Technical Report Documentation Page

1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

FHWAlTX-93+ 1127-3F
4. Title and Subtitle 5. Report Date
May 1993
DETAILING FOR STRUCTURAL CONCRETE 6. Performing Organization Code

7. Author[s) 8. PerForming Organization Report No.

K. Bergmeister, J. E. Breen, J. 0. Jirsa, and M. E. Kreger Research Report 1127-3F
9. PerForming Organization Name and Address 10. Work Unit No. [TRAIS)
Genter for Transportation Research
The University of Texas at Austin 11. Contract or Grant No.
3208 Red River, Suite 200 Research Study 3-5-87/9-1127
Austin, Texas 78705-2650
f-:-:~-----:-:-----:---:-":----------------I
13. Type of Report and Period Covered
12. Sponsoring Agency Name and Address
Texas Department of Transportation Final
Research and Technology Transfer Office
P. O. Box 5080 14. Sponsoring Agency Code
Austin, Texas 78763-5080
15. Supplementary Notes
Study conducted in cooperation with the U.S. Department of Transportation, Federal Highway
Administration. Research Study 77tle: "Reinforcement Detail Design in Structural Concrete"
16. Abstract

This report is the final report in a series which investigates the applications of strut-and-tie modelling
for typical details in structural concrete bridges. It summarizes the state of the art of strut-and-tie
modelling and presents specific recommendations for choosing the critical dimensions and carrying out
detailed computations using such strut-and-tie models. Separate sections treat the overall modelling
and detailing process, checking compression struts, detailing tension ties, evaluating TfT, CCG, CCT
and CIT nodes, and incorporating prestressing forces. The report includes a series of examples
showing application of strut-and-tie models in detailing deep beams, corbels, anchorage zones, dapped
ends, openings, and pretensioned beams. In addition, a number of detailing aids are included in an
appendix.

17. Key Words 18. Distribution Statement

strut-and-tie modelling, details, structural No restrictions. This document is available to the
concrete bridges, critical dimensions, public through the National Technical Information
compression struts, tension ties, anchorage Service, Springfield, Virginia 22161.
zones, beams, corbels
19. Security ClassiF. (of this report) 20. Security ClassiF. (oF this page) 21. No. of Pages 22. Price
Unclassified Unclassified 316

Form DOT F 1700.7 [8-72) Reproduction of completed page authorized

DETAILING IN STRUCTURAL CONCRETE

by

K. Bergmeister
J. E. Breen
J. O. Jirsa
M. E. Kreger

Research Report Number 1127·3F

Research Project 3-5-87/9-1127

"Reinforcement Detail Design in Structural Concrete"

conducted for the

Texas Department of Transportation

in cooperation with the

u.s. Department of Transportation

Federal Highway Administration

by the

CENTER FOR TRANSPORTATION RESEARCH

Bureau of Engineering Research
THE UNIVERSITY OF TEXAS AT AUSTIN

May 1993
 IMPLEMENTA1"ION STATEMENT

This report provides detailed background, specific recommendations

and comprehensive examples to illustrate applications of strut-and-tie
models for detailing structural concrete. It outlines the basis for detailing
complex and unusual structural applications and provides a practical way of
extending results of sophisticated analyses (such as finite element results) to
determining practical construction details. A series of specific
recommendations are provided for determining node dimensions for typical
node applications as well as determining limiting node stresses considering
confinement present. The report is aimed at improving the ability of design
engineers to provide proper reinforcement and anchorage details in
complex and/or unfamiliar design applications. Usage of these procedures
should improve the behavior of concrete structures, allow more efficient use
of reinforcement and result in lower maintenance and repair expenditures.
A series of design examples are presented to illustrate typical applications.
Study of these examples should greatly improve familiarity of design
personnel with strut-and-tie models for detailing structural concrete.

Prepared in cooperation with the Texas Department of Transportation

and the U.S. Department of Transportation,
Federal Highway Administration

iii
 The contents of this report reflect the views of the authors, who are responsible
for the facts and the accuracy of the data presented herein. The contents do not
necessarily reflect the official views or policies of the Federal Highway Administration.
This report does not constitute a standard, specification, or regulation.
There was no invention or discovery conceived or first actually reduced to
practice in the course of or under this contract, including any art, method, process,
machine, manufacturer, design or composition of matter, or any new and useful
improvement thereof, or any variety of plant which is or may be patentable under the
patent laws of the United States of America or any foreign country.

NOT INTENDED FOR CONSTRUCTION,

PERMIT, OR BIDDING PURPOSES

J. E. Breen, P.E. (Texas No. 18479)

J. O. Jirsa, P.E. (Texas No. 31360)
M. E. Kreger, P. E. (Texas No. 65541)

Research Supervisors

iv
 PREFACE

This is the final report in a series of three reports which investigated

applications of strut-and-tie modelling for typical details of structural concrete bridges.
Research Report 1127-1 looked more specifically at the problems of shear and
diagonal tension in the negative moment regions of precast girders for use with drop-
in spans. Research Report 1127-2 summarized a series of tests of typical details used
with dapped beams and several different types of nodes. This report (1127-3F)
presents a summary of the basis for strut-and-tie model use in detailing structural
concrete and includes a series of illustrative examples.

This work is part of Research Project 3-5-87/9-1127 entitled "Reinforcement

Detail Design in Structural Concrete." The research was conducted by the Phil M.
Ferguson Structural Engineering Laboratory as part of the overall research programs
of the Center for Transportation Research at The University of Texas at Austin. The
work was sponsored jointly by the Texas Department of Transportation and the
Federal Higllway Administration.

Liaison with the Texas Department of Transportation was maintained.Jhrough

the contact representative, Ms. Mary Lou Ralls, who was extremely helpful in providing
typical current details from a wide variety of projects. Mr. Eric Munley was the contact
representative from the Federal Highway Administration.

This portion of the overall study was directed by John E. Breen, who holds the
Nasser I. AI-Rashid Chair in Civil Engineering. He was assisted by co-principal
investigators James O. Jirsa, Professor of Civil Engineering (who had primary
responsibility for directing the nodal and dapped beam tests) and Michael E. Kreger,
Associate Professor of Civil Engineering (who had primary responsibility for directing
the negative moment test series). The synthesis of ideas and the development of the
initial draft of this 'final report were the direct responsibility of Dr. Konrad Bergmeister,
Visiting Engineer from the University of Innsbruck.

v
 TABLE OF CONTENTS

Chapter 1 Introduction 1
1.1 General Introduction . 1
1.2 Summary of the History of Detailing . 5
1.3 Objectives . 24
1.4 Scope . 27

Chapter 2 Background . 29
2.1 Concept Background . 29
2.2 Isolate Discontinuity or Detail Region: D - Region . 33
2.3 Elasticity Analysis Method as Load Path Method . 38
2.4 Strut Background . 41
2.4.1 Concrete Compressive Strength Limitation for Struts . 45
2.4.1.1 Effective concrete strength in
compression diagonals . 63
2.4.1.2 Two- and three-dimensional concrete strength 69
2.4.1 .3 Confined concrete strength . 71
2.5 Tie Background . 84
2.5.1 Prestressing Forces . 88
2.6 Node Background . 105
2.6.1 CCC - Nodes . 107
2.6.2 CCT - Nodes . 112
2.6.3 cn - Nodes . 119
2.6.4 TTT - Nodes . 124
2.6.5 Anchorage Requirements in the Nodal Zone . 128
2.7 Model Optimization . 133
2.8 Concrete Efficiency Factors for Design . 135
2.8.1 Unconfined Nodes and Undisturbed Concrete Struts . 135
2.8.2 Compression Diagonals . 136
2.8.3 Confined Nodes . 136
2.9 Anchorage Requirements for Design . 139

Chapter 3 Procedures . 141

3.1 General Analysis - Structural Analysis . 141
3.2 Checking and Dimensioning Concrete Compression . 143
3.3 Checking and Dimensioning Tensile Ties . 147

vii
Chapter 3 Procedures (continued)
3.4 Checking and Dimensioning Nodes:
Determining Anchorage Requirements . 149
3.4.1 Checking and Dimensioning CCC - nodes . 150
3.4.2 Checking and Dimensioning CCT - nodes . 151
3.4.3 Checking and Dimensioning cn -nodes . 157
3.4.4 Checking and Dimensioning TIT - nodes . 160
3.4.5 Curved Tensile Ties . 162

Chapter 4 Design Aids . 165

4.1 Detailing Aids . 165
4.2 Types . 168
4.3 Typical Examples of Detailing Aids . 169
4.3.1 Load Near Support . 171
4.3.2 Corbel Projecting from a Column . 189
4.3.3 Deep Beam with a Hole . 205
4.3.4 Dapped End Beam . 226
4.3.5 Anchorage Zone . 243
4.3.6 Pretensioned Beam with Eccentricity . 257

Chapter 5 Conclusions and Recommendations 271

References 273

Appendix A: Detailing Aids 285

viii
 LIST OF FIGURES

Figure Page

1.1 Typical examples of strut-and-tie models ......................... 3

1.2 Examples of flow of forces and the strut-and-tie model ............. 4
1.3 Details of special concern that may exist in actual structures ........ 6
1.4 Truss action in a cracked reinforced concrete beam ............... 8
1.5 Lower and upper bound solution .... ...... ... ... .. .. ... ... .. . . .. 13
1.6 Truss model and tied-arch model ................................ 15
1.7 Truss model for the force transfer between
the concrete and the steel ...................................... 19
1.8 Basic truss-model cube ........................................ 20
1.9 Tension and dowel force, and hoop stress close to a bar ........... 20
1.10 Two-phase model for aggregate interlock . . ... ..... ......... . . . .. 21
1.11 Semi-continuous members with dapped end ... . ....... ..... .. . .. . 23
2.1 The principle of Saint Venant ................................... 34
2.2 Subdivision of structures into Band D regions using
Saint Venant's principle ........................................ 35
2.3 Suggested subdivision of structure ............................... 37
2.4 Load path method with "u-turn" .................................. 40
2.5 Planar stresses in prism under compression ...................... 42
2.6 The basic compression fields ................................... 43
2.7(a) Fan region at beam support .................................... 44
2.7(b) Bottle shaped struts ........................................... 44
2.8 Stress-strain curves for low, medium and high strength concrete 47
2.9 Relationship of diagonal concrete strut efficiency factor
versus concrete strength ....................................... 49
2.10 Distortional effect .............................................. 51
2.11 Forces acting on edge members of parabolic arches .............. 51
2.12 Mean crack strain . ..... ... ...... .... ...... .... .. .. .. .... . .. . ... 53
2.13 Relationship between the mean crack strain and the strains
in the reinforcement for different angles of inclination
of the diagonal strut ............................................ 53
2.14 Varying strut angle versus efficiency factor for
concrete compression strength ... .. ........ . . ..... ....... . .. .... 55

ix
2.15 Mohr's circle .................................................. 58
2.16 Proposed strut-and-tie model for shear behavior .................. 66
2.17 Comparison of test results with the theoretical approach of
predicting the diagonal compression strength ..................... 68
2.18 Two-dimensional compressive strength .......................... 70
2.19 Typical geometrical data for confined core ....................... 71
2.20 Experimental stress-strain curves of 4 x 16 in.
normal weight spiral columns ................................... 74
2.21 (a) Strip load dimensions .......................................... 76
2.21 (b) Variable dimensions for geometry of the bearing and loaded plate 77
2.22 Bearing stresses versus concrete strength ........................ 78
2.23 Comparison of test results with theoretical approaches
for predicting confined concrete strength ......................... 83
2.24 Various approaches for the tensile strength of concrete ............ 86
2.25 Crack width versus tensile strength reduction factor ................ 86
2.26 The biaxial compressive-tensile strength of concrete ............... 87
2.27 Pretensioned beam ........................................... 91
2.28 Frictional loss along circular curve ............................... 93
2.29 Bond stress distribution at the end of a bond anchorage
of a pretensioned wire .......................................... 96
2.30 Strut-and-tie models for prestressed concrete ..................... 99
(a) Pretensioning force transfer ................................. 99
(b) Overall model . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 99
(c) Eccentric pretension force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 100
(d) Fully plastic strut-and-tie model " 100
(e) Strut-and-tie model for the end liD" region . . . . . . . . . . . . . . . . . . . .. 100
2.31 Strut-and-tie model for prestressed concrete beam
with curved tendon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 101
2.32 Approximation for radial compression component of curved tendon 103
2.33 Prestressed beam with parabolic tendon . . . . . . . . . . . . . . . . . . . . . . . .. 104
2.34 Width of the compression-and-tension chord . . . . . . . . . . . . . . . . . . . . .. 104
2.35 Types of nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 106
2.36 CCC nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 108
(a) CCC node with unequal pressure . . . . . . . . . . . . . . . . . . . . . . . . . . .. 108
(b) Struts created by hydrostatically dimensioned node . . . . . . . . . .. 108
2.37 Dimensions for hydrostatic stress check in CCC node " 109

x
2.38 CCC node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 11 0
2.39 Comparison of the horizontal compression strut width . . . . . . . . . . . . .. 111
2.40 Anchorage detail for CCT node:
Anchorage of reinforcement with anchor plate . . . . . . . . . . . . . . . . . . . .. 113
2.41 Anchorage detail for CCT node with directly anchored bars . . . . . . . .. 113
2.42 Equivalent concrete area approach to define tile tie width . . . . . . . . .. 114
2.43 Proposed tie width by CEB-MC - Draft 1990 . . . . . . . . . . . . . . . . . . . . .. 115
2.44 General information about the tested CCT nodes . . . . . . . . . . . . . . . .. 116
2.45 Comparison of test results with the
concrete efficiency factor of V e = 0.8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 117
2.46 Dependency of the efficiency factors for CCT node . . . . . .. . . . . . . . .. 118
2.47 Comparison of design rationale used for nodal region
of strut-and-tie model and joint of steel truss . . . . . . . . . . . . . . . . . . . . . .. 120
2.48 Geometrical approach to define the strut width for cn node . . . . . . .. 121
2.49 Dependency of the compression strut width for cn node . . . . . . . . .. 124
2.50 Tensile strength of concrete implicitly utilized . . . . . . . . . . . . . . . . . . . . .. 125
2.51 Conservative starting point for computing development length . .. . .. 126
2.52 Positive and development length anchorage details , 127
2.53 Lateral pressure on reinforcement bar . . . . . . . . . . . . . . . . . . . . . . . . . .. 130
2.54 Lateral pressure and the distance "e" to the reinforcing bar . . . . . . . .. 131
2.55 Comparison of a theoretical approach and test results for the
development length of straight bars
with confinement from bearing plates . . . . . . . . . . . . . . . . . . . . . . . . . . .. 132
2.56 Design approach for concrete efficiency factor . . . . . . . . . . . . . . . . . . . .. 138
3.1 Design procedure for concrete structures . . . . . . . . . . . . . . . . . . . . . . .. 142
3.2 Compression fields and strut-and-tie model . . . . . . . . . . . . . . . . . . . . . .. 143
3.3 Comparison of various diffusion angles . . . . . . . . . . . . . . . . . . . . . . . . .. 146
3.4 Proposed diffusion angle for design , 146
3.5 Width of the reinforcing tie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 147
3.6 Assumption for the biaxial compressive-tensile strength . . . . . . . . . . .. 148
3.7 Strut width for smeared node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 149
3.8 Geometrical relation for CCC node for dimensioning process . . . . .. 152
3.9 Geometrical relation for CCC node with borders
not parallel to the compression strut . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 153
3.10 CCT node with single straight reinforcement bar . . . . . . . . . . . . . . . . .. 154
3.11 CCT node with hooked bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 155

xi
3.12 CCT node with multiple reinforcement bar layers . . . . . . . . . . . . . . . . .. 156
3.13 CCT node with single reinforcement bar layer . . . . . . . . . . . . . . . . . . . .. 158
3.14 CCT node with multiple reinforcement bar layers . . . . . . . . . . . . . . . . .. 159
3.1 5 Special anchorage devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 160
3.1 6 Tn node with looped bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 161
3.17 Dimensions for curved tensile ties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 163
4.1 Types of nodes , 169
4.2 Examples to be presented in detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 170
4.3 Elastic finite element analysis: principal stresses . . . . . . . . . . . . . . . .. 173
4.4 Strut-and-tie model for load near a support , 174
4.5 Node 51- CeT node '" .. .. . .. .. . . .. 183
4.6 Node Q. - eec node . .. . .. .. . .. . .. . .. .. 187
4.7 Reinforcement layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 188
4.8 Typical cracking patterns of corbels . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 189
4.9 Possible failures of corbels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 191
4.10 Stress trajectories in a homogeneous elastic corbel . . . . . . . . . . . . . .. 192
4.11 Diagonally reinforced corbels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 193
4.12 Failure mechanism in corbels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 193
4.13 Strut-and-tie model for corbel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 194
4.14 Corbel strut-and-tie model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 195
4.15 Strut-and-tie model results compared with test results . . . . . . . . . . . . .. 196
4.16 Strut-and-tie model for corbel projecting from a column . . . . . . . . . . .. 199
4.17 Anchorage detail for corbel design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 203
4.18 Reinforcement layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 204
4.19 Geometry for determining stress concentration factors . . . . . . . . . . . . .. 206
4.20 Principal tension trajectories and
reinforcement for corner in tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 207
4.21 Strut-and-tie models for deep beam with a hole . . . . . . . . . . . . . . . .. .. 211
(a) Example 4.3 dimensions . . . .. . .. .. . . . . . . ... . . . . . . . . . . . . . . . .. 211
(b) Finite element analysis contours for similar structure
with load placed farther to right . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 212
(c) Model 1, left side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 213
(d) Model 2, left side . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 214
(e) Model 1 - 50% of load (left) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 215
(f) Model 2 - 50% of load (left) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 216

xii
4.21 Strut-and-tie models for deep beam with a hole (continued)
(g) Model 3 - 100% of load (right) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 217
(h) Combined strut-and-tie models " 218
4.22 Reinforcement layout for deep beam with a hole . . . . . . . . . . . . . . . . .. 225
4.23 Reinforcement layout for strut-and-tie model ST1 " 227
4.24 Reinforcement layout for PCI detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 227
4.25 Reinforcement layout for Menon/Furlong - detail . . . . . . . . . . . . . . . . .. 228
4.26 Reinforcement layout for modi"fied strut-and-tie model ST2 . . . . . . . .. 228
4.27 Orthogonal strut-and-tie model " 230
4.28 Diagonal strut-and-tie model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 230
4.29 Proposed strut-and-tie model for dapped end beam . . . . . . . . . . . . . .. 231
4.30 Proposed strut-and-tie angle for dapped end beam . . . . . . . . . . . . . .. 232
4.31 Comparison with the proposed strut-and-tie model
using test results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 233
4.32 Strut-and-tie model for example: dapped end beam . . . . . . . . . . . . .. 236
4.33 Reinforcement layout for dapped end beam . . . . . . . . . . . . . . . . . . . . . .. 242
4.34 Possible configuration for single anchor . . . . . . . . . . . . . . . . . . . . . . . . .. 245
4.35 Concentric single anchor: geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 246
4.36 Comparison of finite element analysis with results from Guyon . . . . .. 247
4.37 Comparison of two different strut-and-tie models
with principal stress vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 249
4.38 Proposed strut-and-tie model for anchorage zone . . . . . . . . . . . . . . . .. 250
4.39 Strut-and-tie model for example: anchorage zone . . . . . . . . . . . . . . .. 253
4.40 Reinforcement layout for anchorage zone . . . . . . . . . . . . . . . . . . . . . . .. 256
4.41 Pretensioned beam: geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 260
4.42 Strut-and-tie model for prestressed concrete . . . . . . . . .. 261
(a) Transfer length forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 261
(b) Model with tendon eccentricity effects . . . . . . . . . . . . . . . . . . . . . . .. 261
(c) "0" region at end . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . .. 262
4.43 Reinforcement layout for pretensioned beam . . . . . . . . . . . . . . . . . . . . .. 269

xiii
 LIST OF TABLES

Table Page

1.1 Statistical data from comparison with different models . . . . . . . . . . . . . . 16

2.1 Analysis usage ................................................ 32
2.2 Statistical parameter for shear tests .............................. 57
2.3 Efficiency factor proposed in the CSA and by Macgregor ........... 60
2.4 Efficiency factors proposed by Schlaich et al. and CEB ............. 62
2.5 Values for 'tRD according to CEB MC - Draft 1990 .................. 64
2.6 Comparison with the proposed strut-and-tie model and test data 67
2.7 Statistical data of the comparison in Figure 2.17 ................... 68
2.8 Strength ratio for three-dimensional
concrete compressive strength .................................. 69
2.9 Statistical data from Figure 2.18 ................................. 70
2.10 (Number inadvertently skipped - Table 2.11 is next following Table 2.9)
2.11 Statistical data for confined concrete strength with various test data ... 81
2.12 Statistical data from Figure 2.23 for confined concrete
with an efficiency factor Ve =0.5 + 15/(f dO.5 . ••••••. . •••••••••. •. . • 82
2.13 Statistical data for confined concrete with
an efficiency factor Ve = 0.5 + 20/(f dO.5 • ••••••. •••••••••. •••. ••. •• 82
2.14 Friction and wobble coefficient .................................. 93
2.15 Loss of prestress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.16 Design steps for 'B' regions of prestressed beams
using the strut-and-tie method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 102
2.17 Statistical data from Figure 2.45 (CeT node) . . . . . . . . . . . . . . . . . . . . .. 118
2.1 8 General information about the tested nodes cn
. . . . . . . . . . . . . . . . .. 122
2.19 Statistical data from comparison in Figure 2.55 . . . . . . . . . . . . . . . . . . .. 131
4.1 Statistical analysis from Figure 4.15 omitting unreinforced specimen .. 196
4.2 Statistical data from Figure 4.31 " 233

xiv
 SUMMARY

This report is the final report in a series which investigates the

applications of strut-and-tie modelling for typical details in structural concrete
bridges. It summarizes the state of the art of strut-and-tie modelling and
presents specific recommendations for choosing the critical dimensions and
carrying out detailed computations using such strut-and-tie models.
Separate sections treat the overall modelling and detailing process,
checking compression struts, detailing tension ties, evaluating TlT, GGG,
CGT and CTT nodes, and incorporating prestressing forces. The report
includes a series of examples showing application of strut-and-tie models in
detailing deep beams, corbels, anchorage zones, dapped ends, openings,
and pretensioned beams. In addition, a number of detailing aids are
included in an appendix.

xv
 CHAPTER 1. INTRODUCTION

1.1 General Introduction

A structure must be safe, serviceable and durable during its lifetime. In

addition, a designer must always be aware that a structure should be practical
and economical to construct and should fulfill not only functional but also
aesthetic needs.

Concrete structure construction is marked by increasingly versatile appli-

cations. The wide spectrum includes precast and cast-in-place concrete with
conventional pretensioned and lor post-tensioned reinforcement. The term
structural concrete is used to define the global spectrum for the different concrete
structures and covers allioadbearing concrete, including reinforced, prestressed
and also pla.in unreinforced concrete, if the latter is part of a concrete structure.
This definition purposely eliminates traditional expressions such as "reinforced
concrete", "prestressed concrete" and "partially prestressed concrete". It was
chosen to emphasize that a traditional "prestressed concrete" beam in fact may
have "active" or "prestressed" reinforcement for a portion of the flexural resis-
tance, "passive" or "non prestressed" reinforcement for the remainder of the
flexural resistance and for local crack control as well as for shear - diagonal
tension resistance, and additional "passive" reinforcement for anchorage zone
resistance to bursting and spalling stresses. The basic design principles for
selection of these various reinforcements are essentially identical at the ultimate
limit state and can be consistently treated by well recognized similar analysis
techniques at the serviceability limit state. It will greatly reduce confusion and
possible error if a consistent treatment can be developed for

1
 2

determination of the internal forces in the members of the selected structural

system. The action forces (external forces) on a structure must be in equilibrium
with the resisting forces (internal forces). Most structural analyses use an elastic
system which combines equilibrium, geometric constraints and continuity to
determine the action forces. These analytical models condense linear structural
members along their centerlines and condense slabs and shells in their middle
plane. Modern computer methods provide rapid and efficient solutions for these
action forces (including support reactions) as well as the determination of
sectional resistance forces such as axial load, shear, bending moment and
torsional moment acting on a specified cross section. Both linear elastic analysis
and simplified plastic analysis procedures have been widely accepted for such
determination. However, to reasonably dimension a concrete structure, choose
cross-section dimensions and specify reinforcement quantities and patterns,
further knowledge of material properties and internal force distributions are
required. For a better understanding of the distribution of internal forces in a
concrete structure, applications of strut-and-tie- models are helpful [1]. Struts and
ties condense the real stress-fields and internal forces of a structure along
straight or curved lines and concentrate the curvatures of the force paths in
nodes, as shown in Fig. 1.1. In these examples dashed internal lines indicate the
compressive struts while solid internal lines indicate the tension ties. The small
circles representthe nodes where the struts andties intersect. Becausethe strut-
and- tie model is a conceptual model, it enables the designerto visualize the flow
of forces within the structure (Fig. 1.2) and is of particularly great assistance in
proportioning reinforcing.

A consistent design approach for a structure is attained when its tension

members, compression members, and the interconnecting nodes with their
specific joining requirements are designed with proper regard to ensure safety,
serviceability and durability.
 a. Moment Opening Comer
3

..: .. ..
-,

J
~
:

b. Moment Closing Corner

) .. :J
. .11
c. Corbel d. OeepBeam

...

Figure 1.1: Typical examples of strut- and- tie- models (from Ref. [2])
 4

·· ..
".1&
,.. , ...
\
\
"
_.,
••• + +++ "".,
..... ..1"'1-++ + -'"'¥ J( "
• '. ~ J. ..,. -t- ..... -t- ~ .. ~
,-.. ... .,. + ...:''',
• .0-. , I ' --0--.
• • .r.,..,.. + + ... of' , .
".
•

,."
• • .t + +

• • • • • +++ 1.'
+ + + + ,.. •

1'9f.""'i~•
•• " • • t • • •" "

.............•
, • • t t • t t ." ,
• • • • • • • • • I , ,
•
tension

., •
1
11·11"

.... .... .
•••••••

11,111'1111'
,
•••••
,.'

",
I
I
compression

11'IIIII'I'll

w(x) = P I (b h)

Figure 1.2: Examples of flow of forces and the strut- and- tie- model
 5

1.2 Summary of the History of Detailing

Design of a concrete structure is a step by step procedure beginning with

estimation of the loads acting 011 the structural system, preliminary proportioning
o'f relative stiffnesses of the structural system, carrying out a detailed and often
iterative general structural analysis, followed by the sizing or proportioning of
members and concluding with the detailing process. In concrete structures,
detailing would encompass:

preparation of drawings showing the size and location of structural

elements and reinforcement, and specifying the required concrete
strength
specification of bar details such as anchorage provisions and locations
of splices and overlaps
specification of time dependent quality control requirements

In detailing a concrete structure particular attention must be paid to the

statical or geometrical discontinuities, shown in Fig. 1.3, such as concentrated
loads (statical) or frame corners, corbels, recesses, holes and other openings
(geometrical). The proper detailing for such areas of special concern is essential
to overall structural integrity. While engineers are usually well trained in analysis
procedures and the basic mechanics of structural concrete, there is not a general
methodology for detailing. This often presents the designer with numerous
difficulties. The different codes and standards (AASHTO [3], ACI 318-89 [4])
propose empirical recommendations for some specific applications. However,
the design standards cannot include the innumerable details that may arise. Most
text books emphasize the basic mechanics and are vague or illusory regarding
details for irregular members.
6

Point Load

""':1 Corner

I
Corbel Hole

Figure 1.3: Details of special concern that may exist in actual structures
 7

Since all parts of a structure including the discontinuity regions are of similar
importance, an acceptable design concept must be based on a physical model
with a logical understanding. Truss models, because of their transparency and
adaptibility to many design situations, are seen as attractive alternatives to
empirical approaches for detailing structural concrete. Truss models for shear
design of reinforced concrete beams were introduced by Wilhelm Ritter [5] in
1899. Ritter introduced his model to dispel the idea that the main function of the
stirrups was to resist horizontal shearing stresses by a dowel-type action for
which vertical wooden pegs were used in timber beams. Morsch [6] in 1902
presented the truss analogy for the design of web reinforcement based on
laboratory tests. In the truss model for shear, the reinforced concrete beam is
represented by an analogous truss. A typical reinforcement scheme in a cracked
reinforced concrete beam will mobilize the truss action as shown in Fig. 1.4(b).
The flexural concrete compression zone is thoug ht of as the top chord ofthe truss
while the tensile reinforcing forms the bottom chord. The top and bottom chords
are connected by stirrups acting as vertical tension hangers and pieces of
concrete between diagonal tension cracks acting as compression struts.

In 1906 Withey [8] introduced Ritter's equation into the American

literature. He found that this equation gave tensile stresses in the stirrups which
were too high when compared with values obtained from actual test results.
Withey indicated that the concrete of the compression zone may carry
considerable shear even after the web below the neutral axis is cracked in
diagonal tension. He also indicated a possible vertical shear transfer by dowel
action of the longitudinal reinforcement.
8

a. Stress Trajectories and B- and D-Regions

r++
-..fe ...... _;.. . . . . -.".~ .... -1!J~ • • • ~••- .. - -e;-
"Ie;> • - •. • •• ._w- -, • e.
.....
. --c: .I,
t(e,_ ..

-...'.:.
• ..- ••- .- •••• ••••• ••• e. •••

....
.:.'
.,..
'
•• • "'.

t.- • •
:-_ •
..-.' .-•

t...,...
•• • • • ••
',~..
'.•
•

'"
••
-.
•e. •• •

",....

b. Truss Model or Strut-and-Tie Model

~p

~~ 4~
P/2 (c) Actual Reinforcement Layout P/2

Figure 1.4 : Truss action in a cracked reinforced concrete beam (from Ref. [7])
 9

In 1909, Talbot [9] presented a study of web stresses, including tests of

188 beams. The conclusions of this report are indeed important. In particular, the
conclusion referring to beams with stirrups said:

Stirrup stresses computed by Ritter's equation appear too high. It is therefore

recommended that stirrups be dimensioned for two-thirds of the external shear,
the remaining one-third being carried by the concrete in the compression zone.
It willbe found that the value ofnominal shearing stress will vary with the amount
of reinforcement, with other factors which affect the stiffness of the beam.

The stiffer the beam the larger the vertical stresses which may be developed.
Short, deep beams give higher results than long slender ones, and beams with
high percentage of reinforcement than beams with a small amount of metal ...
In beams without web reinforcement, web resistance depends upon the quality
and strength of the concrete ...

Unfortunately Talbot's findings were not expressed in mathematical terms, and

became lost as far as design equations were concerned. The National
Association of Cement Users, the forerunner of the present American Concrete
Institute, published its first code recommendations in 1908 [10]. This report was
essentially based on what has later become known as ultimate strength design.
The various sections were dimensioned on an ultimate basis for a load four times
the total working load.
10

In 1927, Richart [11] expressed the shear capacity of concrete beams with

vertical stirrups by an equation of the form:

v =
where
v = nominal unit shearing stress in concrete
r = aJ (s b)
ay = cross sectional area of web reinforcement
s = spacing of web reinforcing bars, measured at right angles to
their direction
b = width of beam
fy = tensile unit stress in web reinforcement
C = factor which varies between 90 and 200 psi (depends upon the
the percentage of web reinforcement used and also on the
quality of the concrete).

This expression indicates that the computed stresses from the truss model were
lower then the measured stresses. The factor "C" was included to express
the additional mechanism for shear behavior, like aggregate interlock (friction),
dowel action etc. These basic ideas found wide use in American design
standards throughout the Twentieth Century. Since the majority of members
designed were subjected to only low or moderate shear levels, an empirical "C"
or concrete contribution (Vc) was introduced to supplement the truss model
capacity (VJ The present US expressions [3, 4] for shear capacity are of the
pattern:

Vu =:;; cj> Vn = cj> (Vc + Vs )

 11

Continued use of the supplementary "Ve" term in US practice was encouraged

by more contemporary leaders such as Hognestad [12] who stated in 1951 that
if designs are made on an ultimate limit state basis the truss model will not result
in safe and economic structures under all conditions met in practice.

Kupfer [13] in 1964 developed an expansion of Morsch's truss analogy by

application of the principle of minimum strain energy.

Forthe last 25 years, researchers in Europe and North America have been
working with the goal of developing aconceptual model to properly representthe
behavior of concrete members subjected to torsion and shear. The main
objectives were to rationalize and at the same time simplify the design
procedures in these areas. Lampert and ThOrlimann [14] developed a conceptual
model based on theory of plasticity. The theory of plasticity provides a
mathematical basis for collapse load calculations. Using a yield condition, a
mathematical description forthe ultimate stresses can be developed. Given a
set of generalized stresses, 0'1' 0'2'" an the yield condition is a function
f(11' 0'2' •• an) = O. The yield condition can be visualized as a surface in n-
dimensional space.

If f < 0, the point determined by the generalized stresses lies within the
surface and does not give yielding. The condition f> 0 implies a point outside the
yield surface which corresponds to stresses that cannot occur. The flow law is
a second major concept in plasticity. The flow law is defined as

Cj = f / ('A, a) = 1,2.. n
c j = generalized strain corresponding to a j

'A, = positive constant

12

Starting from the yield condition and How rules it is possible to derive the
. theorems of limit analysis. The lower bound theorem states (see Fig. 1.5 (a)):
A.i OJ:::;; A.U OJ

A load system based on a statically allowable stress field which does not violate
the yield condition is a lower bound of the ultimate load.

The statically allowable stress distribution must satisfy the equilibrium

equations and the statical boundary conditions. From this it follows that the strut-
and tie- model is a lower bound solution. Use of the lower bound theorem will in
all cases be conservative. The upper bound theorem states (see Fig. 1.5 (b)):

A load system which is in equilibrium with a kinematically allowable mechanism

and compatible with the geometrical boundary conditions is an upper bound of
the ultimate load.

Solutions forthe upper bounds are derived by equating the external work
done to the internal energy dissipation for the assumed mechanism. Upper
bound solutions are generally unconservative. The theory of plasticity states that
there is a unique and exact solution such that both the upper and lower bound
theorems are satisfied. The quality of a plastic analysis is dependent on the
constitutive equations used. These constitutive models of material behavior
define the yield condition which determines failure of the plastic model. The way
constitutive equations (most are empirically derived) are handled by the models
will be discussed in Chapter 2.

Various researchers, including Leonhardt [15] , ROsch [16], Lampert [17],

Grab [18], LOchinger (space truss) [19], MOiler (optimum inclination of the
diagonal members) [20], Neilsen et al. [21], Mitchell and Collins [22], Ramirez
and Breen [23] have worked to refine and expand the method so it is applicable
 13
M1 =-FI/3

M2 = + F I 19

M1 = + 2 F I I 9

(a) lower bound solution

~~
Mplastic

1/3 1/3 1/3

(a) upper bound solution

Figure 1.5: Lower- and upper bound solution

14

to shear, torsion, and the interaction of these actions, as well as bending. The
space strut - and- tie-model with variable angle of inclination of the compression
diagonals departs from the traditional truss model with 45 degree angle
diagonals. The angle is chosen such that in the field where failure occurs, both
the longitudinal and transversal (stirrups) reinforcement will reach their yield
strength. In this case a sufficient sheartransfer by aggregate interlock across the
initial inclined cracks is assumed so that the concrete diagonals can reach their
final inclination under ultimate load. Due to the fact that such shear transfer
across a crack decreases with increasing crack widths (rough crack model [24,
25]), additional considerations become necessary. Hence, limits on the inclina-
tion of the concrete diagonals must be introduced. The model is valid in the
complete range of interaction between general bending, norma.l force, shear
force, and torsion. However, limits must be set in some fashion to preclude initial
compression failures. Recently, MacGregor and Gergely [26], Marti [27],
Schlaich et al. [28], Schlaich and SchaJer [2] have published refined methods for
detailing structures using strut- and- tie- models. In the Canadian CSA-Standard
[29] the compression field theory an idea somewhat simi larto the strut- and- tie-
I

model, was introduced in 1984. Cook and Mitchell [30] published studies on
regions near discontiunities. The strut- and- tie- models were compared with a
nonlinear finite element study and test results.

The "Design and Construction Specifications for Segmental Concrete

Bridges" [31] introduces also the strut-and-tie-model as a design tool for areas
where the strain distribution is non-linear.

For prestressed beams with unbonded tendons, Kordina et al. [32]

comparedtheirtest results with the truss- model (Fig. 1.6(a)) and also with atied-
arch model (Fig. 1.6(b)). According to the truss model, the shear-carrying
capacity increases approximately linearly with the amount of web reinforcement,
whereas with the tied-arch model the shear-carrying capacity depends only on
the load-carrying capacity ofthe arch orthetension chord. The comparison ofthe
 15

test results with the truss model (45-deg truss was assumed) showed that the
shear capacity could be predicted best for unbonded prestressed beams.
Conversely, for a tied-arch model, an insufficient agreement with the test results
was obtained; only the compression-arch failure was considered as a failure
mode. Measurements indicated thatthe stirrup stresses in beams with unbonded
prestressing do not differ in principle from comparable beams with bonded
prestress reinforcement, as far as the shear -carrying system is concerned.

Ics.atitld crock pattern

(a) Truss model (b) Tied-arch model

Figure 1.6 : Truss model and tied-arch model ("from Ref. [32])
16

Similarto the case of bonded prestress, the stirrup forces are reduced by a
part of the shear force carried by the concrete, including parts of the shear
force carried by aggregate interlock, the compression zone, and the dowel
action of the longitudinal reinforcement.

Recent studies by Hartmann, Breen and Kreger [33] compared the

truss model with test results for prestressed concrete girders using concrete
strength in the range of 12,000 psi. The 45 degree truss model gave very
conservative results when using a concrete efficiency factor of 0.5. Table 1.1
contains the statistical comparison of the analysis with different models.

Table 1.1: Statistical data from comparison with different models

Method Experiment/theory: mean standard deviation

ACI [2] 1.18 0.18

CSA [28] 1.74 0.78

Truss: ve = 30 / (f'e)o.s· 1.72 0.26

Truss: v e = 0.5 (f'e) 1.38 0.32

It is interesting that the traditional ACI- AAS HTO shear expressions [3,4] provide
the closest agreement, once again indicating that the use of a supplementary "Ve"
term is important for economy in beams with carefully controlled loading.
Presence of precracking of the webs due to other loading patterns would greatly
affect and could substantially diminish any Vecontribution. One of the problems
in evaluating test data and comparing it witll the truss model is to define the strut
width, strut angle and the efficiency factor of the concrete.
 17

Powers [34] studied prestressed girders with high strength concrete.

Evaluation of the test results and the comparison with the strut-and-tie-model
showthat even for prestressed girders with high strength concrete, failure can be
predicted on the basis of a strut-and-tie-model. For the web-crush ing failure the
limitation of the efficiency factor of the concrete is highly important.

Schaefer [35] and Castrodale [36] have shown that there is good agree-
ment between the truss model and observed test results in both reinforced and
prestressed concrete beams subjected to different loading combinations of
bending and shear which extensively precrack the girders. Experimental results
of the behavior of reinforced concrete beams with various arrangements of
stirrups were compared with the strut- and-tie model by Kotsovos [37]. He
concluded that in general the strut- and- tie- model does not provide any detailed
information with regard to the strength and deformation of concrete.

Related work on this project by Barton [7] studied the application of a strut-
and- tie- model to beams with dapped ends. The various singular nodes that
may occur in the strut- and- tie- models were studied by Anderson [38] and
Bouadi [39]. Their detailed observations are summarized in the accompanying
report 1127-1.

Noguchi and Watanabe [40] applied the strut- and-tie- model based on a
'fin ite element study for the shear resistance mechanisms to beam-column joints
under reversed cyclic loading. The strut-and- tie- model gave good agreement
with the shear stress distribution model for all test specimens. Breen and Stone
[41], Burdet [42] and Sanders [43] investigated strut-and-tie-model approaches
based on elastic 'finite element studies and experimental tests, for the design of
post-tensioned girder anchorage zones.
18

The approaches of the various authors cited differ in the treatment of the
prediction of ultimate load and the satisfaction of serviceability requirements.
Schlaich et al. [28] proposed in general to treat the ultimate limit state and
serviceability in the cracked state by using the same model. This was to be done
by orienting the geometry of the strut- and- tie- model based on elastic stress
fields and by analyzing the resulting strut-and-tie-model structure following the
theory of plasticity. A computer based design approach based on these ideas
was developed by M. Schlaich [44].

