z_mat_manual - Z-set

z_mat_manual - Z-set
Material
& Structure
Analysis Suite
Materials manual
Version 8.6
Z-set 8.6 is distributed by
Transvalor S.A.
Centre des Matériaux
B.P. 87 Ð 91003 EVRY Cedex
France
http://www.zset-software.com
[email protected]
Neither Transvalor, ARMINES nor ONERA assume responsibility for any errors appearing in this
document. Information provided in this document is furnished for informational use only, is subject
to change without notice, and should not be construed as a commitment by the distributors.
Z-set, ZebFront, Z-mat, Z-cracks and Zebulon are trademarks of ARMINES, ONERA and Northwest
Numerics and Modeling, Inc.
c
ARMINES
and ONERA, 2015.
Proprietary data. Unauthorized use, distribution, or duplication is prohibited. All rights reserved.
Abaqus, the 3DS logo, SIMULIA, CATIA, and Unified FEA are trademarks or registered trademarks
of Dassault Systèmes or its subsidiaries in the United States and/or other countries.
ANSYS is a registered trademark of Ansys, Inc.
Solaris is a registered trademark of Sun Microsystems.
Silicon Graphics is a registered trademark of Silicon Graphics, Inc.
Hewlett Packard is a registered trademark of Hewlett Packard Co.
Windows, Windows XP, Windows 2000, and Windows NT are registered trademarks of Microsoft
Corp.
Contents
Introduction
What’s in this manual
Conventions . . . . . .
Introduction to Z-mat
System Requirements
Material Frameworks .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1.1
1.3
1.5
1.7
1.9
1.11
Z-mat General Commands
Z-mat interfaces . . . . . .
Interface files . . . . . . . .
***automatic time .
***behavior . . . . .
***debug . . . . . .
***external storage .
***material . . . . .
***parameter . . . .
***save energies . .
***skip cycle . . . .
Zmaster interfaces . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2.1
2.3
2.5
2.8
2.9
2.10
2.11
2.12
2.18
2.19
2.20
2.21
.
.
.
.
.
.
.
.
.
3.1
3.3
3.5
3.7
3.9
3.11
3.13
3.15
3.17
3.19
Z-mat ANSYS
Zansys ANSYS interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1
4.3
Z-mat MSC-Marc
Z-mat MSC-Marc interface . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1
5.3
Z-mat SAMCEF
Z-mat SAMCEF interface . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1
6.3
.
.
.
.
.
.
.
.
.
.
Z-mat ABAQUS
Z-mat ABAQUS interface
Current status . . . . . .
Site definition . . . . . . .
Interface files . . . . . . .
Output variables . . . . .
Extra files . . . . . . . . .
User additions . . . . . .
Example . . . . . . . . . .
Post calculations . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Z-mat Cosmos
Z-mat Cosmos/M interface . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.1
7.3
Z-mat LS-Dyna
Zlsdyna LS-Dyna interface . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.1
8.3
Behavior functionality
Introduction to Behaviors
Grad-Flux . . . . . . . . .
Material variables . . . . .
Material file . . . . . . . .
BEHAVIOR . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
9.1
9.3
9.5
9.7
9.11
9.13
Material Models
linear elastic . . . . .
damage elasticity . . .
hyper elastic . . . . .
linear viscoelastic . . .
viscoelastic spectral .
hyperviscoelastic . . .
gen evp . . . . . . . .
reduced plastic . . . .
porous plastic . . . . .
mechanical step phase
umat . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10.1
10.3
10.4
10.5
10.6
10.9
10.12
10.13
10.16
10.18
10.24
10.25
Secondary Models
aging . . . . . . . .
aniso damage . . .
becker needleman .
cast iron . . . . . .
bodner partom . .
finite strain crystal
matmod . . . . . .
matmod z . . . . .
memory . . . . . .
non associated . .
visco aniso damage
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
11.1
11.3
11.5
11.7
11.9
11.11
11.13
11.15
11.17
11.19
11.20
11.22
Other Models
coefficient diffusion . .
needleman debonding
crisfield debonding . .
chaboche debonding .
generalized debonding
diffusion . . . . . . . .
linear spring . . . . . .
thermal . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
12.1
12.3
12.4
12.6
12.8
12.10
12.14
12.15
12.16
.
.
.
.
.
.
.
.
.
.
.
variable friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Material Components
ANISOTROPIC DAMAGE . . . . . . . . .
ANISOTROPIC DAMAGE scalar . . . . .
ANISOTROPIC DAMAGE elastic tensorial
COEFFICIENT . . . . . . . . . . . . . . . .
COEFFICIENT MATRIX . . . . . . . . . .
CONDUCTIVITY . . . . . . . . . . . . . .
CRITERION . . . . . . . . . . . . . . . . .
CRITERION anisotropic . . . . . . . . . . .
CRITERION bron . . . . . . . . . . . . . .
CRITERION cast iron . . . . . . . . . . . .
CRITERION hill . . . . . . . . . . . . . . .
CRITERION karafillis boyce . . . . . . . .
CRITERION linear drucker prager . . . . .
CRITERION mises . . . . . . . . . . . . . .
CRITERION modified nouailhas . . . . . .
CRITERION nouailhas . . . . . . . . . . .
CRITERION ratio . . . . . . . . . . . . . .
CRITERION tensile mises . . . . . . . . . .
CRITERION tresca . . . . . . . . . . . . .
CRITERION unsym . . . . . . . . . . . . .
CRITERION 2M1C . . . . . . . . . . . . .
CRYSTAL KINEMATIC . . . . . . . . . .
CRYSTAL ORIENTATION . . . . . . . . .
DAMAGE . . . . . . . . . . . . . . . . . . .
DIRECT KINEMATIC . . . . . . . . . . .
DIRECT KINEMATIC asaro . . . . . . . .
DIRECT KINEMATIC non symmetric . . .
ELASTICITY . . . . . . . . . . . . . . . . .
FLOW . . . . . . . . . . . . . . . . . . . . .
FLOW function . . . . . . . . . . . . . . . .
FLOW gsell . . . . . . . . . . . . . . . . . .
FLOW hyperbolic . . . . . . . . . . . . . .
FLOW modified visco . . . . . . . . . . . .
FLOW sellars tegart . . . . . . . . . . . . .
FLOW sum . . . . . . . . . . . . . . . . . .
INTERACTION . . . . . . . . . . . . . . .
ISOTROPIC . . . . . . . . . . . . . . . . .
KINEMATIC . . . . . . . . . . . . . . . . .
KINEMATIC aniso nonlinear . . . . . . . .
LOCALIZATION . . . . . . . . . . . . . . .
POROUS CRITERION . . . . . . . . . . .
POROUS CRITERION cam clay . . . . . .
POROUS CRITERION elliptic . . . . . . .
POROUS CRITERION elliptic aniso . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
12.17
13.1
13.3
13.5
13.6
13.7
13.9
13.13
13.14
13.15
13.16
13.17
13.18
13.19
13.20
13.22
13.23
13.24
13.25
13.26
13.27
13.28
13.29
13.30
13.31
13.35
13.38
13.39
13.41
13.43
13.44
13.46
13.47
13.48
13.49
13.50
13.51
13.52
13.54
13.57
13.60
13.61
13.63
13.64
13.65
13.66
POROUS CRITERION fkm . . . . . . . . .
POROUS CRITERION gurson . . . . . . .
POROUS CRITERION rousselier . . . . . .
POROUS CRITERION modified rousselier
POROUS CRITERION zhang niemi . . . .
POTENTIAL . . . . . . . . . . . . . . . . .
POTENTIAL associated . . . . . . . . . . .
POTENTIAL coupled recovery . . . . . . .
POTENTIAL delobelle . . . . . . . . . . .
POTENTIAL crystal . . . . . . . . . . . . .
POTENTIAL gen evp . . . . . . . . . . . .
POTENTIAL gen evp2 . . . . . . . . . . .
POTENTIAL suvic . . . . . . . . . . . . . .
POTENTIAL 2M1C . . . . . . . . . . . . .
POTENTIAL z6 gen evp . . . . . . . . . .
SINTERING STRAIN . . . . . . . . . . . .
SLIP INTERACTION . . . . . . . . . . . .
STRAIN NUCLEATION . . . . . . . . . .
THERMAL STRAIN . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
13.68
13.69
13.70
13.71
13.72
13.73
13.74
13.76
13.77
13.80
13.83
13.86
13.87
13.89
13.90
13.91
13.92
13.93
13.94
.
.
.
.
.
.
.
.
.
.
14.1
14.3
14.5
14.6
14.7
14.8
14.10
14.11
14.12
14.13
14.14
Model Simulation
Model Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
****simulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
***test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15.1
15.3
15.4
15.6
Optimization
Introduction . . . . . . . . . . . .
****optimize . . . . . . . . . . .
***compare . . . . . . . .
***compare t file file . . .
***compare i file file . . .
***compare i func file . .
***constraint . . . . . . .
***comparison constraint
***files . . . . . . . . . .
***function . . . . . . . .
16.1
16.3
16.5
16.10
16.12
16.14
16.15
16.16
16.17
16.18
16.20
Modifiers
MODIFIER
MODIFIER
MODIFIER
MODIFIER
MODIFIER
MODIFIER
MODIFIER
MODIFIER
MODIFIER
MODIFIER
. . . . . . . . .
lagrange polar
lagrange rotate
plane stress . .
auto step . . .
runge jacobian
runge rollover .
perturbation .
explicit . . . .
bifurcation . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
***value . . . . . . . . . . . .
***shell . . . . . . . . . . . .
***zrun . . . . . . . . . . . .
****optimize nelder mead . . . . . .
****optimize levenberg marquardt .
****optimize evolution . . . . . . . .
****optimize augmented lagrangian
****optimize single . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
16.21
16.26
16.27
16.28
16.30
16.33
16.35
16.37
Reference
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Environment Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17.1
17.3
17.5
Bibliography
18.1
Index
19.1
Chapter 1
Introduction
Z-set — Non-linear material
& structure analysis suite
1.1
1.2
Z-set — Non-linear material
& structure analysis suite
What’s in this manual
Description:
This Z-mat user-commands handbook covers the command syntax and some of the details for
the different files related to use of Z-mat. This includes the material behaviors, utilities and
extra programs, and interfaces to the different FEA codes which Z-mat works with.
For users who are interested only in the Z-mat interface to another FEA solver the primary
source of information will be in the Examples/Training and this Z-mat User commands
manual. There is supplementary information in the other books however, such as user development, and model theory. The Zmaster program has increasingly expanded its capabilities,
and will continue to explore new interfaces with other software products, The Post computations section in the Z-set manual also has information on making batch post computations
of imported results from various codes.
Handbook Summary:
The following list summarizes the documentation for all of Z-set. As part of our 8.2/8.3
developments, greatly expanding the software documentation is one of our primary goals.
The list below is sorted in what we feel would be an appropriate sequence for the normal
user, starting with installation and reviewing capabilities, to creating input files and eventually
scripting and developing add-ons to the software.
Release Notes/Zmaster The basic overview of the software, installation instructions, and
documentation for the graphical user interface Zmaster on all platforms. The Zmaster
manual also now covers all the base reference chapters such as environment variables,
user parameters, function reference, command line programs, etc.
