on the left represent the input variables. proppant, utilizing UFD in this 0.01 md reservoir
What the analyst wants to see is whether there is with net height of about 200ft. Pumping 300,000
any relationship among these five variables. The two lbs of premium proppant, the optimum fracture half-
squares on the right are the outputs. These are the length should be 1000 ft (fracture height about 450
variables that measure success. ft) and the job should take about 490 minutes to
Although there is no limit on the number of input execute. This would result in an 11-fold PI increase.
variables, the level of difficulty in interpreting the That is, the well would perform 11 times better than
results increases with the number of outputs. For the an undamaged, unstimulated well.
purpose of this chapter, the selected inputs are proppant
Table 12-5 Sample Results with Unified
type, formation number or name, proppant mass,
Fracture Design
estimated ultimate recovery and estimated net pay. The Economy Premium
Premium
two outputs selected are cumulative production after Proppant
Proppant Proppant
Design Execution
90 days and stimulation cost. (More analyses were
Proppant Permeablility 200,000 10,000 200,000
performed but are not shown here.)
Proppant Mass, lbs 300,000 300,000 300,000
In interpreting the results of Data Mining, as
Folds of PI increase 11 7.2 6
shown in Fig. 12-24, there can be many conclusions.
Fracture Half Length 1000 890 464
The circled area in Fig. 12-25 represents the most
Pumping Time 490 359 93
successful fracture treatments based on cumulative
production after 90 days. This same area in proppant
type shows that all successful fracture treatments were If the treatments were performed with the
made with the economy proppant, in the formation economy proppant, then the folds of PI increase
shows that all successful treatments were performed in would have been 7.2, as it is shown in the second
formation 6, the wells have good estimated ultimate column. The last column shows the typical treatment
recovery, good net pay, were performed with moderate- utilizing the premium proppant as it was actually
to high-proppant-mass and had moderate treatment executed. Because the premium, manufactured
costs. In other words, good reservoir quality, stimulated proppant had good sphericity, it was easy to pump
appropriately will give good production. and easy to place without screenout. These treatments
The selected area in Fig. 12-26 shows fair were pumped very fast, in about 90 minutes, resulting
treatments with low cost and poor to fair cumulative in facture half-lengths of less than 500 ft, less then
production. The wells here are from all formations in one-half of the optimum. The resulting folds of PI
the study, all stimulated using the economy proppant, increase of these treatments is around 6, less than
445
Modern Fracturing
Figure 12-24 Self-Organizing Maps set up some parameters for Data Mining analysis
that of the ideal designed treatments using the

economy proppant, and about one-half of that of
the ideal design using the premium proppant. Under
these circumstances, a proper design with a lesser
proppant and utilizing UFD outperformed a poor
design using the best proppants.
446
Figure 12-25 Most successful treatments, based on 90-day cumulative production
447
Modern Fracturing
Figure 12-26 Fair treatments with low cost and low to fair 90-day cumulative production
448
Figure 12-27 Expensive treatments with poor results
449
Modern Fracturing
REFERENCES Rueda, J., and Wolcott, D.S.: Pushing the limits

of hydraulic fracturing in Russia, paper SPE
Agarwal, R., Imielinski, T., and Swami, A.: Mining 90357, 2004.
Association Rules Between Sets of Items in Large Elkins, L.F. and Skov, A.M.: Determination of
Databases, Proc. of Very Large Data Bases, pp. Fracture Orientation from Pressure Interference,
207-216, 1993. Trans. AIME, 219 (1960) 301-304.
Baranov, D., Kontarev, R., Senchenko, D., Faizullin, Finley, S.J. and Lorenz, J.C.: Characterization and
I., Gayfullin, A., Ivshin, A., and Kazykhanova, Significance of Natural Fractures in Mesaverde
N.: Fracture Treatments of Horizontal Wells on Reservoirs at the Multiwell Experiment Site, paper
Western Siberia Oil FieldMaximizing Long- SPE 19007, 1989.
Term Production, paper SPE 107845, 2007. Fisher, M.K., Wright, C.A., Davidson, B.M., Goodwin,
Buchsteiner, H., Warpinski, N.R., and Economides, A.K., Fielder, E.O., Buckler, W.S. and Steinsberger,
M.J.: Stress Induced Permeability Reduction in N.P.: Integrating Fracture Mapping Technologies
Fissured Reservoirs, paper SPE 26513, 1993. to Optimize Stimulations in the Barnett Shale,
Daal, J.A., and Economides, M.J.: Optimization of Paper SPE 77441, 2002.
Hydraulically Fractured Wells in Irregularly Shaped Jones, F.O. and Owens, W.W.: A Laboratory Study
Drainage Areas, paper SPE 98047, 2006. of Low Permeability Gas Sands, paper SPE 7551,
Demarchos, A.S., Chomatas, A.S., Economides, M.J., 1980.
Mach, J.M., and Wolcott, D.S: Pushing the Lorenz, J.C., Branagan, P., Warpinski, N.R., and Sattler,
Limits in Hydraulic Fracture Design, paper SPE A.R.: Fracture Characteristics and Reservoir
86483, 2004. Behavior of Stress-Sensitive Fracture Systems in
Demarchos, A.S., Economides, M.J., Diyashev, I., Flat-Lying Lenticular Formations, paper SPE
Svaykin, V.A.: Analysis of the Performance of 15244, 1986.
Hydraulic Fracturing Treatments and Quantum Kohonen, T.: Self Organizing Maps, Springer,
Design Improvements, paper SPE 94643, 2005. Heidelberg 1997.
Detournay, E. and Cheng, A.: Influence of Marongiu-Porcu, M.: Physical and Economic
Pressurization Rate on the Magnitude of the Optimization of Hydraulic Fracturing, MS
Breakdown Pressure, Proc., 33rd Symposium on Thesis, University of Houston, 2007.
Rock Mechanics, Santa Fe, New Mexico, USA Nolte, K.G,: Determination of Fracture Parameters
(1992), 325333. from Fracturing Pressure Decline, Paper SPE
Earlougher, R.C., Jr.: Advances in Well Test Analysis, 8341, 1979.
Society of Petroleum Engineers, Richardson, Texas, Nolte, K.G. and Economides, M.J.: Fracture Design
1977. and Validation with Uncertainty and Model
Economides, M.J., Hill, A.D. and Ehlig-Economides, Limitations, JPT (Sept. 1991), 1147-1155.
C.A..: Petroleum Production Systems, Prentice Romero, D. J., Valk, P. P. and Economides M. J.:
Hall, NY, 1994. The Optimization of the Productivity Index and
Economides, M.J., and Nolte, K.G.: Reservoir the Fracture geometry of a Stimulated Well with
Stimulation, Third Edition, John Wiley and Sons Fracture Face and Choke Skins, Paper SPE 73758,
Ltd., 2000. 2002.
Economides, M. J., Oligney, R.E. and Valk, P.P.: Rueda, J.I., Mach, J., and Wolcott, D.: Pushing
Unified Fracture Design, Orsa Press, Houston, Fracturing Limits to Maximize Producibility in
2002a. Turbidite Formations in Russia, Paper SPE 91760,
Economides, M.J., Oligney R.E. and Valk, P.P.: 2004.
Applying unified fracture design to natural gas Rueda, J.I., Voronkov, A., and Mach, J.: Optimum
wells, World Oil, (Oct. 2002b), 52-62. Fracture Design under Transient and Pseudosteady
Economides, M.J., Demarchos, A.S., Mach, J.M., Conditions using Constant Fracture Volume
450
Concept, (2005) 67th European Association of

Geoscientists and Engineers, EAGE Conference
and Exhibition, incorporating SPE EUROPE
2005 - Extended Abstracts, pp. 1795-1805.
Sabaev, V.V., Wolcott, D.S., Mach, J.M., Antipina, D.V.,
Haidar, A.M., and Sviyazova, O.O.: Vertically
Fractured Well Performance in Rectangular
Drainage Area, paper SPE 101048, 2006.
Siegfried, R. and Simmons, G.: Characterization of
Oriented Cracks with Differential Strain Analysis,
Journal of Geophysical Research (February 1978)
83, No. B3, 12691278.
Strickland, F.G. and Ren, N.-K.: Predicting the In-
Situ Stress for Deep Wells Using Differential Strain
Curve Analysis, paper SPE 8954, Proc., SPE/DOE
Symposium on Unconventional Gas Recovery,
Pittsburgh, Pennsylvania, USA (1980), 251258.
Teufel, L.W.: Determination of In-Situ Stress From
Anelastic Strain Recovery Measurements of
Oriented Core, Paper SPE 12304, 1983.
Thearling, K.: An Introduction to Data Mining,
white paper, p.4, 2007.
Vakhitova, A.: Analysis of Hydrofracturing Efficiency,
paper SPE 102064, 2006.
Walsh, J.B.: Effect of Pore Pressure and Confining
Pressure on Fracture Permeability, Int. J. of Rock
Mechanics, Min. Sci. & Geoph. Abstracts, 1981.
451
Modern Fracturing
452
Tony Martin is business development manager for international stimulation at BJ Services Company.
Since graduating from Imperial College, London, with an honors degree in mechanical engineering and a
master's degree in petroleum engineering, Martin has spent 17 years in the oil industry and has completed
engineering assignments around the world. Martin's primary interest has been hydraulic fracturing and
stimulation, and he has been involved in production enhancement projects in more than 25 countries.
He teaches fracturing, acidizing and sand control both in-house and externally. A constant theme in this
teaching is the need to de-mystify the world of hydraulic fracturing, in an attempt to make the process
more accessible and less intimidating. He is the author or co-author of numerous SPE papers and has
served on the technical committees for several SPE events. He is also the author of BJ Services Hydraulic
Fracturing Manual.
Martin Rylance is a senior petroleum engineer and engineering adviser with BP Exploration Co. Ltd. He
has been with BP Exploration, their partners and joint ventures (SOHIO, Maxus, TNK-BP) for more
than 20 years since graduating with a BSc (honors) in pure mathematics. During his time with BP, his
positions have included head of fracturing and stimulation for BP Worldwide, Interventions Team leader
with BPX in Colombia, chief engineer with TNK-BP in Moscow, and more recently, project manager
with VICO in Jakarta. Rylance has been involved in all aspects of pumping operations and has worked
in most of the major known tectonic basins in the oil & gas industry. In recent years he has specialized
in fracturing in tectonic regimes, HP/HT environments and mature field re-development. He has been
responsible for implementing many campaigns and more than 5,000 frac and other stimulation treatments
around the world. He has numerous papers and publications with the SPE, AIME and various numerical
and geological societies, and a number of patents. An active member of the SPE, he is also a fellow of the
Institute of Mathematics in London.
Chapter 13 well. One of the editors of this book recalls that on
at least one occasion, the production manager of a
Technologies for company, after having been pressured by upper
Mature Assets management to acquiesce to fracturing, provided a
Tony Martin, BJ Services and candidate well for the treatment; only to find out that
the well was some 25 years old and already producing
Martin Frac-Pup Rylance, BP Exploration
at a 99% water-cut. The logic was that if fracturing
is such a good approach to production enhancement,
13-1 Introduction it should be capable of also turning water into oil?
Another frequently encountered problem is to judge
We usually find oil in new places with old ideas. and assess the appropriateness of fracturing, based
Sometimes we find oil in an old place with a new on the results from just a single well. Fracturing
idea, but we seldom find much oil in an old place is complex, many things can happen, numerous
with an old idea. Several times in the past we thought situations can arise and any negative results may affect
we were running out of oil, whereas actually we were adversely the potential for future deployment of this
only running out of ideas. widely established well completion and stimulation
Parke A. Dickey, 1958 operation in a particular petroleum province.
The evolution of the tip screenout (TSO)
It was not very long ago, in fact as late as 1990, technique in the late 1980s (see Section 4-7.3) and
that many engineers in the petroleum industry its subsequent application has essentially rendered
still considered hydraulic fracturing as applicable every well a potential candidate. Consider this: It
exclusively within the realms of low- to very low- would take more than 400 gas wells within a 0.01 md
permeability reservoirs. While it is still absolutely true permeability reservoir, with massive hydraulic fracture
that hydraulic fracturing remains the only technique treatments, to equal the incremental performance
that can render low-permeability formations economic, of just one correctly designed and successfully
some unfortunately, continue to believe that this is its fractured well in a 10 md reservoir.
exclusive application and remain unaware of major It is therefore clearly not too late to start
new developments over the last two decades that make fracturing in many mature areas where fracturing has
fracturing compelling for absolutely any reservoir. For not yet been attempted (or where fracturing has not
others, fracturing has been the action of last resort: If yet been implemented correctly).
everything else fails, lets try fracturing. The previous chapters of this book have outlined
There have been two major reasons for occasional the majority of the important building blocks,
reluctance to consider this technique. The first of technologies and subjects for consideration in order
these is the effective cost of fracturing, especially to deliver successful gas fracturing operations. This
when performed on an ad hoc basis, one job at a time, chapter discusses the appropriate implementation of
assigning all of the mobilization and set-up costs to this knowledge in an integrated and multi-disciplinary
a single treatment, a situation that is clearly both approach to fracture execution, with particular
undesirable and untenable. The second of these is an emphasis on fracturing application to mature assets, re-
unreasonable phobia that fracturing generally increases fracturing operations and as a remedial technique for
water production, a problem that is quite manageable poorly performing conventional completions.
and indeed a situation that for an industry accustomed
to produced water management should be part of the 13-1.1 Definition of a Mature Asset
normal way of doing business.
All too often, in areas where fracturing has not been How do we define a mature asset? What parameters
employed in the past, if tried for the first time, it would determine if an asset or field fall within this
generally be on a very bad and/or an underperforming category? A number of definitions can be applied at
455
Modern Fracturing
various periods during the life of a field, and these 13-1.2 Minimum Cost & Maximum Value
apply equally to oil and gas, see Babadagli (2005),
and Fig 13-1. One simplistic definition, from the There are various ways in which a mature field may be
Oxford English Dictionary (1989), defines mature re-developed (see Fig. 13-2), and full consideration
as an entity having achieved maximum growth of these techniques should form the basis of a
and development by natural means. For a group selection process which maximizes the overall value of
of operating wells, to which fracturing may their re-development.
be targeted, this definition would appear to
be reasonable for our purposes. Mature Field
Re-Development
-
Fracture Re-Fracture
Infill Drilling or
Existing Wells Fracced Wells
Side-Tracking
(Chap. 13.) (Sec. 13.5)
? ?
Production
?
New New Well
Completion Fracture
? (Conv, HAW,) (Chap. 12).
Figure 13-2 Options available for mature field re-

development
?
When is a field called mature?
Time
During the initial development phase the wells may
Figure 13-1 Various points at which a field may be have been completed with hydraulic fracture treatments
referred to as mature (from Babadagli, 2005) as part of their initial completion; re-fracturing will be
discussed in some detail within Section 13-5. However,
Typically then, mature assets will have if the development is not within conventional hydraulic
undergone initial periods of discovery, delineation fracturing regions such as the US and Canada, then
and development. Through this phase they would it is more likely that the wells would not have been
have grown into a period where a production originally completed with hydraulic fracturing as
plateau and/or maximum delivery rate might the primary completion technique, and this can
have been achieved, the upper constraint of which present a number of significant issues in terms of
is driven by a limited throughput or choke in any potential remedial implementation.
some part of the production delivery system. After Reviewing the literature, it would appear
a period of plateau production, the field/asset is that there are a number of golden rules or
likely to enter an extensive and longer-term decline common themes related to performing remedial
phase, during which time the standard suite of hydraulic fracturing operations on mature fields.
intervention operations may not be able return the These rules should be considered in order to
deliverability to the previous plateau rates. ensure the maximum economic efficiency from
This means that some of the common campaigns within mature gas fields:
characteristics that will be shared by mature assets
can typically include pressure depletion, presence Golden Rules of Efficient Remedial Fracturing
of various scales, increasing produced water and/or Re-Fracturing:
issues and mechanical integrity problems, to name
but a few. Some of these assets may originally Rule 1: The old adage that the better and more
have been classified as oil fields or gas-cycling productive wells are generally the better
schemes, but increasing gas-oil ratio (GOR) and fracturing candidates holds true for
efficient sweep may have resulted in their effective both mature field re-development and
conversion to gas fields. re-fracturing operations.
456
Chapter 13 Technologies for Mature Assets
Rule 2: Ensure that any planned pilot remedial program Frac Phobia or Lack of Knowledge of Fracturing:
has sufficient critical mass (volume/number This is a sweeping and recurring theme throughout
of wells/treatments) to ensure that a learning the petroleum producing world. Hopefully, a
curve can be followed; one-off treatments textbook such as this will help to remedy the
are not generally successful. situation. Non-fractured wells, ambling along, result
Rule 3: Candidate selection should take into in huge amounts of unrealized revenues.
account all aspects that contribute to the Ineffective Initial Completion: Many wells fall into
overall remedial economics. Considering all this category, for example, if not all of the potential
of the contributory factors will minimize pay was completed or the drilling, cementing and
the associated risks. completion fluids were incompatible with the
Rule 4: Do not underestimate the potential impact formation leading to damaged or sub-standard
of mechanical integrity issues within wells. For such wells, where the initial conventional
existing wellbores during a remedial completion approach has resulted in under-delivery,
campaign. Uncertainty and costs associated hydraulic fracturing can be utilized to improve and
with unexpected/unplanned workover stimulate that delivery significantly.
operations can dominate and devastate
treatment economics.
Rule 5: Pre-plan an appropriate Exit Strategy from a
campaign. Treatments should be performed
on a ranked table of candidates, results
monitored carefully and the campaign halted
at an appropriate juncture.
Rule 6: Successful fracturing may show that existing
well completions, such as the sizing of
production tubing, may be limiting incremental
production. In an integrated approach not just
the stimulation but also the well design should
be constantly re-evaluated.
Finally, it should be apparent that the optimum