The concept of a strut- and- tie- model can be used not only for statical
or geometrical discontinuities but also for other load transfer mechanisms like
anchorage provisions, dowel action and force transfer between concrete and
steel. Yankelevsky [45] described a truss model forthe force transfer between the
concrete and the steel by using static equilibrium and compatibility to relate the
forces. By knowing the axial force in the steel (a differential equation was solved
forthe axial force in the steel), the bond shear stress was predicted and was of
an exponentially decaying form, maximum at the bar's ends and minimum at the
specimen's midspan (see Fig. 1.7). Another application of strut- and- tie- models
to details is the three di mensional truss model suggested forthe fracture behavior
of concrete. Rode [46] used a three dimensional truss model cube (Fig. 1.8) for
a computer simulation to study crack opening and crack growth. The model
conception is based on a 1941 idea of Hrennikof'f [47] for the solution of linear
elastic continuum problems by a three dimensional framework method. This
model allows simulations on micro- and macro levels without altering the number
of elements. The basic cell is a truss cube with edge struts, surface diagonal
struts and space diagonal struts. The struts themselves behave linearly elastic
up to given strain rates. On exceeding the maximum tensile or compressive
strain, the affected struts are removed from the system, representing cracks. A
 19

Compression In concrete

Compression in Radial tension In concrete

concrete

Tension In bar

Compression In concrete

Figure 1.7: Truss model for the force transfer between the concrete
and the steel (from Ref. [45])
20

- - edge strut
--- surface diagonal strut
_._.- space diagonal strut

Figure 1.8: Basic truss-model cube (from Ref. [46])

1.00 VINTZELEO~ '•••••..•~.:.tr

d . .",..; ~v
Vou/vtu -1-./ ., ,',,"
...... ~ .. J.~' •.
.•~:.; .;..~ .:-.e.:. "=-' 'J
0
"T
0.75 to ~
P M
a= o.4Sd
0.50

0.5 d=24 mm
0.25 fs y • 440MPa
v:U [17]11: 0.7Vtu [18]
T:=Tsy
0.0 0.25 0.50 0.75 Tu/T u *

Figure 1.9: Tension and dowel force, and hoop stress close to a bar
(from Ref. [48])
 21

single strut represents the stress flow mainly through an aggregate particle,
another one represents the 'flow main Iy th roug h tile mortar matrix, and a third one
is affected by the bond between matrix and aggregate. The strut parameters are
stochastically endowed with values by a computerized random number process.
All quantities of the strut parameters are normally distributed with a variance of
50%.

Soroushian et al. [48] and Vintzeleou [49] studied the dowel action with
regard to bond, tensile strength of concrete, and bar- to- stirrup interaction with
stirrup tension stiffening (see Fig. 1.9).

Different studies were done to investigate the shear transfer mechanism

based on aggregate interlock (Bazant and Gamabarova [50], Divakar, Fafitis and
Shah [51]). The two-phase model (Walraven [52]) has a rational formulation
based on a few assumptions. (See Fig. 1.10.)

AJ]II 11II11111 II II

= a me (Aw + I..l Ay )
= a me (Ay - I..l A w )
= 0.394 fIe 0.56 (matrix compressive strength)
0.4 (friction coefficient)

Figure 1.10: Two-phase model for aggregate interlock

22

The principal factors affecting the aggregate interlock are:

quality of the concrete. Usually the top part of a member, because of the
particle sedimentation and water gain under the coarse aggregate, will
contain weaker concrete.

the size of the crack width. Smaller crack widths lead to larger shear
stresses, but also to more sudden failures.

Based on the interface shear transfer tests conducted by Walraven [52],

an approximate expression to limit the shear stress along the crack was given by
Cook and Mitchell [30]:

Vmax = 2.168 (f'c)o.s / (0.31 + 24 w / (da + 0.63))

Vmax = max shear stress [psi]

w = crack width [in.]
da = maximum aggregate size [in.]

Gamabarova [53] compared the truss model with test data and found that
for thin-webbed I beams the truss-model is a quite conservative approach.
Therefore, more realistically the aggregate interlock should be taken into
account for sheardesign. Thisconclusion reemphasizes the findings of Talbot
[9] and Richart et al. [54] in the 20's and Hognestad [12] in the 50's that a
concrete term was needed to amplify the truss model for economical design of
lightly loaded members. Also Brandtzaeg [54] concluded in his theoretical
"analysis of stresses in a material composed of non-isotropic elements" that the
limiting value of shearing stress is the sum of the shearing strength of the
material and the coefficient of internal friction times the normal stress.
 23

There are several additional load-carrying mechanisms that can supplement the
basic truss model:

dowel action of the longitudinal bars (vertical)

aggregate interlock transfer forces across a crack (vertical and normal)
component of inclined prestressing tendons (vertical)

These components offorcetransfer must eitherbe neglected ortreated indirectly

in the strut- and- tie- model. In lightly loaded members with low levels of shear,
such components are significant and some supplementary design mechanism
continues to be necessary for economy [55]. Current design codes and standards
worksatisfactorilyfortypical "traditional" members such as uniform depth beams
with well distributed loading. Empirical solutions at supports have been devel-
oped to give good designer guidance for such cases. However when irregular
members are used, such guidance is limited. For example, the increased usage
of concrete bridge substructures and superstructures in highly congested urban
areas has caused increasing complexity in bent cap geometry, the introduction
of new cross section shapes and the increasing usage of precast, prestressed
beams cantilevered over a support to a hinge away 'from the support. Simply
supported drop-in beams are placed from one hinge to another hinge at the other
end of the drop-in beam. (See Fig. 1.11.)

Figure 1.11: Semi-continuous members with dapped end

24

This generally necessitates large notches or "daps" being used at the cantilever
span end and the ends ofthe drop-in span. The many unusual bent configurations
in congested urban areas result in many highly loaded short bracket and corbel
applications in reinforced and post-tensioned concrete members. The increasing
usage of pretensioned concrete in unusual long span situations as well as in
massive bent caps creates a host of new applications of reinforced, pretensioned
and post-tensioned members and assemblages. Traditional code rules and
simple reinforcement patterns based on the simple span test specimens utilized
for experimental determination of so manyoftheACI [3] and AASHTO [4] design
provisions do not provide guidance and are not applicable to many of these new
applications.

In order to avoid potentially serious strength or serviceability problems,

better guidelines for proportioning and locating reinforcement are needed. Such
guidelines should considerthefull range of reinforcement from the passive action
of nonprestressed bars to the active action of prestressing tendons, as well as the
case of mixed reinforcement (active and passive) wh ich is becorn ing widely found
in post-tensioned concrete. Comprehensive detailing methodology and guides
are needed in practical detailing.

1.3 Objectives

The general practice in detailing structural concrete has been based on

experience, rules of thumb or highly specific standard details. Most of these
methods do not incorporate conceptual models to assist the designer. The lack
of a consistent, rationa.l method for deta.iling may lead to problems when unique
situations are encountered. Strut- and- tie models serve to provide a rational
framework for a detailing method which may be applied to a variety of structural
components and loading conditions. Restriction of the use of such models to
 25

zones or regions where traditional approaches incorporating a concrete

contribution are inappropriate or uncertain will eliminate the economic penalty
occuring in general usage of stnJt-and-tie-models with lightly loaded members.
In the strut-and- tie model, the actual stress distribution within a structure is
idealized as a static force system consisting of the following basic elements:

struts: representing concrete compression stress fields

ties: representing concrete tension stress 'fields or
bonded reinforcement or
bonded prestressing steel
(unbonded tendons are individual members of the structure
and should be treated correspondingly in the structural analysis)
nodes: representing the intersection area of struts and ties in which the
internal forces are redirected

The quantification ofthe strut- and- tie model as a comprehensive design

tool for structural concrete details is a relatively new undertaking. The focus ofthis
study is on design of details for which no rational design method currently exist.
Therefore traditionally well defined applications such as uniform depth beams,
columns, and slabs with uniform loading patterns are only briefly addressed. This
study has as a basic objective the development of a consistent methodology and
an accompanying comprehensive detailing guide for structural concrete based
on use of refined strut- and- tie- models. It is hoped that the methodology and the
illustrative guide will help designers develop a clearer understanding of the
functioning o'f reinforcement and anchorage details in a wide variety of details in
concrete structures. It is envisioned that the designer will approach the detailing
of a concrete structure using strut-and- tie- models which may be based on an
equilibrium analysis of load paths, on detailed results from a linear finite element
analysis, or by analogy with a steel design procedure. After isolating the
geometrical and/or statical discontinuities regions, the designer has to develop
26

a suitable strut- and- tie- model to carry the applied loads and meet the given
boundary conditions. After selecting and analyzing a strut- and- tie- model, a
major concern are the nodal zones where inclined compression struts, vertical
stirrups, and longitudinal reinforcement intersect. The actual patterns of the
nodes and the limiting stresses in the nodes must be quantified before practical
implementation. Similarly, the allowable or effective compressive stresses in in-
clined compressive struts must be carefully evaluated.
The specific objectives of this overall study are:

(1) To determine the state of the art in structural concrete detailing as

reflected by research conducted and reported in Europe and North
America

(2) To specifically investigate experimentally the applicability of current

AASHTO provisions for shear in the negative moment zones of
pretensioned and post-tensioned composite beams.

(3) To test selected structural concrete details experimentally, such as beams

with dapped ends and node regions

(4) To use the experimental results to refine the strut- and-tie model
especially in terms of nodal zones, material characteristics, and member
continuity

(5) To develop a detailing guide which provides simple models fordesigning

complex details in structural concrete

Objectives (2) and (3) have been reported in detailed in Reports 1127-1 and
1127-2. This report summarizes the efforts to meet objectives (1) and (4) and
presents the detailing guide of objective (5). The gUide should lead to more
consistent, constructible, economical and reliable details.
 27

1.4 Scope

The existing state of knowledge is not sufficient for complete application

of the strut-and-tie-model to complex detailing situations. Therefore, the scope
of this study was to develop a relatively simple approach to show how a designer
can develop a strut- and-tie- model for different applications, to provide some
ideas about model optimization in order to be economical and to provide some
in-depth understanding of the behavior of singular nodes as affected by their
reinforcing and anchorage details. The experimental tests in the accompanying
studies (1127-1,1127-2) included different concrete strengths, degrees of con-
'finement, strut width, reinforcement anchorage details, and strut angles. Also full-
sized dapped beams were studied and compared with the possible use of the
strut- and tie- model.

Chapters 2 and 3 explain the basis for the strut- and- tie- model used for
detailing structural concrete. They present an overview of the proposals devel-
oped by various researchers. In particular this study was aimed at providing
information about:

length of the discontinuity zone (D-region)

model optimization
dimensioning the nodes
configu ration of the stress 'field
allowable concrete stresses or efficiency factor of the nodes
detailing considerations

Chapter 2 and 3 discuss general principles, components, and modeling

techniques as well as dimensioning of the struts, ties and nodes to illustrate the
method and facilitate its use. Recent studies about high strength concrete were
also included and design provisions evaluated. Suggested Design Specification
28

language to implement these procedures are provided. Illustrative design aids

and design examples are presented in Chapters 4 and in Appendix A.

Chapter 5 summarizes the work and presents conclusions and recom-

mendations. The guide should provide support to the structural designer in an
area where current codes and design specifications provide little information.
 CHAPTER 2
BACKGROUND

2.1 Concept Background

The strut- and- tie model is a limit analysis approach to the design of structural
concrete. More specifically, the strut- and- tie model is a static or lower bound
plasticity solution. Marti [27] explains that strut- and tie models represent a possible
equilibrium system offorces within a structure at its ultimate load. While the plasticity
theory behind the strut- and- tie model is quite complex [56], it is primarily used to
establish a rational basis for the method. For most practical applications, it is only
necessary to understand that a properly chosen and dimensioned strut - and- tie
model represents a lower bound (or conservative) estimate of the true capacity of a
structural element assuming other brittle failures such as stability or local crushing
are precluded.

Although development of detailed mathematical verification forthe strut- and-

tie- method is unecessary to understand its application, awareness of the
assumptions is important. The most important of these assumptions are:

1. Failure is due to the formation of a mechanism resulting from yielding

of one or more ties
2. Crushing of the concrete struts should not occur prior to yielding of the
ties. This is prevented by limiting the stress levels in the concrete.
3. Only uniaxial forces are present in the struts and ties
4. All external loads including post- and pretensioning forces are applied
at the nodes of the strut- and- tie model. In the case of distributed loads
and pretensioned strand loads, the model must be adequately
formulated to realistically represent the load distribution
5. The reinforcement is properly detailed so as to prevent local bond or
anchorage failure

29
30

While strut-and-tie-models can be used in detailing any element of a concrete

structure, it generally is more efficient to use traditional methods based on cross-
section analysis for proportioning axial, flexural and web reinforcement in
constant depth beam, column or slab type structures subject to distributed
loading. Irregular shaped zones or zones subject to heavy concentrated loads
are logically targets for the application of strut-and-tie-models. For a succesfull
implementation of the strut- and- tie model a classification of concrete structures
with respect to their geometry and their load bearing behavior is required. Any
concrete structure may be classified by subdividing it (by application of Saint
Venant's principle) into D- (Discontinuity) and B- (Bending) regions [28]. (See
Fig. 1.4.)

Those specific areas for which the Jakob BernOUlli-hypothesis that a

plane section before bending remains plane after bending (linear strain pro'l'i Ie or
plane strain) applies with sufficient accuracy are identified as B-regions. In the B-
regions the elastic principal stresses may be determined directly from the axial,
flexural and shear stresses acting on the member.

Any general region in which the strain distribution in the crossection is

substantially nonlinear due to statical and/ or geometrical discontinuities is
defined as aD-region.

The following classifications can be made with regard to geometry and

loading and used in taking into account the division of B- and D-regions.

(1) Linear structures are structures with two dimensions considerably

smaller than the third (beams, frames). If they are essentially uniformly
loaded, they will consist in substantial part of B-regions.

(2) Plates or deep beams are plane structures with two dimensions
considerably larger than the third with loads acting in their plane (walls,
thin webs of box girders).
 31

3. Slabs and shells are structures with two dimensions considerably larger
than the third with loads acting transverse to their plane or curved middle
plane. If they are predominantly uniformly loaded, they will essentially
consist of B-regions. Strips taken along the principal moment directions
behave and can be treated therefore as linear structures.

If a structure contains B-regions in substantial part, it is usually more

convenient to first determine its sectional load effects (MB , MT' V, N) by use of
conventional elastic analysis.

For uncracked B-regions the internal stresses then can be determined

from the sectional load effects by use of cross-sectional values A, IB' IT' and the
usual laws of mechanics (bending theory). In the case of high compressive
stresses the linear analysis of internal stresses may have to be modified by
replacing Hooke's law with one of the nonlinear material laws. For cracked B-
regions the internal forces are generally determined from the cross-sectional
load effects by application of the standard truss or ordinary cracked reinforced
concrete theory. In B-regions the use of truss models or of strut- and- tie- models
will often be more complex than required.

The forcepaths or the struts and ties of the D-regions can be determined
from the loads applied to the D-regions by equilibrium analysis. If a structure or
member consists of only one D-region, the analysis of sectional effects by a
conventional structural analysis may be omitted and the internal forces or
stresses may be directly determined from the applied loads. If the structure is
externally statically indeterminate, the internal compatibility of stresses should
be considered by first orienting the geometry of the model to the pattern of forces
indicated by the results from a conventional elastic analysis and then possibly
reorienting it thereafter according to the major design intent: emphasis on
ultimate load capacity or on serviceability under working loads.
32

Table 2.1 summarizes the above statements.

Table 2.1: Analysis usage

B-regions D-regions

analysis sectional ana.lysis direct

material behavior linear or nonlinear nonlinear

state I sectional values elastic stress analysis

= uncracked
state " conventional reinforced strut- and- tie- model
= cracked concrete analysis or nonlinear analysis
or truss model

It is usually most convenient to orient the geometry of the strut- and- tie
models to the general pattern of load paths traced by the forces passing through
the member. These load paths can be determined from intuition, experience, or
in unusual cases by examining the elastic stress fields indicated by a finite
element analysis. Design of B-regions is accomplished by ordinary cracked
reinforced concrete theory orby using a specia.l type of strut- and-tie model which
is generally termed the truss analogy. In the truss model for a simply supported
beam the upper horizontal chord represents the concrete compresssion zone.
The lower horizontal chord represents the main tension reinforcement. The
stirrups ofthe beam are lumped together as the truss vertical members. Inclined
compression struts are used to represent the continuous inclined compression
fields in the web of the beam. The strut- and- tie model is proposed as a
generalization of the truss analogy applicable to a variety of design situations.
 33

The truss analogy is a specialized form of the strut- and- tie model and can be
used exclusively in the design of B-regions. Other types of models which apply
to the wide range of D-regions occuring in the structure are then lumped under
the more general category of strut- and- tie- models.

2.2 Isolate Discontinuity or Detail Region: D-region

For the majority of structures it would be unreasonable and inefficient to

model the entire structure with a strut and tie model. Rather, it is a more
convenient and common practice to 'first carry out a general structural analysis.
The general elastic analysis of linear structures results in determination of
external support reactions. Then, from equilibrium methods, sectional effects
(bending moment Ms ' normal forces N, shear forces V, and torsional moments
Mr ) can be determined at any desired section.

It is advantageous to subdivide the given structure into B-regions and D-

regions. In order to roughly find the division lines between B- and D-regions, the
following procedure was proposed by Schlaich et al. [2]. It utilizes the well known
principle of Saint Venant which localizes the effect of concentrated forces as
shown in Fig. 2.1. This procedure is illustrated in the examples of Fig. 2.2.

1. Replace the real structure (a) by the fictitious structure (b) which is loaded
in such a way that it complies with the Jakob Bernoulli hypothesis and
satisfies equilibrium with the sectional forces. Thus, (b) consists entirely
of one or several B-regions. It usually violates the actual boundary
conditions.

2. Select a self-equilibrating state of stress (c) which, if superimposed on (b),

satisfies the real boundary conditions of (a).
34

d= distance between the self-equilibrating forces

Figure 2.1: The principle of Saint Venant (from Ref. [28])

 Column Beam 35

,h;:j
F

(""I" 11111 " " 'j-~

t-J..

(a) ( a)

1I'4Iil"'j'IIIIII'

CIID (b) (b)

F
~F/h
i
d·h
.... td.~

0+- +.<~
~]
r
d·h
.j....fI:LIt
~h-
(c) (c)

,/'.
o
(d) (d)

Figure 2.2 Subdivision of structures into Band D regions

using Saint Venant's principle (from Ref. [28])
36

3. Apply the principle of Saint-Venant to (c) so that the local stresses may
be assumed negligible at a distance "d" from the equilibrating
forces, which is approximately equal to 1.0 to 1.5 of the maximum
distance between the equilibrating forces themselves. This distance defines
the range of the D-region (d).

For practical applications the following approaches illustrated in Fig. 2.3 are
suggested:

(1) Isolate the geometrical or the statical discontinuity regions (zone 1)

For statical discontinuities this is basically the point of application of
concentrated loads
For geometrical discontinuities these are the end points of the irregular
zones: end of an opening, corner line, etc.

(2) Combine overlapping areas of the statical and geometrical discontinuities

(zone 1 in Fig. 2.3)
 37

(3) Add additional lengths (zone 2) equal to 1.5 h or h' on both sides of
the previously isolated discontinuity zone

(4) Assume that the total area of zone 2 + zone 1 + zone 2 is the effective
D-region

The discontinuity region boundaries may also be determined by the use

of stress trajectories (contours of principal stress).For theoretical purposes, the
D-regions end where the stresses introduced by the discontinuities are negligible
or where the strain distribution is linear. Between the D-regions the stress
distribution is essentially uniform and regular. The linear strain profile assumption
of Jakob Bernoulli is applicable (B-region).

B-region
statical discontinuities

1.5h

D-region h o h zone 1

eometrical discontinuites

h'
~~
B-region

h
h = height of the cross section

h' = 1.5 h cos B

= h zone 1 + 1.5 h cosB + 1.5 h cosB (in general)

Figure 2.3: Suggested subdivision of structure

38

2.3 Elasticity Analysis Method vs. Load Path Method

For the D-regions it is necessary to develop an individual strut- and- tie-

model for each application. For very unusual configurations as well as to ensure
good crack control behavior at service stress levels, it is recommended that the
model should be generally based on the principal stress pattern as determined
from an elastic analysis. For unusual cases such an elastic analysis with the
principal stress direction can be computed with an elastic finite element program.
The directions of the struts and ties can be located at the center of gravity of the
corresponding stress fields. It usually makes it more convenient if the principal
stresses are converted to stresses parallel to the structural member borders (O'x'
A strut- and- tie- model based on such orthogonal elastic stresses chosen
O'y' 'txy')'
parallel to the concrete surface generally leads to more economical straight
reinforcement layouts. As a general approach, the reinforcement should be
arranged parallel and/or normal to the concrete surface and take into account
the requirements of controlling possible inclined crack propagations by inclined
bars when needed.

If such elastic analyses are inconvenient or when the general form of the
solution is known from experience, the strut- and- tie- model can also be
formulated by tracing the so called "load paths". The outer equilibrium of the D-
region must be satisfied by previously determining all the loads and reactions
acting on it. For a boundary adjacent to a B-region, the loads on the D-region are
taken from the B-region analysis. The following approach can be used in order
to determine the load paths (see Fig. 2.4).

(1) Compute the elastic stress resultants or ordinary cracked reinforced

concrete forces for the B-region and apply the equivalent forces to the
D-region
 39

(2) Subdivide the discontinuity zone into regions in such a way that the loads
on one side of the discontinuity zone are in unique regions with their
counterpart on the other side of the discontinuity zone. These regions are
the load paths connecti ng the opposite sides and te nd to take the shortest
possible streamlined way across.
The load paths must be single lines and must not cross each other.

(3) Ifthe applied forces are not completely equilibrated with the obvious load
paths, then the resulting loads must follow a U-path as shown for B-B in
Fig. 2.4.

(4) Sketch all load paths (including possible U-paths) and replace them by
polygons made up of compression struts and tension ties.

(5) Add further struts and ties as required for equilibrium at the nodes.

(6) If necessary rearrange the struts and ties with consideration of practicality
of the reinforcement layout.

For very complicated cases, the finite element analysis results can also be
combined with the load path method.
40

8 C B C
a)

I
I
I
II
)-~
T2 / // \\
T / \
I

A lie A c
sh~ceJ I

~-l

b)

.I I i II II
II I
1i ) \ /
L--~ \
"I /
~I.~ IViI /" \ 11+-+1/ \
A tc
--- strut
tie

- load path
anchorage length of the bar

Two models for the same case: (a) requiring oblique reinforcement;
(b) for orthogonal reinforcement.

Figure 2.4: Load path method with "u-turn" (from Ref. [28])
 41

2.4 Strut Background

The struts transfer internal compression forces from node to node. In

general this transfer is via three-dimensional stress fields in the concrete. Due to
compatibility requirements the stress fields tend to spread out between the
nodes and fill the available space. This results in transverse tensile and
compressive stresses which must be considered in tile evaluation of a strut's
strength and which may require provision of local reinforcement (see Fig. 2.5).
However, well tied compression reinforcement will also resist compressive
forces as long as it is restrained from buckli ng. Whi Ie all concrete structu res must
be built and reinforced in three dimensional space, it is usually sufficient to
determine reinforcement separately in two orthogonal planes. This leads to
consideration of two dimensional or planar struts.

Four typical configurations of two dimensional compression fields are

presented in Fig. 2.6. Generally it is safe to determine the strength of
compression struts using one of the four simplified types of stress fields shown
in Fig. 2.6 (a) to (d).

A prismatic strut as shown in Fig. 2.6 (a) is the simplest idealization of a

compressive stress field. The prism is uniform in geometry and has a constant
stress along its length. Prisms are generally used to model stress fields having
uniform parallel stress trajectories.

The fan shaped stress fields shown in Fig. 2.6 (b) are developed at points
of concentrated loading or at supports. Fig. 2.7(a) illustrates a fan region
incorporating a series of trapezoidal struts Wllicl1 act to distribute force to several
stirrups.
42

......
.... - -
-
I""'---
~

compression

(a) compression only

1lj..-r-t--1--;--\-~+--- tension

nll_.JJ,..-f--t----1IT--U._ I I ' r - l - - - compression

(b) real state of stress: compression and tension

Figure 2.5: Planar stresses in prism under compression

 43

'llI B-~

T L

(a) prism (b) fan

(c) bottle (d) confined core

Figure 2.6: The basic compression "fields ("from [2])

44

.
,
I

,,
19

S-Struts
T-Tension Ties

Bottom Chord

Figure 2.7(a): Fan region at beam support

Smeared
N des

,
••
.
.•. ,.•

Singular
Nodes

(b) Strut-and-Tle Mode (c) Corresponding

(a) Stress Fields
Reinforcement

Figure 2.7(b): Bottle shaped struts

 45

In some cases, as a corollary to the trend of stress fields widening between loads,
a stress field may tend to narrow near points of application of concentrated loads
or at supports. This can be modelled by using a bottle shaped strut (Fig. 2.6(c)).
The increase in strut width induces tensile stresses normal to the longitudinal
axis of the strut. This tensile stress must either be resisted by transverse
reinforcement or by the tensile strength of the concrete in order to prevent
cracking. Fig. 2.7(b) shows a bottle shaped strut represented by a secondary
strut- and- tie model for the analysis and provided with local transverse
reinforcement to properly develop the required tie forces. The confined core of
Fig. 2.6 (d) is a transversally reinforced core or prism with a special behavior. The
reinforcement can be spirals, closed stirrups or steel pipe. The reinforced core
develops under load a three dimensional state of stress, which is controlled by
the behavior and the form of the reinforcement and the transversal contraction
of the concrete. It is generally restricted to points of application of very large
magnitude forces or relatively small areas as when post-tensioned tendons are
anchored or when extremely large loads are applied by columns bearing on a
transfer girder.

2.4.1 Concrete compressive strength limitation for struts

In recent years the useful range of concrete compressive strengths in

highway applications has exceeded 12000 psi. Most empirical equations for
concrete structures now in standards such as the AASHTO Specifications [3]
were derived using results 'from tests having concrete compressive strengths
less than 6000 psi. Extrapolating such empirical equations for concretes with
twice the compressive strength of those used in the original formulation can be
dangerous or uneconomical. In this study the proposed efficiency factors for
concrete compressive strength should be applicable to values up to 12000 psi.
Many different test results of high strength concrete specimens were included in
the study. For background a brief literature review on pertinent information
concerning high strength concrete is included.
46

High strength concrete requires extreme care in all steps of the production
process. It has become common practice to specify high strength concrete
strengths at 7,28,56, or 90 days [57, 58]. Economically it is important to know
at the outset of high strength concrete production specifically what strength one
needs and when one needs it. High strength concrete requires a very low water
to cement ratio (=0.25 - 0.45). Therefore, inaccurate estimation ofthe aggregates
water content, which affects the quantity of additional water added at batching,
can result in either balling of the concrete due to lack of mixing water or in too
high a slump. Mixing is critically important as well. For satisfactory performance
all the materials, especially admixtures, must be thoroughly mixed. Curing
becomes more critical in high strength concrete production and proper hydration
must be allowed to prevent shrinkage cracking.

The tensile strength is significantly linked to the curing conditions. The

following relations (concrete compressive strength between 731 0 and 10040 psi)
were given in Ref. [33] with a 10% coe'fficient of variation:

= moist cured

= dry cured

The compressive strength of the aggregate has a major effect on the

strength of high strength concrete. The aggregate must be sufficiently strong to
allow higher concrete strength. Tests reported in [33] indicated that the failure
breaks were going through the aggregate without bond failure. T~lis proved that
the coarse aggregate has a major effect on the strength of high concrete strength.
The physical properties of high strength concrete tend to be somewhat different
than for normal strength concrete. The slope of the stress-strain curve is steeper
and more linear up to about 80% of ultimate capacity (see Fig. 2.8). The
descending branch of the high strength concrete stress-strain curve is steeper.
 47

12 , . . - - - - - - - - - - - - - - - - - ,
LIMESTONE GRAVEL

10 High -Strength High St renglh

8
-
V)
1\
I\
11 Medium h
-
~
Strength
V)
&II

-
ell
~

( /)
6 I\,, II Medium
I~ Strengfh
1
I
""~ " ,,
, I
4
\ Normal- Strength
\.
,'\ Normal
\. \ Strength
2
Shoke Rore: 2000p.-in.(0.05mm)/min. Stroke Rote: 2000p-in.(0.05mm)/min.
4" It a" (102 mm a 203 mml Cylinder 4': as" (102 mm a 203 mm) Cylinder
0 ..1--_""--_-'--_........_ ........ ......

o 2 3 4 5 6 7 0 2 3 4 5 6 7
Strain (. 1000 in./in.) Strain (Jl IOOOin.lin.l

Figure 2.8: Stress-strain curves for low, medium and high strength concrete
(from Ref. [59])
48

It has been stated that the descending branch becomes almost a vertical line.
The ultimate strain at failure is lower than for moderate strength concretes. The
steeper stress-strain curve for high strength concrete means the modulus of
elasticity is higher. The following equation forthe modulus of elasticity has been
proposed [59]:

Ec = (40000) ( f J O•s + 1.0 x 106) (w/145)1.S

w = unit weight of concrete (:::::145 pcf)

The total shrinkage at later ages is about the same as for medium strength
concrete. High strength concrete does, however, see more of its total shrinkage
at early ages than does normal concrete strength. Unit creep tends to be much
lower in high strength concrete. Given the fact that it is stressed to higher stress
levels, total creep stays about the same.

Considerable research has been conducted in effort to determine the

limiting compressive stress for concrete in compression struts in structura.l
members. Because of cracking, aspect ratios other than the value of 2 found in
standard cylinder tests, and the presence of strain gradients, the value of
compressive strength found in a cylinder test may not be appropriate for the
compressive strut in a structural member. Much of this work has focused on thin
webs of beams which is a critical case. Empirical rela.tions for the compressive
strength of concrete struts in beam webs as suggested by, Nielsen et al. [21],
Ramirez and Breen [23], ThQrlimann [60], Collins and Mitchell [61] and MacGre-
gor [62] are summarized in Fig. 2.9.

Several factors influence the value of the effective concrete strength in a

diagonal concrete strut. Due to the fact that the stirrups cross the diagonal
tension cracked concrete, the diagonal concrete strut is then in a biaxial state of
stress (Compression-Tension) which reduces its compressive strength. Another
 0
o·
0
v 49
..-
c
Q) CD
ca m... ...
Q) .t=
B
- CD...
Q)
C
Q)
~
N
c
c
C'l:S
.§ en
~
CI/:S
0
C)
...
Q)
0
0
(!) 0
Cf.I
'EC'l:S -.:: ,S C\J
CD
z a:
:::::1
r= '0
t>
"
C'l:S
~
..-

t t t tt 0
0
0
0
..- -
.-
f/)

.....
Q.

0
.r:.
....
0 C)
0 C
co (1)
....'-en
0
....
(1)
(1)
0 I-
0 U
co C
0
CJ

0
0
0
v

o
o
o
C'\I

q C') co co Ii) C\I .... 0

.... o ci ci ci ci ci ci

concrete efficiency factor

Figure 2.9: Relationship of diagonal concrete strut efficiency factor versus

concrete strength
50

factor is that the redistribution of forces in the member due to the different ratios
of longitudinal to transverse reinforcement may cause the failure crack and the
compressive struts between them to be at an inclination otherthan the 45 degree
angle corresponding to initial diagonal tension cracking of the concrete. Thus the
strut may be crossing previously cracked concrete. Another important factor is
the need to select a very conservative value due to the undesirability of a failure
due to crus~ling ofthe concrete in the web because of its brittle nature. In general,
the effective concrete strength available for use in the struts is chosen as some
portion of the concrete compressive strength flc. The effective strength flc is the
product of an efficiency factor v e and the 28 day cylinder compressive strength.
The efficiency factor should take into account the following parameters:

multiaxial state of stress

disturbances from cracks
disturbances 'from reinforcement
confining reinforcement
friction forces
aggregate interlock after cracking
dowel forces
time dependence

f'ce = Ve f'c

Various proposals for the efficiency "ve " factor have been presented. They are
usually based on tests of continuous compression fields either in rather thin
web beams or rather thin shear panels although some seem to be based
largely on engineering jUdgement. Very little experimental verification exists
for effective compressive stress efficiency factors for use in model analysis or
for use in large panels where shear is not a major concern. Many of the
various proposals for the efficiency factor are summarized in this section. They
basically correspond to the product of a basic efficiency factor and a modifier
to make them applicable to thin webs although this distinction is not always
shown by the various authors. Another factor considered by some authors
was the fact that in the case of torsion the twisting of the beam induces an
additional compression stress into the diagonal. Lampert and Thurlimann [14]
stated that the increase in the diagonal compression test was due to a
distortional effect in the walls of the cross
 51

,. a

Figure 2.10: Distortional effect (from Ref.[23])

x
/
Figure 2.11: Forces acting on edge members of parabolic arches (from Ref. [23])
52

section. Through twisting, the originally plane walls of the section are distorted
to hyperbolic paraboloids (Fig. 2.10) limited by four straight edges. The distorted
wall then constitutes a hyperbolic parapoloid shell subjected to a uniform shear
flow. The entire shell when loaded in this fashion is subjected solely to pure shear
stresses of constant intensity (see Fig. 2.11). These edge shears require edge
members. In the case of truss models these edge members are provided by the
longitudinal chords which are thereby loaded axially. The additional compressive
stresses on the outer surface of the diagonal due to wall distortion must be added
to those obtained from the actual shear flow. As a result ThOrlimann [ 60]
suggests that the maximum value of the compression strength in diagonal
compression struts used be approximately 2400 psi, corresponding to f'eof about
4800 psi. ThOrlimann [60] on the basis of test evidence proposed that the
allowable efficiency factor for the compression stress be:

Ve =0.36 + 696 / f'e [psi] fIe ~ 4800 psi

This upper limit should prevent a premature failure.

Additionally a limit is placed on the inclination of the concrete compression

strut "<Pes", and thereby on the amount of redistribution of internal forces. The flow
rule orfailure mechanism is uniaxial yielding ofthe reinforcement opening upthe
final cracks perpendicular to the crack direction (see Fig. 2.12). Finally the
reinforcement is assumed to be properly detailed so that no local failures are
possible. As shown in [56], within the limits for the angle of inclination of the
diagonal compression strut (26.5° ~ <Pes ~ 6"3:5° see Fig. 2.13), the average
diagonal compression stress can be controlled by limiting the nominal shear
stress independently of the inclination <Pes of the compression diagonals. From
Fig. 2.13, it can be seen that if the compression struts are inclined at 45 degrees,
the mean crack strain and hence the mean crack width are at the minimum value
foryielding of both the longitudinal and transverse reinforcement. The values are
 53

"I" longitudinal

w =

= ~\ (1 + tan 2<l>cs) (yielding of transversal reinforcement)

= £1 (1 + cot 2<l>cs) (yielding of longitudinal reinforcement)

Figure 2.12: Mean crack strain ('from Ref. [56])

,
Onset of YieldinQ
in
t
10 lonQitudinal Stirrup
Reinforcement

8
t R ~ Displccement Parameter
6 £, & Yield Strcin of Reinforcing Steel
5--+-----+~~~~t----r
4

0.5:!: tan ex ~ 2.0

Figure 2.13: Relationship between the mean crack strain and the strains in the
reinforcement for different angles of inclination of the diagonal strut
(from Ref. [61])
54

not exact limits but give a general range for transitions of failure
mechanism.ThOrlimann noted that at !lew / ~ !l values of about 5 the failure
mechanism begin to change, and either shear or flexural failures become
possible without both of the reinforcement types yielding. It is also shown that if
the angle of inclination is greater than 45 degrees, yielding of the stirrups
demands larger mean crack strains. Conversely, forangleslessthan 45 degrees,
yielding of the longitudinal reinforcement requires increasingly larger crack
openings.

A best fit curve of the form k (fIe )0.5 to approximate the equation proposed
by ThOrlimann resulted in the relation [23]):

34
ve = [psi]

Ramirez and Breen [23] suggested that the compressive stress in the
compression diagonals should be less then 30 (f'c )0.5.

A study by Hartmann, Breen and Kreger [33] investigating the shear

capacity of high strength prestressed concrete girders (f'c===12000 psi) showed
that by using this concrete limit the experimental results were 1.72 times
(standard deviation =0.26) higher then the expected results with the 45 degree
truss model (see Table 2.1).

Zimmerli [56] presented a relationship for the efficiency factor as a

function of varying strut angles (see Fig. 2.14) based on test results of beam
webs.
Ve = 0.83 sin 2 <l>cs
<l>cs = angle between the tie and the strut
f'c = 0.7 fcc•5% (5% fractile cube compressive strength)
 55

1.0
~ efficiency factor
0.9

--
a..
0
u
ca
0.8
>-
u
c
.! 0.7
-
.2
( I)
0.6

0.5
10 .20 30 40 50 60 70 80
angle [0]

Figure 2.14: Varying strut angle versus efficiency factor for concrete
compression strength

It was assumed that for an inclination of 45 degrees the effective concrete

strength forthe principal compression strut would be 5/6 of the normal concrete
strength (flc ). The decrease of the efficiency factor for other angles is based on
the changed angle between the compression strut and the principal stress
direction.

Collins and Mitchell [61] suggested that the limiting value of the average
principal compressive stress in the diagonal concrete strut is governed not as
much by the compression strength of the uncracked portions of the strut as by the
capacity of the interface shear transfer mechanisms, such as a.ggregate inter-
lock, to transmit the required shear stress across previously existing cracks.
56

The principal factors affecting the aggregate interlock are:

quality of the concrete. Usually the top part of a member, because of the
particle sedimentation and water gain under the coarse aggregate will
contain weaker concrete
crack width
aggregate strength

Tests by Paulay and Loeber [63], in which the crack width increased
proportionally with the applied load, verified that the stiffness of the aggregate
interlock mechanism gradually decreased as the shear across the interface
increased. Since the aggregate interlock disintegrates with large crack widths,
and the mechanism of sheartransfe r in the diagonally cracked concrete is largely
dependent on the aggregate interlock, it is apparent that the maximum
compressive stress that the diagonal strut can take will be a function of the angle
<Pcs•

Nielsen et al. [21] at the Technical University of Denmark applied a rigid

plastic model for the concrete based on the modified Coulomb failure criteria.
With the assumption of only plane stresses the model gives a square yield locus
with a compressive yield stress of fIe and zero tensile capacity. Based on
experimental results an effectiveness parameter for concrete strength in the
webs was suggested. Nielsen et al. [21] described the experimental facts by
means of an empirical formula of the type

= 725 ::;; cre::;; 8700 psi

 57

f2 (h) = 10.6 (1 + 0.16 / (h)o.s 3 in. ::;; h ::;; 27.5 in.

fs (Jl) = 0.15 Jl + 0.58 Jl ::;; 4.5%

f 4 (a) = 1.0 + 0.17 (afh - 2.6)2 afh ::;; 5.5

fs (eJe t/eJO.2) = 1.1 (1.0 + 0.81 (eJet/eJo.2) for prestressing

eJc = compressive stress in concrete

eJO.2 = steel stress at 2000 micro-strain in the prestressing bar
eJeff = effective prestress stress
a = length of shear span
h = depth of beam
1.1 = As/(bh)
As = longitudinal reinforcement
b = web width of beam

The analysis included 186 test results from normal reinforced rectangular beams
and 19 rectangular prestressed beams without shear reinforcement. The
statistical parameters of the ratio between the experimental and the calculated
ve - va.lues are presented inTable 2.2.

Table 2.2: Statistical parameter for shear tests from [21]

mean value coefficient of variation

norma.l reinforced 1.0 14.5%

prestressed 1.0 8.6%
58

Fordesign Nielsen et al. [21] recommended the use of the conservative straight
line expression

Ve = 0.7 - f'cl (28997) [psi] flc:::;; 8700 psi

Limits are also placed on the angle of assumed strut inclination to prevent too
large a deviation from elastic behavior.

21.8° < <Pes < 45°: beams with constant longitudinal reinforcement
26.5° < <Pes < 45°: beams with curtailed reinforcement

_i

Mitchell and Collins [T9] and the Canadian CSA-Standard A 23.3-

M84 [29] presented a more detailed method for determining the limiting stress
in compression struts based on results oftests on shear panels. (See Fig. 2.15.)

normal strain

Figure 2.15: Mohr's circle

 59

The principal strain may be determined based on Mohr's circle of strain using the
strut angle, principal compressive strain and the strain parallel to the beam axis.
The efficiency factor is related to the principal tensile strength along with the
cylinder compressive strength.

1
v=
e

= principal tensile strain normal to the principal compressive stress

Mitchell and Collins [61] assumed'that the principal compressive strain

(e2 ) in the strut is generally - 0.002 and the principal strain direction is assumed
to coincide with the principal stress direction. The principal tensile strain in the tie
e1 can be computed by compatibility as follows:

= angle between the tie and the strut

The strain ex for the reinforcing bar can be conservatively taken as fy / Eg •

Table 2.3 summarizes the recommendations for the efficiency factor from CSA
(see also [61 ])[29] and [62]

f'c = specified compressive strength

fee = the factored concrete strengths used in checking ultimate limit
state 1
0.6 === partial safety factor === 1.667
(somewhat similar to the ACI and AASHTO ¢ factor)
60

Table 2.3: Efficiency factor proposed in the CSA [29] and by Macgregor [62]

Structural member ve [29] Ve [62]

Truss nodes:

Joints bounded by compressive struts and bearing areas 0.85 0.85

Joints anchoring one tension tie 0.75 0.65

Joints anchoring tension ties in more than one direction 0.6 0.50
If the tie reinforcement is anchored by bearing against
metal plates at the back of the nodal zone, then 0.85 /

Isolated compression strut in deep beams or D-region / 0.50

Severely cracked webs or slender beams with £5=0.002

<Pes= angle between tensiond tie and strut

<Pes = 30 0
0.31 0.25

<Pes = 45 0 0.55 0.45

For buildings of normal importance the load factors for dead load and live load
are 0 = 1.25 (exceptthat if dead load resists overturning, uplift or stress reversal,
then 0= 0.85) and L= 1.5 respectively.
 61

Relating to the Canadian CSA - Standard [29] the CEB-MC - Draft 1990 [64]
gives the following formula for plane stress fields with closely spaced cracks and
no major geometrical disturbance:

1
ve = ~ 1.0
(0.85 + 0.27 ct /~)

f'e = cylinder compressive strength

C, = longitudinal compressive strain (absolute value)
ct = average transverse tensile strain perpendicular to the
compression direction

In the CEB Code-Draft [64] , f2J =1.0 in all cases and the load factors for dead and
live load are 1.35 and 1.5, respectively.
Major skew cracks are not likely, if the theory of elasticity is followed sufficiently
closely during modelling. This means that the angle between struts and ties
entering a singular node should not be too small. However, skew cracks may also
be left over from a previous loading case with different stress situations (creep,
shrinkage, temperature etc.)