Examples/Training This book is essentially the “getting started” documentation for the
software. The book describes Z-mat, simulation, optimization, material models, and
the FEA code use. The examples cover setup of complete models, and are meant to
demonstrate the capabilities with relatively simple examples.
Z-mat User commands This summarizes the command file formats and capabilities of the
Z-mat interface, simulation, optimization and material files for all of Z-mat and Z-set.
Z-set User commands This summarizes the command file formats for the FEA related
capabilities including meshing, FEA solution, post processing, etc.
Developer A guide to the user-extensible features of the software, including scripting,
ZebFront material model development, making plugins, and the C++ programming
API.
Plugins A guide to add-on features available.
Theory Theory manual covering formulations (under development).
Z-set — Non-linear material
& structure analysis suite
1.3
1.4
Z-set — Non-linear material
& structure analysis suite
Conventions
This page summarizes the conventions used for the Z-set input files. An overview of the general
command syntax (command hierarchies) is given in the beginning of the Examples/Training
manual.
• Running of Z-set modules generally requires that a “problem name” be given. Most
input and output data files are based on this name with a variety of suffixes attached.
Henceforth, problem will often be used to indicate the problem name given while running
the commands.
• The characters % and # indicate that the rest of the line is a comment. For example:
***load % external problem loading
• There are no abbreviations allowed in the use of keywords. All keywords and command
names must be written entirely.
• The admissible characters for the names of user variables are: a-z, A-Z, ", +, -, *,
., =, /, , ), (, \, ~
• The text entry is always case-sensitive.
• The use of braces [ ] in the syntax descriptions indicates an option with a default
definition.
• All parameter values used in the input files and described as “real” in the syntax descriptions must have a decimal point. All standard specifications of floating point values
are accepted. Two examples of real values are: 3.0 4.2e-5
• The use of parenthesis () indicates data input of real values in vector form. An example
is: (0.1 0.2 1.0). In most cases the size of such vectors must be compatible with the
overall problem dimension.
The symbol indicates a section where the calculation is sensitive to input data or format.
Z-set — Non-linear material
& structure analysis suite
1.5
1.6
Z-set — Non-linear material
& structure analysis suite
Introduction to Z-mat
Starting with version 7.2 of Zebulon (approx 1997), the program has been modularized into
components relating to finite element methods, and components for material simulation and
optimization. These later programs are grouped into Z-mat which, while of course is natively
implemented with all of Z-set, also includes interfaces to other codes. One of the more
interesting aspects of this is to be able to use Z-set material behaviors as an “extension
set” to codes which support user materials such as ABAQUS/Standard, ABAQUS/Explicit,
ANSYS, LS-Dyna, MARC, and Cosmos/M.
The Z-mat product is thus mainly a library of material behavior routines (constitutive
equations) which can be interfaced with FEA software, and its supporting utilities. In contrast with other FEA software products however, Z-set and Z-mat are built on C++ with a
strong object-oriented design, and many utility programming classes for advanced tensorial
mathematics. Such advanced program design techniques have enabled the software to be
enormously extensible, on many different levels of sophistication.
In developing extensions to Z-mat, a user can make new behaviors using the many abstract
“building bricks” to create a modular, flexible model based on fundamentals. For final use, we
then let the different implemented “bricks” define the specifics. That is to say, the user always
chooses the final model form (e.g specific evolution or state equations) in the input file, after
development is finished. In that way, each model is really a “model class” or framework within
which a user has a high degree of flexibility. One can also add to the “types” available for these
building bricks. This not only lets one work on a smaller specific aspect, but allows extending
the capabilities of every model using that particular “abstract” type. In this book, examples
of the general behavior frameworks are the subject of the chapters Material Models (page
10.3) Secondary Models (page 11.3) and Other Models (page 12.3). The material “bricks” are
described in the Material Components chapter beginning on page 13.3.
In addition to the material behaviors, and the development kit, the Z-mat product also
ships with a number of additional programs bundled. One such program is the ZebFront
pre-processing language which greatly aids the development of user models. Others which
are optionally included are the Simulation module and the Optimization module. These two
combine to make a very powerful material coefficient identification tool1 . One can efficiently
simulate and identify using the exact same model code as is used in the finite element code.
This will virtually eliminate an extra testing/verification step, and ensures compatibility of
behavior between the different stages of material modeling. The optimization works efficiently
on multiple experimental/simulation data sets, and can even be used in conjunction with
structural calculations.
Finally with versions 8.3 and greater, Z-mat has been expanded to include more “method
level” capabilities, such as pre and post processing of results, along with other analysis add
ons such as cycle skipping and user elements. It is our hope that Z-mat will grow to be a
large-scale CAE complimentary product for commercial and academic users.
1
although they are general purpose and can be used with subjects that have nothing to do with materials,
or finite elements
Z-set — Non-linear material
& structure analysis suite
1.7
1.8
Z-set — Non-linear material
& structure analysis suite
System Requirements
System Requirements
Z-mat is an integral part of Z-set, and therefore is implemented on all the Z-set platforms
2 Floating licenses allow the user to employ model simulation and coefficient calibration on
desktop systems, while the Z-mat interface can be used on different platform servers optimized
for large FEA runs. Also, all file formats used by Z-set and Z-mat are platform independent,
including the reading of externally created binary files such as the abaqus .fil or ansys .rst
files. Specifics outlined below relate to the interfaces with the Z-mat available solvers.
External interfaces are, at the time of this writing, available for ABAQUS 6.3-6.4, MARC,
Ansys version 7.x, LS-Dyna 970, and COSMOS/M 2.7. Machine platforms for ABAQUS
include Solaris (32 bit), IRIX (64 bit), HP-UX 11, HP-UX 11/Itanium, AIX 5, OSF1, Linux
i86, Linux Itanium, and windows (NT4/2000/XP). The COSMOS interface is available on
windows platforms only. Z-mat for Ansys is available on Windows, SunOS (64-bit), and
IRIX64. Z-mat for MARC is available on 64 bit IRIX only at this time.
Some of the platforms have limitations due to the methods of the different solvers. Information on these is described in the different Z-mat chapters. Normally we find the interface
for ABAQUS the most robust, with Ansys next, then Cosmos followed by MARC.
The library is compiled as a dynamic shared object which can be linked to other programs.
The library is entirely programmed in C++, but only a system linker (ld) is required for basic
use of the program. Again, some specifics relating to the different platforms are relevant. In
particular for Ansys and Cosmos the user needs to compile a custom executable because we
are not allowed to distribute a pre-linked Z-mat executable. To have user routines attached
to the basic library (plug-ins), a C++ compiler is necessary. Which C++ to use is determined
by the requirements of the interfacing code (e.g. ABAQUS), and one should seek information
from that vendor for compatibility. For use with Z-set only a g++ version is available for all
platforms. All Win32 platforms use Microsoft Visual C++.
On SGI systems the development foundation package may be required on older systems.
This package is supplied with all the MIPSpro compilers, or can be purchased separately.
The package is required for the system libraries and linker which are not supplied with the
standard system. The C++ compiler is required only for user additions to the package. Also,
the software is currently compiled on IRIX 6.5, which will generate an undefined symbol error
for earlier systems. A workaround is now available, so if you need that please contact your
distributor.
2
The development of Z-mat interfaces requires of course operating copies of the mating code at our site
for development and maintenance. It is frequently possible for additional platforms to be added per request
(possibly with a supplemental charge). Please inquire to your vendor regarding any additional such desired
ports.
Z-set — Non-linear material
& structure analysis suite
1.9
System Requirements
1.10
Z-set — Non-linear material
& structure analysis suite
Material Frameworks
Material Frameworks
One thing which needs mentioning right away with regards to the Z-mat documentation is
that the material models are in general not documented with respect to the type of material,
but rather the model framework. Each framework can be considered an assembly of different
user-created “building blocks” which then finalize the full model definition.
The commands for configuring Z-mat can be classified as follows:
• Controls and context are commands which adjust the specific numerical run (Note:
for Zebulon runs these controls are defined in the Z-set users manual, so this section
does not apply). These are all ***-level commands in the Z-mat control file (filename
depends on specific interface chosen).
• Framework This part is determined basically by a selection of a ***behavior modeltype pair. All sub-commands after the behavior declaration will be context-sensitive.
• Components
Z-set — Non-linear material
& structure analysis suite
1.11
Material Frameworks
1.12
Z-set — Non-linear material
& structure analysis suite
Chapter 2
Z-mat General Commands
Z-set — Non-linear material
& structure analysis suite
2.1
2.2
Z-set — Non-linear material
& structure analysis suite
Z-mat interface
Z-mat interfaces
The current version of Z-mat (Z-mat 8.3) supports the following interfaces.
• Z-mat ABAQUS Both ABAQUS/Standard and ABAQUS/Explicit are supported. With
ABAQUS there are perhaps the most comprehensive set of features due primarily to
the very robust user capability within ABAQUS. In this regard, Z-mat is constantly looking for additional paths of interface, and with the 8.3 version we have added significant
ODB input and output functions, and also UEL interfacing for all the Zebulon element
formulations.
• Z-mat ANSYS The Zansys interface has improved immensely in terms of stability and
robustness due to the new ability for ANSYS to load shared libraries dynamically.
• Z-mat MSC-Marc Marc fully supports Z-mat material models, and there is an interface
for results files reading.
• Z-mat SAMCEF SAMCEF is supported by the Transvalor group in France only.
• Z-mat LS-DYNA The LS-DYNA fully supports the user materials numbers 41 and 42 currently. The remaining material numbers are left for the user to optionally select. The
Z-mesh, Z-post, and Zmaster codes allow for inputting of the k file as well as the d3plot
results files. There is currently no output to the LS-DYNA formats.
• Cosmos/M Cosmos is no longer supported because of lacking customer demand. This
port could potentially be re-activated upon request.
Later in the manual these interfaces are discussed in detail.
Z-set — Non-linear material
& structure analysis suite
2.3
Z-mat interface
2.4
Z-set — Non-linear material
& structure analysis suite
Interface files
Interface files
The Z-mat interface file exists as a medium to configure the interface, translation, and provide
extra convergence parameters for the local integration. This interface effectively duplicates
the controls which exist in the Zebulon FEA input file.
All the different Z-mat platforms (target FEA systems) use the same basic controls described in this chapter. Some certain commands may however have functionality specific to
one or some of the codes due to limitations or special features available in the user interfaces.
These differences will be noted in the description section for each command.
Note that the interface file name itself is very dependent on the particular solver which is
being interfaced to. For example, with ABAQUS materials are assigned character names, and
that name will be the same ASCII filename to be opened by Z-mat. On other solvers such
as ANSYS only a material number is given as an identifier, and in that case a convention of
naming the file as: 100+material number+“.txt” (for example 105.txt for material id 5).
Please double-check in the different chapters specific to each solver for further information.
Syntax:
The following commands are available in the Z-mat interface file. Reading of the commands
will stop when a ***return command or the end of file is reached.