choice for a given field may not consist of a single re-
development technique. Variable off-take, reservoir Figure 13-3 Numerous reasons to consider hydraulic
heterogeneity, improved reservoir understanding and fracturing
many other factors may result in a combination of
approaches and techniques providing the most efficient Progressive Deterioration Damage Effects:
and economic means of re-development. Many forms of damage can result in progressive
deterioration of the well and an overall reduced
13-1.3 Motivation for Fracturing performance with time. These include scale
deposition, condensate banking, fines migration
There are numerous reasons why hydraulic fracturing and induced damage from interventions. These
of existing wells may be considered a preferred appear in wells whose performance does not
remedial technique for mature and declining fields: match the expected reservoir behavior; hydraulic
costs associated with an infill drilling program, well- fracturing can overcome some or all of the effects
spacing/drainage issues, etc. (Fig. 13-3). However, of these damage mechanisms. However, alternative
we can readily classify most of these reasons into a techniques must be considered, as well as treatments
number of broad categories. that combine hydraulic fracturing with other
457
Modern Fracturing
remedial approaches (such as including a relative and declining field, the fundamental starting
permeability modifier or scale inhibitor into the point should be the existing reservoir, completion
fracturing fluid or proppant coating). and system performance.
Reservoir Properties: Changes in reservoir
properties are generally driven by a reduction in 13-1.4 New Technologies/Approaches
pore pressure, both near the wellbore and in the
far-field. For example, a reduction in pore pressure A number of recently deployed technologies have
will create a change in stresses, leading to changes radically affected the suite of options available to
in potential fracture geometry (manifested as perform fracturing operations. In addition, a number
additional confinement, fracture re-orientation, of these technologies are particularly suited to mature
etc.) Other effects such as condensate banking, as asset application because they often make the operations
already noted, can result in enhanced impact of the simpler, cheaper and/or quicker.
relative permeability relationship for the reservoir, a Ultra-Lightweight Proppant (ULWP): As discussed
factor that may not have been originally considered in more detail in Section 8-3.4, ULWPs are designed to
or well understood during the original field have neutral density within a fracturing fluid (Rickards
development assessment. In these wells, reservoir et al., 2003). This means that the proppant particles will
maturity enhances the opportunity for fractuirng to remain suspended in the fluid without the requirement
deliver significant improvements. for significant fluid viscosity, resulting in two major
Technology Advances: Since its inception in advantages. The principal benefit is that the proppant
the 1950s, hydraulic fracturing has proven to be a can be readily and efficiently placed using a simple brine-
particularly innovative technique in every aspect of based fluid system. Clearly this will significantly reduce
its application and execution, as reported by Veatch the operational complexity and treatment costs, while
(2007). Technology improvements, new solutions maintaining the ability of the proppant to stay suspended
and direct and flexible response(s) to emerging issues in the fracturing fluid in the fracture. The secondary
have all continued to ensure the generalized and benefit is that the ULWP can be pre-mixed into the
expanding application of this approach. Many of the brine and held in tanks on the location until it needs to
more recent and effective technologies will be further be pumped. This pre-mixing approach eliminates the
discussed within the next Section. need for complex blending, proportioning and metering
System Pressures: There are also many reasons systems on location, as demanded by conventional
why improving part of the flow system can be a valid proppant systems. This approach will also mean that the
reason for implementing hydraulic fracturing within concentration of the ULWP can be precisely controlled
a mature field. Of these reasons, many are related to in a pre-mix situation, thereby avoiding unplanned/
the reservoir fluids (e.g., scale deposition, condensate undesirable events during pumping.
dew-point, coning effects, etc.) and others to the rock The combination of these two benefits dramatically
mechanics of the formation (e.g., pressure-dependent simplifies surface operational complexity, eliminating
permeability, sand production, etc.) Advantages can the blender and proppant-handling equipment in favor
also be achieved by reducing the pressure values within of mixing tanks and a booster pump. This simplification
the completion itself and improving surface gathering is particularly applicable to mature assets, where cost
system efficiency. All of these improvements exploit and complexity can have an exaggerated effect on the
a hydraulic fracture treatments ability to improve a operational economics. In addition, ULWPs have been
wells underlying productivity index (PI) such that the used extensively with slickwater fracturing systems in
additional rate or pressure can be utilized to improve tight and unconventional gas reservoirs (Schein et al.,
overall system performance and efficiency. 2004; and Posey and Strickland, 2005), where the fluid
Therefore, regardless of the particular systems are kept as simple as possible in order to reduce
reason(s) for considering hydraulic fracturing the costs, minimize formation damage and maximize
operations for potential application to a mature fluid recovery (see also Chapter 11).
458
Proppant Partial Monolayer: Early work by Darin packer system is then used to isolate individual
and Huitt (1959) investigated the effects of variation in treatment intervals, upon each of which a fracture
areal proppant concentration on fracture conductivity treatment is then performed down the coiled tubing.
(see Section 8-4.8 for more details). At unusually low During the pumping operations the coiled tubing
areal proppant concentrations (0.09 lbs/ft2 of 20/40 is static, allowing maximum injection pressures
fracturing sand), they were able to show that a fracture considerably higher than the limit for conventional
would provide similar conductivity, at low stresses, to a coiled tubing operations, where the tubing is being
fracture propped with 4.0 lbs/ft2 of frac sand. It was continuously plastically deformed. The advantages of
appreciated that this technique would not be suitable using coiled tubing are that multiple intervals can be
for every well, as the partial monolayer approach rapidly treated (in excess of 20 in a single day) and
means that each individual proppant grain experiences that the coiled tubing can then be used to circulate
a significantly higher stress level. In addition, medium- proppant from the wellbore and even gas-lift the well
to high-permeability wells require an areal proppant back into production. This technique has also been
concentration in excess of 4.0 lbs/ft2. However, for successfully applied in coalbed methane formations
shallow and lower-permeability environments, this and to selectively stimulate specific intervals in a
technique was considered appropriate. multi-interval wellbore (Zemlak et al., 1999; and
However, at the time when Darin and Huitt Gilbert and Greenstreet, 2005). In the case of mature
published their work, fracturing technology was not reservoirs with multiple intervals, this application is
quite capable of delivering this laboratory effect in most readily applicable where multiple by-passed or
a field environment, and the approach was largely unstimulated pay zones may exist.
ignored for upwards of 40 years. More recent advances Fracturing through coiled tubing has not been
however, in both metering and control technology, as widely applied outside the Canadian operations
well as the introduction of reliable ULWP systems and selected areas of the USA. This is because in
(see text above), have made the application of this order for the technique to be most cost-effective, the
approach a practical consideration (Aboud and Melo, additional expense of using coiled tubing has to be
2007). Chambers and Meise (2005) measured similar competitive with the charges associated with the use
performance in monolayer fractured wells, as that of a conventional workover rig. This has limited the
predicted by Darin and Huitts original work. Clearly, application of coiled tubing fracturing technology
placing approximately 0.09 lbs/ft2 of proppant to those areas with numerous shallow intervals and
instead of 4.0 lbs/ft2 will result in substantially relatively high conventional workover costs.
reduced proppant volumes and hence costs per Open-Hole Horizontal Wells: Until quite recently,
treatment, making these techniques an ideal solution the potential of fracturing open-hole horizontal wells
for a mature reservoir environment. was fairly limited and effectively considered to be the
Fracturing Through Coiled Tubing: During worst-case scenario in terms of wellbore remediation.
recent years fracturing through coiled tubing (see Such wells were extremely difficult to fracture with
Section 9-5) has become a well-established method proppant due to issues related to control of fracture
of developing and re-stimulating the shallow, multi- azimuth, prevention of initiation and development of
zone gas reservoirs of southern Alberta and additional multiple fractures and a strong tendency for early screen-
areas of Canada and the USA (Spady et al., 1999; out. Furthermore, the situation was often exacerbated
McConkey and Stromquist, 2000; and Rodvelt et al., and over-complicated by the tendency to drop slotted
2001). This technology is usually applied to mono- liners in open-hole environments.
bores or wells that have not yet been completed, More recently an innovative technique has
after all of the proposed intervals have been initially been successfully developed and applied to placing
perforated. The treatments are generally performed fractures under these conditions (Surjaatmadja et al.,
down large-diameter coiled tubing (usually 2 or 1998; and Love et al., 2001), utilizing a specialized
2 in. OD) with an isolation packer system. The jetting tool positioned on the end of coiled tubing
459
Modern Fracturing
(or jointed pipe). The technique takes advantage of has shown that for effective operational execution, the
Bernoullis stagnation pressure effect, whereby the preferred fracture azimuth should be within 30 of
act of instantaneously arresting a high-velocity fluid the perpendicular of the wellbore. Finally, it should
causes a significant increase in pressure. also be remembered that when operators decide to
remedially fracture a horizontal wellbore, it is because
the well has failed to meet its original production
Annular fluid is
sucked into the expectations i.e. It is already an underperforming
cavity by the jet well. Of course, had the operator planned to fracture
the horizontal well from inception, the well would have
been designed and completed in order to facilitate the
Location of the
deployment of fracture(s) in an optimum, efficient
Hydraulic Fracture
Highest Pressure and economic fashion (see Chapter 10).
continues to extend Unconventional Fracturing Fluids: Tremendous
Low High advances have been made in recent years in the area
Pressure Chart
of unconventional fracturing fluids (see Section
Figure 13-4 Generating a hydraulic fracture with the 7-4.6). Many of these fluids do not rely on the
fluid-jetting approach use of polymers or metallic cross-linkers, and in a
number of cases these fluids are significantly easier
In order to achieve this, the jetting tool, which is to manage than conventional fluid systems. Other
specially assembled for each application, is positioned fluid options offer considerable advantages in areas
adjacent to the interval that is to be fractured. such as formation compatibility and fluid clean-
Clean fluid is then pumped down the annulus at up. Unconventional fracturing fluid technology has
just below the estimated fracturing pressure. At the made a major impact on fracturing mature reservoirs
same time, fluid is also pumped down the tubing or via two fundamentally distinct approaches, and
coiled-tubing at a rate sufficient to generate 4,000 to these are the use of CO2 fracturing systems and
5,000 psi differential pressure across the jetting tool. viscoelastic surfactant systems.
In the wellbore, the Bernoulli effect will cause the CO2 Fracturing Systems: Carbon dioxide (CO2) has
high-velocity fluids pumped through the jetting tool been used in hydraulic fracturing for many years as a
to create a very localized region of higher pressure, partial-phase component, usually to provide a foamed or
which is sufficient to fracture the formation (Fig. energized phase to conventional fluid systems (Crawford
13-4). The pad and proppant-laden fluids are then et al., 1963; Hurst, 1972; and Black and Langsford,
pumped through the tubing or coiled tubing, with 1982). CO2 is particularly suited to this application, as
clean fluids only pumped down the annulus, at a rate it can be pumped in a liquid state by conventional high-
sufficient to maintain a constant annulus pressure. pressure pumping equipment. However, two relatively
Some variations in this technique also pump low recent applications of CO2 fracturing technology have
proppant concentrations down the annulus. Most of significantly advanced the application in low-pressure,
the fluid pumped down the annulus leaks off into the fluid-sensitive and mature reservoir environments (see
formation before reaching the fracture. However, it Section 7-4.5). The first of these applications is the
is estimated that sufficient annulus fluid enters the employment of a CO2 methanol system, eliminating
fracture to dilute the treatment by 20 to 30%, and the use of water entirely (Gupta et al., 2003). Proppant
this must be allowed for in the job design. is slurried directly into the methanol base and mixed
Although this technique is undoubtedly with the CO2 at very high pressures before being
a significant advance in deploying fracturing pumped down the wellbore. Within the formation, the
technology, the most valuable application is realized CO2 turns to gas and the methanol should vaporize,
when performing remedial operations within existing leaving little or no fluid residue and no affects from
horizontal wellbores. In addition, practical experience increased water or hydrocarbon saturation.
460
Another innovative application of CO2 which then imparts viscosity to the system. The
for fracturing is the use of a 100% CO2 system systems are extremely shear-thinning (which gives
as reported by Lillies and King (1982). With them a low friction pressure), easy to break and non-
this method, proppant is mixed directly into a damaging. They have three additional properties that
liquid CO2 base using a special high-pressure make them especially suitable for mature reservoir
refrigerated blending system that looks and operations. First, as they break back to essentially
operates like a bulk CO2 tank with a sand auger water viscosity, they are easy to recover from the
running along the bottom (Fig. 13-5). Before formation; therefore, they are easy to deploy and
the treatment, proppant is loaded into the CO2 apply. Second, the dual-surfactant system incorporates
blender and liquid CO2 is then used to cool the surfactant chemistries that act as biocides, low
proppant down (which must be entirely dry, or it surface-tension modifiers, clay-stabilizers and non-
will freeze solid). Once the proppant is at the correct emulsifiers: just add the two surfactants on the fly,
temperature and pressure for liquid CO2 fracturing, and the system is complete. Finally, some of these
the auger can then be used to add the proppant at systems can be recovered, recycled and re-utilized on
controlled rates to a stream of liquid CO2. Although additional treatments, with obvious potential cost-
treatments are small (each CO2 blender holds saving benefits. The VES systems, however, do have
approx. 30,000 lb only), this technique represents two distinct disadvantages: higher costs and lack of
the ultimate in non-damaging, zero residue, high-temperature stability (approximate temperature
easy-to-recover fracturing fluids. limits are 240 F for the surfactant-brine systems and
180 F for the dual-surfactant systems).
Understanding Aqueous Phase Trapping: It
is possible -- and necessary -- to assess whether
a formation is likely to be sensitive to aqueous
phase trapping phenomenon (also known
as water blocking) that could be caused by
Pre-Cooled introducing an additional interstitial water phase
Proppant
via the fracturing fluid system.
Due to the work of Bennion and colleagues
(1994, 1996, 2002), it is now possible to predict
and potentially mitigate the effects of aqueous phase
trapping effects within gas reservoirs (see Section
7-6.2). Bennion et al. introduced the concept of
Liquid Liquid CO2
CO2 In Slurry
an Aqueous Phase Trapping Index (APTi) and
percentage Bulk Water Volume (%BWV), which
Figure 13-5 Schematic Diagram of the CO2 Blender are calculated as follows:
Viscoelastic Surfactant Systems: The second APTi = 0.25[ log10 ka ] + 2.2 S wi , and (13-1)
unconventional fracturing fluid technology, applied
to mature reservoirs, is that of viscoelastic surfactant
(VES) fluid systems (see Section 7-4.6.1). There are %BWV = S w 100, (13-2)
two derivatives of this system available, the early
surfactant-brine systems (Nehmer, 1988; and Brown et where ka is the uncorrected absolute permeability to
al., 1996) and the more recent dual-surfactant blended air (md), Swi is the initial water saturation (fraction),
systems (Gupta et al., 2005). Both approaches utilize Sw is the current water saturation (fraction) and is
surfactant technology to generate a long, worm- the porosity (fraction). Bennion et al. defined the
like micellar structure within a base water or brine, following criteria based upon these indices:
461
Modern Fracturing
Aqueous Phase Trapping Index for a formation: Contract Duration: Whenever possible, core
fracturing contracts should be tendered and awarded for
APTi 1.0 Unlikely to exhibit the longest duration reasonably possible. Consideration
significant permanent of equipment, personnel and material logistics have
sensitivity to APT shown that short-term contracts are often dominated
0.8 < APTi < 1.0 May exhibit by complex set-up costs and cannot truly demonstrate
sensitivity to APT and deliver baseline economics.
APTi 0.8 Likely to exhibit significant Volume Scope of Work: As noted above, the volume
sensitivity to APT scope of work is important (for a pilot-program) to
maximize learning-curve delivery. Under a longer-
Percent Bulk Water Volume for a formation: term contract award, a reasonably accurate (P50) scope
will allow both parties to apply appropriate resource
%BWV 3.5 Unlikely to exhibit and maximize their overall economics. For example, if
significant permanent a tender specifies between 10 (P90) and 100 (P10)
sensitivity to APT treatments, a service provider will often plan on
2.0 < %BWV < 3.5 May exhibit pricing and support for the P90 case (10 operations)
sensitivity to APT and consider anything else upside. Often a potential
%BWV 2.0 Likely to exhibit significant project will be in direct competition for access to limited
sensitivity to APT resources (people, equipment, technology, etc.) within
both the operator and the service provider companies;
Employing and evaluating these correlations can assist a demonstrably realistic and large scope will play a role
with the decision about to whether to utilize a non- in deciding allocation of such resources.
aqueous base for the fracturing fluid. Technical Scope of Work: If a pilot program has been
performed or earlier applications already demonstrated
13-1.5 Reducing Treatment Costs appropriate solutions, then this should allow a reasonably
tight technical scope of work to be defined. Uncertainty
When employing hydraulic fracturing technology about technology (e.g., fluid chemistry and additives,
to re-develop a mature asset, it is often critical to the treating pressures and rates (HHP requirements), etc.)
well or campaign economics to extract maximum cost generally translates into higher costs for the operator
efficiency from fracture treatments and associated caused by larger inventory (among other issues) for
operations (pre-fracture preparation, post-fracture the service providers. This may not be such a large
restoration to production, etc.) at the earliest possible problem within Canada and the USA; however, in
opportunity. To achieve this within a reasonable overseas and remote locations inventory costs must
time-frame, it is generally desirable for the operating be recovered within the limited number of treatment
company and the service provider(s) to work together opportunities, hence reduced inventory translates into
closely to develop a efficient approach. A number of a better contract for all parties.
specific considerations can help deliver and maximize Equipment Requirements: A fracturing contract
this, including the following: tender should be practical when specifying the necessary
Pilot Program: A suite of pilot treatments are equipment for a new contract. Clearly, brand-new
sometimes required in order to sufficiently demonstrate state-of-the-art equipment will generally command
the true potential for full-scale program implementation. brand-new state-of-the-art pricing. However, a 7-year-
In order for this approach to have any reasonable old fracturing pump is just as capable of delivering a
chance of success, the volume scope (learning-curve treatment as a brand-new one, if it has been reasonably
potential) and quality of the candidates (best producers well-maintained. For small treatments (typically 1 to 3
often make best candidates) require particularly careful tanks), the use of hydration units simply adds complexity
attention and consideration. and cost to the rig-up and is only really applicable under
462
exceptional HSE circumstances. In addition, a well- in limbo, awaiting new contract implementation. In
written contract will readily allow the introduction the majority of circumstances, experienced personnel
of new technology (or removal of redundant items), will have to be transferred from existing projects an
especially when it would benefit all parties. approach more easily justifiable for long-term contracts
Material Efficiency: The efficient deployment comprising a significant scope of work.
of fracturing fluids and proppants can also have a Contractual Structure: There are numerous ways and
significant effect on treatment and project economics. means in which to set up a contract and remuneration
As discussed in Section 4-5, fracture treatments should structure for hydraulic fracturing operations, but Kugler
be designed for optimum CfD, based upon formation et al. (2004) and McIninch et al. (2002) note that the
characteristics. In order to obtain the optimum structure and management of these contracts (in mature
CfD, which implies the most efficient use of fluids field environments) is playing an increasingly important
and especially proppants, an accurate knowledge of role in the levels of success subsequently achieved.
formation permeability is essential. In a mature asset, 1. Responsibilities. It is important that a contract
it is usually possible to analyze well production history reflects the responsibilities and skill-sets that
in order to obtain a reasonably reliable estimate of k. both the operator and service provider bring to
This approach, combined with a prediction of post- the project. For example, it is inappropriate for a
treatment production conditions (allowing an accurate service provider to be held accountable for areas
estimate of proppant permeability requirements, kf, outside their direct control, e.g. height growth,
under actual in situ conditions) will allow the maximum water encroachment, etc. Conversely, it is essential
efficiency of treatment design. that they are held directly accountable for providing
In terms of direct inventory control, it is a reasonable service quality in terms of pumping
considered normal practice for the service provider to service, actual proppant schedule, additive rate,
carry the necessary material to deliver the campaign. etc., compared to a reasonable, pre-defined and
However, in some situations, such as a remote area or agreed set of acceptable tolerances.
where multiple service providers are utilized, it is not 2. Project Management. One of the fundamental
unusual for the operator to take control of and accept building blocks to implementing an efficient
the risk of such inventory control. This is especially and economic mature field re-development
true of proppant material, and such an approach can (Burnstad et al., 2004; and Kiburz and Sullivan,
occasionally provide benefits to all. 1995) is to establish an effective, decisive and
Personnel Requirements: Personnel requirements authoritative project management team at the
should be clearly defined. Although whenever possible earliest possible opportunity. Preferably, this team
local structure and content are desirable, this is not should be set up during the contracting phase and
always possible or practical from the standpoint of success should be empowered to implement whatever
through experience. Before tendering, the operator is necessary in order to ensure the programs
should be clear on the resource and skill-sets that they success. Led by either the operator (historically
can provide, and ensure that the remaining positions the more usual approach) or the contractor
will be provided by the service provider. In addition, (more recently), but including representation
both parties should ensure commitment to a project, from all parties, the roles of the team are wide-
with personnel, through all phases of contracting and ranging. These roles can include management
execution. The oft-applied policy of utilizing contract- and organization of numerous topics, such as
setup engineers for tendering and contracting phases, contractual issues, technical issues, operational
only to have these people to promptly disappear at QA/QC issues, reporting and tracking,
execution time, should be discouraged, as this will result logistical planning, accounting/invoicing issues,
in a lack of continuity and a loss of faith and belief in conflict resolution, HSE assurance and relations
the project as a whole. In practice, neither operating nor with third parties (e.g., local government
service companies keep experienced fracturing personnel agencies, national partners, etc.).
463
Modern Fracturing
3. Risk and Reward (gain-share). This is not a Whatever the final scheme implemented for the
fundamentally new approach to contracting, contract structure, the considerations and choices
(Comrie and Olsen, 1997), but it is undergoing undertaken during the early planning phase will
a resurgence of interest and finding increased potentially have a tremendous effect on the chances of
application with particular emphasis on the re- fracturing program success.
development of mature oil and gas fields. With
this approach, each party can reasonably calculate 13-2 Candidate Selection
potential minimum/maximum exposures and
minimum/maximum margins. For the operator, Opportunity is missed by most people because it is
this allows an opportunity to consider candidates dressed in overalls and looks like work.
that would potentially fail an internal NPV or Thomas Edison
ROI hurdle and, if successful, receive a portion of
otherwise potentially un-recovered reserves. For Smith (2006) best describes hydraulic fracturing
the service provider, this approach could create the as THE Multi-Disciplinary Technology, and if
necessary volume scope to ensure an economically any single issue within fracturing most accurately
efficient campaign and also to demonstrate faith epitomizes this, it would be the complex and multi-
in their own technology by assuming part of the faceted subject area of candidate selection. Successful
risk associated with field deployment. candidate selection for hydraulic fracturing purposes
Critical to the success of such an approach requires the consideration of all of the opportunities
is ensuring that all of the operator(s) partners available on three distinct scales, which
(both private and public) are engaged and agree we will define as follows:
with the decision-making process as the gain- Regionalized (Macro): This we will define as
share approach and scales are determined. It encompassing the generalization of areal well
should also be remembered that risk should be performance, reservoir structure/variation (as
commensurate with reward. understood), and any potential faulting, geological and
4. Volume Discounting. Should significant other larger-scale issues which will potentially affect a
uncertainty exist with the potential scope candidates potential productivity.
(and the breadth of values for P10/P50/P90 Neighborhood (Meso): This we will simply define
cover a significant range) then the application as being related to wells within the immediate vicinity
of a volume discount approach can assist both (i.e. immediate offsets within the drainage-block/
parties. The contract generally reads with the drilling pattern) and wells that are concurrent and
higher pricing as noted with the example under directly laterally correlated.
Volume Scope of Work; however should Localized (Micro): This we will very simply
upside exceed original scope, additional discount define as being both native and related to the
factors are applied on annual or quarterly immediate wellbore under consideration, over
invoicing to reflect the additional volumes. the vertical distribution and within the radius of
This approach ensures that the service provider influence of the wellbore.
recovers the small scope volume at the higher
cost; however, the upside scope is recovered by 13-2.1 Regional Considerations
the operator via a discount.
5. Service Alliances. Where multiple product The first level of consideration should be those issues
and service lines are going to be employed that can affect potential candidate performance and
within a development or re-development, candidate selection, on a regional basis:
creating Alliances is a common approach 1. Reservoir Heterogeneity: As noted by Meehan
considered (Gazi et al., 1995; and Farrell et al., (1988), most reservoirs demonstrate
and McDermott, 1995). some level of heterogeneity in a vertical
464
and an areal sense. The well-level detail 13-2.2 Neighborhood Considerations