The CEB proposed efficiency factors are related to specified safety factors
which are different from those in North America. The following equation will be
used in Europe to compute the effective concrete strength:

fee =Ve fIe C / 1.5

fee = V e fIe 0.667 for short duration load

fee = V e fIe 0.567 for slJstained load

f' e = specified compressive strength

C = coefficient for sustained load =0.85
coe"tficient for short duration load = 1.0
1.5 = partial safety factor for the concrete in compression
62

Table 2.4 gives the efficiency factors proposed by Schlaich et al [28] and
CEB - MC 1990 [64]:

Table 2.4: Efficiency factors proposed by Schlaich et al. [28] and CEB [64]

State of stress and/or reinforcement layout for strut Ve

For undisturbed uniaxially state of compressive stress 1.0

If tensile strains in the cross direction or tensile reinforcement

in the cross direction may cause cracking parallel to the strut
with normal crack widths 0.8

For skew cracking or skew reinforcement 0.6

For skew cracks with extraordinary crack width

Such cracks must be expected, if modelling of the struts
departs extremely from the theory of elasticity's flow of
internal forces. 0.4

In the following subsections, some test results are analyzed with the strut- and-
tie-model in order to evaluate the efficiency factor for compression struts in
cracked webs.
 63

2.4.1.1 Effective concrete strength in compression diagonals

Shear capacity can be analyzed with different models. A number of empirical and
conceptual models have been presented over the course of time. Given all of this
effort, however, a completely satisfactory solution has not been attained. The design
concepts of the ACI [3] and AASHTO [3] - recommendations and those in the CEB-
MC - Draft 1990 [64] have in principle the same structure. The general basis is

Vu = factored shear force at a section

Vc = nominal shear strength provided by concrete
Vs = nominal shear strength provided by shear reinforcement
Vp = vertical component of effective prestress force at section
<1> = strength reduction factor =0.85 for shear
For reinforced concrete there are two equations for V c under normal loading
conditions. One equation is [3]

Vc = {1.9 (fie )0.5 + 2500 (A/(d b w) [Vu d / M u } bw d :s; 3.5 (fie )0.5 d bw
As = area of longitudinal reinforcement
Mu = factored moment at section
bw = web width
d = distance from extreme compression fiber to centroid of longitudinal
tension reinforcement

The second equation for V c in reinforced concrete is

= (fl )0.5
2 b de
Ve w

For the concrete contribution in prestressed members the equation is

64

In both reinforced and prestressed concrete,

Vs = Av fy d / s

In prestressed concrete with inclined tendons,

Vp = Np sin a.

According to CEB [64] the concrete contribution is given as follow

= 2.5 'tRD bw d B

B =

Values for 'tAD for different concrete strengths are given in Table 2.5.

Table 2.5: Values for 't RD according to CEB MC - Draft 1990 [64]

f'c 1740 2320 2900 3625 4350 5075 5800 6525 7250

'tAD 26.1 31.9 37.7 43.5 49.3 55.2 60.9 66.7 72.5

Kordina and Hegger [32] present another fofmulation for the concrete contribution
in prestressed concrete girders.

Vc = {2.5 (f'c )o.s [Asl / (bw d]o.sfY + 0 • 15 0'cp } (d bw)

= N / (Ac ) ::;; 0.4 f'c

N = longitudinal force or prestressing force

= longitudinal reinforcement
 65

An excellent approach is given by Vecchio and Collins [65] to predict the

response of reinforced concrete beams subjected to shear using the modified
compression field theory. Strain softening and tension stiffening effects are taken
into account in the theoretical model.

The strut- and-tie-model for sheardesign is based on some assumptions.

Yielding of both the longitudinal and transverse reinforcement is required. This
requires an upper limit on the diagonal concrete stresses to prevent crushing.
The reinforcement can only resist axial loads. The reinforcement is properly
detailed so that local crushing and bond failures are prevented. The angle of
inclination forthe compression diagonals differ as proposed by different authors.

25° ::; <Pes ::; 65° (Ramirez, Breen [23])

15° ::; <Pes ::; 60° (Mitchell, Collins [22])

The compressive stress in the compression diagonals, fd can be computed

z = distance between stringers

The orientation of the diagonal compression strut and the width of the strut are
the most important factors for a strut-and-tie-model. Fig. 2.16 shows the strut-
and-tie-model based on the idea of a "design zone" for the ultimate behavior
under shear and bending proposed by Kaufmann and Ramirez [66]. A new
formulation of the compression width is given here and the proposed model is
compared with test results. To obtain this strut-and-tie-model, the beam is first
divided into design zones. A vertical tension tie is placed at the location of the
resultant force of the web reinforcement in each design zone. The tension chord
is located at the centroid of the flexural tension reinforcement and the compres-
sion chord is located at the centroid ofthe flexural compression block. Diagonal
concrete members are then placed to complete static equilibrium of the model.
66

TENSION MEMBER WHICH

REPRESENTS RESULTANT
FORCE OF WEB REINFORCEMENT
WITHIN DESIGN ZONE

a
z =d--
2

TENSION

DESIGN ZONE DESIGN ZONE

4:V=O

Figure 2.16: Proposed strut-and-tie-model for shear behavior (from Ref. [66])
 67

In this behavioral model, the strength and angle of inclination of the

diagonal compression members represent the concrete contribution to the
ultimate shear strength. The selection of the angle value is important because it
directly affects the number of design zones in the member and the relative
amount of transverse and longitudinal reinforcement, as well as concrete
stresses in the diagonal strut.
Table 2.6 contains the geometrical data and the concrete efficiency factor
obtained by using the proposed model for different test results by Kaufmann and
Ramirez [66], Johnson and Ramirez [67]. Symbols are illustrated in Fig. 2.16.

Table 2.6: Comparison with the proposed strut- and-tie-model and test data

Ref. f' c Ca epa ep1 Z b Ts V

[67] 5.28 111.9 15.3 28.7 18 6 20.5 0.435

[67] 7.44 161.2 15.3 28.7 18 6 20.5 0.44

[67] 7.44 142.2 15.3 28.7 18 6 20.5 0.39

[67] 8.1 170.8 15.3 28.7 18 6 40.9 0.43

[66] 8.34 287.4 21.8 34.8 24 6 39.3 0.67*

[67] 10.49 191.5 15.3 28.7 18 6 20.5 0.375

* I-beam, all the other had rectangular cross sections

The following equation for the concrete efficiency factor in diagonal compression
struts is proposed (see also section 2.4.1.2 confined concrete strength). As can be
seen in the next section the
68

same equation without the 0.6 reduction factor is used for the basic efficiency factor
for confined concrete strength and for the compression strength in the nodal zone.
The reduction factor 0.6 (in actuality a judgement factor "3/5") reflects the lower
effective concrete strength for severely cracked webs of slender beams. A
differentiation between the higher effective concrete strength in nodal zones and
isolated concrete struts as compared to more uniformly stressed webs is also made
by MacGregor [62] as reflected in Table 2.3. MacGregor and most of those
proposing efficiency factors do not consider reductions for high strength concretes.

The statistical data from the comparison are given in Table 2.7 and shown in Fig.
2.17. 2.0

---
~
experimentltheory=1.0
~xperiment/theory

/"'" ~
.>-
0
1.5

--
G)
.J::

C
1.0
~
~

....
G)
E
't:
G)
Q. 0.5
)(
G)

0.0
5000 7000 8000 1‫סס‬oo 11000

concrete strength [psi]

Figure 2.17: Comparison of test results with the theoretical approach of predicting

the diagonal compression strength

Mean: Std. Dev.:

[1.136 1.272
Minimum: Maximum:
.967 1.68

Table 2.7: Statistical data of the comparison in Figure 2.17

 69

2.4.1.2 Two- and threedimensional concrete strength

Test results from Kupfer [68] showing the two-dimensional compression strength
are summarized in Fig. 2.18. The maximum efficiency factor "v" was 1.498 and
was obtained from tests with solid bearing plates (f'c = 4480 psi). It should be kept
in mind, that this apparent increase in strength is only due to the quite artificial
restraint of the specimen. The other test data indicate that the strength of
concrete undertwo-dimensional state of stress, 0"1 = 0"2' is only 17.8% largerthan
under uniaxial compression. For the three-dimensional state of stress the test
results from Linse [69] with a compressive strength between 4480 and 3620 psi
are presented for different stress ratios in Table 2.8. The results show that the
efficiency factor depends to a large extent on the triaxial stress ratio and the
difference between the three stresses.

The CEB-MC Draft - 1990 [64] proposes that the multidimensional compressive
strength have the following values:

two dimensional compressive strength = 1 . 1 f' c

efficiency factor v e2 = 1.1

threedimensional compressive strength = 3.0 fIe (confined = 3.3 fIe)

efficiency factor v e3 = 3.0

Table 2.8: Strength ratio for three-dimensional concrete compressive strength

strength ratios

-1.0/-1.0/0. 1.1
-1.0/ -0.93/ -0.18 .. 6
-1.0/ -0.49 / -0.14 3.5
-1.0/ -0.50 / -0.25 ..8
-1.0/ -0.26 / -0.09 3.3
-1.0/ -0.25/ -0.12 6.0
-1.0/-0.16/-0.08 2.6
-1.0/ -0.14/ -0.14 4.4
-1.0/ -0.26/ -0.09 3.3
-1.0/-0.1 / -0.05 1.8
70

2------------------,
s
-..
~
~

en
1
B
r?a
51/82=-1 J-1.
51/82=-1 J-.52
en G:J
-
51/52=-1J-.22
e
en

o
2716 4608 4480 8804
concrete strength [psi]

Figure 2.18: Two-dimensional compressive strength

Table 2.9: Statistical data from Figure 2.18

Column 1
Mean: Std. Dev.:
-I.. .
1;.;.;1..::..;.2;;;;;.6-4----,Ir-.1-1~4~--~ ........---.........
Variance:
O_1_3 .l__
Count:
IF-1;;":;2;";";';;';"---

Minimum: Maximum: Sum:

1.123 1.498 15.164

For the two- or three dimensional state of stress, a large number of theoretical
investigations have been carried out in recent years and various models have
been proposed to characterize the multiaxial stress-strain behavior of concrete
. A brief review of some previous recommendations is given here. The Cauchy
model (nonlinear elastic) by Kotsovos [70], the hypoelastic material law by
Stankowski and Gerstle [71], the elasto-plastic constitutive law by Han and Chen
[72] and the bounding surface model developed by Meschke et al [73] and by
Fardis, Alibe, Tassoulas [74] represent typical constitutive modelsfordescription
of the material behavior for multiaxially loaded concrete structures.
 71

2.4.1.3 Confined concrete strength

In many applications substantial confining reinforcement may be present

so as to greatly increase the efticiencyfactorforconcrete in compression. If such
triaxial concrete strength increase from confining reinforcement is taken into
account, then the unconfined concrete portions such as the concrete cover over
the confining steel have to be disregarded in the evaluation of the strut strength.
For triaxially confined concrete strength, Richart, Brandtzaeg and Brown [75]
based the following formula on 112 test results (see Fig. 2.19):

fe3 = f'c + 4.1 flat

flat = 2 As f/ (d s)

As = confining reinforcement area

fy = confining reinforcement yield strength

d = diameter of entire concrete section

db = diameter of bearing area

A = 1t d2 / 4

Ab = 1t db2 / 4

~// '"'Y".h
.,
'"
1
s
,. i)

CI': rY//"

Figure 2. 19: Typical geometrical data for confined core

72

For square compression struts, with longitudinal reinforcement atthe corners as

well as at intermediate points, and with closed ties arranged so that sufficient
lateral support is provided to longitudinal bars, the lateral pressure is reduced
from that of spiral confinement and can be computed by assuming an equivalent
circular compression strut with effective diameter "de" equal to the side of the
confined square core [76] and a substantially reduced confining effect as

= total cross sectional area of the stirrups and ties (cross tie included)

If the square compression strut has no longitudinal reinforcement and the lateral
reinforcement consists of square ties, the effective confinement was found by
Fafitis and Shah [76] to be about 40% of the confinement pressure for a square
compression strut with longitudinal reinforcement.
Recent work work by Ahmad and Shah [77] has shown that spiral reinforcement
is less effective for compression struts of higher strength concrete. The authors
also foundthatthe stress inthe steel spiral at peak load for high-strength concrete
is often significantly less than the yield strength. These conclusions are consis-
tent with results of experimental research at Cornell University. In the study by
Martinez, Nilson and Slate [78], an effective confinement stress "fs(1 - sid)" was
used in evaluating results, where "fs" is the actual stress in the spiral. The term
"(1 - sid)" reflects the reduction in effectiveness of the spiral associated with
increasing spacing of the spiral wires.

fea = flc (AlAb)o.s + 4.0 'Fctlat

fact
lat = f s (1 - sid) 2 As I (d s)

fs = C Jl 2 s I (1t d As) ~ fy

C = compression load

Jl = poisson ratio (= 0.16 - 0.3 up to 12,500 psi)

 73

There is not a general agreement on the effectiveness of spiral steel for

improving the ductility of high strength concrete compression struts, that is, for
increasing the strain limit and flattening the negative slope of the stress-strain
curve past the point of peak stress. The paper by Ahmad and Shah [77] indicates
that con'fining spirals are about as effective in flattening the negative slope of the
stress-strain curve for high -strength concrete as for normal-concrete. However,
the study by Martinez, Nilson and Slate [78] showed significant differences. Fig.
2.20 shows experimental stress-strain curves for different strengths of normal
weight concrete columns with varying spiral reinforcement. Three groups of
curves are identified by the three concrete strength levels studied. Each of these
groups consists of three sets of curves corresponding to three different amounts
of lateral reinforcement. Different behavior for comparable confinement stress is
evident. Not only is the strain at peak stress much less for high-strength concrete,
butthe stress falls off sharply just pastthe peak value. This is seen to be true even
for compression strut "NC169" with a very high confinement stress of 2500 psi.
Caution must be used when applying the results to very high strength concretes.

The basis for design of ACI 318-83 is the following equation

= 0.7 forties
= 0.75 for spirals

When the supporting surface is wider on all sides than the loaded area (Ab) ,de-
sign bearing strength on the loaded area may be multiplied by (A / Ab)o.5, but not
more than 2.
 ...,j
.f::io

-n
cO·
c
(i)
!"
I\J
o
om
o X
-"'0
C (1) 2~ l---l I , I Normal Weight Spiral Columns
3 :::!.
::J
(J)
3
(1) 4 .16 - in. 1102 • 406 • mm) C,lIndtr
'=+:::::J
NCI&91(2500~i9h Sirenoth I Shoke Rote 12,000 p.ln.(0.30mm)lmin.
I~O
as" I

3;; 20 .t-
f: • Unconfined Co'vmn SIrtn9th
:JJ~ (2500) • Effethv. Confinement
(1) (1)
Stress f 2 (1- a/dJ
-. en Axial
........ I Strenolh A.ial
-_.
~~
::::J
Stress.
kli
15 l--I I -, ~
100 Stress.
M Po
g
<
(1)
10 l-"" I #.; I I: l.I
en
o Low Strenoth

+: :=t:::
-0. (8001 50
.J:l>
x
...... ~ H· - - IU IIt - --lLJ[:J! l::k:::t (522)
m al
::::J (244
::::J o II~~
__._ III I I I I I I I I I I I 10
o
...,
I I I I I • I I I

3 " ...III\)
Pl Axial Strain. in. I in.
:E
(1)
cO'
:::r
.....
en
"2.
ii3
 75

Based upon a number of tests of cylinders loaded through rigid bearing

plates, in 1948 Billig [80] developed the following formula for permissible bearing
stresses:

fc=
3'
a 6 f'c (NA b
)113 -
<cf'

In 1952 Komendant [81] published the same formula with the exception that the
cube root was replaced by the square root. This was based on a substantial
number of tests using, again, cylinders loaded through bearing plates.
Four substantial studies performed under the direction of Middendorf [82] were
carried out in 1960. Using both rectangular blocks as well as cylinders ranging
from 6 in. to 16 in. in diameter, Middendorf reaffirmed the recommendations of
Komendant, and recommended the following formula:

He further recommended that the restriction fc3 ~ f c be dropped and the value
be increased to a multiple of fIe' probably 3 f'c . Middendorf [82] concluded that
the recommendations are applicable to concrete with f'c ranging from 4000 to
6000 psi.

Approximate expressions were developed by Hawkins [83] forthe bearing

strength of concrete members loaded through rigid plates.

fc3 = f'c {1 + K / (2 (f'Jo.s [c/b (3-alb)-1])

K = function of aggregates internal coefficient of friction

:::::: 50 - 65
a = length of shorter side of a rectangular plate or side of square plate
b = length of longer side of a rectangular plate or side of square plate
c = side length of a square block
76

For a=b=c and K=50 the increase of the bearing stresses is a function ofthe square
root of the concrete strength. The author proposed for design purpose a K equal to
50 (see Fig. 2.22). The results for block length equal to the plate dimensions are
shown in Fig. 2. 22.

In 1971, based upon further tests, Hawkins [84] recommended the following
formula for strip loading of concrete through rigid plates:

fc3 = 18.5 (flc )0.5 (d/(2 W))0.3

d/2 = distance from the block edge to the centerline of the plate
w = width of the plate (see Fig. 2.21 (a))

w
I

d/2

Figure 2.21 (a): Strip load dimensions

 77

Niyogi [85] discussed the problems associated with the calculation of the
allowable stresses and the probable mechanics of failure. The primary parame-
ters were the geometry of the bearing plate related to the loaded surface and the
plate geometry. Square, rectangular and strip loadings were considered. He pos-
tulated the following formula:

fc3 = f'c { 0.42 (ala' + alb') - 0.29 [(ala' - alb')2 + 5.06 )0-5

Dimensions and definitions are shown in Fig. 2.21 (b). According to Niyogi, the
bearing strength decreased for increasing height and eccentricity of the load.

2a

2a'
I I

.}b' 2b

Figure 2.21 (b): Variable definitions for geometry of the bearing and loaded plate
(from Ref. [85])
78

0
0
0
~
....

~
0
0
0

\ C\I
....

~'\ 0
0
0
0 ='
.... en
....a.
\ ~
0
0
0
co
-...
.s::

-
C)
C
Q)

1\
en

0
0
0
-...
Q)
Q)

U
C

1\
co 0
u
;

0
0
0

i\
~

::J o
. . . o
o
o
o o o o oo 0
o
OC\l
o o o 0 0
o o
~
8o o 0 o 0
....
<0
....
~
..... .... OJ <0 ~ C\I

bearing stress [psi]

Figure 2.22: Bearing stresses versus concrete strength (from Ref. [83])
 79

In another study Niyogi [86] studied various forms of lateral reinforcement.

Among the various forms used, single large diameter spirals gave the highest
efficiency forthe same percentage of steel. Provision o'f reinforcement had some
beneficial effect on the resistance of specimens against initial cracking, depend-
ing on the form and amount of reinforcement and the relative size of the bearing
plate. The cracking load, particularly for specimens reinforced with large spirals,
and forsmall relative loaded areas increased with increasing percentage of steel.
Niyogi also suggested that the concrete efficiency of spirally reinforced concrete
be expressed relative to that of plain concrete by a linear relationship depending
on the percentage of steel.

Fafitis and Shah [76] presented an analytical expression for the stress-
strain curves of confined high strength concrete based on several sets of
experimental data. The peak stress (fes ) and the corresponding strain (ces ) are
given by the following equations.

f'es = 0 [ f'c + (1.15 + 3048 / fIe) flat]

Cc3 = 1.027 10-7 f'c +0.0296 0 flat / f'c + 0.00195

o = 1 + 15 (flat / f'c)3

In general the following stress-strain relationships are proposed:

f = fc3 [1 - (1 - c/cc3 )A] for c s:; cc3

f = fc3 exp[ -k (c - cc3)1.1S] for c;;::: cc3

A = Ec cc3 / fc3
Ec = 33 w 1.S (f'c)o.s
w = weight of concrete [Ibs / ft. 3]
k = 0.17 fIe exp (-0.01 f'a/ ( 1)
01 = 1 + 25 flat/f'cl1 - exp (- f'/6500)9]
80

The analytically predicted values of peak stress compared to the test resu Its (with
concrete strength ofthe specimens from about 3000 psi to about 10,000 psi and
the confinement pressure from about 250 psi to about 3000 psi) gave differences
from 3.5 to 26.5%.

Schlaich et. al. [2] propose the following equation to compute the confined
strength for spiral confining reinforcement:

For square compression struts and square confi nement reinforcement the lateral
pressure can be computed by reducing the equivalent circular compression
strength by 50%

d = equivalent diameter = side length of confined square core

Roberts [89] tested local anchorage zone specimens with spiral con'fining rein-
forcement. Test results of five different authors were compared with theoretical
approaches to determine the best fit function. For the basic concrete efficiency
factor the same term was used as shown in 2.4.1.1 for concrete strength in com-
pression diagonals. The following approach is used for the comparison

fe3 = [0.5 + 15/(f'c )0.5] fIe (AIA b )0.5 + 4 (Acore I A b ) flat (1 - sId)

A = the area of confined concrete concentric with and geometrically similar

to the bearing plate
effective area of bearing plate (Roberts imposed a stiffness require
ment on the plates)
Acore = d 2core 11: I 4

=
 81

For square confining reinforcement with no longitudinal reinforcement and with

lateral reinforcement consisting of square ties (ineffectiveness of ties without
longitudinal reinforcement, was studied by Sheikh [87]) the following approach
was used (for the test results by Muguruma, et al [88]).

fe3 = [0.5 + 15/(f'c )0.5] fIe (NAb )1/2 + 1.0 (Acore I A b ) flat (1 - sid)

Table 2.11 gives the results from the statistical analysis:

Table 2.11 : Statistical data for confined concrete strength with various test
data

author specimens experimenVtheory

mean standard deviation

Roberts [89]: 28 0.95 0.15

Wurm & Daschner [90]: 29 1.05 0.06

Niyogi [85]: 39 1.07 0.19

Muguruma [88]: 25 1.38 0.22

Martinez [78] 11 1.31 0.09

Total 122 1.125 0.24

Different sensitivity analyses have shown that the reduction factor (1 - sid) forthe
confinement strength has a signi'ficant influence. In the following approaches the
squared reduction factor (1 - s/d)2 is used for the comparison. For practical
application the reduction factor forces the designerto use smaller spacings for
confinement reinforcement.
82

The term "(1 - sid)" reflects the reduction in effectiveness of spiral associated with
increasing spacing of the spiral wires. For a better correlation with test data the
second term intheequation was changed. The following approaches were used:

fe=3[05
'
+ 15/(f'c )0.5] f' c (AlA b)0.5 + 4 •0 (Acore I Ab)l af t (1 - s/d)2

Comparison with the test results (see Fig. 2.23 and Table 2.12) shows that the
proposed equation with the effective con'finement strength is a generally conser-
vative and safe approach. The 95 percent limits (X - 2a) would be 0.65 which is
also the minimum actual test result.

Table 2.12: Statistical data from Figure 2.23 for confined concrete with an
efficiency factor "ve = 0.5 + 15/(f'c )0.5"

Mean: Std. Dev.:

11.124 1238
Minimum: Maximum:
.85-' 2.128 134.851

By using a higher concrete efficiency factor "V =0.5 + 20/(f'c )0.5" the statistical
mean of the comparison is 1.05 (see Table 2.13). The 95 percent limits become
0.61 with the minimum actual test result 0.62. Either of the'se efficiency factors
could be used in practice.

Table 2.13: Statistical data for confined concrete with an efficiency factor
ve = 0.5 + 20/(f'c )0.5

Mean: Std. Dev.:

1,·054 1221
Minimum: Maximum:
.021 2 120.40
 o 83
o
....-----.,.-----.,.------:----'"""'i"'---..
...
O
V

o
o
o
...
N

0
0
0
...
0
-
1-

m
-
Q.

c -
.c
0)
8 G)
--
1

-0
co ~

m
,....
0
G)
II G)
~~ ~
u
0 0 0 c
Q) Q) 0
0

--
.t= .t= 0
~~ co .U
c: c:
Q) Q)
E E I
"i:: "i::
Q) Q)
c.c.
>< ><
Q) Q)
0
0
0
v

tt 0
0
0
q &q ON
It) 0 ~
N N ..... .... 0 ci

experiment I theory

Figure 2.23: Comparison of test results with theoretical approaches for

predicting confined concrete strength
84

2.5 Tie Background

The ties in a strut- and- tie model have to be provided either by

reinforcement or by the concrete tensile strength. Availability of the concrete
tensi Ie strength to resist loadings depends to a large extent on the restraintforces
in the concrete and its load history. Microcracks from other former load cases,
thermal stresses or shrinkage may reduce the concrete tensile strength locally
to zero. Considering these circumstances individually in each case,dependence
on the concrete tensile strength has to be chosen very cautiously. The
dependable tensile strength contribution is usually assumed as zero or as only
a small fraction of the theoretical computed tensile strength. For most practical
detailling problems, concrete tensile strength should be ignored. However, in
certain cases such as massive wall sections, such discounting of the tensile
strength contribution may produce very conservative designs. Dependence on
tensile strength should only be utilized for equilibrium forces in those cases
where no progressive failure is possible. If such dependence is made, restraint
forces and microcracks due to shrinkage and temperature have to be taken into
account.

Because of the great incompatibility of their strains at peak values, the

tensile strength of concrete and the yield strength of reinforcement cannot be
counted on jointly for carrying internal forces. However, in those regions where
the tensile strength of concrete is counted on for strength purposes, nominal
reinforcement improves the reliability of the concrete's tensile strength and
thereby contributes to the overall strength and safety of the structure. Although
it is difficult to develop design criteria for this case, it would be even worse to
maintain the formalistic view that the tensile strength of concrete cannot and
therefore must not be utilized. Following the flow of forces in a gap free and
consistent mannerwith strut- and-tie- models will inevitably showthat equilibrium
can frequently only be satisfied if ties ortensile forces can be accepted in places
where, for practical reasons, reinforcement cannot be provided.
 85

Some examples which demonstrate that in fact certain types of members

presently depend on the tensile strength of the concrete:

plain or unreinforced concrete members such as pedestals

slabs without stirrups or other vertical reinforcement yet carrying shear
bond strength and lap splices in reinforcement
concrete joining and fastening elements (anchor bolts, expansion- and
adhesive anchors)

Representative design values fortensile strength obtained from tests and

measurements vary greatly as shown in Fig. 2.24 [91].
The full tensile strength should never be counted since restraint forces and
microcracks have to be taken into account, even in uncracked and unloaded
concrete. In many important cases, the engineer has to deal with larger crack
widths then microcracks (0.002 in.). Although the tensile strength at a section with
an open crack is certainly limited, still sizeable shear forces can be transferred
across large cracked interfaces. Figure 2. 25 shows some test results. If the
tensile strength in concrete structures is used for the analysis, then the stress
peaks at outer fibres or at failure zones may be averaged over a length of
approximately 2 in., but not more than 3 times the largest aggregate size [28].
The design engineer wi II have to decide if, and to what deg ree the tensile strength
can be depended on for carrying load. Although design standards frequently
require that the tensile strength of concrete be neglected, such restrictions are
usually qualified as in ACI Building Code Section 10.2.5 by insertion of the
wording "in flexural calculations of reinforced concrete." [4] Thus, these limita-
tions do not apply to strut-and-tie models at discontinuity regions, per se.
86

--
Ui
a.
800...-----------------------,
6.0(fc)E.5
.r: 1.7(fc)E.67
'0 600 0.44(fc)E.8

-.
cCl)
0.05fc
en
.! 400
Ui
c
oS

...-
CD
CD
U
C
200

o
u

2000 3000 4000 5000 6000 7000 8000 9000

concrete compressive strength [psI]

Figure 2.24: Various approaches for the tensile strength of concrete [91]

1.0
~ experiment

--
"#.

-..
.r:
CD
0.8

0.6
• 1/(1 OOOw+')

c
-CD
en
CD
0.4
enc
CD
t- 0.2

0.0
0.00 0.01 0.02 0.03
crack width [In.]

Figure 2.25: Crack width versus tensile strength reduction factor (from Ref. [53])
 87

If the tensile stress field is crossed by a compression field, the reduced

biaxial strength must be considered. Fig. 2.26 provides a safe assumption based
on test results [68]. Usually, tie forces are resisted by reinforcement placed sym-
metrically about the line of action of the force. The reinforcement must extend the
enti re length ofthe tie and should be properly anchored at the nodes. The amount
of reinforcement must extend the entire length of the tie although the amount may
vary from one set of nodes to the next set of nodes as bars are cut off or bent.
Reinforcement should be proportioned so that at the ultimate design load it will
just reach yield. In order to ensure a ductile failure mode, sufficient yielding must
occurto allow the formation of a mechanism priorto crushing of the concrete. Tie
reinforcement may consist of single or multiple bars or of prestressing bars or
strands. However, in many cases it is not readily apparent whether ties should
consist of a few large reinforcing bars or a large number of smaller bars. A more
theoretical approach indicates reinforcement within a tie should undergo similar
strains in order to act as a unit or a single tie. A practical advice in this matter is
to use normal diameter bars, normal spacings for the bars and take care to
provide sufficient concrete cover. In addition the concrete curing and quality
control during the hydration process plays an important role for development of
efficient concrete tensile strength.

f'c

1.0 1 - - - - - -__

0.25 1.0

Figure 2.26: The biaxial compressive -tensile strength of concrete

88

2.5.1 Prestressing forces

Intensive research work has been done to develop procedures forthe cal-
culation of the flexural strength of prestressed concrete structures. Also a large
number of prestressed concrete specimens have been tested to determine their
strength in resisting shear, or combined moment and shear, with or without web
reinforcement [91].
Prestressed concrete members pose a somewhat different challenge in
detailing. The computation of the ultimate strength of such members is generally
very similarto that of non-prestressed members when appropriate assumptions
and calculations are made to determine the stress level in the tendons at failure.
This stress level depends principally on whether the tendons have been effec-
tively bonded to the concrete. Required quantities of non-prestressed reinforce-
ment to resist anchorage force concentrations, to distribute support reactions,
and to counteract similar concentrated load effects can be effectively determined
using strut- and- tie models for these 0 or discontinuity zones. Use of such
models actually gives superior understanding and very good guidance for the
special reinforcement required. However, whiletheycan be used if desired, strut-
and- tie models applied to the B zones tend to be time-consuming and somewhat
unwieldy in application.
Since strut- and- tie modelling is a plasticity approach, it is quite inefficient
and artifical to use if for checking elastic stresses such as allowable stresses at
service load levels. Such checks are important in prestressed concrete members
in which it may be desirable to prevent or severely limit crack formation at the
service load limit state. It is possible to make such calculations using a type of
strut- and- tie model for B zones, but it does not appear very efficient to do so. In
order to illustrate the nature of the computations, a type of model adaptable to
allowable stress calculations will be shown in this section and used in Example
4.6. Examination ofthe actual calculations will indicate the general impracticality
 89

of such an approach in B regions which are efficiently treated by ordinary

procedures.
Forthe efficient use of strut- and- tie models in prestressed concrete, the
humanly controlled prestressing forces are applied in the same manner as other
loads to the strut- and- tie models. This requires knowledge of their magnitude
and of the effective points of application of both horizontal and vertical compo-
nents. Subsequent changes in the prestressi ng forces in the tendons due to other
load effects are treated as internal forces or tie forces.
In orderto apply the strut- tie- model to a prestressed member some basic
information has to be known. First and foremost is an accurate estimate of the
actual prestressing forces, both initially and after time dependent losses. In the
pretension process the first step is to stretch high strength wires or strands
between end piers of a prestressing bed. After placing the member forms and
casting the concrete, the wires or strands are cut between members. The
resulting release of wire tension Pi is equivalent to applying Pi as an external
compressive force. The strands are usually eccentric and the prestressing force
introduces a moment as well. Because of Pi and the accompanying eccentricity,
rather large calculated tensile stresses develop on the top beam fibers. Except
near the ends, those need not be considered dangerous because nearer
midspan these stresses never exist without the counteracting compression from
dead load moment [55] (see. Fig. 2.27). The stress levels do need checking;
shrinkage of concrete, creep underthe applied prestressing forces, relaxation of
steel and loss-of-steel stress from elastic shortening reduce the applied
prestessing force to an effective prestress Pse •

Shrinkage of concrete is influenced by many factors, such as volume-to-

surface ratio, relative humidity, and time from end of moist curing to application
of prestress. Since shrinkage is time dependent, about 80% will occur in the 'first
90

year [92]. The loss o'f prestress due to shrinkage is the product of the effective
shrinkage "Csh " and the modulus of elasticity of prestressing steel.

Csh = 8.2 10-6 (1 - 0.06 VIS) (100 - RH)

RH = relative humidity
VIS = volume-to-surface ratio

= modulus of elasticity of prestressing steel

 91
n r-----------,
·bJ L_________
I fpi I
-~

(a)
I. Before cuttin& wires becomes - FINAL (with time)

')3 ....
r---------.,
_~IL_ _f._r_·_fp~i_-_I_os_ses_~I!",+,
...., .
..J
. • PU<Pi

I~
t '
I + 11 + 111 Combined. with allowable stresses I + 11 Combined, with allowable stresses

E~'Z3W~
flO I.'~
Comp. < O.GOln ' Tens. <. 6 ..Jf7

(b~ (cJ

Ultimate compo

l JI \ I j
Cracked
(d)

(a) prestresing wire stretched between casting yard abutments

(b) initial stresses at midspan

(c) stresses at midspan under full service load

(d) ultimate strength conditions

Figure 2.27: Pretensioned beam (from Ref. [55])

92

Creep is assumed to occur with the superimposed permanent dead load added
to the member after it has been prestressed. Loss of prestress due to creep is
computed for bonded members from the following expression:

PIcr =Kcr E/E(f.-f)

S c clr cd

Kcr = 2.0 for pretensioned members

Kcr = 1.6 for post-tensioned members
Ec = modulus of elasticity of concrete at 28 days
fcir = stress in the concrete at level of steel due to prestress force
fcir = P/A+Pe2 /I-M w ell
e = eccentricity
I = moment of inertia of the section
A = cross section area
Mw = moment acting due to only its own weight

= average compressive stress in the concrete along the member

length at the center of gravity of the tendon

For unbonded tendons the average compressive stress is used to evaluate

losses due to elastic shortening and creep of concrete losses.
The loss of prestress force strands in a duct due to friction and wobble within a
duct according to the "Design and Construction Specifications for Segmental
Bridges" [31] shall be calculated usi ng the equation:

PIfw -- PIfw,x e (!La + k I)

ex = angle from the curved tendon

k = wobble coefficient
~ = friction coefficient
I = length over curved tendon (see Fig. 2.28)
 93

The proposed friction - "J.L" and wobble coefficient "a" are shown in Table 2.14
[31]:

Table 2.14: Friction - and wobble coefficient (from Ref. [31])

Type of strand Friction coefficient Wobble coefficient

~[I] a [k/ft]

wire or strand in galvanized

metal sheating: 0.15-0.25* 0.0002
high strength bars in galvanized
metal sheating: 0.15 0.0002
wire or strand in internal
polyethylene duct: 0.23 0.0002
wire or strand in straight
polyethylene duct (external to concrete: 0.00 0.00
rigid steel pipe deviators: 0.25* 0.0002

* lubrication will probably be required

Figure 2.28: Frictional loss along circular curve (from Ref. [92]
94

Loss of prestress due to steel relaxation overthe time interval t 1 to t may be estimated
as follows for ordinary stress relieved strand::

PIS = [log ( 24 t - log 24 t 1 ) / 10] x fs / (0.85 fy ) - 0.55

fs / (0.85 fy ) - 0.55 ~ 0

= steel stress level at beginning of time intervall t 1

As proposed in [31] for low relaxation steel a different expression can be used:

PiS = [log ( 24 t - log 24 t 1 ) /45] x fs / (0.9 fy) - 0.55

The value for t 1at the time of anchorage of the prestressing steel shall be taken as
1/24 of a day so that log t 1 at this times equals zero.

It is difficult to generalize the amount of loss of prestress, because it is dependent

on so many factors. For average steel and concrete properties, cured under average
air conditions, the tabulated percentages of Table 2.15 may be taken as represen-
tative of the average losses [92].

Table 2.15: Loss o'f prestress (from Ref. [92])

type of loss pretensioning [%] post-tensioning [%]

elastic shortening 4 1
creep of concrete 6 5
shrinkage of concrete 7 6
steel relaxation 8 8
Total loss 25 20
 95

After an accurate estimate of the prestressed forces is obtained, the

spread of the highly concentrated forces into the member must be approximated
in the strut- and- tie models. Based on comprehensive finite element analyses
by Burdet (42) and comparison with experimental values by Sanders (43) a
conservative values of the diffusion angle was chosen. The proposed compres-
sion strut diffusion angle is 12°. It is the first term from the more comprehensive
expression for the diffusion angle for a compression strut given in Section 3.4.
In post-tensioned concrete the point of application of the major or initial
force is relatively clear. Except for frictional and radial forces in curved tendons,
the post-tensioning forces are applied at the anchorages. However, in preten-
sioned concrete, the initial prestressing forces are distributed over longer
lengths. With both bonded post-tensioned tendons and bonded pretensioned
tendons, subsequent stress changes can be induced by flexural actions.
The development length of the prestressing strands is another important
factor. Two types of bond strength must be considered. The first of these is
referred to as "transfer bond stress" and has the function of transferring the force
in a pre-tensioned tendon to the concrete. Transfer bond stresses come into
existence when the prestressing force in the tendons is transferred from the
prestressing beds to the concrete section. The second type of bond is termed
"flexural bond stress" and comes into existence in pre-tensioned and bonded
post-tensioned members when the members are subjected to external loads
[93J. After cracking, the increase in steel stress above effective prestress
develops flexural bond stress between the steel and the concrete. Flexural bond
stress does not exist in unbonded, post-tensioned members. Transfer bond
involves the Hoyer effect. When a prestresing tendon is stressed, the elongation
of the tendon is accompanied by a reduction in the diameter due to Poissons'
effect [94]. Hoyer [95] pointed out that, on release of the wire from its temporary
anchorage on the prestressing bed, the end of the wire swells as a result of the
recovery of the lateral contraction and develops a wedge effect because the
prestressing force must diminish to zero at the end of the wire. The stress in the
wire is zero at the extreme end and is at a maximum value at some distance called
the "transmission length" from the end of the member (see Fig. 2.29). The length
over which the prestress transfer bond exists is termed the prestress transfer
length, and depends mainly on the amount of prestress, surface condition of the
strand and the concrete strength. Three factors which contribute to bond
performance are adhesion betweeen concrete and steel, friction between
concrete and steel and mechanical resistance [96]. Libby [93J pointed out that
under normal conditions, the transmission length for clean seven-wire strands
can be assumed to be equal to 50 times the diameter of the strand. The
transmission length oftendons can be expected to increase from 5 to 20% within
one year after release as a result of relaxation.
96

warpin; DfCITJSS-&ecfion (/fen'

fransret'St
lenskn

6rm1sfressbetween
wire ant!CIJI1cref!

