ANSYS Polystat Users Guide

ANSYS Polystat User's Guide
ANSYS, Inc. Release 2020 R1

Southpointe January 2020
2600 ANSYS Drive
Canonsburg, PA 15317 ANSYS, Inc. and
ansysinfo@ansys.com ANSYS Europe,
Ltd. are UL
http://www.ansys.com registered ISO
(T) 724-746-3304 9001: 2015
(F) 724-514-9494 companies.
Copyright and Trademark Information
© 2020 ANSYS, Inc. Unauthorized use, distribution or duplication is prohibited.
ANSYS, ANSYS Workbench, AUTODYN, CFX, FLUENT and any and all ANSYS, Inc. brand, product, service and feature
names, logos and slogans are registered trademarks or trademarks of ANSYS, Inc. or its subsidiaries located in the
United States or other countries. ICEM CFD is a trademark used by ANSYS, Inc. under license. CFX is a trademark
of Sony Corporation in Japan. All other brand, product, service and feature names or trademarks are the property
of their respective owners. FLEXlm and FLEXnet are trademarks of Flexera Software LLC.
Disclaimer Notice
THIS ANSYS SOFTWARE PRODUCT AND PROGRAM DOCUMENTATION INCLUDE TRADE SECRETS AND ARE CONFID-
ENTIAL AND PROPRIETARY PRODUCTS OF ANSYS, INC., ITS SUBSIDIARIES, OR LICENSORS. The software products
and documentation are furnished by ANSYS, Inc., its subsidiaries, or affiliates under a software license agreement
that contains provisions concerning non-disclosure, copying, length and nature of use, compliance with exporting
laws, warranties, disclaimers, limitations of liability, and remedies, and other provisions. The software products
and documentation may be used, disclosed, transferred, or copied only in accordance with the terms and conditions
of that software license agreement.
ANSYS, Inc. and ANSYS Europe, Ltd. are UL registered ISO 9001: 2015 companies.
U.S. Government Rights
For U.S. Government users, except as specifically granted by the ANSYS, Inc. software license agreement, the use,
duplication, or disclosure by the United States Government is subject to restrictions stated in the ANSYS, Inc.
software license agreement and FAR 12.212 (for non-DOD licenses).
Third-Party Software
See the legal information in the product help files for the complete Legal Notice for ANSYS proprietary software
and third-party software. If you are unable to access the Legal Notice, contact ANSYS, Inc.
Published in the U.S.A.

Table of Contents
Using This Manual ........................................................................................................................................ xi
1. The Contents of This Manual ............................................................................................................... xi
2. The Contents of the ANSYS Polyflow Manuals ...................................................................................... xi
1. Introduction ............................................................................................................................................ 1
1.1. Objectives ........................................................................................................................................ 1
1.2. How to Characterize Mixing? ............................................................................................................. 1
1.3. Classification of Flows, Capabilities of the Mixing Module .................................................................. 2
1.4. General Explanation on the Way to Solve a Mixing Task ...................................................................... 4
1.5. The Numerical Techniques Involved in the Mixing Module ................................................................. 6
1.6. Examples .......................................................................................................................................... 7
1.6.1. Example 50: The Rectangular Cavity .......................................................................................... 7
1.6.2. Example 51: Coextrusion of Polymers in a Square Channel ........................................................ 7
1.6.3. Example 52: Flat Die ................................................................................................................. 7
1.6.4. Example 46: Periodic Flow Through a Kenics Mixer .................................................................... 8
1.6.5. Example 37: Mixer 2-D .............................................................................................................. 8
1.6.6. Example 65: Adaptive Meshing for Moving Parts; Mixer 2-D ....................................................... 8
1.6.7. Example 91: Dispersion ............................................................................................................ 8
1.6.8. Example 116: Distributive Mixing Analysis in the Mixer GK7N .................................................... 8
1.7. Known Limitations in ANSYS Polystat 2020 R1 ................................................................................... 9
2. The Mixing Theory ................................................................................................................................. 11
2.1. Introduction ................................................................................................................................... 11
2.2. Kinematic Parameters ..................................................................................................................... 12
2.2.1. Kinematic Parameters for 2D Flows ......................................................................................... 12
2.2.2. Kinematic Parameters for 3D Flows ......................................................................................... 13
2.2.3. Statistical Analysis .................................................................................................................. 14
2.2.3.1. Mean and Standard Deviation ........................................................................................ 15
2.2.3.2. Cumulated Probability Function (or Distribution Function) ............................................. 15
2.2.3.3. Density of Probability Function ...................................................................................... 15
2.2.3.4. Percentiles ..................................................................................................................... 16
2.2.3.5. Histograms .................................................................................................................... 16
2.2.3.6. Correlations ................................................................................................................... 17
2.3. Homogenization ............................................................................................................................. 17
2.3.1. Definition ............................................................................................................................... 17
2.3.2. Numerical Method ................................................................................................................. 19
2.4. Distributive Mixing ......................................................................................................................... 20
2.4.1. Distribution Index .................................................................................................................. 20
2.4.2. Distribution in Zones .............................................................................................................. 22
2.4.3. Deviation of Points Concentration .......................................................................................... 23
2.5. Disagglomeration ........................................................................................................................... 24
2.6. Comment ....................................................................................................................................... 29
3. Mixing Tasks in Polydata ....................................................................................................................... 31
3.1. Creating a New Mixing Task ............................................................................................................. 31
3.2. Using the General Menu of a Mixing Task ......................................................................................... 34
3.3. Definition of the Flow Domain ........................................................................................................ 36
3.4. Definition of the Boundary Conditions ............................................................................................ 37
3.4.1. Specification of the Boundary Type ......................................................................................... 38
3.4.2. Specification of a Stopping Plane ............................................................................................ 43
3.5. Definition of the Flow Field ............................................................................................................. 45
3.5.1. Steady-State Flow .................................................................................................................. 45
3.5.2. Time-Dependent Flow ............................................................................................................ 46
Release 2020 R1 - © ANSYS, Inc. All rights reserved. - Contains proprietary and confidential information
of ANSYS, Inc. and its subsidiaries and affiliates. iii
Polystat User's Guide
3.6. Parameters for the Generation of the Material Points ....................................................................... 49

3.7. Parameters for the Tracking ............................................................................................................. 56
3.8. Selection of Properties .................................................................................................................... 58
3.8.1. Parameters for the Kinematic Mixing Properties ...................................................................... 60
3.9. Parameters for the Storage of the Results ........................................................................................ 61
3.10. Definition of Moving Parts ............................................................................................................. 63
3.11. Definition of the Kinematic Analysis .............................................................................................. 64
3.11.1. Management of Properties ................................................................................................... 67
3.12. Definition of the Distribution Analysis ........................................................................................... 69
4. The Polystat User’s Manual ................................................................................................................... 75
4.1. The File Menu ................................................................................................................................. 77
4.1.1. Open... ................................................................................................................................... 78
4.1.2. Options .................................................................................................................................. 78
4.1.3. Read Data .............................................................................................................................. 79
4.1.4. Read mesh ............................................................................................................................. 80
4.1.5. Run ........................................................................................................................................ 81
4.1.6. Draw results ........................................................................................................................... 81
4.1.7. Draw Stats. ............................................................................................................................. 87
4.1.8. WRITE Trajectories .................................................................................................................. 89
4.1.9. WRITE Slices ........................................................................................................................... 91
4.1.10. WRITE Stat. ........................................................................................................................... 93
4.1.11. Save ..................................................................................................................................... 94
4.2. The Properties Menu ....................................................................................................................... 95
4.2.1. Definition ............................................................................................................................... 95
4.2.2. See Properties ........................................................................................................................ 96
4.2.3. Create Properties ................................................................................................................... 96
4.2.3.1. || A || .............................................................................................................................. 98
4.2.3.2. A^x ............................................................................................................................... 99
4.2.3.3. exp ( A ) ......................................................................................................................... 99
4.2.3.4. log ( A ) ........................................................................................................................ 100
4.2.3.5. A + B, A - B ................................................................................................................... 100
4.2.3.6. A / B ............................................................................................................................ 100
4.2.3.7. A * B ............................................................................................................................ 100
4.2.3.8. Rotate ......................................................................................................................... 101
4.2.3.9. Translate ...................................................................................................................... 103
4.2.3.10. Derivate .................................................................................................................... 104
4.2.3.11. Integrate ................................................................................................................... 104
4.2.3.12. k (constant) ............................................................................................................... 105
4.2.3.13. A ( Slice ) .................................................................................................................... 106
4.2.3.14. Concentration ........................................................................................................... 108
4.2.3.15. Min / Max .................................................................................................................. 109
4.2.3.16. Extract ....................................................................................................................... 110
4.2.3.17. Step .......................................................................................................................... 111
4.2.3.18. Instant. Eff. ................................................................................................................. 112
4.2.3.19. Time Aver. Eff. ............................................................................................................. 113
4.2.3.20. Shear Rate ................................................................................................................. 114
4.2.3.21. Melting Index ............................................................................................................ 115
4.2.4. Disagglomeration Properties ................................................................................................ 116
4.2.4.1. Disagglomeration ........................................................................................................ 118
4.2.4.2. Typical size of agglomerates ......................................................................................... 119
4.2.4.3. Fraction of agglomerates of given size .......................................................................... 120
4.2.4.4. Number of agglomerates of given size ......................................................................... 121
iv of ANSYS, Inc. and its subsidiaries and affiliates.
4.2.4.5. Mass of agglomerates of given size .............................................................................. 122

4.3. The Trajectories Menu ................................................................................................................... 123
4.3.1. See the Set of Trajectories ..................................................................................................... 123
4.3.2. CREATE a new set of trajectories ........................................................................................... 125
4.3.3. COMBINE sets of trajectories ................................................................................................. 125
4.3.4. The SELECT one single trajectory Option ............................................................................... 126
4.3.5. SEARCH stagnation points .................................................................................................... 127
4.4. The Slices Menu ............................................................................................................................ 129
4.4.1. See the Set of Slices .............................................................................................................. 130
4.4.2. Automatic Slicing ................................................................................................................. 131
4.4.3. Manual Slicing ...................................................................................................................... 132
4.4.4. Sub - Slicing ......................................................................................................................... 133
4.5. The Statistics Menu ....................................................................................................................... 134
4.5.1. See the Statistical Functions ................................................................................................. 134
4.5.2. Create Statistical Functions ................................................................................................... 135
4.5.2.1. The Property along a trajectory Function ...................................................................... 138
4.5.2.2. The Sum Function ........................................................................................................ 139
4.5.2.3. The Mean + Standard Deviation Function ..................................................................... 140
4.5.2.4. The Correlation Function .............................................................................................. 141
4.5.2.5. The Probability Function .............................................................................................. 142
4.5.2.6. The Auto-Correlation Function of the Concentration Function ...................................... 143
4.5.2.7. The Distance distribution Function ............................................................................... 144
4.5.2.8. The Distribution in zones Function ............................................................................... 146
4.5.2.9. The Sum A <op> Sum B Function ................................................................................. 148
4.5.2.10. The Smoothing Function ............................................................................................ 149
4.5.2.11. The Density of Probability Function ............................................................................ 151
4.5.2.12. The Percentile Function .............................................................................................. 152
4.5.2.13. The Histogram Function ............................................................................................. 154
4.5.2.14. The Segregation Scale Function ................................................................................. 155
4.5.2.15. The Deviation Function .............................................................................................. 156
4.5.2.16. The Points Concentration Deviation Function ............................................................. 158
4.5.3. New disagglomeration functions .......................................................................................... 159
4.5.3.1. The See disagglomeration along a single trajectory Function ........................................ 161
4.5.3.2. The Mean disagglomeration function ........................................................................... 162
4.5.3.3. The Probability function ............................................................................................... 164
4.6. The Help Menu ............................................................................................................................. 165
4.7. Additional Definitions ................................................................................................................... 165
4.7.1. The Slices ............................................................................................................................. 165
4.7.2. The Zones ............................................................................................................................ 166
4.7.3. Weighting ............................................................................................................................ 167
A. The Simulation of the Distribution ......................................................................................................... 169
B. The Global Efficiency of Stretching ......................................................................................................... 171
C. Adaptive Meshing and Mixing Task ........................................................................................................ 173
D. Sliding Mesh Technique and Mixing Task ............................................................................................... 175
Bibliography ............................................................................................................................................. 177
of ANSYS, Inc. and its subsidiaries and affiliates. v
vi of ANSYS, Inc. and its subsidiaries and affiliates.
List of Figures
1.1. Open and Closed Domains ...................................................................................................................... 2
1.2. Example of a Piecewise Steady Flow ........................................................................................................ 3
1.3. Example of a Spatially Periodic Flow ........................................................................................................ 4
1.4. The Flow Simulation ............................................................................................................................... 5
1.5. The Mixing Simulation ............................................................................................................................ 5
1.6. The Statistical Postprocessing ................................................................................................................. 6
2.1. Distribution Function Graph .................................................................................................................. 15
2.2. Density of Probability Function Graph ................................................................................................... 16
2.3. Percentiles Graph .................................................................................................................................. 16
2.4. Histograms Graph ................................................................................................................................. 17
2.5. Correlations Graph ................................................................................................................................ 17
2.6. Mixing Fluids ........................................................................................................................................ 17
2.7.The Correlation Coefficient for Concentration over Distance ................................................................... 18
2.8. Calculated Values of the Correlation Coefficient ..................................................................................... 20
2.9. Distributing Particles from a Small Box .................................................................................................. 20
2.10. Density of Probability over Distance .................................................................................................... 21
2.11. Real and Optimal Distributions ............................................................................................................ 22
2.12. Distributing Particles from a Small Box ................................................................................................ 22
2.13. Zones of the Flow Domain .................................................................................................................. 23
2.14. Points Within a Cluster ........................................................................................................................ 24
2.15. A Representative Volume .................................................................................................................... 25
2.16. Discretized Mass Density Function ...................................................................................................... 25
2.17. Disagglomeration Processes ................................................................................................................ 26
2.18. The Effect of Erosion on the Mass Distribution Function over Time ....................................................... 27
2.19. The Transfer Function for Rupture ........................................................................................................ 28
3.1. The Polydata Menu ............................................................................................................................... 32
3.2. The Create a new task Menu .................................................................................................................. 33
3.3. The MIXING Task 1 Menu ....................................................................................................................... 34
3.4. The Flow domain Menu ......................................................................................................................... 36
3.5. Fluid and Solid Subdomains .................................................................................................................. 37
3.6. The Boundary conditions Menu ............................................................................................................. 38
3.7. The Boundary condition along Boundary 1 Menu .................................................................................. 39
3.8. Inflow and Outflow Boundary Conditions .............................................................................................. 39
3.9. Relative and Absolute Trajectories ......................................................................................................... 40
3.10. Dialog Box for the Entry Boundary ....................................................................................................... 40
3.11. The Corresponding boundary Menu .................................................................................................... 41
3.12. The Updated Boundary condition along Boundary 1 Menu .................................................................. 42
3.13. The Dialog Box for the Spatially Periodic Boundary Parameters ............................................................ 43
3.14. 2.5D Flow with a Stopping Plane ......................................................................................................... 43
3.15. The Parameters for the stopping planes Menu ..................................................................................... 44
3.16. Dialog Box for Defining Coefficients .................................................................................................... 44
3.17. Dialog Box for Specifying Trajectory .................................................................................................... 45
3.18.The Parameters for the stopping planes Menu with Multiple Stopping Planes ....................................... 45
3.19.The Flow definition Menu for Steady-State Flow ................................................................................... 46
3.20. The Flow definition Menu for Time-Dependent Flow ............................................................................ 47
3.21. The Flow definition Menu with Manual Selection ................................................................................. 48
3.22. The Flow definition : list of result files Menu ......................................................................................... 48
3.23.The velocity field 1 Menu ..................................................................................................................... 49
3.24. The Generation of material points Menu .............................................................................................. 50
3.25. The Generation of material points Menu with Multiple Zones ............................................................... 51
of ANSYS, Inc. and its subsidiaries and affiliates. vii
3.26. The Generation Zone 1 Menu .............................................................................................................. 52

3.27. Equidistant Distribution of Points in a 2D Box ...................................................................................... 53
3.28. The Add topo-objects Menu ................................................................................................................ 55
3.29. The Parameters for the tracking Menu ................................................................................................. 56
3.30. The Parameters for the tracking Menu ................................................................................................. 57
3.31. The Trajectory of a Material Point ......................................................................................................... 58
3.32. The Selection of properties Menu ........................................................................................................ 59
3.33. The Parameters for the kinematic variables Menu ................................................................................ 60
3.34. The Storage of the results Menu .......................................................................................................... 62
3.35. The Define moving parts Menu ........................................................................................................... 63
3.36. The moving part #1 Menu ................................................................................................................... 64
3.37. A Kinematic Analysis Mixing Task ......................................................................................................... 65
3.38. The Definition of the KINEMATIC analysis Menu ................................................................................... 65
3.39. Advanced Options for the Definition of the KINEMATIC analysis Menu .................................................. 67
3.40. The Management of properties for the KINEMATIC analysis Menu ........................................................ 68
3.41. A Distribution Analysis Mixing Task ...................................................................................................... 70
3.42. The Definition of the DISTRIBUTION analysis Menu .............................................................................. 70
3.43. Advanced Options for the Definition of the DISTRIBUTION analysis Menu ............................................. 72
4.1. The Polystat Window ............................................................................................................................. 75
4.2. A Cutting Plane Through a Flow Domain ............................................................................................... 76
4.3. The File Menu Options .......................................................................................................................... 77
4.4. The Options Dialog Box ......................................................................................................................... 78
4.5. The Read Mixing files Dialog Box ........................................................................................................... 79
4.6. The Read mixing files Dialog Box ........................................................................................................... 79
4.7. The Read a mesh file Dialog Box ............................................................................................................ 80
4.8. The RUN Dialog Box .............................................................................................................................. 81
4.9. The Graphic Display Window of Polystat ................................................................................................ 82
4.10. The Graphic Display Window Tabs ....................................................................................................... 83
4.11. The Draw Statistical Results Dialog Box ................................................................................................ 87
4.12. The Polycurve Window ........................................................................................................................ 88
4.13. The Save Set of Trajectories Dialog Box ................................................................................................ 89
4.14. Setting Up the Writing of Trajectory Data ............................................................................................. 90
4.15. The Save Set of Slices Dialog Box ......................................................................................................... 91
4.16. Setting Up the Writing of Slice Data ..................................................................................................... 92
4.17. The Save Statistical Results Dialog Box ................................................................................................. 93
4.18. The Properties Menu Options .............................................................................................................. 95
4.19. Positions along a Trajectory ................................................................................................................. 95
4.20. The See / Modify / Delete existing properties Dialog Box ...................................................................... 96
4.21. The CREATE a new PROPERTY Dialog Box ............................................................................................. 97
4.22.Types of Properties that Can Be Created ............................................................................................... 98
4.23. The || A || Settings in the CREATE a new PROPERTY Dialog Box .............................................................. 99
4.24. The A * B Settings in the CREATE a new PROPERTY Dialog Box ............................................................ 101
4.25. Rotation Data .................................................................................................................................... 102
4.26. The Rotate Settings in the CREATE a new PROPERTY Dialog Box ......................................................... 102
4.27. The Translate Settings in the CREATE a new PROPERTY Dialog Box ...................................................... 103
4.28. The Derivate Settings in the CREATE a new PROPERTY Dialog Box ...................................................... 104
4.29. The Integrate Settings in the CREATE a new PROPERTY Dialog Box ..................................................... 105
4.30. The k (constant) Settings in the CREATE a new PROPERTY Dialog Box ................................................. 106
4.31.The A ( Slice ) Settings in the CREATE a new PROPERTY Dialog Box ...................................................... 107
4.32. Transporting a Property from a Slice .................................................................................................. 107
4.33. Defining Concentration Using a Slice ................................................................................................. 108
4.34. The Concentration Settings in the CREATE a new PROPERTY Dialog Box ............................................. 109
viii of ANSYS, Inc. and its subsidiaries and affiliates.
4.35. The Min / Max Settings in the CREATE a new PROPERTY Dialog Box .................................................... 110
4.36.The Extract Settings in the CREATE a new PROPERTY Dialog Box ......................................................... 111
4.37. The Step Settings in the CREATE a new PROPERTY Dialog Box ............................................................ 112
4.38. The Instant. Eff. Settings in the CREATE a new PROPERTY Dialog Box ................................................... 113
4.39. The Time Aver. Eff. Settings in the CREATE a new PROPERTY Dialog Box ............................................... 114
4.40. The Shear Rate Settings in the CREATE a new PROPERTY Dialog Box ................................................... 115
4.41. The Melting Index Settings in the CREATE a new PROPERTY Dialog Box .............................................. 116
4.42. Types of New Disagglomeration Properties ........................................................................................ 117
4.43. The Disagglomeration Settings in the CREATE a new PROPERTY Dialog Box ........................................ 118
4.44.The Typical size of agglomerates Settings in the CREATE a new PROPERTY Dialog Box ......................... 119
4.45. Minimum and Maximum Sizes ........................................................................................................... 120
4.46.The Fraction of agglomerates of given size Settings in the CREATE a new PROPERTY Dialog Box .......... 120
4.47. The Number of agglomerates of given size Settings in the CREATE a new PROPERTY Dialog Box ......... 121
4.48. The Mass of agglomerates of given size Settings in the CREATE a new PROPERTY Dialog Box .............. 122
4.49. The Trajectories Menu Options .......................................................................................................... 123
4.50. The See / Modify / Delete existing sets of trajectories Dialog Box ........................................................ 124
4.51. The Add a new set of trajectories Dialog Box ...................................................................................... 125
4.52. The set C = set A <operator> set B Dialog Box .................................................................................... 126
4.53. The Select a single trajectory Dialog Box ............................................................................................ 127
4.54. The Search stagnation points Dialog Box ........................................................................................... 128
4.55. The Slices Menu Options ................................................................................................................... 129
4.56. Interpolating a New Instant ............................................................................................................... 130
4.57. The See / Modify / Delete existing sets of slices Dialog Box ................................................................. 130
4.58. The Create ‘Automatically’ a list of slices Dialog Box ............................................................................ 131
4.59. The Create ‘manually’ a list of slices Dialog Box ................................................................................... 132
4.60. The Generate a subslicing on a set of slices Dialog Box ....................................................................... 133
4.61. Periodic Subslicing ............................................................................................................................ 134
4.62. Nonperiodic Subslicing ..................................................................................................................... 134
4.63. The Statistics Menu Options .............................................................................................................. 134
4.64. The See / Modify / Delete existing functions Dialog Box ..................................................................... 135
4.65.The Create a new function Dialog Box ................................................................................................ 136
4.66. Types of Functions that Can Be Created ............................................................................................. 137
4.67. The Property along a trajectory Settings for the Create a new function Dialog Box .............................. 138
4.68. The Time Evolution of a Property for Two Trajectories ......................................................................... 138
4.69. The Sum Settings for the Create a new function Dialog Box ................................................................ 139
4.70. The Evolution of the Sum Function of a Property along the Slices ....................................................... 139
4.71. The Mean + Standard Deviation Settings for the Create a new function Dialog Box ............................. 140
4.72.The Mean and Deviation of a Property along the Slices ....................................................................... 140
4.73. The Correlation Function Settings for the Create a new function Dialog Box ....................................... 141
4.74. Exploring the Correlation Between Properties in Two Slices ................................................................ 142
4.75. The Probability Function Settings for the Create a new function Dialog Box ........................................ 142
4.76. The Probability Function of a Property for a Slice ................................................................................ 143
4.77.The Auto-Correlation Function of the Concentration Settings for the Create a new function Dialog
Box ........................................................................................................................................................... 143
4.78. The Auto-Correlation Function for a Slice ........................................................................................... 144
4.79. The Distance distribution Settings for the Create a new function Dialog Box ....................................... 144
4.80. A Distribution Function of Distances Between Material Points ............................................................ 145
4.81. The Distribution in zones Settings for the Create a new function Dialog Box ....................................... 146
4.82. The Sum A <op> Sum B Settings for the Create a new function Dialog Box ......................................... 148
4.83.The Smoothing Settings for the Create a new function Dialog Box ...................................................... 149
4.84. Equal Smoothing .............................................................................................................................. 150
4.85. Centered Smoothing ......................................................................................................................... 150
of ANSYS, Inc. and its subsidiaries and affiliates. ix
4.86. Upwind Smoothing ........................................................................................................................... 150