will be taken into account when considering
localized behavior, but considering changes The next level of consideration is an assessment of the
in reservoir quality with regard to regional quality of the neighborhood and an examination
behavior ensures that any potential trends are of offset well behavior - in other words, looking for
reasonably well understood. poor producers in good parts of the field, or wells
2. Reservoir Continuity: Numerous mechanisms, that are bad producers but not fracture candidates.
on a macroscopic scale, can significantly affect Rather than generating volumes of data about each
fracture efficiency: discontinuities, fault blocks, well, this level should consider just a few important
imbricated formations, fractured crestal areas, bulk properties and values, such as:
compressed flanks, varying drive mechanisms, 1. Offset Well Performance: These are the classical
varying water contacts/perched water, stress color-contoured or bubble map and spot-pattern
rotations, tectonics, fault environment thrust/ variables we are accustomed to considering,
reverse/uplift, etc. It is important to have a based on such parameters as production rate,
thorough knowledge of any regional variations cumulative production, net pay, reservoir
in the formation geology. pressure, productivity indices (normalized),
3. Geographic Information Systems (GIS): One porosity, permeability, produced liquids, etc.
useful way, quickly gaining acceptance, to (see Chapter 12 for examples).
represent and examine large amounts of areally Judicious collection and review of this data
distributed information is the application of GIS can yield significant insight into the quality
systems. The geographical perspective can be of the location of the potential candidate
extremely insightful in terms of understanding well and potentially identify massively
reservoir and well performance. underperforming wells based on consideration
4. Gathering and Production (GAP): It is important of the offsets alone. This data should be visually
to consider the pipeline, gathering and satellite reviewed as well as recorded, in an appropriate
processing capability to produce a candidate format, into a structured database for the
within a region, if a candidate is successfully neighborhood-level information.
fracture stimulated. It is possible to discover that 2. Drainage Shape and Area: Visual inspection of the
due to a mechanical constraint, a well cannot bubble maps created when examining offset well
be initially flowed at full potential in one part performance now need to be considered with
of a field compared to another. Ideally, a nodal respect to preferred fracture direction and the
analysis model will provide guidelines about the drainage shape associated with the offset wells.
existence of any such regional limitations, and Cipolla et al. (2005) and Meehan et al. (1988)
the regional Mscf/d-value variation of a dollar provide excellent examples of the importance of
spent will be well understood. well location choice for infill and/or selection as
a fracturing candidate. This approach provides
Often, an assessment at this first primary level insight into poor locations for fracture stimulations
can either remove one or more entire sections of a and infill wells; conversely, it also provides insight
development from further consideration, or (more into harvest opportunities to recover additional
likely) rank areas of the field so that better regions/ reserves where un-drained areas exist within existing
areas will weigh higher under final consideration. drainage patterns. As noted in Section 13-5.3, the
A structured database should be constructed, post- industry is beginning to pay significantly more
assessment, that contains such weighting information attention to the importance of areal drainage.
to be utilized with a risk-analysis or ranking If the preferred fracture direction is not well
model after the full suite of candidate selection known, then there are numerous methods of
criteria have been considered. obtaining a reasonable determination of its value
465
Modern Fracturing
(Meehan et al., 1988). Occasionally, simply studying performing a candidate-selection process that there
trajectories of wells drilled on natural-walk (i.e. is confidence in the model(s) utilized to make
without any directional BHA) may indicate the these estimations. If this is not the case, then an
minimum stress directional trend. alternative approach should be taken when making
3. Areal Connectivity: These maps may also indicate the any comparative assessments or ranking.
magnitude of areal connectivity that exists among The definition and relative distribution of the
immediate offset wells. Good connectivity across pay (bypassed or existing), gross and net sands,
the neighborhood should be reflected in offset and an early assessment of the potential fracture
behaviors and pore pressure distribution. Poor geometry should also be assessed at this level.
connectivity, of course, would be demonstrated as 2. Pressure Transient Analysis: The localized reservoir
steep value contouring and/or a complete division parameters can be calculated by performing pressure
or isolation trend between two or more types of transient analysis (PTA) as described in Chapter 3.
well response(s) on the plot. However, obtaining this data may be complicated
4. Publicly Available Data: Data recovery should by multi-zone/multi-layer commingling issues
include the immediately operated wells as (especially in mature environments) and it is often
well as competitor wells (which may also be expensive data to recover (in terms of direct costs
offsets) if available. and deferred production).
When available, PTA can provide valuable
All of the neighborhood data available from offsets information regarding in situ reservoir properties
for each potential candidate should be recorded and such as average reservoir pressure, p*, reservoir
available within a structured database. In addition, permeability, k, reservoir level effects (e.g. faults,
for each well/entry a simple connectivity map to fractures, etc.) and the magnitude of skin-damage,
offsets should also be available, potentially via GIS, s, around a well (either induced during the initial
for later manipulation. drilling and completion phases or accumulated
during, and related to, the extensive production
13-2.3 Localized Considerations history of the well).
Whenever reasonably possible, pressure
For the purposes of candidate selection, we should transient data should be obtained and analyzed,
finally determine the localized and singular behavior in order to provide the most accurate assessment
of the candidate under consideration; not surprisingly, of local conditions.
this is the level at which most information generally 3. Production History Matching: The analysis of
exists. As noted in Section 13-1.3, the starting point production decline data and material balance
for an individual well candidate selection should be the approaches have been used for some time, in
performance of a well compared with a zero skin/100% a variety of forms (see Chapter 3), in order
flow-efficiency condition, given assumed/estimated to determine recoverable reserves numbers.
reservoir parameters at that location. However, these approaches often suffer from a
From the familiar equations, as defined in Chapter number of fundamental limitations; therefore,
2, we can appreciate that there are a number of formation during the last thirty years, type-curve analysis,
and reservoir parameters with which we must be familiar using methods established by Carter (1985) and
and comfortable if we are going to be able to accurately Palacio and Blasingame (1993) have enjoyed
predict pre- and post-fracture well behavior: a more widespread application in gas wells, as
1. Reservoir Characteristics: The utilisation of log- noted by Agarwal et al. (1998).
derived permeability, from porosity estimations, Agarwal et al. (1998) also confirmed the findings
forms one of the cornerstones of accurately of Palacio and Blasingame (1993) that the constant
estimating a wells potential deliverability (Barba rate and constant BHFP solutions for gas systems
and Shook, 2004). It is therefore imperative when could be converted to an equivalent constant rate
466
solution for liquids. Having demonstrated this, they pipe-weight/grade/connection requirements for
then went on to show that classical constant-rate fracturing. However, the costs associated with
liquid solutions for PTA can then be applied to gas upgrading the integrity from surface to TD
well conditions and described the three principal for the particular case of fracturing can often
curve types they considered: offset the savings associated with a monobore
Rate-Time Production-Decline Type approach; therefore, it is often disregarded or
Curves: These curves have application to the overlooked. In addition, because many wells
estimation of GIP (gas in place), reservoir are not immediately considered as fracture
permeability and skin factor. candidates (as per the subject matter of this
Rate-Cumulative Production-Decline Type Chapter) and therefore are not appropriately
Curves: These curves have application in the designed, limited mechanical integrity can play
estimation of GIP, ensuring GIP accuracy a very important role in defining the potential
and flow-period identification (transient future fracability of a candidate.
and pseudo-steady state).
Cumulative-Time Production-Decline Type All of the data available for each potential
Curves: These curves are useful when cumulative candidate, from a localized perspective, should be
production data provides a smoother data set than recorded into the structured database. In addition,
rate production data and can be utilized to calculate for each of these parameters, upper and lower
reservoir parameters and reserves data. bounds should be selected if appropriate, and a data-
4. Mechanical Integrity: This may seem an obvious accuracy measure calculated.
consideration when selecting candidates for
hydraulic fracturing, however either poor initial 13-2.4 Risk Ranking and Data Manipulation
or loss of mechanical integrity with time, has often
obscured or over-complicated the upside potential The principal objective, after having completed
of fracture campaigns (Solares, 2001; and Rylance the exercise above, would be to have access to a
et al., 2007). The fracturing industry is littered structured candidate database(s) with two distinct
with the corpses of treatments that have failed due sets of entries: performance-related candidate data
to poor mechanical integrity, not always because and risk-related candidate data. Once these two
the integrity was not considered, but more often integrated data sets are available, the process of
because the magnitude of the effect of losing candidate selection may begin.
integrity was considerably underestimated. 1. Risk Ranking: Prior to performing any candidate
There are a number of reasons why mechanical ranking exercise, it is important to determine the
integrity should be of such concern, and many risk valued-drivers to the project.
of these are discussed in Chapter 5, including Each application may have any number of
the importance of tubing-stress analysis and drivers, each of which will have a relative level
primary cementing isolation. For example, of importance, depending upon the particular
during recent years there has been a growing circumstances. For example, drivers might
trend to complete wells with much simpler include pre-fracture production rate (i.e., risked
monobore-style completions with casing/ production), post-fracture production rate,
tubing cemented back to surface. Although increased reserves, $/Mscf, ROI, NPV, mechanical
this reduces the time, cost and complexity of a complexity, CAPEX limitations, delivery-time,
well, it generally removes the ability to deploy etc. The relative importance of a number of these
a fracturing string and hence exposes the entire factors must be appreciated prior to performing a
completion to fracturing screen-out pressures. candidate selection process, as they will provide
This is manageable, of course, (Macfarlane and the individual skins for the ranking mechanism
Mackey, 1998) by careful consideration of the during the candidate selection process.
467
Modern Fracturing
2. Data Manipulation: As discussed in Section 13-2.5 Case Histories and Results

12-5, in many cases, the oil and gas industry
will have a strong tendency to manipulate Silver bullets are for Hammer Horror Films and not
these databases (or more likely spreadsheets), fracture candidate selection. With fracture candidate
with conventional tools that will be limited selection there is no one-size-fits-all approach as a
to and constrained by the experience and pre- process and certainly no single consideration that
conceived notions/expectations of the person provides the 100% solution. This is just as true with
who is performing the assessment. candidate selection for re-fracturing (see Section 13-
As a simple first-pass approach to a fracturing 5.2) as it is for initial fracturing.
candidate selection exercise, this is not an However, not all is doom and gloom. Clearly from
unreasonable approach. However the availability the previous chapters of this book we can see that
of more modern methods offers the opportunity there are a variety of different approaches to choose
to examine the data, and their relationships, from or apply and it is the selection or combination of
in other ways and to maximize the concealed approaches which will provide the particular solution in
potential. There are a number of different each case. We will now consider a number of different
emerging artificial intelligence approaches that cases, where appropriate but increasingly complex,
have been applied in this way in recent years solutions have been applied.
(Mohaghegh et al., 2000) to the manipulation of 1. Candidate selection with no or limited available data:
typical fracturing data sets. These approaches are An extreme situation, where there are numerous
quite varied in their requirements and their upside unknowns and an effective candidate selection
potential (Oberwinkler et al., 2004, and Stundner method is extremely difficult to achieve. However,
and Oberwinkler, 2004), and include approaches Utegalyev et al. (2006) demonstrated that by using a
such as fuzzy logic, neural networks, genetic multi-disciplinary team, creating and implementing
algorithms, self-questioning matrices, pattern a learning culture, and concentrating on variables
recognition, self-organizing maps, etc. that could be directly affected (e.g,. the operational
It should be acknowledged that these aspects of the process), they were able to perform a
applications need to be applied with caution, highly successful stimulation campaign.
as there are a number of preparatory steps 2. Candidate selection using mixed conventional/
that can affect the quality of the outcome advanced approach: Grieser et al. (2006) describe a
and their usefulness, such as: case whereby a conventional approach, including
Inadvertent/subliminal filtering of the data prior cross-plots, bar-charts, cumulative frequency
to database entry (i.e. pre-determined conceptions plots, etc., did not determine any definitive
related to what are or are not the important trends or correlations. They then proceeded to
variables to be considered, thereby circumventing perform a more detailed data-mining exercise
a neutral data gathering process). and applied the Self-Organizing Map approach
Poor-quality, spurious, inaccurate or wholesale (see Section 12-5.5). The results were clear, with
absence of data will also render a number of definitive trends, which could then be applied to
these approaches potentially ineffective. Like all improve the future program.
calculations, the dictate that garbage-in = garbage- 3. Candidate selection using advanced approaches:
out should not be overlooked. Although performed in order to primarily
identify re-fracturing candidate potential, the
However, it has been increasingly reported that using work reported by Oberwinkler and Economides
these methods with other selection techniques, such as (2003) is an excellent demonstration of the
type-curve analysis and economic risk-ranking, provides measure of application that can be achieved
yet another level of confidence and a robust tracking with the artificial intelligence approaches
method for the candidate selection process. when quality data is available.
468
13-3 Fracture Design in Mature Fields can be given lower priority; conversely, wells with higher
or stable pressures, lower water cuts and/or higher skin
Those who cannot remember the past are factors can be given higher priority.
condemned to repeat it. Finally, reservoir engineers will often have a significant
George Santayana, The Life of Reason, period of production history data with which to fine tune
Volume 1, 1905 the reservoir simulation model and understanding. This
will allow for the potential identification of by-passed
There are a wide variety of treatment options reserves and allow the wells to be ranked according to
available for mature reservoirs in fact just as remaining recoverable reserves. Obviously, the wells
many as there are for treating new reservoirs. All of with the greatest remaining reserves are better candidates
the conventional choices are at the users disposal; for production if these reserves can be accessed by
however, a number of key issues often make mature fracturing. For instance, a well with significant reserves
field fracturing an effective proposition. located in a direction perpendicular to the fracture
Primarily, there is often significantly much azimuth will not see greatly improved access to these
more information and history available regarding reserves from a fracture treatment. Such a situation
a mature asset by its very definition than there is may require an infill well to be drilled. Conversely, an
about a potential new development. This data can accurate knowledge of the fracture azimuth can help
be comprehensive and include much insight into refine candidate selection and treatment design because
permeability and pressure variation and distribution, knowledge of the optimum dimensionless fracture half-
rock mechanical properties and the tendency of length (xf/re) is a crucial design parameter.
reservoir fluids to produce/promote organic and Perforations and Workovers: Mature wellbores
inorganic scale deposition when contacted by foreign will potentially have multiple perforated intervals
fluids, among other things. All of this information and completions in place; often, these will have a far
can be used to make fracturing treatments much more significant effect on a treatment design than any
more effective. However, more than anything else, formation characteristics. Usually, the well will require
an improved knowledge of reservoir permeability workover and re-perforating as part of the preparation,
and variation can dramatically improve the results and this is discussed in more detail in Section 13-4.5. This
and efficiency obtained from fracturing operations, preparatory work will provide an opportunity to achieve
allowing each fracture to be designed to produce the significant flexibility and optimization because now the
optimum CfD (see Section 4-5). treatments can target specific intervals. In addition, re-
Potentially, there may also be an available wealth perforating presents the potential to manage and control
of information from previous hydraulic fracturing the point of fracture initiation (see Section 6-2.1).
operations within the field. Important parameters Fracture Treatment Design: In general, the same
such as in situ stress magnitude and variation, leak- rules and procedures should be followed for mature
off behavior and rock-mechanical properties may field fracture design as for any development (see
already have been recorded from treatments in offset Section 4-5). However, there are some specific points
wells. At the very least, archives from such disciplines that will need to be considered:
as drilling can be examined to provide potential 1. In many formations it is very important to
bounding information such as leak-off tests (LOTs) consider the potential movement of the gas/water
and formation integrity tests (FITs), etc. contact(s). Open-hole wireline logs indicating
The ability to efficiently select individual wellbores these will probably have been run and processed
and candidates for zonal fracturing will also be much for original fluid contacts when the well was first
improved, as once again, significantly more data is drilled. It is not uncommon, in a depleted gas
available to assist with the selection process. Wells (or reservoir, for the GWC to have migrated upward,
blocks of the field) with lower reservoir pressure, higher and it may be worthwhile to consider performing
water cuts and/or a history of fines or scale production some form of surveillance in order to re-determine
469
Modern Fracturing
the location of these contact(s) before even selecting In these cases, when creating the fracture design,
fracturing as a potential re-development method. any uncertainty related to design half-length
Ideally, fracture treatments in gas reservoirs should and volumes of proppant should err on the side
be designed to avoid penetrating any rock with of caution. In brief, it is worthwhile to slightly
significant water saturation. Although there over-design the frac half-length and proppant
are techniques available to help mitigate water volumes in order to compensate for any potential
production if this occurs, these approaches will underestimation of requirements.
only potentially reduce the water production levels
and not eliminate the problem entirely (Borodin, 13-4 Depletion Considerations
et al., 2006, and Pietrak et al., 2005).
2. Ensure that the positive and negative aspects The weak can never forgive. Forgiveness is the
of pressure depletion have been thoroughly attribute of the strong.
considered for each candidate. This is discussed in Mohandas (Mahatma) Gandhi, Indian Political
further detail within Section 13-4. Leader (1869 1948)
3. Consider the potential economics of limiting
the fracturing operations to merely skin by-pass Probably the single most important consideration
fracturing treatments (see Section 4-7.5). Mature when fracturing in a mature environment will be
reservoir intervals can often have quite high skin the reduction in pore pressure that has taken place
factors associated with them; an assessment of since the onset of production, and the influence
these skins may show that these zones can be that this will have on any subsequent operations.
effectively stimulated with relatively small fracture The following sections consider some important
treatments (Rae et al., 1999). factors this reduction will affect.
4. As with water contacts, be aware of the potential
effect that the onset of water production may 13-4.1 Pore-Pressure Considerations
have on the overall economics. Excessive water
production can result in numerous negative Changes in Closure-Stress: Hubbert and Willis
effects, including: (a) A decrease in the effective (1957) demonstrated that fracture pressure gradient
proppant pack permeability (Vincent et al., 1999), was a function of overburden stress gradient, pore
(b) Formation deconsolidation and failure, (c) pressure and stress ratio, accepting a number of
Formation and deposition of inorganic scales, (d) generalized simplifications. They proposed that the
Excessive proppant back-production and (e) A in situ horizontal stress is generated solely by the
major effect on the vertical lift performance of the uniaxial vertical strain (assuming no tectonic input),
completion and/or surface gathering system. and that the horizontal stress can be expressed as a
5. Wherever possible, fracture treatments should be function of the overburden by:
designed to provide the maximum areal recovery

from the field. This will be achieved by taking c = (v pr ) + pr , (13-3)
(1 )
consideration of (a) The effects of the dominant
fracture azimuth, (b) The placement of longer where c is closure stress, is Poisson's ratio, v
fractures where wells have greater separation and is vertical stress, is Biot's poroelastic constant
vice versa and (c) Ensuring that un-depleted areas and po is wellbore pressure. As the pore pressure
of the field receive appropriate attention. decreases from an initial state, the closure stress
6. It is also worth considering, when finalizing the will drop proportionately. Differentiating Eq. 13-
fracture design, that this may be the last significant 3 with respect to the pore pressure (and assuming a
intervention within this well. This is especially true Biots constant of 1) gives the following:
of remote or offshore locations, where preparation,
mobilization and logistics dominate the economics. c (13-4)
= 1 .
pr (1 )
470
For a Poissons ratio of 0.25, this will mean that a 1 As the pore-pressure reduces, the in-situ stress in the
psi reduction in pore pressure will reduce the localized sand reduces which will increase the equivalent stress
closure stress by 2/3 psi. This localized reduction in barrier between the shale and sand. This increase in the
pore pressure and closure stress can have a dramatic stress barrier results in a more readily achievable fracture
influence on fracture geometry and efficiency. half-length (for the same fluid efficiency or fluid-volume).
One of the most important aspects of a change This half-length xf should then be modelled for a range of
in the closure stress profile is the manner in which barrier magnitudes to generate an equivalent frac Half-
this change will affect the growth, geometry and Length Graph, the shape of the xf vs. pore-pressure decline
therefore deliverability of a hydraulic fracture graph (i.e. straight line, curve-up or curve to asymptotic)
treatment. In situations where fracturing attempts, will be determined by the actual sand-thickness, initial
early in field life, resulted in a lack of confinement barrier value, leak-off parameters etc.
and poor fracture aspect ratio, a reduction in pore Once this simple graphic has been generated
pressure (and hence closure stress) can result in it may then be utilised to help maximise the
improved fracture geometry based on the change efficiency, timing and selection of the proposed
in the stress profile and potentially improved fracture/re-fracture candidates.
contrasts with any shale-barriers. Conversely, any uneven depletion within a laminated
When considering the impact of pore-pressure sand-body may result in preferential proppant placement
depletion on fracture geometry, it may be worthwhile across the depleted sands at the expense of reserves
constructing a simple graphic (similar to the one provided recovery within higher pore-pressure regions.
in Figure 13-6), based on the available knowledge of In mature fields, it is therefore always important
the formation stresses, pore-pressure depletion and to consider how such a pore-pressure reduction
previous formation fracturing experience. will affect fracture geometry. This is particularly
true if earlier attempts at hydraulic fracturing were
2500 7000
unsuccessful due largely to lack of confinement and
Stress Contrast, psi and
Fracture Half-Length, ft
2000 6500 hence poor fracture aspect ratios.

Closure Stress, psi
1500 6000
Asymmetric Fracture Growth/Fracture Reorientation:
The previous section briefly considered the effects of
1000 5500
pore-pressure reduction within a single payzone regarding
500 5000 fracture confinement (i.e. one-dimensional consideration)
0 4500
or if preferential depletion is present say between two-
0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 distinct pay-zones (then this could be considered as two-
Pore-Pressure Reduction with Production, psi/ft dimensional). However, such depletion also has an areal
Closure Stress in sand
Sand-Shale Stress Contrast
Upper Bound Half-Length Potential
Lower Bound Half-Length Potential
or three-dimensional component as described in several
case studies, such as those reported by Mukherjee et al.
Figure 13-6 Example of the impact of pore-pressure (1995) and Cipolla et al. (2005).
reduction on stress, stress-contrast and achievable
To a certain extent, the fracturing community has
fracture half-length xf potential.
largely overlooked or disregarded this areal effect, due
As in Section 13-4.1, assuming a Poisson ratio mostly to a lack of diagnostic evidence. However, more
of 0.25 then we can assume that for every 1 psi recently, fracture re-orientation and asymmetric fracture
reduction in reservoir pore-pressure, that we have 2/3 growth have been more easily demonstrable through the
psi reduction in stress within the main sand body. increased gathering of micro-seismic and tilt-meter data
We can also assume that the shale (pore-pressure) (Cipolla and Wight, 2002). The gathering of this data has
does not reduce equally or that any reduction in been largely driven by the increased activity within the
pore-pressure (and hence stress) is offset by an sphere of unconventional gas (described within Chapter
increase due to the hanging wall effect as applied 11) and re-fracturing operations (which will be described
by von Winterfeld et al., (2005). further within Section 13-5.4).
471
Modern Fracturing
Stress Dependent Permeability: An additional 2. Minimization of broken-fluid surface tension:

consideration when planning fracturing operations This is generally achieved with a specially
in mature fields is the potential for formation designed surfactant system. These systems are
deliverability to have actually reduced over time due usually blended from several surfactant-based
to a pressure-dependent permeability effect (Vairogs chemicals and can consequently be optimized
et al., 1971). A significant amount of work has been for other purposes as well, such as promoting
performed in this area over the years, a large part of desirable wettability characteristics, preventing
which has indicated a bias to a more pronounced effect the formation of emulsions and controlling clay
in lower-permeability environments and reservoirs instability. Adding methanol can also reduce the
with distinct natural fracturing. broken-fluid surface tension.
3. Minimization of fracture fluid leak-off: Improved
13-4.2 Fracturing Fluid Selection overall fluid efficiency will of course reduce the
total volume of treatment fluid to be returned.
The selection of fracturing fluids for gas-reservoirs is dealt In addition, an efficient filter-cake deposition
with extensively in Chapter 7 of this book. However, some will ensure that the majority of this leaked-off
aspects specific to fracturing mature and low-pressure gas fluid is actually fracturing fluid filtrate. This is
reservoirs will be discussed here in more detail. an extremely desirable outcome because whole
The success of hydraulic fracturing treatments within gel loss to the formation should be minimized if
low-pressure reservoirs can be heavily influenced by the possible (for numerous reasons), especially in a
efficiency of recovering the fracturing fluid after the low-permeability environment.
treatment. This recovery consists of two distinct phases: 4. Minimization of base water volume: This can
The first is the process of returning broken fracturing fluid be readily achieved by replacing some or all of
and gel residue out of the formation and proppant pack the base water with methanol, which has little
and into the wellbore; the second consists of returning or no interaction with most reservoirs and
this fluid back to the surface, where it can be disposed of, hence is relatively easily recovered, often in the
and this subject will be dealt with in Section 13-4.4. form of a vapor. Another way to minimize the
The principal additional consideration to enable rapid water content is to replace part of the aqueous
and efficient fluid recovery from depleted formations, phase with CO2 or N2 (or both). The use of a
over and above normally pressured formations, is that the foamed fracturing fluid (generally accepted
return fluid must have the maximum possible mobility to be effective above 60% by volume of the
ratio. This equates to ensuring that the minimum possible gaseous phase at downhole conditions) can have
energy is expended to ensure efficient fluid recovery. several advantages for low-pressure reservoirs.
Consequently, the guidelines for fracturing fluids for These advantages include:
depleted formations are very similar to those for very Dramatically reduced fluid leak-off, and hence a
low-permeability formations: reduction in the overall fluid volume required to
1. Minimization of broken-fluid viscosity: Clearly, place the same amount of proppant
the more viscous the broken fracturing fluid and Reduced water percentage within the foamed
filtrate, the more energy required to initiate flow slurry (e.g. 100 bbl of downhole volume will
within the fracture and the formation. Minimizing contain only 30 to 40 bbl of water-based
broken fracturing fluid viscosity can be achieved by fluids in quality foams)
a variety of approaches, such as reducing polymer Potentially improved proppant transport
loading, eliminating cross-linked/complex polymers capability (leading to a reduction in the overall
altogether (e.g. using slickwater treatments or required polymer loadings)
viscoelastic surfactant systems) and using efficient Easier fluid recovery because the energy stored
and multiple breaker mechanisms, such as polymer- within the gaseous phase provides a boost to the
specific enzymes (Ely et al., 2003). localized pore pressure
472
However, using foam does also have potential declines in reservoir pressure during the life-of-
disadvantages, including: field; consequently, there is always the possibility
Increased surface treating pressure caused by of an initial increase in the effective stress on the
increased overall wellbore friction loss and a loss of proppant as the field undergoes depletion (McLennan
hydrostatic head and Roegiers, 1982).
Increased overall treatment costs, a factor that can
8,000
be critical in the relatively sensitive economics
7,000
associated with mature developments
6,000
Increased overall treatment complexity
5,000
Pressure, psi
4,000
5. Elimination of aqueous phase: A number of 3,000
operators have successfully fractured low-pressure 2,000
gas formations using water-free fracturing systems 1,000
based on methanol, oil (such as condensate, 0
1 7 13 19 25 31 37 43 49 55 61 67
diesel, kerosene or frac oil) and even 100%
Elapsed Time, Months
CO2 (e.g. Craft, et al., 1992; Hernandez et al.,
1994; Mazza, 2001; Gupta, 2003; and Cramer Pore-Pressure Decline
Drawdown Case A
et al., 1991). Such fluid systems can be extremely Effective Stress Case A
Drawdown Case B
formation-friendly and very easy to recover, and Effective Stress Case B
Drawdown Case C
consequently are becoming increasingly popular. Effective Stress Case C
Drawdown Case D
Effective Stress Case D
However, there are generally significant cost
and availability issues to be overcome, as well
as safety issues related to pumping flammable Figure 13-7 Change in effective stress on proppant with
pore-pressure decline and drawdown
and low-flash-point fluids. In addition to
this, oil-based fracturing fluids should not be This is more easily explained with reference to
utilized in gas formations containing residual Fig. 13-7, which demonstrates the effective stress on
oil saturation (i.e. wet gas reservoirs) proppant for a range of drawdowns - in a formation at
because the increase in liquid hydrocarbon 10,000 ft with an initial stress gradient of 0.70 psi/ft
saturation level can lead to oil-wetting and and a pore-pressure gradient of 0.43 psi/ft, undergoing
hydrocarbon phase trapping. depletion over four years (with a minimum bottomhole
flowing pressure of 150 psi). As in the discussion within
13-4.3 Proppant Selection Section 13-4.1, we have assumed here a Biots constant
of 1 and a Poissons ratio of 0.25.
Chapter 8 deals extensively with proppant selection Conversely, when the drawdown is limited by
and provides detailed criteria for selecting proppants. the available pr and subsequently begins to decline
However, once again in mature reservoirs, there are some (see Fig. 13-6), the effective stress will also begin to
specific details that will need to be considered. decrease. Pathak et al., (2004) described the effects
Change in Effective Stress: As discussed in Section on the giant Indonesian Arun gas field as reservoir
4-3.2, a decline in reservoir pressure can significantly pressure declined from an initial pressure of over
affect the effective stress experienced by the proppant 7,000 psi in 1971 to just below 700 psi. This reduction
under producing conditions. During decline, produced such a massive change in formation stresses
while the drawdown (pr po) remains unchanged, that the field experienced numerous wellbore failures.
the effective stress on the proppant will gradually Obviously, this is an extreme example; however, it
increase until the drawdown drops due to the limited does demonstrate that the expected reservoir depletion
available bottomhole pressure. In gas reservoirs, it should be a crucial consideration in proppant selection
is possible for a formation to experience very large for life-of-field fracture behavior.
473
Modern Fracturing
Although a decrease in effective stress might possibly, reduced temperature) can result in the
appear as though it would benefit proppant subsequent deposition of inorganic scales from
permeability, this is not necessarily always the case. formation water and in some very isolated instances
To start with, the reduction in proppant permeability organic scales from condensates.
caused by high initial effective stress is primarily due Decreasing Drawdown: As the reservoir pressure
to fines generated within the proppant pack (due to declines through maturity, the drawdown and
large localized contact stresses between individual productivity will also follow an eventual declining
proppant grains). These fines will not disappear as trend. This reduction in drawdown and productivity
the effective stress drops, and so the proppant will will mean that the velocity profile of the gas within
not regain significant permeability. In addition, as the the fracture will diminish, and various material(s)
effective stress drops, so does the stable-arch force that may have previously been produced efficiently
holding the proppant grains together; this can lead through the proppant pack (such as liquids and
to severe and abrupt proppant flow-back problems, fines) may begin to build up and cause a reduction in
especially if a liquid phase is also present. proppant pack conductivity.
It is worth noting that the potential worst-case Proppant Selection for Declining Pore Pressure: All of
scenario can often arise immediately after fracture the factors noted above mean that the proppant pack
stimulation, during the clean-out and flow-back stage must be designed for the conductivity and strength
(see Section 13-4.4). It is not uncommon to see wells that will be faced during the life of field. The peak
being blown dry during this phase because of poor requirements of the proppant pack can often be in
inflow. This is an extremely undesirable situation excess of the original static needs for efficient initial
because if the BHFP is zero, then the effective stress production of the reservoir. This is supported by the
on the proppant is at a maximum (near the wellbore). numerous re-fracture case histories (e.g., Bakhtar et al.,
Low-pressure, low-permeability formations must 1985; and Mukherjee et al., 2000) that have exhibited
be treated with patience and care during the clean- significant benefits simply from utilizing improved
out phase, in order to ensure that the near-wellbore proppant systems. In short, when selecting proppant
proppant pack is not permanently damaged. material for fracturing, the transient production
Increasing Liquid Production: As demonstrated by conditions of the reservoir should be considered - not
Vincent et al. (1999), multiphase flow can have just just the initial production state.
as significant an effect within a proppant pack as it
can within a formation. Propped hydraulic fractures 13-4.4 Cleanout and Flowback
placed initially within dry gas reservoirs can often
lose the majority of their effective conductivity due With any fracturing operation, the wellbore cleanout and
to retrograde condensation as the reservoir pressure fluid recovery phase is an essential part of the fracturing
declines, or due to water production as the reduced process. Indeed, the cleanout should be considered
reservoir pressure alters the gas-water contact as vital as pumping the pad or including the breaker.
or both. Proppant material should therefore Fracturing operations should not be commenced unless
be selected (and the fractures designed) with this the facilities and capabilities are in place to flow the well
in mind, so that a fracture can remain effective back with immediate effect. Moreover, the equipment
throughout the life-of-field. and personnel required for the flow back of the well
Increasing liquid production will also have should be rigged up and ready to perform prior to
other profound effects on the proppant pack. For fracturing execution because a premature screen-out
example, higher liquid production will increase may require rapid recovery of wellbore fluids.
drag forces within the proppant pack, increasing In order for a propped fracture to be present, a
the potential for proppant flowback (see Section new volume has to be created within the formation or
8-10). In addition, increased liquids production, formations to provide the additional capacity for the
coupled with depleted reservoir pressure (and also, fracturing fluid and proppant that will be pumped within
474
the interval. This volume is created by compressing the time a well is shut in during fracturing fluid recovery,
rocks, fluids and gases within the formation immediately it can become increasingly difficult to initiate fluid
adjacent to the fracture face. This region of compressed recovery. This is particularly true if gas is also being
formation represents significant accumulated energy recovered with the fracturing fluid, as the multi-phase
that can be used to recover the fracturing fluids and flow effects within the formation will tend to decrease
additives. However, this energy will rapidly disperse the effective permeability to both phases.
after shutdown, and it is therefore important to make It should also be noted that additives incorporated
maximum use of this energy by flowing back as soon as within fracturing fluids to prevent formation damage
possible after the treatment. such as non-emulsifiers, clay stabilizers and low
In depleted reservoirs cleanout of the wellbore surface-tension additives will have a limited life
and recovery of the fracturing fluid becomes even and are in general designed for a few hours exposure
more important because there is less inherent energy to formation conditions, rather than days or weeks.
available to recover the fluid. In general, the following The longer the fracturing fluids are left within the
practices should be observed: formation, the less effective these additives will
Immediate Flowback (Pre-Closure): There are certain be and the greater the potential for permanent
situations, such as very low-permeability formations formation or fracture damage.
(where there is significant elapsed time to fracture Managed Flowback: The well should be flowed
closure) or adverse fracture geometry situations (where back as rapidly as can practically and safely be
there is potential for proppant convection away from achieved, with the available surface equipment and
the pay zone and wellbore), when consideration should drawdown. Application of a maximum drawdown
be given to immediate flowback or forced closure, as ensures the most efficient use of the stored energy
defined by Ely et al. (1990). However, it is generally within the formation. However, too much draw-down
accepted that forced closure will lead to reduced can create other issues (e.g. liquid loading or proppant
fracture width, and therefore a reduced conductivity, crushing); therefore, a reasonable compromise must
immediately at the wellbore and that consequently be reached that allows the maximum continuous
this approach has a limited range of application and drawdown, achieving the maximum flow-rate while
should be considered carefully. minimizing proppant pack damage. Generally, the
Immediate Flowback (Post-Closure): Turn the most common practice appears to be setting the
well around and begin the flowback process as soon choke to produce at a high flow rate while being
as possible after the treatment has been shut down prepared to adjust the choke size to maintain the rate
(but after fracture closure has been confirmed/noted). as the producing conditions vary.
Design the fluids to break quickly, and take fluid Another important aspect of flowback is the
samples during the treatment for breaker testing. prevention of early gas breakthrough within the
As soon as the last proppant-carrying stage has proppant pack or at the wellbore, as ideally the
broken (and the fracture has closed on the proppant fracturing fluid would be produced first, followed
pack), the flowback can begin. This can be assisted by the gas. Although early gas production can be
by ramping up the breaker toward the end of somewhat helpful in keeping the well flowing, it can
the treatment, providing a faster break in the later also result in a permeability reduction due to multi-
fluid stages than in the earlier ones. Note that it is phase flow effects and the by-passing of fracturing fluid.
desirable to have the fracture close prior to the fluid Sherman and Holditch (1991) have demonstrated that
stage breaking, in order to avoid excessive proppant cumulative fracturing fluid recovery was significantly
settling away from the wellbore interface. affected by early gas breakthrough and that wells with
Continuous Flowback: After the fracturing fluids lower fluid recovery ultimately produced less gas.
have begun producing back toward the wellbore, it is Continuous, uninterrupted and managed flowback
very important to keep the fluids moving and not to conditions will help promote maximum fluid recovery
halt the process until the well is fully cleaned up. Each and minimize early gas breakthrough.
475
Modern Fracturing
In general, this means that the choke should be (Spady et al., 1999; McConkey and Stromquist,
adjusted continuously to maximize the fluid flow rate. 2000; and Rodvelt et al., 2001). Gavin (2000)
When the well is flowing 100% liquids, the choke can described the process and benefits of using coiled
usually be opened further to increase the liquid rate. tubing to fracture shallow gas zones, and Cawiezel et
However, after gas starts breaking through, opening the al. (2004) described the fracturing fluid requirements
choke further will often result in an increased gas cut and for successful coiled tubing fracturing.
reduced liquid rate. In such a situation it may be more Coiled Tubing Fracturing: Recently, there have
beneficial to actually reduce the choke size. been a number of tremendous advances in coiled
Uniform Flowback: As with any fractured well, tubing fracturing technology (see Section 9-5),
any sudden large changes in the drawdown should in addition to the open-hole fracturing technique
be avoided; this is especially true during post-fracture described in Section 13-1.4. These techniques, which
clean out and flow-back. Vreeburg et al. (1994) are mainly designed for cased and cemented wellbores,
demonstrated that stress cycling of proppant packs can be classified into four main groups:
could lead to aggravated proppant flowback. Sudden 1. Isolation Packer, Pump down CT: This is the
changes in drawdown, such as when rapidly opening conventional coiled tubing fracturing technique
or closing a tree valve, can produce sudden changes in that has been used with so much success in
pore pressure, effective stress and drag forces within the southern Alberta and numerous parts of the USA
formation around the fracture. Therefore, whenever (Spady et al., 1999; McConkey and Stromquist,
possible, choke sizes should be staged slowly up or 2000; and Rodvelt et al., 2001). This technique is
down as required, in order to minimize the overall described in detail in Section 13-1.4.
effects on the proppant pack. 2. No Packer, Sand Jet, Pump down Annulus: With
Volume of Flowback: As discussed in Section 13-3.1, this technique, a small-diameter coiled tubing
designing the treatment with the minimum necessary string (typically 1 or 1 in. OD) is used with
fluid volume is essential for treating low pore-pressure a sand jetting tool because initially, the well is
formations. The less fluid that is pumped into the not perforated. The sand jet is used to cut a set
formation, the easier it is to recover. of perforations or slots in the lowermost interval.
Coiled Tubing Flowback: Coiled tubing is often After this has been completed, the treatment is
used for nitrogen-assisting gas wells back to production then performed through the CT-wellbore annulus,
after a fracture treatment. The coiled tubing is run to an allowing the coiled tubing string to be used as a
optimally calculated depth, and a gas (typically nitrogen) dead leg in order to monitor a live BHTP. After
is pumped into the wellbore, reducing the hydrostatic the treatment is flushed, a sand plug is set over
head. Coiled tubing can also be used to clean out the the holes/slots, and the jetting tool is moved up to
wellbore sump/rat-hole and remove excessive proppant the next interval. The process can be repeated as
from a wellbore (e.g. after a premature screen-out) or from many times as necessary, moving up the wellbore,
sand plugs used to isolate multiple intervals. Therefore it although the nozzles on the jetting tool have to
is generally common practice, in many areas, for a coiled be replaced (typically) every 5 or 6 intervals. After
tubing unit to follow a frac spread around a field, cleaning all of the intervals have been treated, the coiled
up wells and returning them to production. tubing is then used to circulate the sand plugs
and excess proppant from the wellbore. If a post-
13-4.5 Mechanical Deployment treatment workover is necessary (for example, in
order to complete the well with tubing), drillable
As noted in Section 13-1.4 coiled tubing has been or inflatable bridge plugs can be set above the
used for a number of years now to successfully fracture top interval in order to provide a temporary
the shallow gas wells of southern Alberta as well as isolation. Fig, 13-8 shows a typical hole cut
the Appalachians; in addition this approach has been with the sand jet system and Fig, 13-9 shows a
extensively applied to coalbed methane recovery typical nozzle after four intervals.
476
4. Packer, Selective Perforating, Pump down Annulus:

This technique is very similar to the one
above, but instead of using a sand-jetting tool
positioned above the packer, selective perforating
guns are run below the packer (Tolman et al.,
2006). These guns are usually controlled via an
electric line contained within the coiled tubing;
at the required depth a controlled number of
the perforating charges can be fired. The coiled
tubing is then lowered so that the packer can
be set below the perforations and the treatment
performed through the CT-wellbore annulus.
The fundamental difference from the technique
described in the previous approach is that there
Figure 13-8 Typical hole cut by a sand jetting tool (3
bpm, 15 seconds). Photo courtesy of BJ Services, 2007. is a practical limit as to how many intervals can
be effectively perforated in a single run.
Multi-Zone Intervals: Coiled tubing fracturing