***automatic_time
***behavior
***debug
***external_storage
***material
***parameter
***save_energies
***skip_cycle
***state_var_engineering_shear
***state_var_no_change
***state_var_real_shear
***suppress_temperature
***symmetrize_tgmat
***verbose
***plane_stress_modifier
***zero_dt_for_first_dt
These commands and their options are the subject of the rest of this chapter. Some of
the commands are simply switches and thus do not have any sub-commands. These simple
commands will be discussed in the following.
***state var no change This command indicates that the user does not require changing
the Z-mat internal format of√shear variables to a real measure. Shear components output
are
√ multiplied by a factor 2 from the actual tensor component (t12 in the output is
2t12 ).
Z-set — Non-linear material
& structure analysis suite
2.5
Interface files
***state var engineering
shear Transform shear variables to be output with an addi√
tional factor of 2. (e.g. γ12 = 212 ). This is now the default.
√
***state var real shear Divide the 2 term out of the shear components so the output
is the real component of the tensor.
***suppress temperature Eliminate the setup for temperature as a parameter. This
optimizes slightly the computations so constant coefficients may be assumed.
***symmetrize tgmat Make an extra step to symmetrize the material tangent returned
from the Z-set behavior.
***verbose verbose message outputs.
***plane stress modifier This command may be used to de-activate the automatic
detection of the plane stress condition and allows the use of an explicitly defined
plane_stress modifier of the behavior instead. Note that automatic plane stress treatment is not implemented for all behaviors, and that in the case of anisothermal loadings
convergence may be very slow, In those two cases the plane_stress modifier method
should be preferred.
***zero dt for first dt use a zero ∆t for the pre-step test (with zero strain increment).
This pre-step is used by ABAQUS to get the initial tangent modulus. Time-dependent
materials may be non-linear if the time step is > 0 even with zero strain increment,
because of stress relaxation under fixed displacement.
The Z-mat material file:
The Z-mat material file is the input which determines which constitutive model from the
available models in Z-mat to use, and then specifies the particular coefficients accordingly.
This file is the same as input for Zebulon, and for the Z-sim. The input format is the subject
of chapters 4 and 5.
Note:
By default, the input file name for the material file is the same as the Z-mat interface file, so
the behavior definition can follow the Z-mat interface commands.
. . . continued
2.6
Z-set — Non-linear material
& structure analysis suite
Interface files
Example:
The following small material file is a simple example of Z-mat use (taken from the ABAQUS
interface tests which puts together some of these concepts. Many more examples will of
course be given in the different code-specific chapters and around the discussion of each subcommand.
***suppress_doing_first
***state_var_no_change
***suppress_temperature
***material
*integration theta_method_a 1.0 1.e-10 1500
*initialize_variable
epcum 0.43
X11
7.529e-03
X22
-1.065e-02
X33
3.118e-03
***behavior gen_evp
**elasticity isotropic
young 2.1e5
poisson 0.3
**potential gen_evp ep
*flow plasticity
*criterion mises
*isotropic nonlinear
R0 200.
Q 2000.0
b 0.26
*kinematic nonlinear X
C 25500.0
D 81.
***return
Z-set — Non-linear material
& structure analysis suite
2.7
Interface files
***automatic time
***automatic time
Description:
This command is used to give parameters controlling the time stepping based on the convergence and gross change in material variables. Variables which control the time step may be
taken from the FLUX or VINT data members.
Note:
This automatic time-stepping controls the global convergence stepping, which means that if
there is a local divergence or violation of a limit variable, the current increment will be thrown
away and the global solution time step reduced. One can use the auto_step beahvior modifier
to control local-only automatic time stepping (see page 14.8). This command depends on the
capabilities of the FEA solver used. ABAQUS allows the material to control the next time
step to be used, while Cosmos and Ansys only allow the material to sub-cut the time step by
2.
Syntax:
The syntax for the automatic time stepping control is the following:
***automatic_time
[ *security factor ]
[ *divergence div ]
*limit var1 val1 [ . . . varN valN ]
where the following parameters are used:
factor real value giving the maximum time increase factor for a well converged time step.
The value must be greater than one.
div
dividing factor for a diverging increment. The global solution step will be re-run with
a time step smaller than the rejected one by this factor.
var1
a character name of a variable of the problem. This can be from the FLUX, or VINT
variables. If the name is the base name of a tensorial or vector set of variables, the limit
will be placed on all the components of that variable (e.g. if sig is given all the stress
components will be used).
val1
real value for the limit of the last-given variable name. Limit names/values must be
given in pairs.
Example:
Here is a small typical use for viscoplasticity type problems. The time step will be limited both
in the increment of viscoplastic strain (evcum to changes less than 0.1% and stress component
changes to less that 15 MPa. All components of the stress tensor will be checked to be within
the limits.
***automatic_time
*limit
evcum
1.e-3
sig
15.
*divergence 2.0
*security
1.2
2.8
Z-set — Non-linear material
& structure analysis suite
Interface files
***behavior
***behavior
Description:
This command is more fully described in the chapter Material Models and Material Components. This command lets the user include in-line a material definition in the Z-mat interface
file. If this command is not used, the material file will be sought using the *file or *standard
sub-commands of ***material.
Be aware of the following items when defining a Z-mat behavior.
• Behavior definitions have the “external parameter” value temperature available always
for coefficient definitions, thermal strains, and similar coefficient taking material objects.
With ABAQUS, additional field variables can be defined using the field options.
• One may use the Lagrangian modifiers described on page 14.5 to transform the behavior
into finite strain when the conditions for calculating the deformation gradient in a UMAT
have been met (see ABAQUS user manual 25.2.261 ).
• For use with ABAQUS/Explicit one must add the material modifier explicit.
1
from version 5.6 handbook
Z-set — Non-linear material
& structure analysis suite
2.9
Interface files
***debug
***debug
Description:
This command indicates that debug output will be included during the run. This is primarily
for user-defined functions and behavior models using the prn set of C++ functions. The
command will also print out the active list of auto-load keywords so one may verify if a
particular function is loaded with the Z-mat/ABAQUS link.
Output from the debug statements will be in a file named OUT located in the given directory
(full path). If no path is given, and the library is unable to determine the problem directory,
the full path of the output file is /tmp/OUT. This path sometimes helps to get around the
problem copying to a scratch directory by ABAQUS.
Syntax:
***debug [ path ]
[ *local_debug ele gp ]
[ *flags flags ]
[ *limit_debug_time st end ]
where path is the path to use for storing the OUT debug file.
*local debug localize the debug output to a given element number and a given Gauss
point number.
*flags Set flags for the internal int DeBuG. This is used control what gets printed. One
can use the prn2 to make certain debug selections. See the developer manual for more
information.
*limit debug time Limit the time for debug output. Give the start and end time for
debug output in the solution time scale.
Example:
The following example is taken from material input file e3danis located in test database
directory Z − mat/umat v61.
***debug
***material
*integration runge_kutta 1.e-3 1.e-3
2.10
Z-set — Non-linear material
& structure analysis suite
Interface files
***external storage
***external storage
Description:
This option is used to specify that the state variable storage is to be made in a separate file
instead of using the FEA solver’s internal database. The command currently works with the
ABAQUS and ANSYS interfaces.
With this command, there is no limit to the number of state variables which can be used
in a Z-mat model. In some cases we find that the codes efficiency can be improved as well by
using external storage, because of the buffering and relief from some copying operations.
Syntax:
***external_storage
*buffer_size sz
*file db-tmp-file
*full_path db-tmp-file
*vars list
*buffer size The buffer size determines the file buffering given in number of integration
points. This means that only every sz integration points will the storage be serialized
to disk.
*file Specifies the external file to be used for the storage. This file name is relative to the
current working directory.
*full path Specifies a fully qualifies filename for the external storage. Note that ABAQUS
copies your input problem to a temporary directory, so do not use relative path names
for this one.
*vars Specifies the variables which will be made available for output. Zpreload can be used
to determine the variable naming which is required here (the Zebulon type names, not
sdv##. In the screen or log output the sdv-type naming will be printed, for example:
done with material file reading...
** real state variables
sdv(1)
epcum
sdv(2)
epi22
Example:
The following lines are an excerpt from a Z-mat material input file doghri st, which demonstrates the usage of the external storage keyword. The material file can be found in the
test database in /Z − mat/umat v61.
***external_storage
*file oo.store
*vars epi22 epcum
*buffer_size 5000
Z-set — Non-linear material
& structure analysis suite
2.11
Interface files
***material
***material
Description:
This command marks the definition of the materials in a structure to be studied. The behavior
of each material is defined in a file with special syntax (see the chapter Material Behavior).
The purpose of this command is therefore to define the material file names, associate these
files to different element sets, and specify other global applications on top of a material model
such as rotation of material coordinates or give local integration methods.
Syntax:
***material
[ *file file ]
[ *standard std-file ]
[ *integration
]
[ *rotation
]
[ *initialize_variable ]
The sub-commands for ***material pertain to the behavior defined in the current material file only. Note that the function of this command is somewhat different than the
equivalently named command used in the Z-set .inp file.
A last comment: Always Verify Your Materials. The behaviors supplied in the Z-mat
library are compatible with the simulation program, so there is no excuse to not validate the
material behavior with a given set of coefficients.
2.12
Z-set — Non-linear material
& structure analysis suite
Interface files
***material
*integration
*integration
Description:
This option determines the local integration method for a material behavior.
Syntax:
*integration
method params
The allowable methods are summarized in the table below:
CODE
runge kutta
DESCRIPTION
explicit Runge-Kutta integration with automatic time stepping based on integration error
theta method a
implicit generalized midpoint integration; this method normally supplies the best tangent matrix
theta auto a
automatic time stepping in the implicit θ-method
runge kutta The Runge-Kutta method implements a second order explicit integration with
automatic time stepping. Variables are normalized to allow varied variable magnitudes
in “stiff” sets of equations. The method takes two real parameters. These are the
convergence criteria followed by a minimum value for normalization. Standard RK
error calculation for each integrated variable will be normalized by either the increment
of the variable or this second parameter, whichever is greater. the resulting error is
compared with the first parameter.
The Runge-Kutta integration with the gen_evp material behavior provides a tangent
matrix in models with a single inelastic deformation. This matrix is however not consistent with the integration scheme, and thus yields less than optimal global convergence.
The explicit integration also performs poorly in heavily time-dependent problems such as
viscoplasticity. However, some complex models are only implemented with this method.
theta method a The θ-A method is the standard integration for the majority of material
laws requiring integration.
x(t + ∆t) − x(t) = ẋ (t + θ∆t) ∆t
This method requires 3 parameters to describe the convergence. These are first the θ
value (real) followed by the residual required for convergence (real) and the maximum
number of local iterations in the integration (integer).
The value for θ must be greater than zero and less than one. It is strongly advised
to use theta values of 1 for time independent (plastic) materials, and 1/2 for time
dependent (viscoplastic) problems. Time independent plasticity will normally show
strong oscillations about the solution for values of θ less than 1.