4.87. Downwind Smoothing ...................................................................................................................... 150
4.88.The Density of Probability Function Settings for the Create a new function Dialog Box ........................ 151
4.89. The Density of Probability Function of a Property for a Slice ............................................................... 152
4.90.The Percentile Settings for the Create a new function Dialog Box ........................................................ 152
4.91. Percentiles along the Slices ............................................................................................................... 153
4.92. The Histogram Settings for the Create a new function Dialog Box ....................................................... 154
4.93. The Histograms for a Slice .................................................................................................................. 154
4.94. The Segregation Scale Settings for the Create a new function Dialog Box ........................................... 155
4.95. The Segregation Scale along the Slices .............................................................................................. 155
4.96. The Deviation Settings for the Create a new function Dialog Box ........................................................ 156
4.97. The Deviation along the Slices ........................................................................................................... 157
4.98. The Points Concentration Deviation Settings for the Create a new function Dialog Box ....................... 158
4.99. Types of New Disagglomeration Functions ........................................................................................ 160
4.100.The See disagglomeration along a single trajectory Settings for the Create a new function Dialog
Box ........................................................................................................................................................... 161
4.101. Disagglomeration Functions for Two Time Steps .............................................................................. 162
4.102.The Mean disagglomeration function Settings for the Create a new function Dialog Box ................... 162
4.103. The Mean Disagglomeration Function for a Slice .............................................................................. 163
4.104.The Probability function Settings for the Create a new function Dialog Box ....................................... 164
4.105. The Help Menu Options ................................................................................................................... 165
4.106. The Define a slice Dialog Box ........................................................................................................... 166
4.107. Defining a Position Zone ................................................................................................................. 166
4.108. The Define a zone Dialog Box .......................................................................................................... 167
1. Disconnected Subdomains .................................................................................................................... 175
2. Overlapping Subdomains ...................................................................................................................... 176
x of ANSYS, Inc. and its subsidiaries and affiliates.
Using This Manual
1. The Contents of This Manual
The ANSYS Polystat User’s Guide tells you how to set up a MIXING task in ANSYS Polydata and how
to use the ANSYS Polystat module for statistical postprocessing of results.
Important:
Under U.S. and international copyright law, ANSYS, Inc. is unable to distribute copies of the
papers listed in the bibliography, other than those published internally by ANSYS, Inc. Use
your library or a document delivery service to obtain copies of copyrighted papers.
2. The Contents of the ANSYS Polyflow Manuals

The manuals listed below form the ANSYS Polyflow product documentation set. They include descriptions
of the procedures, commands, and theoretical details needed to use ANSYS Polyflow products.
• The Polyflow User's Guide explains how to use ANSYS Polydata and ANSYS Polyflow to set up and solve
a problem.
• The Polyflow Tutorial Guide contains a number of example problems with complete detailed instructions,
commentary, and postprocessing of results.
• The Polyflow in Workbench User's Guide explains how to use the ANSYS Polyflow application within
ANSYS Workbench.
• The Polymat User's Guide explains how to use the ANSYS Polymat module for material property evalu-
ation.
• The Polystat User's Guide (p. 1) explains how to set up a MIXING task in ANSYS Polydata and
how to use the ANSYS Polystat module for statistical postprocessing of results.
Related video help can be found on the ANSYS How To Videos page.
• The GAMBIT manuals teach you how to use the GAMBIT preprocessor for geometry creation and mesh
generation.
• The CFD-Post User's Guide explains how to use CFD-Post to examine your results.
For details on how to access the ANSYS Polyflow manuals, see ANSYS Polyflow Documentation in the
separate Polyflow User's Guide.
of ANSYS, Inc. and its subsidiaries and affiliates. xi
xii of ANSYS, Inc. and its subsidiaries and affiliates.
Chapter 1: Introduction
1.1. Objectives
The mixing process is prevalent in industry. There are many different kinds of mixers, and mixing can
be done as a continuous or batch process, as in Banburry mixers, Kenics mixers, extruders, stirred vessels,
and so on. The mixing objectives vary as well: distribution of pigments or other compatibilizers, gener-
ation of interfaces between different fluids in order to enhance chemical reactions, and so on.
The main objective of this manual is to offer you the ability to quantify the mixing in the process of
your interest. This manual contains a set of objective parameters that are relevant for different situations.
The way that mixing processes are quantified is constantly evolving, so Polystat allows you to not only
use existing parameters, but define new parameters as well.
Polystat simulates mixing for various flows, but mixing situations are so complex, that numerical simu-
lation cannot take into account all the real phenomena existing in such processes. The following sections
explain the assumptions and hypotheses that are needed to perform a simulation.
1.2. How to Characterize Mixing?

There are various ways to define the process of mixing. Dankwertz in the 1950s analyzed the mixing
as a homogenization process of a concentration field. If initially, two different fluids are separated in
adjacent zones by a barrier and the barrier is then removed, as time passes, the local concentration of
each fluid disperses everywhere within the fluid. With perfect mixing, the concentration tends to the
same value everywhere in the fluid domain. To quantify this homogenization process, Dankwertz defined
two parameters: the segregation scale and the intensity of segregation. The segregation scale is the
average thickness of the striations present in the flow domain. The intensity of segregation is the
standard deviation of the concentration around its mean. These parameters are used when there are
only two fluids to mix, and when their proportions in the flow domain are approximately equivalent.
In the 1980s, Ottino defined other parameters based on the Continuum Mechanics Theory. He showed
that mixing is a process that increases the interface that exists between fluids, but instead of measuring
the surface of the interface (which is almost impossible in complex flows), he measures the local increase
of infinitesimal surfaces distributed everywhere in the flow domain.
However, these parameters are not very useful when analyzing the distribution of small amounts of
pigments, tracers, and so on, in the flow domain (small percentage of the volume of the total flow do-
main). Based on the work of Manas-Zloczower, a new parameter, , is defined in order to measure this
process. Suppose a set of particles is placed in a small zone of the flow domain and over time these
particles move in the flow and distribute throughout the flow domain. The parameter measures the
deviation of the current distribution with respect to a perfect distribution of particles in the flow domain.
Another technique that is similar to the previous one is available. The flow domain is divided into a set
of adjacent (and non-overlapping) zones. Initially, a set of particles is placed in a small box in the flow
domain and they distribute progressively. Over a period of time, the number of points in each zone are
of ANSYS, Inc. and its subsidiaries and affiliates. 1
Introduction
counted. If each zone contains a number of points proportional to its surface/volume, then there is
good distributive mixing.
A third method to estimate distributive mixing, is to evaluate the local points concentration in various
locations of the flow domain and compare these concentrations with a perfect points concentration
(where the same number of points per unit volume are found everywhere in the mixer). A new para-
meter, , measures the deviation in points concentration.
Dispersive mixing is another important aspect of the mixing process, it concerns the break-up of drops
into small droplets or the disagglomeration of solid particles in a matrix. The stress applied on the drops
or on the solid particles by the matrix is the mechanism that can lead to dispersion. If the stresses are
high enough to compete with the surface tension of drops or with internal mechanical resistance of
solid particles, dispersion occurs. Dispersion is improved if there is some elongational effect in the flow.
This information is available by adding postprocessors when defining the set-up for the flow calculation.
Polystat includes a model to calculate the disagglomeration process along multiple trajectories.
The Polydata mixing task is a general and accurate method to calculate all of these parameters in a
single simulation. The main steps of this method are:
1. Calculate the flow pattern.
2. Compute the trajectories of a large set of material points and calculate the local deformation of the ma-
terials along these trajectories (as well as other relevant properties).
3. Analyze the results using the statistical tools available in Polystat to obtain a global, objective overview
of the mixing evolution.
1.3. Classification of Flows, Capabilities of the Mixing Module

The mixing task in Polydata does not work for all types of flow systems, there are some limitations.
The following are concept definitions:
• Open / Closed domain: A closed domain is a domain with no entry and no exit of fluid. An open domain
has a fluid inlet and a fluid outlet.
Figure 1.1: Open and Closed Domains
• Steady state / Time dependent flow: For steady-state flow, the flow does not change over time. The gen-
eral case for a time dependent flow (also known as transient flow) is a flow that evolves continuously with
time. The Polydata mixing task is an effective tool for studying these types of flows. However, for transient
flows, the flow domain must not change with time except for flows with moving parts simulated using the
2 of ANSYS, Inc. and its subsidiaries and affiliates.
Classification of Flows, Capabilities of the Mixing Module
mesh superposition technique (for additional information, see Advantages of the Mesh Superposition
Technique in the Polyflow User's Guide). If the flow is transient, the current flow must be calculated and stored
at successive time steps. These two flows are noted as flow( ) and flow( ). Two techniques are implemented
to calculate the particles’ path, because velocity field is needed at intermediate times.
– In the piecewise steady flow case, the flow is assumed to be steady between time steps and (with
and flow( ) is equal to flow( )). This assumption is valid if inertia is neglected and if the boundary
conditions change abruptly. One example is an oscillating square cavity:
Figure 1.2: Example of a Piecewise Steady Flow
In this case, two velocity fields alternate. The first state lasts for DT1 seconds; the upper wall moves
to the right while the other walls are at rest. The second state lasts for DT2 seconds; the lower wall
moves to the right while the other walls are at rest. This type of flow achieves far better mixing
than a steady-state flow.
– In the more general second case, the flow changes continuously between and . The flow at time ,
is a linear combination of flow( ) and flow( ).
(1.1)
where
(1.2)
and , .
• Spatially periodic flows: The flow is spatially periodic if there is an initial location where the flow field can
be calculated and where the velocity field in the inflow section is equal (exactly) to the velocity field in the
outflow section. A spatially periodic flow is only possible in an open domain.
Introduction
Figure 1.3: Example of a Spatially Periodic Flow
The flow field is repeated infinitely in space: the flow field in the next module is the same as in the
current module, and in the previous module, and so on.
The mixing task has the following limitations:
• The domain must not change with time. For example, if you want to analyze the mixing in a single screw
extruder, you must assume the screw to be fixed and the barrel to be rotating. If there are moving internal
parts, the mesh superposition technique must be used to simulate their motion.
• For a piecewise steady flow, there must not be any inertia.
• The flow must always be incompressible.
• There is no void formation in the flow. The flow domain is completely filled with the same fluid: if you want
to mix two or more fluids, they must have the same rheological behavior, no diffusion nor chemical reactions
between them, and no interfacial tension.
Even though this is a mixing problem, the flow calculations are identical to that of a single homogeneous
fluid. The time dependence of a set of mixing parameters is examined without making any distinction
between the fluids to be mixed.
There are no limitations on:
• The fluid model: generalized Newtonian or visco-elastic models are available.
• The dimensional complexity of the problem: 2D planar, 2D axisymmetric, 2D 1/2 planar (3 components for
the velocity field), 2D 1/2 axisymmetric (swirling flows), 3D.
• The thermal complexity of the problem: isothermal or non-isothermal simulations are possible.
1.4. General Explanation on the Way to Solve a Mixing Task

There are three major steps in the calculation of a mixing task problem; calculating the flow, calculating
a set of trajectories, and performing a statistical analysis of the results.
1. Flow simulation.
You must define a finite element mesh (via GAMBIT, ICEM CFD, Patran, and so on). Next, define an
F.E.M. task in Polydata to calculate the flow ONLY. Run Polyflow with the data file, to obtain a
General Explanation on the Way to Solve a Mixing Task
Polyflow result file containing the velocity field, the shear rate, and other fields of interest. If the
flow is transient, it is recommended that you save the Polyflow results files at exacts time steps,
. However, if the flow is piecewise-steady with successive boundary conditions ( small, in
loop or not), it may be easier to perform Polydata sessions (one for each specific set of boundary
conditions). Run Polyflow times (once for each data file), to obtain result files, each one con-
taining a set of fields specific to a particular set of boundary conditions.
Figure 1.4: The Flow Simulation
2. Mixing simulation.
Return to Polydata to define a MIXING task. Specify the mesh, the velocity fields to use and the
initial position of the material points and other properties to evaluate along the trajectories. Next,
run Polyflow with this data file. Polyflow randomly generates the initial position of a set of material
points and calculates their trajectory in the flow domain. Polyflow also calculates the evolution of
some properties and kinematic parameters (temperature, viscosity, stretching, rate of stretching,
rate of dissipation, and so on) along these trajectories, and generates a results file.
Figure 1.5: The Mixing Simulation
3. Statistical postprocessing.
Use the Polystat postprocessor to analyze all of the trajectories. You will calculate the time evolution
of global mixing parameters such as the segregation scale or the evolution of the mean stretching
or other parameters.
Polystat is built so that you can define new parameters and test them. New parameters are created by
combining existing parameters in various ways . Polystat can also

be used to analyze processes other than mixing, for example, the quality of a molten glass exiting a
Introduction
furnace. It is also possible to visualize the spatial distribution of a kinematic parameter at a given time,
or in a cutting plane, or along a given trajectory. You can visualize the statistical functions of these
parameters using Polystat, Excel, or another similar postprocessor.
Figure 1.6: The Statistical Postprocessing
1.5. The Numerical Techniques Involved in the Mixing Module

Typically, trajectories are calculated by the time integration of the equation with a Euler explicit
scheme. This method is sufficient if you are only interested in the successive positions of material points.
A more accurate numerical technique is required if you want to know the precise deformation accumu-
lated along these trajectories.
The mixing module combines two techniques: explicit fourth order Runge-Kutta, and instead of integ-
rating the motion of a particle in the real space a coordinate transformation is performed. The trajectory
of the parent element is integrated with the Runge-Kutta method.
(1.3)
Successive positions of the particle in real space is calculated using:
(1.4)
The mixing module algorithm proceeds as follows:
1. Initialization:
a. Find the element , which contains the initial position .
b. Find the local coordinates, , of in element .
2. The algorithm solves by the following loop unless there is a problem or a required stop:
a. Integrate Equation 1.3 (p. 6), until a boundary of element is crossed.
b. If a boundary of is crossed, the time step of integration is adapted in such a way that the position
is on the boundary.
c. Once on a boundary of in :
Examples
i. Search the element adjacent to for where to continue the integration. This element is denoted
as .
ii. Determine the local coordinates in element of the current position ;
iii. Return to step 2.a.
Knowing the iterative procedure that the mixing module uses is important because you must define
some of the numerical procedures that this algorithm uses in Polydata. For example, you must define
the coefficient NBELEM, which indicates the mean number of integration steps necessary to cross an
element (For additional information, see Parameters for the Tracking (p. 56)).
1.6. Examples
Below are short descriptions of the Polyflow examples devoted to mixing. These examples are contained
in the ANSYS Polyflow Examples Manual, which you can find on the ANSYS Help site.
1.6.1. Example 50: The Rectangular Cavity

This example is the tutorial of the mixing module
In this example, you compare the mixing efficiency of a steady-state flow with a piecewise steady
flow. This is a 2D planar and isothermal flow problem.
This example explains in detail how to use Polystat: a) creating new properties, b) defining a slicing
on time, c) defining the statistical functions needed for the comparison of the two cases. It also explains
how to extract useful information from the statistical curves.
Keywords: piecewise steady flow, distributive mixing, reorientation process, mixing efficiency, Polystat,
statistical analysis
1.6.2. Example 51: Coextrusion of Polymers in a Square Channel

Two viscoelastic fluids are injected in a channel and they have identical rheological properties but
different colors. The axial evolution of the interface between these fluids in the square channel is
analyzed. The axial evolution of the segregation scale is calculated in order to quantify the progressive
deformations of the interface in the channel.
Keywords: 2D 1/2 planar flow, viscoelasticity, coextrusion, secondary motion, Polystat, concentration
field, segregation scale
1.6.3. Example 52: Flat Die

In this example, you analyze a 3D steady-state non-isothermal flow through the die section of an
extruder. The focus is on the residence time distribution and the melting characteristics of the matter
leaving the die.
Keywords: 3D steady state flow, non-isothermal, flat die, Polystat, residence time distribution, melting
index, preset statistical analysis
Introduction
1.6.4. Example 46: Periodic Flow Through a Kenics Mixer

In this example, you analyze the distributive mixing generated by a Kenics mixer. As the complete
flow domain is too large to be used, the flow calculation is reduced to a single mixing element of
the mixer. The flow field is assumed to be spatially periodic. The generation of striations through
successive mixing sections and the efficiency of the process are analyzed.
Keywords: 3D steady state flow, periodic boundary conditions, static mixer, Polystat, distributive mixing,
preset statistical analysis
1.6.5. Example 37: Mixer 2-D

This example simulates the 2D transient flow produced by the rigid rotation of two cams in a batch
mixer. The dispersive mixing capability of the mixer is evaluated. Forces and torque along the cams
are also evaluated.
Keywords: mesh superposition technique, batch mixer, transient flow problem, forces and torque,
dispersive mixing, mixing index, eigen values of the stress tensor, shear rate, vorticity, Polystat,
preset statistical analysis
1.6.6. Example 65: Adaptive Meshing for Moving Parts; Mixer 2-D
This example simulates the 2D transient flow produced by the rigid rotation of two cams in a batch
mixer. As in example 37, the mesh superposition technique is used to take into account the rotation
of the cams. Adaptive meshing is used to refine the mesh locally in the neighborhood of the cams,
in order to get a more accurate evaluation of the flow field. As the mesh changes with time, it cannot
be used anymore to perform tracking of material points; you must save the flow field in a set of CSV
files first, which you map onto a new, uniformly refined mesh. The tracking is eventually performed
on this new mesh with the mapped flow field.
Keywords: mesh superposition technique, batch mixer, time dependent flow, adaptive meshing,
conversion form CSV files, mixing task, statistical analysis
1.6.7. Example 91: Dispersion

This example presents the models of erosion and rupture in a simple shear flow. You analyze the effect
of various functions and parameters of these models.
Keywords: dispersion, disagglomeration, erosion, rupture, Polystat
1.6.8. Example 116: Distributive Mixing Analysis in the Mixer GK7N

This example simulates the 3D transient flow produced by the rigid rotation of two rotors in a batch
mixer (mixer GK7N). As in Examples 37 and 65, the mesh superposition technique is used to take into
account the rotation of the rotors. The distributive mixing capability of the mixer is evaluated.
Keywords: mesh superposition technique, batch mixer, time dependent flow, distributive mixing, mixing
task, preset statistical analysis, Polystat
Known Limitations in ANSYS Polystat 2020 R1
1.7. Known Limitations in ANSYS Polystat 2020 R1

This section lists limitations that are known to exist in ANSYS Polystat. Where possible, suggested
workarounds are provided.
• Mesh
– ANSYS Polystat does not support mixing tasks on CutCell meshes.
– ANSYS Polystat does not support mixing tasks on meshes with subdivided elements. See Appendix C: Ad-
aptive Meshing and Mixing Task (p. 173) for a method to circumvent this limitation.
– ANSYS Polystat does not support mixing tasks if the flow field is meshed using the sliding mesh technique.
See Appendix D: Sliding Mesh Technique and Mixing Task (p. 175) for a method to circumvent this limitation.
– When using the Mesh Superposition Technique (MST) with moving parts (with or without slippage along
the moving parts), the trajectories and kinematics of the deformation may be affected when material
points (massless particles) are close to the surface of the moving parts.
– When using MST with moving parts, it may happen that some of the material points get overlapped by
a moving part. This is due to the accumulation of small errors along the trajectory of the material points.
There are two ways (which can be combined) to limit this problem: decreasing the time step between
two successive flows and refining the mesh of the flow domain.
– ANSYS Polystat does not support mixing tasks if the mesh of the flow domain changes with time (changes
in coordinate position of the mesh vertices and/or number and shape of mesh elements). This does not
include the use of the Mesh Superposition Technique with moving parts on a transient flow field, because
the underlying mesh remains constant.
• Boundary conditions
– ANSYS Polystat does not support mixing tasks on flow fields defined by non-conformal boundaries.
– In some cases, material points (massless particles) may stop on an impermeable boundary (wall, free
surface, plane of symmetry) due to stagnation of the flow field in close proximity to these boundaries.
This can have a dramatic effect on the flow statistics, particularly if you are analyzing an open flow field
(with entry and exit of fluid). When slicing from entry to exit you should confirm that the number of
samples in each slice remains constant. If the concentration of samples decreases by more than 10–15%,
it may be necessary to refine the mesh.
• Material properties
– ANSYS Polystat does not support mixing tasks on compressible fluids.
Chapter 2: The Mixing Theory
2.1. Introduction
In polymer blending, a minor component is generally present as drops (or filaments) in a continuous
phase of a major component. Mixing is a process of deformation and rupture of the drops but also a
process of distribution of those drops in the whole flow domain. Good mixing is characterized by small
and identical drops distributed uniformly throughout all of flow domain.
Deformation of drops is promoted by the viscous stress exerted on the drops by the flow field and
counteracted by the interfacial stress . is the interfacial tension and , the local radius. The capillary
number is useful to characterize mixing.
(2.1)
For a given pair of polymers, a critical Capillary number may be found. It corresponds to the situation
where the viscous stress competes with the interfacial stress. The drop is extended and finally breaks
up into smaller droplets. This process is called dispersive mixing. Note that an extensional flow field is
more efficient for breaking up drops into droplets than a shear flow is.
If the capillary number is much higher than the critical capillary number, then the viscous stress overrules
the interfacial stress, and the drop is extended but does not break up. This process is called distributive
mixing. Conversely, if the capillary number is much lower than the critical capillary number, then the
interfacial stress dominates and the drop is only slightly deformed.
In general, mixing begins with a distributive step (drops are deformed passively), followed by a dispersive
step (drops break up into droplets), and finally by the distribution of the droplets in the flow. The fol-
lowing paragraphs focus on distributive mixing (Kinematic Parameters (p. 12) and Homogeniza-
tion (p. 17)) and on distribution of material points into the flow domain (Distributive Mixing (p. 20)).
Dispersive mixing can also be analyzed. You can add postprocessors to the flow calculation:
• Mixing index (or 'flow number') — indicates if the flow is locally a rigid motion (mixing index = 0), a shear
flow (mixing index = 0.5), or in extension (mixing index = 1).
• Eigen values of the extra-stress tensor — this field gives access to the main component of the stress that
stretches and breaks the drops.
Once the flow and those postprocessors are determined, you can calculate the evolution of the mixing
index and the main stress along the trajectories of material points. With these data you can evaluate
the fraction of the matter experiencing a given stress value and evaluate the efficiency of the dispersive
mixing.
The process where solid particles are broken by erosion or rupture in smaller parts due to stresses applied
on them by a matrix (carbon black or silica in a rubber matrix, for example) is called dispersive mixing.
A new model has been developed by B. Alsteens and V. Legat [2] (p. 177) to simulate this process of
disagglomeration and is available in Polystat. The description of the model is found in Disagglomera-
tion (p. 24).
The Mixing Theory
2.2. Kinematic Parameters