offers numerous advantages, not the least of which
is the speed at which multiple intervals can be
treated (20+ in a single day). However, its use is only
economically justifiable in areas where a) The wells are
reasonably shallow (usually less than approximately
5,000 ft); b) Operations are very time critical; and c)
Workover rigs are expensive or not readily available.
In many areas, these conditions are not met; for
instance, outside of North America it is usual to have
the fracturing equipment on contract, rather than
on call-out, and so the operator is not paying for the
Figure 13-9 Typical sand jetting nozzle after cutting four crew and equipment on an hourly basis.
intervals. Photo courtesy of BJ Services, 2007.
In the majority of situations, a mature field
re-development involving hydraulic fracturing
3. Packer, Sand Jet, Pump down Annulus: With this will require some kind of well intervention (see
technique, perforations or slots are cut with a sand Chapter 5 as well). Potential reasons for this will
jet as described above. However, with this particular include the removal of a fragile completion that
approach, a packer is deployed on the coiled tubing cannot withstand the tubing forces associated with
below the sand jetting tool (Tolman, et al., 2006). fracturing, removal or isolation of sub-surface
The first (and lowermost) interval is then treated as control valves, removal of downhole pumps, fishing
before. After the holes have been cut for the second of debris, wellbore cleanout, re-perforating and the
interval, the packer is then set below the second set isolation of multiple intervals. Therefore, availability
of holes and the treatment is pumped through the of a cheap form of well intervention is usually a
CT-wellbore annulus. After finishing the treatment, pre-requisite for this kind of project.
the annulus is then circulated to clear it of proppant, Consequently, the use of a workover rig is generally
the packer is unseated, the BHA is moved up to the an integral part of fracturing operations for mature field
next interval and the entire process is repeated as re-developments and it is not uncommon to have two
many times as necessary. or even three workover rigs per frac spread. In such a
477
Modern Fracturing
case, one rig might be preparing a well for fracturing Re-perforating for Fracturing: Perforations provide a
operations, one potentially on location during vital connection between the wellbore and the fracture
fracturing and/or one performing post-treatment work and are extremely important in terms of fracture behavior
such as clean out and re-completion. (see Sections 6-1 and 6-2) both during placement
Treating wells completed with a single interval and when on production. Only a limited number of
is relatively operationally straightforward, if the factors can be 100% directly controlled during a typical
completion possesses the necessary integrity and the fracture treatment; of these, the choice and manner of
wellbore is free of debris, fish and scale. This kind perforating probably has the largest single influence on
of well can be easily treated without a workover the success of treatment placement. Unfortunately of
intervention. However, these wells are generally course, unless the well is an infill and has been newly
rare. Usually, wells are completed with multiple drilled; it is unlikely that in a re-development there will
perforated intervals and will require some kind of be an opportunity to significantly change the location
completion-related intervention. and distribution of the existing perforated interval(s). In
If coiled tubing isolation methods are not the case of mature assets, most of the wells to be fractured
economically viable, then the best approach for will already have been perforated (multiple times) across
isolating multiple intervals is to use a dedicated string of existing zones; however, some opportunity may exist
tubing (referred to as the frac string) and a retrievable to optimize fracture initiation if significant amounts of
squeeze packer. The packer can be set above the interval by-passed pay have been identified.
to be treated and provide the necessary isolation from In many cases, it is often necessary to re-perforate
the intervals immediately above the zone of interest. formations prior to re-fracturing operations (Olsen,
Isolation below the zone can be provided by one of 1991, and Lisigurski and Rowe, 2006). In both of these
several reliable and proven methods such as cement examples, it was found that existing intervals had to be
plugs, sand fills and bridge plugs (retrievable packer- re-perforated in order to prevent/reduce the onset of a
type, drillable/millable and inflatable). premature screenout. This suggests that the existing
However, care must be taken with this approach perforations were either partially blocked, completely
if the upper packer is positioned between sets of blocked or that original perforation design (e.g. entry-hole
perforations, with perforations open to the annulus diameter) was not appropriate for the size of proppant
above the packer. First, it will be very difficult to material being pumped. The presence of blocked
maintain pressure on the annulus in order to provide perforations is not uncommon, as phenomena such as
additional downwards force on the packer. In such inorganic scales, asphaltenes, waxes, fines, perforating
circumstances, it is not uncommon to run several drill debris and solids from injected or workover fluids can
collars or joints of drill-pipe above the packer (Rylance often easily accumulate in the tunnels.
et al., 2007) in order to provide sufficient weight on Poor-quality perforations/perforation friction
the packer. Careful calculation of packer hydraulics is difficult to distinguish from tortuosity based
must be made; otherwise, a screenout could cause the solely on a simple single-rate pump in test because
packer and tubing to move out of the well. The second both will manifest as a sharp drop in BHTP as soon
consideration is the potential for communication as the pumps are shut down. In order to accurately
while fracturing, between the lower and upper sets of allocate the pressure-drop to the appropriate root-
perforations. This can this result in the equalization of cause, it is generally necessary to perform a step-
pressures across the packer (causing the hydraulic hold- down test (preferably with access to BHTP), as
downs to become ineffective); in addition, proppant detailed in Section 4-2.1.8. After this step-down
could be placed into the annulus above the packer, test has been completed, it is possible to use Eq.
making it very difficult to recover the packer/string 4-2 to quickly estimate either the number of open
from the well. Therefore, this technique should only perforations (assuming a value for the entry-hole
be used when the well has a good cement bond and diameter) or a value for the average entry-hole diameter
sufficient separation between the intervals. (assuming the number of open perforations).
478
Alternatively, it is often possible to input the step- fracturing can be covered by the same broad groupings.
down test data directly into one of the commercially However, in the case of re-fracturing there are some
available fracture models, where the same analysis can additional considerations, which are specifically related
be quickly and easily performed. to the original fracturing operations.
Re-perforating can provide several benefits, Ineffective Initial Completion: Potential reasons
including: for poor initial fracture performance are numerous,
1. A relatively clean connection between the including inadequate or inappropriate design, poor
wellbore and the formation, which will almost on-site quality control, inappropriate original fluid/
certainly be cleaner than the connection proppant choice(s), poor treatment fluid recovery and
provided by the old, used perforations by the occurrence of insurmountable mechanical issues
taking advantage of any improvements in during placement. Whatever the reasons may be,
perforating charge technology. any fracture that is not optimum with respect to the
2. An opportunity to employ the recommended formation requirements potentially presents itself as a
principles for perforating for fracturing, as candidate for re-fracture consideration.
detailed in Section 6-2, in order to minimize Transient Damage Effects: As noted in Section 13-
the potential for tortuosity and the possibility 1.3, transient damage effects can reduce the reservoir
of a premature screenout. performance with time; this is also true of producing
3. The ability to initiate a fracture without viscous hydraulic fractures. Scales and fines can invade and
gel collecting the accumulated detritus found block the existing proppant pack and reduce the
in the existing perforations and injecting it effective fracture conductivity, or long-term continuous
back into the formation and fracture tip with proppant flowback may eventually negate the fracture
potentially disastrous results. potential. Alternatively, perhaps a well has had a
4. The ability to perforate a limited interval or workover performed and the kill fluids and fluid-loss
number of intervals and target specific sections additives may have permanently damaged the near-
of pay, as the fracture will almost certainly wellbore fracture conductivity. Once again, any fracture
propagate away from the new perforations whose in situ properties have degraded to the point that
even if these are in the middle of a long it has become the choke between the formation and the
section of existing perforations. wellbore is a potential re-fracture candidate.
Transient Reservoir Properties: A reduction in pore
The only real downside to re-perforating is the pressure will create a change in stresses leading to
cost and additional operational complexity. increased effective stress on the proppant pack; it is quite
feasible that the proppant pack may deteriorate under
13-5 Re-Fracturing Operations these conditions as proppant crushes and conductivity
reduces. Another example would be in a condensate-
Once more into the breach, dear friends, once more. banking environment, where it is possible/probable
William Shakespeare, Henry V, act 3, sc. 1. that the original fracture treatment was not adequately
designed to cope with the unusual flow conditions and
Given that a formation was suitable to be developed fracture requirements to cope with two-phase flow. Under
with a hydraulic fracture treatment during the initial these conditions, the initial fracture (although potentially
completion phase, there is always the possibility that being optimal), does not provide an efficient solution for
it will become a re-fracture candidate at a later date. late-life field/well behavior; re-fracturing may therefore
Re-fracturing operations continue to be a staple provide additional potential in these cases also.
approach/consideration when enhancing production Technology Advances: During the past 50 years
from existing fields (Wolhart, 2004). In the same way fracture understanding and application has been
that we categorized reasons for fracturing in mature undergoing constant development, as recently
fields (see Fig. 13-3), the potential for considering re- reported by Veatch (2007). A natural byproduct
479
Modern Fracturing
of this will mean that significant incremental Re-Fracture Productivity: There are many case history
improvements have been made in all aspects of examples in the literature assessing the results of re-
fracture execution, providing significant potential for fracture programs in order to evaluate their overall
efficient re-fracturing. economic success. The majority of these studies
demonstrate success, as reported by Wolhart (2004),
13-5.1 Re-Fracturing Case Histories yet re-fracturing appears to have a poor reputation with
most operators in terms of its efficiency.
A good starting point, when considering a potential Of course, an immediate assessment of re-
re-fracturing campaign, is to study the wide and fracture benefits is often made by directly comparing
varied experience(s) of others. Investigations of trends, initial and post-treatment productivity. However,
successes and failures of re-fracs (Conway et al., 1985, there may be any number of drivers for considering
and Elbel and Mack, 1993) and the GRI funded re-fracturing, (as well as improving the recoverable
studies (ARI, 1996) provide some critical insight reserves, there is establishing or sustaining stable
into the most important areas of consideration. More flow, improving wellbore de-watering capability,
recently the work reported by Ely et al. (2000) and improving gathering and compression efficiency,
Reeves et al. (1999) presents effective application of etc.), and it is important that these form part of
the historical lessons learned. any post-treatment assessment.
A recurring sentiment from these references is
that no one categorization will generate an effective Re-Fracturing Overall Economics: The make-
re-fracturing candidate list; no silver-bullet exists. or-break measure of any re-fracturing campaign
Rather, the most successful means of selecting and will be the overall economics the program presents
ranking candidates for a re-fracturing campaign are upon completion (or shut-down).
multi-faceted and multi-disciplinary. These selection Some obvious considerations (Pospisil et al., 1992,
processes/considerations include: and Sencenbaugh et al., 2001) will be major drivers in
the overall viability of these programs, e.g. post-fracture
Fracture Design and Fluids/Proppants: There are production and decline rates, gas prices, working
two distinct and independent areas of consideration interest, availability of tax credits, etc. Some of these
for fracturing design and fluid and proppant are not within the influence of the operator; however,
utilization: initial treatment choices and re- all should be considered. Of greater concern are those
stimulation treatment choices. areas that seem to be consistently under-estimated
Initial Fracture Treatment: A number of papers in their potential to de-rail an efficient re-fracturing
refer to the various problems with initial fracture design program (e.g. workover/wellbore integrity, management
and the material choices for these treatments (Parrot appreciation/understanding of what risk actually
and Long, 1979; Hunter, 1986; Pospisil et al., 1992; means, surface gathering system impact, program
and Sencenbaugh et al., 2001). These papers refer to a planning/ranking, exit strategy, etc.)
multitude of problems arising from the initial treatments, Any planned re-fracturing program should be
such as ineffective formation coverage, use of excessive based upon a risk-ranked table of candidates and
gel concentrations and inappropriate proppant choices opportunities, where the operator has weighted
for life-of-field reservoir conditions. the issues which have the greatest impact for that
Re-Fracturing Treatment: As well as remediating poor particular campaign. The program should then be
initial fracture placement, re-fracturing will always hold executed and the table continuously updated on the
significant potential because of continuous technological fly as operations proceed, lessons are learned and
advancements and improved understanding/ additional data are gathered. There are numerous
interpretation. An example of careful consideration and ways that this can be achieved, one of the more
identification of actual fluid and proppant requirements efficient being the tracking system promoted by
is presented by Shaefer (2006). Martins et al. (1995). Upon eventual departure from
480
an agreed set of successful metrics, the program 4. Evaluate Economics based on Drivers and Costs:
should then be halted (permanently or temporarily) Clear understanding of the economics is
until the candidate table and agreed metrics essential for a re-fracturing campaign. All of
are updated and reviewed. the drivers and goals should be assessed before
the program begins. Even more importantly,
13-5.2 Candidate Selection for Re-Fracturing as the re-fracture program develops, the
measured results must be quickly compared to
Candidate selection for re-fracturing is just as reliant expectation, and the economics for the program
on input data quality and availability as is candidate must be adjusted as the ranked candidate list
selection for an initial fracturing treatment (as noted wells are stimulated. One of the recurring
in Section 13.2) with, of course, the caveat of the faults during re-fracturing campaigns is that
added complication of a highly complex and multi- they fail to learn from the present campaign
variate fracturing operation having already been rapidly enough to affect the economics or
performed on the well. magnitude of the scope of work.
Conway et al. (1985) determined that candidate
consideration for potential re-fracturing operations could Artificial Intelligence in Candidate
be readily broken up into four simpler stages: Selection: Within Section 13-2.4, the role of
1. Estimation of the Remaining Reserves: This can be Artificial Intelligence techniques in candidate
achieved in a variety of ways, from straightforward selection was briefly discussed (e.g. fuzzy-logic,
material-balance approaches through complex neural-networks, genetic algorithms, self-
reservoir and gridded simulations. However, questioning matrices, etc.) In the case of re-
as noted in Section 13-2.3, the use of type- fracturing candidate selection, these approaches
curves such as those presented by Agarwal would appear to be even more appropriate and
et al. (1988) appears to be the most popular effective. This should come as no surprise, given
and accurate current approach. that the comprehensive structured data/database
2. Assessment of Fracture Quality/Parameters: When that should be available from the initial fracture
performing candidate selection for potential treatments greatly increases the quantity,
re-fracturing, it is most important to be able quality and relationships for the artificial
to differentiate between a good well (original intelligence methods to consider.
candidate) with a bad fracture and a bad well Mohaghegh et al. (2000) provide an excellent
(initial candidate). The Agarwal-Gardner type- introduction to the general application of a number
curve approach, as noted above, will also provide of these new approaches, to this particular subject
an assessment of the fracture parameters, as will matter, and this is expanded upon further by the
the approach of Reciprocal of Productivity work presented by Oberwinkler and Economides
Indices or RPI approach as referred to by Shaefer (2003). Finally, Moore and Ramakrishnan
and Lytle (2001), and Cramer (2003). (2006) outlined the overall procedure and
3. Evaluate Fracturing Parameters/Success: There have reinforced the well-reported understanding
been numerous reported cases of fracture data-set that a candidate selection process for re-
manipulation (Parrot and Long, 1979; Hunter, fracturing should be tailored to an individual
1986; Reese et al., 1994; and Oberwinkler and field case and not generalized.
Economides, 2003) aimed at determining either
poorly fracture stimulated wells or common 13-5.3 Re-Fracture Re-Orientation
factors to success. In order for this approach
to be an efficient process that provides realistic As noted by Elbel and Mack (1993), there are
results, the quality and availability of the data two principal reasons for a secondary fracture
input must be carefully managed. treatment to re-orient:
481
Modern Fracturing
Proppant Induced: Creating a hydraulic fracture orthogonal stress. If this stress reduction exceeds the
will induce stresses within a formation perpendicular initial horizontal stress difference, then the direction of
to the minimum principal stress direction. The the minimum horizontal stress will become the direction
magnitude of this induced stress will be equal of the maximum, and vice versa.
to the additional net pressure on the proppant
pack at closure but will decrease rapidly with
distance from the fracture face.
Note: This effect can cause a potential re-orientation Isotropic
Stress Point
of the re-fracturing treatment if the magnitude Reversal
Region
of the newly induced stress is sufficient to x f
overcome the relative difference between the x f
two original horizontal stresses.