Reasonable values of convergence range from 10−6 to 10−10 . Values which are too large
usually lead to poor global convergence. Too small values will not converge due to
numerical roundoff (10−12 is about the limit). Convergence will rarely take more than
25 iterations, and should not take more than 50. If this is the case, there may be some
Z-set — Non-linear material
& structure analysis suite
2.13
Interface files
***material
*integration
error in the integration (make a bug report), or the material parameters are excessive
(damage laws may provoke this). If the local iterations are greater than 50 it is probably
better to reduce the global iterations or use automatic time stepping (global or local).
The default integration is dependent on the material law used. Most behaviors modeling
plastic or viscoplastic materials use a default of the θ-method with theta = 1.0 ,eta = 1.e-9
and max iteration = 200.
Example:
% plasticity or large deformation
*integration theta_method_a 1.0 1.e-9 50
% difficult viscoplastic case
*integration theta_method_a 0.5 1.e-6 100
% complex law
*integration runge_kutta 1.e-3 1.e-3
2.14
Z-set — Non-linear material
& structure analysis suite
Interface files
***material
*rotation
*rotation
Description:
This material option is used to change a coordinate systems by rotation. It is used here to
simplify specification of some materials (anisotropy, etc), but the syntax is general. Other applications using the rotation object include specification of grain orientations for polycrystals
(see page 13.61).
There are currently two methods for specifying a rotation. These are by vectors of the
rotated coordinate axes in the global coordinate system, and by Euler angles (used for crystal
orientation for example). The first case is displayed in the following figure:
y’
y
y
x1
apply
x’
apply
x’
x1
x3
x
x
z’
y’
z
For the material rotation of this section, the material gradient will be rotated (rotation is
applied) before being integrated by the material behavior. For small deformation mechanics,
this would be a rotation of the strain tensor.
0tot = RT tot R
The material behavior then solves for the flux in terms of the new gradient, which is 0tot → σ 0
for the mechanical problem. Afterwards the flux is rotated to the global coordinates again:
σ = Rσ 0 RT
Rotation by giving Euler angles is similar. The significance of the three angles is given in the
following figure:
Z
Z’
Z’
φ
Y’
Y’
φ2
Y’
Y
φ1
X
X’
X’
X’
. . . continued
Z-set — Non-linear material
& structure analysis suite
2.15
Interface files
***material
*rotation
Syntax:
For rotations specified using coordinate axes:
*rotation
[ x1 x∗ y ∗ [z ∗ ] ]
[ x2 x∗ y ∗ [z ∗ ] ]
[ x3 x∗ y ∗ [z ∗ ] ]
The arguments x1, x2, x3 indicate the components of direction vectors for the transformed
coordinate frame. Exactly one direction is required in 2D problems, and two directions are
required in 3D. The order of definition is not important. The local coordinate system may be
assembled with any of the geometrical axes. The input vectors will also be normalized by the
program to automatically make unit vectors.
Using the notation here that t1 is the first direction vector defined, and t2 is the second
(for 3D problems), direction vectors of the coordinate system are defined as follows: The first
vector is collinear to t1. The second vector is a vector in the plane defined by t1,t2 and is
perpendicular to t1. The third direction will always be calculated using the vector product
of the first two vectors. The t vectors will replaced by those given by you using the x1, x2,
x3 choices.
For rotations specified using Euler angles:
*rotation
2.16
φ1 φ φ2
Z-set — Non-linear material
& structure analysis suite
Interface files
***material
*initialize variable
*initialize variable
Description:
This command is used to give initial values to the variables of a material. Currently, only
constant (uniform) values are allowed. This command is used in the place of the initializing
commands in ABAQUS.
Syntax:
The syntax for the automatic time stepping control is the following:
***initialize_variable
var1 val1
...
varN valN
The syntax is free format, except for the fact that variable name and initial values must
be given in pairs. The definable parameters are:
vari
vari
a character name of a variable of the problem. This can be from the FLUX, or VINT
variables. Each component of a tensorial or vector variable must be initialized separately.
real value for the limit of the last-given variable name.
Example:
This is an example for the problem 4022301 of ABAQUS. The input file is the following (this
is for a custom behavior integrating the same Ziegler model as in ABAQUS:
***material
*integration theta_method_a 1.0 1.e-7 1500
*initialize_variable
epcum 0.43
X11 128.0
X22 -181.0
X33
53.0
***behavior ziegler_test
...
***return
Z-set — Non-linear material
& structure analysis suite
2.17
Interface files
***parameter
***parameter
Description:
This command controls the translation of ABAQUS field variables to Z-mat external parameters which can be used in coefficient dependancies. Remember that the temperature
parameter is always active in Z-mat.
Syntax:
***parameter
[ **ambient_temperature val ]
[ **field_variable var-name location ]
[ **initial_value val ]
**ambient temperature sets the ambient temperature used as T0 . This is important for
thermal strain calculations (see page 13.94). In the event that this command is not
used, an additional state variable is added with fixed value of the temperature at the
beginning of the problem (if there are dependancies on temperature in the behavior that
is). It is probably desirable to use this command therefore if the initial temperature
field is constant.
**field variable add a new field variable to the problem, with index location in the list
of fields defined within the ABAQUS problem. This index starts with 1.
*initial value optional command with a similar meaning as **ambient temperature
has for temperature.
Example:
The following example illustrates the use of parameter keyword. It is an excerpt of file ts1
located in test database in Z − mat/umat v61.
***parameter
**ambient_temperature 125.0
**field_variable humidity 2
**initial_value 0.25
2.18
Z-set — Non-linear material
& structure analysis suite
Interface files
***save energies
***save energies
Description:
This command indicates stores the elastic energy for output. The command is a bit simplistic,
but extracts from the material the variable eel and returns
1
sse = σ : el
2
if el does not exist in the material (not normally the case for standard materials, but possible)
an error will occur, and the option should be removed.
Z-set — Non-linear material
& structure analysis suite
2.19
Interface files
***skip cycle
***skip cycle
Description:
The ***skip cycle command is used to give cyclic based extrapolation of the material state
to allow skipped cycles for structures loaded with many cycles.
Syntax:
***skip_cycle
*check_with_component list-of-components
*file sdv-file
2.20
Z-set — Non-linear material
& structure analysis suite
Zmaster interfaces
Zmaster interfaces
Description:
Many of the different Z-mat platforms have integrated functionality for their mesh input
and results files within the other Z-set products including the Zmaster GUI program. This
software is what NW Numerics uses exclusively for validation and problem processing for all
the different interfaces. Generally the Z-mat user could be interested in the following:
• Opening results files for visualization and rendering. The following simple commands
are examples:
Zmaster -odb my_calculation.odb
Zmaster my_calculation.fil
Zmaster ansys_calc.rst
• The 8.3.6 version includes the Simulation GUI interface for setting up and working with
material simulation and fitting work.
• Mesh files can be imported and saved to different formats with sets and other boundary
condition data preserved.
• Many of the Z-mat validation test cases use a post processing step to extract X-Y data
in an automated manner. Zmaster can be used to plot general ASCII file data using
the Plot button:
Zmaster doghri.test
More detail on the file format translators and use of Zmaster is given in the Release
Notes / Zmaster handbook.
Z-set — Non-linear material
& structure analysis suite
2.21
Zmaster interfaces
2.22
Z-set — Non-linear material
& structure analysis suite
Chapter 3
Z-mat ABAQUS
Z-set — Non-linear material
& structure analysis suite
3.1
3.2
Z-set — Non-linear material
& structure analysis suite
Z-mat interface
Z-mat ABAQUS interface
Description:
The commands described in this chapter are used to enable a Z-mat material model for use
within ABAQUS1 .
Syntax:
Z-mat jobs are launched via the Zmat command line script only.
% Zmat [ opts ] problem ←The command syntax for Zmat is given in more detail in the Release Notes & Zmaster
handbook, including discussion of all the command line switches used to control the specifics
of the Z-mat launch.
The script will initially launch ABAQUS, and then be called again by the ABAQUS preprocessor in compile and link mode. By default the launch submits the job to the
ABAQUS queue, and therefore the script exits rapidly (even though the job continues in
the background). To see the status of the calculation, look at the problem.log file (e.g.
tail -f myprob.log), or run with the -fg switch.
ABAQUS Input:
The definition of a material for the Z-mat behaviors always uses an external file to establish
the model components and coefficients. Z-mat never uses the material parameters defined
in the ABAQUS .inp file. This is because the Z-set coefficient definitions are much more
flexible in their definitions, and are integral to the actual structuring of the material model
to be used. Global dependencies are established using the temperature parameter, or other
global field variables.
Other commands are available in addition to the behavior definition which control the
local integration method (implicit mid-point, Runge-Kutta explicit, etc), variable initializations, automatic time stepping parameters depending on the maximum allowable variable
increments, and the local rotations.
Linking summary:
The Z-mat library is delivered as a dynamic shared object which can be linked to other
programs. The library is entirely programmed in C++, but only a system linker (ld) is required
for basic use of the program. The Zmat script prepares the proper link command in place of
using the C++ or Fortran command lines. On Windows platforms no development software is
necessary at all.
To have user routines attached to the basic library, a C++ compiler will be necessary,
as a second add-on shared library must be prepared2 . Since the standardization of C++, the
compiler requirements are greatly reduced from what was previously the case. In most systems
1
The Z-mat interface is also known as Z-aba or ZeBaBa in some older circles. The Z-mat name is however
the official product name and symbolizes an approach which is a step beyond classical single code UMAT
solutions, and one which is tied to the extensive Z-set software.
2
actually there is no limit to the number of plugins which can be used. There is a specific naming convention
for Z-mat plugins however. For unix systems the user-library should begin with libZmat and on windows the
DLL should start with zmat.
Z-set — Non-linear material
& structure analysis suite
3.3
Z-mat interface
Z-mat plugins can be made which whatever C++ compiler is convenient. Normally the user
would be best advised to use the most modern version possible. The general compiler level
requirements specified by ABAQUS, Inc will most likely not be absolutely necessary, though
we validate that those compilers do in fact work so they can be used as a guideline.
User plugins will always be compiled and linked before running ABAQUS, and the enduser will therefore not need any additional development tools in the same manner as running
Z-mat without the plugins.
3.4
Z-set — Non-linear material
& structure analysis suite
Current status
Since the initial releases, many improvements have been made for the ABAQUS interface of
Z-mat. Some issues still remain however.
Some of the additions with this version include:
• Continuing streamlining and performance improvements. The distributed domain
solvers in ABAQUS at versions 6.5 and above are well supported and are proven to
give decent scalability with Z-mat.
• Support for ABAQUS/Explicit is complete.
• Improved automatic time stepping.
• Improvements in the simulation/optimization tools, including support for finite strain
problems.
• Compatibility with the rest of Z-set, making one installation for sites with Z-set and
Z-mat.
• External disk storage for state variables. This allows models with thousands of state
variables.
• Finite strain has been improved, including now the addition of hyperelastic and Hyperviscoelastic models.
• Direct reading of the .fil format is greatly improved. Abaqus ODB files can also be
read and written by Zpost/Zmaster and Zebulon. This data file interface is without
question the most robust of all the Z-mat interfaces.
• All the Z-mat test cases use Z-post.
• Support for all Abaqus 6.x versions.