One way to measure mixing is to quantify the capacity of the flow to deform matter and to generate
interface. The following theory section neglects interfacial forces — the interface is passive and no
break-up into droplets can occur.
For 2D flows, the interface between fluids is a line. Instead of calculating the evolution of this interface
(a very complex and impossible task to perform because of the exponential growth of the interface),
the stretching of infinitesimal vectors attached to a large number of material points distributed in all
the flow domain is calculated. As the points move in the flow, the vectors are stretched. The stretching
and the rate of stretching of these vectors are properties that vary by time and location in the flow
domain. Statistical analysis of the results provide a global overview of the process. Using this method,
you have an objective and quantitative evaluation of the mixing of any process. You can, for example,
find areas of poor mixing in the domain (low stretching instead of exponential increase). For 3D flows,
the interface is a surface and the stretching of infinitesimal surfaces attached to material points is cal-
culated.
Let and denote the domain occupied by the homogeneous fluid at time 0 and , respectively.
The motion of the fluid is described by the relationship
(2.2)
where denotes the position of a material point in and in . The symbols and denote the
deformation gradient and the right Cauchy Green strain tensor between both configurations. The velocity
gradient and the rate of deformation tensor at time are denoted by and , respectively. For later
use, note that
(2.3)
The dot denotes the material time derivative.
2.2.1. Kinematic Parameters for 2D Flows

Consider in , a material fiber with a unit orientation which deforms into a material fiber
with a unit orientation at time . With denoting the length stretch . It follows [6] (p. 177)
that:
(2.4)
while is given by
(2.5)
Good mixing quality requires high values of throughout time and space. A local evaluation of the
efficiency of mixing [7] (p. 177) is given by the ratio
(2.6)
where
(2.7)
The values of this instantaneous efficiency are always included in the interval [-1; 1]. Through manip-
ulation it follows:
Kinematic Parameters
(2.8)
Note that is a local measure along the path of a material point. The time-averaged efficiency is
defined as
(2.9)
However, there is another way to define a mean efficiency over time:
(2.10)
The physical interpretation of Equation 2.10 (p. 13) is the following: for one material point, at time
, is the ratio of the final stretching obtained at time over the total mechanical dissipation
until time .
The global efficiency over all of the material points initially distributed in the flow is defined as:
(2.11)
This global efficiency is the ratio of the output—the mixing obtained—(the total stretching of the
matter until time ) over the input—the "energy" —(the total mechanical dissipation until time ).
2.2.2. Kinematic Parameters for 3D Flows

For 3D flows, you can calculate the local stretching of infinitesimal surfaces by the mean of the area
stretch .
In the initial configuration , consider an infinitesimal surface with a normal direction . This
surface deforms with time. At time , the surface is noted , with a new normal direction . The
area stretch , is the ratio of the deformed surface at time over the initial surface .
(2.12)
For an incompressible fluid:

(2.13)
If the fluid is incompressible, the normal direction to the surface is:
(2.14)
The Mixing Theory
A good mixing quality requires high values of throughout time and space. A local evaluation of the
efficiency of mixing [7] (p. 177) is given by the ratio
(2.15)
The values of this instantaneous efficiency are always included in the interval [-1; 1]. After some
transformations:
(2.16)
Note that is a local measure along the path of a material point; the time-averaged efficiency is
defined as
(2.17)
However, time-averaged efficiency can also be defined as:
(2.18)
As for Equation 2.10 (p. 13), the physical interpretation of Equation 2.20 (p. 15) is the following: for
one material point, at time , is the ratio of the final stretching obtained at time , over the
total mechanical dissipation until time .
Like for 2D kinematic parameters, a global efficiency over all the material points initially distributed
in the flow is defined:
(2.19)
2.2.3. Statistical Analysis

An initial orientation is assigned to material points (which does not need to be identical for all
points). While tracking the material points as a function of time, successive values of , , and
are also calculated. A global representation of , , and is again obtained by associating a ma-
terial point with a small rectangle of size — the color of the rectangle is associated with the
value of the field to be represented. The quality of the representation increases with the number of
points ( ).
When the number of material points is sufficiently large, you may proceed with a statistical treatment
of the calculated quantities.
Kinematic Parameters
2.2.3.1. Mean and Standard Deviation

You can calculate the time evolution of the mean and its standard deviation for any scalar kinematic
parameter α:
(2.20)
(2.21)
2.2.3.2. Cumulated Probability Function (or Distribution Function)
The distribution function associated with the scalar field is defined by the quantity ,
as follows:
(2.22)
where the right-hand side is the probability that the field be smaller than at time .
A new graph of the distribution function is calculated at every time, (see figure below).
Figure 2.1: Distribution Function Graph
2.2.3.3. Density of Probability Function
Based on the distribution function of a scalar field , the density of probability function
is defined as follows,
(2.23)
The function is the frequency with which the value of is in the range at
time . A new graph of the density of probability function is calculated at every time (see figure
below).
The Mixing Theory
Figure 2.2: Density of Probability Function Graph
2.2.3.4. Percentiles
An easier representation of the mixing progress is based on the time dependence of percentiles.
For the field , is defined such that
(2.24)
indicates that, at time , % of the material points have a value of lower than , as
shown in the figure below:
Figure 2.3: Percentiles Graph
Using the percentiles, you can study the evolution of the mixing for specific fractions of the popu-
lation of material points. For example, you can determine the value of the length of stretch reached
by the 5 or 10% of the points with the lowest stretching; these percentiles can easily show local
defects in the stretching.
2.2.3.5. Histograms
Another way to represent the frequency of values of a field is to define histograms. You specify
a set of intervals of values of and obtain the percentage of the points population that have a
value of in each interval at time (see figure below):
Homogenization
Figure 2.4: Histograms Graph
Here, % of the points population has a value of between and at time .
2.2.3.6. Correlations
Once the number of material points is sufficiently large, you can examine the correlation between
fields, either at the same or at different times.
You have to define two times ( and ), and two fields ( and ). For every material point, you
plot in abscissa and in ordinate. Analysis of the graph reveals a possible correlation
between the fields.
Figure 2.5: Correlations Graph
2.3. Homogenization
2.3.1. Definition
Suppose you want to mix two fluids (A and B) and both fluids occupy two separated zones of the
flow domain at time (see the following figure).
Figure 2.6: Mixing Fluids
The Mixing Theory
Let denote the concentration of fluid A throughout the mixing process. Since no diffusion
occurs between fluids A and B, equals either 0 or 1 and remains constant along the trajectory of a
material point. Its evolution is governed by the transport equation
(2.25)
The concept of concentration allows introduces the idea of a segregation scale [5] (p. 177), [8] (p. 177).
At time , consider a set of pairs of material points separated by a distance . For the th pair at
time , let and denote the concentrations at both points of the pair. Also, let denote the av-
erage concentration of all points and the standard deviation. At time , the correlation coefficient
for the concentration is defined as follows,
(2.26)
The function gives the probability of finding a pair of random points with a relative distance
and with the same concentration.
Figure 2.7: The Correlation Coefficient for Concentration over Distance
The previous figure shows a typical graph of as a function of . Let be such that . When
, you cannot predict whether the members of the pair have the same concentration or not. The
segregation scale is defined as
(2.27)
It is easy to understand that is a measure of the size of the regions of homogeneous concentra-
tion. decreases as mixing improves.
Homogenization
While quantities such as , , and are proper to the flow, irrespective of the initial concentration,
the segregation scale is a quantity that is affected by the flow but it is highly dependent upon
the initial concentration distribution.
Note:
Dankwertz defined another parameter, the intensity of segregation:
(2.28)
Because of our assumptions (the concentration attached to any material point remains
constant with time), this parameter does not change with time and will not be calculated.
2.3.2. Numerical Method

A set of material points are uniformly (but randomly) distributed in all of the flow domain at time
. Each point has an associated concentration depending on its situation in a zone of fluid A
or a zone of fluid B .
As time progresses, the new location of each point is calculated. To visualize the current state of the
concentration field, each point is plotted in the domain as a small pixel (if 2D flow) with the appro-
priate color. Once the successive coordinates of material points are stored, the concentration at time
is calculated corresponding to another set of initial conditions. A limitation of this method is in the
size of the material point identified by the pixel, but the number of points can be increased with the
computation time.
If material points are tracked, the average and the standard deviation of the concentration are
calculated as follows:
(2.29)
(2.30)
where is the concentration (0 or 1) of the material point .
is the maximum distance between two material points in the flow domain. nbval values of the
correlation function for distances uniformly distributed between 0 and are calculated. To calculate
one value of correlation function (for a distance ), pairs of points are randomly selected if
their relative distances are in the interval . The correlation coefficient
is completed by a linear interpolation through the discrete calculated values. The segregation scale
is then calculated by numerical integration on the basis of Equation 2.27 (p. 18).
The Mixing Theory
Figure 2.8: Calculated Values of the Correlation Coefficient
There is a limitation to this method. The mean size of the pixels is finite because the number of points
is finite. Therefore a segregation scale that is smaller than this characteristic size cannot be accurately
calculated. If the segregation falls below that size, the average thickness of the striations in the flow
is smaller than the size of the pixels and the concentration field appears as a random distribution of
pixels of the two colors; There are no more continuous lines of one color. Another problem of the
segregation scale is that it cannot detect a local defect in the flow. It is a global index of the quality
of mixing. Essentially, the segregation scale depends on the size of the flow domain.
The parameters of this method are:
• The initial distribution of the zones of fluid A and B.
• : the maximum possible distance between two points in the flow domain.
• : the number of pairs of points necessary to calculate one value of the correlation function.
• : the number of values of the correlation function to calculate.
2.4. Distributive Mixing
2.4.1. Distribution Index

Suppose you want to distribute a cluster of particles initially concentrated in a small box (see figure
below). It is assumed that the particles do not affect the flow field and that there is no interaction
between them.
Figure 2.9: Distributing Particles from a Small Box
Distributive Mixing
The flow distributes this set of points as a function of time. It is important to define a distribution
index to quantify this process. Its definition is based on the work performed by Manas-Zloczower
and her colleagues [9] (p. 177), [10] (p. 177).
At time there are points distributed in the flow domain.
Option 1: These points can form pairs of points. For each pair of points and , their
inter- distance is calculated. The maximum inter-distance will be of the order of the
diameter of the mixer.
Option 2: The inter-distance, between each point and its closest neighbor, is stored.
Thus, there are only distances. Using this method of calculation helps to better discriminate dis-
tributive capacities of similar mixers. The maximum inter-distance will be of the order of ,
where is the volume of the mixer.
With this set of distances, you can calculate the density of probability function on the distance :
the probability to find a pair of points (chosen randomly) such that their inter-distance is included in
range at time is: .
Figure 2.10: Density of Probability over Distance
Now suppose you have randomly distributed the same number of points throughout the flow domain.
You can assume that this is an ideal distribution. Using the same tools, you can calculate the function
for this optimal distribution. It is noted .
The distribution index is defined as the deviation of the function (real distribution) from the
function (optimal distribution):
(2.31)
As the distribution improves, the index decreases. This index is dimensionless — it is independent
of the size of the flow domain. The evolution of depends on the initial position of the box. Another
important parameter is the number or material points to distribute; Careful analysis must be done to
measure its influence.
Two other parameters can also be evaluated:
The Mixing Theory
• The difference of the means.
(2.32)
• The difference of the standard deviations.
(2.33)
Figure 2.11: Real and Optimal Distributions
Because these parameters are global indices, they do not detect local defects.
2.4.2. Distribution in Zones

The distribution index is used to quantify distributive mixing. Using this next method, you can detect
zones in the mixer where material points are missing, and where there is an excess of points.
As for the distribution index, there is a cluster of particles initially concentrated in a small box (see
figure below). It is assumed that the particles do not affect the flow field and that there is no interaction
between them.
Figure 2.12: Distributing Particles from a Small Box
As a function of time, the flow distributes this set of points.
The flow domain is defined as a set of adjacent and non-overlapping zones. In this case there are
four zones, shown below.
Distributive Mixing
Figure 2.13: Zones of the Flow Domain
The same number of points , is then distributed randomly throughout the flow domain (it is assumed
that this is the optimal distribution).
You determine the number of points in each zone, for both distributions at time . Based on these
numbers, you can evaluate a relative error of distribution for each zone, .
(2.34)
is the number of points of the real distribution included in zone , at time . is the number
of points of the optimal distribution included in the same zone.
If is zero for a zone, the optimal number of points are in that zone.
If is negative for a zone, there is a lack of points in that zone as compared to the optimum.
If is positive for a zone, there are too many points in that zone as compared to the optimum.
A global index based on all the zones is defined as,
(2.35)
The number of points and the zone partitioning influences the indices described above. When com-
paring two different mixers, it is recommended that you keep the ratio number of points and zones
constant. In order to have relevant results, this ratio should be higher than 100.
2.4.3. Deviation of Points Concentration

As for the distribution index, the goal is to quantify distributive mixing. However, for this parameter
there is need to compute a perfect points distribution.
As for the previous parameters, there is a cluster of particles initially concentrated in a small box (see
figure below).
The Mixing Theory
Figure 2.14: Points Within a Cluster
As a function of time, the flow distributes this set of points.
At time , there are points distributed throughout the flow domain.
For each point , the number of neighbors, , within a sample radius are determined. Depending
on the dimensionality (dim) of the cluster of points at time , you can evaluate the local points con-
centration :
(2.36)
where is the number of points of the cluster around at a distance smaller than the radius.
With a perfect distribution of the points, the points are equally distributed throughout the entire flow
domain. That is, there are the same number of points per unit volume found everywhere in the domain.
The perfect points concentration, , corresponds to the number of points divided by the volume of
the flow domain. You can adapt this method for determining the perfect points concentration for
other situations (for additional information, see The Points Concentration Deviation Function (p. 158)).
The standard deviation of points concentration at time is evaluated as
(2.37)
Where is the number of points in the cluster at time . correspond to the location of point at
time and is the perfect points concentration.
Note that the points concentration is only evaluated at positions where there are points. There is no
points concentration evaluation in zones of the mixer that do not contain material points. If distributive
mixing improves, the points concentration deviation should decrease. At perfect distribution, the
same points concentration is found at any location in the cluster, and the deviation is zero.
2.5. Disagglomeration
Evaluation of the dispersive mixing of solid particles in a fluid matrix is done by studying the evolution
of the size of the agglomerates [2] (p. 177), [1] (p. 177). Consider a set of agglomerates of different sizes
at the start of the mixing in an internal mixer. If the volume of the mixer is subdivided into smaller
Disagglomeration
representative volumes, , each representative volume contains agglomerates of different sizes, as il-
lustrated below:
Figure 2.15: A Representative Volume
If the number of agglomerates is large enough in volume , this distribution of agglomerates sizes
can be summarized in a mass density function , where is the time and is the size (mean dia-
meter) of the agglomerate. Its unit is [1/µm].
It is discretized by a piecewise-linear curve:
Figure 2.16: Discretized Mass Density Function
The mass fraction of agglomerates of size in interval is: .
This mass fraction distribution changes with time, as the volume attached to the point moves in
the mixer, because of erosion and rupture taking place during mixing. It is assumed that there is no
transfer of agglomerates between adjacent volumes, due to the high viscosity of the matrix.
Erosion is a slow and continuous process observed for all admissible sizes of agglomerates. This process
generates a lot of small particles. Rupture occurs when a critical stress is reached and is observed for
large particles. This process generates two or more agglomerates.
Erosion and rupture depend on the size of agglomerates, the shear rate and the shear stress. The fol-
lowing text distinguishes between the two types of solid particles: the aggregates and the agglomerates.
Aggregates are the smallest particles that cannot be eroded or broken anymore. Agglomerates are larger
particles formed of a number of aggregates linked together by cohesive forces.
The Mixing Theory
Figure 2.17: Disagglomeration Processes
Erosion
The erosion model implemented in Polystat is based on the work of Collin and Peuvrel-Disdier [4] (p. 177)
on the dispersive mixing of carbon black agglomerates N234 in a SBR matrix. Due to the large variety
of models and raw materials, it is possible to adapt or modify the implemented model (accessible in
the CLIPS file disagglomeration.clp, which can be found in $POLYFLOW/bin directory):
the corresponding functions are interpreted by Polystat at run time. Consider hereafter the case of a
low concentration of carbon black pellets, meaning that erosion due to friction between pellets is
neglected — only erosion due to hydrodynamic forces is considered.
During erosion, small particles are removed continuously from the agglomerates, which diminish in
size. This removal can occur if the shear stress is above a given threshold . In this case, assume
that the agglomerates are roughly spherical and the variation of size, , of the agglomerate after
seconds can be described as follows:
(2.38)
where is a coefficient of proportionality, is the local shear rate, is the size of the agglomerate at
time , and + is the new size of the agglomerate at time + .
Based on the assumption of mass conservation, the mass distribution function is:
(2.39)
As the total mass of solid particles is constant in the control volume , the mass fraction of aggregates
increases as the mass fraction of agglomerates decreases. This can be seen in the next figure, which
shows the effect of erosion on an initial set of agglomerates with sizes ranged between 15 and 25 mi-
crons. They are mixed in a matrix with a viscosity of 11000 Pa.s. A constant shear rate of 10 s-1 is applied
and the mass distribution function is plotted every 25 seconds. Observe the shift to left, the widening
and flattening of the Gaussian curve centered initially at 20 microns as erosion develops. Also, note the
increasing peak at the extreme left of the graph, in the small sizes, which corresponds to the generation
of aggregates.
Disagglomeration
Figure 2.18: The Effect of Erosion on the Mass Distribution Function over Time
Rupture
For the modeling of ruptures, (also based on [4] (p. 177)) it is assumed that an agglomerate ruptures
into a few fragments if it is submitted to a shear stress that is higher than a critical shear stress over a
given time period (called rupture time). This critical shear stress for the rupture function of size can
be written:
(2.40)
However, the inverse relationship is of more interest; which agglomerates break for a given shear stress
. Equation 2.40 (p. 27) becomes:
(2.41)
All agglomerates with a size greater than can break for a shear stress .
In this model, it is assumed that the rupture time is a constant (the default value is set to 0.1 s in the
disagglomeration.clp file) and that only a fraction of all particles of a given size will break if
rupture criteria are met. It is easily understood that not all particles of a given size have the same cohesion
or same impregnation level (infiltration of the matrix inside the agglomerate due to diffusion). This is
why you can specify the rupture rate — the fraction of particles that break when rupture criteria are
met (however, the default value is set to 1 in the disagglomeration.clp file, meaning that all
agglomerates break).
To model how a set of agglomerates of size will break in numerous fragments of various sizes, it is
assumed that the volumes of the fragments follow a Gaussian distribution between 0 and the parent
agglomerate volume. On average, the parent agglomerates (of size ) are cut into two fragments of
equal volume, leading to a mean size of 0.8 . This behavior can be modified by changing the corres-
The Mixing Theory
ponding function in the disagglomeration.clp file. The transfer function for rupture is shown in
the following figure, with a rupture rate of 1:
Figure 2.19: The Transfer Function for Rupture
Here are the details for implementing the rupture model. Associate each discretized agglomerate size
with an induction time initialized to zero. As the agglomerate progresses along its trajectory, the
shear stress is evaluated. Based on Equation 2.41 (p. 27), all the classes where ruptures can occur are
calculated. For classes < , their induction time is reset to zero. For the other classes, their associ-
ated induction time is increased by the local time step . For all classes where the induction time is
greater than the rupture time, the rupture occurs and the corresponding induction time is reset to zero
(the induction time is reset to zero for agglomerates that do not break, a phenomenon that occurs
when the rupture rate is not 100%). Classes where , but with an induction time less than the
rupture time, the agglomerates can fracture, but they do not break. Their mass frequency increases,
but their induction time, , must be modified, because it is assumed that incoming fragments come
with their own (zero) induction time.
The following rule is applied,
(2.42)
where is a function of the ratio of mass frequency of class at previous time and current time
+ . By default, is defined as:
(2.43)
The list of functions used in the erosion and rupture models available in the disagglomeration.clp
file:
• DNSPRB: define the initial mass density distribution function (default = Gaussian distribution between 15
and 25 µm).
Comment
• EROSION_MODEL: to define the function of erosion model (see Equation 2.38 (p. 26)).
• TRANSFER_RUPTURE: define the transfer function for rupture model.
• CRITICAL_SIZE: define the minimum size of agglomerates that can break for a given shear stress (see
Equation 2.41 (p. 27)).
• RUPTURE_TIME: define the amount of time during which the stress must be above the required threshold
in order to rupture.
• RUPTURE_RATE: define the fraction of agglomerates that will actually break if the rupture criteria are met.
• MODIFY_INDUCTION_TIME: define the function (see Equation 2.43 (p. 28)).
2.6. Comment
All of the mixing parameters discussed here are always defined as evolving with time. This kind of
representation is well suited for flow in a closed domain, where the mixing evolves with time.
But what if the flow occurs in an open domain, such as in a single screw extruder, or in a Kenics mixer?
In this case, the mixing quality evolves from the entry of the machine to the exit. To analyze this process,
a set of points is generated in the plane section of the entry. Then their trajectory through the machine
is calculated, until they reach the exit. For statistical analysis, a set of slicing planes are generated that
are uniformly distributed from the entry to the exit. For each slice, the intersections with the trajectories
are determined. At these intersections, the values of the kinematic parameters are interpolated. For
each slice, you can then calculate the mean value of a field , or the distribution function of a field ,
and so on. As the slices are sorted from the entry to the exit, you can analyze the evolution of the
mixing, slice by slice.
Chapter 3: Mixing Tasks in Polydata
3.1. Creating a New Mixing Task

In order to create a mixing problem, you must have defined and solved the Navier-Stokes equations
on the flow domain in a previous session. You will therefore have one or more result files containing
the velocity field. In this session, you will use these velocity fields as data for solving a mixing problem.
To begin, launch a Polydata session and read a mesh, to display the top level menu shown in Fig-
ure 3.1: The Polydata Menu (p. 32).
Mixing Tasks in Polydata
Figure 3.1: The Polydata Menu
Next, click Create a new task to open the menu shown in Figure 3.2: The Create a new task Menu (p. 33).
Create a new task
Creating a New Mixing Task
Figure 3.2: The Create a new task Menu
Perform the following steps to create the initial mixing task:
1. Specify that it is a mixing task.
MIXING task
2. Select one of the following to specify the type of the flow field that was previously calculated.
Steady-state problem
Time-dependent problem
3. Select one of the following to specify the nature of the geometry of the problem. Note that mixing tasks
are not available for 2D or 3D blow molding simulations.
2D planar geometry
2D axisymmetric geometry
2D 1/2 planar geometry
2D 1/2 axisymmetric geometry
4. Click Accept the current setup to open the mixing task menu, which is described in Using the General
Menu of a Mixing Task (p. 34).
3.2. Using the General Menu of a Mixing Task