!JonrJlengflt fj

site/stress

II I L10i0H"'- CD/1~~ stress

Figure 2.29: Bond stress distribution at the end of a bond anchorage of a pre-
tensioned wire [94]
 97

In order to apply the strut- and- tie- model to pretensioned concrete a

model has to be developed for the force transfer into the concrete over the
transmission length. Such a transfer length model is shown in Fig. 2.30(a). Use
of such a model will indicate special lateral and vertical reinforcement is required
in the transmission zones to resist the splitting tensile components T p' A second
model can be developed for the internal forces induced through the eccentricity
of the applied prestressing forces. Figure 2.30(b) shows a suggested model
which can be used to compute effects of various load cases for prestressed
members.. The internal force combined load action case shown in Fig. 2.30(b)
is the sum of the strut-and- tie-forces developed by external loading w (dead-
and live load) and the prestressing forces. As mentioned by Schlaich et al. [28]
the prestressing steel will serve as regular reinforcement, if it is bonded with the
concrete. The remain ing force in the tension chord T can be taken by the reserve
capacity of the prestressing steel "Tp " and the capacity of any additional non-
prestressed regular reinforcement "Ts "'

The reserve capacity "Tp" of the prestressing steel which is still available for use
as a tensile chord after prestressing is equal to its yield force (under ultimate load
the prestressing steel is strained beyond its yield strength) minus the prestress-
ing force which is applied to the member as a load

Tp =Ap fpy - P

P = prestressing force = Ap fpi

fpi = effective prestress level in the prestressing steel
fpy = yield strength of the pretressed steel
Ap = area of the pretressed steel

A major problem in the calculation of "elastic stresses" using the strut- and- tie
models is that the fully plastic strut- and- tie model does not adequately represent
98

the compatibility effects so important at the service load state. A good example
is the effect of tendon eccentricity. As shown in Fig. 2.30(c), the elastic stress
distribution can result in tension on the tob fiber. If the fully plastic strut- and- tie
model shown in Fig. 2.30(d) is used, the free rotation of the joints possible in the
assumed fully plastic members results in concentration of stresses in the lower
chord only and no forces in the upper chord, verticals or diagonals. From
knowledge of compatibility this is clearly inadmissible although equilibrium is
satisfied.
If the end region of the eccentrically prestressed member is treated as a
'0' region as shown in Fig. 2.30(e), and if the boundary forces corresponding to
the calculated elastic stresses at the boundary with the 'B' region are applied as
indicated in Fig. 2.30, a very realistic force path can be used to constructthe strut-
and- tie model shown. T, is the tensile force while C2 is the offsetting compression
force. This '0' region model clearly indicates that tensile reinforcement is
required on the top fiber area and along the support face jf concrete tensile
stresses are not considered to adequately carry the tensile force. This tensile
force, T 1 ,is the force which must be applied to any strut- and- tie model used
to represent the prestressing effect (see Tp in Fig. 2.30(b)). In Fig. 2.30(b), the
chord forces shown are nominal forces. The load factors as well as the material
reduction factors have to be taken into account in design. In post-tensioned
members, if no bond is provided after prestressing, the prestressing steel cannot
be considered as reinforcement. The tendon force is applied as an applied force
or the tendon is considered as a constant force tie. Figure 2.31 shows the
proposed strut-and-tie-models for a prestressed concrete member with curved
or harped tendons.
 Cp1 =P / (2 cos ~1 )
Cp2 =P /2
Tp1 = (P tan ~1 ) / 2 99

,- -.
1 I
-
-----
-Cp2' -
--- - - - ;-
--- --
_I I

-- -
_Cp1 1 1

~p1 P
•I ,
I

I I

-.... -.... Tp1 -.... -.... -....

-.... -.... -.... -.... -.... -....
Cp1 ....
L - - -- -....
"'::. - ---
-.... -....
--

transfer length = 50 CPb .1

Proposed compression strut diffusion angle:

(a) Pretensioning force transfer

r
h

L
-I
Tp = wt b (- P I At + P e ~ I It)
Cp = we b (- PI Ac - P eYe I !c)

(b) Overall Model

Figure 2.30: Strut-and-tie-models for prestressed concrete

100

P Pec
- A -+ -1-
p J.....~---------J=~-:..:.--~--Jl
....~. . . -

(c) Eccentric Pretension Force

o o o o o o

p p

(d) Fully Plastic Strut -and - Tie Model

....... ,
.......
'r-----
I
I

./
.J-- ------.4-__\. ~-- C 3
/"
p .L..:~ -t!===~

(e) Strut-and-Tie Model for the end "0" region

 101

.:. :::,.;

e = f(x); Z = f(x)

rh

L
P en Cr1 = P (sin ')'1 + lin 'Y2)

~p

rh -Z

L~~~~~~I----'- A
A

I- ·1
Tenllon chord = k T (load) • C (pr-Itr...lng)
Compr••• lon chord = IT (pre.tr..lln;) - C (load)

Ver1lcal chord = k Tv (load) • Cr (pre.tr...lng)

Figure 2.31: Strut-and-tie-model for prestressed concrete beam with

curved tendon
 102
Should it be desired to use strut- and- tie models in 'B' regions it can be done, if
inefficiently. This section presents a summary of the design steps involved in
such use of the strut- and-tie approach for prestressed concrete members. Table
2.16 show the analysis procedures for prestressed beams using the strut and tie
model. The symbols are related to Fig. 2.30 and Fig. 2.31 and are defined as
follows:

Table 2.16: Design steps for 'B' regions of prestressed beams using the strut-
and-tie-model

step load situation selected model strength limits

(1 ) Pretensioned wires cut a= 12° Cp1 = P 1 (2 cos 12°)

Ap1 = (7 dJ2 1t 14
Cp1/Ap1)~v.fc or
service state limits
(1 ) Bonded or immediately
Unbonded after prestressing ~ =12+31(alh) anchorage zone
Post- analysis (see 4.2.6)
Tensioned

(2) Pretensioned
Bonded or Prestress + compression chord 1: [T(prestr.)- C(load)-T]
Unbonded Dead Load ~ 0 (tension)
Post- tension Chord 1: [T(load] - C(prestr.)]
Tensioned ~ b WI v. * f'c
or service state values

(3) Pretensioned
Bonded or Prestress + compression chord I [T(prestr.)- C(load) - T]
Unbonded Dead load ~bwc v.· f c or
Post- + Live Load service state values
Tensioned tension chord 1: [T(IOad)- C(prestr.) -1]
~ b WI 3 (fJ0.5 (tension)
verticals 1: [Tv(load] - Cr(prestr.)
~T

(4) Pr.etensioned
Bonded Post- ultimate compression chord t C (load) ~ b we v. * flc
Tensioned
tension chord t T (load) s T + P

(5) Unbonded
Post- ultimate compression chord t C (load) S b Wc v: f c
Tensioned
tension chord tT(load)sT+P
 103

Load situations:

I = prestress alone = f se =(fpi - losses)

dead load
" =
III = live load

ve* = 0.45 (ACI 318-89-chapter 18) [4] (service state)

proposed strut-and-tie-model: v e= f( f'c ) (see Section 2.9)

T = T p + Ts

Tp = A p fpy - P

Ts = regular reinforcement capacity

= As fSY
fpy = yield strength of the pretressed steel
Ap = area of the pretressed steel
fSY = yield strength of the regular reinforcement
Ap = area of regular reinforcement
P = prestressing force

For curved tendons the radial compression component "Cr" can be computed from
the deviations between the tangents to the curve (Fig. 2.32). The section length
should be chosen according to the spacing of the vertical tension members (stirrups
or lumped stirrups).

P as a tension force I Cr
I
~ I
""""""IIIII1!!!!!Z:::::~!I~!!!!!!!!!!!!!!!!!!!!!~P~a~s a tension force

Figure 2.32: Approximation for radial compression component of curved tendon

104

= vertical compression component ofthe curved prestressing tendon

= spacing of the vertical chords from the chosen strut-and-tie-model

If the prestressed tendon is following a parabolic curve a uniform load over the
length of the span can be computed (see Fig. 2.33):

Wb = 8 P h / (L)2

P = prestressed force
h = sag of parabola
L = length of span

- - _ Parabolic'tendon _ _ -0
----.,....---
-11
L

Uniform load

to.
Concrete IS freebody

Fig. 2.33: Prestressed beam with parabolic tendon (from Ref. [92]

The width of the compression- and tension chord can be found according to Fig.
2.34. The half ofthe width is the distance from the center line of the chords to the
outside fiber of the member.

cJ2.=
. ..---;;;; -.--
r
ompression chor we
--L
z ""':::::~ """"== ~

L tension chord
'w_==
- wtr
,
Fig. 2.34: Width of the compression - and tension chord
 105

2.6 Node Background

The nodes of the strut- and- tie- model represent the locations of change
of direction of internal forces, which in the structure occurs over a certain length
and width in the node region. The intersecting strut- and- tie forces have to be
linked together and balanced in equilibrium in the node region.
If one of the struts or ties represents a concentrated stress field (e.g. near a single
load, a support or concentrated reinforcement) the deviation of forces tends to
be locally concentrated and the node region is relatively small. These kind of
nodes are called "singular nodes" and have to be dimensioned with special
care.The special studies about CCT (Com pression-Com pression-Tension) - and
cn (Com pression-Tension-Tension)- nodes given in Ref. [38,39] and summa-
rized in Report 1127-1 provide inform ation upon which design recom mendations
are based.
Splices or joints of overlapping reinforcement are the special but frequent case
of nodes occuring between two ties or reinforced struts for which specific rules
based on comprehensive tests have to be used.
Where wide concrete stress fields join each other, the node region extends over
a considerable length of struts and ties. Such "smeared nodes" need not be
checked for safety, if the same D-region contains a singular node.

Numerous possibilities exist for detailing nodes. In all cases, the flow of forces
can be visualized by strut- and- tie- models with singular nodes at the deforma-
tions of the bar. Bond is in fact a load transfer mechanism involving both
compressive and tensile stresses.

Singular and smeared nodes may be grouped into subsets relating to the
type of elements which they join. Four different kind of nodes can be worked out
from a strut- and- tie- model (see Fig. 2.35).

CCC: Compression- Compression- Compression

CCT: Compression- Compression- Tension
cn:Compression- Tension- Tension
Tn: Tension- Tension- Tension
106

(a) CCC·Nodl (b) CCT·Node

(e) CTT·Node (d) TTT·Node

Figure 2.35: Types of nodes

Evaluation of the nodal regions includes checking the nodal boundary stresses
and determining reinforcement development requirements for nodes which con-
tain tension ties. Each of these steps requires the determination of the physical
boundaries of the node. The dimensioning of nodes is largely determined by two
constraints:

All the lines of actions of struts and ties as well as any external forces must
concide
The widths and relative angle of the struts and ties constrain the nodal
geometry
 107

If the nodal geometry can be varied it should be chosen to minimize the stresses
in the nodal region. This is accomplished by selecting a geometry in which the
stresses along the border of the node do not exceed the limiting value of the
effective concrete strength (fce=ve f'c).ln orderto get a state of planar hydrostatic
stress, the geometry should be selected so that the stresses on all the node faces
are equal. Both principal stresses within the nodal region would then equal the
stress at the boundary of the node [27, 28].

2.6.1 CCC· Nodes

For a GGG-node under a hydrostatic stress state the strut forces are pro-
portional to their width and the sides of the node are perpendicular to the axis of
each of the struts. It should be recognized that the geometry of the model may
not allow for equilization of the boundary stresses. Such a situation is shown
in Fig. 2.35a. Following Schlaich and Schafer [2], this stress state is tolerable if
the maximum ratio of stresses between any two sides does not exceed 2.0. In
order to get an hydrostatic state of stress, the geometry of the node can be
changed as shown in Fig. 2.36b. The intersection ofthe strutcenterlines actually
lies outside the nodal region in this case. Bottle-shaped struts are often used
where one of the nodal boundaries is fixed as in the case of a node adjacent to
a bearing plate. A reduction of the width of the struts is required to produce a
hydrostatic state of stress. In this case a more convenient approach proposed by
Schlaich and Schafer [2] can be used to check the concrete strength in the nodal
zone (see Fig. 2.37):

ao = a * tan <PC2 * tan <Pes / (tan <PC2 + tan <Pes )

crb = G1 / (a b)
crco = Go / (ao b)

A hydrostatic stress in a GGG-node is only given if the compression struts are

perpendicular to the node sides.
108

(a) eeG-node with unequal pressure (from Ref. [7])

(b) Struts created by hydrostatically dimensioned node (from Ref. [7])

Figure 2.36: eee Nodes

 109

1r...·I-------a ----~~I

)'
W2

/
0"1 =02 =0"3 =0"0 (hydrostatic stress)

a * tan 1P2 * tan 1P3

aO· =
tan 1P2 + tan 1P3

for 1P2 = <>3 = IP

aO = a I 2 tan <>

Figure 2.37: Dimensions for hydrostatic stress check in eee-node

110
For design purposes some general rule has to be adopted to check the stresses
in a eeG-node. The stresses from the struts without bearing plates must be
checked with dimensions relating to the bearing plates. It is very useful to
subdivide the node under the bearing plate into two sub-nodes as shown in Fig.
2.38. A key assumption is the distance from the centerline of action of the force
to the sub-nodes.The distance proposed is the quarter width of the beari ng plate
(a/4). By studying the strut- and- tie- model shown it can be seen that for eee-
nodes with only one bearing plate, the angle of inclination of the struts relative to
the plate becomes an important factor. In order to conform with test resultsof
post-tensioned anchorages, the eee-nodes are best split into two parallel sub-
nodes [42, 43] as shown in Fig. 2.38.The strut- and- tie- pattern depends on the
bearing plate width, a, because for increasing width a decrease in the required
tie force "T" must follow. The design concrete efficiency factor v recommended
in Section 2.8 is a lower bound solution and compared to test results gives a safe
approach. The stresses underthe bearing plate should recognize the degree of
local confinement and be checked utilizing the proposed expressions of Roberts
given in Section 2.4.
a/2 CC1 a/2
.-
<: ) < )-

I. a/4 a/4
1
I II

Cc2 ;
~4~-r-{~,....
~ j;Y ,
&4
Cc3
;' I ,
; I "
fC IT ~
I I I
I I I
I I I
I I I
I I I
I I I
I

Figure 2.38: eee-node

 111

The two approaches to compute the stresses in a CCC - node are compared in
Fig. 2.39. It can be seen that for the "hydrostatic" stress solution the width of the
horizontal compression strut is dependent on the compression angle and on the
width of the bearing plate. Forthe "quarter width ofthe bearing plate" solution the
horizontal compression strut width is dependent only on the bearing plate width.
The comparison is made for equal compression angles with the following
equations:

- hydrostatic stress solution: aD = a tan <1>2 tan <1>3 / (

tan <1>2 + tan <1>3 )
- quarter width of the bearing plate solution: aD = a / 2

a = bearing plate width

aD = horizontal compression strut width

3.0
.c
:c ---lil'- hydrostatic stress solution
'j
--+- aG-a12
~ 2.5
a.
Dl
c
;:
as 2.0
III
.e
oS
"0
'j 1.5
c
0
"ii
..
ell
III
ICl. 1.0
E
0
u
iii
C 0.5
2
;:
0
.c
0.0
0 20 40 60 SO 100
compression strut Ingle to)

Fig. 2.39: Comparison of the horizontal compression strut width

112

2.6.2 CCT - Nodes

Forthe CCT-nodes, two different reinforcing details can occur. Theoretically, but
rarely occurring in practice, the anchorage of reinforcement can be developed by
anchoring the tie forces from behind with an anchor plate (Fig. 2.40). The usual
and more practical case is anchorage by providing sufficient development length
behind the node as shown in Fig. 2.41. When using an anchor plate, the deter-
mination of the node geometry is clear. Special care sl10uld be taken to provide
adequate bending strength and stiffness in the anchor plate (plate bending
results in higher bursting forces) and to provide a proper connection with the tie.
A smooth surface for the tie where it crosses the node is theoretically betterthan
a good bond quality because strain compatibility with the bonded bar will tend
to crack the concrete. Bearing plate anchorage of tie forces usually means
diversion of compression fields. The compression stresses of the stress fields
concentrate on the steel plate's surface, if the tie is developed in this way. The
curvature of a deviated compression field is largest at the origin immediately
adjacent to the bearing plate.

In the more usual case of reinforcing bars directly anchored without

plates, either straight bars, hooks or loop anchorages may be used. Loop anchor-
ages with confining direct pressure as from a bearing or direct load point are
preferred. Hooks shall preferably be placed to have confining pressure trans-
verse to the hook plane. Sufficient anchorage lengths have to be provided within
as well as behind the node, if necessary. Anchorage begins where the compres-
sion struts (see Fig. 2.41) meet the surface of the bar. The bars should extend
to the other end of the node region in order to engage the outermost fi ber of the
deviated compression strut.

For the effective widths of the struts- and- ties different proposals can be
found in the literature. The equilvalent concrete area approach [23,24,56] (see
Fig. 2.42) describes the width as follow:

w3 = co h = P fy h / ( veftc)
 113

..
Bearing plate

Figure 2.40: Anchorage detail for GGT-node:

Anchorage of reinforcement with anchor plate (from Ref. [7])

A wcl

7
':"'~l:::::.~o:-- Com pre ssIon St ru t

...
wT3
1-

Bear1ng plate

wc1 = wc2 sin «b+ wT3 cos' %.

Figure 2.41: Anchorage detail for GGT-node with directly anchored bars
114

wh

t-
(1-2 W)h

Tension Tae
_L T~ wh
B

v
r
Figure 2.42: Equivalent concrete area approach to define the tie width (from
Ref. [56])

Further proposals in the GEB-MG - Draft 1990 [64] suggest that dimensions of
GGT-nodes are dependent on factors such as the relative magnitude of stress
fields and the amount of tie reinforcement. For instance, where 0'2 is less than 0'1
and there are multiple layers of reinforcement, the width I W3 " ofthe tensile tie may
be as much as 20 % of the length (for slabs 20 % of the span length) or width of
the entire D-region (see Fig. 2.43).

The fundamental aspects should allow the designer to determine the

geometry of the CCT-node for varying reinforcement distribution and anchorage
details (several layers, loops, hooks etc.). The experimental portion of the study
by Bouadi [39, 131] provided information about the behavior and transfer of
forces within the CeT-rtode as well as the ultimate strength. Geometric and
reinforcing details for the test specimens are shown in Fig. 2.44. Test resu Its with
a concrete strength in the range from 2360 to 4680 psi showed crushing of the
concrete struts only for the lower concrete strength specimen. In all
the other
 115

I
w3 = O.21d •

1---J-D-regiOn - - / ~1.D.regiOn i
Figure 2.43: Proposed tie width by GES-MG - Draft 1990 [64]

cases anchorage failure was obtained. The approach sllown in Fig. 2.41 ,to
define the geometry of GGT-nodes for anchored reinforcing bars anchored by
development length behind the node, is based on the test results from Souadi
[39]. In his specimens, the compressive forces and the tensile force in the rein-
forcement bar were increased simultaneously. All specimens experienced post-
yield failures including strut crushing, cover splitting, and gross slippage of rein-
forcement. In order to find the concrete strength efficiency factor for the GGT-
node the specimens with concrete failure are compared in Fig. 2.45. The
statistical data from the comparison in Fig. 2.45 are shown in Table 2.17. Included
in the comparison are the different geometry of the nodes with different reinforc-
ing details (forfurtherinformation aboutthe study see [39]). The limiting concrete
strength in the strut used in the compression was based on an efficiency factor
of v e = 0.8.
116 I Tension Tie

g
soecimen
.
I fc (psi) !Beann
Plate
I Bars I Layout
I
I 23~0 Full 13-;\,5'8.#51~
I l~
L.FO 2470

LHT 2490 I Half 13-#5'8-#61~

LHO 2600
I Half 13 "'- -1~
8 ..
-":l,."t>~

LF/-R 2610

Specimen fc ( psi) I Bearing Bars Layout

Plate
I
Hr=T
I
4860
I
Fun 1 3-#5,8-#61 ~ ~,.:a

HFO-SS I 5005
I
Full 6-#7
Ig
HF0-HS 5015
I
Full
I
6-#7
Ig
H~O-SL I 5025
I
Full
I
6-#7
~
HFO-HL I 5025
I
Full 6 #7
I~ .~'._':'#

Specimen 1'c - 28 day failure load loading plate ["] failure mode

LFT 2360 psi 260 kips 14*12; 12*12 crushing of concrete

LFO 2360 psi 260 kips 14*12; 12*12 bond failure
LHT 2360 psi 240 kips 9"12; 6*12 crushing of concrete
LHO 2360 psi 240 kips 9"12; 6*12 crushing of concrete
LFT 2360 psi 345 kips 14*12; 12"12 crushing of concrete
HFT 4860 psi 538 kips 14*12; 12*12 splitting of the side
HFO-SS 4860 psi 450 kips 14*12; 12*12 splitting of the side
HFO-HS 4860 psi 415 kips 14*12; 12"12 splitting of the side
HFO-SL 4860 psi 470 kips 14*12; 12*12 splitting of the side
HFO-HL 4860 psi 433 kips 14*12; 12"12 splitting of the side
I

Figure 2.44: General information about the tested GGT-nodes (from Ref. [131])
 117

A CCT-node can be analysed by checking the concrete strength after finding the
geometry based on the approaches shown in Fig. 2.41. In order to optimize the
CCT-node both stresses at the C1 and C2 faces should be the same (hydrostatic
stress). The stress atthe strut C2 face depends on the strut width "WC1 " the tie width
" WT3 " and the angle "<I>es". Fig. 2.46 shows the geometric inter-relation of these
factors with various strut angles "<I>es". The relation may be used for dimensioning
the width of the strut or to change the strut angle. The best way to design a CCT-
node is to strive for hydrostatic stress (<11 = <12 = <13 = 1.0 where <1j is the stress on
node side i) which leads to an optimal efficiency. The following equation can be
used to find the optimal solution (see Fig. 2.46):

1.6
--a-- experimentltheory • smaller bearing plate

1.4
• experimentftheory=1.0

>-
I-
0
. •
.,.-.
~
---
1.2
CIJ
-
~
.t:

1.0
c:
CIJ
E
".:
CIJ 0.8
Co
)(
CIJ

0.6

0.4
A1
I .
A2 82 A1·A
specimen

Figure 2.45: Comparison of test results with the concrete efficiency factor of
v e = 0.8
118

x,: Column 1
Mean: Std. Dev.: Std. Error. Variance:
11.173 1·138 1.069 1.019
Minimum: Maximum: Ran e: Sum:
1.338 4.691

Table 2.17: Statistical data from Figl.Jre 2.45 (CCT-node)

2.0
1.8 t:J W3=1.0w2
1.6 • W3=O.5w2
1.4 a W3=2.0w2

1.2
,..- 1.0
- b

~
0.8
0.6
0.4
0.2
0.0
0 20 40 60 80
angle between strut and He = ~s

Figure 2.46: Dependency of the efficiency factor's for CeT - node

 119

2.6.3 en - Node
The CIT-node is an intersection of a concrete compressive strut and two
tensile ties. In steel trusses, as shown in Fig. 2.47, bolts, welds, and gusset plates
are sized to safely transfer load between the members. In contrast, a CIT-node
in a concrete member must rely on anchorage, bond, and other internal force
transfer mechanisms to transfer strut and tie forces. Anchorage is achieved by
providing proper development length orin special circumstances by attaching the
reinforcement to bearing plates or other fixed components.

The following approaches are proposed in order to find the geometrical

constraints for the CIT-node. The definition of the effective width plays an
important factor in the dimensioning process. For the relatively rare case of a
CIT-node with anchor plates, the widths of the plates are given as dependent
constraints which tend to fix the width of the unknown compression strut. The
more practical and generally occurring case is the CIT-node without a bearing
plate. For this case the approach of Fig. 2.48 is proposed in order to define the
effective strut width. Ascan beseen, it issimilartothe approach usedforthe CCT-
node.

The efficiency factorforthe CIT-node was investigated in an experimen-

tal study by Anderson [38]. Table 2.18 shows several parameters of the test
specimens. Two specimens, one with normal strength concrete (HHSR: f'c= 5780
psi) and one with low strength concrete (LHSR), used reduced bearing plate area
(4 in. instead of 10.6 in). Specimen LFAC: f'c= 3920 psi was the only specimen
in this study that was SUbjected to unequal forces in the tension ties. The purpose
of the unequal force was to induce a different compression strut angle into the
specimen. A 30 degree angle from the horizontal was chosen so the force in the
longitudinal steel would be approximately 1.73 times the force in the transverse
reinforcement.
120

.--
t /
- -

Steel Truss Joint Design CTT·Node·Strut-and·Tie Model

Considerations

Strength of Members
Adequacy of Connections

Figure 2.47: Comparison of design rationale used for nodal region of strut- and-
tie- model and joint of steel truss (from Ref. [7])
 121

Tension Tie

I 11 I

oOiIOL-----:~=---Com pression Strut

wc1 = wT2 sin $cs+ wT3 cos epcs

Figure 2.48: Geometrical approach to define the strut width for Cn-node
122

~, 5~ ••

typical Fun WIdth Reduced Bearing Surface for

BearIng Surface SpecImens HHSR and LHSR

Speelmen 1';" 28 day failure load strut width rI anchorage detail; failure mOde
contlning transverse
reinforcement (yes-no)
HFSR..A 7010 psi 127.4klps 10.6" 12 1800 hook; yes none-cap.ofsetup
HFSFl-B 5780 psi 137.51¢s 10.6 ·12 180° hook; yes non~.ofsetUp

HHSA 5780 psi 139 kips 4·12 .1aoo hook; yes none-ca.p. of setup

HFSB 5780 psi 138.1 k~s 10.6 a 12 straight bar. yes gross slip-trans.
HFNC 5780 psi 132.5 kips 10.6 • 12 180° hOOk; no cover splitting
LFSR .3720 psi 117.4 kips 10.8· 12 180° hOok; yes development-trans.
LHSR 3720 psi 130.2l¢s 4 ·12 '80° hOok: yes stnJt crushing
LFNC 3720 psI 117.8 kips 10.6 "12' , BO° hOOk; no cover splitting
LFAC 3920 psi 165.4 kips 10.6 "12 180a hook; yes development-long.

Table 2.18: General information about the tested CTT-nodes (from Ref. [33])
 123

In the tests, general strut failures did not usually occur. The reinforcing
anchorage detail was primarily responsible for limiting the ultimate load. How-
ever, for design purposes the actual efficiency factor for the concrete compres-
sive strength is of interest. Only one specimen (LHSR: flc = 3720 psi) failed by
concrete crushing. The bearing plate stress was 3836 psi. By using a concrete
efficiency factor of 0.8 and by taking into account the smaller bearing plate width
(4") compared to the compression strut width (6.37") the experiment 1 theory-
ratio can be computed:

= C/~ = 184131/{12*4) = 3836 psi

= 3836 psi :::;; 0.8 * 3720 {6.37/4)0.5 = 3754 psi

experiment / theory - ratio = 3836/3754 = 1.02

For this specimen with a concrete strength of 3720 psi a concrete efficiency factor
of 0.8 could be safely used. The efficiencyfactorsforCCT-and CTT-nodes must
produce members in which the critical section will exhibit ductile behavior under
extreme overload. This is done by ensuring that actual failure would occur only
after the reinforcement yields. In order to guaranty ductile behavior, it is
necessary to place a limit on the failure state stress levels in the concrete.

When designing a CTT-node the reinforcement in both ties should yield at the
same time. In orderto 'find the optimum strut angle for a given reinforcement pattern,
the following approach can be used (see Fig. 2.49). It is a geometrical approach and
is based on the compression strut width. Since the compression strut width "w1" is
dependent on the tension tie widths "w2" and "w3" shown in Fig. 2.48, the optimal
concrete efficiency for a CTT-node is given by the angle with the largest compression
strut width "w1". The compression strut width can be computed:

w1 = w2 sin <Pes + w3 cos <Pes

Fig. 2.49 shows the compression strut width "w1" for various strut angle" <Pes "and
three different tension tie width ratios.
124

-...
3: 2.0
1.8
/" • .~
3:
0 1.6
~
..
ca 1.4 •
...
oJ:
"'C
1.2
"3:
....:::s 1.0

... 0.8

•
C/)
0.6 W3 = 1.0w2
C
0
"enC/) 0.4 •• W3 =0.5w2
W3 = 1.5w2
..
Q)

Co
0.2
E
0
0.0
u 0 20 40 60 80
angle between strut and tie

Figure 2.49: Dependency of the compression strut width for eTT-node

2.6.4 TIT-nodes

In those rare cases where the tensile strength is used as a tension tie, some
global understanding about tensile strength has to be formulated.

Although is it difficult to develop design criteria forthe case of concrete tensile

ties, it would be even worse to maintain the formalistic view that the tensile strength
of concrete cannot and therefore must not be utilized. Tracing the flow of forces in
actual structures, to be gap free and consistent with strut- and- tie- models will
sometimes show that equilibrium can only be satisfied if ties or tensile force can be
accepted in places where, for practical reasons, reinforcement cannot be provided
and the tensile strength of concrete is implicitly utilized (see Fig. 2.50).
 125

tensile strength utilized

VAI-------,Anchor bolts---~

Tensile force Tensile force

Figure: 2.50: Tensile strength of concrete implicitly utilized

The tensile strength of concrete is relatively low, about 5 to 15% of the

compression strength. The tensile strength is more difficult to measure and the
variance is greater than for compressive strength. For the biaxial- and triaxial
range the tensile strength is assumed to be equal to the uniaxial tensile strength
[24, 69]. While concrete tensile strength may playa part in the force transfer
mechanism, it is generally more convenient to neglect its contribution. This is
prudent for design purposes as the tensi Ie strength of the concrete is very small
relative to that of reinforcement. Also, the action of creep, shrinkage, external and
internal thermal stresses and other load patterns may cause cracking which
would inhibit the development of concrete ties. For most practical detailing
problems, concrete tensi Ie strength may be ignored. For those cases where the
tensile strength is needed, a value of

fct = 3 (fc)0.5
can be used.
126

Ifthetensile forces are transferred with reinforcing bars, the anchorage re-
quirements became important. Anchorage is achieved by providing proper
devolopment length or in special circumstances by attaching the reinforcement
to bearing plates or other fixed components. The key to determining anchorage
requirements is selecting the point at which the reinforcement must be fully
developed. When the ties at a node are to be fully developed, a conservative
approach is to assume that the development length for each layer of tie
reinforcement is assumed to begin at the intersection point of the different ties
with the confined joint boundaries (see Fig. 2.51).

~
T
a

Figure 2.51: Conservative starting point for computing development length

Because the behavior of the tension controlled nodes (CTI- and TTT-
node) is influenced by the tie anchorage details, it is appropriate to make a
distinction between anchorage details that may be chosen. In Fig. 2.52, tie
anchorages have been separated by type.
 127

Positive anchorage details are those which do not rely appreciably upon
bond stresses to resist the applied tensile force and include end plates and
continuous reinforcing details. The positive anchorage detail must be designed
so that the tie force is distributed over a sufficient area to prevent the node zone
from being overstressed in compression. End plates and continuous reinforce-
ment details are attractive from a design standpoint because they are fai'rly easy
to evaluate.

Development anchorage details are those which are anchored with bent
bars (hooks), bond strength or a combination of both. Development anchorage
details are normally more economical, easier to fabricate and to place in the
formwork. The disadvantage of the development anchorage details is thelonger
required anchorage length.

Figure 2.52: Positive and development length anchorage details

128

2.6.5 Anchorage Requirements in the nodal zone

All nodal zones are influenced by the tension tie anchorage details. If the
applied tensile force is connected to bearing plates and does not rely appreciably
upon bond stresses, then the tie actually provides a compression strut in terms
of its action on the nodal zone.
Barton [7] and Anderson [38] call anchorages with end plates or
continuous looped reinforcement "positive anchorage details." The positive an-
chorage must be designed so that the compression resulting from the tie force
is distributed over a sufficient area to prevent the node from being overstressed.
However, such positive anchorages are not necessarily required nor are they
always desirable or practical construction alternatives for anchoring tensile ties.
Except for small diameter reinforcement, positive anchorage details are more
expensive and more difficult to construct than standard details such as straight
bars or hooks. Where the transfer of strut- and tie- forces is felt to be so abrupt
that sufficient bond anchorage forces cannot be developed, end plates or
continuous reinforcement details should be provided. However, the designer
may often choose to not use positive anchorage details if there is some flexibility
in placing the tie reinforcement. For details with straight bars and hooks the
designer must check the development length requirements of the tensile tie
rei nforcement. Sufficient development length should prevent splitting of the con-
crete cover and the resulting anchorage failure. Tepfers [97] suggested the
following approach to prevent splitting ofthe concrete cover for short anchorages
without transverse reinforcement. The cracking resistance, f'be' lies between the
following limits, suggested by Tepfers:

l'be = fer (c + db/ 2) / (1.664 db tan co) elasto-cracked model

f' be = fer (2c) / (db tan co) plastic model

f' be = cracking resistance

f' be = tensile strength
C = concrete cover
db = diameter of reinforcing bar
co = average angle between the transverse cracks and the axis of the
bar ::= 45°
 129

Bar development length "I d" is the necessary embedment to assure that a bar
can be stressed to its yield point with some reserve to ensure membertoughness
under specific containment conditions. The necessary length is a function of a
number of variables, mainly of the bond strength and confinement from both
concrete cover and transverse reinforcement.. A great amount of research work
has been done in the area of development length (Tepfers, R. [98], Jirsa, J; Lutz,
L.; Gergely, P. [99], Orangun, C.; Jirsa, J.; Breen, J. [100]). The radial stress in
the concrete surrounding a bar being developed can be regarded as a water
pressure acting against a thick - walled cylinder having an inner diameter equal
to the bar diameter and a thickness "c" (the sma.ller of the clear bottom cover cb
or half the clear spaci ng Cs to the next adjacent bar). Based on a comprehensive
review of a broad range of test results, the following equation forthe development
length ( Id ) in terms of the stress in the bar at the critical section ( fs ), the bar
diameter ( db)' concrete strength (f c), cover ( c ) to diameter ratio, and transverse
reinforcement amount ( A tr ), yield strength ( fyt ) and spacing ( s ) were proposed
by Orangun, Jirsa and Breen [100].

=
A modified form of this equation in terms of a series of modifiers is the basis for
the recent changes in splice and development length design provisions in ACI
318-89 [101].

In CCT - and CTT - nodes the reinforcing bars are under lateral pressure
from the compressive struts (see Fig. 2.53). When lateral pressure is applied the
vertical component of the radial pressure tends to be balanced by the lateral
pressure. The bond strength increases approximately in proportion to the square
root of the lateral pressure. In addition, the distance between the bearing plate
and the reinforcement bar, e, has an important effect as shown in the study by
Lormanometee [102]. Different experimental studies were evaluated to develop
a formulation for a possible reduction of the development length for a reinforce-
ment bar under lateral pressure. Only tests in which failure occurred before the
bars yielded were included. The lateral pressure acts similar to the action of
transverse reinforcement. The overall strength of a splice with transverse
reinforcement and lateral pressure can be expressed as follows:

u =
130

I vertical cracking plane

I

~""""""I,.LLL.J"""""'''''''''''',] lateral pressure

horizontal cracking plane

lateral pressure

I
I
Figure 2.53: Lateral pressure on reinforcement bar

u = (f'C)0.5 (1.2 + 3 c / db + 50 db / 's )

utr = (f'c)0.5 [A tr fyl / (500 s db)]
ulP = (f'c)0.5 [( fn )0.5 (200 - e2 ) /1000]

The development length can then be computed:

{ 1.2 + 3 c / db + (Atr fYI ) / (500 S db )+ [(fn)0.5 (200 - e2 ) /1000 ]}

(Atr fyl ) (500 s db)

/ :::; 3.0
[(fn)0.5 (200 - e2) /1000] :::; 6.0 (see Fig. 2.54)

The comparison with test results from Lormanometee [102] and Schmidt-Thro,
Stockl and Kupfer [1 03] are shown in Fig. 2.55 and the statistical data is shown
in Table 2.18. The proposed relationship is conservative for all except one of the
test results and is generally quite conservative. A multiplying factor of 1.25 is
required to make the results consistent with the current ACI and AASHTO
expressions which indirectly introduce a cj> factor as 1 / cj> = 1 /0.8 = 1.25.
 131

I Lateral presssure e

Figure 2.54: Lateral pressure and the distance "e" to the reinforcing bar

Development length anchorage details include straight and hooked bars.

Forthese details the designer must check the development length requirements
of the tension tie reinforcement. For CCT-nodes confining reinforcement had
only a low effect (===2%). Similarly, forthe CIT-node in which the transverse re-
inforcement anchorage hooks were turned nearly parallel to the longitudinal bars
(but not closed), the ultimate load decreased by only a maximum of 4% in com-
parison with closed confined reinforcement.
By using hooks instead of long bars for the anchorage, the ultimate load de-
creased by 8% for CCT (specimen C2 and 02) and for CIT (specimen HFSR-
A) nodes. Using a transverse U forthe second tie in CIT-nodes provided lateral
confinement, but prying action at the 90 0 bend can produce splitting cracks. In
order to control splitting cracks of the end cover it is suggested that the
longitudinal reinforcement be extended a short distance (===s/2 or 2 in.) past the
transverse reinforcement.

Table 2.19: Statistical data from comparison in Figure 2.55

X1: Column 1
Mean: Std. Dev.: Std. Error: Variance: Coet. Var.:

F3;.;..2..;.;75'----r-1·9;.;;.;6;..;;7~'---I_·1_9 _JIL...·9_35 ___'_I29_.532 .L.-. _

Miniroom: Maxiroom:
.91 5.17 85.14
 0
0
0
('I)

0,
132 or-
II
~ ~
0 0
Q) Q)

-
.c.
~
c
Q)
.c.
Sc
Q)
E E
'i:: 'i::
Q) Q)
0. 0.
x
Q)
x
Q)
0
0
0
C\J .....
-
t t U)

......
Q.

...:::s
(1)

U)
U)

...
(1)

Q.

-...as
...as
(1)

0
0
0
or-

I-........,r----r-......,.-......,.---r---r---r---r------,------+ 0
o

NoelU / ~uew!Jedxe
Aq 4lDueJlS u! 4~DUeJ~S

Figure 2.55: Comparison of a theoretical approach and test results for the
development length of straight bars with confinement from bearing
plates
 133
2.7 Model optimization

Since strut- and- tie models are lower bound solutions for the actual load
carrying capacity of a structure, any correctly formulated and correctly detailed
strut- and- tie model should safely carry the design loads applied. In many cases
different models can be developed for the same external load configuration.
Doubts could arise as to whether the most efficient model has been chosen. In
selecting the model, it is helpful to realize that loads try to use the path with the
least forces and deformations. This simple criterion for optimizing a model may
be formulated as follows:

I F. I. c . = minimum
I I ml

Fj = force in strut or tie i

Ij = length of member i
cmi = mean strain of member i

This equation is derived 'from the principle of minimum strain energy for linear
elastic behavior of the struts and ties after cracking. The contribution of the
concrete struts can generally be omitted because the strains of the struts are
usually much smaller than those ofthe steel ties. Since reinforcing ties are much
more deformable than concrete struts, Schlaich et al [28] propose that the model
with the least and shortest ties is the best.

As a more general approach for model development the following consid-

erations are important constraints:

ease of fabrication
equilibrium
ductility
serviceability

In many cases, practicality and ease of fabrication will have the greatest influence
upon the configuration ofthe design model. Models which result in details that are
overly congested or difficult to fabricate should be avoided. The reinforcement
134
pattern for the D region should be compatible with the reinforcement scheme
used in adjacent portions of the structure. In order to satisfy the requirements of
the theory of plasticity, a model must be in equilibrium under the applied loads.
However, if the selected strut and- tie- model is to fully develop, the load carrying
capacity of the strut- and- tie- elements and the rotational capacity of the nodes
must not be exceeded before the ties yield. Furthermore, acceptable servicea-
bilityat usual working load levels requires that crack widths be limited by provision
of sufficient, closely spaced reinforcement in regions of high tension and hence
cracking. Attention must be paid to elastic analysis predictions of high tension
zones to ensure crack control reinforcement is appropriate. In addition to the
accepted standardsfortlexural reinforcement distribution and both minimum and
maximum bar spacings, minimum reinforcement to control shrinkage, creep and
thermal stresses should be provided.

A more sophisticated optimization process would recognize that the

fabrication and placement costs of the local confining reinforcement and addi-
tional anchorage at the nodes is substantially higherthan the costs of longitudinal
reinforcement. A more realistic approach would provide for different unit costs
for the major classifications of reinforcement. The following approach can be
used:

II = length of the longitudinal reinforcement

It = length of the transversal reinforcement
Id = development length
C1 = cost per unit length of the longitudinal reinforcement
Ct = cost per unit length of the transversal reinforcement
::= 1.2 times the cost of G,
eN = cost per unit length of the node reinforcement
::= 1.5 times the cost of G,

= GGG-node = 0.0
GGT-node = 1.0
CD-node = 2.0
TIT-node 3.0 (reinforced TIT-node)
 135
In order to prevent extreme strut angles (which may result in excessive
cracking), the angles between compression struts and tension ties should be
limited to between 25 and 65 degrees. In summary, some guidelines for model
optimization have been proposed. The designer should taken into account
practical considerations in combination with the proposed strut- and- tie- model
principles in the development of a suitable model.

2.8 Concrete efficiency factors for design

2.8.1. Unconfined Nodes and Undisturbed Concrete Struts

The effective concrete strength in the various compression fields or struts

is less than the concrete cylinder strength. General agreement between the
theory developed forthe strut- and- tie- models and test results is only obtained
if a concrete efficiency factor is introduced to limit the concrete capacities in
nodes and struts. When compared with a large number ottest results (Fig. 2.23)
the following functions gave acceptable results.

= 0.5 + 15 / (flc)ooS
= 0.5 + 20 / (f'c)ooS

As a design simplification, a linear relationship falling between these two

functions was chosen as shown in Fig. 2.56. The basic efficiency factor should
be taken as 0.8 for concrete compressive strengths up to and including 4000 psi.
For strengths above 4000 psi, the efficiency factor should be reduced continu-
ouslyat a rate of 0.05 for each 2000 psi of strength in excess of 4000 psi, but
the efficiency factor should not be taken less than 0.65.