• Zebulon elements can be used as a UEL with output to an ODB file or other Zebulon
supported output. This includes full state variable name and typing information.
• Optional post processing step to translate all the Z-mat SDV# variable names to be the
proper named and typed variable in a new ODB file.
Z-set — Non-linear material
& structure analysis suite
3.5
3.6
Z-set — Non-linear material
& structure analysis suite
Site definition
Site definition
There are a number of site definition files which must be configured to make the Z-mat program
additions work with ABAQUS. These will most likely be configured by NW Numerics during
the installation.
The location of the abaqus executable to use is adjustable. There are two possibilities for
defining the path. The first is to edit the file $Z7PATH/lib/Zmat/ABAQUS_ROOT and make an
entry for the machine name (the name given via hostname) and the path to the ABAQUS
installation directory. The second method is to define the environment variable ABAQUS_ROOT
to point to the abaqus directory.
When launching the script will automatically link the Z-mat library and all plugins with
the libZmat (unix) or zmat (win32 DLL) prefixes found in the standard search path (see the
Zmaster/release manual reference section on search paths).
Unix configuration:
The most important issue for the Z-mat programs to work is the definition of an environment
variable Z7PATH to point to the root directory of the Z-mat distribution (see release notes).
The next important thing is to source the $Z7PATH/lib/Z7_cshrc file to set up additional
environment variables. These two definitions can be put in a users .cshrc file for the C-shell.
The second most important environment variable is the Z7MACHINE variable. All the scripts
which access binary executables use the Z7MACHINE variable to determine where and which binary to run. By default (if the variable is not set before sourcing the $Z7PATH/lib/Z7_cshrc
file) the machine type will be set as setenv Z7MACHINE ‘uname‘. Sometimes this is not sufficient to distinguish an architecture, so the user may add additional machine types. Additional
architectures may be defined for different machines in the $Z7PATH/lib/MACHINE_TYPES file.
This file is also used to define different compilers to be used. The Release Notes/Zmaster
manual has more information in the Configuration chapter.
Windows configuration:
Aside from specifying the ABAQUS_ROOT position, no special work is needed for running Zmat
on windows platforms. Only a command line interface is available however.
Z-set — Non-linear material
& structure analysis suite
3.7
Site definition
3.8
Z-set — Non-linear material
& structure analysis suite
Interface files
Interface files
The fact that Z-mat is developed for use in Z-set (Zebulon FEA) natively, and that the
interface with codes like ABAQUS is based on a Fortran subroutine implies that a certain
amount of translation back and forth is required. Also, there are additional inputs required to
control Z-mat outside of ABAQUS. The following figure outlines The interaction of different
programs, and the use of input files.
Z−sim
Zebulon FEA
FEA Code
(e.g. ABAQUS)
FEA code input
(e.g. ABAQUS .inp file)
UMAT
FEA/Z−mat
translation and
control routine
Z−mat input
file
Integration
method
Z−mat
behavior
integration
Z−mat behavior
file (model definition,
coefficients)
The ABAQUS .inp file:
The use of an externally defined material behavior in ABAQUS requires a number of special
entries to be made in the ABAQUS .inp file. Some of these entries are superfluous given
the structure of the Z-mat library, but are nevertheless required to satisfy ABAQUS’s umat
interface. The entries are:
• *MATERIAL command to define the different materials of the problem. The name given
to this command (using *NAME=) will be the name of an external file containing a Z-mat
behavior definition (see previous section).
• *DEPVAR command which establishes the storage per integration point for the material
variables. This value must be greater or equal (best case) of the size determined by
Z-mat for the behavior. There is no way except for user intervention to size this value,
so the line must be included. If the value is too small for the behavior, an error message
is printed in the .log file, and the calculation terminates.
• *USER MATERIAL this defines that there is a user behavior. As as parameter to this
command, one must give a CONSTANTS option, with at least one coefficient. This
coefficient is not used in the Z-mat behavior however.
Z-set — Non-linear material
& structure analysis suite
3.9
Interface files
ABAQUS commands for initializing the state variables, giving a material orientation, and
material coefficients are not used with Z-mat. There are instead alternatives which may be
defined in the separate Z-mat material file (see the commands starting at page 2.12).
3.10
Z-set — Non-linear material
& structure analysis suite
Output variables
Output variables
All the material variables should be available for output in ABAQUS calculations if the
**external_storage option is not used. This should include the GRAD, FLUX, VINT and VAUX
variables as shown with the Zpreload utility. The variables will all be stored as SDV variables.
Caution
There are some tricky spots which remain in the Z-mat output which may not be obvious
to the new user. These are summarized below:
• The
√ shear components of symmetric tensional variables are natively stored with a factor
of 2, which is removed by default at each exit of the z-mat routine. This operation takes
up some CPU however, so it can be suppressed by using the ***state_var_shear_alter
command. Also the default transformation for state variables is to change them to
the real tensorial shear component. Engineering output is available by using the
***state_var_engineering_shear option. This is true for variables in the integrated
vector and the auxiliary variables (VINT and VAUX variables). The GRAD and FLUX variables always have the factor removed however.
• The integrated and auxiliary variables (VINT and VAUX) are stored sequentially in the
vector prefixed sdv in ABAQUS.
• For mechanical problems, the stress variable is named S and the strain E with all behavior
as in ABAQUS standard.
• Some extra variables such as the energy output variables are not yet calculated.
Z-set — Non-linear material
& structure analysis suite
3.11
Output variables
3.12
Z-set — Non-linear material
& structure analysis suite
Extra files
Extra files
The compilation and linking of UMAT routines (Z-mat included) are controlled by parameters
in the abaqus_v6.env file (located in the Site directory of the ABAQUS distribution.
Alterations must therefore exist in the active abaqus_v6.env file to link with the Z-mat library. An abaqus_v6.env to do this is supplied in the $Z7PATH/lib/Zmat/ directories. There
is a default abaqus_v6.env file, and the option to have specific files for different machines
(example files exist and it should be obvious how to manipulate these). The abaqus_v6.env
file which is needed will be copied into the calculation directory in order to be the first read
configuration file, and can thus be verified by looking at that file after an execution (possibly
add comments to make the env file creation clearer).
The user can add configuration options in their customized personal files also. The standard abaqus_v6.env file will not be used in this case, replacing that with the file named
zebaba_v6.env (replaced only for Zmat jobs). The Zmat program will search first in a users
home directory, followed by the local execution directory.
The environment file shipped in the Z-mat for HP-UX is for example:
link_sl="Zmat -link_v6 /usr/ccs/lbin/ld64 -b -ashared_archive +k +n
+FPD +vnocompatwarnings +pd L +pi D +s -c%E +h%U -o %U %F %A %B %L
-lpthread -lcps -lcl"
compile_cpp="Zmat -compile_v6"
compile_fortran="Zmat -compile_v6"
In the standard Z-mat validation tests in the directory test/Z-mat/umat_v61 we have
the custom file zebaba_v6.env with the following:
ask_delete=OFF
split_dat=ON
For personal files one may wish to remove the ask_delete parameter.
Z-set — Non-linear material
& structure analysis suite
3.13
Extra files
3.14
Z-set — Non-linear material
& structure analysis suite
User additions
User additions
Extensions to Z-mat:
User functions and classes may be added to the Z-mat library without limit. This is one of
the advantages of working with Z-mat, as a site can create a library of user functionality in
combination with Z-mat, and use it repeated without distributing the source files for each
problem run. Also, the user classes will be automatically available to the simulation and
optimization methods, creating a coherent extensible modeling environment.
In order to make user-additions to Z-mat, a custom shared library must be compiled
and installed into the proper location with the proper name. For running Z-mat with other
ABAQUS, the shared library must be named libZmat*.so where * is replaced by a user
designated character filename. On HP-UX systems the suffix is .sa and on AIX it is .a.
Files fitting the above wildcard name will be automatically loaded using the system dlopen
command (or equivalent). The search path is:
• the current working directory.
• the path pointed to by the environment variable ZEBU_PATH
• the default binary location $Z7PATH/PUBLIC/lib-$Z7MACHINE
More detail about the compiling process is given in the developer handbook, and there is
always a pre-configured user-project in the installation directory $Z7PATH/User-project .
ABAQUS user routines:
The Z-mat package uses the ABAQUS UMAT function to interface with that FEA solver.
This does not prevent one from programming extra user routines to run along with Z-mat.
If the user routines are unrelated to the material (such as user loads or MPCs), an extra
Fortran file can simply be appended to the Z-mat interface by using the Zmat option -uf.
An example is provided in the +$Z7PATHtest/Z-mat/UMAT+ directory as problem 4020201.
The execution is for example:
Zmat -fg -uf mpc.f 4020201
which will link in the user routines in mpc.f.
The user can override the default UMAT interface as well, by modifying the file
+$Z7PATHlib/Zmat/mech-sd.c+ and using the -UF command switch.
Z-set — Non-linear material
& structure analysis suite
3.15
User additions
3.16
Z-set — Non-linear material
& structure analysis suite
Example
Example
This is an example material file which can be used with Z-mat. Note that all input is case
sensitive!
This example has the material definition in the file e3danis. The following lines are the
material related entries for the ABAQUS .inp input file:
*MATERIAL,NAME=e3danis
*DEPVAR
20
*USER MATERIAL,CONSTANTS=1
0.0
*SOLID SECTION,MATERIAL=e3danis,ELSET=A1
Note that this syntax is very sensitive to positions, and connectivity. Commands are not
case sensitive, but filenames may be. The command *DEPVAR allocates in ABAQUS storage
for the material variables at each Gauss point. There is no way to dimension this quantity
from the UMAT routine, so the user is obliged to enter the correct number corresponding to
the material model used. Information concerning this quantity will be printed to the .log
file generated by the ABAQUS output.
***material
*integration runge_kutta 1.e-3 1.e-3
***behavior gen_evp
**thermal_strain isotropic
alpha temperature
2.0e-06
-0.1
1.0e-06 1000.1
**elasticity isotropic
young 200000.
poisson 0.25
**potential gen_evp ev
*criterion mises
*flow norton
n 4.0
K 500.
*kinematic nonlinear % x1
C 10000.0
D 100.0
*isotropic constant
R0 100.e+00
***return
For this material file, the output of the Zpreload program shows the material variables
and their associated ABAQUS output variables.
============================================
Flux Name:
Z-set — Non-linear material
& structure analysis suite
3.17
Example
sig11
sig31
sig22
sig33
sig12
sig23
Grad Name:
eto11 eto22
eto31
eto33
eto12
eto23
var_int Name:
eel11(sdv1) eel22(sdv2) eel33(sdv3) eel12(sdv4) eel23(sdv5)
eel31(sdv6)
evcum(sdv7)
al111(sdv8) al122(sdv9) al133(sdv10) al112(sdv11) al123(sdv12)
al131(sdv13)
var_aux Name:
evi11(sdv14) evi22(sdv15) evi33(sdv16) evi12(sdv17) evi23(sdv18)
evi31(sdv19)
============================================
3.18
Z-set — Non-linear material
& structure analysis suite
Post calculations
Post calculations
Description:
With version 8.3 all the Z-set post computations can be used with ABAQUS results files,
and are in fact used for each of the Z-mat validation examples. It is sufficient to use the
***data_source selection to choose an import format for the results files. The post processing
environment contains a large number of very convenient curve extraction routines which can
be used to automate analysis post processing. There are also a large number of both global and
local post computations which can be applied for failure analysis etc. The post computations
are now part of the Z-mat bundle.