Clicking the Accept the current setup menu item in the Create a new task menu opens a menu like
the one shown in Figure 3.3: The MIXING Task 1 Menu (p. 34).
Figure 3.3: The MIXING Task 1 Menu
Perform the following steps in this general menu:
1. Define the flow domain (not necessarily the complete domain defined by the mesh), as described in
Definition of the Flow Domain (p. 36).
Definition of the flow domain
2. Define the type of boundaries of this domain, as described in Definition of the Boundary Conditions (p. 37).
There are several kinds of possible boundary conditions: for example, a material point can cross an entry,
but not a wall.
Definition of the boundary conditions
3. Define the velocity fields (necessary to calculate the trajectory of the material points), as described in
Definition of the Flow Field (p. 45).
Definition of the velocity fields
4. Define the zones of the flow domain where the initial positions of the material points are generated, as
described in Parameters for the Generation of the Material Points (p. 49). The initial positions can be ran-
domly generated or distributed in an equidistant manner inside each generation zone.
Using the General Menu of a Mixing Task
Generation of the material points
5. Specify the numerical parameters necessary for the precision of the trajectory calculation in the flow, as
described in Parameters for the Tracking (p. 56)
Parameters for the tracking
6. You have the option of selecting the properties (such as mixing parameters, or pressure, temperature, and
so on) you want to calculate along the trajectories, as described in Selection of Properties (p. 58).
Selection of properties
7. Specify how to store the results of the simulation, as described in Parameters for the Storage of the Res-
ults (p. 61). You will use the created result files to make a global analysis of the mixing with Polystat.
Storage of the results
8. You have the option of defining moving parts that overlap the flow domain, as described in Definition of
Moving Parts (p. 63). This step can be necessary if the mesh superposition technique has been used previ-
ously to calculate the flow field.
Definition of moving parts
9. After you have defined the mandatory information needed for the mixing task (that is, performed
steps 1.–5. and 7.), you can enable and customize a preset statistical analysis. These analyses generate
typical sets of statistical results and allow you to proceed directly to postprocessing in ANSYS Poly-
stat after the solver has completed the calculation. The two preset statistical analyses that are
available are referred to as the following:
• the kinematic analysis
The kinematic analysis is focused on the kinematic parameters of the material points, and can
calculate results for a variety of functions (such as probability functions and density of probability
functions) on a set of properties (such as shear rate and stretching) that you select.
• the distribution analysis
The distribution analysis is focused on the spatial distribution of the material points, and can
calculate the global distribution index, the axial distribution index, and the transfer rates between
different parts of the mixer. Note that the distribution analysis requires that you have only one
generation zone, which must be shaped like a box: if the domain is closed, the box must be located
inside the flow domain; if the domain is open, the box must be located on the inflow boundary
and have no thickness in the direction of flow. See Parameters for the Generation of the Material
Points (p. 49) for details on creating the generation zone.
To enable a preset statistical analysis, repeatedly click Modify task type at the top of the menu
until the text of the option shows that the task type is set to the kind of analysis you want.
Modify task type
Depending on your task type selection, the following menu items become available; these additional
options allow you to customize the analysis to suit your problem, as described in Definition of the
Kinematic Analysis (p. 64) and Definition of the Distribution Analysis (p. 69).
Definition of a KINEMATIC analysis
Definition of a DISTRIBUTION analysis
3.3. Definition of the Flow Domain

Clicking the Definition of the flow domain menu item in the mixing task menu opens the menu shown
in Figure 3.4: The Flow domain Menu (p. 36).
Definition of the flow domain
Figure 3.4: The Flow domain Menu
With this menu, you can remove (or add) subdomains of the whole mesh in order to define the flow
domain, as shown in the example below:
Definition of the Boundary Conditions
Figure 3.5: Fluid and Solid Subdomains
The flow domain is the union of the subdomains s2, s3, and s4. The subdomains s1, s5, and s6 (which
are solid parts of the problem) should be removed from the upper list of the menu.
3.4. Definition of the Boundary Conditions

Clicking the Definition of the boundary conditions menu item in the mixing task menu opens the
menu shown in Figure 3.6: The Boundary conditions Menu (p. 38).
Definition of the boundary conditions
Figure 3.6: The Boundary conditions Menu
With this menu, you can do two things: you can select each boundary and specify the type as necessary
(such as wall, entry, or exit), as described in Specification of the Boundary Type (p. 38); and you can
use the button next to the Add, modify or remove stopping planes label to add "stopping planes",
as described in Specification of a Stopping Plane (p. 43).
3.4.1. Specification of the Boundary Type

Selecting a boundary from the list in the Boundary conditions menu and clicking the Modify button
opens a menu like the one in Figure 3.7: The Boundary condition along Boundary 1 Menu (p. 39).
Figure 3.7: The Boundary condition along Boundary 1 Menu
Boundaries are defined as non-penetrable by default. This type covers the following cases: walls, axes
of symmetry, and free surfaces. It specifies that a material point cannot cross the boundary. If, for
numerical reasons, a material point does cross the boundary, a specific flag attached to it is set and
the calculation of the trajectory is interrupted (the stopping is ABNORMAL).
Besides non-penetrable boundaries, the following types are available for selection:
• inflow
An inflow is a part of the boundary where the fluid enters into the flow domain. Normally, a mater-
ial point cannot cross an inflow boundary (that is, a material point cannot go against the flow).
• outflow
An outflow is a part of the boundary where the fluid exits the flow domain. When a material point
reaches an outflow boundary, the calculation is stopped. In this case, the stopping is NORMAL (as
opposed to ABNORMAL).
Figure 3.8: Inflow and Outflow Boundary Conditions
• spatially periodic flow options
The remaining options are specific to spatially periodic flows. Remember that a problem is spatially
periodic if there exists an elementary "module" on which the flow field can be calculated and where
the velocity field in the inflow section is equal (exactly) to the velocity field in the outflow section.
The flow field is repeated infinitely in space.
With this kind of flow, when a material point reaches the outflow, the calculation can continue by
injecting this particle back into the inflow; the particle will travel several times through the same
module. There are therefore two systems of coordinates: the first is "relative" to the coordinates in
the module; the second is "absolute" and is attached to the real trajectory in the real infinitely re-
peated domain. See Figure 3.9: Relative and Absolute Trajectories (p. 40) for an illustration.
Figure 3.9: Relative and Absolute Trajectories
In order to calculate trajectories in spatially periodic flows, the inflow must be connected with the
outflow. Note that due to a limitation of the program, the geometrical dimensions and the mesh
distribution in the inflow and the outflow must be equal.
To define the spatially connected boundaries, you should begin by clicking the boundary that is
the entry of the flow domain in the Boundary conditions menu. Next, select Entry boundary (in
a spatially periodic flow) in the menu that opens.
Entry boundary (in a spatially periodic flow)
The dialog box shown in Figure 3.10: Dialog Box for the Entry Boundary (p. 40) will then open.
Figure 3.10: Dialog Box for the Entry Boundary
Here you will enter the number of loops that the material points can do in the flow domain: the
calculation of a trajectory is stopped when the material point reaches the exit after this number
(or if the lifetime of this particle has expired).
After you click OK in the previous dialog box, the menu in Figure 3.11: The Corresponding boundary
Menu (p. 41) will open, where you can select the exit boundary that is connected with the entry.
Figure 3.11: The Corresponding boundary Menu
Finally, you will return to the previous menu, which has changed as shown in Figure 3.12: The Up-
dated Boundary condition along Boundary 1 Menu (p. 42).
Figure 3.12: The Updated Boundary condition along Boundary 1 Menu
You must now specify how the coordinates of the points in the entry section should be transformed
in order to relate to the corresponding coordinates in the exit section.
The general relation between points in the two sections is:
Click the Parameters to connect spatially periodic boundaries menu item and then specify the
coefficients of the rotation matrix and the translation vector in the dialog boxes that open
(see Figure 3.13: The Dialog Box for the Spatially Periodic Boundary Parameters (p. 43) for an ex-
ample).
Parameters to connect spatially periodic boundaries
Figure 3.13: The Dialog Box for the Spatially Periodic Boundary Parameters
Note:
When you choose the exit boundary connected to the entry boundary (of a spatially
periodic flow), the boundary type of the exit boundary is automatically updated. It is
not necessary to define it again.
3.4.2. Specification of a Stopping Plane

You can define “stopping planes”, so that the calculation is stopped when the material point reaches
a defined spatial plane. The purpose of adding stopping planes is to save CPU time when you are
only interested in the mixing in a fraction of the domain.
Another situation in which you might want to use stopping planes is for 2.5D planar flows, where
the mesh is 2D but the velocity field has 3 components (Vz is perpendicular to the mesh). For such
flows, there are two systems of coordinates: the "relative" coordinates for the trajectory in the plane
of the mesh, and the "absolute" coordinates in the real flow domain (3D). A stopping plane allows
you to continue to calculate the trajectory until the material point reaches (in the absolute system)
the given plane (see Figure 3.14: 2.5D Flow with a Stopping Plane (p. 43)).
Figure 3.14: 2.5D Flow with a Stopping Plane
Clicking the button next to the Add, modify or remove stopping planes label in the Boundary
conditions menu will open the menu shown in Figure 3.15: The Parameters for the stopping planes
Menu (p. 44).
Add, modify or remove stopping planes
Figure 3.15: The Parameters for the stopping planes Menu
If you click Add a new stopping plane, you have to respond to the dialog boxes that open, entering
the coefficients that define the plane (for example, see Figure 3.16: Dialog Box for Defining Coeffi-
cients (p. 44)) and specifying whether it is valid for points to cross the stopping plane (see Fig-
ure 3.17: Dialog Box for Specifying Trajectory (p. 45)).
Add a new stopping plane
Figure 3.16: Dialog Box for Defining Coefficients
Definition of the Flow Field
Figure 3.17: Dialog Box for Specifying Trajectory
You must be careful when defining a stopping plane: in the algorithm of the trajectory calculation,
after every time step of integration, the current position of the material point is entered in each plane
equation; if the results are all positive then there is no crossing, and the calculation continues. Other-
wise, the calculation is stopped and the trajectory is saved.
Figure 3.18: The Parameters for the stopping planes Menu with Multiple Stopping Planes (p. 45) shows
how the menu looks after the definition of multiple stopping planes.
Figure 3.18: The Parameters for the stopping planes Menu with Multiple Stopping Planes
Note that you can Modify a stopping plane or Remove a stopping plane, as necessary.
3.5. Definition of the Flow Field
3.5.1. Steady-State Flow

Clicking the Definition of the velocity fields menu item in the mixing task menu for a steady-state
flow simulation opens the menu shown in Figure 3.19: The Flow definition Menu for Steady-State
Flow (p. 46).
Figure 3.19: The Flow definition Menu for Steady-State Flow
Click the Enter the result filename and specify the name of the result file containing the velocity
field (in the Polyflow result file format). If the file is unformatted, click the Switch to binary (unformat-
ted) file menu item; the note at the top of the menu will reflect your choice and the menu item will
change to Switch to ascii (formatted) file.
3.5.2. Time-Dependent Flow

Clicking the Definition of the velocity fields menu item in the mixing task menu for a time-dependent
flow simulation opens the menu shown in Figure 3.20: The Flow definition Menu for Time-Dependent
Flow (p. 47).
Definition of the Flow Field
Figure 3.20: The Flow definition Menu for Time-Dependent Flow
Perform the following steps:
1. Select one of the following to specify the kind of flow.
Piecewise steady flow
Continuous transient flow
2. Specify the list of result files that contain the successive flow fields. There are two ways to specify this:
• Automatic selection of result files
If Automatic selection of result files is selected (the default), you use the menu items at the
bottom of the menu to specify the time step (constant) between two successive flows, the
prefix and the format of the Polyflow result files (one particular file for each flow), the number
of result files to be read, and the number ID of the first flow. For example, in Figure 3.20: The
Flow definition Menu for Time-Dependent Flow (p. 47), the flow is defined in 10 files, named
res.1, res.2, and so on, until res.10. The prefix is “res”. The first file is res.2. The time
step is 1 second between successive flows. Note that all these files must exist: there cannot be
missing result files from the specified sequence. If the lifetime of the material points is greater
than time_step * number_of_result_files, the succession of velocity fields are used in a loop.
• Manual selection of result files
If Manual selection of result files is selected, the menu items change, as shown in Fig-
ure 3.21: The Flow definition Menu with Manual Selection (p. 48).
Figure 3.21: The Flow definition Menu with Manual Selection
You must specify one by one the list of Polyflow result files that define the flow. To begin, click
Modify the list of result files to open the menu shown in Figure 3.22: The Flow definition : list
of result files Menu (p. 48).
Modify the list of result files
Figure 3.22: The Flow definition : list of result files Menu
In this menu, you can add, remove, and modify the result files, as well as change the order.
Parameters for the Generation of the Material Points
The parameters for a particular velocity field are summarized in a menu like the one shown in
Figure 3.23: The velocity field 1 Menu (p. 49).
Figure 3.23: The velocity field 1 Menu
You define a new velocity field in this menu, by specifying the name of the result file that stores
this field and whether it is formatted. Finally, you enter the time of use of the current velocity
field. If the lifetime of the material points is greater than the sum of all times of use, the succes-
sion of velocity fields will be used in a loop.
Note:
The manual selection mode is not practical if many (such as a dozen) result files must
be specified.
3.6. Parameters for the Generation of the Material Points

Clicking the Generation of the material points menu item in the mixing task menu opens the menu
shown in Figure 3.24: The Generation of material points Menu (p. 50).
Generation of the material points
Figure 3.24: The Generation of material points Menu
This menu allows you to define zones where the material points are initially positioned. In order to have
a well-defined mixing problem, you must have at least one generation zone. Note that if you plan to
use the distribution analysis to generate a statistical analysis of the mixing, you must create only one
generation zone, which must be shaped like a box: if the domain is closed, the box must be located
inside the flow domain; if the domain is open, the box must be located on the inflow boundary and
have no thickness in the direction of flow.
As soon as more than one generation zone is defined, a new property is added to the list of properties
evaluated along the trajectories and enabled in the Selection of properties menu (see Selection of
Properties (p. 58)). This new property, named “label”, is constant along each trajectory and indicates
the zone ID of the material point’s origin.
When you have generated one or more zones (as shown in Figure 3.25: The Generation of material
points Menu with Multiple Zones (p. 51)), you can modify the parameters of zones or delete zones.
Figure 3.25: The Generation of material points Menu with Multiple Zones
The following are definitions for the terms used in the menu shown in Figure 3.25: The Generation of
material points Menu with Multiple Zones (p. 51):
• The index is the identification number of the zone. For example, in the previous figure the second zone
(that is, ID=2) is the flow domain.
• The intensity identifies the relative number of points that are generated in the various zones. The previous
figure shows two defined zones: every time 4 points are generated in the second zone (flow domain), only
1 point is generated in the first zone (a box in the flow domain).
You can add a new generation zone by using the menu shown in Figure 3.26: The Generation Zone 1
Menu (p. 52) opens.
Add a new generation zone
Figure 3.26: The Generation Zone 1 Menu
You have the following options:
• You can modify the name of the generation zone (this is optional and local to Polydata).
Modify the name of this generation zone
• You can make a selection from the following to specify how the zone is generated:
– This zone is all the flow domain
The generation zone is the entire flow domain. The initial positions of material points are randomly
generated in the flow domain. This option is well suited for flows in a closed domain.
– This zone is a box included in the flow domain
You define a box by its two corners (xmin, ymin, zmin) and (xmax, ymax, zmax). This box must
have a non-empty intersection with the flow domain. Two modes of generation are available. With
the first mode, the initial positions of the material points are randomly generated in the box. With
the second mode, referred to as the “equidistant distribution”, the material points are initially dis-
tributed at equal distance d (which you specify) between neighboring points. If the flow domain
is 2D, the points are distributed at vertex positions of a lattice of identical equilateral triangles (see
Figure 3.27: Equidistant Distribution of Points in a 2D Box (p. 53)). If the flow domain is 3D, the
points are distributed at locations corresponding to the centers of close-packed equal spheres. The
generated points that are outside the flow domain will be rejected.
Figure 3.27: Equidistant Distribution of Points in a 2D Box
– This zone is all the inflows of the flow domain
In a previous menu, you have defined the boundaries that are the inflows of the flow problem.
With this option, the initial positions of material points are randomly generated in these inflows;
this option is well suited for flows in an open domain.
– This zone is a topo-object
With this option, you can define a zone by selecting a topo-object from a list. If the list does not
have appropriate topo-objects, it is possible to add new ones, as described in a later step.
– This zone is one inflow of the flow domain
In a previous menu, you have defined the boundaries that are the inflows of the flow problem.
With this option, you select one of these inflow sections. The initial positions of material points are
randomly generated in the selected boundary. This option is well suited for flows in an open domain.
– This zone is a CSV slice included in the flow domain
With this option, you can specify your own initial distribution of points, saved in a comma-separated
values (CSV) file. After selecting this option, use the dialog box that opens to specify the name of
the CSV file.
A question dialog box will then ask if you agree to have all fields reinitialized to default values.
Click Yes if you just want the coordinates of the points to be read from the CSV file. The other
properties will be reinitialized to their default values: time = 0; logarithm_of_stretching = 0;
space_integration = 0; and so on.
Click No in the question dialog box if you want Polyflow to read not only the coordinates of the
points in the CSV file, but also some other properties. If they are found, they will be used to initialize
the property, otherwise the default values will be used for initialization. The properties that are
read in the CSV file are as follows (note that the coordinates must be the first property written in
the file):
“COORDINATES” (in upper case, always with three components)
“time” “logarithm_of_stretching”
“space_integration” “direction_of_stretching”
“label”
All the points defined in the CSV file that are included in the flow domain will be tracked. The
other points will be rejected.
Another use of this option is the following: assume that you divided your flow simulation into
several sections, to simplify and reduce the size of the problem. The flow in each section is evaluated
separately. If you then want to track material points across all the sections, you would begin by
defining a first tracking through the first flow section. Next, with the generated mixing files, perform
a slicing in Polystat at the exit of the first flow domain and save the slice in a CSV file. You can
then define a second mixing task in Polydata, on the second flow section; the generation zone will
be the CSV file you created in Polystat. In this way, you will continue the tracking in the second
flow section. In such a case, it is useful to click No in the question dialog box, in order to keep the
history of deformation and stretching throughout the whole flow domain.
• You can modify the intensity of the generation of points, in order to generate a higher frequency of points
in one zone in comparison with the others. Clicking this menu item opens a dialog box, which you can use
to modify the intensity factor of the created zone. A higher intensity value results in a higher frequency of
point generation in that zone.
Modify the intensity of the generation of points
Important:
If some of the generation zones are based on CSV files, you must set the intensity of each
zone equal to the number of material points starting from that zone. Then be sure to set
the number of result files and the maximum number of trajectories per file in the Storage
of the results menu (as described in Parameters for the Storage of the Results (p. 61))
such that their product is equal to the sum of the intensities.
Important:
If some of the generation zones are defined as boxes with equidistant distribution, it is
difficult to estimate the exact number of points that will eventually be generated in those
zones. That is why an approximation (an upper bound) of the number of points that will
be generated is provided to you when defining the box generation zone. This number can
be used to appropriately define the intensity of generation, the number of result files, and
the maximum number of trajectories per file, as described in the previous note.
Note that the management of topo-objects is done in the Generation of material points menu:
• Create a new topo-object
You can Create a new topo-object by clicking this menu item and using the menu shown in Fig-
ure 3.28: The Add topo-objects Menu (p. 55).
Figure 3.28: The Add topo-objects Menu
To define a topo-object, first specify two objects by making a selection from the list and then clicking
the appropriate button (for example, Select object 1). Note that an object like Subdomain
1*Boundary 5 refers to the intersection of Subdomain 1 and Boundary 5. Then select an operator
(intersection or union), enter a name for the new topo-object, and click create.
Important:
When using the intersection operator, you must be sure that the resulting object is not
empty, as there is no check to avoid such a situation. Running Polyflow with this kind of
degenerated generation zone will lead to a fatal error.
• Modify the name of a topo-object
You can Modify the name of a topo-object by clicking this menu item and then using the Modify
name of topo-objects menu that opens. You should select an existing object in the list, click Select
object, enter a new name, and then click modify.
• Delete a topo-object
You can Delete a topo-object by clicking this menu item and then using the Delete topo-objects
menu that opens. You should select an existing object in the list, click Select object, and then click
delete.
Enter a time interval (Tmin, Tmax)
When the mixing task is transient and if the flow is open (i.e., you have at least one inflow and one
outflow boundary), it is possible to specify the time interval for the generation of points. This is performed
in the Generation of material points menu using the Enter a time interval (Tmin, Tmax) setting.
You are prompted to enter a minimum and a maximum value for the time interval. The generation of
points can be performed in any type of generation zone (a box, in all inflows, etc.).
3.7. Parameters for the Tracking

In order to calculate the trajectory of a material point, you have to specify a set of numerical parameters.
These parameters have different purposes, such as deciding when to stop the calculation or when there
is a problem. Clicking the Parameters for the tracking menu item in the mixing task menu opens the
menu shown in Figure 3.29: The Parameters for the tracking Menu (p. 56).
Parameters for the tracking
Figure 3.29: The Parameters for the tracking Menu
You must modify the following two parameters:
• Modification of TIMAX
TIMAX is the lifetime (in seconds) of the material points. Their trajectories will be calculated until that
time is reached. In order to lower the CPU time of your simulation, do not use unnecessarily large
values.
• Modification of VNORMX
Parameters for the Tracking
VNORMX is the maximum velocity magnitude in the flow field, and is used to detect stagnation points
in the flow. A stagnation point is a point that has a negligible velocity in comparison with the max-
imum velocity. By default, “negligible” is defined as one millionth of the maximum velocity.
You can modify some numerical parameters acting on the accuracy of the integration scheme by
clicking Options and using the menu shown in Figure 3.30: The Parameters for the tracking Menu (p. 57).
Options
Figure 3.30: The Parameters for the tracking Menu
You can modify the following parameters:
• Modification of NBELEM
NBELEM affects the number of integration steps used to cross one element of the finite element
mesh: the default value of 3 requests that, in the mean, any material point crosses one finite element
in 3 steps. This parameter is important for the accuracy of the calculation of kinematic parameters:
higher values for NBELEM result in better accuracy. It is not necessary to set NBELEM higher than 3,
as the intrinsic accuracy of the method is reached with this value and higher values would be expensive
in CPU time.
• Modification of EPSPNT
EPSPNT is the tolerance on a distance. Two points are identical if the distance between these points
is smaller than EPSPNT.
• Modification of EPSVIT
EPSVIT is the tolerance on a velocity. EPSVIT is used to determine if a point is a stagnation point: if
the local velocity is lower than EPSVIT * VNORMX, then the material point at time is a
stagnation point.
• Modification of EPSTIM
EPSTIM is the tolerance on a time step. A time step smaller than EPSTIM is considered to be zero; it
will be used to stop the iterative Newton-Raphson procedure that finds the time step needed to
reach the border of the current finite element containing the point.
Figure 3.31: The Trajectory of a Material Point
Remember that the trajectory of a material point is calculated piece by piece: the velocity is integrated
in a finite element until the border of this element is reached. Then the adjacent element where the
calculation is to continue is identified, and so on. It is therefore important to precisely determine the
time step needed to reach the border of the current element containing the material point.
3.8. Selection of Properties