= ve l'c

= 0.8 for f'c ~ 4000 psi

= 0.9 - 0.25 fIe / 10000 for 4000 < fIe < 10000 psi

= 0.65 for f'c ~ 10000 psi

136

This basic efficiency factor can be usedforchecking compressive fields and short
struts within unconfined or !lightly confined nodes. as well as applications where
the compressive struts act over undisturbed or uncrushed concrete as occurs in
many wall type applications where no tensile cracking is expected.

2.8.2 Compression Diagonals

As shown in 2.4.1.1, the effective concrete strength in the compression diagonals

must be reduced with an additional factor of 0.6. This is particularly important in
thin web members and cores where fairly wide cracks must be crossed by tile
struts.

= a.6ve

There are many reasons why the efficiency factor for compressive diagonals is
less than the global efficiency factor for unconfined nodes and undisturbed
compression fields. The web strength might depend somewhat on

the stirrup spacing in the longitudinal direction and the resultant control of inclined
web cracking. In addition the effective strength of the web may be reduced
because of cracks developed in early loading stages and having directions other
than that of the final cracks [21,28,62,65,66]. Finally, in beams and girders the
compression zones are highly concentrated and the struts in the web concrete
have a corresponding concentraion of load which may lead to more local failure
of the concrete at a stress level which as an average over the web is less then
the effective compression strength in more uniform compression fields.

2.8.3. Confined Nodes

In certain applications such as post-tensioned anchorage zones, a very

large amount of local confinement may be desirable to allow the safe develop-
ment of very high compression stresses in a local zone node. In such cases the
confined node effective compressive stress is designated as fe3 and may be
determined as:
 137
(a) For nodes confined with closely spaced spiral reinforcement:
f = v f'c (AlAb)o.s + 4.0 (Accre 1 A b ) flat (1 - s/d)2 ~ 2.5 fIe
es e

The term "(1 - s/d)2 reflects the reduction in effectiveness of spiral associated
11

with increasing spacing of the spiral wires.

(b) For nodes confined with orthogonal reinforcement such as closed square
hoops and with longitudinal reinforcement to anchor the corners of the hoops:
f = V fIe (AlAb)o.s + 2.0 (Acore 1 A b ) flat (1 - s/d)2 ~ 2.5 f'c
es e

(c) For nodes confined with orthogonal reinforcement such as closed square
hoops but without longitudinal reinforcement anchoring the corners ofthe hoops:
f
as
=va f'c (AlAb)o.s + 1.0 (Acore 1 A b ) flat (1 - s/d)2 ~ 2.5 fIe

In (a), (b), and (c) the following limits apply:

AlA b ~ 4
1 ~ Accre 1 A b ~ 3

In (a), (b), and (c), the symbols are de'fined as:

va = basic efficiency factor as defined in Section 2.8.1

f' c = concrete compressive strength
A = area of confined concrete concentric with and geometri-
cally similar to the bearing plate
Ab = effective area of the bearing plate
Accra = area of the confined strut
As = cross sectional area of the con'fining reinforcement
flat = 2 As fs 1 (s d )
fy = yield strength of confining reinforcement
d = diameter of confined core
s = pitch or spacing of confining reinforcement
fs = design stress in con·fining reinforcement
= fy for fIe ~ 7000 psi
= C !.1 25 1 (1t d As ) ~ fy for f' c > 7000 psi
!.1 = Poisson ratio ( :::: 116 or 0.17 for f'c up to 10000 psi)
C = compression loads
138

o
8C\I

1---...,....----;---.,....---+----,,.......--+-----,.--+0
co co
c:i c:i

effi ci ency factor

Figure 2.56: Design approach for concrete efficiency factor

 139

2.9 Anchorage requirements for design

Generally the development lengths for straight bars and for hooks should be
taken as recommended in AC1318-89 considering such effect as concrete cover, bar
spacing, and transverse confining reinforcement. Since the ACI 318-89 provisions
neglect the often beneficial effect of local bearing pressures such as occur at regions
where direct loads are applied or direct supports are provided, such local confine-
ment can be considered if for design purposes the development length for straight
bars is computed as:

{ 1.2 + 3 c / db + (Atr fyt ) / (500 S db )+ [(fn)O.5 (200 - e 2 ) /1000 ]}

3.0

o 6.0

The proposed formula take the lateral pressure into account whenever the
distance between the closest bar surface and bearing plate, e, is 14 in. or less..
140
 CHAPTER 3. PROCEDURES

3.1 General Analysis - Structural Analysis

Dimensioning is in principle an iterative process. Figure 3.1 shows the general

procedure for designing and dimensioning concrete structures. For many
conventional structures and even for large numbers of regions in unusual
structures, the strain profiles will be linear and many of the regions will be B-
regions. It will ordinarily be simpler and quicker to dimension the B-regions with
conventional sectional analysis and design procedures as given in ACl[4] and
AASHTO[3] standards. However, in D regions strut- and- tie models should be
used. Steps for detailing using the strut- and- tie models are also indicated in Fig.
3.1 in an iterative loop.

From the flow of forces an appropriate strut- and- tie model is chosen and loaded
with the applied forces and boundary forces. After computing the strut and tie
forces the compression struts, tension ties and the nodal zones have to be
dimensioned. There is a close relation between the detailing of the struts bearing
on the node, of the ties anchored in the node, and the node itself, because the
detail of the node chosen by the design engineer affects the flow of forces. This
method implies that the structure is designed according to the lower bound
theorem of plasticity. Since concrete permits only limited plastic deformations,
the strut- and- tie- model has to be chosen in a way that the deformation limit is
not exceeded at any point before the assumed state of stress is reached in the
rest of the structure. This is especially important for the main members, which
carry a significant portion of the load. According to Schlaich et al. [28] it is
desirable that the struts and ties follow the elastic flow paths closely with a
deviation of at most 15° from the elastic principal stress directions. The proposed
design recommendations are applicable to either prestressed or conventionally
reinforced concrete members. The general assumptions for the application of the
strut- and - tie- model in the design procedure are:
Yielding of the reinforcement is required prior to concrete - or anchorage
failure
the ties transfer only unixial forces and neglect dowel action, aggregate
interlock, tensile strength across cracks etc.
141
142

General Structural System

Determine Loads including prestressing forces

Estimate Members Sizes and Dimensions

Divide the Members into B- and D-regions

Dimension B-regions with Sectional Analysis

Develop Strut- and- Tie- Model for D-region

Compute Strut- and- Tie- Forces

Dimension Reinforcement for Ties

Check Concrete Stresses at Node Zones

Determine Tie Anchorage Requirements

Optimize Strut- and- Tie- Model

Check Serviceability Control under Working Loads

Figure 3.1: Design Procedure for Concrete Structures

 143

For the D-region it is of considerable value to find out how much

reinforcement is neededand where it should be placed. The most outstanding
contribution of strut- and- tie modelling is the ease and rapidity of determining
reinforcement requirements. Checking of nodes and struts is far more complex
and involved. For the highly loaded members the strut- and- tie- model should
follow the elastic load path rather closely. For low loaded D-regions a larger
deviation than 15 degrees from the elastic prinicipal stress directions is permis-
sible (max 45°). Also the angle between struts and ties entering a singular node
should not be too small in orderto prevent skew cracks (~ 15° and preferably 25°)
and diagonal crushing of the concrete prior to yielding of the reinforcement.

3.2 Checking and Dimensioning Concrete Compression Struts

The struts in the model are resultants of the compresion stress fields. The
path of the compressive forces may be visualized as the flow of compressive
stresses with varying sections perpendicular to the force path direction [104]. As
the strut- and- tie- model is an idealization of the real structure, the struts are
assumed as straight and concentrated at the nodes. The straight line of a com-
pression strut can be re'fined for higher stressed struts and some possible tensi Ie
forces can be counted as shown in Fig. 3.2.

••• itA";) 1-'

~
\I)V

•••
J
1/ ? I' ~
I
•• h I

••• ,
1\, 'I
IJ
,
• ~,
~ "'~
Figure 3.2: Compression fields and strut- and- tie- model
144

In orderto be consistent with the factored load design methods of AASHTO and
ACI, load factors must be incorporated into the force calculations and 0 factors
must be incorporated into the resistance calculations. Forconcrete compression
struts 0 factors as used in concrete column design seem most appropriate.
o = concrete strength reduction factor
= 0.75 members with spiral reinforcement conforming to
Sec. 10.9.3 (ACI 318-89)
= 0.7 for other reinforced members (ACI 318-89)
f' e = concrete compressive strength based on standard 6-in. x
12-in. cylinders at 28 days age

For dimensioning purposes the following approaches are proposed:

a) Compression fields, like fan, bottle, prism (Fig. 2.6a, b, c):

cr ~ 0 fee =0
Ve fIe
Ve = concrete efficiency factor
Ve = 0.8 for fIe ~ 4000 psi
Ve = 0.9 - 0.25 fIe /10000 for 4000 < fIe < 10000 psi
Ve = 0.65 for fIe ~ 10000 psi

b) Compression diagonal struts

cr ~ 0 0.6 fee = 0 0.6 VefIe

c) Confined Compression fields (Fig. 2.6d) /

cr ~ 0 flece
flece = [(Ve fIe (AlAb)o.s + a (Accre / Ab ) flat (1 - s/d)2)] ~ 2.5 fIe

= 4.0 for spiral confinement

= 2.0 for square closed hoop confinement anchored with
longitudinal reinforcement
= 1.0 for square closed hoop confinement without longitudi
nal reinforcement anchorage
flat = lateral pressure = 2 fyAJ (d s) for f'e ~ 7000 psi
=2fs As /(ds) for fIe;::: 7000 psi
 145

fs = C Il 2 s 1 (1t d As) :::; fy

C = compression load
Il = poisson ratio (::::: 1/6 or 0.17 up to 10,000 psi)
A = area of confined concrete concentric with and geometrically similar
to the bearing plate
~ = area of the bearing plate
Acore = area of the confined strut
A/Ab :::; 4
1:::;Acore IA:::;3
''b

In determining the spread or diffusion of concentrated forces it is necessary to

assume or define a strut diffusion angle, the angle witl1 respect to the strut axis
at which the com pression force spreads out from the edge of a bearing plate. For
the strut diffusion angle for heavily loaded members or under bearing plates
(anchorage zone) the following proposals are given. MacGregor [62] proposes
a diffusion angle of 15 degrees. An elastic 'finite element analysis by Burdet [42]
shows for various ratios of bearing plate width to compression field width from
0.1 to 0.9, diffusion angles vary between 27 to 22 degrees. The experimental
study by Sanders [43] gave somewhat lower values. Figure 3.3 show the various
approaches with the proposed equation (see Fig. 3.4)

diffusion angle [deg] = 12 + 31 (alh)O.5

a = bearing plate width (see Fig. 3.2)

h = compression field width (see Fig. 3.2)

In case of doubt, or if the detail being considered is especially critical, larger

diffusion angles can be used. This will lead to larger values of the tension force
in the ties. The location of the elastic resultant can be estimated with sufficient
accuracy for design by using an apparent diffusion angle of "12 +31 (alh)o.5 ". The
solution with diffusion angle of 26.5 0
(slope: 1:2) proposed by Burdet [42] will
lead to conservative answers when compared with the elastic finite element
analysis for cases having ratios a/h larger or equal to 0.15. If no other information
about the alI1 ratio is available, the diffusion angle can be estimated for design
by using an apparent diffusion angle of 21.8 0 (slope 2 : 5).
146

30

-
Cl
25
.
- ~
... m experiment
Gl
~
.!!!
Cl
c
CI:I
20
-- ---
.:. .:,

\ ~ ~
•
a
0
angle = 15°
FEM
proposal
15
~~ ~
~
~

10 I I ,
0.1 0.2 0.3 0.4 0.5
bearing plate width I compression field width

Figure 3.3: Comparison of various diffusion angle

24
-Cl
Gl 22 1:::1 12+3/...J(87ii}
"..... .
.!!!
...Cl~
c
0
20

18
,
" ~
~
-
U;
:::l
:=
"-
16
. .,..
-.
:::l
14
0

12 I I

0.0 0.2 0.4 0.6 0.8 1.0

bearing plate width I compression field width

Figure 3.4: Proposed diffusion angle for design

 147

3.3 Checking and Dimensioning Tensile Ties

The reinforced concrete ties are essentially linear or one-dimensional

elements between two nodes. All of the tensile force has to be transfered by re-
inforcing ties. From the known tensile forces found by equilibrium at the node
regions, the dimensions ofthe reinforcement can be computed. In orderto select
normal spaces and reinforcing bar diameters it is necessary to determine the
effective width of the reinforcing tie (Fig. 3.5).

Figure 3.5: Width of the reinforcing tie

The di mensioning is a check against the yield strength of the rei nforci ng bars and
prestressed tendons

fy = yield strength of the reinforcing bar

D.fpy = availiable part (non-prestressed portion) ofthe yield strength ofthe
prestressed steel
A5 = area of the reinforcing bar
Ap = area of the prestressing tendons
<1> = 0.9 for steel tensile ties and 0.65 for concrete used in tension
148

Prestress forces are to be applied to the strut- and- tie model as external loads
(external force pair) with friction forces in the transmission zone (Fig. 2.31) in the
analysis and dimensioning. Only the available remainder of the yield strength
above the effective prestress force can be used for carrying tensile forces from
the strut and tie model (internal forces) (see Table 2.16).
After selecting the required spacing for the reinforcing bars and prestressing
tendons the width of the tie is determined as the outside dimension of the rein-
forcement layers. The width is necessary for dimensioning the node regions.
For those instances when it is desirable and permissible to count on the tensile
strength of the concrete to carry equilibrium forces where no progressive failure
seems possible, the following approach can be applied (width of the tension tie
assumed as 1 in.)

3..JfC h
= depth of the tension tie

If the tensile stress field is crossed by a compression field, the reduced biaxial
strength shown in Fig. 3.6 must be considered.

:Me ~[I :::::::::-::::==---_

0.25 f'c 1.0 f'c

Figure 3.6: Assumption for the biaxial compressive-tensile strength

The maximum angle "a" between the compression and the tensile field is
arc tan a = ;ffC/ 24

Forthe resulting tensile - compression forces a parallel bounding failure curve

can be assumed.
 149

3.4 Checking and Dimensioning Nodes: Determining Anchorage

Requirements

D-regions usually contain either smeared or singular nodes. The singu-

lar nodes are more critical and need more attention. The following dimensioning
and anchorage requirements must be applied to either smeared or singular
nodes. The stress peaks in smeared nodes are less critical because a greater
amount of surrounding concrete is normally available. For multiple, widespread
reinforcement layers it is difficult to choose the node width (Fig. 3.7).

Ca/2

r-
! aI2 --1
1llllllllllj'tllltf

1---w2/2-- ~-w2/2----l

w1 = w2 cos 4>1

Figure 3.7: Strut width for smeared node

150

Forthe concrete efficiency factors and the anchorage requirements in smeared

nodes the rules for singular nodes (CCC, cn, CCT, Tn) should be applied.
3.4.1 Checking and Dimensioning CCC· nodes

The following factored load stress limits are proposed (see Section 2.8)

(j ~ " flee
fee = Ve <
f' c
- 2.5 f' c
Ve = concrete efficiency factor
(3
= concrete strength reduction factor
= 0.75 members with spiral reinforcement conforming to
Sec. 10.9.3 (ACI 318-89)
= 0.7 for other reinforced members (ACI 318-89)
f' c = concrete compressive strength based on standard 6-in. x
12-in. cylinders at 28 days age

a) Unconfined Nodes without bearing plates

ve = 0.8 for fIe ~ 4000 psi
Ve = 0.9 - 0.25 fIe 11 0000 for 4000 < f'c < 10000 psi
Ve = 0.65 for f'c ~ 10000 psi

b) Confined nodes
(j ~ (3 flcce
f'cce = [(Ve fIe (A1A b)o.s + a (Accre 1 Ab ) flat (1 - s/d)2)] ~ 25
. f' c

= 4.0 for spiral confinement

= 2.0 for square closed hoop confinement anchored with
longitudinal reinforcement
= 1.0 for square closed hoop confinement without longitudinal
reinforcement anchorage
lateral pressure = 2 fyAJ (d s) for fIe ~ 7000 psi
= 2 fsAsl (d s) for fIe> 7000 psi
 151
c) Unconfined nodes with bearing plates
(j ~ <j> f be
fbe = v e f'c (AlAb)o.s ~ 2.5 f'c
AlA b ~ 4
d) Triaxially confined nodes
The increase in strength due to three-dimensional states of compressive
stresses may be taken into account if the simultaneously acting transverse com-
pressive stresses are considered reliable. This may be particularily appropriate
if supplementary transverse prestressing is applied.
When threedimensional compressive strength is appropriate
(j ~ <j> fc3
fc3 ~ 2.5 f'c
The dimensioning for CCC-nodes based on the proposed strut width ap-
proaches given in Section 2.6.1 can be determined as shown in Fig. 3.8. For a
more general application with borders ofthe compression strut assumed to be not
parallel, the strut width "w3" at a certain distance "y" from the concrete surface
can be computed by using the proposed geometrical approaches shown in Fig.
3.9.
In a CCT-node where the tension reinforcement is welded or bolted to the
anchor plate, the stress configuration in the nodes is then similarto those in CCC-
nodes. A smooth surface of the tie where it crosses the node is better than hig h
bond from deformations because strain compatibility with the bonded bar will
tend to crack the node's concrete. The proposed CCC-node strength can be
applied forthis kind of load transfer. In addition the anchor plate has to be checked
for bending strength and the welding connection with the tie must also be
checked.

3.4.2 Checking and Dimensioning CCl· nodes

For CCT-nodes the width of the strut can be found by considering

geometrical constraints such as bearing plates and by assuming that the
effective width of the tensile tie is governed by the dimensions from the inside to
the outside reinforcement layer (wT). With a single layer of reinforcement wT is
taken as the bar diameter as shown in Figs. 3.10 and 3.11. With multiple layers
of reinforcement wT is taken as shown in Fig. 3.12.
152

C1 = C2 * cos ~2 + C3 * cos cj>3

C1 = C1' + C1"
C1' 1a' = C1" 1a"

=
C4 C3 * cos cj>3
C4 = C2 * cos cj>2

a (tan cj>2 * tan cj>3)

aO = - - - - - - - - -
tan ~2 + tan 413

w4 =a 12
If w4 ~ aD then hydrostatic stress
CO ~ v * fc * b * aD

wC2 =an cos cj>2 + a 12 sin c!l2

wc3 =a' cos ~3 + a 12 sin 4>3

Figure 3.8: Geometrical relation for GeG-node for dimensioning process

 153
a'

I"

T -i
1 :
w4
.' " '
e"oI aO

~j

1
I
1
1
1
.·1

w4 =a /2
wc3=w3 +2w3' for w3' S w3"
vic3 = w3 + 2 w3" for w3' ~ w3"

w3 = a / 2 sin 4>3 + a' cos 4>3

tan (4)3' • <1>3)

2
w3'= [4 y. a (sin <1>3) • a' sin 2<1>3]
4 cos <1>3

ex = arctan [a / (2a')]

tan (4)3 • <1>3")

w3" = ( 8y • [tan 4>3 • 2tan (4)3· a)] * [a sin 24>3 + 4 a (cos 4>3)2 ])
8 cos <1>3

Figure 3.9: Geometrical relation for eee-node with borders not parallel to
the compression strut
154

The concrete stresses should be checked:

01 ;5; '" fee (AIAb)o.s ;5; 2.5 f'c
02;5; '" fee (AIAb)o.s ;5; 2.5 f'c
AlAb ;5; 4

= v e f' c
= 0.7

The efficiency factorforthe concrete compression strength should be taken as:

ve = 0.8 for flc;5; 4000 psi
Ve = 0.9 - 0.25 fIe /1 0000 for 4000 < fie < 10000 psi
V
e
= 0.65 for fie;;::: 10000 psi

•
wT= db

w2 =w1 sin cP + wT cos cP

Figure 3.10: CeT-node with single straight reinforcement bar

 155

The nodal zone must also fulfill the requirement for minimum develop-
ment length, concrete cover and bar spacing limits. Test results [39] show that
vertically oriented hooks decrease the ultimate load of the CCT-nodes by 4 to 8%
as compared to straight bars with full development lengths. This decrease is
probably not significant given the other uncertainties in the deign process. The
advantage of hooks is that the required anchorage length can be minimized (Fig.
3.11). With multiple layers of reinforcement, the available Id can be taken from
the intersection of each bar layer with the nodal zones (Fig. 3.12).

+
l--/2 C1
·1- w1/2

-'d
I· a .1
wT= db

w2 = w1 sin <I> + wT cos <I>

Figure 3.11: CCT-node with hooked bar

156

114---.. -a
·1
wT = n db + (n-1) s
n = number 01 reinforcing bar layers
s = clear bar spacing
w2 = w1 sin 4>+ wTcos 4>

Figure 3.12: CCT-node with mUltiple reinforcement bar layers

 157

3.4.3 Checking and Dimensioning cn - nodes

The dimensioning process for the CTT- node is similar to the proposed
approach for the CCT-node. The strut width can be computed from the geomet-
rical boundaries or widths of the tension ties (see Figs. 3.13 and 3.14).

For CTT-nodes the concrete compression efficiency factor should be

taken as:
0'1 :::; o flee (AIAb)o.5 :::; 2.5 f'e
AlAb :::; 4

= Ve f' e
= 0.7

The efficiency factor for the concrete compression strength should be taken as:
ve = 0.8 for fIe:::; 4000 psi
V
e
= 0.9 - 0.25 f'e / 10000 for 4000 < fIe < 10000 psi
V
e = 0.65 for fIe ~ 10000 psi

Test results [38] shown that the outside layers of reinforcement close to
a surface of the member are the most critical. Major cracks which initiate at the
surface and generally follow the theoretical strut angle decrease the bond
strength. Reinforcement should be provided across all planes of weakness to
control cracking. Confining reinforcement normal to planes of hooks and bends
is especially important [38]. Fig. 3.13 show singular tensile ties and Fig. 3.14
show multiple tensile ties for crr - nodes.
158

d'
~r-
t
T1

w2_~
.--
1-..
-----'W1 -----.I

w1 = n' d~ + (n' -1) s'

w2 = d~ or conservatively: w2 =0
n' = number of reinforcing bar layers
s' = clear bar spacing

wc = w1 sin <p + w2 cos cP

Figure 3.13: eTT-node with single reinforcement bar layer

 159

d'
~r-
t
11

~ wc/2

"><C2
s'

"" '>
wc/2

1-------'W1 ----~
w1 =n' d~ + (n'-1) s'

w2 =n" d~' + (n" -1) s"

n = number of reinforcing bar layers
s = clear bar spacing
wc = w1 sin <I> + w2 cos Q>

Figure 3.14: Cn-node with mUltiple reinforcement bar layers

160

3.4.4 Checking and Dimensioning TTT • nodes

For TIT-nodes the anchorage requirements have to be checked. It must

be evident that satisfactory behavior and adequate strength can be attained only
by the efficient interaction of concrete and steel. In details where the length
available for end anchorage is very short, special devices, such as illustrated in
Fig. 3.15, may be required to ensure the development of the reinforcing bar
strength. For TIT-nodes the largest tensile tie should be anchored with looped
- or hooked bars. This is illustrated in Fig. 3.16.

Figure 3.15: Special anchorage devices (from Ref. [105])

 161

III 1

III

t
Figure 3.16: TIT - node with looped bar
162

3.4.5. Curved Tensile Ties

When tensile ties are curved some special considerations for detailing must be
given. The out of axis deviation force induced when the tie tries to straighten
under the applied tension must be anchored back into the member by closely
spaced "tie back" stirrups. The stirrup spacing "s" must be so selected that the
cover will not break away between two stirrups when the curved bar tends to
straighten. Leonhardt [106] has suggested that by considering the approximate
flexural stiffness of the curved bars and by limiting the tensile stress in the cover
concrete, the stirrup spacing should not exceed the following values:

S =:;; 2 ,1s / [1 - (f'cl c R) / (420 d b2)] when (S, + db ) ~ 3 c

s =:;; 2 ,1s / [1 - (f'cl Sl R) / (840 db2)] when (sJ + db ) < 3 c
,1s = 3 db or c whichever is smaller
S1 = parallel bar clear spacing (see Fig. 3.17)
flCI = concrete tensile strength
R = radius of curvature

The equation are based on the assumption that the working stress in the curved
bar does not exceed 34 ksi.The stirrups can be omitted when the radius of the bar
is large enough so that the cover concrete will be sufficient to supply the radial
tensile force.

R for f'c = 4000 psi

R for flc ~ 4000 psi

In order to prevent splitting in the plane of the bars, Leonhardt [106] proposes
some minimum cover unless transverse reinforcement is proVided (see Fig.
3.17).

c for f'c~ 4000 psi

 163

- stirrups
---s----..

curved tensile tie

lL.
T

"
stirrup-

curved tensile tie

Figure 3.17: Dimensions for curved tensile ties (from Ref. [105])
164
 CHAPTER 4. DESIGN AIDS

4.1 Detailing aids

Recent advances in the understanding of the behavior of concrete

structures have resulted in more sophisticated methods of analysis. Computer
oriented analysis and plotting techniques enable the elastic or inelastic analysis
of highly indeterminate structures to be carried out speedily. However, an
elaborate analysis becomes worthless if the computations are not translated ir:lto
successful structures. The design process is a sequential and iterative decision-
making process. The 'first step is to define the needs and priorities which a
building or bridge has to fullfil. These may include functional requirements,
aesthetic considerations, and economic aspects. Based on the cc;mstraints,
needs and priorities, possible layouts of the architectural, structural and other
sytems are developed. Preliminary cost estimates are made and the final choice
of the system to be used is based on how well the overall design satisfies the
prioritized needs within the budget available. Once the overall layout and general
structural concept have been selected, the structural sytem can be designed to
ensure structural adequacy. Based on the selected design a structural analysis
is carried out to determine the external and internal forces in the structure. When
a structure or structural element becomes unfit for its intended use, it is said to
have reached a limit state. The limit states for concrete structures can be divided
into three basic groups.

Ultimate limit states: These involve a structural collapse of part or all of the
structure. Such a limit state should have a very low probability of
occurrence since it may lead to loss of life and major economical losses.
(100 year lifetime probability == 10-5 == B = 4.2)

Serviceabil ity lim it states: These involve losses of the functional use of the
structure but not collapse of any part of the structure.
(100 year lifetime probability == 10-3 == B = 3)

165
166

Special limit states: This class involves majorbut repairabledamagetothe

structure due to abnormal conditions, such as long-term physical or
chemical instability
(100 year lifetime probability ::= 10-3 "'" B = 3)

For normal concrete structures the determination of acceptable levels of

safety against occurence of each limit state are carried out by the building code
or design specification authorities. They specify the load combinations and
safety factors to be used in checking the limit states.

Since many repetitive computations are necessary to proportion concrete

structures, handbooks containing tables or graphs of the more common quanti-
ties are available from several sources. The American Concrete Institute and the
Prestressed Concrete Institute [107] pUblishes its Design Handbook in several
volumes, the German Concrete Group publishes its yearly "Betonkalender"
[108], the Concrete Reinforcing Steel Institute publishes the CRSI Handbook
[109]. In recent years, specialized computer programs have been replacing
design aids in many applications [110].

Detailing consists not only of the preparation of plans giving concrete

dimensions, reinforcement placing drawings and reinforcing bar details, but in-
corporates the whole thought process through which the designer enables each
part of the structure to perform safely under the various limit states. This chapter
gives some background for typical application of the strut- and-tie model. It is
intended to assist in establishing design parameters for some specific applica-
tions. These design guidelines should help a designer develop and dimension a
strut- and- tie- model and apply the model to different situations. This chapter
draws on the analytical and experimenta.l results presented in the earlier
sections. It uses these results to develop design procedures for concrete
structures. Overall, it should lead to a better understanding of the force flow in
D-regions and the designer should have substantially improved knowledge
regarding the design process.
 167

A foremost practitioner of strut- and- tie model detailing is Dr. Jorg

Schlaich. He and his coworkers have developed a series of detailing aids
[1,2] for comonly occurring situations. A number of these aids are given in
Appendix A herein as further assistance to the designer interested in
applying strut- and- tie models.

The general design approach ensures a reasonably ductile behavior by incorpo-

rating relationships to preclude premature anchorage or concrete failure and
requiring that reinforcement yield substantially before final failure. One ofthe ad-
vantages of the strut- and- tie- model is that both prestressed and reinforced
concrete structures can be treated with the same model. Experience indicates
that the models are of most use in the 'D' regions of a structure.

The required checks for a concrete structure are:

ultimate limit state: factored loads and reduced nominal resistances

serviceability limit state: working loads and acceptable stress or
deformation states
durability:· technological aspects and requirements

The proposed strut- and-tie- model is based on the plasticity analysis for cracked
concrete and gives a lower bound forthe ultimate limit state analysis as long as
premature anchorage or concrete strut failures are precluded.
168

4.2 Types

The detailing process for D-regions begins with isolating the D-regions
'from the B-regions in a structure and development of a preliminary strut- and- tie
model. In order to find the appropriate strut- and- tie- model the load path can be
traced, general knowledge of appropriate models can be used, or in complex
cases results of an elastic 'finite element analysis should be used.

All the design factors specified in this section are based on the 28-day
design concrete compression strength. Concrete strengths up to about 12000 psi
have been studied and included in the proposed parameters. However, the
present design methods, such as those in the AASHTO - Specification [4]- and
the ACI Building Code [2] were actually developed using concrete strengths
varying mostly from 2500 to 6500 psi. For higher strength concrete the average
ratio of proportional limit stress to ultimate strength under uniaxial loading ranges
from 55 to 84 percent [58]. Tests at the University of Texas at Austin by Khana
[111] indicate that for normal strength concrete the proportional limit under
uniaxial load is generally between 40 and 43 percent of the ultimate strength.
The principal reinforcing steels available have nominal yield strength
between 40 and 75 ksi, with 60 ksi being the most widely used. The ACI Code
318-89 [4] provides that the specified yield strength will be the stress correspond-
ing to a strain of 0.35%. All standard bars are deformedround bars, designated
by size from #3 to #18: this number corresponds roughly to the bar diameter in
one-eights of an inch.
The principal prestressing tendon materials are seven-wire strand,
smooth wire, smooth bars and deformed bars. The yield strength varies between
120 and 270 ksi [55]. A minimum amount of reinforcement is necessary to ensure
distributed cracking and should be placed to avoid infrequent wide cracks. In
critical cross sectional areas, crack prediction formulas can be used to distribute
reinforcement to avoid wide cracks.
 169

4.3 Typical Examples of Detailing Aids

The design examples given in this section are typical D-region details.
Figure 4.2 shows several different examples which will be developed in subse-
quent subsections. For clarity in presentation no nominal resistance reduction
factors (<I» or load factors have been used in the following examples. For actual
design purposes the nominal concrete strength "v f'c·' must be further reduced
with the appropriate "<I>" or resistance reduction factor. The design loads must be
increased with the appropriate load factors. Since these vary from code to code
and with actual applications, they are not included herein in the interest of
simplicity.

Nodes shown in Fig. 4.1 can occur in the different examples. Equilibrium
must be established in the nodes. The forces depend on the choice of their
position and are known from the boundary conditions of the B-regions. If nodes
with more then 3 forces occur the principle remains the same.

cee CCT cn TIT

I C2 C1/
T1&-; C1
T2

y-n
I

C1 ftc « C3
C2
T2 13
/

T2 T3
AC2
:1 ~ ~1 ~1
c~,
C3
4JC2
T1
T1V . T2V ·
Figure 4.1: Types of nodes
170

I-----------------T--------~~~-----l
I Load near a aupport: I Corbels projl'Cting ---. I
, -i I
I tram a column: II'

f-:I "F: ';~ t>

I
I
I
t I
I '
I
I R> -i
~,
1
I
I
I Y I O..egion a·region El I 0-<09"'" -+ I
: -' ~T:- 4• i · -l :
I
I
t-i rout>.
I I.a~
I
I
I ----l
1.5h ....B 1
I
I I I
I I ~~"'" I
I LJ~ I I
r--------~1-------I-----------------~
I I I 1
II m

I Deep beam with a hole P, IF=poP'

1'"' I Dapped end beam: I

1
I D ~::" l
"T j
i : ,£ __ I
I .J__
0 ~'O" I-~r. + B
i
I
I 1 I -E=-l .-j~ I
I = -;- I [,.5h~ L1.5h_'J 1
'I t,: et ll -- I I . I
~~ ~~ I I
IL t.-,-i-i I , +I I
~

I I I
I Anchorage zone: I 1
I I I
I I I
:r I Pretensloned beam wllh eccentricity :

: J r '-l ~.- ..~.""-: 11111111111111111-11111111111111111 i

: : ep.§i j2_C p
:

I I I
I I I. I
: L ~Oft:l~h I : :
L ~ ~

Figure 4.2: Examples to be presented in Detail

 171

4.3.1 Load near a support

For a heavy point load located near a support as shown in Fig. 4.3 and
Fig. 4.4, the proposed strut- and- tie- model is a logical approach to represent
the flow of forces indicated by the elastic analysis results shown in Fig. 4.3. If the
load is applied at a distance from the support smaller then the height of the
specimen (x::;; h), then a compression 'field between the support and the load
provides the primary internal force mechanism. This is also the reason why no
traditional shear reinforcement is needed in the area between the point of the load
application and the support. The strut- and- tie- model chosen reflects that the
primary compression strut between the external forces is a bottle shaped
compression field. The highly loaded bottle shaped compression field can be
represented with the local strut- and- tie- models shown in Fig. 4.4. The local ties
(T1) are dependent on the force diffusion angle 1/<1>1 11 and the compression force.
The T1-force in the tension ties can be provided for practical purposes by using
equilibrium to proportion orthogonal vertical and horizontal ties. The detailed
calculation in Example 4.1 shows that the tie forces in the bottle compression
field can be of large magnitude and have to be taken into account.

The bearing plate forces IIFII were divided into two individual forces with
separate nodes. The magnitude of each force is determined from the overall
analysis according to the proportion flowing to the left support and that flowing to
the right support. In order to get uniform compression in the struts the bearing
plate width has to be also subdivided into two dependent widths matching the
compression forces F' and FII . The example shows that the new strut- and- tie-
angle based on the widths a' and all for this case does not have a significant
influence on the strut- and- tie- forces. The difference is only about 1%. It could
have a significant influence if the two compression struts IICall and IIC4 11 had to
carry similar forces and the bearing plate is much larger (see Section 4.3.5
anchorage zone). The inclined compression struts outside of the support region
are assumed to be at an inclination of 45° as traditional in shear design (<1>7 = 45°).
Thus the shear panel length becomes equal to the height Z in panels to the right
of section cd in Fig. 4.4.
172

Example 4.1: Load near a support

Design a beam end for the member shown in Fig. 4.4 to transfer a vertical re-
action load applied within a distance x ::;; d to a supporting column. The load to
be transferred is 200 kips. Member dead load is neglected for clarity in this
example. Use f c = 5,000 psi and fy = 60,000 psi.

The computational steps are:

Estimate member sizes and dimensions

Divide member into B- and D-regions (see Fig. 4.4)
Develop a strut- and- tie- model (see Fig. 4.4)
Compute the external forces
Compute the strut- and- tie- angle
Compute the strut- and- tie- forces
Dimension reinforcement for ties
Determine anchorage requirements
Check concrete stresses

To prevent large crack widths under working loads some arbitrary reinforce-
ment spacing limits are applied in the final design layout in Fig. 4.7.
 173

... -
- - " . '''''--'\j'- -r
-~//T
.,/////! "
+
,

x/I / t \ --...
, . , / ..- /" /<' /11 '\ - - -
, I ;t )C x - - -
Illxx;xI!IY i
\
......
1-1-1-><><'1-1-111 + ......
...
I /'1'1-'1-)<)<1/ /
-;. f
I Iff. '1><><)< I- I / I
1111'1'/-)<'1-'1-'11 I ~

li%19)<)<)(X'"
--- --- --- ---- ---
'/-XXX)()( ~ / / / / /

t h~~~'/-xxxx'!- .... .- .-
"" .-
l h~'-I/><XXx.~,
l /,~II-XXx.~,,-.
-- -
--
~ lll;<x~~,,~-
t ~ '11>< 'x, ,,'-~
- -- -- - -
t%lxx,
tL!/~------~
! '
It!~--=-
.!~I \ I I

Figure 4.3: Elastic finite element analysis: principal stresses

 Example: Load near support

174
F

T
f
r ==J

h
.....L

0- region

F =F' + F"

htf d f

z=3/4 h

llt A
l; ~

I~ x+1.5h ~I
Figure 4.4: Strut and tie model for load near a support
 175

load and dimensions:

F = 200 kips
1= 300 in.
h = 40 in.
x = 20 in.
a=10in.
b = 12 in.
z 0= 0.9 d = 0.9 * 0.85 * h = 3/4 h = 30 in.
Minimum clear cover = 1.5 in.

External forces:

=
A F (I-x) /I
A = 186.7 kips
F'=A
B=Fx/l
B = 13.3 kips
F" = B

Strut and tie angle:

tan ¢a = z 1 x =0.75 hI x
tan <pa = 0.75 * 40/20
tan ¢a = 1.5: arctan 1.5 =,56.3° (old aa)
epa = 56.3°

aIh =0.25
61 = 12 + 3 I-J(a I h) =18°
176

F' F"
Internal forces:

CaCa = Alsin q>a = 186.71 sin 56.31

= 224.4 kips
'. •

C1 =Ca/(2cos81)=118.0kips
C2 = Cal2 = 112.2 kips
T1 = (Cal2)tan 81 = 36.5 kips 1f4--~:==-_al-a~_~_~a"J ._
F' = Ca sin q>a = 186.7 kips
F" = F - F' = 13.3 kips
a" = F" * al F
a" = 13.3 * 10/200 = 0.66 in.
a' = a - a" = 9.34 in.
.·I.. . .f-----:~ _
a'/2 ~I

newq>a
New strut and tie angle:

q>a = arctan [z 1 (x - al2 + a'/2)]

q>a = 56.75 0 (new aa)
q>4 = arctan [z 1 (x - a/2 + a"/2)]
q>4 = 62.9 0

q>7 is chosen to be 45°

 177

New Internal forces:

Ca = Nsin <1>a = 223.2 kips

C1 =Ca I (2 cos 81) = 117.3 kips
C2 =Cal2 = 111.6 kips
T1 = Cal2 * tan 81 =36.3 kips (Negligible change)

F' = Ca sin <1>a = 186.7 kips F"

F" = F - F' = 13.3 kips (= F" (old)

C3 C5
T2 = A I tan <1>a = 122.4 kips ~O~)-
C3 = T2 = Ca cos <1>a = 122.4 kips Q ~~
C4
C4 = F" I sin 62.9 = 14.9 kips
C5 = C3 - C4 cos 62.9 = 115.6 kips
\
T3 = C4sin 62.9 =F" = 13.3 kips
T5=T3
C5
C7 = T3 I sin <1>7 = 18.8 kips
C6 = C5 - C7 cos <1>7 = 102.3 kips

T4 = C5 = 115.6 kips
178

Compute the reinforcement for the ties:

From Fig. 4.4, two T1 ties are required.

T1 =36.3 kips
T1 h = T1 sin epa =30.4 kips
Ash = T1 h / fy = 0.51 in. 2 (per each T1 tie) Grade 60
total Ash = 1.02 in. 2 < 6 - #4 = 1.20 in. 2
For the horizontal componentT1 h 3 pairs of #4 bars are spaced evenly
on the sidefaces (see Fig 4.7)
T1 v = T1 cos epa =19.9 kips
Asv = T1v / fy =0.33 in. 2 (per each T1 tie) Grade 60
total Asv = 0.66 in. 2 < 4 - #4 = 0.80 in. 2
For the vertical component T1v two NO.4 U stirrups were used in the heavy
shear span.
Additional #4 U stirrups are placed under the load and just outside the
support (see Fig. 4.7).

For the horizontal tensile tie T2 try NO.6 bars.

T2 = 122.4 kips
As2 = T2/ fy = 2.04 in. 2 < 5 - #6 =2.20 in. 2
Use five NO.6 bars in the flexural tension zone (see Fig. 4.7).

T3 = T5 = 13.3 kips
As3 = As5 = 13.3/60 =0.22 in. 2 <2 - #3 = 0.22 in. 2
Use #3 U stirrups
Since ep7 is assumed at 45°, spacing of these stirrups can be z =30 in.
However, such wide spacing is unwise since major diagonal cracks could
form between such widely spaced stirrups. Stirrup spacing should be
restricted to z/2 or 30/2 = 15 in. #3 is the smallest practical size. Use #3 U
@ 15" (see Fig~ 4.7).

For the continuing horizontal tensile tie T4 try the same No.6 bars as for T2
(flexural reinforcement)
=
As4 = 115.6/60 1.93 in. 2 < 5 - #6 =2.20 in. 2
 179

Anchorage requirements:

The horizontal #4 bars and the vertical #4 stirrups provided to take the
T1 tie forces should be well anchored by hooking the horizontal bars around
the stirrups and hooking the vertical stirrups around the bottom and top bars.
No special check would be required for a member of this size as such
hooked stirrups could be easily developed.

The main flexural reinforcement (Tie T2) anchorage at the support

needs careful examination. Using the provisions of ACI 318-89, for a clear
cover over all reinforcement of 1.5 inches and with #4 stirrups, the effective
cover below and outside the #6 bars is 2 inches. The clear spacing between
the 3- #6 bars in the bottom layer is 2.88 inches. Thus the cover is greater
than 2 db and the clear spacing is greater than 3 db satisfying ACI 318-89
Sec. 12.2.3.1 d. Hence, a multiplier of 1.0 is used with the basic
development length.