Initial Fracture
xf
Production Induced: During post-fracture
production, tensile stresses are induced both parallel
and perpendicular to the fracture face, offset by
Isotropic
induced compressive stresses further afield. The tensile Point
stresses induced on a plane perpendicular to the Re-Fracture
fracture plane are initially of greater magnitude than
those parallel to the fracture plane. With continuing
Figure 13-10 Re-fracture reorientation concept (Siebrits
production at a constant BHFP, the location of the
et al., 1998)
induced compressive stresses moves away from the
wellbore, perpendicular to the fracture face. In turn, In this case, the re-fracture treatment will now
as the compressive region retreats, the magnitude of propagate orthogonally to the original fracture and
the tensile stresses reduce proportionately. grow to a half-length of xf at which point it will reach
Note: This effect can cause a potential re-orientation the isotropic (or balance) point. Once past this point
of the re-fracturing treatment if the difference the re-fracture treatment will begin to re-orient itself
between the pore pressure-induced stresses is back to the far-field stress regime over the remaining
greater than the difference between the two half-length, xf. The distance to the isotropic point will
original horizontal stresses. be a function of the initial horizontal stress contrast,
Note: The transient nature of the location/magnitude the initial fracture penetration, the production
of these induced stresses will mean that there is transient and the reservoir permeability.
an optimum timing to the secondary fracture Clearly, there is the possibility of significant
treatment, in order to achieve maximum post-fracture production potential with re-fracturing
secondary fracture penetration (Elbel treatments, when the additional opportunity
and Mack, 1993). exists that the re-fracture might include a certain
measure of fracture reorientation. There have been
The second concept can be more readily understood numerous case studies of reorientation impact,
with reference to Fig. 13-10, which is taken directly effects and the principal and defining
from Siebrits et al. (1998). The figure shows a horizontal parameters. A number of the most important
cross-section through a vertical well containing an initial considerations have been listed here:
fracture of half-length xf; continued production (post- 1. Horizontal Permeability Anisotropy: In the
frac) will result in a localized pore pressure change with presence of a distinct horizontal permeability
an elliptical distribution. As noted, this pressure depletion anisotropy or secondary permeability such
will result in the horizontal stress component (parallel as natural fracturing, the reorientation of a
to the initial fracture) declining more quickly than the re-fracture treatment presents the distinct
482
possibility of a dramatic post-fracture production There are a number of reasons why re-fracture
improvement if the original fracture treatment stimulation(s) are or can be highly successful and
was created in an unfavorable direction to such economical ventures. Re-fracturing for fracture
anisotropy (Wright and Conant, 1995). reorientation is one of these and is quickly becoming
2. Angle of Re-fracture Reorientation: The outline a growing application. However, the complexity
theory presented and expanded upon by of efficiently choosing re-fracture candidates,
Siebrits et al. (1998) and Siebrits et al. (2000) where the principal or lone driver is re-orientation,
discusses the symmetric situation when the re- should not be underestimated; this is candidate
fracture treatment is created orthogonally to selection in multiple dimensions, with offsets,
the original fracture. In reality, uneven areal production histories and transient behavior playing
depletion, off-balance fracture growth and the an increasingly intricate role.
potential creation of shear-failure planes (Li et
al., 2006) can result in a substantial variation of 13-5.4 Improved Treatment Design
the re-fracture angle with respect to the original
treatment (Wright et al., 1994a, 1994b). The last 20 years or so have seen unprecedented
3. Drainage Shape, Drainage Area and Depletion: advances in fracture treatment design, described in detail
It can be seen from the discussion(s) above and in Sections 4-5 and 13-3 (and, of course, throughout
Section 13-3 that the relationships between this book); patently, re-fracturing operations will equally
the well spacing, drainage shape, drainage benefit from this process of continual innovation.
area, reservoir permeability (magnitude and With the benefits of hindsight, post-fracture behavior
directionality), original fracture geometry and associated data, it is often possible to review older
and depletion stream lines are complex and treatments and readily find errors or inefficiencies in the
interrelated in terms of potential re-fracture original treatment design or approach. Furthermore, the
reorientation behavior (Cipolla et al., 2005). interval between an initial treatment and a re-fracture
It is well understood, however, that maximum operation does not have to be very great or require a
re-fracture reorientation impact, is achieved massive technological leap. One of the primary factors
when the effective drainage area is significantly that has contributed to the continued popularity of
widened as a result of the treatment (i.e. hydraulic fracturing is just how robust and forgiving this
the aspect ratio is reduced). technology can be. Many fracture treatments continue
4. Timing of the Secondary Fracture: The transient to be poorly designed; however, there is sufficient
nature of the re-fracture reorientation mechanism production benefit (from even a poorly designed frac)
also means that there is an optimum timing for the approach to be economic. This commonplace
associated with the secondary fracture treatment lack of fracture optimization, although detrimental to
execution. The timing of the re-fracture initial post-fracture performance, is the very focus of a re-
treatment can be crucial in achieving a reasonable fracture treatment philosophy. Consequently, one of the
economic return (from reorientation behavior prerequisites that an engineer will attempt to quantify
alone); a simulation is necessary to ensure that when considering a re-fracture treatment is the quality
the opportunity for effective re-fracturing (for and optimality of the original fracture treatment design
reorientation effect) still exists. (Elbel and Mack, 1993; and Reeves et al., 1999).
5. Off-balance Fracture Geometries: As noted earlier Theoretically, it should be considerably more
in Section 13-3, and by Mukherjee et al. (1995), straightforward to design a re-fracture treatment than
off-balance/asymmetric fracture growth can also an initial treatment, given the data garnered during the
be associated with pore-pressure depletion and original fracture treatment and potentially significant
therefore re-fracture candidates. The potential for production and reservoir history available from the well
asymmetric fracture growth is an added level of and field. Most important of all available additional
complication to the reorientation effect. data will be the improved understanding of the average
483
Modern Fracturing
reservoir permeability, which is essential to efficiently not be a suitable re-fracture candidate. However, if
determine the optimum CfD for the fracture (see suitable improvements in the treatment design can be
Section 4-5.2). If such data is not readily available, it is made (improved proppant, fluid system and/or fracture
recommended that either a flow and shut-in pressure geometry), then the treatment may well be successful
transient test be performed (see Sections 3-6 and 3-14) if sufficient recoverable reserves remain to justify the
or that production data be used to perform a rate-time expense of the re-fracture stimulation.
reciprocal productivity index analysis (see Section 3-7.3) With this type of treatment, it is highly recommended
to determine k (Shaefer, 2006). to flush the existing proppant away from the wellbore,
In addition, because the producing conditions will and then subsequently perform a minifrac to evaluate
also be well-recorded and -understood, it should be tortuosity and fluid leak-off (see Sections 4-2.1.5 and
possible to obtain a much more accurate knowledge 4-2.1.6, respectively). Furthermore, in those particular
of the effective proppant permeability requirements cases where the initial treatment had experienced a
(see Sections 7-6 and 8-6), significantly improving the premature screenout, Sauer et al. (2003) recommend
accuracy of the optimum fracture dimensions. using acid soaks to remove polymer residues, caustic soda
In the case of a well with multiple intervals, an soaks to break up resin-coated proppant packs (separated
economic decision must be made between treating each by suitable spacers, of course) and a series of injection
zone individually or attempting a limited-entry type re- tests and proppant slugs to ensure the near-wellbore
fracture treatment over a number of these at one time. It region is free of proppant. After all reasonable steps have
may also be better to isolate sections of long perforated been taken to clean the near-wellbore region and flush
intervals, rather than trying to treat the entire section. the proppant deep into the formation, the re-fracture
Certainly, an appropriate selection of these strategies will treatment can be designed based on the data generated
allow re-fractures to be placed deterministically along the by the injectivity testing, in exactly the same fashion as if
wellbore, targeting reserves, productivity and previously the original fracture did not exist.
unstimulated pay rather than at the stochastic fracture Fracture Re-Orientation Case: In the case of opening
locations produced by the previous methods. a new fracture, the accepted practice in the industry
When designing a re-fracture treatment, the two is to design the re-fracture treatment as if the original
distinct re-fracture situations (re-opening an existing fracture were not there. If the recommended diagnostic
fracture and fracture reorientation) should be considered injection test has been performed (see above), then this
as separate and particular cases. In order to differentiate data should be used as the basis for the design, together
between these two classes of re-frac, it is highly with any data gained during the producing life of the
recommended to perform a short pumping test consisting well. In the absence of additional data, the design should
of 3 to 4 wellbore volumes of slickwater pumped above be based on successful offset treatments, the initial
the fracture pressure. After pump-in, the pressure decline treatment (if it was successful) and any improvements
should be monitored closely; a brief analysis of pISI and in treatment design (fluids, proppant and/or fracture
pc will quickly reveal if the initial fracture has been re- geometry) that can be made.
opened (similar pressure response to the initial frac) or
if fracture re-orientation has taken place (significantly
different pressure response).
Re-Opening an Existing Fracture: If the initial fracture References:
is simply being re-opened, then the most important
question is whether or not this well is a good candidate Aboud, R.S., and Melo, R.C.B.: Past Technologies
for re-fracturing where the production enhancement Emerge Due to Lightweight Proppant Technology:
opportunity is simply related to the ability to improve Case Histories Applied on Mature Fields, paper
on the previous treatment design. A close examination SPE 107184, 2007.
of the original fracture treatment must be made, and Advanced Resources International Inc., Assessment
if there is no significant upside, then this well may of Technology Barriers and Potential Benefits of
484
Restimulation R&D for Natural Gas Wells, Final Field Development using Propped Hydraulic
Report to Gas Research Institute, GRI-96/0267, Fracturing, paper SPE 88604, 2004.
1996. Carter, R.D.: Type Curves for Finite Radial and Linear
Agarwal, R.G., Gardner, D.C., Kleinsteiber, S.W., and Gas Flow Systems: Constant Terminal Pressure
Fussell, D.D.: Analyzing Well Production Data Case, paper SPE 12917, 1985.
Using Combined-Type-Curve and Decline-Curve Cawiezel, K.E., Wheeler, R.S., and Vaughan, D.R.:
Analysis Concepts, paper SPE 49222, 1998. Specific Fluid Requirements for Successful Coiled
Babadagli, T.: Mature Field Development - A Review, Tubing Fracturing Applications, paper SPE
paper SPE 93884, 2005. 86481, 2004.
Bakhtar, K., Barton, N.R., Rakop, K., and Jones, A.H.: Chambers, R. and Meise, K.: Comparison of Fracture
Modeling Fracture Permeability Around a Well Geometries Utilizing Ultra-lightweight Proppants
During Depletion, paper SPE 13671, 1985. Provide Evidence that Partial Monolayers Can
Barba, R.E., and Shook, R.A.: Post Frac Evaluation Be Created: A Case History, paper SPE 96818,
of Multi Zone Fracture Treatments Using the 2005.
Completion Efficiency Concept, paper SPE Cikes, M.: Long-Term Hydraulic-Fracture
90483, 2004. Conductivities Under Extreme Conditions, paper
Bennion, D.B., Bietz, R.F., Thomas, F.B., and Cimolai, SPE 36878, 1996.
M.P.: Reductions in the Productivity of Oil and Cipolla, C., Peterman, F., Creegan, T., McCarley, D.,
Low Permeability Gas Reservoirs due to Aqueous and Nevels, H.: Effect of Well Placement on
Phase Trapping, J. Can. Pet Tech. 94-09-05, Production and Frac Design in a Mature Tight Gas
November 1994. Field, paper SPE 95337, 2005.
Bennion, D.B., Thomas, F.B., Bietz, R.F., and Bennion, Cipolla, C.L., and Wight, C.A.: Diagnostic Techniques
D.W.: Water and Hydrocarbon Phase Trapping to Understand Hydraulic Fracturing: What? Why?
in Porous Media Diagnosis, Prevention and And How? paper SPE 59735, 2002.
Treatment, J. Can. Pet Tech. 96-12-02, December Comrie, P., and Olsen, T.: A Risk Sharing Alliance
1996. Breathes New Life into a Mature North Sea Field,
Bennion, D.B., Thomas, F.B., Ma, T., and Schulmeister, SPE Paper 3822, 1997.
B.: Gas Productivity Reductions due to Water- Conway, M.W., McMechan, D.E., McGowen, J.M.,
Based Phase Trapping Effects in Low Permeability Brown, D., Chisholm, P.T., and Venditto, J.J.:
Gas Reservoirs, The Pet. Soc. 2002-064, June Expanding Recoverable Reserves Through Re-
2002. Fracturing, paper SPE 14376, 1985.
Black, H.N., and Langsford, R.W.: Energized Craft, J.R., Waddell, S.P., and McFatridge, D.G.:
Fracturing with 50% CO2 for Improved CO2-Foam Fracturing with Methanol Successfully
Hydrocarbon Recovery, paper SPE 9705, JPT, Stimulates Canyon Gas Sands, paper SPE 20119,
135-140, January 1982. 1992.
Borodin, E.G., Vakhroushev, P.E., Stolyarov, S.M., Cramer, D.D.: Evaluating Well Performance and
and Kalfayan, L.J.: Applications of Relatively completion Effectiveness in Hydraulically Fractured
Permeability Modifiers To Control Water Cut Low-Permeability Gas Wells, paper SPE 84214,
Following Hydraulic Fracturing in Western Siberia 2003.
Oilfields-Russian Case History Study, paper SPE Cramer, D.D., Dawson, J., Ouabdesselam, M.: An
102679, 2006. Improved Gelled Oil System for High-Temperature
Brown, J.E., King, L.R., Nelson, E.B., and Ali, S.A.: Fracturing Applications, paper SPE 21859, 1991.
Use of a Viscoelastic Carrier Fluid in Frac-Pack Crawford, H.R., Niell, G.H., Bucy, B.J., and Crawford,
Applications, paper SPE 31114, 1996. P.B.: Carbon Dioxide a Multipurpose Additive
Burnstad, R.G., Martin, A.N., Stemberger, D.J., for Effective Well Stimulation, paper SPE 571,
and Purwanto, P.: A Case Study of a Mature JPT, 237-242, 1963.
485
Modern Fracturing
Darin, S.R., and Huitt, J.L.: Effect of a Partial Field Applications in the Western Canadian
Monolayer of Propping Agent on Fracture Flow Sedimentary Basin, paper SPE 97211, 2005.
Capacity, AIME Petroleum Transactions 1291-G, Hernandez, J.M., Fernandez, C.T., and Scianca, N.M.:
1959. Methanol as Fracture Fluid in Gas Wells, paper
Elbel, J.L., and Mack, M.G.: Refracturing: SPE 27007, 1994.
Observations and Theories, paper SPE 25464, Hubbert, K.M., and Willis, D.G.: Mechanics of
1993. Hydraulic Fracturing, Trans., AIME, 1957, paper
Ely, J.W., Arnold III, W.T., and Holditch, S.A.: New SPE 686.
Techniques and Quality Control Find Success in Hunter, J.C.: A Case History of Refracs in the Oak
Enhancing Productivity and Minimizing Proppant Hill (Cotton Valley) Field, paper SPE 14655,
Flowback, paper SPE 20708, 1990. 1986.
Ely, J.W., Tiner, R., Rothenberg, M., Krupa, A., Hurst, R.E.: Gas Frac A New Stimulation Technique
McDougal, F., Conway, M., and Reeves, S.: Using Liquid Gases, paper SPE 3837, 1972.
Restimulation Program Finds Success In Kiburz, D.E., and Sullivan, S.C.: A Methodology to
Enhancing Recoverable Reserv, paper SPE 63241, Optimize Fracturing Treatments in a Tight-Gas
2000. Field of Northwest Louisiana, paper SPE 25994,
Ely, J.W., Martin, M.A., Duenas, J.J., and Trythall, 1995.
J.R.: Enhanced Fracture Stimulation in the Deep Kugler, R.L., Stefanescu, D., Louboutin, M., and
Morrow Sandstone in Western Oklahoma, paper Wegemer, K.: Rehabilitation of Mature Gas
SPE 80932, 2003. Fields in Romania: Success Through Integration of
Farrell, S., and McDermott, J.R.: An International Management Processes and New Technology, Acta
Perspective on Risk/Reward Contracting: Montanistica Slovaca, Volume 9, Issue 3, 2004.
Comparison of U.S., Middle-East and U.K. Li, P., Song, Z., and Wu. Z.: Study on Reorientation
Alliances, paper SPE 30376, 1995. of Refracturing in Ordos Basin A Case Study:
Gavin, W.G.: Fracturing Through Coiled Tubing Chang 6 Formation, Yanchang Group, Triassic
Recent Developments and Case Histories, paper System in Wangyao Section of Ansai Oil Field,
SPE 60690, 2000. paper SPE 104260, 2006.
Gazi, N.H., Hottman, W.E., Logan, J.L., and Verrett, Lillies, A.T., and King, S.R.: Sand Fracturing With
L.C.: Alliances and Partnering: A New Relationship Liquid Carbon Dioxide, paper SPE 11341,
Between Oil/Gas Producing Companies and 1982.
Service Companies, paper SPE 29793, 1995. Lisigurski, O., and Rowe, G.C.: Practical Steps to
Gilbert, J., and Greenstreet, C.: Applications of Increase Production and Reserves in Mature Gas
Pinpoint Fracturing in the Cooper Basin, Australia, Fields: Hugoton and Panoma, Texas County,
paper SPE 97004, 2005. Oklahoma, U.S.A., paper SPE 102259, 2006.
Grieser, B., Shelley, B., Johnson, B.J., Fielder, E.O., Love, T.G., McCarty, R.A., Surjaatmadja, J.B.,
Heinze, J.R., and Werline, J.R.: Data Analysis of Chambers, R.W., and Grundmann, S.R.:
Barnett Shale Completions, paper SPE 100674, Selectively Placing Many Fractures in Openhole
2006. Horizontal Wells Improves Production, paper
Gupta, D.V.S., Miechwiadowicz, G., and Jerat, A.C.: SPE 74331, SPEPF, 219-224, November 2001.
CO2 Compatible Non-Aqueous Methanol Macfarlane, M.S., and Mackey, P.A.: Monobores
Fracturing Fluid, paper SPE 84579, 2003. Making a Difference to the Life Cycle Cost of a
Gupta, D.V.S.: Field Application of Unconventional Development, paper SPE 50046, 1998.
Foam Technology: Extension of Liquid CO2 Martins, J.P., MacDonald, J.M., Stewart, C.G., and
Technology, paper SPE 84119, 2003. Phillips, C.J.: The Management and Optimization
Gupta, D.V.S., Leshchyshyn, T.T., and Hlidek, B.T.: of a Major Wellwork Program at Prudhoe Bay,
Surfactant Gel Foam/Emulsions: History and paper SPE 30649, 1995.
486
Mazza, R.L.: Liquid-Free CO2/Sand Stimulations: Palacio, J.C., and Blasingame T.A.: Decline-Curve
An Overlooked Technology - Production Update, Analysis Using Type-Curves Analysis of Gas Well
paper SPE 72383, 2001. Production Data, paper SPE 25909, 1993.
Meehan, D.N., Horne, R.N., and Aziz, K.: Effects Pathak, P., Fidra, Y., Avida, H., Kahar, Z., Agnew, M.,
of Reservoir Heterogeneity and Fracture Azimuth and Hidayat, D.: The Arun Gas Field in Indonesia:
on Optimization of Fracture Length and Well Resource Management of a Mature Field, paper
Spacing, paper SPE 17606, 1988. SPE 87042, 2004.
McConkey, S., and Stromquist, M.: Integration of Parrot, D.I., and Long, M.G.: A Case History of
Conventional Fracturing, Coiled Tubing and Massive Hydraulic Refracturing in the Tight
Retrievable Tool Technology, paper SPE 60709, Muddy J Formation, paper SPE 7936, 1979.
2000. Pietrak, M.J., Stanley, F.O., Weber, B.J., and Fontenot,
McIninch, B., McIntyre, A., Gulrajani, S., and Norris, J.S.: Relative Permeability Modifier Treatments
M.: New Relationship Between Oil Company on Gulf of Mexico Frac-packed and Gravel-packed
and Service Company Rejuvenates a Mature North Oil and Gas Wells, paper SPE 96945, 2005.
Sea Gas Field, paper SPE 78327, 2002. Pospisil, G., Lynch, K.W., Pearson, C.M., and
McLennan, J.D., and Roegiers, J-C.: How Rugen, J.A.: Results of a Large-Scale Refracture
Instantaneous are Instantaneous Shut-In Pressures, Stimulation Program, Kuparuk River Unit, Alaska,
paper SPE 11064, 1982. paper SPE 24857, 1992.
Mohaghegh, S., Reeves, S., and Hill, D.: Development Posey, D., and Strickland, B.: The Effect of Using
of an Intelligent Systems Approach for Restimulation a Lightweight Proppant in Treatment of a Low
Candidate Selection, paper SPE 59767, 2000. Permeability Dry Gas Reservoir: A Case Study,
Moore, L.P., and Ramakrishnan, H.: Restimulation: paper SPE 97998, 2005.
Candidate Selection Methodologies and Treatment Rae, P., Martin, A.N., and Sinanan, B.: Skin Bypass
Optimization, paper SPE 102681, 2006. Fracs: Proof that Size is not Important, paper SPE
Mukherjee, H., Poe, B., Heidt, H., Watson, T., and 56473, 1999.
Barree, R.D.: Effect of Pressure Depletion on Reese, J.L., Britt, L.K., and Jones, J.R.: Selecting
Fracture Geometry Evolution and Production Economic Refracturing Candidates, paper SPE
Performance, paper SPE 30481, 1995. 28490, 1994.
Mukherjee, H., Poe B.D., Jr., Heidt, J.H, Watson, T,B., Reeves, S., Hill, D., Hopkins, C., Conway, M.,
and Barree, R.D.: Effect of Pressure Depletion Tiner, R., and Mohaghegh, S.: Restimulation
on Fracture-Geometry Evolution and Production Technology for Tight Gas Sand Wells, paper SPE
Performance, paper SPE 65064, SPEPF, 15, 3, 56482, 1999.
pp144 150, August 2000. Rickards, A.R., Brannon, H.D., Wood, W.D.,
Nehmer, W.L.: Viscoelastic Gravel Pack Carrier Fluid, and Stephenson, C.J.: High Strength, Ultra-
paper SPE 17168, 1988. Lightweight Proppant Lends New Dimensions
Oberwinkler, C., and Economides, M.J.: The to Hydraulic Fracturing Operations, paper SPE
Definitive Identification of Candidates Wells for 84308, 2003.
Refracturing, paper SPE 84211, 2003. Rodvelt, G., Toothman, R., Willis, S., and Mullins,
Oberwinkler, C., Rythammer, G., Zabgl, G., and D.: Multiseam Coal Stimulation Using Coiled
Economides, M.J.: New Tools for Fracture Design Tubing Fracturing and a Unique Bottomhole
optimization, paper SPE 86467, 2004. Packer Assembly, paper SPE 72380, 2001.
Olsen, K.E.: A Case Study of Hydraulically Refractured Rylance, M., Doig, T.W., Hoff Jr., R., and Knopp,
Wells in the Devonian Formation, Crane County, S.: Remedial Frac Operations in Deep Tectonic
Texas, paper SPE 22834, 1991. Wells: Mechanical and Operationsl Issues, paper
Oxford English Dictionary, (OED), 20 Volume, Second SPE 105205, 2007.
Edition, 1989. Sauer, P.W., Burns, R.A., Skees, J.L., Aud., W.W., Gentry,
487
Modern Fracturing
B., and Wing, C.: Re-Fracturing: Evaluation, Surjaatmadja, J.B., Grundmann, S.R., McDaniel,
Design, and Implementation of a Chester Oil Well B., Deeg, W.F.J., Brumley, J.L., and Swor, L.C.:
In SW Kansa, paper SPE 80916, 2003. Hydrajet Fracturing: An Effective Method for
Sencenbaugh, R.N., Lytle, D.M., Birmingham, T.J., Placing Many Fractures in Openhole Horizontal
Simmons, J.C., and Shaefer, M.T.: Restimulating Wells, paper SPE 48856, 1998.
Tight Gas Sand: Case Study of the Codell Tolman, R.C., Carlson, L.O., Kinison, D.A., Nygaard,
Formation, paper SPE 71045, 2001. K.J., Gloss, G.S., Sorem, W.A., and Shafer, L.L.:
Schein, G.W., Carr, P.D., Canan, P.A., and Richey, Method and apparatus for stimulation of multiple
R.: Ultra Lightweight Proppants: Their Use and formation intervals, US patent No. 7,059, 407,
Application in the Barnett Shale, paper SPE issued June 2006.
90838, 2004. Utegalyev S.A., Duzbayev, S.K., Kulbatyrov, S.V.,
Shaefer, M.T., and Lytle, D.M.: Fracturing Fluid and Nadezhdin, S.V.: Optimized Stimulation
Evolution Plays a Major Role in Codell Refracturing Solutions for a Mature Field in Kazakhstan, paper
Success, paper SPE 71044, 2001. SPE 98055, 2006.
Shaefer, M.: Awakening an Old Field A Case Study Vairogs, J., Hearn, C.L., Dareing, D.W., and Rhoades,
of a Refracturing Program in the Greater Green V.W.: Effect of Rock-Stress on gas Production
River Basin, paper SPE 101026, 2006. From Low-Permeability Reservoirs, paper SPE
Sherman, J.B., and Holditch, S.A.: Effect of Injected 3001, 1971.
Fracture Fluids and Operating Procedures on Veatch, R.W.: A Historical Perspective of Hydraulic
Ultimate Gas Recovery, paper SPE 21496, 1991. Fracturing, SPE HFTC Keynote Speech, 2007.
Siebrits, E., Elbel, J.L., Hoover, R.S., Diyashev, I.R., Vincent, M.C.: Proving It A Review of 80 Published
Griffin, L.G., Demetrius, S.L., Wright, C.A., Field Studies Demonstrating the Importance
Davidson, B.M., Steinsberger, N.P., and Hill, of Increased Fracture Conductivity, paper SPE
D.G.: Refracture Reorientation Enhances Gas 77675, 2002.
Production in Barnett Shale Tight Gas Wells, Vincent, M.C., Pearson, C.M., and Kullman, J.:
paper SPE 63030, 2000. Non-Darcy and Multiphase Flow in Fractures:
Siebrits, E., Elbel, J.L., Detournay, E., Detournay- Case Studies Illustrate the Dramatic Effect on Well
Piette, C., Christianson, M., Robinson, B.M., and Productivity, paper SPE 54630, 1999.
Diyashev, I.R.: Parameters Affecting Azimuth and Von Winterfeld, C., Babajan, S., Amer. A. and
Length of a Secondary Fracture During a Refracture Marsden, J.A.: Geomechanics Analysis of the
Treatment, paper SPE 48928, 1998. Crestal Region of an Omani Gas Field, Paper
Smith, M.B.: Hydraulic Fracturing: THE IPTC 10231, 2005.
Multidisciplinary Technology, SPE Distinguished Vreeburg, R-J., Roodhart, L.P., Davies, D.R., and
Lecturer Presentation 108827, 2006. Penny, G.S.: Proppant Backproduction During
Solares, R.: Efficient Technology Application to Hydraulic Fracturing: A New Failure Mechanism
Optimize Deep Gas Well Completions in the For Resin-Coated Proppants, paper SPE 27382,
Khuff and Jauf Formations Requiring Hydraulic 1994.
Fracturing in Saudi Arabia, paper SPE 68206, Wolhart, T: Hydraulic Fracturing Re-Stimulation,
2001. Distinguished Lecturer, SPE Presentation 101458,
Spady, D.W., Udick, T.H., and Zemlak, W.M.: 2004.
Enhancing Production in Multizone Wells Wright, C.A., and Conant, R.A.: Hydraulic Fracture
Utilizing Fracturing Through Coiled Tubing, Reorientation in Primary and Secondary Recovery
paper SPE 57435, 1999. from Low-Permeability Reservoirs, paper SPE
Stundner, M., and Oberwinkler, C.: Self-Organising 30484, 1995.
Maps for Lithofacies Identification and Permeability Wright, C.A., Conant, R.A., Stewart, D.A., and
Prediction, paper SPE 90720, 2004. Byerly, P.M.: Reorientation of Propped Refracture
488
Treatments, paper SPE 28078, 1994a.

Wright, C.A., Stewart, D.W., Emanuele, M.A.,
and Wright, W.W.: Reorientation of Propped
Refracture Treatments in the Lost Hills Field,
paper SPE 27896, 1994b.
Zemlak, W., Lemp, S., and McCollum, R.: Selective
Hydraulic Fracturing of Multiple Perforated
Intervals with a Coiled Tubing Conduit: A Case
History of the Unique Process, Economic Impact
and Related Production Improvements, paper
SPE 54474, 1999.
489
Modern Fracturing
490
NOMENCLATURE
a large half-axis of the drainage ellipsoid formed by a horizontal well

a variable used in Nolte G function analysis to describe how constant the fracturing fluid viscosity
remains
a characteristic fracture length
a spacing between cleats or fractures
a mass fraction of ash in coal
A area
Aa horizontal wellbore drainage area shape factor
Ac cross-sectional area
Ac contact area
Af area of fracture face (includes both sides of fracture)
Aperf perforation cross-sectional area
Ar aspect ratio
APTi aqueous phase trapping index
b channel width
b Langmuir constant
b1 fractured well radial damage in (equivalent to rs)
b2 width of fracture face damage
bpss y-axis intercept of material balance time plot
bs penetration of damage normal to the fracture face
B formation volume factor
B directional closure quality
Bg formation volume factor for gas
Bgi formation volume factor for gas at initial reservoir conditions
Bo formation volume factor for oil
Bw formation volume factor for water
Bx directional closure quality in the x-direction
By directional closure quality in the y-direction
BWV bulk water volume, percent
c compressibility
c proppant concentration, mass per unit volume of slurry
cadded proppant concentration, as added to clean fluid (i.e. pounds of proppant added)
cf compressibility of formation (rock matrix)
cfrac fracture compressibility
cg gas compressibility
co original formation compressibility
ct total (overall) formation compressibility
ct average total (overall) formation compressibility
cw compressibility of wellbore contents
cw compressibility of formation water
C wellbore storage coefficient
C Kachanov parameter
491
Modern Fracturing
Ca proppant concentration required to produce monolayer

CA Dietz shape factor
Cc compressibility-controlled leakoff coefficient
Ccc critical crosslinking concentration
Cd discharge coefficient
Cf fracture conductivity
CfD dimensionless fracture conductivity
CfD 0.1 optimum dimensionless conductivity at Nprop = 0.1
CfD,opt optimum dimensionless fracture conductivity
CL leakoff coefficient
CTRANS transport coefficient
Cr dimensionless fracture conductivity (for Cinco-Ley, et al., 1978) (= CfD /)
Cv viscosity-controlled leakoff coefficient
Cw wall building-controlled leakoff coefficient (the wall-building coefficient)
C* minimum concentration of polymer in water at which intermolecular interaction occurs
C** minimum concentration of polymer in water at which molecular interpenetration occurs
d distance
d minimum dimension of elliptical fracture
d diameter of fracture tip
d lattice spacing
D diameter
D turbulence coefficient
Dc pipe grade perforation diameter correction factor
Dp perforation diameter
Dprop average proppant grain diameter
Dsol solid particle diameter
Dtbg tubing diameter
e natural exponent
e slope of Nolte-Smith analysis plot
E Youngs modulus, modulus of elasticity
E plane strain Youngs modulus
Edyn dynamic Youngs modulus
Eg gas expansion factor
Egi gas expansion factor at initial reservoir conditions
Et Youngs modulus of damaged zone
f Fanning friction factor
fs volume fraction of proppants in a slurry
fs,max maximum volume fraction of proppants in a slurry
F force
Fopt optimum F-function for fractured horizontal wellbore, as defined by Equation 10-7
FL Nolte linear after-closure leak-off function
FR Nolte radial after-closure leak-off function
g intermediate function used in Nolte G function analysis
g acceleration due to gravity
492
gf frac gradient
gob overburden stress gradient
G Nolte G function
G total gas volume currently in place in the reservoir
G shear modulus
G elastic energy release rate
G1c critical energy release rate, failure mode 1
Ga total volume of gas adsorbed in the formation (usually coal)
Gc Nolte G function at fracture closure
Gc average in-situ gas content for coalbed methane reservoirs
Gdyn dynamic shear modulus
GF total original free gas (i.e. in cleats in coalbed methane reservoir) in place
Gi total gas volume initially in place in the reservoir
Gp total gas volume produced from the reservoir
GR total remaining free gas (i.e. in cleats in coalbed methane reservoir) in place
h height
h net height
h depth of embedment
h average reservoir net height
hf fracture height
hn height of nth zone
hp fracture height in pay
H total vertical depth (TVD)
Iani vertical-to-horizontal permeability anisotropy
ISP slurry properties index
Ix penetration ratio
J productivity index (PI)
JD dimensionless productivity index
JDdamaged damaged well dimensionless productivity index
JDmax maximum dimensionless productivity index
JDTH dimensionless productivity index for transversely fractured horizontal well
JDTH1 dimensionless productivity index for one transverse fracture in a horizontal well
JDTHt total dimensionless productivity index for transversely fractured horizontal well (= n JDTH1)
JDV dimensionless productivity index for fractured vertical well
Jratio productivity index ratio
Jo original unfractured (s = 0) productivity index
k permeability, single-phase permeability
k reservoir permeability
k1 permeability of the radial damage zone for fractured well
k2 permeability of the fracture face outside of the damage zone
k3 permeability of fracture face inside the radial damage zone
ka uncorrected average permeability to air
kavg average permeability of the reservoir including skin effects
kD reference permeability measured under Darcy flow conditions
493
Modern Fracturing
ke effective permeability under non-Darcy flow conditions