Clicking the Selection of properties menu item in the mixing task menu opens the menu shown in
Figure 3.32: The Selection of properties Menu (p. 59).
Selection of properties
Selection of Properties
Figure 3.32: The Selection of properties Menu
In this menu, you can start with Parameters for the kinematic variables, which allows for the setting
of the numerical parameters necessary for the calculation of the mixing parameters evolving along the
trajectories (see Parameters for the Kinematic Mixing Properties (p. 60) for details). Note that the default
numerical parameter settings are appropriate in many applications, so this step may not be necessary.
The next step is to enable the properties you want to store in the mixing result files, by clicking the
appropriate menu item. All enabled properties will be listed at the top of the menu. Note that the time
and the coordinates are always saved.
The first property, the space integration , is the length of the trajectory up to the current time :
(3.1)
The following properties are kinematic parameters that are defined in The Mixing Theory (p. 11): rate
of stretching, rate of dissipation, direction of stretching, and logarithm of stretching.
• For 2D flows:
– The rate of stretching is equal to .
– The rate of dissipation is equal to .
– The direction of stretching is the vector .
– The logarithm of stretching is the natural logarithm of .
• For 3D flows:
– The rate of stretching is equal to .
– The rate of dissipation is equal to .
– The direction of stretching is the vector .
– The logarithm of stretching is the natural logarithm of .
The following properties are based on the fields stored in the result files needed for tracking: pressure,
velocity, and temperature. These properties come from the flow calculation. If these fields are not in
the result files, they are initialized to zero.
The following properties are useful to evaluate the accuracy of the calculation: divergence of velocity
and determinant of F.
The following properties are based on fields stored in the result files needed for tracking: viscosity,
mixing index, and first eigenvalue of tensor . These properties are postprocessors defined as additional
sub-tasks to the flow calculation. If these fields are not in the result files, they are initialized to zero.
These three fields are especially useful for the analysis of the dispersive mixing. The mixing index M in-
dicates whether locally the flow is “rigid” (M = 0), a shear flow (M = 0.5), or an extensional flow (M =
1). Moreover, the first eigenvalue of the extra-stress tensor indicates the local stress, which is an im-
portant parameter for the evaluation of the capillary number.
You should enable only the properties that are necessary, in order to save time and memory.
3.8.1. Parameters for the Kinematic Mixing Properties

Clicking the Parameters for the kinematic variables menu item in the Selection of properties
menu opens the menu shown in Figure 3.33: The Parameters for the kinematic variables Menu (p. 60).
Parameters for the kinematic variables
Figure 3.33: The Parameters for the kinematic variables Menu
Parameters for the Storage of the Results
As a material point moves in the flow, a small volume of matter attached to it will deform. To calculate
its stretching and rate of stretching, you need to specify the direction in which this stretching is
measured (as explained in the theoretical background). You can select from the following two options:
• Random initial direction of stretching
You do not specify a direction: the computer will randomly generate an initial direction of
stretching. This direction will be different from material point to material point.
• Imposed initial direction of stretching
You specify a direction (dx, dy, dz): every material point has the same initial direction.
Over time, the two options are, in fact, equivalent from a statistical point of view.
You can also select from the following to define the way in which tensor (the gradient of deformation
tensor) is calculated. You can choose to leave tensor unconstrained, or you can constrain it such
that the determinant of tensor always remains 1:
No constraint over the tensor F
Constraint over the tensor F : detF = 1
2D simulations do not present any problems, as the method guarantees that the determinant of
will remain 1 (incompressibility); it does not matter whether you choose to constrain the tensor or
leave it unconstrained.
For 3D flows, on the other hand, you cannot be sure that the tensor will remain 1 along a trajectory
of a material point. If you do not constrain the tensor , a future analysis of the results in Polystat
will show if the calculations were accurate (that is, if det remains 1 along the trajectories). In the
other case, unfortunately, if you constrain such that det remains 1 along trajectories, you cannot
be sure that this will improve the accuracy of the results; for this reason, it is recommended that you
retain the default setting in which is not constrained.
3.9. Parameters for the Storage of the Results

Clicking the Storage of the results menu item in the mixing task menu opens the menu shown in
Figure 3.34: The Storage of the results Menu (p. 62).
Storage of the results
Figure 3.34: The Storage of the results Menu
With this menu, you will specify how the calculated trajectories and evolution of different properties
are stored. These results can be stored in several files, in order to be analyzed by Polystat. The menu
items allow you to revise the settings for the following:
• result files
You can modify the prefix of the result files that are created, which is set to mixing by default. The
full name of the result files will be of the following form: [prefix]_ x .mix, where x varies from
1 to the total number of files (which you can specify). You can also specify whether the files are
formatted (ASCII) or not (binary). Despite the fact that the unformatted files are smaller in size (and
so take up less memory space), the default option is formatted.
• trajectories and CPU time
You can specify the maximum number of trajectories that are stored in a single result file, as well as
the maximum CPU time (in hours) that is spent before the current file is closed and results begin to
be stored in the next file.
In this way, files are continuously generated: you can begin the statistical analysis (on a short popu-
lation of material points), while at the same time Polyflow continues to work. If you are happy with
the results, you can interrupt Polyflow. You have the option of running Polyflow again with the same
data file; it will begin to generate a new result file without erasing the result files from an old session.
• storage
You have to select one of the five available ways to store trajectories:
Definition of Moving Parts
– You can store all of the calculated points of the trajectories. Note that this can be very expensive in terms
of memory space.
– You can store the positions of the material point every dt seconds exactly, where dt is the time step specified
in a later step.
– You can specify that successive stored positions have a minimal time step of dt seconds, where dt is the
time step specified in a later step.
– You can store a position every time the length of the trajectory has increased a distance of dl, where dl
is the displacement specified in a later step.
– You can store a new position if the increase in the trajectory length is greater than dl, where dl is the dis-
placement specified in a later step. The increase in this case is the difference between the trajectory length
for the current stored position and the trajectory length for the previous stored position.
• time step
If you specified that you wanted the material point positions stored based on dt (as described previ-
ously), you can modify the exact or minimum time step.
• displacement
If you specified that you wanted the material point positions stored based on dl (as described previ-
ously), you can modify the exact or minimum displacement.
3.10. Definition of Moving Parts

Clicking the Definition of moving parts menu item in the mixing task menu opens the menu shown
in Figure 3.35: The Define moving parts Menu (p. 63).
Definition of moving parts
Figure 3.35: The Define moving parts Menu
With this menu, you specify the moving parts (if any) overlapping the flow domain. This data is used
to remove particles that are randomly generated in the flow domain and also included in a given
moving part.
Important:
You must specify the moving parts in the same order as in the previous flow task.
First, click the Creation of a new moving part menu item to open the menu shown in Figure 3.36: The
moving part #1 Menu (p. 64).
Creation of a new moving part
Figure 3.36: The moving part #1 Menu
For this moving part, you specify its domain and its title.
3.11. Definition of the Kinematic Analysis

If you have set the task type to KINEMATIC analysis in the mixing task menu (by clicking the Modify
task type menu item repeatedly, until the task type is defined as shown in Figure 3.37: A Kinematic
Analysis Mixing Task (p. 65)), you can click the Definition of a KINEMATIC analysis menu item to open
the menu shown in Figure 3.38: The Definition of the KINEMATIC analysis Menu (p. 65).
Definition of the Kinematic Analysis
Figure 3.37: A Kinematic Analysis Mixing Task
Definition of a KINEMATIC analysis
Figure 3.38: The Definition of the KINEMATIC analysis Menu
The Definition of the KINEMATIC analysis menu allows you to modify the default settings of the kin-
ematic analysis (as described in Using the General Menu of a Mixing Task (p. 34)), in order to customize
the preset statistical analysis of the mixing to fit your problem. Note that at any time you can revert to
the default settings by clicking Reset to default values.
To begin, you can modify the parameters for the slicing of the trajectories, including the number of
slices, the starting position (for an open domain), the starting time (for a closed domain), the increment
between slices, and the direction of slicing. For a closed domain, the slices represent moments in time;
such slicing allows you to analyze the time evolution of the stretching distributions in a batch mixer,
for example. For an open domain, the slices represent parallel spatial planes that are perpendicular to
a coordinate axis, which begin at the inflow boundary and proceed to the outflow boundary; such slicing
allows you to analyze the spatial evolution of the residence time distributions in a static mixer, for ex-
ample.
The Management of properties menu item allows you to specify which properties are evaluated in
the functions, by adding or removing them from the list. See Management of Properties (p. 67) for
details.
You also have the ability to select the statistical functions that will be evaluated for a specified set of
properties. You can choose to evaluate the mean and standard deviation functions, the percentile
functions (which are 10%, 50%, and 90%, by default), the probability functions, and the density of
probability functions (with or without smoothing). Note that if you added concentration to the list of
properties using the Management of properties menu item, the previous functions (mean and standard
deviation functions, and so on) will not be calculated for concentration as they are for the other prop-
erties; instead, you have the option of enabling the Evaluation of the segregation scale, which will
only be calculated for concentration.
You can click Show advanced options to add menu items to the menu (see Figure 3.39: Advanced
Options for the Definition of the KINEMATIC analysis Menu (p. 67)). These additional menu items allow
you to modify some of the numerical parameters involved in the evaluation of the statistical functions.
Show advanced options
Definition of the Kinematic Analysis
Figure 3.39: Advanced Options for the Definition of the KINEMATIC analysis Menu
If you want to make major modifications to the slicing method, you can disable the automatic settings.
This action will make menu items available that allow you to modify the slicing type and the slicing
coordinates. Note that such modifications are not recommended, unless you have extensive expertise
in the subtleties of Polyflow.
The remaining menu items allow you to modify the number of values used to represent the various
curves, as well as the number and values of the percentiles.
3.11.1. Management of Properties

Clicking the Management of properties menu item in the Definition of the KINEMATIC analysis
menu opens the menu shown in Figure 3.40: The Management of properties for the KINEMATIC ana-
lysis Menu (p. 68).
Management of properties
Figure 3.40: The Management of properties for the KINEMATIC analysis Menu
This menu shows the list of properties that will be evaluated as part of the kinematic analysis. The
addition of the phrase + stats indicates that for that property, statistical functions will be evaluated
for each slice; otherwise, the properties are just evaluated along the trajectories. To add or remove
+ stats, simply click the menu item associated with that property.
You can Remove a property from the list, if there is a property that is not useful for your analysis.
You can also Add a property to the list. Besides the properties selected by default (as shown in
Figure 3.40: The Management of properties for the KINEMATIC analysis Menu (p. 68)), you can add
the following:
• space_integration
• direction_of_stretching
• pressure
• temperature
• divergence_of_velocity
• determinant_of_tensor_F
• viscosity
• mixing_index
• first_eigenv_of_tensor_T
• melting_index
Definition of the Distribution Analysis
• concentration
• min(shear_rate)
• min(first_eigenv_of_tensor_T)
• max(first_eigenv_of_tensor_T)
• min(mixing_index)
• max(mixing_index)
Note that in order to be evaluated, the following properties must exist in the Polyflow result file(s)
used to perform the computation of the material points trajectories: velocities, pressure, temperature,
viscosity, mixing_index, and first_eigenv_of_tensor_T. Other properties are derived from these. For
example, the melting index is a function of time, viscosity, and temperature; therefore, these last
three properties must exist in the Polyflow result file(s) if you want to evaluate the melting index.
Still other properties are evaluated during the computation of the trajectories, such as the logarithm
of stretching.
If you add concentration to the property list, you must define zones where the concentration field is
set to 1. If a material point has its starting position in one of these zones, it will be assigned a unit
concentration; otherwise, it will be assigned a null concentration. The initially assigned concentration
value is transported unchanged along the trajectory for every point, and the diffusion and the
chemical reactions are not taken into account. You will be prompted to define the concentration
zones when you add concentration to the list. You also have the option of defining zones at a later
point by clicking Define zones where concentration=1 in the Management of properties for the
KINEMATIC analysis menu. Note that a zone is a box defined by the coordinates of its two extreme
corners, (Xmin, Ymin, Zmin) and (Xmax, Ymax, Zmax).
3.12. Definition of the Distribution Analysis

If you have set the task type to DISTRIBUTION analysis in the mixing task menu (by clicking the
Modify task type menu item repeatedly, until the task type is defined as shown in Figure 3.41: A Distri-
bution Analysis Mixing Task (p. 70)), you can click the Definition of a DISTRIBUTION analysis menu
item to open the menu shown in Figure 3.42: The Definition of the DISTRIBUTION analysis Menu (p. 70).
Figure 3.41: A Distribution Analysis Mixing Task
Definition of a DISTRIBUTION analysis
Figure 3.42: The Definition of the DISTRIBUTION analysis Menu
The Definition of the DISTRIBUTION analysis menu allows you to modify the default settings of the
distribution analysis (as described in Using the General Menu of a Mixing Task (p. 34)), in order to cus-
tomize the preset statistical analysis of the mixing to fit your problem. Note that at any time you can
revert to the default settings by clicking Reset to default values.
To begin, you can modify the parameters for the slicing of the trajectories, including the number of
slices, the starting position (for an open domain), the starting time (for a closed domain), the increment
between slices, and the direction of slicing. For a closed domain, the slices represent moments in time.
For an open domain, the slices represent parallel spatial planes that are perpendicular to a coordinate
axis, which begin at the inflow boundary and proceed to the outflow boundary.
Next, you can select and define the indices and transfer information you want to evaluate. These values
are calculated using the position data of the material points that originate in two separate generation
zones. The first generation zone is the single box zone you created for the mixing task; the points that
originate in this zone are tracked as they become distributed throughout the flow region as the solution
progresses. The second generation zone is created internally by Polyflow, by taking the same number
of points as are in the first zone, and randomly distributing them throughout the whole flow region (if
the domain is closed) or the whole entry section (if the domain is open); these points are also tracked
as their distribution changes as the solution progresses.
You have the option of enabling the calculation of the global distribution index. This index compares
the distribution of the first generation zone with that of the second generation zone, by using the in-
tegral method of the “Deviation” function described in The Deviation Function (p. 156). The distribution
calculation only considers points that are within a certain distance of each other; the Modify max.
distance between points menu item allows you to define this distance. For a closed flow domain, you
should set this distance to be greater than the maximum distance possible, as defined in Equa-
tion 4.20 (p. 145).
For closed domains, you can enable the calculation of the axial distribution index. This index compares
the point densities of probability associated with the two generation zones along a particular axis, by
using the integral method of the “Deviation” function described in The Deviation Function (p. 156). When
the evaluation of this index is enabled, you can set the axial direction. Note that for a mixer with moving
parts, it is recommended that you set the axis to be the axis of rotation of the moving parts.
You can evaluate the transfer of the points of the second generation zone as they move from one half
of the domain to the other. This information is particularly useful for 3D batch mixers with closed flow
domains. You can divide the domain into halves relative to the coordinate axes: the positive direction
corresponds to the direction from left to right, the negative direction corresponds to the direction
from the upper half to the lower half, and the positive direction corresponds to the direction from
back to front. For each direction, two functions are calculated. Each function represents a fraction with
respect to time or axial position, depending on whether you have a closed or open domain, respectively.
For example, the fractions associated with the Evaluation of the Left-Right Transfer (X dir)? menu
item include the left-left fraction (that is, the fraction of the points that originate in the left half of the
domain and still remain there) and the left-right fraction (that is, the fraction of the points that originate
in the left half of the domain and move to the right half ). Because the fluid is incompressible and the
mixer is fully filled, only the left-left and left-right fractions are calculated, because the right-right and
right-left fractions are complementary; this statement holds true for all three transfer directions. For
more information about these functions and how to interpret the curves that result from a distribution
analysis, see Example 116 in the ANSYS Polyflow Examples Manual, which you can find on the ANSYS
Help site.
Note that if the domain is open, you cannot evaluate the transfer of points in the direction of slicing.
If any of the transfer evaluations are enabled, you must define the coordinates for the extreme corners
of the smallest box that could surround the mixer by using the Modify surrounding box for transfer
functions, XYZ min, and/or XYZ max menu items.
You can click Show advanced options to add menu items to the menu (see Figure 3.43: Advanced
Options for the Definition of the DISTRIBUTION analysis Menu (p. 72)). These additional menu items
allow you to modify some of the numerical parameters involved in the evaluation of the statistical
functions.
Show advanced options
Figure 3.43: Advanced Options for the Definition of the DISTRIBUTION analysis Menu
If you want to make major modifications to the slicing method, you can Disable automatic settings.
This action will make menu items available that allow you to modify the slicing type, the slicing coordin-
ates, and second generation zone. Note that such modifications are not recommended, unless you have
extensive expertise in the subtleties of Polyflow.
The remaining menu items allow you to modify the number of values used to represent the various
functions.
Chapter 4: The Polystat User’s Manual
After the calculation of a large set of trajectories (performed with Polyflow), you have to use Polystat
in order to treat and / or visualize those results to obtain a global overview of the mixing process in
the current flow.
To perform such statistical treatment several steps must be performed in a specific order: the main
menu of Polystat appears as follows.
Note:
The Help option summarizes this analysis process and gives information on the way to
contact us if necessary (telephone, Email and fax number).
Figure 4.1: The Polystat Window
The major steps are the following:
1. In the File menu, you can read the results generated by Polyflow (the data necessary to Polystat).
2. In the Properties menu, you can ask the program to calculate new parameters evolving along the traject-
ories. For example, you can define any concentration field, or a new mixing index. These new parameters
are always a combination of existing parameters (those calculated in Polyflow and stored in the mixing
result files). You will see later the different possibilities accessible to you.
3. In the Trajectories menu, you can select a subset of trajectories on which you will perform the statistical
treatment. For example, you can eliminate all the trajectories that terminated abnormally (for example,
on a wall). You will see later the different possibilities accessible to you.
The Polystat User’s Manual
4. In the Slices menu, you determine the way to slice the selected trajectories. For example, suppose you
are analyzing the flow through a cylinder, like the one shown in the following picture:
Figure 4.2: A Cutting Plane Through a Flow Domain
To analyze this flow, place a set of material points in the inflow section, and calculate their trajectory
until they reach the outflow section. You will then cut the trajectories with planes disposed regularly
from the entry to the exit (this is the slicing step). In each plane, you will calculate statistical func-
tions. Those functions will evolve from entry to exit, and show the way the mixing changes.
Note:
If the flow occurs in a closed domain, you want to know the time evolution of statistical
functions and the slicing will be done on the time.
5. In the Statistics menu, you define the set of statistical functions that you want to calculate on the defined
set of slices.
6. Return to the File menu, and click Run. By clicking this option, you order Polystat to calculate all of the
desired data (new properties, the subset of trajectories, the set of slices, the set of statistical functions).
This calculation can take a while to complete. Once the calculation is complete, you can analyze, visualize
and store your results.
The following explains every Polystat menu, option, and dialog box, in detail.
The File Menu
4.1. The File Menu

Figure 4.3: The File Menu Options
Open...
allows you to read an old Polystat session file.
Options
allows you to specify some parameters for reading trajectories, for slicing and for visualization.
Read Data
allows you to read the files containing the trajectories calculated by Polyflow. All the trajectories and the
kinematic parameters calculated along those trajectories are stored in Polystat.
Read mesh
allows you to read the file containing the finite element mesh (Polyflow format) used to calculate the flow
and the trajectories. This mesh is used only by the option Draw results to visualize a property in a slice
through the flow domain.
Run
calculates all of the objects defined earlier (new properties, new sets of trajectories, new sets of slices, new
statistical functions).
Draw results
allows you to visualize one selected slice/trajectory in the flow domain: you will see the spatial repartition
of any property in this slice/trajectory. The drawings can be saved as .jpeg or .png files.
Draw Stats.
allows you to visualize the calculated statistical functions and save them as .jpeg or .png files.
WRITE Trajectories
allows you to select a set of trajectories to save on files (in the csv file format; see the Polyflow User's
Manual for more details).
WRITE Slices
allows you to select a set of slices to save on files (in the csv file format; see the Polyflow User's Manual
for more details).
WRITE Stat.
allows you select statistical functions to save on files (in the crv file format; see the Polymat User's
Manual for more details).
Save
allows you to save your current Polystat session.
Quit
exits the program.
4.1.1. Open...
This option allows you to read an sav file containing a previous Polystat session. It is possible to
pursue an interrupted session on the same set of mixing files or to apply same statistical treatment
on a new set of mixing files (in this case, the mixing files must have same name, same type (ASCII or
binary) and with the same number of mixing files). See Save (p. 94) for additional information on the
list of objects saved (or not saved) in the Polystat session file.
4.1.2. Options
Figure 4.4: The Options Dialog Box
You can specify two parameters:
• The Maximum number of instants per trajectory: default value =1000. If a trajectory has more instants
than this maximum, Polystat interpolates the properties along the trajectories in order to limit the instants
to that number. It may be useful to increase this number, if you know that some trajectories are very long
and have many more instants stored in the mixing file.
• The Method of intersection between trajectory and slice: by default, Polystat computes all the intersec-
tions. This may cause some bias in the statistical functions, if you perform a slicing in a space direction
and if there are backflows. In that case, it is mandatory to use the Find first intersection method.
The File Menu
4.1.3. Read Data

Figure 4.5: The Read Mixing files Dialog Box
This window shows the list of mixing files already read.
If you want to add several mixing files to the list in a single command, click Add all.
The Read mixing files dialog box appears.
Figure 4.6: The Read mixing files Dialog Box
• Here you must specify the prefix of the mixing files, their first and last indexes and their format. If the
mixing files are not in the current directory, click the Browser button to search their location with a spe-
cific file browser. The names of the mixing files are built as follows: [prefix]_[i}.mix, where is the
current index. Once those data are entered, click OK to close the window and read the mixing files. Click
Cancel to close the window without reading any files.
Note:
Polystat first looks for mixing files with the current naming system before searching for
files with the old naming system ([prefix].[000i]).
• To add only one file, click Add and select your file using the file browser that appears. It is not possible
to read the same file twice. When a file has been selected, a new window appears asking if this file is
formatted or not; don't make any mistake, because an incorrect selection will interrupt your session.
• It is possible to remove files from the list. To remove one file from the list, select it in the list, and click Del.
To remove all the files, click Del All. In both cases, Polystat asks for a confirmation of your choice.
When the list of files seems complete, click OK to return to the main window.
4.1.4. Read mesh