Id = 1.0 Idb = 0.04 Abfy/,fu = 0.04(0.44) 60,00011'5,000 = 14.9 inches.

Should the #6 bars need to be hooked, the basic development length

for the hooked bar from ACI 318-89 Sec. 12.5.2 is

Idh = Ihb= 1200 db /..,/fC =1,200 (0.75) rvE,OOO = 12.73 inches.

However, since the side cover over the hook is less than 2 1/2 inches,
ACI 318-89 Sec. 12.5.4 would require that stirrup ties be spaced along the
entire Idh at a spacing of not more than 3 db, or 2.25 inches. This would
make placement of concrete very difficult and so hooked bars are not very
desirable here.

In order to illustrate the effect of confinement due to the bearing plate

and stirrups provided, the more complex development length equation will
also be checked.
180

I " 1.25 db {to I [ 4 (t: t" ]_-5_0 ..L-} _

d
{ 1.2 + 3 c / db +( Au fyt )/ ( 500 S db ) + [ ( fn t S
( 200 - e 2 ) /1 000 ]}

2. layer -KP:l=+J

~2
r
1. layer

----:r------
0.5 in.
L1 e1
c = 1.5 in.

Id=8in. for a=10in.

The initial estimate of the bearing plate size was based on a length a = 10
inches and a width equal to the beam width b = 12 inches. The bearing stress at
reaction A is thus f n = A / ab = 187k / (10) (12) = 1.56 ksi, well within the concrete
bearing capacity. From the clear cover selected (1.5 inches minimum), the #4 stirrups
chosen for the Tl v ties, and the spacing between layers selected (s = 1.5 inches), the
values of e1 and e2 can be calculated as 2.0 inches and 4.25 inches, respectively.

I " ( 1.25 )( 0.75)[ (60,000/45,000) -50 ]'----_ _

d
[ 1.2 +(3 * 2.0/0.75 ) + ( ( 2 * 0.20 * 60,000 )/ ( 500 * 15 * 0.75 ) )
2
+ ( -/1560 (200 - 2.0 )/1000 )] = 7.2 in.

Hence, if anchorage at the support becomes critical, the beneficial effect of

confinement can be considered. This reduces the necessary development length for
the lower layer of bars to 7.2 in. and for the upper layer to 7.4 in., say 7.5 in. for
both.
 181

If the main flexural reinforcement runs to within 1.5 in. (for cover) of the
end of the beam as shown in Fig. 4.7 and Fig. 4.5, the length of bar ava.ilable for
meeting the requirements in node a is 15 - 1.5 + (0.75 / 2) (cot 56.8) = 13.75
in. for the lower layer and 15 - 1.5 + ( 0.75/2 + 2.25) (cot 56.8) = 15.22 in.
for the upper layer. Thus the bars in the upper layer clearly exceed the 14.9 in.
required for Id according to ACI 318-89. The bars in the lower layer provide only
92% of the required Id according to the ACI 318-89 provisions* but do provide
188% of the required Id when the local confinement due to the bearing plate is
considered. Hence these straight bars can be considered effectively anchored
as detalled.

* (There is certainly not this degree of accuracy implicit in the computations.)

182

Concrete stresses at node zones:

concrete strength: fIe = 5 ksi

efficiency factor: v e = 0.9 - 0.25 (5,000) I 10,000 = 0.775

Normally the compressive stress in nodes need only be checked where concen-

trated forces are applied to the surface of the structural member; e.g. below

bearing plates, anchor plates and over supports.

Node g (Fig. 4.4): CCT - node

See Fig. 4.5 for node geometry.

a = 10 in. b = 12 in.
<j>a = 56.8° Ca = 224.4 k
w2 = 2 db + S = 2 * 6/8 + 1.5 = 3.0 in.
wa = a sin <j>a + w2 cos <j>a = 10 sin 56.8 + 3 cos 56.8 = 10.0 in. (See
Fig. 4.5)
crca = Ca I (wa * b) :5 Ve fIe
crca = 224.4 I (10.0 * 12) = 1.87 ksi:5 0.775 * 5. = 3.88 ksi
cra = A/(a*b):5V e f'c [ac b/(ab)]0.5
cra = 186.7 1(10 * 12) =1.56 ksi:5 0.775 (10 110)°·5 * 5.
cra = 1.56 ksi :5 3.88 ksi with full width bearing plate

a = 10"
 183

T1v

2-#4 A wa

'7
c
.--
w2 T2
+ ~
cC '---t-r-r......-'!'...-T""'T'"..-J,_d~ .....

•
•
~Bearing plate
l.-~--J

wa =a sin~a + w2 cosQa

Figure 4.5: Node a - CCT - node

184

Node 12 Fig. 4.4): CCC node

See Fig. 4.6 for node geometry

a = 10 in.
a' = 9.34 in.
a" = 0.66 in.
b = 12 in.
a'
<l>a = 56.8°
<1>4 = 62.9°
Ca = 223.2 k
F = 200 k
C3 = 122.4 k
/1
/ /<l>a I
C4 = 14.9 k
/
C5 = 115.6 k /

= F/(a*b)~vef'c

= 200/ (10 * 12) = 1.67 ksi ~ 0.775 * 5. = 3.88 ksi OK

(Jcwa = Ca / (wa * b) ~ V e f'c

0: = arctan [a / (2a')] = 28.16°

wa = a cos (a' - <l>a) / (2 sin a') = 9.3 in.
or
wa = a' cos <l>a + a / 2 sin<l>a = 9.3 in.
Note: wa was computed as 10.0 in. at
the lower node - this is
approximately correct)

<Jewa = 223.2 / (9.3 * 12) = 2.0 ksi

~ 0.775 * 5. = 3.88 ksi OK
 185

a ew4 = C4/ (w4 * b) ::;; ve f'c

d: = arctan [a / (2a")] = 82.48°

w4 = a cos (a"- <1>4) / (2 sin a") = 4.75 in.

a ew4 = 14.9/ (4.75 * 12) = 0.26 ksi ::;; 0.775 * 5. = 3.88 ksi OK

<J'c3 = C3/ (w3 * b) ::;; V e fIe

w3 = a/2 = 5 in
crew3 = 122.4 / (5 * 12) = 2.04 ksi ::;; 0.775 * 5. = 3.88 ksi OK

a ewS = C5/ (w5 * b) ::;; V e fIe

w5 = a/2 = 5 in.

a ews = 115.6/ (5 * 12) = 1.93 ksi ::;;0.775 * 5. = 3.88 ksi OK

The assumption that w5 = a/2 assumes a hydrostatic type node. Note that the
computed uniform stress in C5 is only 1.93 ksi or only 45% of the 0.85 f'c as-
sumed at failure in compression zones under ACI or AASHTO [3,4]. This indi-
cates that the initial assumption z = 30 in.. was very conservative. The actual z
atthis load level is more like 40 - 2/3 (5.0) - 3.28 = 33.4 in. Calculations
could be revised on this basis but since everything has checked OK at the lesser
assumed z, forces would be decreased. Therefore, this analysis is conserva-
tive.

Concrete stresses in compression diagonal struts:

Since this application of load results in a heavy diagonal strut rather than a
compression field, it might be useful to check the main strut system for the crite-
ria a::;; 0.6 v e f'c = (0.6) (0.775) (5) = 2.33 ksi .
Since the nodes were checked for the concentrated Ca force, it is not
likely that C1 or C2 will govern. However, to check the level of stress in the
186

struts take C2 = 112.2 k distributed over a strut width x at a distance a 'from

the center of the lower node. Since 81 was assumed at 18°, x = wa + 2 a tan 8
= 10 + (2) (10) tan 18 = 16.5 in. a2 =C2 / b x = 112.2/ (12) (16.5) =0.57 ksi <
2.33 ksi (OK)

Details:

The complete layout of reinforcement is shown in Fig. 4.7. It is empha-

sized that many different arrangements could be used depending on the strut-
and- tie model selected.
 ...
.. a'
a
-
-, a" .
187

+
aJ4

I
I
I
I
I
l-a'l2+a"I2---1

w3 =w5 =aJ2

ct = arctan a I (2a')

et' = arctan a I (2a")

wa =a cos (a' - $3 ) I (2 sin a')

or
wa = a' cos cPa + aI2 sin cPa

=
w4 a cos (cr" - 4>4) I (2 sin a" )
or
wa4 = a"cos 4>4 + all. sin 4>4

Figure 4.6: Node .12: eee - node

188
,~A

l' ",..--2 - #4 bars

r.----40"--~~I

~---------D - region = 80" ---------~

Section: A· A

T· 375

8.25· .....-#4

40"
+
+8 .25·
t
3.5"

~
8.25·

+-
8.25·
1
1.5"
,
~.25
~ 12· ~I

Figure 4.7: Reinforcement layout

 189

4.3.2 Corbel projecting from a column

A corbel is a short member that cantilevers out from a column or wall to support
a load (see Fig. 4.8). The corbel is generally built monolithically with the column
or wall. The term corbel is generally restricted to cantilevers having shear
span-to-depth ratios, aid ~ 1.0 [62]. The design of corbels and brackets is based
primarily on experimental results. The strut- and- tie- model allows one to
visualize the flow of forces in the typical D - region. Hagberg [112] also proposed
a truss analogyfordesign of concrete brackets. Experimental and nonlinear finite
element analysis studies were done by Cook and Mitchell [113] for corbels with
vertical and horizontal loads.

1. crack

2.crack
F

Figure 4. 8: Typical cracking patterns of corbels

190

In tests, corbels display several typical modes of failure, the most common of
which are:

yielding of the tension tie

failure ofthe end anchorages of the tension tie, either underthe load point
or in the column
failure of the compression strut by crushing or shearing
local zone failure under the bearing plate

When using a reinforcement tie hooked downward, as shown in Fig.

4.9(a), the concrete outside ofthe hook may split off. In orderto avoid this problem
closed ties extending past the loading platemay be used, or straight bars may be
used but should be anchored at thei r ends by weldi ng them to a cross bar or plate.
If the corbel is too shallow at the outside end, there is a dangerthat cracking may
extendthroughthecorbel as shown in Fig. 4.9 (b). ForthisreasonACI [4] requires
that the depth of the corbel be at least 0.5 d at the outside edge of the bearing
plate.

A total of eight corbels, divided into four series with concrete strength
ranging from about 6000 psi to 12000 psi, were studied by Yong, Closkey and
Nawy [114]. Only vertical ldad was applied. The specimens which had no steel
reinforcement (f c = 10260 psi and 12630 psi) had failure at the interface of the
corbel and the column as shown in Fig. 4.8. The failure plane started at the
reentrant corner of the horizontal surface of the corbel and the vertical face of the
column. All the other specimens had almost identical behavior when subjected
to failure under a vertical load. A flexural crack (crack 1) 'first started at the
reentrant corner and propagated slightly into the column. Close to failure another
crack (crack 2) appeared at the inner edge of the bearing plate and propagated
at a faster rate than the initial crack towards the interface of the column and the
sloping face of the corbel.
 191

It is obvious that the horizontal reinforcement is a very important factor. Also

some additional reinforcement has to be placed under the concentrated load.
The compression strut develops additional tension as indicated earlier for bottle
compression struts. The same type of force flow was found in the previous
Example 4.1 "load near support".

a. failure crack outside of the hook b. failure crack through the corbel

Figure 4.9: Possible failures of corbels

192

The distribution of principal stresses in a trapezoidal corbel and its

supporting column, obtained by Franz and Niedenhoff [115] from photoelastic
models, is presented in Fig. 4.10. The load received from a gantry girder is
simulated. An evaluation by Park and Paulay [105] of that study reveals the
existence of four conditions:
The shape of the corbel has little effect on the state of the stresses. In a
rectangular corbel, the outer corner opposite the load point is virtually
stress free.
The compression force along the sloping edge of the corbel is also
approximately constant, indicating that a diagonal compression strut
develops.
The inclined tensile stresses arising from the change of direction of the
compression force are very small.
The tensi Ie stresses along the top edge are almost constant between the
load point and the column face.
tenlile trajeccoriel

matnitude of
the principal
ten"te nreuel

matnit ude of tM
comprenive nrellel 'e
parallel to the face

Figure 4.10: Stress trajectories in a homogeneous elastic corbel

(from Ref. [115], Figure 'from Ref. [105])
 193

In the traditional approach to the problem, one would have relied on the
consideration of shear stresses. Indeed, corbels have often been reinforced with
diagonals, as shown in Fig. 4.11 , to take a substantial part of the shearing force.
The investigations of Franz and Niedenhoff [115] have conclusively proved the
inefficiency of this approach.

~---I--
-- -J..--
1
--~---
I

Figure 4.11: Diagonally reinforced corbels

(from Ref. [115], Figure from Ref. [105])

Park and Paulay [105] identified failure mechanisms as shown in Fig. 4.12 from
the tests by Kriz and Raths [116].

<-I (hI leI 141 lrl /fI

(a)lIexurai tension (b) diagonal splitting (e) a1iding 8hear (d) anchorage splitting (e) aullhing due to bearing (I) horizontal tension

Figure 4.12: Failure mechanism in corbels: [From Ref. 105]

194

Thestrut-and-tie- modelforacorbelshown in Fig. 4.13isproposedandassumes

that the concrete stress at the lower reentrant corner reaches the effective
concrete strength [28]. A typical CCC -node is found in the lower corner and the
tensile force in the upper chord can be computed without knowing the angle of
inclination of the main compression strut. The flow of forces and typical CCC-
node dimensions are shown in Fig. 4.14.

Tie CCT-node
;'
;

Strut - -r-- ;;

I
I
Strut I

Figure 4. 13: Strut- and- tie- model for corbel

Figure 4.15 shows eight test results from [117] with a concrete range from 5690
to 12630 psi, six test results from [118] with concrete strengthts from 4200 to
5057 psi and one test from [113] with 5858 psi compared with the proposed strut-
and- tie- model. The statistical analysis forthe comparison is shown in Table 4.1.
The proposed model is generally conservative and reasonably accurate.
 195

f
Fv

Fh..

CCC - node

C3

w2= Fv
__
C2 b '{f'e

Figure 4.14: Corbel strut- and- tie model

196

2.0....------------------------,
u = unreinforced - 0 - - experiment/theory
• experiment/theory=1.0

o
...
>.
0
1.54-----------------------1
caJ::
.-

...
11-
-
QJ

u
.Q C 1.0 4----44r---.......---ila..---.........----..-----.....--.....,
Q) QJ
ca
.~ .;:
E
.::: QJ
::J c..
~ 0.5 -+---------------------------1

0.0 -l---r-----r---.,.-----,--r----r----r---.----r-.....,
4000 6000 8000 10000 12000 14000·
concrete strength [psil

Figure 4. 15 : Strut- and- tie- model results compared with test results

X1 : Column 1
Mean: Std. Dev.: Std. Error: Variance: Count:

11 . 082 1. 155 1. 04 1·024 115

Minimum: Maximum: Sum:
.93 , .442 , 6.224

Table 4.1: Statistical analysis from Fig. 4.15 omitting unreinforced specimen
 197

Example 4.2 : Corbel projecting from a column

Corbel projecting from a column

Design a pair of corbels to transfer vertical reactions of 200 kips and

horizontal reactions of 40 kips to a supporting column. Use f'c = 8000 psi
and fy = 60,000 psi. The computation steps are:
Estimate member sizes and dimensions
Divide member into Band D regions (see Fig. 4.16)
Develop a strut- and- tie model (see Fig. 4.16)
Compute external forces
Compute the strut- and- tie angle
Compute the strut- and- tie forces
Dimension requirement for ties
Determine anchorage requirements
Check concrete stresses
Assume clear concrete cover over all reinforcement is 1.5 in.

Load and dimensions:

Fv = 200 kips
Fh = 40 kips
h=12in.
g' = 8 in.
g"= 6 in.
b = 12 in.
a= 5 in.
f = 5.5 in.
i = 9 in.
d "'" 0.75 (g' + g") = 10.5 in.

Since tie reinforcement often has to be placed in multiple layers, cover and
spacing requirements suggest that a conservative estimate be used for the
effective depth, d. The basic strut- and- tie model shown in Fig. 4.16 is
chosen for computing the strut and tie forces. The compression struts are
again assumed as "bottle" struts. The location of node b is uncertain and
depends somewhat on the location of C4. After C4 is located the node b will
be chosen on a 45° angle inward from the reentrant corner.
198

Strut and tie angle:

w4 == C41 ( b Ve f'c )
w3 = w4 I tan <j>a
<j>a = arctan (d - w3/2) I (f + w4/2)

w3, w4, <j>a are unknown but interrelated

C3
 199

f~1 1.5 h

Fv Fv

1.5 h cosB

~
Figure 4.16: Strut- and- tie model for corbel projecting from a column
200 Internal forces:

C4 = Fv = 200 kips

concrete strength: f'c = 8,000 psi

concrete efficiency factor: v a= 0.9 - 0.25(8,000) 110,000 = 0.7
Minimum w4 = Fv 1 (b v f'c) = 200 1 (12 * 0.7* ?O) = 2.98 in..
Thus it would be acceptable to use w4 = 3.0 in. However, cover over the main
column bars indicates C4 should be at least 2 in. from the column face. Use
w4= 4.0 in..

With the location of C4 at w4/2 or 2.0 in. inside the column face, the location of
node b will be chosen 2.0 in. inward from and 2.0 in. above the reentrant corner.
This establishes the angle of inclination of the strut Ca:

Tan ~a = (d - 2.0) 1 (f + 2.0) = (10.5 - 2.0) 1 (5.5 + 2.0)

= 1.133
~a = 48.6°
Ca = Fv 1 sin ~a = (200) 1 sin 48.6° = 266.6

The main tension tie T2 then must balance the horizontal component of Ca as
well as equilibrate the lateral tension force Fh.

T2 = Ca cos ~a + Fh = (266.6) cos 48.6 + 40

= 216.3 k
C3 = Ca cos ~a = 266.6 cos 48.6 = 176.3 k

Determine the diffusion angle for C1 :

a/ i = 5/ 9 = 0.55
e = 12 + 3 / -.j 0.55 - 16.0°
T1 = (Ca / 2) * tan 16 = (266.6/2) (tan 16)
T1 = 38.2 kips
 201
Reinforcement for Ties:

T1 = 38.2 kips

T1 h =T1 sin <j>a =28.7 kips

Ash =T1h / fy =0.5 in. 2

Since there are 2 T1 forces to be provided for each corbel

total Ash = 1.0 in. 2 < 6 - #4 (Grade 60) = 1.20 in. 2 (Use 3 closed ties)

T1 v ;;; T1 cos <j>a =25.3 kips

Asv =T1v /fy = 0.42 in. 2

total Asv = 0.84 in. 2 = 4 - #4 (Grade 60) = 0.80 in. 2 (Use 1 closed tie and

1 pair of bent bars which must be hooked over

and anchored by the topmost horizontal #7 closed tie)

T2 =216.3 kips
As = T2 / fy = 3.60 in. 2 = 6 - #7 (Use 3 closed ties (Grade 60) = 3.60 in. 2

Concrete stresses at node zones:

Node g: CCT - node (Figures as in Example 4.1)

a =5 in.
(Ja = Fv / (a * b):::;; [a b / (a b)]0.5 V fie
c e

(Ja = 200/ (5 * 12) =3.3 ksi:::;; 0.7 * 8. * (1)°·5

(Ja = 3.3 ksi :::;; 5.6 ksi OK

(Jca = Ca / (wa * b):::;; V fIe

wa = a sin <j>a + w2 cos <j>a = 5 * sin 48.6 + 2.5 cos 48.6 = 5.40
(Jca = 266.6/ (5.40 * 12) = 4.11 ksi:::;; 0.7 * 8. = 5.6 ksi OK
202

Node.Q: GGG - node w4 = 4.0 w3 = w4 /tan $a = 3.52

w1 = ..J4.02 + 3.522 = 5.32

cr ca = Ga/w1 b::;; 0.7f'c = 5.6 ksi

= 266.6/ (5.32) (12) = 4.18 ksi < 5.6 ksi OK

cr3 = G3 / w3 b = 176.3/ (3.52) (12) = 4.17 ksi < 5.6 ksi OK

cr4 = G4/w4 b = 200/ (4.0) (12) = 4.17 ksi < 5.6 ksi OK
 203

Anchorage requirements:

Three closed #7 ties were used for the main horizontal reinforcement. Because
the tension tie in the strut- and- tie- model is assumed to be stressed to the yield
strength intension between the loading plateandthecolumn, itmustbeanchored
in the node zone gand outside the bearing plate forthattension. The closed ends
ofthe ties should be sufficient positive anchorage. If straight bars were used they
could be welded to an angle or bar at right angles to the tie (see Fig. 4.17), or be
welded to a transverse reinforcing bar of the same diameter as the tie.

bearing plate

angle welded to bar

ezzzzo07ZZZZZOZZZZzo,

L.....----.Jlr-----/

Figure 4. 17: Anchorage detail for corbel design (from Ref. [62])
 1--"- - 1 2 " ----.,-,

204

column reinforcement
::.II:I!!I::: .····:.·!i;·11::;j!1:1:::::ii:l1·!i::i:l·1.::1:lii::;; '::::::::::"J---g"--j

5.5"-1
I 18"

Fv Fv
A A
Fh ..... a
.1
2 closed ties
(1 MKA t 1MKB) #7
-r=~~~;i$:~~~~:::=-~-~-y-~-~~:$::~~~~
~~i)::':::::::::< ::::-:.:~:~:;:.:::::::. ::::.:.:::::::::;:::;::: ";"':::"::;:;:;:::::::::':::::::::::;:;:;:::::;:;:;' - L--'=-1
1 closed tie
(MKA) #7 8"

Jt
3@3"

#4 closed
tie (V)
6"
2 - #4 bent bars

is''

Section: A • A
~
2 closed ties #7 - MKA
closed ties #4
1 closed tie #7 - MKB

-r1
3@2" 7"

~l 1 5"

#4 bent bar pair

/--3@3---!
1.5" 1.5"
/ 1.. . - - - - - - - 2 7..- - - - - - - -..~1

Figure 4.18: Reinforcement layout

 205

4.3.3 Deep beam with a hole

A deep beam is a beam in which a significant amount of the load is carried to the
supports by a compression thrust joining the load and the reaction rather than
through flexural action. This occurs when a concentrated load acts closer than
about 2 d to the support, orfor uniformly loaded beams when the span- to- depth
ratio, 111/ d ll , is less than about 4 to 5. The ACI Code [4] speci'fies that deep beam
action must be considered when designing for flexure if III / d ll is less than 5/2 for
continuous spans or 5/4 for simple span (see also Ref. [105]).
Cook and Mitchell [113] did some experimental verification and non liner finite
element analysis of a uniform loaded T-beam with a hole in the web. The finite
element analysis by Schlaich et al [28] and the experimental study by Cook and
Mitchell [113] show that tensile forces are acting especially around the corner.
This leads to the assumption that a diagonal tie has to be placed in the
discontinuity zone in the tension region. The stress concentration factor for an
infinite plate with a rectangular hole and subjected to biaxial stress is the highest
such factor for all different forms of openings [119]. For a finite-width member
with infinite thickness and a circular hole under biaxial tension the stress
concentration factor is given by Ling [120] as:

K = 12/(7 - 5 v)
v = 0.16 for concrete
K = 1.93
Forthecorresponding case of a semi-infinite body, Tsuchidaand Nakahara[121]
developed stress concentration factors. The values with Poisson's ratio of 0.25
and r = radius of circular hole and m = distance from center of cavity to surface
are:

rIm = 0.5: K = 2.32

rIm = 0.8 K = 3.3
For rectangular openings the following mathematical results, with specific data
obtained by computer, have been published: 0'1 = (J2
206

tension: 0'1

I
~r

tension: 0'2 - a

b=a; rib = 0.1 K = 4.88

b=a; rib = 0.3 K = 2.76

Figure 4.19: Geometry for determining stress concentration factors

In addition, for b/a = 1/2, the stress concentration factor reaches the low value
of 1.5 for the ellipse. In general it has been shown that the outside fiber at the hole
has at least approximately 2 times higher stresses. If the radius of the rounding
becomes smaller at the corners of rectangular openings, K increases rapidly.
 207

Schlaich et al [28] proposed that, for this kind of problem, two separate
strut- and- tie- models should be developed, each with a carrying capacity of 50%.
One model should follow the elastic principal stress trajectories with a diagonal
tension tie and the second model should have strut and ties parallel to the
borders. From a practical standpoint it is very inconvenient to place diagonal
reinforcement in many structures.
Test results from Shah [122] and Gaynor [123] for tests on reinforced
concrete in-filled shear walls with openings gives some indication that first
cracking appears near the openings. In orderto prevent large cracking and for
crack control under service load it seems reasonable to round off the corners
(see Fig. 4.19: rib"", 0.3) and for geometrical discontinuities subjected to biaxial
tension a quantity of diagonal reinforcement equal to about 118 of the orthogonal
reinforcement should be provided as an addition. In the general literature on
design [62, 105, 113, 124] no detailed information about the required amount of
diagonal reinforcement is given. As shown in Fig. 4.20, such diagonal reinforce-
ment follows the principal tensile stress directions closely and should be very
effective in controlling the reentrant corner crack width.

T t = 1/8 (max T, or Tt ) principal tension trajectories

Figure 4.20: Principal tension trajectories and reinforcement for corner in tension
208

Example 4.3: Deep beam with a hole

Design a deep beam with an overall depth of 16'8" and an overall length of
27'6" to transfer a vertical load of 500 kips applied 14'7" from the left edge.
The wall is supported on simple supports located 25'0" on centers (see Fig.
4.21 (a)). The supports are 20" X 15" columns and the wall thickness is 15".
There is a large hole 65 in. square located in the lower left corner. The hole
has 30 in. of concrete below it and its left edge is 30 in. from the center of the
left support. Concrete strength is 7000 psi and Grade 60 reinforcement is
used.

Load and dimensions:

F = 500 kips
I = 300 in.
f = 30 in.
j = 65 in.
h = 200 in.
e = 30 in.
k = 65 in.
m = 160 in.
a = 20 in.
b = 15 in.

External forces

B = 267 kips
A = 233 kips
 209

In developing a strut- and- tie model for this type of unusual structure,
it is very use'ful to consider the elastic stress pattern indicated by a finite
element analysis. A general pattern for a similar type of problem with the
load located considerably more to the right has been provided by Schlaich
et al [28] and is shown in Fig. 4.21 (b). Based on the pattern of elastic
stresses shown, it can be seen that significant tension acts in the diagonal
direction at the upper right corner of the opening and lesser tension acts on
the diagonal at the lower left corner of the opening. The thrust to the left of
the opening is skewed slightly to the right, inclining towards the opening's
upper left corner. Schlaich et al [28] have suggested two possible strut- and-
tie models for the left side of the structure as shown in Fig. 4.21 (c) and (d).
They suggest the left reaction be considered as split 'on a 50-50 basis
between these patterns. One minor problem with these suggested models
is that there is no tie reqUired beneath the opening. Inclining the thrust
towards the upper corner of the opening in the section to the left of the
opening would result in a tie requirement beneath the opening. The part of
the wall to the right of the applied load is basically a straightforward case
with the thrust being transferred by a bottle-shaped strut to the reaction at B
and the lateral component of the strut force tied back by a lower tie. For this
particular example, variations of the models suggested by Schlaich et al
were adopted. For the portion of the wall to the left of the centerline of the
applied load, it is assumed that the load carrying capacity will be shared
equally (50-50) by the Strut- and- Tie Models shown in Fig. 4.21 (e) and Fig.
4.21 (f). The geometry of the models and the resulting strut compressions
and tie tensions are shown on each figure. Note that in Fig. 4.21 (f) C8 was
assumed as a compression strut but in the solution (performed using a
microcomputer program for a 2D truss based on SAP) was found to have a
very low level of tension. Similarly TIl was assumed as a tension tie but
analysis indicates a small amount of compression. While the two models to
the left could be combined, it is easier to proportion reinforcement using the
two separate models. The much simpler section to the right of the load is
shown in Fig. 4.21 (g) with the right reaction, 267 k, and an equal part of the
load applied to a bottle strut and major tension tie.
210

Strut and tie angle

z :::::; 7/8 h = 175 in.

Tan <j)b= 175/135 = 1.296
0
~b = 52.35
alh = 10/200 = 0.025
12 + 3/-V(a/h) ~ 25
0
81 =
81 = 12 + 3 / -v(0.05) = 25.4 0
~ 25° Use 25°

Internal forces

Cb = B/sin <j>b = 267 / sin 52.35° = 337.2 kips

Cb/(2 cos 81 ) = 186 kips

C14 =
(Cb/2) tan 81 = 78.6 kips
T14 =
T13 = Cb cos <\lb = 206 kips
 211

a =20"
I'" ~1500k
15"
. m= 160" 140"
. 15- .
-I ,
I -, .11 III

A t B
....... 1 I- "I
a = 20" b = 20·

1=300"

(a) Example 4.3 Dimensions

Figure 4.21: Strut- and- tie models for deep beam with a hole
212

, ,
• ,. ,
\ ,
... .. "l- I I I
.,. ,.
-. -. ",. ••
.... .... ... •
.. '1l -.. .... ... ...
'I. If.. -...
.,
~~

"
• .

--
"-
.. ...."
-
Principal Compression
- - - Principal Tension

Fig. 4.21 (b) Finite element analysis contours for similar structure with load
placed farther to right (From Ref [28])

Figure 4.21: Strut- and- tie models for deep beam with a hole
 213

A1 =O.SA

q ..
/// C1 =O.SC
/ /
/ /
---e( /
:lEi' I'P
TS' I

A1 = O.SA

Fig. 4.21 (c) Modell, left side (From Ref [28])

Figure 4.21: Strut- and- tie models for deep beam with a hole
214

A2 = O.5A

_ - -0 1-_.._-
8° - - - -- -- --- --- I C2= 0.5C
. .~--- I
I
I
I
I
I
I
I
I
T9 I
I
I
I T2 = 0.5T

A2 =0.5A

Fig. 4.21 (d) Model 2, left side (From Ref [28])

Figure 4.21: Strut- and- tie models for deep beam with a hole
 215

C1 = 104.1 k
o=
C2 = 110.5k T'-

C3 = 120.ak
T1 = 134.9 k
T2 = 25.6 k

/
.:' C3

i
/.,
..
•
lo
<.0

'----_6_5"_......1

T2

155"

Fig. 4.21 (e) Model 1 - 50% of load (left)

216

116.3k
Cl= 59.40
B = 13.80

Compression
C4 = 116.7k
C5 = 164.5k
C6 = 123.7k l£
---
"'
---------- ----68
R
C7 __ -6
.Q.
::.-
Cl
:
C

1.80
=103
k
Tension
T3 = 103.0k
T4 = 103.0k
T5 = 103.0k
.
. T11 ••••.
.. .
C7 = 111.3k T6 = 206.0k
C8 = -8.3 k (T) o
. .
. ,
, T7 = 203.7k
C9 = 154.4k T8···••99 T8 = 206.9 k
...
..·C5 T12 •••·66 : C10
C10 = 91.5k ... ... T9 = 203.8k
C11 = 145.7k .... ·~450 T10 "-\450 T9 T7 ••••..:. T10 ~ 116.3k
C12=103.0k ,
,
o , ~ T11 = -9.0 k(C)
C13 = 145.7k
.' , ,.... T12 = 87.5k
• T6

~ ...
.'
0"

... C11 : C12

,

,
,'C4
•' T5

.....
.-
,I
T4
450
-)".'
.'
.
/C13
o

T = 103k
•,, .' T3

116.3k
155"

Fig. 4.21 (f) Model 2 - 50% of load (left)

 217

m
IF = P + F" = 500k
F' =233k t F" = 2S7k
NodeC
I II I

.. ~~ Cb = 337.2 k
C14 = 18Sk
',\,
,."
' \'C14 C15 = 1S8.6k
T13 = 20S k
, '\, \Cb'
' T14 =78.S k

'X\
\
\
.\ \\C15
\
\
z h = 200"

\
r \ \

~~ ,
. ,
\

T13
'~\
,.
$b'~
I I I
•N
\
A Node b 8=2

f j 135"

Fig. 4.21 (g) Model 3 - 100% of load (right)

Figure 4.21: Strut- and- tie models for deep beam with a hole
 ....N
00

F = F' + F" = soak

C1 _ \,

-- --
_ - - C7
-----
--- 1 \ "'\ .....
I
I
\ , .....C14
\ '\ , '
T11C / J~C2 ~ C10 ..........
C6.1/ I \ '\ '
/ I \ ''\ \
T8 I \C14 , T14 \C15
I \ . \
T7 "\!.. \ '\. \\
\ \
J
,
I
, I
I 10 / /
/C11 :~C2+C12
.
I
'C15/ \\
\\ \ T14

/ \
,
C4'
I
I'C3
TS
/1
I \
\
,'\
\
'\
\
\C14
I
I I T4 / I C2 ' .Q 14 \, \
I '
II /C13
,// ,I '..... ..........'\\\'
I.' .I I T13 '....:
T2 T2+T3 \

A= 233k B = 267k

Fig. 4.21 (h) Combined strut- and- tie models

Figure 4.21: Strut- and· tie models for deep beam with a hole
 219

Reinforcement for ties: (All Grade 60)

T13 = 206 kips As = 206/60 = 3.38 in. 2

Use 6 - #7 = 3.60 in. 2 in lower right section (See Fig. 4.22)
T14 = 78.6 kips X 2 = 157.2 k
T14v = T14 cos <j>b = 157.2 cos 52.4 0 = 96 k
Asv = 96/60 = 1.60 in. 2
T14H = T14 sin <j>b = 157.2 sin 52.4 0 = 125 k
Ash = 125/60 = 2.08 in. 2

The vertical reinforcement, 1.60 in. 2 , needs to be spaced over a width of

approximately 11 ft. or approximately 0.15 in. 2 per foot.

The horizontal reinforcement, 2.08 in. 2 , needs to be spaced over a height of

approximately 14 ft. or approximately 0.15 in. 2 per foot.

Both of these values are quite low. The ACI Building Code Sec. 14.3
prescribes 0.0012 for minimum vertical wall reinforcement percentage and
0.0020 for minimum horizontal wall reinforcement percentage based largely
on shrinkage and temperature considerations. Using the average would
indicate minimum reinforcement

As = 0.0016 (12) (15) = 0.29 in. 2/ft.

If this reinforcement is placed in 2 curtains, it would require 0.15 in. 2 / ft. in

each curtain. This would be nicely furnished by #4 at 16 in. spacing.

Use - 2 curtains at #4 spaced horizontally and vertically at 16 in. (See Fig.

4.22).
220

Anchorage requirements:

#7 bars for T13 ties at node b (Fig. 4.22) (CCT node)

The 6 - #7 bars required can be efficiently placed in 2 layers of 3
bars each with a clear cover of 2 in., a clear spacing greater than 2 in.
and the vertical #4 bars bent below the #7 bars. This results in 2-1/2 in.
of concrete below the bars. From Fig. 3.12 and the dimensions shown
in Fig. 4.21 (a), assuming 2 in. clear cover over the tails of the #7 bars,
w1 = 20 in. and the length available to satisfy Id requirements is
w1 + 5 - 2 = 23 in. From ACI 318-89 Ch. 12, Idb = 0.04 Ab fy ...}f'c =
(0.04) (0.60) (60,000) / ...}7000 = 17.2 in. In a 15 in. thick wall with
clear cover of 2 in., two #4 bar vertical curtains, and three #7 bars in a
layer, the cover is 2-1/2 in. and the clear spacing is 3.7 in. which is
greater than 3db. Therefore, a factor of 1.0 is used.

Id = 17.2 in. < 23 in. available OK

No hooks required. Anchorage for the other end of these bars at nodes
(a) and (c) (Fig. 4.22) will be checked as part of the left portion.

#4 bars for T14 ties

The required Id for these bars is short. Idb = (0.04) (0.20)
(60000) / -./7000 = 5.7 in. Since cover is 4 db and spacing is 32 db
in one direction and 20 db in the other, Id = 1.0 Idb = 5.7 in. > (.03)
(0.5) (60000) / ...}7000 = 10.8 in. which governs here. Clearly there is
no problem along the top or right edge of the wall. As a good detailing
practice to provide confinement for the main tension tie, the vertical #4
bars should be U type hairpins and enclose the #7 bars in the main
tension tie (see Fig. 4.22).
 221

Check Node (b) (See Fig. 4.21 (g)) - CCT Node

From Figs. 3.12 and 4.21 (g), w1 = 20 in., db = 0.88, n = 2, s = 2 in.,

and $ = 52.4°
wT = 2 (0.88) + (1) (2) = 3.76 in.
w2 = (20) sin 52.4 + 3.76 cos 52.4 = 18.1 in.
O"cb = Cb / (w2 * b) < V e f'c
acb = 337.2/ (18.1 * 15) = 1.24 ksi < 0.725(7) = 5.075 ksi OK
0"1 = B / (a * b) < Ve f'c [acb / ab]O.5
= 267 / (20 * 15) = 0.89 ksi < 0.725 (7) = 5.075 ksi 0K

Check right part of Node (c) (See Fig. 4.21 (g)) - CCC Node

From Figs. 3.9 and 4.21 (g), $3 = 37.6° , $3' = 62.6° , $3" = 12.6° ,
= (267/250) (10) =
a = 20 in., a' 10.68 in.
Checking for w4 = a/2 = 10 in.
C4 = Cb * cos $3 = 337.2 cos 37.6 = 267 k
0-4 = c4/ (w4 * b) ~ Va f'c
= 267/ (10 * 15) = 1.78 ksi < 0.725 (7) = 5.08 ksi OK

Check strut cb at y = a = 20 in.

a3 = Cb / [(w3' + w3 + w3") * (b)] < 0.6 V e f'c

= 267 / (22.02 * 15) = 0.80 ksi < (0.6) (0.725) (7) = 3.05 ksi
OK
222

Left portion of Wall (Fig. 4.21 h)

Reinforcement for Ties (All Grade 60)
The two strut-and-tie models, which were proportioned to share the load on
the left part of the structure equally, are superimposed and combined on Fig.
4.21 h to expedite choice of reinforcement. Several of the members in each
model are concurrent or nearly so (i.e., 12 +T3, C2 +C12, etc.). In 'these
cases the required forces can be added and a single tie proportioned to take
the total load. The following tie reinforcement is required:
Tension Tie Force, K As req'd, si As provided, si
T1 135 2.25 4 - #7 = 2.40
T2 26 0.44 2 - #7 = 1.20*
T2 + T3 129 2.15 4 - #7 = 2.40
T4 103 1.72 6 - #7 = 3.60
T5 103 1.72 6 - #5 = 1.86
T6 206 3.44 6 - #7 = 3.60
17 204 3.40 6 - #7 = 3.60
T8 207 3.45 6 - #7 = 3.60
T9 204 3.40 6 - #7 = 3.60
T10 117 1.95 4 - #7 = 2.40
T11C 9C 0
T12 88 1.47 6 - #5 - 1.86
* Two other #7 bars required for T2 & T3 are carried to support A to avoid
bar cutoff in tensile zone complications.
In satisfying these requirements, bars may be provided which
simUltaneously cover several cases. For instance T4, T6 and T8 are best
satisfied by a single set of vertical bars. The maximum of these forces
requires 3.45 si which can be nicely satisfied by 6 - #7 = 3.60 si. Similarly,
the 6 #7 bars chosen for T13 are more than adequate for T2 +T3. Four are
required to be bent up for T1 and two of them can continue to support A
satisfying T2 amply. They are supplemented by 2 #7 in the T2 - T3 zone
which must be lap spliced to the T13 bars.
 223

The horizontal tie above the opening (T7 - T9 - T10) is nicely satisfied by 6
#7 with two of them bent down across the corner of the opening to cater to
any tensile stress raisers at the reentrant corner. Lastly, the tensile ti~s T5
and T12 require smaller #5 bars well distributed overthe tie zones. Again to
counteract very local tensile stresses similar to those shown in Fig. 4.20 at
the lower corners of the opening, inclined #4 bars are placed on the
diagonal at each of these corners. All bars are arranged as far as possible
in two curtains of reinforcement with a minimum of 2 in. of cover. Final bar
patterns are shown in Fig. 4.22. Note that while no tension reinforcement is
required in the C3-C4 strut areas to the left of the opening, minimum
reinforcement for columns (reduced to 1/2% to reflect that the section is much
larger than required for the compression load) is provided to control time
dependent deformations and for general ductility. As = (0.005)(30)(15) =
2.25 si. Use 4 - #7.
Anchorage Reguirements
When checking the right side of the wall it was determined that for #7 bars, Id
= 17.2 in. At support A at least 23 in. > Id is available. The smeared nodes
at the right ends of the T5 and T7 ties require only normal Id embedment
past the node, as does the upper ends of the T12 and T8 ties. In the more
critical cases at the left end of the T10 and T5 ties and at the bottom of the T4
tie, positive anchorages are provided by looping the ends of the bars. The
orthogonal curtains of #4 bars provided for the T14 ties require no further
check as they have ample length to satisfy Id requirements.
224

Check Node (a) (See Fig. 4.21 h) CCT Node.