keq equivalent permeability
kf in-situ proppant pack permeability, fissure permeability
kf, e equivalent (i.e. under turbulent flow) proppant pack permeability
kfi initial fracture or fissure permeability
kf, n nominal (i.e. under Darcy flow) proppant pack permeability
kfs damaged fracture permeability
kF equivalent permeability of proppant incorporating inertial flow effects (Forcheimer
permeability)
kg permeability to gas
kg,Swi permeability to gas at secondary irreducible water saturation
kH horizontal permeability
kM effective permeability of proppant pack under multiphase flow conditions
ko original formation permeability
kperf perforation permeability
ks permeability of the damaged or stimulated formation around the wellbore or fracture
kV vertical permeability
kx permeability in the x-direction
ky permeability in the y-direction
kz permeability in the z-direction
k* equivalent cement permeability
K bulk modulus
K consistency index
K power law fluid consistency index
K Herschel-Buckley fluid consistency index
K1 stress intensity factor, failure mode 1
K1c critical stress intensity factor, failure mode 1 (fracture toughness)
K1ceff effective fracture toughness
K1clag fracture toughness produced by fluid lag effect
Knwb proportionality constant used for calculating near wellbore friction pressure
Kdry dynamic bulk modulus under drained conditions
Kdyn dynamic bulk modulus
Kperf proportionality constant used for calculating perforation friction pressure
KVE volume-equalized consistency index
Ks average dynamic bulk modulus of materials composing rock,
Kwellf proportionality constant for calculating wellbore friction pressure
average distance between microcracks
L length of horizontal wellbore
L length of pipe
Lgravel the difference between drilled radius and screen radius
m mass
m slope or gradient
m1wing mass of proppant contained in one fracture wing
m
slope term for closed chamber test pseudo-pressure plot
494
m
cpr slope term for specialised production analysis plot
m
cpl slope term for specialised plot for identifying the end of linear flow in production analysis
mbf bilinear flow gradient
mlf linear flow gradient
mL slope of the linear flow after closure analysis plot
mR slope of the radial flow after closure analysis plot
ms slope of material balance time plot
m(p) gas pseudo-pressure function
M molecular weight
M constrained axial modulus
n number, number of fractures
n flow behaviour index
n fracturing fluid power law exponent
n Herschel-Buckley exponent
Nperf number of perforations
Nprop proppant number
Nprop,e equivalent proppant number
NRe Reynolds number
NRe Reynolds number for power law fluid
NRe,VE volume equalized Reynolds number
NRe,w wall Reynolds number
NRe,w wall Reynolds number for power law fluid
p pressure
p average reservoir pressure
p1hour bottomhole wellbore pressure 1 hour after shut-in
pavg average pressure
pc closure pressure
pD dimensionless pressure
pe constant outer reservoir pressure
pext extension pressure, fracture extension pressure
pf pressure in fracturing fluid (usually inside main body of fracture)
phead pressure due to hydrostatic head
pi initial pressure, reservoir pressure
pif breakdown pressure
pif,lower lower boundary for breakdown pressure
pif,upper upper boundary for breakdown pressure
pinj wellhead injection pressure
pISI instantaneous shut-in pressure (bottomhole)
pISI,surf instantaneous shut-in pressure (surface)
piw bottomhole injection pressure
pL the pressure at which sorption induced strain equals half of the maximum volume (0.5 VL) (the
Langmuir pressure)
pm Nolte G function analysis match pressure
pnet net pressure
495
Modern Fracturing
pnet,n actual net pressure, caused by n multiple fractures

po reference pressure, original pressure
pob overburden pressure
ppc pseudocritical pressure
ppc1 pseudocritical pressure for whole mixture
ppc* corrected pseudocritical pressure
ppipe friction friction pressure of fluid in tubing or pipe
ppr pseudo-reduced pressure
pr pore pressure
pR pressure at radius R from wellbore
pSC pressure at standard conditions (usually 14.7 psi or 101.353 kPa)
ptip net pressure is the non-wetted part of the fracture tip
ptf flowing (producing) wellhead pressure
pw bottomhole wellbore pressure
pwo initial (original) bottomhole wellbore pressure
pwf bottomhole flowing wellbore pressure
pz process zone pressure
p* average reservoir pressure extrapolated from Horner plot
P matrix-shrinkage Langmuir curve-fitting parameter
q gas flow rate
q liquid flow rate
qcem flow rate through or along a cement sheath
qf gas flow rate after fracture treatment
qi average injection rate
qi,1wing average injection rate into one fracture wing (usually = qi)
qideal gas flow rate from laminar flow, open hole vertical gas well
qideal ideal gas flow rate through an orifice (i.e. Cd = 1)
qDd normalized material balance gas flow rate
qDdi integrated normalized material balance gas flow rate
qDdid derivative of the integrated normalized material balance gas flow rate
qL fluid leakoff rate
qn gas flow rate during nth flow period
qo gas flow rate before fracture treatment
qref initial calculated gas flow rate for closed chamber test
qSC gas flow rate at standard conditions (usually 14.7 psi and 519.67 R)
qTH gas flow rate from transversely-fractured, horizontal well
qv gas flow rate from fractured vertical well
r radius
r radius of indentation
rcsg external radius of casing
re reservoir radial extent
reH equivalent radial flow drainage radius for horizontal wellbore
rhalo radius of halo effect
rp storage ratio
496
rp radius of zone of plastic deformation at fracture tip
rp ratio of fracture area in permeable formation (i.e. net pay) to total fracture area
rs radius of the damaged or stimulated zone around the wellbore
rw wellbore radius
rw effective wellbore radius
R radius
R distance from point of fracture initiation
R universal gas constant
Rf fracture radius for radial fracture model
Rn actual fracture radius, caused by n multiple fractures
s skin effect, skin factor
s rate-dependent skin effect
s apparent skin
sc choke skin factor
sd fractured well composite damage skin effect
sf skin effect after fracture treatment
sfc choked fracture skin effect
sf,opt optimum skin effect after fracture treatment
so skin effect before fracture treatment
sp skin effect due to perforations
Scc critical condensate saturation
Sp spurt loss coefficient
SRE rock embedment strength
SUC ultimate strength under uniaxial compression
SUT ultimate strength under uniaxial tension
Sw water saturation
Swi initial water saturation
t time
t reciprocal of pseudocritical temperature
t effective time
ta pseudo-time
tasuper superposition time
tc closure time
tD dimensionless time
tDA dimensionless time for bounded flow for reservoir of drainage area A
tel time at the end of linear flow
ter time at the end of radial flow
texp time at which a specific point on the fracture face was exposed to fluid leakoff
ti time to end of injection (pumping time)
tm material balance time
tma material balance time pseudo-time
tmDd normalized material balance pseudo-time
tp longitudinal (compression) wave transit time
tpa pseudo-time at shut in
497
Modern Fracturing
tpad pad fluid pumping time

tpss time at which pseudosteady state flow occurs.
ts transverse (shear) wave transit time
T temperature
T absolute temperature
Tpc pseudocritical temperature
Tpc* corrected pseudocritical temperature
Tpc1 pseudocritical temperature for whole mixture
Tpr pseudo-reduced temperature
TSC temperature at standard conditions (usually 519.67 R or 288.706 K)
Ttf flowing (producing) wellhead temperature
TOC total organic carbon
u velocity
u fracture propagation rate
uL velocity of leakoff fluid perpendicular to fracture face
uopt natural logarithm of CfD,opt
up longitudinal (compression) wave velocity
us transverse (shear) wave velocity
U elastic energy
v velocity
v particle settling velocity in pseudoplastic fluid
vavg average velocity
vavg foam average velocity of multiphase flow in foam
vg gas velocity
vl superficial velocity of liquid phase in foam
vsol hindered settling particle velocity
vt Stokes law terminal velocity
V volume
Vc sorption of gas (usually in coal)
Vf volume of one propped wing (= Vp/2)
Vi injection volume
Vi,1wing injection volume into one fracture wing (usually = Vi)
VL volume of fluid lost due to fluid leakoff
VL gas volume at infinite pressure for coalbed methane (the Langmuir volume)
VL,C total fluid loss volume due to filtration-based fluid loss
Vmpc pure coal Langmuir volume
Vp proppant bulk volume in the pay
VP total gas adsorbed per unit mass in reservoir rock, at current reservoir pressure, pr
Vr reservoir drainage volume
Vs volume of fluid lost due to spurt loss
Vw wellbore volume
VRo vitrinite reflectance
w width, fracture width
w0 width at the center of the fracture
498
wave average propped fracture width
wetch acid fracture average etched width
wf propped fracture width
wfs damaged propped fracture width
wmax maximum created fracture width
wn actual fracture width, caused by n multiple fractures
wopt optimum fracture width
Wp cumulative water production
x linear distance
xe well drainage dimension
xe channel half-length in short direction
xf fracture half length
xf distance from wellbore of isotropic point
xf half length of fracture re-oriented by depletion-based stress reversal
xf,eff apparent (effective) fracture half length
xfs damaged fracture half length
xfD dimensionless fracture half length
xfnew equivalent fracture half length in homogenous space
xfopt optimum fracture half length
X ratio of dimensionless productivity indexes for transversely fractured horizontal well to fractured
vertical well
X specific velocity (= v/)
Xe distances in the x-direction from the well to the boundary in anisotropic space
Xfluid specific velocity of fluid
Xg specific velocity of gas
Xnew distance corresponding to Xe in homogenous space
Xt total (overall) specific velocity of multiphase flow
y linear distance
y mole fraction
ye well drainage dimension
ye channel half-length in long direction
yeD dimensionless well drainage in the y-direction
Ye distances in the y-direction from the well to the boundary in anisotropic space
Ynew distance corresponding to Ye in homogenous space
z linear distance
ze well drainage dimension
Z gas deviation factor
Zi gas deviation factor at initial reservoir conditions
ZSC gas deviation factor at standard conditions
Z* gas deviation factor modified for coalbed methane reservoirs as per King (1990)
Zi* Z* at initial reservoir conditions
499
Modern Fracturing
SYMBOLS
variable used in Nolte G function analysis
Biot-Willis poroelastic constant
the difference in azimuth between the plane of the wellbore axis and the main fracture plane
Langmuir adsorption constant
rate of change of cleat or fracture compressibility
1 conversion constant for units (= 1 for coherent system, such as SI)
s the change in matrix volume resulting from the desorption or adsorption of gas (the matrix
shrinkage/swelling coefficient)
non-Darcy flow rate coefficient
geometry factor used in linear elastic fracture mechanics
well deviation from the vertical
exponent used in calculating particle hindered settling velocity
g effective non-Darcy coefficient to gas
nwb exponent used for calculating near wellbore friction
s geometry factor used in Nolte G function analysis
w exponent used for calculating pipe friction
specific gravity
shear strain
fluid shear rate
pressure drop multiplier for multiphase flow in a proppant pack
0 pressure drop multiplier for multiphase flow in a proppant pack for saturated gas in single phase
flow
acid specific gravity of acid
fluid specific gravity of fluid
g specific gravity of gas, relative to air
hc specific gravity of gas, relative to air, for gas containing nitrogen, CO2 and/or H2S
prop specific gravity of proppant grains
w wall shear rate
non-Darcy correction factor for permeability
time step for derivative calculation
K1c change in fracture toughness
L length of cement sheath
m(p) gas pseudo-pressure function difference
m(p) gas pseudo-pressure function difference for fractured well pressure build-up
m(p)im impulse derivative of Kuchuk
m(p)q gas pseudo-pressure function difference for fractured well pressure build-up after flow rate q
p pressure differential
pfriction ratio of fluid friction pressures with and without solids
pnear wellbore total pressure loss due to near-wellbore effects
ppf perforation friction pressure loss
ptort pressure loss due to tortuosity
ptotal friction total friction pressure losses during pumping (including near wellbore effects)
500
pUB optimum underbalanced perforating pressure differential
pwf flowing wellbore pressure difference (drawdown)
ta pseudo-time since shut-in
tae equivalent time
tD dimensionless time
Wichert and Aziz correction factor
strain
h horizontal strain
h, min minimum horizontal strain
h, max maximum horizontal strain
l the maximum matrix shrinkage strain for coal formations (at V = VL, the Langmuir volume)
max linear strain at infinite pore pressure on an unconstrained sample (usually coal)
x, y, z strain in the x-, y- and z-directions, respectively
diffusivity coefficient
efficiency, fluid efficiency
prop proppant placement efficiency
angle
angle between the initiating perforation and some reference point on the circumference of the
well
opening time distribution factor (UFD analysis)
wavelength
viscosity
[] intrinsic viscosity of particles in slurry
0 viscosity at zero shear rate
a apparent viscosity
e equivalent Newtonian viscosity
f fracturing fluid apparent viscosity
fluid viscosity of the slurry base fluid
g gas viscosity
g average gas viscosity
p plastic viscosity
r ratio of apparent viscosities of slurry and clean fluid
s viscosity of solvent
slurry overall viscosity of slurry
sp specific viscosity
Poissons ratio
dyn dynamic Poissons ratio
density
b bulk density
f fluid density
foam foam density
g gas density
l density of liquid phase in foam
501
Modern Fracturing
n density of nth zone

p proppant absolute density
r ratio of the densities of slurry and clean fluid
s slurry density
stress, hydrostatic stress
effective stress
1, 2, 3 mutually-perpendicular principal stresses (such that 1 > 2 > 3)
c critical stress
h in-situ stress, horizontal stress
h, max maximum horizontal stress
h, min minimum horizontal stress
o original hydrostatic stress
r radial stress
t tangential stress
T tensile strength of rock
v vertical or overburden stress
x,y,z stress in the x-, y- and z-directions, respectively
xx,yy,zz principal stress in the x-, y- and z-directions, respectively
y yield stress
* reference stress at laboratory test conditions
x* reference stress at laboratory test conditions in the x-direction
y* reference stress at laboratory test conditions in the y-direction
shear stress
fluid shear stress
o threshold shear stress for Herschel-Buckley fluid
xy,xz,yx,yz,zx,zy shear stresses acting on the x-y, x-z, y-x, y-z, z-x and z-y planes respectively
y yield shear stress for fluids
yp yield point for Bingham plastic fluids
porosity
o original formation porosity
p proppant pack porosity
p,min minimum proppant pack porosity
angle of internal friction
factor used in after closure analysis (= 16/2)
storativity coefficient
length of non-wetted zone
502
INDEX Cleat: 388-9, 392-3
permeability: 388, 392-3
A system: 385, 388-90, 396
Acid Closure pressure: 81-2, 97-9, 106-7, 109, 145, 148,
fracturing: 159-60, 205, 254, 338-40, 372, 375-6 154, 160, 284, 286, 289, 291-2, 297, 299, 301-4,
solubility: 300, 337-8 314-6, 301, 348-50, 435, 471
system: 159, 339-40 Coal: 3, 10, 134, 159, 384-5, 387-98, 410, 418-23
Adsorption: 246, 385, 387, 390-2, 400, 419, 422-3 matrix: 385, 387-8, 390-1, 395, 422
Anisotropic reservoir: 152-3, 436 Coalbed methane: 3, 9, 19, 154, 383-5, 416, 419,
Aqueous phase trapping: 266, 268, 274, 485 421-3
Areal proppant concentration: 148-50, 303, 459 Coiled tubing: 204, 242, 370, 372-3, 375-6, 378, 421,
Arps decline analysis: 70-2, 75, 87 459-60, 476-7, 486-8
fracturing: 476-7, 487
B Completion design: 169, 171, 188, 190, 193, 223
Bacteria: 257-9, 330-1, 335 Complex fracture growth: 37, 219-21, 223
Ball sealers: 160, 203, 205, 342-4, 376 Contaminants: 35, 271, 273, 301, 337
Barnett Shale: 158, 162-3, 165, 202, 222-3, 387, Convection: 110, 189, 305-6
400-20, 422, 434, 437-41, 450, 486, 488 Cooke correlation: 26-7, 36, 284-5, 310-1, 319
Bauxite: 284, 291-2, 294, 300, 302, 415 Critical pressure: 350, 388
Borate: 229-30, 242, 244-6, 252, 256-7, 259-60, 272, Crosslinked fluids: 230, 240, 243-246, 273-5, 278-9,
276, 279, 312, 318, 392 294, 307, 318, 331, 333, 349, 358, 392, 403, 409,
fluids: 244, 246, 249 417
ion: 244-5, 259 system: 331-3
Bounded flow: 45, 49, 52, 70-1, 73-6 Crosslinker: 160, 260, 313, 333-5
Bourdet: 50, 55, 87-8 Crosslinking: 228, 244-6, 252, 256, 259, 333
Brady sand: 290-1, 300, 302 time: 245, 333, 335
Breakdown pressure: 79, 93, 123-4, 435, 450 Cumulative production: 34, 318, 399, 410, 415, 437,
Brittle fracture: 117, 127-9 439, 443-5, 447-8, 465
Bulk density: 132-3, 136, 296, 299, 300
D
C Data mining: 427, 441-2, 444-5
Candidate selection: 457, 464, 466-9, 481, 483 Deformable
Carbon dioxide: 35, 170, 191, 236, 242-4, 247-53, 272, particles: 110, 315-6, 320
276, 326-7, 356-7, 387-8, 395-6, 418-9, 460-1, proppant: 316, 393
472-3, 485-6 Deformation: 116-7, 121, 132, 137, 230, 316, 343
sequestration: 396, 420, 422-3 Derivative plot: 49, 51-2, 54, 56-8, 60-1, 63, 65, 68-9,
Cased wells: 202-3 71-2, 76, 79, 80, 82-6, 107, 397
Casing: 41, 157, 170-2, 174-5, 180, 183-5, 192, 197, Desorption: 242, 250, 385, 387-8, 390-2
202-5, 213, 330, 343, 345-7, 351, 374-6, 396-7 desorption pressure, critical: 391, 420
CBM Deviated wells: 16, 181, 188, 197, 200, 203, 208, 209,
fracturing: 158-9 222, 439
reservoirs: 384-398, 417-8, 420, 423 Differential pressure: 177, 241, 273, 313, 343-4, 353,
wells: 391-3, 396-9, 418 371, 460
Cement sheath: 172-6, 178-9 Dimensionless
Ceramic proppant: 264, 284, 291-2, 294, 297, 299, fracture conductivity: 25-6, 47, 102, 142-4, 160, 221,
301, 313, 314 287, 433
503
Modern Fracturing
pressure: 21, 43 rate: 114, 156