Figure 4.7: The Read a mesh file Dialog Box
• Enter the name of the mesh file in the Initial mesh: text box (Polyflow mesh format). If the file is not in
the current directory, click the Browser button to browse to the file’s location.
• In the lower part of the window, you can enter specific data in order to visualize the motion of moving
parts (calculated using the mesh superposition technique in a previous Polyflow run).
1. Answer yes to the question Moving parts motion?.
The File Menu
2. Enter the prefix of the result files in the First index : number-entry field (Polyflow result files). Select
the files that contain the successive flow fields.
3. Enter the number of files in the Last index : number-entry field. The result file names will be built as
follows: [prefix].[id], where id ranges from 1 (first index) to (last index).
• Clicking OK closes the window and reads the files.
• Clicking Cancel closes the window without reading any files.
• Clicking Reset reinitializes the internal data structure containing the mesh and the successive positions
of the moving parts.
Warning:
Be careful not to make any mistakes in the format of the files, because a wrong answer
can interrupt your session definitively.
Note:
It is absolutely necessary to read a mesh before any visualization of your results.
4.1.5. Run
Clicking Run calculates all of the objects you defined elsewhere (properties, trajectories, and so on).
This process is computationally expensive, so you must confirm that you want to run the calculation.
Figure 4.8: The RUN Dialog Box
When you click Yes Polystat will successively calculate the new properties, the new sets of trajectories,
and the sets of slices and the statistical functions.
The results are available for visualization once the calculations are complete.
4.1.6. Draw results

This option allows you to visualize the results of your calculations. You can display the spatial distri-
bution of a property for a specific slice or for a given trajectory. Clicking Draw results opens the
Graphic Display window. The left side has tabs that allow you to specify what to see in the graphics
window on the right side. The graphics toolbar – located above the graphics window– contains a set
of buttons that allow you to manipulate the view (for additional information, see Graphics Toolbar
in the Polyflow User's Guide).
Figure 4.9: The Graphic Display Window of Polystat
There are two zones in the Graphic Options window. The first zone contains two buttons, Draw and
Close.
• Draw allows you to update the drawing with the current set of options you selected.
The File Menu
• Close closes the Graphic Display window; the current graphic options are saved.
The second zone contains three tabs, Slice, Prop, and Mesh.
Figure 4.10: The Graphic Display Window Tabs
Enter the three tabs in the following order: Mesh, then Props, then Slice.
1. In the Mesh tab, specify which parts of the mesh you want to display.
The following types of objects populate the Existing domains list.
The Existing domains list shows the domains available for display.
• Domains
• Domain intersections (for example, the intersection between domains 1 and 2 is written Domain1*Do-
main2)
• Intersections of domains and boundaries (for example, the intersection between domain 1 and the
boundary 3 is written Domain1*Boundary3)
• The perimeter of the previous object (the perimeter of Domain1*Boundary3 is written $(Do-
main1*Boundary3)
To visualize a domain from this list, select the domain(s) you want to display, and click Add.
The selected domain(s) will disappear from this list and appear in the Domains to draw list.
If you want to change the color of a domain, select it from the list of Domains to draw and click
button.
If you do not want to see a domain, select it from the Domains to draw list and click Del.
(Clicking Del All removes all of the domains from the Domains to draw list; Clicking Add All
displays all of the domains.)
2. Specify the property(s) you want to see on the slices or along the trajectories in the Prop tab.
a. Select a coordinate field (used to position the pixels or the segments in the domain).
b. Select a property to display.
Once you choose property, you can see the range of values for this property in the Min. value
and Max. value boxes. If it is a vector field, the range is based on the magnitude of the field.
You have the ability to change this range of values if needed.
Clip to Range is enabled by default, so areas where the value is outside of the specified range
are left empty. If you disable this option, values below Min. value will be displayed in the color
that represents the lowest value on the color scale, and values above Max. value will be displayed
in the color that represents the highest value on the color scale.
3. In the Slice tab, specify the set of slices or trajectories you want to visualize.
You can also specify a possible link between successive slices and successive positions of moving
parts.
a. Select a starting slice or trajectory from the Set of slices (or traj.) list.
b. To see successive slices or trajectories, enter a positive integer in the Step : number-entry field.
c. Click the First button in the Draw group box, and Polystat will draw the first slice or trajectory in
the graphics window.
d. Clicking the Next button increases the current slice or trajectory by the specified increment.
Alternatively, clicking the Animate button will automate the process of clicking the Next
button.
e. Enter a positive integer in the Enter pause : number-entry field to define the pause time between
two successive drawings.
The File Menu
Once the animation has begun, the Animate button changes to the Stop button.
Note:
If the mesh superposition technique was used in Polyflow to calculate a flow with
moving parts, it is possible to see the location of those moving parts at different time
steps. To do so, enable Set into motion?, then follow above steps (a) – (d). You have
the responsibility to define the correct set of slices corresponding to the motion of
the moving parts: you have to define an automatic slicing on time, with an increment
corresponding to the time step existing between two successive positions of the
moving parts. Also, you have to choose correctly the first slice, the first moving parts
position, step slice and moving parts step. (The default values are rarely correct.)
To save the visualized slices or trajectories, use the controls in the Export group box. To save a
series of individual images, select Image Series and click Save. In Windows, you also have the
option of saving an animation: select Animation, specify the Frame Rate, and click Save.
The Options group box at the bottom of the Slice tab provides the following options for the
graphics window:
• In the Color set group box, enabling the Show option displays the color scale. Enabling the Logar-
ithmic Scale option and changes the scale intervals from linear to logarithmic.
• Specify the pixel size in the Enter pixel size number-entry box.
• The Enter pause number-entry box allows you to specify the waiting time between two successive
drawings displayed when you click Animate.
• In the Logo group box, you can specify whether or not the ANSYS logo is displayed by using the
Show Logo option. You can also specify whether the color of the logo is black or white by making
a selection from the Color list.
Here is a list of shortcuts that can be used to modify the graphics window and can be easily in-
voked when the graphics window is active.
• Appearance:
– s: displays topological objects in shaded mode.
– w: displays topological objects in wireframe.
• Axes:
– a: removes axes from display.
– A: adds axes to display.
• Background:
– b: sets background color to light blue.
– B: sets background color to black.
– W: sets background color to white.
– G: sets background color to gray.
• Views:
– x: changes the view to be perpendicular to the plane from a positive coordinate; repeatedly
pressing this key will cause the view to alternate with the view from a negative coordinate.
– y: changes the view to be perpendicular to the plane from a positive coordinate; repeatedly
– z: changes the view to be perpendicular to the plane from a positive coordinate; repeatedly
• Scaling:
– X: doubles the scaling of the view in the dimension.
– Y: doubles the scaling of the view in the dimension.
– Z: doubles the scaling of the view in the dimension.
– Ctrl-x: reduces the scaling of the view in the dimension by half.
– Ctrl-y: reduces the scaling of the view in the dimension by half.
– Ctrl-z: reduces the scaling of the view in the dimension by half.
• Resize:
– r: removes the scaling factors and fits the display into the window.
– R: fits the display into the window while retaining the scaling factors.
The File Menu
4.1.7. Draw Stats.

Figure 4.11: The Draw Statistical Results Dialog Box
Here you can select the functions you want to visualize with Polycurve. Polycurve is a charts editor
linked to Polystat: its main purpose is to allow you to create and visualize charts (for additional in-
formation, see Defining and Plotting Curves in the Polymat User's Guide). A chart can contain several
curves. You can add and remove a curve, change a title, a color, the range of the axes, and so on.
1. Select one function from the upper list and the corresponding set of curves are displayed in the lower
list.
2. Select the curves you want to display and click OK.
A chart is automatically created to display your selected curves in Polycurve.
To visualize curves of another function in another chart:
1. Click New Chart in the Polycurve window.
2. Click Statistics, the Draw Statistical Results dialog box opens and you can select another statistical
function and another set of curves, as explained above.
Click Close to close the Polycurve window and to go back to the main Polystat window.
Figure 4.12: The Polycurve Window
The File Menu
4.1.8. WRITE Trajectories

Figure 4.13: The Save Set of Trajectories Dialog Box
Use this dialog box to select the set of trajectories that you want to save in formatted files.
1. Select one set in the upper list.
2. Click the button.
Figure 4.14: Setting Up the Writing of Trajectory Data
In the above dialog box you can choose the coordinates and time used when writing trajectory data
(useful in some cases where absolute and relative coordinates exist, or if you defined your own co-
ordinates or time).
You can also select the file format.
• The default format is CSV. The CSV file format is a common format for tabulated data that can be read
into spreadsheet programs such as Excel (for additional information, see Polyflow User's Guide in the
Polyflow User's Guide). Polystat will generate one file per trajectory; the names are built as follows: [pre-
fix]_[trajectory index].csv.
• The second format is the FV format. You save the selected set of trajectories in a single FVP file, that can
be loaded in FieldView. This option is useful if your simulation is steady-state and if you want a better
graphic treatment of your trajectories (for transient simulations, see WRITE Slices (p. 91)). Polystat will
generate one file with the name: [prefix].fvp.
• The last format is the CFD-Post format, which saves the selected set of trajectories in a single TRK file that
can be loaded in CFD-Post. This option is useful if you want a better graphic treatment of your trajectories.
Polystat will generate one file with the name: [prefix].trk.
Once you have selected the format of the file(s) that you want to output, enter the prefix of the files
in the Enter the prefix of the file(s) : text box, or select it by clicking the Browser button.
Once the files are written, they appear in the list named Selected set of trajectories in the Save Set
of Trajectories dialog box.
When you have finished, click OK.
The File Menu
4.1.9. WRITE Slices

Figure 4.15: The Save Set of Slices Dialog Box
In this dialog box, you can select the set of slices you want to save in formatted files.
1. Select one set of slices from the Select a set of slices : list.
This opens the following dialog box.
Figure 4.16: Setting Up the Writing of Slice Data
In the above dialog box you can choose the coordinates and time used when writing trajectory data
(useful in some cases where absolute and relative coordinates exist, or if you defined your own co-
ordinates or time). The selected property for the COORDINATES will be named COORDINATES and
appears as the first property in the file. This can be useful if you need to read a CSV file as a generation
zone (see Parameters for the Generation of the Material Points (p. 49) for additional information).
You can also select the file format.
• The default format is CSV, which saves the selected set of slices in a set of CSV files. The CSV file format
is a common format for tabulated data that can be read into spreadsheet programs such as Excel (for
additional information, see Polyflow User's Guide in the Polyflow User's Guide). Polystat will generate one
file per slice; the names are built as follows: [prefix]_[slice index].csv.
• The other format is FV, which saves the selected set of slices in a single FVP file that can be loaded in
FieldView. This option is useful if your simulation is transient and if you want a better graphic treatment
of your trajectories (for steady-state simulations, see WRITE Trajectories (p. 89)). Polystat will generate
one file with the name: [prefix].fvp.
Once the files are written, they appear in the list named Selected set of slices in the Save Set of
Slices dialog box.
When you have finished, click OK.
The File Menu
4.1.10. WRITE Stat.

Figure 4.17: The Save Statistical Results Dialog Box
In the Save Statistical Results dialog box, you can select the functions you want to save in formatted
files.
1. Select one function the Select a function : list.
A file browser appears.
3. Enter the file name or the prefix of the file.
Once the files are generated they appear in the Selected functions : list.
Click OK to return to the main window.
Depending on the type of function to write, Polystat can generate one or several files: for example,
if you select a probability function, there exists a function of the type for each slice, and
Polystat will generate a file for each one.
The name of these files are built like this: you specify the prefix of the files, and Polystat add an index
(if necessary) and a suffix depending on the kind of result you want to save:
see a property along a trajectory : [prefix]_see[trajectory index].crv
correlation between two fields : [prefix]_corr.crv

segregation scale : [prefix]_seg.crv
deviation from an ideal distribution : [prefix]_dev.crv
sum of a property : [prefix]_sum.crv
mean of a property : [prefix]_mean.crv
standard deviation of a property : [prefix]_stdv.crv
deviation of points concentration : [prefix]_pcd.crv

operator on functions : [prefix]_opr.crv
auto-correlation on concentration field : [prefix]_crc[slice index].crv

probability : [prefix]_prb[slice index].crv
density of probability : [prefix]_dns[slice index].crv
distance distribution : [prefix]_dsp[slice index].crv
histograms : [prefix]_hst[slice index].crv
percentiles : [prefix]_pct[percentile index].crv
distribution in zones :
global deviation [prefix]_zonGI.crv
deviation for zone i [prefix]_zon[zone index].crv
+ zones partitioning,
points concentration : [prefix]_[slice index].csv
(as a set of csv files!)
The .crv files that are generated are ASCII files that can be plotted in ANSYS Polymat and ANSYS
Polycurve (see Defining and Plotting Curves in the separate Polymat User's Guide for details), or
viewed in other spreadsheet applications. The files have the following format:
• The optional header is the first 5 lines:

line 1 : <empty line>
line 2 : title or name of the current curve
• The next lines are the set of points defining the curve. On each line there is one pair of (x,y) values: the
format is 14 characters per value, with 7 digits after the dot. A white space separates the two numbers.
next lines :
±0.0000000E±00 ±0.0000000E±00
±0.0000000E±00 ±0.0000000E±00
...
4.1.11. Save
Save allows you to save a file with a sav suffix in your current Polystat session. In the session file
are saved all information regarding properties (those existing in the mixing files, but also the created
ones), set of trajectories, set of slices and of course statistical functions.
Note:
This session file only contains mesh and result information (no graphical options).
The Properties Menu
4.2. The Properties Menu
4.2.1. Definition
Figure 4.18: The Properties Menu Options
Using the Properties menu, you can define new properties by combination of the existing ones. A
list of values for a small set of parameters (also named properties) is stored along each trajectory, as
explained in the next drawing:
Figure 4.19: Positions along a Trajectory
For each particle, at each stored position (time and so on), you can store several of
the following properties in the mixing files (from Polyflow):
• the current time, the position (coordinates), the space integration (length of the trajectory from the initial
position);
• labels, if several generation zones are defined;
• a concentration field, if you defined one when specifying a kinematic mixing template.
• a porous domain field: values are equal to zero when points are located in the fluid region; equal to one
when points are located in a porous domain (using a Darcy model). This field is always stored in the mixing
file and is equal to zero everywhere when there is no porous domain defined.
• the natural logarithm of stretching (note that here the words stretching and elongation are equivalent)
(ln or ln ), the direction of stretching ( or ), the rate of stretching ( or ), the rate of
dissipation ;
• the pressure, the velocity, the temperature;
• the determinant of the tensor , the divergence of the velocity. Both parameters give information about
the accuracy of the calculation (as the flow is supposed to be incompressible, the determinant of must
remain equal to one and the divergence of the velocity equal to zero).
• the mixing index, the first eigenvalue of tensor , the vorticity.
4.2.2. See Properties

After the reading of the mixing files, if you select SEE / MOD / DEL properties in the Properties
menu of the main window, you will see the list of the existing properties:
Figure 4.20: The See / Modify / Delete existing properties Dialog Box
It is possible to modify or to remove from the list one or several properties (this is only true for the
properties created in Polystat).
If you want to modify some data of a property, select it from the list, and click Modify. The window
that served for the creation of that property will appear and there you can modify some data. If you
want to store the modified data, click OK. Otherwise, click Cancel.
To remove a property from the list, select it from the list, and click Delete. To remove all the created
properties, click Delete All. In both cases, Polystat asks you to confirm your choice.
4.2.3. Create Properties

Based on existing properties, it is easy to define new properties.
1. Select CREATE a new property from the Properties menu of the main window.
The following window appears.
The Properties Menu
Figure 4.21: The CREATE a new PROPERTY Dialog Box
2. Click the button at the right of the combo box presenting <Select Type>.
A drop-down list appears showing the properties that can be created.
3. Select the appropriate property from the list.
Figure 4.22: Types of Properties that Can Be Created
4.2.3.1. || A ||
This method allows you to calculate the amplitude of a property (the velocity, for example), or more
precisely, the evolution of the amplitude of a property along all the trajectories. After selecting ||
A ||, the following window appears:
The Properties Menu
Figure 4.23: The || A || Settings in the CREATE a new PROPERTY Dialog Box
The following data is necessary in this case:
1. Select a vectorial or scalar property (the data).
2. Name the new property (the result). There cannot be two properties with the same name.
3. Click OK to agree with the data. The new property will be stored in the list of existing properties. If
you click Cancel, the property will not be stored.
For the other methods, the process is the same. What changes is the number and the type of the
data needed to calculate the new property.
4.2.3.2. A^x
A^x: the new property is the property A exponent x. Select property A (the data), to enter the ex-
ponent x, and to give a name to the new property (the result).
4.2.3.3. exp ( A )
exp ( A ): the new property is the exponential of the property A. Select property A (the data), and
give a name to the new property (the result).
4.2.3.4. log ( A )
log ( A ): the new property is the natural logarithm of the property A. Select property A (the data),
and give a name to the new property (the result).
4.2.3.5. A + B, A - B
A + B, A - B: the new property is the addition or the subtraction of the properties A and B. Select
the two properties A and B (the data), and give a name to the new property (the result). The two
properties must have the same type (scalar or vectorial).
4.2.3.6. A / B
A / B: the new property is the division of the property A by the property B. Select the two properties
A and B (the data), and give a name to the new property (the result). The property A can be a
scalar or a vector, but the property B must be a scalar.
4.2.3.7. A * B
A * B: the new property is the multiplication of two properties A and B. Select the kind of multiplic-
ation and the two properties A and B (the data), and give a name to the new property (the result).
The Properties Menu
Figure 4.24: The A * B Settings in the CREATE a new PROPERTY Dialog Box
There are 3 kinds of product:
• Scalar product (dot product): both properties A and B are vectors. The result property is a scalar:
(4.1)
• Vectorial product (cross product): both properties A and B are vectors. The result property is a vector:
(4.2)
• Other product: you can multiply one scalar by another scalar (the result is a scalar), or one vector by a
scalar (the result is a vector).
4.2.3.8. Rotate
Rotate: this method allows you to rotate a vectorial property: you have to select the property A
to rotate and the time (the data), some data specifying the rotation (rotation axis, angular velocity)
and to give a name to the new property (the result). This property allows you to change, if necessary,
the frame of reference (one in rotation with respect to the other).
The data needed are visualized in the following picture:
Figure 4.25: Rotation Data
The creation dialog box appears as follows:
Figure 4.26: The Rotate Settings in the CREATE a new PROPERTY Dialog Box
If the property A is a coordinate field, click the corresponding button. The resulting property C will
be:
(4.3)
The Properties Menu
where is one point of the rotation axis, and , the matrix of rotation at time .
However, if the property A is not a coordinate field, click the A is another field button. The resulting
property C is:
(4.4)
4.2.3.9. Translate
Translate: this method allows you to translate a vectorial property. Select the property A to translate
and the time (the data), some data specifying the translation (direction and amplitude of the
translation velocity) and to give a name to the new property (the result). This property allows you
to change, if necessary, the frame of reference (one in translation with respect to the other).
The creation window appears like this:
Figure 4.27: The Translate Settings in the CREATE a new PROPERTY Dialog Box
If the property A is a coordinate field, click the corresponding button. The resulting property C will
be:
(4.5)
where is the translation velocity and , the time.
However, if the property A is not a coordinate field, enable the A is another field option. The res-
ulting property C is:
(4.6)
4.2.3.10. Derivate
Derivate: this method allows you to derive a property along trajectories. Select the property A to
derive, the time (the data), and give a name to the new property (the result).
Figure 4.28: The Derivate Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.11. Integrate
Integrate: this method allows you to integrate a property along the trajectories (in time). Select
the property A to integrate and the time (the data) and give a name to the new property (the result).
The Properties Menu
Figure 4.29: The Integrate Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.12. k (constant)
k (constant): this method allows you to create a property that is constant along all the trajectories.
You have to select the type of the constant (scalar, vector with 2 components, vector with 3 com-
ponents), specify its value, and give a name to the new property (the result).
Figure 4.30: The k (constant) Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.13. A ( Slice )
A ( Slice ): this method allows you to transport the value of a property along the trajectories. You
have to select the property to transport, define a slice (see The Slices (p. 165)), and give a name to
the new property (the result).
The Properties Menu
Figure 4.31: The A ( Slice ) Settings in the CREATE a new PROPERTY Dialog Box
The next picture explains the concept defined above:
Figure 4.32: Transporting a Property from a Slice
In general, you want to visualize the current value of a property (for example, the temperature)
attached to a material point at its current position. However, for the study of mixing, it may be
useful to view the results differently. For example, if you study the stretching of a set of particles
in a Kenics mixer, it may be useful to see at the initial position of the particles, their final stretching
position. Using this method, you can detect zones in the inlet where the upcoming stretching will
be good or bad.
4.2.3.14. Concentration
Concentration: you can define any concentration field in an initial configuration. As the concentration
field is constant for a material point (no diffusion, no chemical reactions), the value of the concen-
tration is transported along the trajectories without changing. You have to define the initial config-
uration and give a name to the new property (the result).
The initial configuration is composed of two things; a slice and a list of zones.
1. Specify a slice (see The Slices (p. 165)).
2. Specify a list of zones (see The Zones (p. 166)). In this slice, the instants included in a zone have a
concentration value of 1. The instants external to all the zones have a concentration of 0.
Figure 4.33: Defining Concentration Using a Slice
When defining a concentration property, the following window appears:
The Properties Menu
Figure 4.34: The Concentration Settings in the CREATE a new PROPERTY Dialog Box
Every zone must have a different name.
• To define a new zone, click Add.
• To modify an existing zone, select it from the list and click Mod.
• To delete an existing zone, select it from the list and click Del.
4.2.3.15. Min / Max

This method allows you to calculate the minimum (or the maximum) of a property A along traject-
ories. The following parameters must be defined:
1. The property A of interest.
2. Specify whether you want the minimum or the maximum of the property.
3. Name the resulting property.
4. Specify if you want the absolute extreme value of property P along the whole trajectory (option 0 -
> t_end) or a time evolving extreme value of property P (option 0 -> t):
• option 0 -> t_end: for a given trajectory T, at any time
• option 0 -> t: for a given trajectory T, at time
Figure 4.35: The Min / Max Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.16. Extract
This method allows you to create a new scalar property by extracting a component of a vectorial
property A. The following parameters must be defined:
1. The vectorial property A.
2. The component , or to extract .
3. The name of the resulting property.
The Properties Menu
Figure 4.36: The Extract Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.17. Step
This method allows you to create a new property C, defined as follows:
• In H mode:
= Amplitude if property A at time threshold value
if property A at time threshold value
• In 1-H mode:
if property A at time threshold value
= Amplitude if property A at time threshold value
The following parameters must be defined:
• The property A
• The H or 1-H mode
• An amplitude
• A threshold value
• The name of the resulting property
Figure 4.37: The Step Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.18. Instant. Eff.

This method allows you to create the instantaneous efficiency of mixing based on the rate of dis-
sipation and the rate of stretching. As seen in The Mixing Theory (p. 11), this efficiency is defined
as:
• For 2D flows:
(4.7)
• For 3D flows:
The Properties Menu
(4.8)
This property is only accessible if the rate of stretching and the rate of dissipation have been calcu-
lated along the trajectories. The following dialog box appears when you select this method.
Figure 4.38: The Instant. Eff. Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.19. Time Aver. Eff.