From Figs. 3.12 and 4.21 h, WT = 20.88 in., W1 = 20 in., and 4>b = 75 0

(Conservative Assumption) w1 = 20 sin", + 20.88 cos", = 24.7

crca = (C3+C4) / w1 b = (120.8 + 116.7) / (24.2)(15) = 0.64 ksi « (.725)(7).

Check Node (d) [Intersection of T1, T2, C2, T13] CIT Node. Detailing
continuous bars has satisfied T requirements. C2 has a force of 11 Ok and
must be basically equilibrated by the bends of the 4 #7 bars in the T1 tie.
For this 55 0 bend, a standard inside bar diameter of 6 db would result in a
bend contact area of approximately (7) (0.88) 1t/4 = 4.84 in. for each layer. An
extremely conservative estimate of node pressures would be:

crcd = c2 / w1 b = (11 O) / (4.84)(15) = 1.52 ksi < (7.25)(7) = 5.08 ksi

Note that all node cases are far from critical.

 .. .
Bars shown as
I
I
.. 67.5" I
I
52.5" .. ..
I
I
55" . I
2 x _ indicate
one set in each
face curtain
r-

~2x3-#7 16"
Looped at ~
2@8"--- I......-

5" 2x3-#5_
'l\
r\ - _2@3"
bottom 1
16"

~-
I--
#7 /
~nt 2@5"

, l V"
2 x 1 -#7 Bent
...;...
/
/ \.. \..
\. - I
)-2 x 2
I
#7
I

V ., ~t
/ '~ 2r'

2x 2#7
2X1-Q
" Looped
\. \.
~~ ,
w/#3 Column
Ties @ 14" clc

!!;: /"'-- /
<C~'\
~ t-
"-2 x 3 -#5
Looped at left

2 x 2-#7
Bent up ,
62.5"
E

7 '\ /
p

Figure 4.22: Reinforcement layout for deep beam with a hole N

N
U'I
226

4.3.4 Dapped end beam

Dapped beams have applications in both building and bridge structures.

Flexural reinforcement may be prestressed. A typical application for dapped end
beams is as drop-in spans for semi continuous beams. The load distribution may
vary significantly. The aspect ratio of the dapped end may differ depending upon
the specific application. Barton [7] did some experimental veri'fication on dapped
beams designed with different models. A total of four details were tested using
two beams. Each detail was designed for an ultimate load capacity of 100 kips
and a concrete compression strength of 5000 psi. The different design models
were:
strut- and- tie- model
modified strut and tie model
PCI design method
Menon/Furlong design procedure
The reinforcement layouts for the different models are shown in Fig. 4.23 to Fig.
4.26. All four specimens cracked at the reentrant corner at load levels ranging
from 20 to 33% of ultimate load. Maximum crack widths at an approximate service
load of about 60 percent of ultimate were in the range of 0.009 to 0.012 in. The
control of the diagonal crack at the reentrant corner is of primary concern for
serviceability. Based on the performance of the PCI - detail (grouped vertical tie
reinforcement) and the modified strut- and- tie- model (about 37% more horizon-
tal and 50% less vertical reinforcement), it appears most efficient to place the
location of the vertical and horizontal tie reinforcement as close as possible to
the interface between the dap and the full section. In addition, the vertical tie
reinforcement shou Id be placed in a closely spaced group. Also the corner should
be rounded in order to decrease the stress concentration. Studies by Cook and
Mitchell [113] using both rectangular and inclined corners in dapped end beams
indicate that the inclination helps to prevent cracks at the reentrant corner. The
two most common strut- and- tie- models for dapped end beams are shown in
Fig. 4.27 and Fig. 4.28. The selection of a particular model is a compromise
between ease of fabrication and fidelity to the elastic principal
stress
 227

~---ttt+-- ~. •

r-ffir-t-- 2 IIOVS OF • " lIARS

IM~~~~+2.
L....:----.-------;------.,................ 2'
,.1,,:111--,. -1--
1
-. -sP-ACE-S-1--'-'-sP-.-a-s- -
sPACES

First cracking load: 19% of predicted failure load

Ultimate load: 147 % of predicted failure load - anchorage failure of the
stirrups and concrete compression failure

Figure 4.23: Reinforcement layout for strut- and- tie- model ST1
(from Ref. [7])

, " lIARS

~~~=*==fl:===:j~I""2'
2- 'f.- t=='

2 IIOV$ aF • " _5

First cracking load: 28% of predicted failure load

Ultimate load: 152% of predicted failure load - concrete compression failure

Figure 4.24: Reinforcement layout for PCI deta.il (from Ref. [7])
228

~ . ",.

t:='

I~~~~~"
$TAAP A$$'T

!-.,......J:::::::::l.---r----...,------:-'- 2'

I' sPACE$

First cracking load: 22% of predicted failure load

Ultimate load: 121 % of predicted failure load - concrete compression failure

Figure 4.25: Reinforcement layout for Menon/Furlong - detail

(from Ref, [7])

,.",'

~~~~,.
-lh-----r--~---,-'--- "
I' "'ACt$
2'

First cracking load: 29% of predicted failure load

Ultimate load: 127% of predicted failure load concrete compression failure

Figure 4.26: Reinforcement layout for modified strut-and-tie model ST2

(from Ref. [7])
 229

directions. The orthogonal model results in a reinforcement pattern which is easy

to place and is well suited to the overall reinforcement pattern of the beam.
However, the model forces the loadpath to deviate substantially 'from the elastic
stress directions. This is also the reason that Schlaich et al [2] propose to place
diagonal reinforcement as shown in Fig. 4.28 (should carry less then 70% of the
overall capacity). The diagonal reinforcement is more difficultto place and anchor
properly but follows the elastic principal stress directions closely.
The quantity of horizontal reinforcement anchored at the bearing plate in
Fig. 4.27 is determined based on the angle of the inclined compression strut.
ObViously, selecting a steep angle reduces the required amount of horizontal
reinforcement. Based on force measurements by Barton [7], the compression
strut angle ranged between 45 and 55 degrees and tended to increase as load
was increased close to the ultimate load.
Anchorage of horizontal reinforcement within the dapped end may be
provided by welding a portion of the reinforcement to the bearing plate at the
bottom and by the use of continuous hoops. Of interest also is the anchorage of
the other end of the horizontal reinforcement. Test results by Barton [7] indicate
reinforcement outside of the second vertical tie developed significant force only
for load levels greater than the design load. In both specimens ST1 and ST2,
the measured forced in the second set of vertical ties, corresponding to T4 in Fig.
4.27, were very low at the design ultimate stage. This was probably due to
substantial tension carried by the concrete and the effects of other contributions
such as aggregate interlock and dowel action. However, by the actual ultimate
load on both specimens, the measured second vertical tie force in both speci-
mens had achieved the level predicted by the models. Since the STM is a lower
bound plastic model, this fulfills all assumptions. Different strut- and- tie- models
were compared with test results from [7] to evaluate a design approach. The
recommended strut- and- tie- model is shown in Fig. 4.29 and Fig. 4.30 and the
statistical data comparing the test results with the computed ultimate loads using
strut- and-tie models is given in Table4.2 and in Fig.4.31. In Report 1127-1 [131]
it was noted that the strut angles in the test specimens ranged between 45° and
55° and tended to increase as load increased. Hence, an angle of 55° is used in
the recommended model.
230

T
i

1
:r2 =T4 = A

Figure 4.27: Orthogonal strut- and- tie- model (from Ref.[2])

T i

1
T1 = AI sin ep1

Figure 4.28: Diagonal strut- and- tie- model (from Ref.[28])

 231

Dapped end beam: Strut- and- tie- forces

T
i

1
Strut- and- tie- forces:

Ca=A/sln c>a T1 = A/tan ¢a

C1 = Ca cos $8/ cos ell1 T2 =Ca sin I\la - C1 sin ¢1

C2 =121 sin 92 T3 =T1 + C2 cos lj>2

C3 = (C1 cos $1 + C2 cos $2) 1 cos ~ T4 = C1 sin 91 + C2 sin ¢2 - C3 sin $3

C4 =T4/sln~ T5=T4/tan ¢4

C5 = C3 cos ~ T6 =T7 + C4 sIn ¢4 - C6 sin ¢46

C6 =(C4 cos ~ - T3) 1cos 96 T7= C3 sinlj>3

C7 =T61 sin 97 T8 =T5 + C7 cos 97

C8 = C5 + C6 cos cps + C7 cos ¢7 Control: C8 = T8

Figure 4.29: Proposed strut- and- tie- model for dapped end beam
 232

Dapped end beam: Strut- and- tie- angle

1 - - - :- ; a - - - - - + - - t

Proposed strut- and- tie

angle:

Proposed geometry:

lIa~1.15

Figure 4.30: Proposed strut- and- tie- angle for dapped end beam
 233
2.0
=
c concrete crushing -0- experiment/theory
=
t test was stopped ......- experiment/theory =1.0
~

o ~ 1.5 III"
~-
+:i0
~ Q)
.J::.
"'C ...
co-
°e
Q) Q) 1.0 --
"'E
co._
E Qj
+:io..
-x
:::> Q) 0.5
specimen 1 =strut-tie model ST1
specimen 2 =P.C.1. - detail
specimen 3 -Menon/Furlong - detail
specimen 4 =modified strut-tie model ST2
.0.0 1 2 3 4
specimen

Figure 4.31: Comparison with the proposed strut-and-tie model using test results
from Barton [7].

Table 4.2: Statistical data from Fig. 4.31

X, : Column 1

Mean: Std. Dev.: Std. Error: Variance: Coef. Var.: Count:

1.37 .13 .37 --- 0.09 4

Minimum: Maximum: Range: Sum: Sum Squared: # Missing:

1.21 1.52 .31 5.47 7.55 a

234
Example 4.4: Dapped end beam
Design the end region of the dapped end beam shown in Figure 4.32 to carry a
concentrated midspan load of 300 kips. Dead load of the span may be neglected in
this example. Dap details are shown in Figure 4.32. Use f: = 6000 psi and Grade
60 reinforcement with a minimum cover of 2 in.

Load and dimensions:

(See Fig. 4.29 and Fig. 4.30 for symbol definitions)
F = 300 kips
= 200 in.
h = 30 in.
f = 5 in.
= 14 in.
= 16 in.
a = 10 in.
b = 15 in.
e = 5 in.

The beam may be divided into D regions near the ends and a central B region (a D
region may be used under the load but is not checked here). The B regions will be
more efficiently handled by ordinary section design procedures.

External forces:
A = B = 150 kips

Strut and tie angle:

Where the D region meets the B region, estimate the effective lever arm z as:
z = 3/4 h = 22.5 in.

From Fig. 4.30:

¢a =
¢2 = 45° 235
¢4 = 45°. From the geometry of Figure 4.30, one can compute
¢2 = Arctan r( 2.39 + 8.33 ) / 8.33 ] ::; 52.1 °

¢3 = Arctan t 11.75 - 10.72) / 11.75 ::; 5.0°

Internal forces:
Ca = 150 / sin 55° = 183.1 kips
C1 = Ca cos ¢a / cos ¢1
C1 = 148.5 kips
To find t/J2

,~

""
,',/
I'
,,,","
,

8.33 Tan 45 = 8.33

," " ,,"
, I'
,
,

4&',' , ,,"
,
,
,,
~'.67T,"55=2-" 2.39 "
" ~8.33
al6 =1.67 aI3 +e = 8.33

Tan clI2 = (8.33 + 2.39) I 8.33

To find t/J3

16"
5
 Example: Capped end beam:

236

[ ~.region I
A B· region
I D.regiO~B
1_~_1_.5_h~ 1= 200 II ~_1_.5_h"_~1

0- region

T 16"
. r---
T7
-cs----

;;
......
""
"
...... ......

T2 ,C,!/" .C6 "

I

'C2 ... -;,""\... ~ ,/

1 I ~~;~,r-T3 " --j /
T9

T H ""

li :l__"
T4 C4 " T6 ",,"

~
14"

~O" L21"_<I>4
(b)
__T5 • -----TB----~,.·~ --'
.5"+6"
..-.----D·region:
I.... 1.5 h = 45"

Figure 4.32: Strut- and- tie- model for example: dapped end beam
 237

T1 = A / tan <1>a
T1 = 105 kips

T2 = Ca sin <1>a - C1 sin <1>1

T2 = 45 kips

C2 = T2/ sin <1>2

C2 = 57.0 kips

C3 = (C1 cos <1>1 + C2 cos <1>2) / cos <1>3

C3 = 140.5 kips

T3 = T1 +C2 cos <1>2

T3 = 140.0 kips

T4 = C1 sin <1>1 + C2 sin <1>2 - C3 sin <1>3

T4 = 137.7 kips To find <1>6
C4 = T4/ sin <1>4
C4 = 194.8 kips 30 - 3 - 3.5 =23.5 1
C5 C3 cos <1>3
---------0
/
= <\>6/
C5 = 140.0 kips ,<
/"---
/
T5 = T4/ tan <1>4 T7 /
T5 = 137.7 kips
11.75" 90 - <\>6 <"
/

T7 = C3 sin <1>3 7\
T7 = 12.2 kips //<1>6 .
(i-------L
'-../

<1>6 = arctan (11.75/ 23.5) = 26.6°

C6 = (C4 cos <1>4 - T3) / cos <1>6

C6 = -2.5 kips

T6 = T7 + C4 sin <1>4 - C6 sin <1>6

T6 = 151.1 kips
238

C7 = T6/ sin <1>7

C7 = 213.6 kips

T8 = T5 + C7 cos <1>7
T8 = 288.7 kips

C8 = C5 + C6 cos <1>6 + C7 cos <1>7 = 288.8 kips (o.k.)

Check C8 =T8 by method of sections

T8 = C8 = 150(1.67 + 8.33 + 11.75 + 23.5) /23.5 =289.5 kips

Check OK
T9 = C6 sin <1>6 + C7 sin <1>7 = 149.9 kips
From method of sections T9 =A = 150 Check OK

In practice, once geometry is defined a simple truss program on a computer

could be used for these calculations.

Reinforcement for tensile ties:

All bars Grade 60

T1 = 105 kips
As = 1.75 in 2 :s; 6 - #5 = 1.86 in. 2

T2 = 45 kips
As = 0.75 in 2 :s; 4 - #4 = 0.80 in. 2

T3 = 140.0 kips
As = 2.33 in 2 :s; 8 - #5 = 2.48 in. 2

T4 = 137.7 kips
As = 2.30 in 2 :s; 12 - #4 = 2.40 in. 2

T5 = 137.7 kips
As = 2.30 in 2 :s; 4 - #7 = 2.40 in. 2
 239

T6 = 151.1 kips
As = 2.52 in 2 :::;; 14 - #4 = 2.80 in. 2

T7 = 12.2 kips
As = 0.20 in 2 :::;; 1 - #4 = 0.20 in. 2

T8 = 288.7 kips
As = 4.81 in 2 :::;; 8 - #5 + 4 - #7 = 4.88 in. 2

T9 = 150 kips
As = 2.50 in 2 :::;; 14 - #4 = 2.80 in. 2

A possible bar arrangement is presented in Fig. 4.33. Note that the T6 bars
are run full height and provide substantial excess for the T7 bars. Note a.lso
that the #4 bars required for T6 and T9 are provided as Groups of c10sed 3 -
W stirrups and 1 - U stirrup. The U stirrups provide a desirable transverse tie
completely across the bottom flange. Where W stirrups are provided a short
U is desirable on the bottom flange. (A W stirrup is a four-legged stirrup

CJLJ)
240

Anchorage requirements:

The very congested conditions in the dap make it difficult to effectively

anchor all the reinforcement. The 8 #5 bars provided for the T1 and T3 reinforce-
ment should have positive anchorage by closed loops at the support. In addition,
the lower layer should be welded to the bearing plate, if possible. The minimum
development length for a Grade 60 #5 bar under ACI 318-89 provisions would be
14.5 in. There simply is not room available in the small dap to rely on develop-
ment length, unless one relies on the confinement present, since the T1 force
must be fully developed above the bearing plate which is only 10 in. long.

Similarly, the #7 bars provided for the T5 force need to be developed

within the CTT node at the bottom corner of the full depth section. This is also a
very congested corner. Under ACI318-89, a standard #7 hook would have a
basic development length of 13.5 in. However, the 2" clear cover over the #4
stirrups provides 2-1/2 in. side cover so that a mUltiplier of 0.7 may be used
reducing the length to 9.5 in. However, the highly grouped T4 reinforcement
greatly reduces the width w1 of the node (see Fig. 3.14). It is highly desirable to
space stirrups at no more than 3d b = 2.64 in. throughout the hook development
length to allow use of an additional 0.8 factor reducing Idh to 7.6 in. This can be
easily done by using the 3 stirrups required for the T4 tie at 1-1/2 in. on centers
and then introducing 2 extra confining stirrups at 2-1/2-in. spacing. One of these
can be counted towards the T6 tie force so only one additional stirrup is required.

Development of the other bars is routine.

Concrete stresses at node zones:

Concrete strength: flc = 6000 psi

Concrete efficiency factor: v = 0.75

The only critical appearing nodes are at (a) and (b) as shown in Fig. 4.32. The
other nodes have substantially more area for node development.

Check node (.91: CCT - node See Fig. 3.12 for typical geometry
based on proposed bar arrangement
of Fig. 4.33
 241
In this case because of the special strut-and-tie arrangement caused by the
angle change <l> , shown in Fig. 4.30, it would be unconservative to use the full
bearing plate width, a, for w1, as shown in Fig. 3.12. Instead a width al2 + al6
=2/3 a will be used for w1 since that should be. 'fully effective in developing the
T1 force shown in Fig. 4.30.

w1 = 2/3 (10) = 6.7 in.

wT = 2(0.625) + (2 - 1) (1.25) = 2.50 in.
wa = w1 sin 55° + wT cos 55° = 6.92 in.
(wa is the same as w2 in Fig. 3.12)

<>ca = Ca / (wa * b) :::;; ve flc

<>ca = 183.1/ (6.92 * 15) = 1.76. ksi :::;;0.75 * 6. =4.5 ksi OK

Check node (Ql: CIT - node See Fig. 3.14 for typical geometry
based on proposed bar arrangement
of Fig. 4.33

w1 (T4) = 3.5 in. (grouped #4 stirrups 3(0.5) + 2(1.0»

w2 (T5) = 1(0.88) + 2(0.62) + 2(1.12) = 4.36 in..
wc = w1 sin 45° + w2 cos 45° = 5.55 in.

<J'c4 = C4 / (wc4 * b) :::;; Ve flc

0'c4 = 194.8/ (5.55 * 15) = 2.34 ksi :::;;0.75 * 6. = 4.5 ksi OK

242
3 - #4 bar
2 #4 bars
2" 3" 3" I /

/ I
I

I
I III

I
#4 stirrups f-.
8 #5 bars wI looped
@3"
"-
r- ends at support -
#4W stirrups

&
3@4- (T6)-
~ / ~
>
v
T 2@2+ ~ #4W stirruos

aring~e
J
~ 3@S" (T9)
(T6)
welded to be
• I I
I
-1'"
\
-
3 grouped #4W
- M'_ •
I
I
\
\
- • \=
\
-
4#7 bars 8#5 bars
stirrups 1 on I" vertical hooks straight bars
centers
(T4)

Section: A-A

3"

T 13~
4
"

30" ~ 2"

T
.....-1--#5 bar
_ _-J--._ #5 bar

#7 bar
2.9"

~--15"--..,

Figure 4.33: Reinforcement layout for dapped end beam

 243
4.3.5 Anchorage Zone:

In post-tensioning concrete, the tensile force of the tendon is introduced

into the concrete by means of end anchorage devices. Normally, the size of the
anchorages is limited by the structural member size. This results in very large,
highly concentrated forces which have to be introduced into the concrete.
Anchorage devices transfer the posttensioning forces 'from the tendon wires,
strands or bars to the concrete. Anchorage devices are in the form of steel plates
or castings. The forces applied to the anchorage device will typically be very
large. As a consequence, the concrete in the anchorage zone is subjected to high
tensile (bursting, spalling and bending) and compressive (bearing) stresses.
Confining reinforcement is used to increase the uniaxial compressive strength of
the concrete to a three-dimensional state of strength around the anchorage
device. The anchorage zone can be divided into the local zone immediately
surrounding the anchorage device where the force is transferred 'from the device
to the concrete, and the general zone where the highly concentrated force
spreads out or diffuses into a linear stress gradient across the entire cross-
section. The local zone is strongly influenced by the specific characterstics of the
proprietary anchorage system. The state of stress in the local zone is highly
complex and nonlinear, with a three-dimensional state of stresses. The general
zone is the part of the anchorage zone that is more distant from the anchorage
device and generally extends for a distance about equal to 1.0 to 1.5 times the
depth of the section. In this zone the distribution of stresses induced by the
posttensioning force becomes more uniform. According to Roberts [89] the
extent of the local zone is defined as the greatest of the following:
the maximum width of the local zone
the depth of the confining reinforcing, but no greater the 1.5 times
the width of the local zone

In order to consider a bearing plate as rigid, Roberts [89] indicates the thickness
"t" must be:
t (3 fb n2 / (0.75 fy )0.5
n = 1/2 of the diagonal or diameter of the plate minus the radius of the
wedge plate
= 0.85 F / A
= gross area of plate
= introduced compression force
244

If the plate cannot be considered rigid, it may be used but the effective bearing
area shall be calculated as the area within a perimeter projected trom the
perimeter of the wedge plate through the bearing plate at a 45 degree angle.
The behavior of the anchorage zone is controlled by the concrete strength and
by the reinforcement. The layout of the reinforcement and the tensile capacity
have asignificant influence on the ultimate capacity and on its behavior at service
state. Different failure modes can occur, either in the local or in the general zone.
Failure in the local zone occurs in the immediate vicinity of the anchorage device.
The surface of rupture is often in the shape of a pyramid or cone, delimited by
crushed concrete. The failure is caused by an insufficient bearing strength of the
concrete, by lack of confining reinforcement or by combination of both. The failure
in the genera.l zone is caused by the incapability of the transverse reinforcement
to resist the bursting forces at the time of cracking or during subsequent loading,
or by excessive compressive stresses in the concrete. A bending failure can be
induced in the anchorage zone by the eccentricity of the post-tensioning force
with respect to the overall cross section. This failure is caused by insufficient
tensile capacity of the bending reinforcement.
 245

Burdet [42] indicates that single anchorconfigurationscan be loaded con-

centrically or eccentrically as shown in Fig. 4.34. The geometry ofthe tendon can
be parallel to the axis of the anchorage zone or inclined, and also curved. Other
external forces like transverse post-tensioning or transverse external forces can
act on the anchorage zone.

a.concentric b. eccentric c. inclined

d. inclined and curved e. transverse post-tensioning f. transverse reaction

Figure 4.34: Possible configuration for single anchor ( from Ref. [42])

Many different authors have studied the behavior of anchorage zones.

Closed form elasticity solutions were presented by Guyon [125]. Among many
others Magnel [126], Lenschow and Sozen [127], Schleeh [128], Stone and
Breen [41], Burdet [42], Roberts [89], Sanders [43], Leonhardt [129] have
presented theoretical and experimental solutions for local and general anchor-
age zones. However, the distinction between local zone and general zone was
not clearly made until the recent work of Roberts [89], Sanders [43], and Burdet
[42].
246

The different studies generally concentrated on the spalling forces and bursting
forces. The spalling forces are the tensile stresses acting in areas of the concrete
close to the end surface on either side of the anchorage device. These stresses
are essentially induced by the condition of compatibility of displacements. Guyon
recommended as a design value 4% of the applied load as the corresponding
reinforcement in the form of a ·fine mesh, located as close to the face of the
concrete as possible. Burdet [42] shows this to be conservative. Since such
compatability induced forces cannot be determined from an equilibrium based
strut- and- tie model, the Guyon value is recommended for loaded-end face crack
control.

Figure 4.35: Concentric single anchor: geometry

 247

The bursting stresses are the tensile stresses acting transversely to the axis of
the tendon at a certain distance ahead of the anchorage. Bursting stresses are
caused by the transverse spreading of the concentrated post-tensioning forces
overthe entire cross section. Figure 4.35 shows the geometry and nomenclature
forthe simplest case, that of a concentric single anchor. Figure 4.36 shows some
comparison of Burdet's finite element ~nalysis [42] and Guyon's analysis. For
design purposes Guyon [125] and Leonhardt [129] presented the following
equation to compute the total bursting force:
T = '1 F (1 - a I d)
y = 0.25 (Guyon)
= 0.30 (Leonhardt)

Many specifications have used similar equations. Good agreement with the
finite load analysis can be seen from Fig. 4.36.

0.3 . .
....
.. .
.. .
.. ..
..
....
0.2 '.
.
...... ..
~ ' ....,
...... FE-concentric
.•• Guyon 0.3
Normalized '. .......... ......,
.~

. ... .
Bursting Force _.. Guyon 0.25
'. .. ..
Tburst I P •....
'.

'. ".. , - Guyon Analysis

.
O1 .....''.. , ,
'
,
'
..
....
~
~

O+---f--f--f--+--+--+--+--+--+--+
o 0.1 0.2 0.3 0,4 0.5 0.5 0.7 0.8 0.9 1
Relative Plate Size
alh

Figure 4.36: Comparison of finite element analysis with results from Guyon
(from Ref. [42])
248

Two different strut- and- tie- models are shown for the concentric single anchor
zone case in Fig. 4.37. Since the equations of equilibrium express the overall
equilibrium of the structure, both logically must give the same answer. The thrust
line model of Fig. 4.37(b) gives about 20 percent lowerstrain energy at ultimate
load level [42] than the simple strut- and- tie- model shown in Fig. 4.37(a). This
indicates more efficiency but is more related to the length of the transverse ties.
Since in actual detailing, the ties would be extended towards the outer
edges in both cases, this efficiency would not be practically developed and
either can be used.
 249

·• ....
••
• •• • •
••• • •• ••
•I • • • •• ••
•• •
•
•
•• •• •,
I

,
I
•• •t • t

,• ,• • •I t •t I
I I
•• I
I
I
I
I
I
I I
I
I
•
I
• •I I
• I
I
I I
I
'I
•t I
I I
t I
I I I I
I I
1 I I t

(a) Simple Model (b) Thrust Line Model

Figure 4.37: Comparison of two different strut- and- tie- models with principal
stress vectors (from Ref. [42])

For design purposes the simple model of Fig. 4.37(a) can be used to
determine the required reinforcement as shown in Fig. 4.38. The
expressions for the angle of spreading of the compression force and the
location of the centroid of the reinforcement shown in Fig. 4.38 are based on
the results of Burdet [42]. The reinforcement should not be placed too close
to the anchor. The thrust line model of Fig. 4.37(b) forces the designer to
spread the reinforcement out more over the entire length of the D-region.
When using the simple strut- and- tie model the reinforcement should be
spread in a zone from 0.2 h" to 1.2h" [42]. Some additional transverse
reinforcement should be placed normal to these stirrups to resist the
spreading of the forces in the principal plane normal to this figure in the third
dimension (see example: " I # 3 bar).
250

h
TF
a
L
--II~
C1

T1
I h/2

~ 'd1~
I 0.2h ,
~r--reinforcement area = 1.0 h ~

I~
1-01_ _------D·region: 1.5 h -------~

Proposed strut angle and

depth of the tension ties:

<I> 1 = 12 + 3 / ~ (alh)

d1 = h / (4 tana) measured from centerline of bearing face of plate wittl

a = 23.6
d1 = (h - a) or 4 tan <1>1

.
Figure 4.38: Proposed strut- and- tie- model for anchorage zone
 251

Example 4.5: Anchorage zone

Design the reinforcement required in the post-tensioning anchorage zone shown

in Fig. 4.39 to carry a maximum applied post-tensioning force of 500 kips applied
to the centroidal axis of a beam with an overall height of 36 in. and a width of 12
in. The bearing plate is sufficiently rigid. It has a height of 8 in. and a width of 8
in. Assume concrete strength at time of stressing will be 6000 psi. Check both
local zone for proper confinement and general zone for both transverse bursting
reinforcement and spalling reinforcement. Both bars and spiral can be assumed
as Grade 60.

Load and dimensions:

F = 500 kips
h = 36 in.
b = 12 in.
a = 8 in.
d = 8 in. (diameter of spiral)
s = 1.5 in. (pitch of spiral)

Spalling Force:

Estimate S = 0.04 F = 0.04 (500) = 20 kips

Required As = 20/60 = 0.33 si
Use 2 #4 bars = 0.40 si
These spalling forces are resisted by the #4 stirrups and #3 ties next to the face
as shown in Fig. 4.40.

General Zone: See Fig. 4.39.

Strut- and- tie- model:

<1>1 = 12 + 3/(alh)o.s
<1>1 = 18.40
d1 = (h - a) 1 (4 tan <1>1)
= 21.0 In.
252
Internal forces:

C1 = FI (2 cos 4>1) = 2501 cos 18.40

C1 = 263.5 kips
CO = F 12 tan 4>1 = 250 tan 18.4
CO = 83.2 kips
C2 = F/2
C2 = 250 kips

T1 = C1 sin 4>1 =
T1 = 83.2 kips

Reinforcement for tensile ties: (Grade 60)

T1 = 83.2 kips
As = 83.2/60 = 1.39 in 2 8-#4 = 1.60 in. 2

These bars must be distributed over a zone from 0.2 h (7.2 in.) to 1.2 h (43.2 in.)
from the loaded face select #4 closed stirrups. One additional stirrup is located
as close to front face as cover requirements allow to provide reqUired spalling
reinforcement. The 4 - #4 stirrups which satisfy the required 8 - #4 bars for T1
are then spaced at 8". This results in locations 10.5 in., 18.5 in., 26.5 in., and
34.5 in. as measured from the front face. One a.dditional stirrup is provided at
42.5 in.

Out of plane:

A similar check must be made in the other principal plane. However, since the
bearing plate width in that direction a2 = 8 in. and h2 = b = 12 in., there is appre-
ciably less tensile bursting force. From Fig. 4.36, with a/h 2 = 8/12 = 0.67, the
bursting force T = 0.25 F (1 - alh) seems quite accurate.

= (0.25) (500) (1- 0.67) = 41.2 kips

= 41.2/60 = 0.69 in. 2

This bursting force can be handled by a series of #3 ties (As = 0.11 in. 2). Six of
these ties spaced in a region from 0.2 h2 = 2.4 in. to 1.2 h2 = 14.4 in. should
control such transverse splitting. These #3 ties are shown on Fig. 4.40. They
are also assisted by the spiral in this region.
 253

Example: Anchorage zone

36"
T
8"
..L
F I 18"

14 d1 ~I

I. D-region : 54" ~~

~1 = 12 + 3/'0/(8/36) = 18.4°
d1 =21 in.

Figure 4.39: Strut- and- tie- model for example: anchorage zone
254
Anchorage requirements:

The #4 bars required forthe T1 ties are adequately developed by bending them
around #4 longitudinal bars placed in the corners.

Concrete stresses at node zones:

concrete strength: 6000 psi

efficiency factor: v = 0.9 - 0.25(6000) / 10000 = 0.75

The critical node in anchorages is usually at the anchor plate. The geometry is
indicated in Fig. 3.8.

Check: eCe-node

Horizontal compressive stress in eeG-node

Since w4 > aD, hydrostatic stress

w4 = a/2

w4 = a/2 = 4 in.

b = 12 in.

(jeo = CO / (w4 * b) ~ve fIe

(jca = 83.2/ (4 * 12) = 1.73 ksi ~ 0.75 * 6. = 4.5 ksi OK

Vertical bearing plate stress with

spiral confining reinforcement (See Sec. 3.4."1 b for limits)

(jca = F 1 (a * b) ~ V e fIe (A 1 Ab)o.s + 4 (Aeon 1 A b) feon

teen = [2 As fy 1 (d 5)] (1 - s/d)2

d = 8 in.

s = 1.5 in.

As = (3/8)2 1t 14 = 0.11 in. 2 (Assumes a #3 spiral)

feen = 2 * 0.11 * 60,000 (1-1.5/8)2 { (8 * 1.5)

teon = 726 psi = 0.726 ksi

 255

<J'ca = 5001 (8 * 12) = 5.21 ksi ~ 0.75 * 6. (12/8)°·5+ 4 * [7t8 2 /(4*8 2 )*O.726

(jca = 5.21 ksi ~ 5.51 + 2.28 = 7.79 ksi OK

In detailing of the spiral, a reasonable length is to extend the spiral to a distance

from the bearing plate equal to 1.5a = 12 in. Specific criteria for performance
evaluation of anchorage devices with confining reinforcement is given by Roberts
[89].

Compressive Strut Stresses:

In highly stressed anchorage zones, the stress level in the compressive

struts can be quite high. This is one of the most likely applications in detailing of
structural concrete where compression may occur. As the strut leaves the node,
the force is diffusing and the strut widens as shown in Fig. 3.9. Surdet [42]
suggests that the diffusion of the strut is helpful and that the strut should be
checked at a level (y in Fig. 3.8) corresponding to the effective plate width, a.
For this example <1>3 = 18.4 0 , the angle of spreading can be assumed as a 1:2
slope or diffusion angle of 26.5 0 •

From the geometry of Fig. 3.9, when w4 = a/2 = 4 in., the strut width at a
depth y = a = 8 in. can be found since a' = a/2 = 4 in. Since w3' > w3" ,
w3 + 2w3" = 16.47 + (2)1.26 = 18.99 in.

cr1 = C1 1 w1 b = 263.51 18.99 (12) = 1.16 ksi <

0.6 v e f'c = (0.6) (0.75) (6) = 2.70 ksi OK

Note that at this depth, y = a, the strut stress is essentially the same as the
stress at the end of the general zone cr = 500 1 (36) (12) = 1.16 ksi.
256
·I..- A

36"

2-1~ ~A

I 12" --I #4 straight bars

i----------54"--------t~~1
1.....

Section: A· A

-.-
3"
8" \."':':':::':~-#4stirrups

36"
t-
12"
1:','.+----'"'"'+-11-- ' " #3 ties
l'PIt"'''t--#3 spiral
L
•-L
8"

3" 1~~~-#4bar
.2" I- 8"-/2,,1
~12"~

Figure 4.40: Reinforcement layout for anchorage zone

4.3.6 Pretensioned beam with eccentricity
257
As a last example and in order to illustrate the application of strut- and- tie
models in both 'D' and 'B' regions, a prestressed concrete beam will be analyzed
using the strut- and- tie model as shown in Section 2.5.1. Use of the strut- and-
tie model in B regions may be more cumbersome than ordinary sectional
analysis. By prestressing, forces are artificially created with the help of hydraulic
jacks. Tllese forces act as loads on the prestressing stell and as loads on the
concrete. The designer chooses the tendon profile, the type and the magnitude
of the prestressing forces in such a manner that these artificial loads change the
internal force path created by the actual loads. Proper prestressing ordinarily can
prevent the formation of tension cracks under working loads. Also, the deflection
under working loads can be greatly reduced, because prestress usually puts
camber into a member under dead loading. A simple span beam is prestressed
by introducing a negative moment to offset the expected actual positive moment
and at the same time intoducing a longitudinal compression to offset the tensile
stresses from bending moment. Continuous beams are prestressed in a similar
fashion but for best results require an effective eccentricity above middepth in
negative moment zones. In ordinary computations, when continuous beams are
prestressed, secondary moments are introduced because the reactions prevent
full movement under the action of the prestress. One of the major advantages
of strut- and- tie models is that the prestress is introduced as forces acting on the
structure and their effect is directly considered for the actual boundary conditions.

Example 4.6: Pretensioned beam

The pretensioned beam with end eccentricity from Ref. [55] is investigated
here with the strut- and- tie- model for a concrete strength f'e =5000 psi, n =7,
wire tension fpi = 135 ksi* and a creep and sh rinkage loss of 35 ksi. The concrete
strength when tendons are released is assumed to be fie =4000 psi. Dimensions
are shown in Fig 4.41. In this example, the actual strands are initially lumped
together as if one supersized strand for simplicity. In actuality the 1.5 in. 2 would
have to be provided by 10- 1/2-in.-diameter strands distributed on 2-in. centers
which would result in a slightly higher centroid. In checking the chord stresses

"This assumed initial prestressing steel stress is about the lowest value which might be used
effectively in prestressing. This leaves an equal reserve (fpu - fpi = 270 - 135 = 135) available
for overloads before ultimate.
258
in the lower chord, this larger area corresponding to distributed strands will be used.
Further assume M d = 65 K-ft. and M 1 = 80 K-ft.

For the uncracked section, the steel is transformed as (n-1) As:

A tot = 10 * 20 + 1.5 (7-1 ) = 209 in. 2
y' from middepth of the concrete cross section
y' = -9 * 7/209 = -0.3 in. (below)
'tot = 10 * 20 3 /12 + 200 * 0.3 2 + 9 * 6.7 2 = 7087 in. 4

After cutting the tendons the compression and tension chord forces the strut-and-tie
model shown in Fig. 4.42(b) can be computed:
Pi = (1.5) (135) = 202 kips
LC p1 = Pi / (2 cos 12°)
= 202 / (2 cos 12°) = 103.3 kips

In order to check the compressive stresses resulting from the application of this
concentrated force by the distributed strands, it is assumed that the centroid of the
strands is 3 in. from the bottom and that they are fully distributed over the width.
Thus AP1 = (2) (3) (10) = 60 in. 2
For f c'= 5000, va = 0.9 - 0.25 (5000/10000) = 0.775
For fe' = 4000, va = 0.8. For simplicity USE va = 0.75 throughout this example
Ve fc' = 0.75 * 4000 = 3000 psi
LC p1 / Ap1 = 103.3 / 60 = 1.72 ksi < 0.75 * 4000 = 3.0 ksi OK

Of substantial concern is the need for lateral and vertical reinforcement throughoutthe
transfer length to resist the tension forces Tp 1 shown in Figure 4.42(a).

LTp1 = LP1 / 2 tan ¢P1 = 202 / 2 (2.5 / 8) = 31.6 kips

As = 31.6 / 60 = 0.53 si. Use 5 #3 = 0.55 si.

Note that these tension forces exist laterally as well as vertically so that only 1 leg on
the bottom of each stirrup runs transversely to resist the lateral component. Thus
As for each stirrup is 0.11 si. This reinforcement must be distributed within the
transfer lengths of 50 db = 25 in. as shown in Fig. 4.43
 259

If the '0' region at the end is isolated as shown in Fig. 4.42(c) and the combined
stresses due to the prestress and its eccentricity are computed from
PIA + Pee I I, the values given in Fig. 4.42(c) are found. Applying these
stresses as forces T1 = C2 and C3 = Pi at their respective centroids as indi-
cated, it is very easy to construct the force path and strut- and- tie model shown.
This clearly illustrates that if tensile strength of concrete is not to be relied on, an
area of steel As =26/60 = 0.43 si should be provided in the end regions close
to the top of the beam. Two #4 bars are provided as shown in Fig. 4.43. They
also are useful for positioning and anchoring the stirrups. This '[j' region also
indicates the need for a similar area of vertical reinforcement at the support. The
closely spaced #3 stirrups provided over the support to work locally to resist
strand splitting forces also work nicely over the full depth to provide this resis-
tance. The advantage of strut- and- tie modelling in the '0' regions is clear from
these types of calculations.
260
Example: Pretensioned beam with eccentricity

/..-A

'..-A
I

1=51 ft I.
Wdl+1I = 0.446 k 1ft

Mdl =65 k-ft =780 k - in.

M
II
=80 k-ft =960 k - in.

Section: A - A

17" 20"

-+J
~10'~ ~10"~

Figure 4.41: Pretensioned beam: geometry (from Ref. [55])

 Cp1 I: PI (2 COS ~p1)
Cp2 I: P 12

Tp1 I: (P tan ~1 ) 1 2
261

.- -.
, .......Cp1
.....•
,
I
I
I
I
- - .... .... ....
-Cp2' - ; -
,
- - - ,- ....-
, , 4lP1 .... , , ,
I ."..
I I
P I I

..... ..... Tp1 ..... ..... ..... .....

..... ..... .....
..... Cp1 ..... L -- - -- - - - -- -
..... .....
..... ..... ..... .....

transfer length. 50 cpb .1

Proposed compression strut diffusion angle:

(a) Transfer Length Forces

r
h

L
-I
Tp = Wt b ( -P I A,. + P e Yt IIJ

Cp = we b ( -P I ~ - P e Ye lie)

(b) Model with tendon eccentricity effects

Figure 4.42: Strut- and- tie model for prestressed concrete

262
T1 = 1/2 (1.0) (5.2) (10) =26
C2 = is taken equal to T 1
I
centroid of I
I

shaded triangle ft-1.00

\

T =26 k
N
LO

"\" <j>T

"'" N
'" , C2 =26 k I
LO

T =26k ',------------------------------~ K~~

I
I
IT..:
I
1<------\
• I
I I
I I

i C3= 202 k !
Pi = 202
k
-------- _~--------------------------- . . I.- r -
I
: : v
• •
: I ~
'---""E=~r---------.....,; \
-2.82
~ 6.94 :> centroid of compression
I force outside shaded triangle

plate
ct Choice is arbitrary
as long as angles
do not get to small.
In this case, chosen so
<j>T = 45°

(e) "0" Region at end

Figure 4.42: Continued

 263

Top chord: tension from prestressing forces after cutting the wires

It can be seen from Fig. 4.42(a) that the prestress force is applied to the lower
chord gradually over the transfer length, 50 db = 25 in. It is assumed that this
can be approximated as three equal forces, Pi, each located about 8 in. apart.
However, when compared to the length of the beam, the critical zone fortension
on the top chord can be effectively checked with full prestress and no dead load
moment. This is slightly severe but practical. The prestressing loads are applied
to the overall beam as any other load would be. This load case illustrates one
problem with strut- and- tie modelling. As previously shown in Fig. 2.30(d), if a
simple truss model is used assuming free articulation at all joints, the application
of a horizontal force concentric with the centroid of the lower chord to a simply
supported, articulated, simple truss does not produce top chord forces. However,
in anormal mechanics analysis, it is assumed that plane sections remain plane
so that the conditions of deformation compatibility are introduced. These are not
part of an equilibrium or plastic analysis. Thus in the highly elastic prestressed
beam at service load conditions, these compatibility considerations are neces-
sary and some beam analysis concepts must be introduced.