Example:
An example from the Z-mat tests follows, generating an ASCII datafile with the stress strain
behavior as computed at node 2345 of the structure. Note that some mesher operations can
be included in the post computation as well to generate nsets for example which were not
included in the abaqus model during the pre-processing stage.
****post_processing
***data_source fil
**open doghri.fil
***global_post_processing
**file node
**output_number 1-999
**nset ALL_NODE
**process curve doghri.test
*precision 4
*node 2345
eto11 sig11
****return
Note:
For Zansys the capability and input is quite the same thing, except that the file format should
be ***data_source rst and the problems .rst file should be placed in the open input.
Z-set — Non-linear material
& structure analysis suite
3.19
Post calculations
3.20
Z-set — Non-linear material
& structure analysis suite
Chapter 4
Z-mat ANSYS
Z-set — Non-linear material
& structure analysis suite
4.1
4.2
Z-set — Non-linear material
& structure analysis suite
Zansys interface
Zansys ANSYS interface
Description:
The Zansys functionality mirrors very much the same principals as in the Z-mat for ABAQUS
port, and so we refer the user to that documentation as well. This section provides a getting
started guide to Zansys.
Syntax:
% Zansys [ opts ] problem ←Getting started:
Perform a standard installation of Z-set, including the binaries and shared files. The install
location will henceforth be referred to as the Z7PATH. The launch procedures for Zansys will
also need to be able to locate the ansys
Note:
In versions prior to 8.3.6 it was necessary to perform a user build and link process in order
to generate a user executable of ansys. The newer versions interface with ANSYS via shared
libraries, and henceforth the user needs to do nothing special for the interface to work. There
are some test cases in the %Z7PATH%\test\Zansys\INP directory. The Z-mat material files
are named 10#.txt where # is the material number selected in the test using ansys commands
such as:
TB,USER,4,0,0
TB,STATE,4,,40
MPCHG,4,1
which would set the Z-mat material for element group 1 to be read in the file 104.txt.
Currently all the commands listed in the Z-mat handbook for the ABAQUS interface are
supported with ANSYS as well, except the use of multiple field variables. Temperature is
available as a parameter, and can be set using commands like:
BFUNIF,TEMP,523.0
The name TEMP will be re-mapped to temperature in the Z-mat files. To try the test cases,
do for example:
z:
cd %Z7PATH%\test\Zansys\INP
Zansys plast3
which launches the ansys GUI from which the input file can be loaded:
/INPUT,plast3,inp
Elements and output
Z-mat for ansys must be run using the 18x class elements. In order to get output for the
state variables in the Z-mat material model, an additional command must be issued to get
the variables stored to the output database, such as:
Z-set — Non-linear material
& structure analysis suite
4.3
Zansys interface
OUTRES,SVAR,ALL,
To plot the variable one can use the command:
PLESOL,SVAR5,1
Note:
With user materials in ANSYS the solver is by default set to not extrapolate integration point
stresses to the nodal points. In fact ANSYS only does this extrapolation by default for linear
materials, while most other codes assume that a least-squares fit extrapolation method should
be done in all cases. In order to activate extrapolation the following lines can be used before
the SOLVE command is issued:
ERESX,YES
With the Z-set RST file reader direct integration point visualization is supported, so unfortunately there is no refined solution for this issue.
4.4
Z-set — Non-linear material
& structure analysis suite
Chapter 5
Z-mat MSC-Marc
Z-set — Non-linear material
& structure analysis suite
5.1
5.2
Z-set — Non-linear material
& structure analysis suite
Zmarc interface
Z-mat MSC-Marc interface
Description:
The command described in this chapter allows to use a Z-mat behavior within a MSC-Marc
analysis. The interface with the Z-mat library is implemented by means of the uvscpl Marc
user subroutine. The choice of uvscpl to build the interface, against other candidate subroutines available to implement user materials, has been motivated by its flexibility. Note however
that Zmarc capabilities are by no means restricted to time-dependent creep behaviors.
Syntax:
% Zmarc -j problem [ run marc options ]
←-
where problem.dat is the name of an MSC-Marc input data file.
The Zmarc script is a simple copy of the standard run_marc script, where commands
needed to link automatically the Z-mat package to build a custom marc executable have been
added. Therefore, standard options of the run_marc procedure are also available with the
Zmarc command.
MSC-Marc Input:
The definition of a material for the Z-mat behaviors always uses an external file to establish
the model components and coefficients. Z-mat never uses the material parameters defined in
the Marc .dat file.
In the current implementation the definition of the Zmat behavior should be given in a
file named problem.zebulon.
Several additions to a standard Marc .dat file are necessary to activate the use of Zmat.
Some of them cannot be accessed by means of the Mentat graphical interface and should be
done directly in the .dat file. Those commands are listed hereafter:
• Parameters section
– The state vars command should be added to specify the number of material
variables needed by the Zmat behavior. The Zpreload utility can be used to
compute the number of variables required. In any case, Zmat outputs to the
.log file the correspondence between the Marc user state variables and Zmat the
behavior components. If not enough state variables are available for the behavior
the analysis is stopped. Otherwise a warning is printed if more variables than
necessary have been specified.
state vars,20,20,
– The creep procedure should be activated by means of the creep command, even if
no creep effects are involved in the analysis. This is needed by Marc, because the
uvscpl user subroutine can only be used within the creep algorithm.
creep,0,0,1
• Model definition section
Z-set — Non-linear material
& structure analysis suite
5.3
Zmarc interface
– The visco plas parameter of the isotropic or orthotropic command must be
selected, to activate the use of material integration by means of the uvscpl user
subroutine. Note that either isotropic or orthotropic materials can be selected
indifferently, since values of the material coefficients given as arguments of these
commands are ignored by Zmat, that reads the material definition in a separate
problem.zebulon file.
isotropic
,
1,visco plas,isotropic,0,0,0,0,material1,
2.E+4,3.E-1,8.E-6,2.E-5,0.0,0.0,0.0,0.0,
1 to 1000
– To benefit from the consistent tangent matrix calculated by most Zmat behaviors,
and accelerate the convergence of the global equilibrium iterations, the full newtonraphson algorithm should be selected by setting the 6th parameter of the control
command to 1.
control
1000,10,2,0,0,1,1,0,1,0,
0.001,0.,0.1E-8,0.,0.1,0.,0.1E-4,1.E-12,
– Output of material variables to the Marc results file
The post command allows to select which material variable should be stored in
the results files. In this case the element codes that should be given as argument
of the post command, is a negative integer value such as -varid, where varid is
the number of the user variable required for output.
post
14,16,17,0,0,19,20,0,1,0,
301
461
311
391
-30,,d1
-31,,d2
-32,,d3
The above command adds user variables 30, 31 and 32 to the .t16 result file.
The corresponding values will then be accessed by the graphical post-processor of
Mentat. Within Mentat the names given to those state variables will be respectively
d1, d2 and d3, as defined by the post command.
Note that this output capability makes use of the user subroutine plotv to interpret
correctly the user element codes. An appropriate plotv code is provided in the
Zmarc package. However, this means that the user cannot redefine this particular
user subroutine for its own purpose.
• History definition section
5.4
Z-set — Non-linear material
& structure analysis suite
Zmarc interface
The use of the auto step procedure for adaptative load step control is strongly advised.
Moreover, starting with Marc2003, auto load doens’t seem to handle anymore user material implemented with the uvscpl subroutine. Also, as described in the next section,
the enhanced scheme of this procedure can be used to provide Marc some feedback
about local integration results obtained in the Zmat library. This option should then
be activated as well by setting the 9th parameter of the second data block to 1.
auto step
0.02,0.4,,,0.0001,0.1,100,6,1,
,10,0,,,,,,,,,2,
Automatic time-stepping and Zmat local integration:
For complex material behaviors and/or large strain increments, non-convergence may occur during integration of the behavior equations by Zmat. Without a proper procedure, such
local divergence will cause either premature convergence (especially in the case of anisothermal analysis under purely thermal loadings), or run-away newton iterations until the specified
maximum number of recycles is reached and increment cutback eventually occurs.
Unfortunately there is nothing available in uvscpl (as in other user material subroutines)
to signal a local divergence back to Marc and force an immediate increment cutback.
However, the enhanced scheme of the auto step command, that allows to specify user
criteria to control the step-size, may be used to implement some kind of emergency procedure.
The input data necessary to implement this mechanism from the .dat file is not trivial, and
is described hereafter:
• activate the enhanced scheme of the auto step procedure by setting to 1 the 9th parameter of the second data block,
• choose to use user criteria as limits (versus targets) by setting to 0 the third parameter
of the third data block,
• add a criterion on a Zmat variable, that will cause increment cutback in case of its
violation during local integration. To select a particular user state variable, a code such
as 13x100 + state variable id should be used as first argument of the fourth data block.
auto step
0.02,0.4,,,0.0001,0.1,100,6,1,0
,10,0,,,1.2,,,,,,2,
1330,cmc
10.,0.01
In the above example, the user criterion used to monitor convergence is defined on user
state variable number 30 (element code 1330). The criterion defines that this variable at any
integration point of the element set named cmc, should not increase of a value of more than
10.0 (defined in the fifth data-block) over an increment. If the increase of the variable exceeds
the limit value specified, the auto step procedure will discard the load increment, and the
step size will be cut back accordingly.
Z-set — Non-linear material
& structure analysis suite
5.5
Zmarc interface
More precisely, denoting by ∆t the current step size, ∆v the maximum variation of variable
v found by local integration during the current increment cycle, and vmax the limit value
specified, if ∆v > vmax a new icrement will be attempted with a lower step size calculated as:
∆t
vmax
∆v
.
Additional commands may then be used in the Zmat problem.zebulon file, to control
increment cut-back in case of local divergence. Those commands are the following ones:
***divergence_variable 30 20.
The above command select variable 30 to monitor local convergence, and specify that in
the event of divergence a ∆v value of 20. should be sent back to Marc (ie. two times the vmax
limit value defined in the auto step command). Hence the effect will be to force a division
by 2 of the global step size in the event of local divergence.
When using this scheme, care must be taken to carefully select the vmax limit value.
Typically the value should exceed likely variations for the variable selected, and the
***divergence_variable increment value set accordingly to produce an increment cut-back
of the required size.
Example:
The following listing summarizes the options needed in a visco.dat Marc input file when
using the Zmarc interface.
title visco
$ parameters section
...
state var,19,19
creep,0,0,1
end
$ model definition section
...
isotropic
,
1,visco plas,isotropic,0,0,0,0,material1,
20000.,0.3,8.E-6,2.E-5,0.0,0.0,0.0,0.0,
all_element
...
post
14,16,17,0,0,19,20,0,1,0,
301,
311,
-7,,evcum
...
control
100,10,2,0,0,1,1,0,1,0,
0.001,0.,0.1E-8,0.,0.1,0.,0.1E-4,1.E-12,
5.6
Z-set — Non-linear material
& structure analysis suite
Zmarc interface
$ history definition section
auto step
0.02,0.4,,,0.0001,0.1,100,6,1,
,10,0,,,,,,,,,2,
1308,all_element
10.,0.01
...