This method allows you to create the time averaged efficiency of mixing based on the time, the
rate of dissipation and the rate of stretching. As seen in The Mixing Theory (p. 11), this efficiency
is defined as:
• For 2D flows:
(4.9)
• For 3D flows:
(4.10)
This property is only accessible if the time, the rate of stretching, and the rate of dissipation have
been calculated along the trajectories. In the creation window, the default values for the properties
are correct. The following dialog box appears:
Figure 4.39: The Time Aver. Eff. Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.20. Shear Rate

This method allows you to create the instantaneous (local) shear rate, based on the rate of dissipation
defined in The Mixing Theory (p. 11). For 2D and 3D flows, the shear rate is defined as
(4.11)
This property is only accessible if the rate of dissipation has been calculated along the trajectories.
In the creation dialog box, the default values for the properties are correct. The dialog box appears
as follows:
The Properties Menu
Figure 4.40: The Shear Rate Settings in the CREATE a new PROPERTY Dialog Box
4.2.3.21. Melting Index

This method allows you to create the melting index, based on the viscosity and the temperature
. The concept of a melting index has been introduced by Beerkens et al. [3] (p. 177) to characterize
the quality of the glass melting in a furnace. The melting index evolves along a given trajectory
(starting at point ) as a function of temperature, viscosity, and time. For 2D and 3D flows, it is
defined as
(4.12)
This property is only accessible if temperature and viscosity have been calculated along the traject-
ories. In the creation dialog box, the default values for the properties are correct. The dialog box
appears as follows:
Figure 4.41: The Melting Index Settings in the CREATE a new PROPERTY Dialog Box
4.2.4. Disagglomeration Properties

Select NEW Disagglomeration properties from the Properties menu. It is possible to create properties
directly related to the model of disagglomeration of solid particles in the specialized dialog box that
opens (for additional information, refer to The Mixing Theory (p. 11)).
The Properties Menu
Figure 4.42: Types of New Disagglomeration Properties
It is not possible to evaluate size, fractions, or number of agglomerates without defining a property
of type disagglomeration.
4.2.4.1. Disagglomeration
Figure 4.43: The Disagglomeration Settings in the CREATE a new PROPERTY Dialog Box
With the Disagglomeration property, you will know the time evolution of the mass fraction distri-
bution for a set of agglomerates of various sizes.
As explained in The Mixing Theory (p. 11), the mechanisms of erosion and rupture depend on time,
shear rate, and viscosity of the matrix.
First you must define these properties.
Note:
The Polyflow solver evaluates the rate of dissipation instead of the shear rate, which
is equal to .
Next, You must specify the number of classes of agglomerates you want to evaluate. You must
specify the size of the largest agglomerates and the maximum size of the aggregates (particles that
cannot be broken in smallest pieces anymore).
Finally, you must specify which CLIPS file contains the definitions for the transfer functions for
erosion and rupture mechanisms, the kinetics of erosion and rupture (function of shear rate, viscosity,
size of agglomerates), and the initial distribution function of agglomerates size. As this file is inter-
The Properties Menu
preted during the calculation, it is very easy to modify these functions as the understanding of the
disagglomeration improves, or to test new ideas.
4.2.4.2. Typical size of agglomerates

Figure 4.44: The Typical size of agglomerates Settings in the CREATE a new PROPERTY Dialog
Box
With this function it is possible to evaluate the minimum, mean, or maximum size of agglomerates,
without taking into account the aggregates. To achieve this, a new distribution function is evaluated,
which ranges between the maximum size of aggregates ( ) and the maximum size of agglomerates
( ):
(4.13)
The mean size of agglomerates is the following:
(4.14)
where
(4.15)
The minimum and maximum sizes are evaluated as shown in the picture below. They corresponds
to the 5th percentile and 95th percentile of the function:
Figure 4.45: Minimum and Maximum Sizes
4.2.4.3. Fraction of agglomerates of given size

Figure 4.46: The Fraction of agglomerates of given size Settings in the CREATE a new PROPERTY
Dialog Box
With this property, you can evaluate the mass fraction of agglomerates, at a given time, having a
size between a given range [ ; ]:
The Properties Menu
(4.16)
4.2.4.4. Number of agglomerates of given size

Figure 4.47: The Number of agglomerates of given size Settings in the CREATE a new PROPERTY
Dialog Box
With this property, you can evaluate the number of agglomerates, at a given time, having a size
between a given range [ ; ]:
(4.17)
where
(4.18)
and is defined in Equation 4.16 (p. 121)

is the volume of a sample able to contain a large number of agglomerates of
various sizes
is the mass fraction of agglomerates included in the sample
is a shape factor ( assuming that agglomerates are spheres, , if they

are cubes.)
4.2.4.5. Mass of agglomerates of given size

Figure 4.48: The Mass of agglomerates of given size Settings in the CREATE a new PROPERTY
Dialog Box
With this property, you can evaluate the mass of agglomerates, at a given time, having a size
between a given range [ ; ]:
(4.19)
where is defined in Equation 4.16 (p. 121)

is the volume of a sample able to contain a large number of agglomerates of
various sizes
is the mass fraction of agglomerates included in the sample
is the density (mass per unit volume) of pure agglomerates
The Trajectories Menu
4.3. The Trajectories Menu

Figure 4.49: The Trajectories Menu Options
The SEE / MOD / DEL trajectories option allows you to manage the list of existing sets of trajectories.
With the CREATE a new set of trajectories option, you will define a new set as a subset of an existing
one, based on a selection criterion. With the COMBINE sets of trajectories option, you will define a
new set by selecting trajectories with Boolean operations between two existing sets. Using the SELECT
one single trajectory option, you can select a single trajectory from an existing set based on a selection
criterion. SEARCH stagnation points allows you to detect stagnant points close to fixed or rotating
walls.
4.3.1. See the Set of Trajectories

After the creation of new properties, you have the ability to select the set of trajectories on which
you will do some statistical treatment. Before any creation of such sets, if you select the option SEE
/ MOD / DEL trajectories option from the Trajectories menu of the main window, you will see the
list of the existing sets of trajectories:
Figure 4.50: The See / Modify / Delete existing sets of trajectories Dialog Box
In this list, you can see three different types of sets of trajectories:
• The All_trajectories set contains all the trajectories read in the mixing files.
• The Interrupted_trajectories set contains all the trajectories the calculations of which have been inter-
rupted (for different possible numerical problems) and read in the mixing files.
• The set_from_filename sets contain the trajectories stored in the mixing file filename.
You cannot modify or delete those sets. If you remove a mixing file (in the Read Data dialog box),
the list of existing sets is adjusted automatically.
If new sets of trajectories have been created, it is possible to modify or to remove them from the list.
If you want to modify some data of a set, select it from the list and click Modify. The creation dialog
box of that set will appear. If you want to store the modified data, click OK; Otherwise, click Cancel.
To remove one set from the list, select it from the list and click Delete. To remove all the created
sets, click Delete All.
Based on existing sets of trajectories, it is easy to define new sets. There are four possibilities:
• the creation of a new set based on a condition to respect
• the combination of two sets to create a new one
• the selection of a single trajectory, based on a criterion of proximity
• the selection of trajectories of stagnant points
4.3.2. CREATE a new set of trajectories

By selecting this option from the main window, the following dialog box appears:
Figure 4.51: The Add a new set of trajectories Dialog Box
Using this method, you will select a subset of trajectories that respect a condition. There are 3 kinds
of conditions:
• A trajectory is selected if, the property always has values in the specified interval. For example, you can
select the trajectories that always have the determinant of F between .99 and 1.01 (rejecting inaccurate
trajectories).
• A trajectory is selected if, the property never has values in the specified interval.
• A trajectory is selected if, the property has a value at least once within the specified interval. For example,
you can select the trajectories that cross a specified box in the flow domain.
If the selected property is a scalar, the interval is defined in rectangles 1 and 4 (the others are set to
zero).
4.3.3. COMBINE sets of trajectories

Wen you select this option, the following dialog box appears:
Figure 4.52: The set C = set A <operator> set B Dialog Box
Using this option, you can combine two sets of trajectories A and B in order to create a new set C.
For example, in the above dialog box, the sets All_trajectories and Interrupted_trajectories and
the Intersection operator are selected. That means that the new set C will contain the trajectories
that belong to the two sets A and B.
If you select the Union operator, the new set C will contain all the trajectories of the two sets A and
B (there is no duplication of the trajectories that belong to the two sets).
If you select the Minus operator, the new set C will contain the trajectories of the set A that do not
belong to the set B.
If you select the Difference operator, the new set C will contain the trajectories of the set A that do
not belong to the set B and the trajectories of the set B that do not belong to the set A. This operation
can be written as: C = ( A Minus B ) Union ( B Minus A ).
4.3.4. The SELECT one single trajectory Option

By selecting this option, the following dialog box appears:
Figure 4.53: The Select a single trajectory Dialog Box
With this method, you will select a single trajectory from a set that is the closest to a specified position.
You have to specify the following data:
• a set of trajectories in which you will search one trajectory
• a property
• a position
• an option for the selection
If you choose Initially, the trajectory with its initial position closest to the specified position is se-
lected. If you choose Independently of the moment., you are not specifying any peculiar moment.
The selected property is not necessarily a coordinate property. For example, you can select the tra-
jectory that has an initial velocity the closest to the value (1,2,3).
If the selected property is a scalar, enter the position in rectangle 1 (rectangles 2 and 3 are set to
zero).
4.3.5. SEARCH stagnation points

By selecting this option, the following dialog box appears:
Figure 4.54: The Search stagnation points Dialog Box
Using this method, you can select points having a specific behavior: the points remaining close to
their initial position, or those coming back—after a specific period—close to their initial position.
You must specify the following data:
• a set of trajectories in which you will search stagnation points
• a coordinate field
• a time field
• a threshold distance:
A point is stagnant if the distance between its initial position and some of its successive positions
are less than this threshold distance.
• a period :
– To detect stagnation points close to a fixed wall, only the positions , for time , corresponding to a
multiple of will be tested, from until (stagnation time).
For example, will be tested at times
– If you want to detect stagnation points close to a rotating wall (of a moving part, for example), you
have to specify the period of rotation of the moving part: all positions will be tested from
until .
• a stagnation time, :
The Slices Menu
must be in the range [0, max. residence time].
• a squeezing factor :
It is a way to weight specifically some components of the coordinates when computing distance
between and : distance
By default, the squeezing factor is set to (1, 1, 1).
4.4. The Slices Menu

Figure 4.55: The Slices Menu Options
After you create new sets of trajectories, you must define a list of slices that cut the trajectories. This
list is necessary to analyze the evolution of the mixing from one slice to the next, from the beginning
of the process to its end. By default, no set of slices exists. Once you have created these sets of slices,
you can select one set on which you will perform some statistical treatment.
Based on an existing set of trajectories, it is easy to define a set of slices. Three possibilities exist:
• Automatic Slicing
• Manual Slicing, one by one
• Sub-Slicing, which is the generation of a new list of slices based on each slice of an other list
Remember that a trajectory is a set of instants ordered in time. Each slice will contain a set of instants
that are the intersections of the slice and the trajectories. If the intersection of a trajectory and a slice
is not a stored instant, a new instant is created by interpolation with the previous and the next instants
that surround the intersection, as explained below:
Figure 4.56: Interpolating a New Instant
4.4.1. See the Set of Slices

If you select SEE / MOD / DEL sets of slices in the Slices menu of the main window, you will see the
list of the existing sets of slices:
Figure 4.57: The See / Modify / Delete existing sets of slices Dialog Box
If you want to modify some data of a set, select it from the list, and click Modify. The window that
served for the creation of that set will appear; then modify some data. Click OK to store the modified
data.
To remove one set from the list, select it and click Delete. To remove all the created sets, click Delete
All.
The Slices Menu
4.4.2. Automatic Slicing

Figure 4.58: The Create ‘Automatically’ a list of slices Dialog Box
Automatic Slicing automatically generates a list of ordered slices.
1. Select a set of trajectories where the slicing will be done.
2. Specify the first slice of the list, select a property, a position and a direction for the plane.
3. Enter the number of slices you want and the distance (increment) between two successive planes.
4. Enter the name of the new set.
Automatic slicing is based on a single property and all the slices are parallel to each other. This is not
the case for the manual method.
4.4.3. Manual Slicing

Figure 4.59: The Create ‘manually’ a list of slices Dialog Box
Using this method, you manually generate a list of ordered slices, one at a time. Note that Polystat
respects the order in which the slices were defined when it calculates the statistics.
1. Select a set of trajectories where you want to perform the slicing.
2. Specify each slice of the set, one by one.
Each slice must have a different name, and you can modify or delete existing slices. You can also
modify the order of the slices.
3. Name the set of slices.
This slicing is not based on a single property; each slice can be defined on a different property. The
slices are not necessarily parallel to each other. This method is more general, but will also consume
more of your time.
The Slices Menu
4.4.4. Sub - Slicing

Figure 4.60: The Generate a subslicing on a set of slices Dialog Box
Suppose you have a 3D unsteady flow and you want to visualize the spatial distribution of the
stretching in a plane cutting across your flow domain. You define a slice with the coordinates as the
selected property, while the instants along this slice can occur at various times. You want to distribute
these instants among a list of time intervals and look at the spatial distribution of the stretching for
one time interval. This is called subslicing: you defined a list of slices (temporal) on another list of
slices (spatial).
To define a subslicing, you first select an existing set of slices. Then you define your slicing data (as
for an automatic slicing). What differs is the number of subslices (for each slice) and if the subslicing
is periodic or not.
Define as the number of subslices by slice. After subslicing, the subslice indices will be:
• for slice 1, the subslices are numbered from 1 to
• for slice 2, the subslices are numbered from to
• for slice 3, the subslices are numbered from to
• and so on
If you select Periodic, the instants are gathered in the following way: for a given slice, the instants
included in the series of intervals named are gathered in subslice .
Figure 4.61: Periodic Subslicing
If you do not select the Periodic option, the instants are gathered in the following way: for a given
slice, the instants included in the interval named are gathered in the subslice .
Figure 4.62: Nonperiodic Subslicing
4.5. The Statistics Menu

After the creation of properties, set of trajectories, and sets of slices, you must define the statistical
functions you want to calculate on these objects. There are no functions defined by default. General
functions can be created by selecting CREATE a new function, while functions specific to the disag-
glomeration model can be found in NEW disagglomeration functions.
Figure 4.63: The Statistics Menu Options
4.5.1. See the Statistical Functions

If you select SEE / MOD / DEL functions in the Statistics menu of the main window, you will see
the list of the existing functions:
The Statistics Menu
Figure 4.64: The See / Modify / Delete existing functions Dialog Box
If you want to modify some data of a function, select it from the list and click Modify. The dialog box
that you used to create that function will appear. Make the changes you want, and click OK to store
the modified data.
To remove one function from the list, select it from the list click Delete. To remove all of the created
functions, click Delete All.
4.5.2. Create Statistical Functions

If you select CREATE a new function in the Statistics menu of the main window, the following dialog
box appears:
Figure 4.65: The Create a new function Dialog Box
Then click the button at the right of the combo box presenting <Select Type>. A drop-down list
appears showing the functions that can be created. Finally, select the appropriate statistical function
from this list.
The Statistics Menu
Figure 4.66: Types of Functions that Can Be Created
Click Cancel to return to the main menu.
There three kinds of functions available:
• functions based directly on the instants of slices
• functions based on other functions
• functions based on a set of trajectories
To create a new function, click the corresponding item in the list.
4.5.2.1. The Property along a trajectory Function

Figure 4.67: The Property along a trajectory Settings for the Create a new function Dialog
Box
Using this function, you can see the time evolution of a given property calculated along any trajectory
you want.
You must specify which set of trajectories will be used, which property to follow, and the time. You
also need to name the new function.
If you visualize this function, the X axis is time and the Y axis is the property of interest. In this
graph, you have a new function for each trajectory:
Figure 4.68: The Time Evolution of a Property for Two Trajectories
The Statistics Menu
4.5.2.2. The Sum Function

Figure 4.69: The Sum Settings for the Create a new function Dialog Box
To calculate the sum function of a property, you need to specify which set of slices will be used
and to select a property. You also have to give a name to the new function.
If you want to weight the sum function of the local velocity, you must select the velocity field.
Additional information on weighting is available in Weighting (p. 167).
You obtain the evolution of the sum of a property along the slices. If you visualize this function,
the X axis is the index of the slice, and the Y axis is the property .
Figure 4.70: The Evolution of the Sum Function of a Property along the Slices
4.5.2.3. The Mean + Standard Deviation Function

Figure 4.71: The Mean + Standard Deviation Settings for the Create a new function Dialog
Box
To calculate the mean and the standard deviation of a property, you need to specify which set of
slices will be used and to select a property. You also have to give a name to the new function.
If you want to weight the mean and the standard deviation in function of the local velocity, you
must select the velocity field. Additional information on weighting is available in Weighting (p. 167).
You obtain two curves: one for the evolution of the mean of a property along the slices, and another
for the evolution of the standard deviation. If you visualize these functions, the X axis is index of
the slice and the Y axis is the property .
Figure 4.72: The Mean and Deviation of a Property along the Slices
The Statistics Menu
4.5.2.4. The Correlation Function

Figure 4.73: The Correlation Function Settings for the Create a new function Dialog Box
To analyze a possible correlation between two properties in two slices:
1. Select a set of trajectories.
2. Define two slices by clicking successively on the a first slice and a second slice buttons.
3. Specify the properties associated with each slice.
4. Name the new function.
If you want to visualize this function, on the X axis is the property associated with the first slice,
and on the Y axis is the property associated with the second slice.
Figure 4.74: Exploring the Correlation Between Properties in Two Slices
4.5.2.5. The Probability Function

Figure 4.75: The Probability Function Settings for the Create a new function Dialog Box
To calculate the evolution of the probability function (also named distribution function) of a property,
you need to specify a property, and which set of slices will be used. You also need to enter the
number of values to represent this function and give it a name.
If you want to weight the function depending on the local velocity, select a velocity field (see
Weighting (p. 167) for more information).
The Statistics Menu
You obtain a probability function for each slice of the set. If you visualize this function for a given
slice, the property is on the X axis and the probability on the Y axis.
Figure 4.76: The Probability Function of a Property for a Slice
4.5.2.6. The Auto-Correlation Function of the Concentration Function

Figure 4.77: The Auto-Correlation Function of the Concentration Settings for the Create a
new function Dialog Box
To calculate the evolution of the auto-correlation function of the concentration, you need to specify
which set of slices will be used. You must select a coordinate property that will serve to calculate
the distance between pairs of points in a slice and you must also select a concentration field.
You also need to specify the number of values necessary to represent the auto-correlation function.
The new function also requires a name.
You obtain an auto-correlation function for each slice of the set. If you visualize this function for a
given slice, you find the distance on the X axis and the auto-correlation function of the concentration
on the Y axis.
Figure 4.78: The Auto-Correlation Function for a Slice
4.5.2.7. The Distance distribution Function

Figure 4.79: The Distance distribution Settings for the Create a new function Dialog Box
To calculate the evolution of the distribution function of distances between material points:
1. Specify which set of slices will be used.
The Statistics Menu
2. Select a coordinate property that will serve to calculate the distance between pairs of points in a slice.
3. Enter the number of values to represent the distribution function.
You have the possibility to choose between two methods:
• the first determines the distance distribution between all pairs of points. The maximum distance measured
will be about the size of the flow domain, if the flow domain is closed.
• the other determines the distance distribution only for pairs of points that are close neighbors. The
maximum distance measured, for a closed flow domain, will be:
(4.20)
where is the volume of the flow domain and , the number of material points. The search of
neighboring points will be done only for distance between 0 and a value you specified. It is re-
commended that you choose a value greater than the maximum distance evaluated above.
In all cases, it is recommended that you have identical, if you want to compare different
flow domains or set-ups on a same flow domain.
You obtain a distribution function for each slice of the set.
If you visualize this function for a given slice, you find the distance on the X axis and the distribution
function on the Y axis.
Figure 4.80: A Distribution Function of Distances Between Material Points
4.5.2.8. The Distribution in zones Function

Figure 4.81: The Distribution in zones Settings for the Create a new function Dialog Box
To calculate the evolution of the distribution of material points initially concentrated in a box:
1. Specify two set of slices: the first one contains points coming from the real distribution, the second
one contains points coming from an optimal distribution.
2. Select a coordinate property used to evaluate in which zones the points are included.
3. Specify the coordinates of the center of the zones in an ASCII file. The format of this file is as follows;
on each line, there are the three coordinates (X,Y,Z) of one center. If the flow domain is 2D, the Z is
set to zero.
+0.0000000e+00 +0.0000000e+00 +0.0000000e+00
+0.0000000e+00 +0.0000000e+00 +0.0000000e+00
+0.0000000e+00 +0.0000000e+00 +0.0000000e+00
With this function, you obtain for each slice:
• for each zone, the evolution of the deviation of the real distribution (compared to optimal distribution)
If you want to save the function, the filename containing this curve is built as follows: [pre-
fix]_zon[zone index].crv.
The Statistics Menu
(4.21)
where ( ) is the number of points of the real (optimal) distribution included in

zone , and is the total number of points in the real distribution.
If is zero, that means that the right number of points is included in zone .
If is negative, that means that the number of points included in zone is smaller
than the optimum: there is a lack of points in that zone.
If is positive, that means that the number of points included in zone is larger than
the optimum: there is an excess of points in that zone.
• the evolution of the global deviation of the real distribution (compared to optimal distribution)
(4.22)
If you wants to save the function, the filename containing this curve is built as follows: [pre-
fix]_zonGI.crv.
• additional properties
Additional properties are also evaluated, but only available through the save functions option.
Files are saved in csv format (one file per slice_index) containing, the zone partitioning, the
deviation from optimal distribution, and the concentration of points in each area of the flow
domain. To see these fields, enter Polydata, read the mesh and select convert old csv files, enter
the name of the csv file and save them in files readable by your graphic postprocessor.
The concentration of real points will be evaluated as follows: for each point of the optimal distri-
bution, the number of points of the real and optimal distributions included in a sphere of radius
and center are determined. The concentration at point is:
(4.23)
where = number of points of the real distribution sphere( , ), and = number of points
of the optimal distribution sphere( , ).
If the sample size method you chose is Automatic evaluation, the radius of the sphere is evaluated
as follows. The smallest box that surrounds all the points of the optimal distribution is determined.
Then its volume , is calculated. A typical distance between points is , where is
the number of instants. Eventually, the radius of the sphere is set to 3 .
However, if the sample size method you chose is User specified, you can impose your own value
for the radius (that must be a positive real number).
4.5.2.9. The Sum A <op> Sum B Function

Figure 4.82: The Sum A <op> Sum B Settings for the Create a new function Dialog Box
With this method, it is possible to combine two sum functions with an arithmetic operator (addition,
multiplication, division and subtraction). Select the two sum functions, the operator, and enter the
name of the new function.
The calculation is performed as follows: Define a list of X-values, distributed linearly along the X
axis and enclosed between the X-minimum and the X-maximum of the two functions. Then, for
each X-value, search for the Y-value of the two functions:
and .
The Y-value of the result function corresponding to x will be:
An example of this method is located in Appendix B: The Global Efficiency of Stretching (p. 171).
The Statistics Menu
4.5.2.10. The Smoothing Function