=
The effective top chord depth, WI' is estimated as having a centroid about as far
from the outer fiber as the centroid of the strands, 3 in. Since the beam has
uniform width and since the 'final stress distribution is assumed uniform, WI = 6 in.
and the distance from the section centroid to the centroid of this chord is 10.3 -
3 =7.3 in. (One can see that lumping of all top chord fibers into a single chord
reduces the accuracy of outer fiber stress calculations done in the traditional
PIA + Mcll manner.)

= 6 * 10 (- 202 I 209 + 202 * 6.7 * 10.3 I 7087) = 60.0 kips

264
Bottom chord: compression from prestressing forces after cutting the wires
Cp = web (- P / A tot - P * e * y c / 'tot )
Cp = 6 * 10 (-202/209 - 202 * 6.7 * 9.7/7087) = 169.1 kips

Check end section where dead load moment is zero under effect of prestress forces:

Top chord:
Tp = 60 kips ~ 6Y f/ At = [6Y 4000 (6) (10) ] / 1000 = 22.7 kips N.G.
At this section the tensile stresses on the top exceed permissible.
The strands should be draped, blanketed, or the cross-section changed.
The same result would be found in a conventional analysis.

Lower chord: Cp = 169.1 kips~0.75f: A b = (.75)(4)(6)(10) = 180 kips OK

From dead load:

Tdl = M dl / z
14.3 (from Ref. [55], or as a first approximation: z = 3/4 h)
= T dl = 780/14.3 = 54.55 kips

Check centerline section under effect of dead load combined with prestress forces:

Top chord:
T (prestr.) - C(load) < 3Y f: At (tension)
60 - 54.55 - 0 = 5.55 kips (tension ~ 3 Y 4000 (6) (10 / 1000 (tension))
= 11.38 kips OK

Bottom chord:
T (load) - C(prestr.) ~ - b w t f:* Ie
54.55 - 169.1 = 114.55 (compression)
~ -6 * 10 * 0.75 * 4 = -180 kips (compression) OK

*Use of ..;e f: which is an ultimate term may be inconsistent with service load
tensile stress checks. Note compression is OK with 0.6 f: as well.
 265

From dead and live load:

Cdl +1I = Tdl+1I = Mdl +U/ Z

Z = 14.3 (from Ref. [55], or as a first approximation: z = 3/4 h)
Cdl +1I = T dl +1I = 1740/14.3 = 121.7 kips

prestress drops (owing to losses) to 202 - (35 * 1.5) = 149.5 kips

Check centerline section under effect of dead load, live load, and prestress after
losses:
266

top chord:

/
T (prestr.) - C(load) ~ - b we V f c
60 * 149.5/202 - 121.7 = -77.6 (compression)
~ - 10 * 6 * 0.75 * 5 = -225 kips (compression) OK

tension chord:

C(prestr.) - T(load) ::;; b wr 3 (f'e )0.5

= 6 in. ::;; 2 in. (For a conservative check, the distance
wr is here limited to the strand spacing
from the outer fiber)
169.1 * 149.5/202 - 121.7 = 3.45 (tension)
::;; 10 * 2 * 3 * (5000)°·5 = 4.24 kips
(tension) OK

vertical chord:

Tv = A - w* x
Tv = 11 .37 - 0.446 * 14.3/12
Tv = 10.84 kips

In this case the tendon is straight so Cr = 0

Tv (Ioad)- C r S; T
10.84 - 0 < #3 U stirrup = 2 * 0.11 * 60 = 13.2 kips

Bennet, Abdul-Ahad and Neville [130] recommend that in reinforced and

prestressed concrete beams in which a large shear is caused by loads varying
or moving in position, the stirrup spacing should not exceed 0.75 h, where h is the
effective depth of the beam.

Use #3 U stirrups at 15 in. spacing as shown in Fig. 4.43.

 267

Obviously the numbers and reinforcement would change if load factors and
resistance factors were applied. However, the principles would be the
same. It can be seen that the allowable stress checks using strut-and-tie
models are more cumbersome than conventional sectional analysis
procedures.

In order to illustrate the check for ultimate conditions, assume ACI

Building Code load factors and cj> = 1.0. Then Mu = 1.4(780) + (1.7)(860) =
2724 in.k. Cu = Tu. = Mu / z = 2724 / 14.3 = 190.5 kips. This load
condition is truly plastic so the basic strut- and- tie model applies. Checking
at the centerline with an effective prestress of 149.5 kips:

I
Top Chord C (load) < b we v e f c
190.5 < (10) (6) (.75) (5)
= 225 o.
Bottom Chord L T (load) < T + P = (Apf py - Pel + p.
190.5 < (1.5) (250) = 375 kips O~

•Apf py - P = A p (f py - f p) is the reserve capacity in the tendons after prestressing.

This is the tension that can be developed in the tendons above that developed during
prestressing.
268

Concrete stresses at support node: CCT - node

This check will be pertormed with the unfactored conditions as an illustration.
Again, in reality, proper load factors and <j> factors would be required. For defini-
tions refer to Fig. 3.12. In this case there would be 3 layers of 1/2" strands, each
2 in. on centers. Thus wT = 3(0.5) + (2)(1.5) = 4.5 in. The inclined compres-
sion struts were assumed at an angle =45°. C1 =A = (51)(0.446)(1/2) = 11.37
<j>

kips. Thus C2 = C1/sin 45 = 11.37/sin 45 = 16.08 kips.

a = 8 in.

wT = 4.5 in.

w2 = a sin <I> + wT cos <I> =8 sin 45 + 4.5 cos 45 = 8.83 in.

O'C2 = C2 1 (w2 * b) ~V fie
e

O'C2 = 16.081 (8.83 * 10) =0.18 ksi ~ 0.75 * 5. =3.75 ksi OK

O'a = A / (a * b) ~V f' c [ac b / (a b) ]0.5

e
O'a = 11.371 (8 * 10) =0.14 ksi ~ 0.75 (10/8)°·5* 5. = 4.19 ksi OK
 269
I+-A
~ .. 4@S"
-I I
:::;:::::::.;.:::;:;:;:;:::;:::;:::::;:::::;:.;.:-:::::;:::;:::;:;:::;:;:;:::;:.:.:.:::;:::;:.:.:.:::::::;:;:-:.:::;:::::::.:.:::::::::::.:.:::::

r
20"

1 I0I.l.""""-11

(..15"~
I
I
• •
U 4 .~A
3 ';f+-~- - - - - 5 1 ft. 1

Section: A - A

#4

20"

prestressing
steel10t "¢
2"

2"

2"

Figure 4.43: Reinforcement layout for pretensioned beam

270
 CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS

In this report a methodology was suggested and an evalua.tion made of the

strut- and- tie- models for detailing structural concrete. Comparison with experi-
mental results from details as well as from single node zones were reported. The
strut- and- tie- model represents a useful tool for detailing concrete structures. It
emphasizes the internal force flow and provides some in-depth understanding of
the behavior of geometrical and statical discontinuities. It is principally of value
in dimensioning and detailing these 'D I or discontinuity regions. While it can be
used in the more regular IB' or Bernoulli regions where linear strain profiles are
encountered, it is not as advantageous as ordinary structural concrete design
procedure in those regions. After discussing the general principles, components
and modeling techniques as well as dimensioning of the struts, ties and nodes
in chapters 2 and 3, illustrative design examples were presented in chapter 4.
Several typical strut- and- tie patterns are furnished in Appendix A.

For the majority of concrete structures it would be unreasonable and

inefficient to model the entire structure with a strut- and- tie- model. It is
advantageous to subdivide the given structure into B-regions and D-regions.
After computing the elastic stress resultants or ordinary cracked reinforced
concrete forces for the B-regions, the equivalent forces should be applied to the
D-regions. Load paths can be sketched based on experience, design aids,
experimental results or a finite element analysis. The strut- and- ties can be
rearanged with consideration of practicality of the reinforcement layout. The
proposed design recommendations are applicable to either prestressed or non-
prestressed reinforced members.

The general assumptions for the application of the strut- and- tie- model
in the design procedure are:

271
272

yielding of the reinforcement is required prior to concrete or anchorage

failure
the ties transfer only unixial forces and neglect dowel action, aggregate
interlock, and tensile strength across cracks.
Dimensioning is an iterative process. The compression struts can be checked by
using the proposed approaches for concrete strength < 12500 psi. Detailed rules
forthe strut limits are given in Section 3.2. Detailed rules for geometry limits and
admissible stresses for unconfined and confined nodes are given in Section 3.4.

In this report, the design procedures based on the strut- and- tie- model
and the proposed detailing approaches are illustrated with a series of design ex-
amples. In addition, several strut- and- tie- models (from Ref. [2,28]) which may
be useful to the designer when detailing '0' regions in concrete structures are
included in Appendix A.

Study of the desig n examples indicates that use of the strut- and- tie model
is an extremely efficient way of detailing reinforcement in '0' regions. The
calculations are relatively simple and straightforward and give the designer
substantial insight. In contrast, the checks of struts and nodes are laborious and
somewhat subjective. It was noted that in many applications these strut and node
stresses were not close to controlling design. Hopefully, further application and
familiarity with the method will give designers a "feel" for when detailed strut and
node calculations are required and when they can be assumed as not governing.
6. References

1. Schlaich, J.; Weischede, D.: A practical Method for the Design and
Detailing of Structural Concrete (in german). Bulletin d'lnformation no.
150. Comite Euro- International du Beton, Paris, 1982163 pp.

2. Schlaich, J.; Schafer, K.: Konstruieren im Stahlbetonbau (Design

for Concrete Structures). In: Betonkalender, Berlin 1989, pp. 563-715

3. AASHTO: American Association of State Highway and Transportation

Officials. Standard Specifications for Highway Bridges. 13th. edition,
Washington, D.C., 1983

4. ACI 318-89: Building Code Requirements for Reinforced Concrete, Detroit,

1989

5. Ritter, W.: Die Bauweise Hennebique (The Hennebique System).

In: Schweizerische Bauzeitung, Bd. XXXIII, No.7, ZOrich, 1899

6. Morsch, E.: Der Eisenbetonbau, seine Theorie und Anwendung (Reinforced

Concrete, Theory and Application), Stuttgart, 1902

7. Barton, D.: Design of Dapped Beams using the Strut- and - Tie- Model.
Master's Thesis, The University of Texas at Austin, 1988

8. Withey, M.: Tests of Plain and Reinforced Concrete, Series of 1906 and 1907.
Bulletin of the University of Wisconsin, Engineering Series, Vol. 4, No.2,
1907 - 1908

9. Talbot, A.: Tests of Reinforced Concrete Beams: Resistance to Web

Stresses, Series of 1907 and 1908. Bulletin 29, University of Illinois
Engineering Experiment Station, 1909

10. Henley, H.: Report of the Committee on Laws and Ordinances. National
Association of Cement Users. Vol. 4, 1908, pp. 233 - 239

11. Richart, F.: An Investigation of Web Stresses in Reinforced Concrete Beams.

Bulletin 166, University of Illinois, Urbana, 1927

273
274

12. Hognestad, E.: What do we know about Diagonal Tension and Web
Reinforcement in Concrete? Circular Series 64, University of Illinois,
Engineering Experiment Station, 1952

13. Kupfer, H.: Erweiterung der Morsch'en Fachwerkanalogie mit Hilfe der
Formanderungsarbeit (Expansion of Morsch's Truss Analogy by
Application of the Principle of Minimum Strain Energy) CEB-Bulletin 40,
Paris, 1964

14. Lampert, P.; ThOrlimann, B.: Ultimate Strength and Design of Reinforced
Concrete Beams in Torsion and Bending. IABSE, No. 31-1, ZOrich, 1971,
pp.107-131

15. Leonhardt, F.: Reducing the Shear Reinforcement in Reinforced

Concrete Beams and Slabs. In: Magazine of Concrete Research, Vol. 17,
No.53, 1965, p. 187

16. ROsch, H.: Ober die Grenzen der Anwendbarkeit der Fachwerkanalogie
bei der Berechnung der Schubfestigkeit von Stahlbetonbalken (On the
Limitations of Applicability of the Truss Analogy for the Shear Design of
Reinforced Concrete Beams), Festschrift F. Campus"amici et alumni,
Universite de Liege, 1964

17. Lampert, P.: Postcracking Stiffness of Reinforced Concrete Beams in

Torsion and Bending, ACI, Detroit, SP-35-12, 1973

18. Grab. J.: Ultimate Strength of Beams with thin walled open crass-
sections. Bericht 56, Institut fOr Baustatik und Konstruktion-ETH, ZOrich,
1970

19. LOchinger, P.: Bruchwiderstand von Kastentragern aus Stahlbeton Linter

Torsion, Biegung und Querkraft (Ultimate Strength of Box-Girders in
Reinforced Concrete under Torsion, Bending and Shear).lnstitut fOr
Baustatik und Konstruktion-ETH ZOrich, Bericht Nr. 69, 1977

20. MOiler, P.: Plastische Berechnung von Stahlbetonscheiben und -balken

(Plastic Analysis of Walls under Bending and Shear). Institut fOr Baustatik
und Konstruktion-ETH ZOrich, Bericht Nr. 83, 1978
 275

21. Nielsen, M.; Braestrup, N.; Jensen, B.; Bach, F. : Concrete Plasticity:
Beam Shear-Shear in Joints -Punching Shear. Specialpublikation -
Dansk Selskab for Bygningsstatik, Lyngby, 1978, 129 pp.

22, Mitchell, D.; Collins, M.: Diagonal Compression Field Theory - A rational
Model for Structural Concrete in pure Torsion. ACI-Journal, Vol. 7, No.8,
1974, pp. 396-408

23. Ramirez, J.; Breen, J.: Proposed Design Procedures for Shear and
Torsion in Reinforced and Prestressed Concrete. Research Report 248-
4F, CenterforTransportation Research-The University of Texas at Austin,
1983,254 pp.

24. Daschner, F.; Kupfer, H.: Versuche zur SchubkraftObertragung in Rissen

von Normal- und Leichtbeton (Sheartransfer across cracks in concrete).
In: Bauingenieur 57, Berlin 1982, pp. 57-60

25. Dei, Poli, S.; Gambarova, P.; Karakoc, C.: Discussion on the papers
"Shear Transfer across cracks in R.C. due to Aggregate Interlock and to
Dowel Action (MRC, No. 126) and "Shear Transfer in R. C." (MCR, No.
130) by Millard, S. and Johnson R., Magazine of Concrete Research, Vol.
38, No. 134, 1986,pp.47-51

26. MacGregor, J.; Gergely, P.: Suggested Revisions to ACI Building Code
Clauses Dealing with Shearin Beams. ACI-Journal, Vol. 74, NO.1 0,1977,
pp.493-500

27. Marti, P.: Truss Models in Detailing. In: Concrete International, V. 7,

No.12, 1985, pp. 66-73

28. Schlaich, J.; Schafer, K.; Jennewein, M.: Toward a Consistent Design of
Structural Concrete. PCI-Journal, Special Report, Vol. 32, No.3, 1987,
pp.74-150

29. CSA: Design of Concrete Structures for Buildings. CAN3-A23.3-M84,

Canadian Standards Association, Rexdale, 1984, 281 pp.

30. Cook, W.; Mitchell, D.: Studies of Disturbed Regions near Discontinuities
in Reinforced Concrete Members. ACI-Structural Journal, 1988,
pp.206-216
276

31. Design and Construction Specifications for Segmental Bridges. Draft

NO.3 Post-tensioning Insitute, Phoenix, 1987

32. Kordina, K.; Hegger, J.; Teutsch, M.: Shear Strength of Prestressed
Concrete Beams with Unbonded Tendons. ACI-Structural Journal, Vol.
86, No.2, Mar-April 1989, pp. 143-149

33. Hartmann, D.; Breen, J.; Kreger, M.: Shear Capacity of High Strength
Prestressed Concrete Girders. Research Report 381-2. Center for
Transportation Research-The University of Texas at Austin, 1988,244 pp.

34. Powers, T.: Shear Behavior of Prestressed Beams in Negative Bending.

Master's Thesis, The University of Texas at Austin, 1989

35. Schaefer, T.: Verification of a Refined Truss Model for Shear Design in
Reinforced and Prestressed Concrete Members. Master's Thesis, The
University of Texas at Austin, 1981

36. Castrodale, R.W.: A Comparison of Design for Shear in Prestressed

Concrete Bridge Girders. Master's Thesis, The University of Texas at
Austin, 1982

37. Kotsovos, M.: Reader Comments: Toward a Consistent Design of

Structural Concrete. PCI-Journal, 1988, pp. 171- 177

38. Anderson, R.: Behavior of cn -Nodes in Reinforced Concrete Strut- and-

Tie- Models. Master's Thesis, The University of Texas at Austin, 1988

39. Bouadi, A.: Behavior of CCT-Nodes in Structural Concrete for the Strut-
and- Tie- Model. Master's Thesis, The University of Texas at Austin,
1989

40. Noguchi, H.; Watanabe, K.: Shear Resistance Mechanisms of Beam-

Column Joints under Reversed Cyclic Loading. In: IABSE Colloquium,
Delft, 1987, pp. 511-522

41. Stone, W.; Breen J.: Design of Post-Tensioned Girder Anchorage Zones.
PCI Journal, Vol. 29, No.1, 1984, pp. 64-109, and No.2, 1984, pp. 28-61

42. Burdet, 0.: Analysis and Design of Anchorage Zones for Post-tensioned
Concrete Bridges. Ph. D. Dissertation, The University of Texas at Austin,
May 1990
 277

43. Sanders, D.: Design and Behavior of Post-tensioned Concrete Anchor

age Zones. Ph. D. Dissertation, The University of Texas at Austin,
August 1990

44. Schlaich, M.: ComputerunterstOtzte Bemessung von Stahlbetonscheiben

mit Fachwerkmodellen. Dissertation, ETH-ZOrich, 1989

45. Yankelevsky, D.: Bond Action between Concrete and a Deformed Bar- A
New Model. ACI-Journal, No. 82-13, 1985, pp. 154-161

46. Rode, U.: Framework Model for the Simulation of Fracture Behavior of
Concrete. In: IABSE Colloquium, Delft, 1987, pp.221-227

47. Hrennikoff, A.: Solution of Problems of Elasticity by the Framework

Method. Journal of appl. Mech. 8, 1941

48. Soroushian, P.; Obaseki, K.; Rojas, M.; Sim, J.: Analysis of Dowel Bars
acting against Reinforced Concrete Core. ACI - Journal No. 83-59,
1986, pp.642-649

49. Vintzeleou, E.; Tassios, T.: Mathematical Models for Dowel Action under
Monotonic and Cyclic Conditions. Magazine of Concrete Research, Vol.
38, No. 134, 1986, pp. 13-2~ 45.

50. Bazant, Z.; Gambarova, P.: Crack Shear in Concrete: Crack Band
Microplane Model. Journal of Structural Engineering, ASCE, Vol. 110,
No.9, 1984, pp. 2015-2035

51. Divakar, M.; Fafi'tis, A, Shah, S.: Constitutive Modeling of Rough

Interfaces in Sliding Shear. Trans. of the 2nd Int. Cont on Constitutive
Laws for Engineering Materials: Theory and Applications, Tucson, 1987,
pp. 1027-1034

52. Walraven, J.: Fundamental Analysis of Aggregate Interlock. Journa.l of the

Structural Division, ASCE, Vol. 107, No. 11, 1981, pp. 2245-2270

53. Gambarova, P.: Modelling of Interface Problems in Reinforced Concrete.

In: IABSE Colloquium, Delft, 1987, pp. 1-16

54. Richart, F.; Brandtzaeg, A.; Brown, R.: A StUdy of the Failure of Concrete
under combined Compressive Stresses. Bulletin No. 185, University of
Illinois, Urbana, 1928, pp. 7-103
278

55. Ferguson, M.; Breen, J.; Jirsa, J.: Reinforced Concrete Fundamentals.
John Wiley & Sons, 5th Edition, 1987

56. ThOrlimarrn, B.; Marti, P.; Pralong, J.; Ritz, P.; Zimmerli, B.: Anwendung
der Plastizitatstheorie auf Stahlbeton (Plasticity in Concrete Structures).
Vorlesung, ETH- ZOrich, 1983

57. Drake, K.: High -Strength Concrete in Seattle."High Strength Concrete,

SP-87, ACI, Detroit, 1985, pp. 21-34

58. ACI Committee 363. State-of-the-Art Report on High Strength Concrete.

ACI Journal, Vol. 81, No.4, 1984,

59. Carrasquillo, R.L.; Slate, F.O.; Nilson,A.H.: Microcracking and Behavior

of High Strength subjected to Short-term Loading. ACI-Journal, Vol. 78,
No.3, 1981, pp. 179 -186

60. ThOrlimann, B.: Lecture Notes from Structural Seminar. The University of
Texas at Austin (see: Ramirez [23])

61. Collins, M.; Mitchell, D.: Design of Disturbed Regions. Chapter 9.

Prestressed Concrete Basics, Canadian Prestressed Concrete
Institute, 1987, pp. 386-429

62. MacGregor, : Reinforced Concrete - Analysis and Design, Prentice-Ha.II,

1988

63. Paulay, T.; Loeber, P.: Shear Transfer by Aggregate Interlock, Shear in
Reinforced Concrete, ACI SP-42, Vol. 1, 1974

64. CEB-MC 1990: Contributiontothe Revision of The CEB-MC 1990, Comite

Euro-International du Beton, Ad hoc Commission IVNI, 1986

65. Vecchio, F.; Collins, M.: Predicting the Response fo Reinforced Concrete
Beams Subjected to Shear Using Modified Compression Field Theory.
ACI - Structural Journal, Vol. 85, No.3, May-June 1988, pp. 258 - 268

66. Kaufmann, K.; Ramirez, J.: Re-evaluation of the Ultimate Shear Behavior
of High-Strength Concrete Prestressed I-Beams. ACI- Structural Journal,
Vol. 85, No.3, May-June 1988, pp. 295 - 303
 279

67. Johnson, M.; Ramirez, J.: Minimum Shear Reinforcement in Beams with
Higher Strength Concrete. ACI - Structural Journal, Vol. 86, No.4,
July-August 1989, pp. 376- 382

68. Kupfer, H.; Hilsdorf, H.; ROsch, H.: Behavior of Concrete under Biaxial
Stresses. ACI-Journal, Vol. 66, No.8, August 1969, pp. 656-666

69. Linse, D.: Festigkeits- und Verformungsverhalten von Beton unter

dreiachsiger Beanspruchung (Strength- and Deformation Capacity of
Concrete under triaxial Stresses). Forschungsarbeiten, Lehrstuhl und
Institut fOr Massivbau, TU-MOnchen, 1973, pp. 25-27

70. Kotsovos,M.: A Brittle Fracturing Material, Material and Structures.

RILEM, Vol. 17, No. 98, 1984

71. Stankowski, T.; Gerstle, K.: Simple Formulation of Concrete Behavior under
Multiaxial Load Histories, ACI-Journal, Vol. 82, No.2, 1985

72. Han, D.; Chen, W.: A Nonuniform Hardening Plasticity Model for Concrete
Materials. Mechanics of Materials, No.4, 1985

73. Eberhardsteiner, J.; Meschke, G.; Mang, H.: Comparison of Constitutive

Models for Triaxially Loaded Concrete. In: IABSE Colloquium, Delft, 1987,
pp.197-208

74. Fardis, M.; Alibe, B.; Tassoulas, J.: Monotonic and Cyclic Constitutive Law
for Concrete. Journal of Engineering Mechanics, Vol. 109, No.2, 1983,
pp.517-536

75. Richart, F.; Brandtzaeg, A.; Brown, R.: A Study of the Failure of Concrete
underCombined Compressive Stresses. Univ. of Illinois, Urbana. Bulletin No.
185,1928, pp. 7 - 103

76. Fafitis, A.; Shah, S.: Lateral Reinforcement for High-Strength Concrete
Columns. ACI, Detroit, SP-87, 1985, pp. 213-232

77. Ahmad, S.; Shah, S.: Stress-Strain Curves of Concrete Confined By Spiral
Reinforcement. ACI-Journal, Vol. 79, No.6, 1982, pp. 484-490

78. Martinez, S.; Nilson, A.; Slate, F.: Spirally-Reinforced High-Strength

Concrete Columns. Research Report No. 82-10. Department of Structural
Engineering, Cornell University, Ithaca, Aug. 1982

79. ACI Manual of Concrete Practice. Part1-1988, Detroit, 1988, Chapter 6

280

80. Billig, K.:AProposalfora Draft Code of Practice for Prestressed Concrete.

Cement and Concrete Association, London 1948

81 Komendant, A.: Prestressed Concrete Structures. Me Graw-Hill Book

company, Inc., 1952, pp. 172 -173

82. Middendorf, K.: Anchorage Bearing Stresses in Post-tensioned Concrete.

Journal of the Amel;can Concrete Institute. 1960, pp. 580- 584

83. Hawkins, N.: The Bearing Strength of Concrete Loaded through Rigid
Plates. Magazine of Concrete Research, Vol. 20, No. 62, 1968, pp. 31- 40

84. Hawkins, N.: The Bearing Strength of Concrete for Strip Loadings.
Magazine of Concrete Research, Vol. 22, No. 71,1971/6

85. Nijogi, S.: Bearing Strength of Reinforced Concrete Blocks.

Journal of the Structural Division, ASCE. ST5, 1975, pp. 1125 - 1137

86. Nijogi, S.: Concrete Bearing Strength - Support, Mix, Size Effect. Journal
of the Structural Division, ASCE. ST8, 1974, pp. 1685 - 1702

87. Sheikh, S.: Effectiveness of Rectangular Ties as Confinement Steel in

Reinforced Concrete Columns. Ph.D. - Thesis, Univ. of Toronto, 1978,
256 pp.

88. Muguruma, H.; Watanabe, F.; Iwashimizu, I.; Mitsueda, R.: Ductility
Improvement of High Strength Concrete by Lateral Confinement.
Reprinted from Transactions of the Japan Concrete Institute, 1983,
pp. 403 - 410

89. Roberts, C.: Behavior and Design of the Local Anchorage Zone in Post-
Tensioned Concrete. Master - Thesis, The University of Texas at Austin,
1990

90. Wurm, P.; Daschner, F.: Versuche Ober Teilflachenbelastung von

Normalbeton (Tests SUbjected to partial area loading of concrete).
Deutscher Ausschuss fOr Stahlbeton, Heft 286, Berlin 1977

91. Ramirez, J.; Breen, J.: Review of Design Procedures for Shear and
Torsion in Reinforced and Prestressed Concrete. Research Report 248-
2, Center for Transportation Research-The University of Texas at Austin,
1983, 186 pp.
 281

92. Lin, T.Y.; Burns. N.H.: Design of Prestressed Concrete Structures. John
Wiley & Sons,1981, pp. 12 - 130

93. Libby, J. R.: Modern Prestressed Concrete. Design Principles and

Construction Methods. Third edition, Princeton 1971, pp. 167 - 171

94. Leonhardt, F.: Prestressed Concrete. Design and Construction. Berlin,

1964, pp. 79 - 84

95. Hoyer, E.; Friedrich, E.: Beitrag zur Frage der Hafspannung in Eisen-
beton bauteilen, (Contribution to Questions Regarding Bond Stresses in
Reinforced Concrete Structural Members). Beton und Eisen, Vol. 38, No.
6,1939, pp. 102-110.

96. Hanson, N.; Kaar, P.: Flexural Bond Tests of Pretensioned Prestressed
Beams. ACI - Journal. 1959, pp. 784 - 802

97. Tepfers, R.: Lapped Tensile Reinforcement Splices. Journal of the

Structural Division, ASCE, Vol. 188, No.1, January 1982, pp. 283-301

98. Tepfers, R.: A Theory of Bond Applied Overlapped Tensile Reinforce

ment Splices for Deformed Bars. Chalmers Univ. of Technology, Gote-
borg, Publication No. 73.2, 1973,328 pp

99. Jirsa, J.; Lutz, L.; Gergely, P.: Rationale for Suggested Development,
Splice, and Standard Hook Provisions for Deformed Bars in Tension. In:
Concrete International, Vol. 1, No.7, 1979, p. 47

100. Orangun, C.; Jirsa, J.; Breen, J.: Reevalauation of Test Data on Develop
ment Length and Splices. ACI-Journal, 74, No.3, 1977, p. 114

101. ACI 318-89: Building Code Requirements for Reinforced Concrete,

Chapter 12, Detroit, 1989

102. Lormanometee, S.: Bond Strength of Deformed Reinforcing

Bar under Lateral Pressure. MS - Thesis. The University of Texas at
Austin, 197464 pp.

103. Schmidt-Thro, G.; Stockl, S.; Kupfer, H.: Verankerung der Bewehrung am
Endauflager bei einachsiger Querpressung (Anchorage of Reinforce-
ment at an End Bearing with Uniaxial Lateral Pressure). Deutscher Auss-
chuB fOr Stahlbeton, Heft 389, Berlin, 1988, pp.11-98
282

104. Kotsovos, M.: Compressive Force Path Concept: Basis for Reinforced
Concrete Ultimate Limit State Design. In: ACI Structural Journal,
Vol. 85, No.1, Jan-Feb 1988, pp. 68 - 75

105. Park, R.; Paulay, T.: Reinforced Concrete Structures. Wiley-Inter-

science, New York, 1975, p. 671

106. Leonhardt, F.: Das Bewehren von Stahlbetontragwerken. (Reinforcing of

Concrete Members) In. Beton-Kalender 1971 , Berlin, pp. 303 - 398

107. PCI Design Handbook: Precast and Prestressed Concrete. Third edition.
PCI, Chicago, 1985

108. Betonkalender : Taschenbuch fOr Beton-Stahlbeton- und Spannbeton

bau sowie die verwandten Facher.Berlin

109. Concrete Reinforced Steel Institute (CRSI) Handbook. Fourth edition.

Chicago, 1984

110. EI-Rahman, M.,A.;Shaw, W.J.D.: Computers in Concrete Technology. In:

Concrete International. 1989/8, pp. 27 - 31

111. Khana, S.: Stress-Strain Characteristics and Failure Mechanism of PIC.

Master-Thesis. The University of Texas at Austin, 1981

112. Hagberg, T.: Design of Concrete Brackets: On the Application ofthe Truss
Analogy. ACI-Journal, Proceedings V. 80, No.1, Jan-Feb 1983, pp. 3 -12

113. Cook, W.; Mitchell, D.: Studies of Reinforced Concrete Regions Near
Discontinuities. Department of Civil Engineering and Applied Mechanics.
McGill University, Structural Engineering Series Report No. 87-3, 1987.
153 pp.

114. Yong, Y.; McCloskey, D.; Nawy, E.: Reinforced Corbels of High-Strength
Concrete. In: High-Strength Concrete SP-87-11. ACI, Detroit, 1985,
pp. 197 - 212

115. Franz, G.; Niedenhof, H.: The Reinforcement of Brackets and Short Deep
Beams. Cement and Concrete Assocation, Library Translation No. 114,
London, 1963
 283

116. Kriz, L.; Raths, C.: Connections in Precast Concrete Structures-Strength

of Corbels. In: Journal of the Prestressed Concrete Institute. Vol. 10,
No.1, January 1965, pp. 16 - 61

117. Mattock, A.; Chen, K.: Soonswang, K.: The Behavior of Reinforced
Concrete Corbels. In: PCI - Journal, V. 21, No.3, May-June 1976,
pp. 18 - 42

118. Chakrabarti, Pinaki; Farahani, Davood; Kashou, Shihadeh: Reinforced

and Precompressed Concrete Corbels - An Experimental Study. In: ACI
- Structural Journal. No. 86-S37, 1989, pp. 405 - 412

119. Peterson, R.: Stress Concentration Factors. New York. 1974. pp. 208

120. Chih-Bing Ling: Stresses in a Stretched Slab Having a Spherical Cavity.

ASME, Vol. 81, Series E, Applied Mechanics Section, 1959, pp. 235

121 Tsuchida, E.; Nakahara, I.: Three-Dimensional Stress Concentration

around a Spherical Cavity in a Semi-Infinite Elastic Body. Bull. Japan Soc.
Mech. Eng., Vol. 13, 1970, pp. 499

122. Shah, S.N.: Evaluation of infill wall strengthening schemes for nun-
ductile reinforced concrete buildings. Master's Thesis, The University
of Texas at Austin, 1989

122. Gaynor, P.: The effect of openings on the cyclic behavior of reinforced
concrete shear walls. Master's Thesis, The University of Texas at Austin,
1988

124. Leonhardt, F.: Vorlesungen Ober Massivbau. Teil1 bis Teil6. Springer-
Verlag, Berlin,1984

125. Guyon, Y.: Prestressed Concrete. New York, 1953

126. Magnel, G.: Prestressed Concrete. New York, Third edition 1954,345 pp.

127. Lenschow, R.; Sozen, M.: Practical Analysis of the Anchorage Zone Problem
in Prestressed Beams. In: ACI - Journal, November 1965, pp. 1421 - 1437

128. Schleeh, W.: The Stress Pattern at the Support of a Prestressed Concrete
Beam. In: Beton- und Stahlbetonbau. Vol. 58, No.7, Berlin, 1963,
pp. 157 - 171
284

129. Leonhardt, F.: Prestressed Concrete - Design and Construction. Berlin.

Second edition. 1964,677 pp.

130. Bennet, E.W.; Abdul-Ahad, H.Y; Neville, A.: Shear Strength of Reinforced
and Prestressed Concrete Beams Subjectto Moving Loads. In: PCI- Journal.
Nov-Dec 1972, pp. 58 - 69

131. Barton, D.L.; Anderson, A.B.; Bouadi, A.; Jirsa, J. 0.; and Breen, J.E.: "An
Investigation of Strut-and-Tie Models for Dapped Beam Details," Research
Report 1127-1, Center for Transportation Research, The University of Texas
at Austin, May 1991.
 Appendix A
Detailing Aids
(from Ref. [1] and [2])

Deep beam with one single load on both sides

Deep beam with two single loads on both sides

Deep beam with three single loads on both sides

Deep beam with three single loads on one side

Deep beam 'with one single load in the middle: d ~ I

Deep beam with one single load in the middle: d > I

Deep beam with two single loads in the middle of.one side

Deep beam with distribuited compression load

Deep beam with distribuited tension load

Cross section under torsion

Frame corner with positive moment

Frame corner with negative moment

Single load in the middle of a deep beam

Support

The stepped beam

285
286
I~ ~I
F

C1 = F 1 (2 cos ep1)
T2 =CO =F 1 2 tan ep1
ep1 =12 + 31 "(all)
 287

F F

C1 =F / cos ~1
T1 =C2 =F tan ~1
~1 =12 + 3/..J(aII)
 288
I~ ~I
F F

F F F

I.
C1 = F 1 (2 cos 4>1)

co =F 1 2 tan 4>1
T2 + C2 = F tan 4>1

4>1 = 12 + 31 "(all)

Ford/l S 1:T1 =T2/3

For dll ~ 2: T1 = T2 I 2
 289

F F

C1 = F / (2 cos ep1)

co = F /2 tan ep1
12 + C2 = Ftan ep1

ep1 = 12 + 3/ ~(aIJ)

For dll s; 1: 11 = T2 / 3
For dll ~ 2: T1 =T2 / 2
 290
2F

d 5; I

Fo-------T1 --------"'i\

t
F
t
F

co =F tan e
T1 =F tan e
T2 =F tan 4>11 cos e
C1 = F 1( 2 cos 4>1 cos e)
C2 =F 1 (2 cos e)
4>1 =12 + 311(aII)
 291

/
J CO ~
,
C1 .~
c

I
/

T2
v- C1,
I
I I
C2
I I
I I d> I
/L_-C3---~
C4 C4
;--v /
,~
,,
,/ 1"1
,
t
F
t
F

co = C3 = F tan 4> 1

11 = T2 = F tan 4> 1

C1 =C4 = F / cos 4> 1

C2=F

4>1 =12 + 3/.f(aJI)

 292
F

d~ I

:T1 ~~

t
O.313F 1.375F
t
O.313F

~a~ ~a~
1-
. . - - ------1---..- - ----.1

T1 =1.375 F tan e co =0.313 F tan 8

T2 =0.313 F tan 4>11 cos 8 C1 = 0.16 F 1 (cos 4>1 cos 8)

T4 = 1.375 Ftan 4>11 cos e C2=0.16 F/(cos 8)

C3 = 0.69 F 1(cos 4>1 cos 8)

T3 = 1.062 F tan e
C4 =0.69 F 1(cos e)
.s.z..
..... - -
, ~-31-_:--:Q.
f 4>1 = 12 + 3/I(an)
0.313 F 1.375 F 0.313 F
 293

I I
I I
C3 C3
• •
I I
J- - - C2 - - -~
J \
J \ d
I C1
C1
. \ z::: 3/4 d
I \
'r-J "8
6,

It
, \
\
T1

t
F
tF

J
arctan e =1/(4 z)::: 1/(3 d)
F = C3 = w 1/2

T1 = F tan e
C1 = F /cos e
C2 =T1 = F tan 6
294 1--
I
114
1 4

--C2---
I \
I \ d
I C1
C1
I \ z "" 3/4 d
I T2 T2 \

8,
,
f------+----T1 -----1---'""'"1

.1
arctan e = II (4 z) "" II (3 d)
F = C3 = w 1/2

T1 = F tan 8

C1 = F Icos 8

T2=F
 295

- - ---
-- - - -- ,,
- ,,
--- -
T2
,, T2 ,
,,
--
T1 _C1 T1

T2
, ,.
""
"C2
"

- ... ...
~
r - lIr
J

, !y-. J~b hO st -.
~
,1.
~
t

,
~
~th
..: ~ ~ r
~th llr

t t
~bo---~l
Tt =M.' (2 bO ho) Ts= S 1(2 z)

th = hO 16 =
T1 TItan 5 T I (bO sin 5 cos 5) ~ v f~
for 45°
tb =bO 16 T1 =T T I (hO sin 5 cos 5) ~ v f~

T2 = T tan 5
for 45°
T2=T
296

M =C1 "z T3 =T1 "tan a2

C1 =T1 T3:: 0.3 T1

~ v f~
C1 I 2 "w1 " b T4 =T1 " tan (cx2) * sin (180-28)/ sin 8

C2 =T1 I cos a2
a2 =15.18 0
T2 =T1 "sin (8 - (2) I cos a2
 297

)
M
d
/
M = C1 * z

C1 =11

C1 1 2 * w1 * b ~ v f~

C2 = C1/sln e

C2/2*w2*b~ vf~

C3 = C1/tan e
12 =11/sln e
298
co

I
I f ~ '\ "
T1 T1
d
C1 F C1

1 ~ ,

1/ 45:\ T2 T2
~,
.L..:::....1
I C3 C3 I
I '4>1 1\ I
I \
C2 ..-T3 - . , C2
d
I I . I I
I I I I
I C4 C4 I
I I
I I I

I
I.

T1 =F /4 co= F /4

T2 =F 14 C1 = 2 F 112

C2=F/4
T3 =F 14 tan <1>1
C3 =F 14 cos 4>1

C4 =F /4
.. h

W61· z •
r 299

~I

1
f

r- a
F
1.5 d

1 g

z=0.75 h y =0.75 g / (2 cos 91 sin 8 )

arctan f3 =0.75 g / (f + 0.25 h - w6) x=O.75g

w6 ~ C6 / (2 b v ~ ) v =[x2 + y2 - 2 x y cos (90-8+91) ]0.5

91 = 12 + 3 /-./ (a/f) arcsin ex =y sin (90 - B + 91) / v

12 =F / tan (B+91) T1 =T3 = C1 sin 91/ cos (ex - B)

C1 = F / sin (B+91) C2 =C3 =F / sin (8+91)

T5=2F

C6= F
300

C1 =T1 C6 =2 C1 - 2 cos 45° T4

T6 =1.732 C6 - 2.732 C11 -1.366 T9
=
T2 T1 • tan r
for r= 15° T7 =2 C6 - 2 C11
=
T2 T1 * 0.268
T8 =0.707 T9
T3 =C1 * tanr / cos 45°
for r= 15° C11 (b * 2 * w1) ~ v f~
T3 =C1 * 0.379
C11 I (b * 2 * w11) ~ v t;
T4 =C1 (cos r - sin r) I (2 cos r cos 45°)
for r= 15°
T4 = C1 * 0.518

=
T5 C1 I cos 45° - T4
for r= 15°
=
T5 C1 * 0.896