The Zmat behavior is defined in a separate visco.zebulon file included hereafter. The
cooresponding model is a typical Chaboche viscoplasticity model. Note that the value of the
young’s modulus will indeed be 150000. as defined in the Zmat file, and that the value of
20000. given after the isotropic command of the .dat file has no impact whatsoever on the
results.
***material
*integration theta_method_a 1. 1.e-9 100
***divergence_variable 8 20.
***behavior gen_evp
**elasticity
young 150000.
poisson 0.3
**potential gen_evp ev
*criterion mises
*flow norton
n 7.
K 1200.
*kinematic nonlinear
C 126000.
D 380.
*isotropic constant
R0 10.
***return
Using the Zpreload utility on the above material file would produce the following output:
$ Zpreload visco.zebulon
Reading behavior in file: visco.zebulon
============================================
Flux Name:
sig11 sig22 sig33 sig12 sig23
sig31
Grad Name:
eto11 eto22
eto31
Z-set — Non-linear material
& structure analysis suite
eto33
eto12
eto23
5.7
Zmarc interface
var_int Name:
eel11(sdv1) eel22(sdv2) eel33(sdv3) eel12(sdv4) eel23(sdv5)
eel31(sdv6)
evcum(sdv7)
al111(sdv8) al122(sdv9) al133(sdv10) al112(sdv11) al123(sdv12)
al131(sdv13)
var_aux Name:
evi11(sdv14) evi22(sdv15) evi33(sdv16) evi12(sdv17) evi23(sdv18)
evi31(sdv19)
============================================
done with material file reading...
Temperature not needed...
...
This allows to select the number of state variables needed by the behavior (19 variables in
the example) that should be given as argument of the state var command. State variable
number 7 is the cumulated plastic strain (named evcum) for this particular Zmat behavior.
This material variable will be stored in the Marc results file using the appropriate post
elem var code (-7 in this case). This variable is also used to monitor local divergence (1308
code to define the user criterion used in the auto step procedure), and a corresponding
***divergence_variable is included in the Zmat file to control the increment cut-back in
case of local divergence. Note that Marc always use the first state variable to store the
temperature, such that the state variable id given by Zpreload must be incremented by one
to select the proper material variable.
Finally, note that this shift in the state variable id, is automatically taken into account
for the post command (the shift is done in the plotv user subroutine included in the Zmarc
release).
5.8
Z-set — Non-linear material
& structure analysis suite
Chapter 6
Z-mat SAMCEF
Z-set — Non-linear material
& structure analysis suite
6.1
6.2
Z-set — Non-linear material
& structure analysis suite
Zsamcef interface
Z-mat SAMCEF interface
Description:
The command described in this chapter allows to use a Z-mat behavior within a SAMCEFMECANO analysis. The interface makes use of the OVMAXX user subroutine that allows to
implement user-defined behavior within SAMCEF.
Syntax:
% Zsamcef problem ←where problem.dat is the name of a SAMCEF input data file.
The Zsamcef script activates a non-standard MECANO executable with name:
$Z7PATH/Zsamcef/mecano_zmat_$Z7MACHINE
Please verify that this file is indeed included in your Zset distribution.
A new user module (module "mecano_zmat" associated to module Id "zm") should be
declared in the SAMCEF environment by means of the samrc.ini configuration file.
The samrc.ini file in the user home directory is automatically updated by the Zsamcef
script with the command required to declare this new user module:
module*zm.me: mecano_zmat $Z7PATH/Zsamcef/mecano_zmat_$Z7MACHINE
Note that a MECANO calculation with a Z-mat behavior can alternatively be launched
by the following standard SAMCEF command, that chains the bacon mesher with the
mecano_zmat user module:
% samcef ba,zm problem n 1 ←MECANO Input:
The main Z-mat modification needed in a standard MECANO input file concerns the .MAT
command used to define material properties.
Syntax is the following, where the BEHA parameter that usually allows to specify the name
of a standard MECANO material behavior is replaced by a ROUTIN parameter followed by the
name of the Zmat material file :
.MAT NOM "MATERIAU"
ROUTIN "zmat_fname"
!
Elastic parameters definition is needed
!
but values are not meaningful
YT 10.
NT 0.3
!
For anisothermal problems thermal expansion
!
coef A should be set to 0. and defined
!
in the Z-mat file by means of a **thermal_strain object
A 0.
where zmat_fname is the name of a Z-mat material file. Note that:
Z-set — Non-linear material
& structure analysis suite
6.3
Zsamcef interface
• definition of some elastic properties (coefficients YT and NT) is required in the SAMCEF
input file. However, values given for those coefficients have no incidence on the results,
since the actual definition of the elasticity coefficients is given in the Z-mat material file
• for anisothermal problems the value of the thermal expansion coefficient needed to
calculate thermal strains should be given in the Z-mat file, and it is safer to set the A
coefficient to zero in the SAMCEF input file.
Z-mat interface file:
Most of the commands are common to the various Z-mat interfaces and are described in
the Z-mat interface file section (page 2.5). However, some commands are not supported by
the SAMCEF interface, while others are specific to this port. Hence, the various commands
allowed in the Zsamcef interface file are summarized hereafter. Only options that are indeed
specific to SAMCEF will be described in detail.
Syntax:
[ ***debug ]
[
*local_debug ip ]
[ ***automatic_time ]
[
*limit name1 vmax1 ]
...
[
*limit namei vmaxi ]
[ ***save_tensor tname ]
[ ***save_scalar sname ]
[ ***needs_temperature ]
***material
[
*file fname ]
[
*integration ...
]
[
*rotation ...
]
[
*initialize_variable ... ]
[
*dim dim ]
***behavior
...
***return
***debug This command indicates that debug output will be generated during MECANO
execution. Output will be stored in a file named ”fname.msg”, where ”fname” is the
name of the Z-mat interface file.
*local debug ip
This subcommand restricts debug output to the integration point number ip given as argument. ip is an integration point counter managed internally by Zsamcef, incremented
during the loop on the elements and resetted to zero at the beginning of each newton
global iteration. Note that the OVMAXX SAMCEF user subroutine doesn’t provide any
information on the element/integration point number that could allow a more meaningful selection mechanism. By default debug output will be generated for all integration
6.4
Z-set — Non-linear material
& structure analysis suite
Zsamcef interface
point of the FE mesh, which will result in huge ouput files and may dramatically slow
down the calculation.
***automatic time This command may be used in conjunction with the SAMCEF automatic time step procedure based upon material integration error. To enable this option
the following parameters must be given as arguments of the .SUB command in the
SAMCEF input file:
.SUB ...
VISC 1 ! activates material automatic time step
PRCV 1. ! material allowable error. note that default value
! of 0.1 is not compatible with the Z-mat mechanism
SREF 1. ! material ref norm (default value)
...
*limit namei vmaxi
This command indicates that, during a loading increment, the increase of the namei
material variable should not be greater than a given value of vmaxi. Several *limit
commands acting on different material variables may be added if necessary. The material
error returned by Z-mat to SAMCEF will be calculated as follows:
ERRLOC = max
i
∆vi
vmaxi
where ∆vi is the increase of variable number i, and the max is taken over all material
variables vi specified by a *limit command. This ERRLOC value will then be compared
by SAMCEF to the PRCV parameter of the .SUB command in order to estimate the
appropriate time step size. Note that PRCV should always be set to 1.0 in this particular
context.
***save tensor tname
***save scalar sname
Output of user-behavior material variables to the SAMCEF results files is restricted to
one tensor and/or scalar only. Corresponding FAC codes are 1399 and 1499 that should
be requested by an appropriate .SAI command in the SAMCEF input file:
.SAI ...
ARCHIVE ALL_ELEMENTS STYPE 1399 ! user scalar
1499 ! user tensor
...
The ***save_tensor and ***save_scalar commands then allow to specify which Zmat variable will be saved in the results files.
***needs temperature Z-mat automatically detects that the temperature should be stored
in the state variables when a material coefficient dependence on this parameter is defined in the behavior. However, if all material coefficients are constant, temperature is
Z-set — Non-linear material
& structure analysis suite
6.5
Zsamcef interface
removed from the state vars management to cut down storage requirements. In this case
the thermal strains calculated by the **thermal_strain object will always be zero.
The ***needs_temperature command may then be used to force storage of the temperature and allow thermal strain calculation even if no coefficients are temperaturedependent.
***material As decribed in the Interface file section (page 2.12), this bloc of commands
is used to set integration methods parameter (command *integration), define local
axis for the calculation of material quantities (command *rotation), initialize material
variables, etc... A different file may also be specified that will contain the actual behavior
definition (***behavior commands) by using the *file command. This allows to
separate the interface commands, that may be specific to the FEA code, and the material
coefficients. Default is to look for the ***behavior definition in the interface file.
*dimension dim
Second order tensors passed in by SAMCEF as arguments of the user-material subroutine OVMAXX always have 6 components (storage of a 3D symmetric second order
tensor in a vector), regardless of the problem dimension. Theorically, the particular
kinematic hypothesis used in the calculation (3D, axisymetric, plane strain, plane stress
...), can be known at the level of the OVMAXX routine by means of the IHYP argument.
Unfortunately, in the current SAMCEF version (SAMCEF v10.1), this argument is not
correctly initialized during the loading phase of the Z-mat behavior. Therefore, by
default, all Z-mat objects will be initialized with a 3D size. A *dimension 2 command
may then be used to bypass this problem and cut-down material state variables storage
requirements for 2D problems.
Note also that for the same reason, it is currently necessary to add an explicit
plane_stress modifier to the behavior, or a **plane_stress switch in the gen_evp
assembly, in order to properly take into account the plane stress hypothesis.
6.6
Z-set — Non-linear material
& structure analysis suite
Chapter 7
Z-mat Cosmos
Z-set — Non-linear material
& structure analysis suite
7.1
7.2
Z-set — Non-linear material
& structure analysis suite
Zcosmos interface
Z-mat Cosmos/M interface
Description:
Currently we are not providing any implicit interface for Z-mat with Cosmos. Users are
encouraged to contact the distributor if they require a separate interface for Cosmos. Unfortunately demand for this port has been extremely limited.
Like many of the older style interfaces, the user interface for Cosmos/M required compilation of a custom executable and therefore requires specifically the compiler and development
environment recommended by SRAC. The build scripts are still supplied with Z-mat distributions, and it is very likely a user could modify that for current versions of Cosmos. The
last supported Cosmos release was 2.8.
Z-set — Non-linear material
& structure analysis suite
7.3
Zcosmos interface
7.4
Z-set — Non-linear material
& structure analysis suite
Chapter 8
Z-mat LS-Dyna
Z-set — Non-linear material
& structure analysis suite
8.1
8.2
Z-set — Non-linear material
& structure analysis suite
Zlsdyna interface
Zlsdyna LS-Dyna interface
Description:
The Zlsdyna port applies to the user material facility within the explicit dynamics code LSDyna. Implicit integration modes are not supported for this interface, so the port is strictly
explicit and therefore somewhat different from the other codes.