Figure 4.83: The Smoothing Settings for the Create a new function Dialog Box
To create a smoothing function:
1. Select the function to smooth.
2. Select the Type of smoothing you want and specify some Parameters (as explained below).
3. Enter the number of values to represent the smoothed function (a good practice is to use the same
number of values that represent the data function).
4. Name to the new function.
The type of the result function will be the same as that of the data function. Note that it is impossible
to smooth a function that is already a smoothed function.
The method of smoothing is as follows: to calculate one value of the result function, you calculate
the mean of values that surround it in the data function. This process can be iterated several times.
The calculation of the mean can be weighted in different ways (type of smoothing).
With Equal smoothing, the current Y-value and all its neighbors have the same weight.
Figure 4.84: Equal Smoothing
With Centered smoothing, the current Y-value has the highest weight, and the weight decreases
linearly as the distance to the current X-value increases.
Figure 4.85: Centered Smoothing
With Upwind smoothing, first consider the Y-value of the farthest neighbor. The side with the
highest Y-value is weighted more (to the left or the right side of the current X-value).
Figure 4.86: Upwind Smoothing
With Downwind smoothing, first consider the Y-value of the farthest neighbor. The side with the
lowest Y-value is weighted more (to the left or the right side of the current X-value).
Figure 4.87: Downwind Smoothing
In general, good results are obtained with centered smoothing and with the following parameters:
two neighbors at left, two at right, and two iterations.
The Statistics Menu
4.5.2.11. The Density of Probability Function

Figure 4.88: The Density of Probability Function Settings for the Create a new function Dialog
Box
To calculate the evolution of a density of probability function:
1. Specify which probability function to use.
2. Enter the number of values to represent this function (a good practice is to use half the number of
values that represent the probability function).
There is a list of probability functions (one for each slice of a sorted list) and you obtain a density
of probability function for each slice :
(4.24)
If you visualize this function for a given slice, the property values (the property has been chosen
earlier when the probability function has been defined) are on the X axis and the density of prob-
ability is on the Y axis.
Figure 4.89: The Density of Probability Function of a Property for a Slice
In order to avoid wiggles, it is a good practice to smooth the probability function before the calcu-
lation of the density of probability. You must also smooth the density of probability function before
visualization.
4.5.2.12. The Percentile Function

Figure 4.90: The Percentile Settings for the Create a new function Dialog Box
To calculate the evolution of percentiles, you must specify which probability function to use. You
must also define a list of percentiles and give a name to the new function.
The Statistics Menu
To add a percentile to the list (ranged from 0 to 100), enter a value in the line tagged <enter new
value> and press Enter.
To delete one percentile, select it from the list, and click Del.
From the list of probability functions (one for each slice of a sorted list), you obtain the evolution
of a percentile along the slices (one curve for each percentile). If you visualize all of the percentiles,
you have the index of the slice on the X axis and the value of the property (chosen earlier when
the probability function was defined) on the Y axis.
Figure 4.91: Percentiles along the Slices
4.5.2.13. The Histogram Function

Figure 4.92: The Histogram Settings for the Create a new function Dialog Box
To calculate the evolution of histograms, specify which probability function to use. Define a list of
intervals by entering a set of values in the List of values box in the line tagged <enter new value>
and press Enter. To delete one value, select it from the list, and click Del. You must name the new
function.
From the list of probability functions (one for each slice of a sorted list), you obtain a histogram
function for each slice. If you visualize this function for a given slice, the specified intervals of values
(you chose the corresponding property when you defined the probability function) are on the X
axis and the percentage of instants of the slice having a property value in each interval is on the
Y axis.
Figure 4.93: The Histograms for a Slice
The Statistics Menu
4.5.2.14. The Segregation Scale Function

Figure 4.94: The Segregation Scale Settings for the Create a new function Dialog Box
To calculate the evolution of the segregation scale, you need to specify which auto-correlation
on concentration function to use. You must name the new function.
From the list of auto-correlation functions (one for each slice of a sorted list), you obtain the evol-
ution of the segregation scale along the slices. If you visualize this function, the index of the slice
is on the X axis and the value of the segregation scale is on the Y axis.
Figure 4.95: The Segregation Scale along the Slices
4.5.2.15. The Deviation Function

Figure 4.96: The Deviation Settings for the Create a new function Dialog Box
This function will allow you to calculate the gap existing between two density of probability functions.
The second one is an optimal function. There are three different methods to evaluate the deviation,
.
• integral ( abs ( f - f_opt ) ) / 2:
(4.25)
• difference of the means:

(4.26)
• difference of the standard deviations:

(4.27)
where
(4.28)
and
The Statistics Menu
(4.29)
This distribution index defined in The Mixing Theory (p. 11) corresponds to the result of the first
method.
In all cases, you must specify the density of probability function for the real distribution and the
density of probability function for the optimal distribution. Next, you select a method and name
the new function.
You obtain the evolution of the deviation along the slices. If you visualize this function, the index
of the slice is on the X axis and the value of the deviation is on the Y axis.
Figure 4.97: The Deviation along the Slices
The steps to simulate and analyze the distribution of matter are explained in Appendix A: The
Simulation of the Distribution (p. 169).
4.5.2.16. The Points Concentration Deviation Function

Figure 4.98: The Points Concentration Deviation Settings for the Create a new function Dialog
Box
This function allows you to calculate the deviation of points concentration in a slice compared to
a perfect points concentration. This deviation is evaluated for each slice in the selected set of slices.
You must select a set of slices and a coordinate field, and you must introduce the points concen-
tration for a perfect distribution in each slice. This perfect distribution may be evaluated as follows:
• For a 3D open flow domain, you perform a slicing in space, in the direction of the flow; therefore, each
slice will be a surface. It is assumed that each slice has the same perfect points concentration, which is
equal to the number of points in a slice divided by the area of the part of the slice that cuts the flow
domain.
• For a 3D closed flow domain, you perform a slicing in time; therefore, each slice will be a volume. It is
assumed that each slice has the same perfect points concentration, which is equal to the number of
points tracked divided by the volume of the flow domain.
• For a 2D open flow domain, you perform a slicing in space, in the direction of the flow; therefore, each
slice will be a line. It is assumed that each slice has the same perfect points concentration, which is equal
to the number of points in a slice divided by the length of the part of the slice that cuts the flow
domain.
The Statistics Menu
• For a 2D closed flow domain, you perform a slicing in time; therefore, each slice will be a surface. It is
assumed that each slice has the same perfect points concentration, which is equal to the number of
points tracked, divided by the area of the flow domain.
In order to evaluate the points concentration in each slice, you must specify the radius of the sample
around each point; all points in the neighborhood of point at a distance smaller than this radius
will be taken to evaluate the local points concentration at position . This radius must be chosen
carefully; if it is too small, no points will be found and the local concentration will not be relevant;
if it is too large, the global concentration will be identical for all points in the slice, and will not
change from slice to slice. A first estimation is to take as the radius a tenth of a typical distance in
the slice (for example, the diameter of a rotor or of a screw).
You can also specify the minimum number of points needed to evaluate the points concentration
at a position ; the default is set to 3. If the number of neighboring points is smaller than this
minimum, Polystat progressively increases the radius (for that position only), until it has the correct
number of neighbors.
Next, you must specify the dimensionality of the samples. If the samples are straight line segments
(dim=1), surfaces (dim=2) or volumes (dim=3). This is needed to evaluate the local points concen-
tration , as explained in Deviation of Points Concentration (p. 23).
The deviation in a slice is evaluated as follows:
(4.30)
where is the number of points in the slice , { } corresponds to the location of point in slice
, and is the perfect points concentration.
You must name the function.
4.5.3. New disagglomeration functions

If you select New disagglomeration functions in the Statistics menu of the main window, you see
a window showing the list of functions that can be calculated on properties derived from properties
linked to the disagglomeration model.
Figure 4.99: Types of New Disagglomeration Functions
The first three functions are specific to disagglomeration model and are explained in the next sections,
while the last three functions are identical to general statistics functions (see explanations in The
Smoothing Function (p. 149), The Percentile Function (p. 152), and The Histogram Function (p. 154)).
The Statistics Menu
4.5.3.1. The See disagglomeration along a single trajectory Function

Figure 4.100: The See disagglomeration along a single trajectory Settings for the Create a
new function Dialog Box
With this function, you see the time (or displacement) evolution of the disagglomeration property
calculated along any trajectory you want.
1. Specify which single trajectory to use.
2. Select the disagglomeration property to see.
3. Specify the time or space interval where the disagglomeration property will be stored.
If you visualize this function, the size of agglomerates is on the X axis and the density of probability
to find agglomerates of a given size is on the Y axis. In this graph, you have a new function y=f(x)
for each time (or displacement) step.
Figure 4.101: Disagglomeration Functions for Two Time Steps
4.5.3.2. The Mean disagglomeration function

Figure 4.102: The Mean disagglomeration function Settings for the Create a new function
Dialog Box
1. Select a set of slices.
2. Select a disagglomeration property.
3. Select a type of statistical treatment: Minimum, Maximum, Mean, Standard deviation, Standard
deviation for values below the mean, and Standard deviation for values above the mean.
The Statistics Menu
With this function and the Minimum treatment selected, a min disagglomeration function is eval-
uated for each slice, based on the disagglomeration property known at each point included in the
slice.
(4.31)
With this function and the Maximum treatment selected, a max disagglomeration function is
evaluated for each slice, based on the disagglomeration property known at each point included in
the slice.
(4.32)
With this function and the Mean treatment selected, a mean disagglomeration function is evaluated
for each slice, based on the disagglomeration property known at each point included in the slice.
(4.33)
Figure 4.103: The Mean Disagglomeration Function for a Slice
With this function and the Standard deviation treatment selected, a standard deviation disagglom-
eration function is evaluated for each slice, based on the disagglomeration property known at each
point included in the slice.
(4.34)
With this function and the Standard deviation for values below the mean treatment selected,
another standard deviation disagglomeration function is evaluated for each slice, based on the
disagglomeration property known at each point included in the slice.
(4.35)
With this function and the Standard deviation for values above the mean treatment selected, a
last standard deviation disagglomeration function is evaluated for each slice, based on the disag-
glomeration property known at each point included in the slice.
(4.36)
4.5.3.3. The Probability function

Figure 4.104: The Probability function Settings for the Create a new function Dialog Box
Select a density of probability function and give a name to the new function.
With this function, for each slice is evaluated, on the probability function as the integral of
a density of probability function on a property .
Additional Definitions
(4.37)
4.6. The Help Menu

Three options are available in the Help menu:
• The Polystat User’s Guide ... option opens the ANSYS Help to the table of contents of the Polystat User's
Guide.
• The About ... option opens a window that displays installation details, including the external version numbers
of the ANSYS products you have installed (which change with each release) and the internal build ID of
Polystat (which may or may not change with each release). When speaking with your technical support en-
gineer, it is the internal build ID that is more relevant.
Figure 4.105: The Help Menu Options
4.7. Additional Definitions
4.7.1. The Slices

A slice is the set of instants of the trajectories that respect some specified condition. For example, a
condition is, the time seconds or the position must be included in the plane ,
and so on. The general way to define a condition is the following: for a property , the value of
must be included in a plane (this plane is defined by a position and a normal direction).
When defining a slice, the following dialog box appears:
Figure 4.106: The Define a slice Dialog Box
1. Select a property where the condition will apply.
2. Specify the plane to respect by entering the position and normal direction in the number boxes.
3. Name the slice.
If the selected property is a scalar, the value to respect is specified in the rectangle 1 (the second and
third number boxes are set to zero). The normal direction must be set to (1,0,0).
4.7.2. The Zones

A zone is an interval of values for a specified property. For example, for a vector property (as the
position), you have to define the two extreme corners of a box.
Figure 4.107: Defining a Position Zone
If the property is a scalar, define the minimum and the maximum values of the interval.
When defining a zone, the following dialog box appears:
Additional Definitions
Figure 4.108: The Define a zone Dialog Box
1. Select a property where you will define the zone.
2. Specify the interval: enter the minimum values in the boxes numbered 1 to 3, and the maximum value
in the boxes numbered 4 to 6
3. Enter the name of the zone.
If the selected property is a scalar, the minimum value is specified in the 1st number box and the
maximum value is specified in the number 4 box (boxes 2, 3, 5 and 6 are set to zero).
4.7.3. Weighting
The slicing must been defined on a coordinate field, in order to weight a property by the velocity.
The general formula for the sum, the mean and the standard deviation are:
(4.38)
(4.39)
(4.40)
where the index indicates the instant of slice , is the value of at instant , and is the
weight at instant .
To calculate the probability function , the set of pairs is sorted:

(4.41)
where is the number of instants in slice .
Finally the probability to find a value below is:
(4.42)
The value of the weight is:
• if no weighting:
• if weighting everywhere:
Note:
This should only be

used if you are sure
that all of the
velocities in a slice
have the same
general direction as
the normal of the
slice. Otherwise, this
leads to a negative
weight.
• if weighting when is positive: if

, otherwise
where is the dot product of the velocity at instant by the normal of the slice .
Appendix A. The Simulation of the Distribution
Let's suppose the flow to be 3D, steady-state, in a closed domain.
To perform the analysis of the distribution, the following steps are necessary.
1. We calculate the flow (velocity and pressure).
2. We calculate in the mixing task 1 the real distribution:
We suppose the material points to be initially concentrated in a small volume. We have to calculate
their trajectory for a given time interval. We store these trajectories in files named real_0001 to
real_000X (for example).
3. We calculate in the mixing task 2 the optimal distribution:
We suppose the material points to be initially distributed in all the flow domain. We have to calculate
their trajectory for a very short time interval (infinitesimal amount of time). We store these traject-
ories in files named opti_0001 to opti_000X.
4. In Polystat, we read the files real_0001 to real_000X, and the files opti_0001 to opti_000X.
5. We define two sets of trajectories: the first one, named "real_set" contains all the trajectories from the
real_* files; the second one, named "opti_set" contains all the trajectories from the opti_* files.
6. We define two sets of slices: the first one, named "real_slicing", is a slicing on the time (N slices, every Δ
seconds), for the "real_set" set of trajectories. The second one, named "opti_slicing", is one slice defined
for time t=0 and on the "opti_set" set.
7. We define two Distance Distribution functions: the first one, named "real_distribution", is based on the
"real_slicing" set of slices. The second one, named "opti_distribution", is based on the "opti_slicing" set
of slices.
8. We define one Deviation function, to calculate the deviation of the real distribution ("real_distribution"
function) from the perfect distribution ("opti_distribution" function).
Appendix B. The Global Efficiency of Stretching
Let's suppose the flow to be 2D, steady-state, in a closed domain.
We will explain how we can calculate the time evolution of the global efficiency of stretching (see
chapter 2 for the definition of this parameter).
For example, for the linear stretch, at time , this efficiency is:
(1)
To calculate such a parameter, the following steps are necessary:
1. In the mixing files are calculated the stretching and the dissipation rate along trajectories.
2. We have to define, in Polystat, a new property, the cumulated dissipation, which is the time integration
of the dissipation rate:
(2)
This new parameter depends on the material point and on time.
3. We perform a slicing on the time ( slices, every Δ seconds).
4. We calculate the sum function of the stretching :
(3)
5. We calculate the sum function of the cumulated dissipation :
(4)
6. Finally, we divide the two sum functions to obtain the global efficiency of stretching:
(5)
Appendix C. Adaptive Meshing and Mixing Task
Example 65 explains the procedure to follow in the case adaptive meshing has been used to evaluate
the flow field. Indeed, two limitations of the mixing module forbid to use meshes whom some parts
have been sub-divided. Firstly, in case of transient simulations, the mesh of the flow domain can/will
change with time: new elements and new nodes are created/deleted. Secondly, as the mesh is refined
locally, non-conformity appears: sub-divided elements are adjacent to non sub-divided elements. As
the tracking of material points does not allow such cases, we propose the following procedure:
• In Polydata, during the definition of the flow task, we ask for the output of CSV results files.
• In Polyflow, we evaluate the transient flow, using adaptive meshing; at each time step, a CSV file is generated
that contains the current flow field.
• In Polydata, we perform a conversion of the CSV files into new Polyflow results files onto a new uniformly
refined mesh (the mesh of the flow domain will remain constant in time); at the end of this step, we have a
new mesh and a set of res files (1 res file per time step) compatible with this new mesh.
• In Polydata, we define a mixing task, based on the new mesh and the new results files.
• In Polyflow, we compute the trajectories of a set of material points.
• In Polystat, we perform a statistical analysis on these trajectories.
Appendix D. Sliding Mesh Technique and Mixing Task
When sliding mesh technique is used, two limitations of the mixing module forbid the use of the
computed flow field to perform tracking of material points. Firstly, in the case of transient simulations,
the mesh can / will change with time (position of the nodes). Secondly, the mesh is non-conformal. As
the tracking of material points does not allow such cases, we propose the following procedure:
• In your mesher application, define a mesh constituted of different subdomains for the flow domain: one
subdomain is fixed, while the other(s) is (are) rotating. In the figure below, the flow domain is composed of
disconnected subdomains 1 to 3. The subdomain #1 is fixed, while subdomain #2 is rotating with the left
cam and subdomain #3 is rotating with the right cam. The internal part of the cams are not be meshed.
Figure 1: Disconnected Subdomains
• In Polydata, during the definition of the flow task, we use the sliding mesh technique; we define motion of
the rotating subdomains. Connected boundary conditions are needed between tangential boundaries of
fixed and rotating subdomains. Moreover, we ask for the output of CSV results files.
• In Polyflow, we evaluate the transient flow, using sliding mesh technique; at each time step, a CSV file is
generated that contains the current flow field.
• In your mesher application, define a new mesh, where the flow domain is conform, and with new subdomains
overlapping the flow domain and corresponding to the moving parts (as if we had to define subdomains
for the mesh superposition technique). The position of the moving parts must correspond to the initial po-
sition of the moving parts used for the simulation of the flow field!
Sliding Mesh Technique and Mixing Task
Figure 2: Overlapping Subdomains
• In Polydata, we perform a conversion of the csv files into new Polyflow results files onto this new mesh (the
mesh of the flow domain will remain constant in time); at the end of this step, we have a set of res files (1
res file per time step) compatible with the new mesh.
• In Polydata, we define a mixing task, based on the new mesh and the new results files.
• In Polyflow, we compute the trajectories of a set of material points.
• In Polystat, we perform a statistical analysis on these trajectories.
Bibliography
[1] B. Alsteens. "Mathematical Modelling and Simulation of Dispersive Mixing". PhD thesis. Université
Catholique de Louvain, Louvain-le-Neuve, Belgium. 2005.
[2] B. Alsteens and V. Legat. “A Model for the Disagglomeration of Carbon Black in Rubber Matrix”. Proceed-
ings of the 6th Conference Eurheo, Erlangen, Germany. September 2002.
[3] R.G.C. Beerkens, T. van der Heijden, and E. Muijsenberg. "Possibilities of Glass Tank Modeling for the
Prediction of the Quality of Melting Processes". Ceram. Eng. Sci. Proc.. 14(3–4). 139–160. 1993.
[4] V. Collin and E. Peuvrel-Disdier. "Dispersion Mechanisms of Carbon Black in Elastomers". Conference on
European Rubber Research – Practical Improvement of the Mixing Process, Paderborn Germany.
2005.
[5] P. V. Dankwertz. "The Definition and Measurement of some Characteristics of Mixtures". Applied Science
Research. 3. 279–296. 1952.
[6] J. M. Ottino. "The Kinematics of Mixing: Stretching, Chaos and Transport". Cambridge University Press.
1989.
[7] J. M. Ottino, W.E. Ranz, and C. W. Macosko. "A Framework for Description of Mechanical Mixing of Fluids".
AIChE Journal. 27. 4. 565–577. July 1981.
[8] Tadmor and Gogos. "Principles of Polymer Processing". John Wiley and Sons. 1979.
[9] T. H. Wong and I. Manas-Zloczower. "Two-Dimensional Dynamic Study of the Distributive Mixing in an
Internal Mixer". Intern. Polymer Proc.. IX. 1. 3–10. 1994.
[10] H-H. Yang and I. Manas-Zloczower. "Analysis of Mixing Performance in a VIC Mixer". Intern. Polymer
Proc.. IX. 4. 291–302. 1994.

ANSYS Polystat Users Guide

Uploaded by

Copyright:

Available Formats

ANSYS Polystat Users Guide

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ANSYS Polystat Users Guide

Uploaded by

Copyright:

Available Formats

ANSYS Polystat User's Guide

ANSYS, Inc. Release 2020 R1

© 2020 ANSYS, Inc. Unauthorized use, distribution or duplication is prohibited.

U.S. Government Rights

Published in the U.S.A.

3.6. Parameters for the Generation of the Material Points ....................................................................... 49

4.2.4.5. Mass of agglomerates of given size .............................................................................. 122

3.26. The Generation Zone 1 Menu .............................................................................................................. 52

4.86. Upwind Smoothing ........................................................................................................................... 150

2. The Contents of the ANSYS Polyflow Manuals

1.2. How to Characterize Mixing?

1. Calculate the flow pattern.

1.3. Classification of Flows, Capabilities of the Mixing Module

The following are concept definitions:

Figure 1.1: Open and Closed Domains

Figure 1.2: Example of a Piecewise Steady Flow

Figure 1.3: Example of a Spatially Periodic Flow

The mixing task has the following limitations:

• For a piecewise steady flow, there must not be any inertia.

• The flow must always be incompressible.

There are no limitations on:

• The fluid model: generalized Newtonian or visco-elastic models are available.

1.4. General Explanation on the Way to Solve a Mixing Task

Figure 1.4: The Flow Simulation

Figure 1.5: The Mixing Simulation

combining existing parameters in various ways . Polystat can also

Figure 1.6: The Statistical Postprocessing

1.5. The Numerical Techniques Involved in the Mixing Module

Successive positions of the particle in real space is calculated using:

The mixing module algorithm proceeds as follows:

a. Find the element , which contains the initial position .

b. Find the local coordinates, , of in element .

a. Integrate Equation 1.3 (p. 6), until a boundary of element is crossed.

ii. Determine the local coordinates in element of the current position ;

iii. Return to step 2.a.

1.6.1. Example 50: The Rectangular Cavity

1.6.2. Example 51: Coextrusion of Polymers in a Square Channel

1.6.3. Example 52: Flat Die

1.6.4. Example 46: Periodic Flow Through a Kenics Mixer

1.6.5. Example 37: Mixer 2-D

1.6.7. Example 91: Dispersion

Keywords: dispersion, disagglomeration, erosion, rupture, Polystat

1.6.8. Example 116: Distributive Mixing Analysis in the Mixer GK7N

1.7. Known Limitations in ANSYS Polystat 2020 R1

– ANSYS Polystat does not support mixing tasks on CutCell meshes.

– ANSYS Polystat does not support mixing tasks on compressible fluids.

2.2. Kinematic Parameters

The dot denotes the material time derivative.

2.2.1. Kinematic Parameters for 2D Flows

However, there is another way to define a mean efficiency over time:

2.2.2. Kinematic Parameters for 3D Flows

For an incompressible fluid:

If the fluid is incompressible, the normal direction to the surface is:

However, time-averaged efficiency can also be defined as:

2.2.3. Statistical Analysis

2.2.3.1. Mean and Standard Deviation

2.2.3.2. Cumulated Probability Function (or Distribution Function)

Figure 2.1: Distribution Function Graph

2.2.3.3. Density of Probability Function