productivity index: 26, 29, 30, 32, 142-4, 151-3, 428, loss control: 176-7, 253, 255, 374
433-4 pressure: 98, 105, 114, 146, 207, 217, 350-1
time: 21, 43, 151-3 properties: 113, 183, 227, 354
Dominant fracture: 217, 219, 353 recovery: 15, 110, 243, 248, 262, 346, 355, 472, 475
Drawdown: 24, 26, 29, 49, 52, 54, 64-6, 78, 156, 184-5, rheology: 145, 229, 235, 241, 277, 305
197, 264, 473-6 system: 102, 109-10, 268, 312-3, 330, 332-5, 356,
458, 461, 473, 484
E velocity: 27, 140, 213, 229, 285
Effective viscosity: 127, 185, 189, 227, 230, 245, 306, 331, 344,
Fracture area: 283, 306 354, 358
length: 144, 151, 154-5, 283, 357 Fluids: 19, 97-8, 138-41, 169-70, 172, 176, 183-4,
proppant pack permeability: 26, 147, 470 189-90, 192, 201-2, 227-8, 242-54, 270-3, 323,
stress: 301, 313-4, 436, 473-4, 476 357-8, 475
wellbore radius: 47-8, 288 Foam fracturing fluids: 241, 262, 278
Elastic deformation: 117, 121, 128 Forced closure: 356-7, 475
Energized fluids: 242-3, 248, 349, 351, 357 Forecast of performance: 8, 9, 16, 34-5, 71, 76, 79,
Enzymes: 249, 258-9, 261, 472 430-1, 436
Equivalent Formation
cement permeability: 173-4 boundaries: 125, 183
newtonian viscosity: 232-3 characterization: 182-3
proppant number: 160, 365 compatibility: 227, 252, 254, 460
wellbore radius: 150-1, 288 damage: 134, 168, 170, 183, 190, 197, 227, 229, 231,
Excess pressure: 59, 97, 343, 396 233, 235, 237, 239, 241, 265, 267-9
evaluation: 163-5
F fluids: 79, 135, 190, 248, 262, 273
Fann viscometer: 240-1, 246, 279, 331 permeability: 30-2, 90, 94, 101, 114, 145, 147-8, 154,
Fetkovich: 70-2, 88 156, 174, 177, 180, 242, 275, 287, 429
plot: 70-5, 80-1 rock: 156, 182-3, 262, 268
Finite conductivity fracture: 47, 51, 58, 65-6, 68-9, 71 stresses: 201, 471, 473
Flow rate: 34, 42, 45, 50-1, 93-4, 140, 173-4, 180, 185, water: 250, 273, 355, 474
187, 193, 210, 215-7, 231-2, 273, 285 Formations: 33, 93, 100-1, 113-5, 123-4, 126-7, 154,
Flowback: 162, 228, 316, 355, 357, 377, 474-6 166, 202, 242-3, 247-9, 251-3, 348-9, 355,
Flowing pressure: 152, 185, 186, 187, 285, 301, 313, 369-70, 445
317, 367, 429, 473 Forward simulation: 57-8, 62-5
Fluid Fracture
additives: 329, 333-5 acidizing: 159-60, 163, 277
efficiency: 82, 102, 109, 147, 221, 323, 347-9, 357, area: 102, 107, 109, 160-1
374, 471-2 azimuth: 15, 363, 434-5, 459, 469, 487
entry: 180-1, 186 calibration tests: 89, 90, 163
flow: 20, 29, 88-9, 93, 101, 140, 165, 186, 203, 205, characteristics: 115, 286, 450
233, 319, 356 closure: 83-5, 106-8, 349, 356, 475
friction: 96, 108, 141, 353 conductivity: 47-8, 65, 80, 148, 154-7, 159, 254, 262,
invasion: 124, 178, 267 283, 285, 287-9, 291, 293, 295, 301-9, 319-20
lag: 10, 130-1, 163 design: 30, 36, 103, 144-5, 148, 232, 309, 311, 330,
leakoff: 101, 106-9, 114, 154, 348, 435 350, 359, 450, 469-70, 480
504
models: 312, 317-8 width: 27, 47, 98, 100, 118, 127, 129-30, 154, 156-7,
extension pressure: 97, 104 220-1, 229, 233, 303-5, 313-4, 367-9, 388-9
flow capacity: 288, 303, 319-20, 359, 486 Fracture-to-well connectivity: 201-3, 205-7, 209, 211,
fluid: 205, 277, 330, 334, 343-4, 351-2, 354, 486 213, 215, 217, 219, 221, 223
leak-off: 352-4, 472 Fracture toughness, apparent: 111, 127
system: 331, 334-5 Fractured
fluids: 352, 360, 418, 488 gas wells: 25, 29, 30, 68, 176, 277, 285, 287, 365-6,
geometry: 26, 37, 41, 45, 82, 87, 97-8, 101, 103-9, 430
111-2, 115, 125, 221-2, 434, 471, 484 horizontal wells: 29-32, 369
gradient: 122, 124, 176, 301, 338, 344, 352, 377 low-permeability gas wells: 161, 358, 485
growth: 98, 203, 207-8, 214, 222, 224, 350, 356 reservoirs: 87, 181-2, 220, 392
half-length: 41, 43, 47, 63-4, 75-6, 79, 81, 87, 129, vertical wells: 27-8, 40, 43, 45, 47, 51, 61, 67, 76,
160, 264, 287, 318, 349, 444, 471 78, 81, 87, 161-2, 363, 368-9, 377, 378-9, 406,
height: 25, 68-9, 114-5, 125-6, 129, 143, 146-9, 233, 427-32
285, 350, 367, 445 Fracturing: 2, 3, 13-6, 92-4, 116-8, 153-4, 157-9, 200-3,
containment: 79, 126-7 275-8, 284-7, 363-9, 373-9, 413-5, 427-9, 433-5,
growth: 126-7, 203, 350-1, 354, 414 454-7, 477-9
initiation: 100, 111, 126-7, 206-8, 217, 219, 221, 228, applications: 102, 255-7, 289-92, 305
423, 435, 469, 478 companies: 335, 348, 355-6
length: 41, 49, 63, 68, 74, 77, 79, 104, 109, 125, 130, equipment: 111, 251, 324, 328, 372, 405, 477
146, 149, 156, 264, 285, 288, 317, 444 fluids: 96-7, 101-2, 110, 113-4, 126, 130, 226-33,
models: 98, 113, 114, 115, 125, 233, 397, 416 235, 239, 241-3, 247-51, 253-5, 257-9, 261-3,
network: 206, 212, 404, 416 267-9, 271, 273-9, 294-5, 312, 330, 475
orientation: 121-3, 435, 450 fluid, base: 235-6, 243, 248
plane: 118, 207, 209, 213, 346, 371, 482 additives: 205, 254, 258
pressure: 89, 164, 207, 338, 351, 394, 484 pressure: 97
propagation: 15, 98, 103, 114, 123, 126, 127-9, 146, selection: 268, 275, 472
149, 203, 206-7, 350, 351, 435 systems: 277, 350, 403
radius: 109, 125, 220-1 pressure: 89, 159, 164, 194, 209, 343, 351, 450
rate: 343, 352 pressures: 89, 94, 164, 222, 359
reorientation: 219, 482-4 process: 14, 96, 111-2, 278, 474
simulators: 64, 99, 104, 112-5, 119, 127, 141-2, 147, pumps: 111-2, 323, 327
150, 154 treatment: 142, 179, 235, 242-3, 251, 323-4, 329,
skin: 150, 209-10 364, 366, 402, 413
stimulation: 94, 156, 161, 180, 206, 290, 343, 345, Friction
365, 404, 423, 474 factor: 33, 141, 232, 235-7
treatment: 32, 45, 203, 283, 344 pressure: 99, 140-1, 343, 351
tip: 79, 94, 96, 98, 110-1, 114, 126, 128-31, 154, 156,
166, 207, 220, 229, 479 G
toughness: 98, 106, 113-4, 126-7, 129-31, 147, 163 Gas
treatment: 3, 45, 79-81, 104, 112, 204, 214-5, 217, formations: 242, 251-2, 267-8, 434, 473
219, 221, 253, 332-5, 346-7, 352-4, 374-6, 406 migration: 172, 177-8, 374
design: 222, 283, 397, 469, 483 reservoirs: 14-5, 37, 42, 154, 183, 262, 265, 274, 363,
volume: 98, 114, 156, 289 366, 369, 378, 383, 398, 413, 417
walls: 229, 235, 253-4, 283, 307 Geo-pressured zones: 383
Gravel pack: 193-4, 211
505
Modern Fracturing
Gravity: 60, 68, 121, 145, 148, 152, 235, 239, 291-2, fracturing: 344, 376
294, 299, 300, 306-7, 339-40, 367, 415, 429 Linear
Guar: 227, 243-6, 249, 255-7, 259-61, 277, 334 elastic fracture mechanics: 127-9
flow: 47-50, 53, 56, 62, 69, 71, 74-7, 80-5, 107, 155,
H 162, 211, 287, 427-8
Hard formations: 123, 303 formation flow: 427-8
High-permeability: 27, 58, 67-8, 184, 190, 194, 209, Liner: 181, 201, 205, 347, 371, 373-5
264, 284 LNG: 2, 4, 5, 10-3, 16, 18, 362, 426
formations: 25, 94, 111, 155, 164, 221, 287-8, 305, Longitudinal fractures: 363, 371, 374, 378
348, 355, 358 Low-permeability: 31-2, 57, 59, 61-3, 65, 68-9, 228,
fracturing: 94-5, 155, 265, 352 352, 356, 359, 378
gas formations: 111, 356-7 formations: 93-4, 110, 153, 184-5, 265, 348, 356,
reservoirs: 71, 109, 211, 264, 368 455, 474
wells: 57-8, 155, 347, 459 gas reservoirs: 59, 161, 195, 274, 276, 305, 357, 485
High-pressure: 170, 326, 334, 416 reservoirs: 25, 57, 88, 94, 109, 224, 229, 264, 267,
High-rate gas wells: 67, 155, 190 355, 368-9, 379, 488
Horizontal wells: 59, 61-2
stress: 97, 100, 122, 153, 206, 208, 221, 363, 388-9, Low-pressure reservoirs: 349, 414, 472
393, 434-6, 470, 482 LWC proppants: 284, 292, 316
wellbore: 224, 347, 371-2, 378, 394, 405, 417 460
wells: 3, 36-7, 47, 94, 135, 200, 208, 222-3, 347, M
363-71, 373-9, 409-10, 412, 434-5, 441-2, 459 Manifold: 326, 329
Horner plot: 46, 52, 54, 56, 62, 83-4 Mesh: 192, 255, 290-2, 298, 300, 302, 310, 403-4, 407,
Horsepower: 95, 243, 324-5, 327-9 415, 440
Hydrate formation: 15, 188-9 proppant: 285, 312, 404, 415-6
Hydrates: 189-90, 243, 268, 383 sand: 220, 304, 312, 393, 404, 411, 416
Methane: 16, 19, 270, 387-8, 390-1, 395, 418-9,
I 421-3
Index of anisotropy: 364 production: 395-6, 418-20, 422
Infinite-conductivity fracture: 47, 58, 62-3, 79 Micro-fractures: 131, 212-3, 413-4
Initial Mid-field tortuosity: 218
reservoir pressure: 62, 64, 398-9, 430 Minimum stress: 231, 434-5
water saturation: 398-400, 461 Multi-rate test: 51, 67
Intermediate-strength proppants: 284, 291, 292, 301 Multi-stage fracturing: 342, 344, 347
IPR: 33-4, 184-5 Multiple
curve: 184-6 fractures: 100, 116, 194, 201, 203-4, 208, 213, 217,
Irreducible water saturation: 265-6, 273 219-22, 229, 352, 365, 369, 374, 378, 459
hydraulic fractures: 214, 217, 219-21, 224, 353, 370,
J 373
Jetting tool: 204, 372, 376, 460, 476 transverse fractures: 152, 378
K N
Katz: 23, 36-7 Natural fracture: 77, 158, 170, 181-2, 208, 219-20,
228-9, 243, 249, 348-9, 386-8, 403-4, 408-9,
L 412-4, 435-6, 441
Limited entry: 185, 203, 343-4, 353, 355, 396-7 system: 181-2, 386, 414
506
Near-wellbore: 104, 159, 208, 215-7, 266 tunnel: 24, 100, 183, 185, 206, 211-2, 372, 375
area: 20, 30, 106, 110, 184, 196, 214, 217-9, 262, 352, Permeability
357, 412, 484 anisotropy: 24, 76, 80, 152, 364, 435-6
complexity: 201, 214 range: 24, 32-3, 74, 156
damage: 143, 183, 262, 358 ratio: 143-4, 389-90
fracture tortuosity: 201, 214, 217-8, 229, 274, 346, PKN fracture geometry: 103, 125
352, 375 Plastic deformation: 111, 120, 124, 128-9, 131, 134
friction: 97-8, 106-7, 213-6, 218 Polymers: 227, 235, 242-3, 247, 252, 255-61, 331, 334,
pressure: 190, 192, 201, 206 392, 403, 460
Net pressure: 97-9, 103, 107, 110, 112-5, 126, 130, Pore pressure: 97, 122, 178, 184, 389-90, 421, 451,
147, 156, 221-2, 342, 350-2, 367-9, 434, 482 458, 470-1, 473, 476, 479
plot: 115, 350-1 depletion: 471, 483
Newtonian fluid: 103, 138-40, 232, 235, 237, 238, Power law fluids: 103, 139-40, 147, 230, 231, 305
305 Pressure
Non-newtonian fluids: 138, 277, 305 analysis: 87-8, 358-9
buildup test: 89, 287, 436
O depletion: 174, 192, 395, 456, 470, 482, 487
Oil-based drawdown: 89, 143, 287
fluids: 227, 247, 327, 473 gradient: 19, 20, 187, 270, 285, 386
Open hole: 24, 25, 30, 179, 183, 185, 194, 201, 205, loss: 41, 97, 232, 235
342, 369-370, 372-3, 394, 459 match: 106-7, 112-4, 115-6
completion: 184, 369-73 response: 46, 69, 81, 103, 105, 107, 112-3, 127, 142,
Optimal dimensionless fracture conductivity: 25-6, 350, 484
428 support: 169, 287-8
Ottawa sand: 69, 283, 285, 290-1, 300, 302-3, 307, transients: 41, 89, 154
316 Pressure-dependent leak-off: 82-4, 397
Oxidizers: 246, 261 Proppant
bulk density: 299, 300
P concentration: 95-6, 106, 115, 141, 145, 147-9, 154,
Packer: 184, 205, 213, 344, 345, 351, 371-3, 476-8 156, 206, 217, 249, 303-5, 314, 354, 404, 417
Pad fluid: 127, 273, 352-3 crush: 302, 311
Particles: 164, 192, 237-9, 291-2, 296, 299, 300, 302, distribution: 147, 306-7
304, 306-7, 314-6, 319 embedment: 134, 219-20, 304, 314, 393
Perforations: 15, 24-5, 97, 99-101, 105, 106-7, 111, flowback: 110, 201, 293, 314-6, 320, 474
127, 149, 183-5, 202-212, 209, 215-9, 222, 333, grains: 155, 284, 292-3, 301, 303, 393, 474
343-6, 352-4, 370, 374-6, 396-7, 418, 476-8 laden fluid: 329, 376
damage: 206, 222 mass: 32, 318, 443, 445
design: 353, 355 material: 145, 152, 229, 234, 284-5, 302, 463, 474,
friction: 96, 98-9, 215-7, 354 478
pressure: 97, 343, 376 number: 26, 29, 68-9, 111, 142, 144-5, 148-9, 289,
length: 25 365-6, 428-9, 433
orientation: 201, 206, 208 pack: 41, 79, 109-10, 152, 242-3, 249-50, 259-61,
phasing: 25, 100, 209, 374-5 286, 292-3, 297, 301-3, 308-10, 312-5, 392,
restriction pressure: 353 474-6, 479
schemes: 202-3, 206 conductivity: 68, 87, 289, 297, 303, 393, 474
skin: 209-10 damage: 264, 305
507
Modern Fracturing
permeability: 25-6, 30-2, 228, 262, 264, 300, 302, Relative permeability: 20, 158, 262, 266, 268, 385,
306, 367-8, 392, 429 392, 439
particles: 229, 235, 237, 241, 293, 295, 303, 307, Reservoir
314-6, 458 conditions: 27, 283, 285, 313, 317, 323, 354-5, 433
permeability: 25, 94, 111, 155-6, 285, 308, 313-4, development: 398, 427, 429, 431, 433, 435, 437, 439,
474 441, 443, 445, 447, 449, 451
placement efficiency: 145, 147-9 fluids: 135, 155, 339, 458, 469
quality: 300, 317 permeability: 24-5, 30-1, 41, 69, 144, 212, 263-4, 357,
sample: 300-1, 337 366-9, 378, 414, 430, 435, 466-7, 469, 482-4
schedule: 146-9 pressure: 19-21, 33, 35, 57, 61, 89, 90, 106, 123-4,
selection: 124, 316, 320, 473-4 159, 204, 211-2, 332, 367, 388, 413-4, 473-4
settling: 234, 277-8, 307, 356, 358 pressure, average: 33, 72, 151, 182, 400, 466
size: 206, 229, 237, 297, 314-6, 337, 411 properties: 41-2, 48-9, 57, 60, 62-3, 68-9, 72, 75-6,
slugs: 106, 161, 202, 215, 217-8, 222-3, 275, 352-3, 79, 82, 135, 160, 344, 398, 413, 458
358-9, 484 quality: 405, 437, 440-1, 443, 465
stages: 95, 342, 354, 412 rock: 98, 278, 330, 344, 352-3, 408
transport: 109-10, 146, 217, 228-30, 243, 247, 253, stimulation: 88, 93-4, 162, 222-3, 275, 450
255, 274, 278, 299, 305-7, 320, 359 Reservoirs: 3, 9, 13, 16, 19, 31-2, 90, 181, 352, 363,
capability: 230, 306-7 369-70, 383, 387, 413-4, 436-7, 472
types: 41, 283-4, 289, 292, 301, 312-4, 317, 318, 337, Resin-coated proppants: 110, 201, 261, 292-3, 315,
415, 417, 445 317, 320, 488
volume: 145, 147, 154, 249, 405-6, 432, 470 Retained permeability: 145, 148, 152, 312-3
Proppants: 3, 31, 36, 109, 132, 237-8, 283-7, 289-99, Reynolds number: 20, 26, 29, 33, 232, 235-6, 285
301-3, 313-5, 317-20, 335-8, 420, 423, 440-1, Rock types: 135-6
463 Roundness: 290, 298-9, 338
Pseudo-pressure: 36, 42-5
Pseudosteady state: 53, 428, 430 S
Pumping San juan basin: 384, 387-8, 392-3, 395-7, 412, 418-9,
fluids: 166, 279, 345 421
rate: 102, 354-5 Sand: 32, 44, 192-3, 196, 283-4, 289-90, 292, 294,
Pumps: 15, 95, 97, 104, 106, 111, 218, 325-7, 329, 300-1, 304, 306-8, 317-8, 323-5, 406, 415-6,
478 471
P-wave: 136-7 jet: 476-7
proppants: 299, 301, 310, 312
R Sand-based proppants: 292, 294, 301
Radial Sand control completions: 171, 192-4, 197
flow: 23, 28, 44-53, 56-8, 60-71, 73-4, 78-85, 210, Sand-free rate: 193
308, 348, 367, 427-8, 436 Sandstone: 132, 136-7, 181, 183, 383
permeability: 436 Scanning electron microscopy: 268-70, 274, 278-9
Radius: 124-5, 130, 134, 170, 217, 367, 464 Shale reservoirs: 243, 384-6, 400-2, 408-10, 412-4,
Re-fracturing: 164, 415, 456, 468, 479-81, 482-5, 416
486-8 Shear
campaign: 480-1 rate: 137-40, 229, 231, 233, 235, 237-8, 241
operations: 455-6, 471, 478-9, 483 rates: 139, 233, 237, 239-40
treatment: 480, 482-4 stress: 118-9, 137-9, 230-1, 239, 241
Shearing time: 233-4
508
Sieve analysis: 296-7, 336-7 U
Sintered bauxite proppants: 291 Ultra-lightweight proppant: 164, 294-5, 300, 304-8,
Skin effect: 24, 47-8, 89, 90, 165, 262-3 320, 415-6, 458, 485, 487
Slickwater: 158, 229, 243, 272, 307, 347, 404, 406, Unconsolidated formations: 118, 146, 156, 284, 317,
409-10, 411, 412, 415-7, 440, 484 391
fracturing: 254, 274, 403-4, 411, 413-6, 422, 440 Unconventional
Slurry: 95, 145, 157, 172, 176-9, 229, 234, 237-8, 253, fluids: 242, 250-1, 272
261, 305, 326, 351-2, 374 fracturing fluids: 226, 242, 460
rate: 96, 99, 104-7, 148, 218 Under-pressured wells: 252, 356
Soft formations: 79, 95, 123, 134, 159, 163 Unified fracture design: 25, 67, 88, 111, 142, 160, 162,
Sphericity: 290-2, 298-9, 338 264, 319, 428, 445, 450
Stage fracturing: 343-6 Unified fracture design, applications: 36, 88, 450
Staged field experiment: 162, 358, 360
Step-down test: 98, 105, 215, 217-8, 224, 348, 478 V
Step rate tests: 104, 164, 223, 347 VES foams: 251, 272
Step-up test: 104-5 Viscoelastic fluids: 228, 242, 249, 250, 275, 392
Stress: 93-5, 100-1, 113-4, 116-23, 126-9, 131-4, 150, Viscometers: 138, 239, 332-3
156-7, 175-6, 181, 207, 209, 230-1, 297-8, 301, Viscosity: 21, 42, 93, 102, 107, 109, 118, 124,
303, 313, 315-6, 389, 434-5, 458, 471-2 137-9, 227, 229-35, 237-40, 243-55, 257, 331-3,
anisotropy: 434-5 339-41
far-field: 100, 207, 216 Viscosity, apparent: 139-40, 229, 231-5, 333
regime: 77, 122-4
Subsidence: 171, 195-6 W
Supercritical fluid: 248-9 Water-based fracturing fluids: 227-8, 243, 246-7, 254,
Superposition plot: 57-8, 63 262, 266-7, 271, 330, 357, 472
S-wave velocities: 136-7 Wellbore
construction: 169, 171, 181-2
T friction: 96, 159, 215, 349
Tight gas pressure: 59, 95, 124, 175, 179
formations: 153-4, 242, 246, 248, 347-9, 355-8 radius: 43-4, 173
wells: 41, 87, 90, 152, 155, 247, 276 stability: 370, 394, 420
Tortuosity: 96-102, 105-7, 110, 170, 206-8, 210, storage: 44, 47, 50, 52-3, 59, 60, 67-9, 89, 90
213-8, 222, 323, 333, 348, 352-3, 358, 364, coefficient: 44, 58, 68
373-5, 478-9 trajectory: 187, 210
Transport proppant: 228-9, 247, 250, 276, 284, 319, volume: 59, 116
332 White sand: 301, 312-3, 415-6
Transverse fractures: 28, 30, 32, 152-3, 208, 365,
368-9, 371, 373, 378 Z
Treating pressure: 96-8, 104, 107, 124, 179, 218, 283, Zirconium: 245-6, 254, 260
339, 344, 350-1, 353-4, 377, 397, 417, 473 Zonal isolation: 172-4, 179, 203, 205, 370-6, 378, 392,
Treatment schedule: 112, 115-6, 142 413
TSO: 25, 146, 148, 156, 349, 352, 455
Turbulence effects: 22-5, 27-9, 31-2, 111, 155, 364-5,
367-8
509

Modern Fracturing CD

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Modern Fracturing CD

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modern Fracturing CD

Uploaded by

Copyright:

Available Formats

Modern Fracturing

Enhancing Natural Gas Production

Graphic design and production: Jay Clark

Published by: Energy Tribune Publishing Inc.

for orders and customer service enquires contact: info@energytribune.com

ISBN 978 1 60461 688 0

Printed and bound by Gulf Publishing Co.

Joe Mach - February 2007

Michael J. Economides, University of Houston

Bob Bachman, Taurus Reservoir Solutions David Mack, Marathon Oil

Michael J. Economides and Tony Martin - September 2007

Table 1-2 The Worlds Largest Natural Gas Gazprom's Share

350 www.gazprom.com, ET, 2007)

The reason that Russia has given Gazprom control

over its natural gas is the so-called social obligation.

exporting the rest (2002 numbers from EIA and ET,

production peak, substantial natural gas exploitation 400

on the North Slope (about 100 Tcf ) and Cook 0

There are two major hydrocarbon producing areas Guadalupe Sonora

horizontal well technologies. Hermosillo

Figure 1-7 Natural gas resource base in Alaska (Williams

Qatar is a small, independent nation on the western

Nevertheless, a wide set of reasons leads us to believe 600

that these projects cannot even be considered marginally 500

competitive to LNG, especially when compared to the 400

economically superior LNG shipped from the recently 300

developed fields and facilities in countries such as Qatar, 200

Russia, Australia and Indonesia. In fact, as is usual for 100

large construction projects, the technical feasibility of 0

North Slope natural gas exploitation must be weighed

against the inexorable balance of the economics. This

is the bottleneck where all the advocated Alaskan gas

QatarGas to deliver 15 Mta of LNG to the UK market.

companies and national operating companies have Supply, 2006a.

One way to prevent condensate formation is

2-4 Effects of Turbulence For an isotropic formation, k equals the horizontal

units of reciprocal rate. One of the implications is h 50 ft Z 0.91 0.9

Case 2: p = 500 psi

in shots per foot (SPF). They also quantified the effect

2-5.1 Unique Needs of Fracture 2-5.2 Turbulence Remediation

where kf,n is the nominal permeability (under Darcy

value of is obtained from: 1 -5.7 13.1 9251 218

increase would be expected in an oil well flowing 25 -5.4 88.4 37020 91

Figure 2-5 Comparison of gas production rates from non-

2-5.3 Multi-fractured Horizontal Gas Wells

6. With the calculated JD,max and sc, calculate the

where JDV is the dimensionless productivity

intersecting a horizontal well (Fig. 2-8). JDTHt is the total

3 fractures intersecting a horizontal well are generated (and

and reducing the strong turbulence effects in higher-

Results and Analysis 0.1 5

for Formation Permeability from 0.01 to 10 md 0

300,000 lbm 4.0

0.01 1.42 0.755 0.19 3.0

0.1 0.332 0.235 0.95 2.0 k =0.1 md

In Eq. 2-44 through 2-47, p is in psia, q is in Mscf/

2-8.1 Pseudocritical Pressure, ppc Tpc* = Tpc1

Tpc = 187 + 330 g 71.5 g2 (2-50) = K exp X Y , (2-56)