The user is referred to the LS-Dyna documentation sections under the *MAT chapter of
the user commands manual for topic *MAT USER DEFINED MATERIAL MODELS, and also to information included in the Appendix A of the 970 user manual.
Syntax:
% Zlsdyna [ opts ] problem ←Compatibility:
The Zlsdyna interface is tested at the time of writing with LS Dyna 970 which requires building
a custom executable from the development kit. The interface has been tested with both single
process and MPP versions of LS-Dyna. Unfortunately because of licensing restrictions NW
Numerics is unable to re-distribute modified binaries of LS-Dyna, so the compilation obligation
falls on the end user. In some cases NW Numerics has been able to assist in this compilation
process, so please inquire with the distributor.
This procedure imposes the following very strict requirements:
• Precisely the same compiler as defined by LSTC must be available and used for the
build process.
• The development kit must be obtained from LSTC. There are different development
kits for the single and MPI based distributed domain solver.
• Check with NW Numerics on our testing level with the platform chosen. Because of the
many different computer/distribution levels possible there is likely some configuration
work to be done.
Note: With LS Dyna 971 MPP versions there will be a more robust “plug-in” style interface will be implemented similar to ABAQUS and ANSYS. In that case the Zlsdyna installation
and version tracking will become a negligible effort. All aspects of the user launch process
will remain the same however.
Getting started:
Perform a standard installation of Z-set, including the binaries and shared files. The install
location will henceforth be referred to as the Z7PATH. In the user-compiled case the LS-Dyna
executable will be located in a common library directory. With the shared library method
used with 971 it will be necessary to specify the location of the LS-Dyna executable via the
Z7 DYNA ROOT environment variable.
There are some test cases in the Zlsdyna/ directories in the validation test database.
For LS-Dyna the Z-mat interface file name is standardized to be umat41 or umat42. Note that
because of the need to have an automated testing environment with all the LS-Dyna validation
cases in the same directory, the testing program copies a Z-mat interface file prob.zmat to
Z-set — Non-linear material
& structure analysis suite
8.3
Zlsdyna interface
umat41 where prob is the basename of the prob.k input before running Zlsdyna. In fact this
behavior is a convenienace option of the Zlsdyna command, with the -auto switch. One can
equivalently set the environment variable Z7 DYNA AUTOCOPY to be equal to yes.
Input file change:
Like all the other Z-mat interfaces, there are some lines in the user’s input deck (the .k file)
which must be changed to indicate that a user-material is being used, and to specify the
amount of state variable storage. The following is an example:
*MAT_USER_DEFINED_MATERIAL_MODELS
$
mid
ro
mt
1
8.930
41
$
ivect
ifail ithermal
0
0
0
$
p1
p2
1.3
0.433
lmc
2
nhv
40
iorth
0
ibulk
1
ig
2
The mt data entry specifies the user routine to use, and the nhv entry specifies the number of
state variables required.
Zmaster interface:
The d3plot results files can be read via Zmaster directly as an alternate choice for post
processing, though the standard LS-Dyna prepost is quite a nice environment and there may
not be so much reason to do this. Because there is no default suffix an added switch -d3d
is needed to indicate an LS Dyna file. When a numbered series of d3plot files is to be read
only the 1st file (unnumbered) should be specified on the command line.
Zmaster -d3d d3plot
Either the Zmaster Mesh, Plot, or Results buttons can be used.
Zpost interface:
Probably more interesting than running in Zmaster for commercial users is the ability to do
post processing or results file translations. The following is an example for extracting time
displacement data at a node in a batch task (execute with Zrun -pp ):
****post_processing
***data_source d3plot
**open d3plot
***global_post_processing
**file node
**output_number 1-999
**nset ALL_NODE
**process curve control_energy.bar-impact.test
*precision 3
*node
1333 U3
****return
8.4
Z-set — Non-linear material
& structure analysis suite
Chapter 9
Behavior functionality
Z-set — Non-linear material
& structure analysis suite
9.1
9.2
Z-set — Non-linear material
& structure analysis suite
Introduction to Behaviors
Introduction to Behaviors
Material behaviors in Z-mat use by far the most dynamic and object oriented input of all the
Z-set modules. This is because behaviors are a set of constitutive equations which get built
from fundamental “building bricks” of sub-models. There is widespread re-use of all these
“bricks” between different behavior models. It is therefore necessary to adjust our input
syntax (or at least documentation of it) to allow for more flexibility.
Classes:
The most important concept for this chapter is the notion of a class of permissible options,
of which the user will choose one or more objects from that class in the material definition.
A class is an abstract notion of basic functionality, and will henceforth be denoted by
the following convention: EXAMPLE CLASS indicating that an object of type EXAMPLE CLASS is
permitted. In the handbook, one will find another section with heading EXAMPLE CLASS which
describes the different types allowable for the class.
An example is an ELASTICITY class which handles the calculation of stress from a strain,
usually employing a 4th order tensor of various modulus coefficients. Because of this very
general function, many behavior models allow for an ELASTICITY instance. In turn, the
elasticity class has many possible types which one can select to fill in an elasticity object
(isotropic, orthotropic, etc).
**elasticity <ELASTICITY>
The keyword **elasticity used here is in fact specific to the behavior model. While we
frequently see the same keyword indicating classes between behaviors, this is not necessarily so.
The function of each sub-command will be described in the discussion of the behavior model
itself. What defines the syntax which follows this line is the ELASTICITY notation, indicating
an elasticity matrix should be input. Again, the use should then go to the ELASTICITY section
of the handbook, and investigate the possible models for this option.
Many behavior models also include the possibility of multiple instances of certain keywords. These possibilities will be described in the section of the containing class (behavior is
a “material piece” class, but there are many others which can contain sub-pieces themselves).
The standard notation for multiple instances of a class is to include three dots following the
data entry line:
*kinematic <KINEMATIC>
...
This means that the material at this location can take any number of *kinematic entries, in
any order, and with no restriction on type.
Example:
An example file follows, which is of a simple elasto-viscoplastic model with non-linear isotropic
hardening, and two kinematic hardening components. Note that the behavior has taken
two objects, an elasticity matrix and a “potential” (dissipation potential with an inelastic
deformation associated to it). The material behavior is gen_evp which stands for generalized
Z-set — Non-linear material
& structure analysis suite
9.3
Introduction to Behaviors
elasto-viscoplastic (see page 10.13). The potential has in turn taken on a number of subobjects: criterion, flow, kinematics and an isotropic hardening. This behavior actually allows
for more than one **potential instance, so very complex behaviors are possible.
***behavior gen_evp
**elasticity isotropic
young 260000.
poisson 0.3
**potential associated ev
*criterion mises
*flow norton
n 7.0
K 400.
*kinematic linear
C 15000.0
*kinematic nonlinear
C 6000.0
D 100.0
*isotropic nonlinear
R0 130.0
Q 20.0
b 500.0
***return
Another important thing to notice here is the coefficient C is entered after each *kinematic
entry. These coefficients are distinct in the behavior because they belong strictly to the
kinematic objects. This totally illuminates the possible conflicts inherent with hard-coding
the model to have for example C1 and C2.
9.4
Z-set — Non-linear material
& structure analysis suite
Grad-Flux
Grad-Flux
The compatibility of material behaviors with the element is determined dynamically based
on their gradient (or primal) and flux (dual) variables. These can also be thought of as the
material input-output combination. These variables are also observable state variables and
observable associated forces. Some of the primal-dual variable combinations in Z-mat are as
follows1 :
eto - sig
DESCRIPTION
small deformation –σ
F - sig
updated Lagrangian large-strain F–σ
dT - q
thermal analysis
dC - J
diffusion analysis
CODE
The process relative to element integration is shown schematically below. Non-linear finite
element analysis consists of a loop over all elements in order to fabricate the global DOF
residual vector (and possibly fabricate the stiffness if needed). In turn, there is a loop within
the element over the integration points in order to numerically integrate the elements volume
integrals. At each integration point, the new increment in primal variable is calculated using
the increment of degrees of freedom, and their derivatives (this is where grad comes from),
and this is passed to the material behavior with the current value of state variables.
MAT_DATA
Gradient
Delta gradient
grad
delta_grad
Internal variables
IV
Auxiliary variables
AV
External parameters
EP
BEHAVIOR
integrate()
Flux
Tangent matrix
1
Z-mat used for other codes such as ABAQUS establish a similar relationship, where the UMAT interface
adjusts itself according to the material primal-dual couple
Z-set — Non-linear material
& structure analysis suite
9.5
Grad-Flux
9.6
Z-set — Non-linear material
& structure analysis suite
Material variables
Material variables
As introduced in the previous section, material behaviors in Z-mat are dynamically “created”
by the user through the assemblage of material sub-objects. The example given of a gen_evp
behavior on page 9.4 was certainly structured this way. Of interest to the user is what variables
are contained in the model, and what is available for output 2 .
Principally all material objects (behaviors and their sub-component classes) have the possibility of the following variables:
grad is the gradient or primal observable input variable.
flux
is the flux or dual output variable. Normally thermodynamically conjugate with the
grad.
var-int integrated state variables. These define the current state of the material, and are
the subject of the integration method. Normally the more var-int variables, the higher
the cost of local integration.
var-aux auxiliary variables; normally used for output, or maintaining state information on
a total basis (secondary to var-int).
ext-param External parameters. These are imposed using ***parameter statements in
Zebulon, or with field variables in other codes.
coefs material coefficients. Can depend on any of the above! Note however that the
integration method or implementation restrictions can limit such dependence.
Each variable is assigned a name, but unlike other codes where the relationship between
model options is hard-coded, the naming scheme is not known a priori. Much of the final
naming is up to the user. The example on the next page demonstrates how the names are
constructed.
Variable attributes:
When asking for specific output of the material variables, some additional attributes are
available. These are summarized as follows:
scal::fabs
DESCRIPTION
absolute value of scalar
tens::mises
von Mises equivalent of tensor
tens::trace
Trace of tensor
tens::p1
Principal eigen values of tensor; also ::p2 and ::p3
vec::eq
Equivalent of vector
CODE
2
Because of the dynamic nature of the object construction, it is often difficult to strictly define the names
of all the stored variables ahead of time. Users are thus strongly advised to observe the stored variables by
using the -v command line switch or **verbose output option.
Z-set — Non-linear material
& structure analysis suite
9.7
Material variables
An example of accessing these secondary variables in a Zebulon output statement follows:
***output
**component sig::mises sig::p1 sig::p2 sig::p3
sig11 sig22 sig33 sig12
eto11 eto22 eto33 eto12
Example:
Taking the gen_evp material file example given previously (page 9.4), and running it with
verbose set (on by default in Z-mat for ABAQUS, or with a -v switch running in Zebulon or
Z-sim) gives the following output:
============================================
Flux Name:
sig11 sig22 sig33 sig12
Grad Name:
eto11 eto22 eto33 eto12
var_int Name:
eel11 eel22 eel33 eel12
evcum
al111 al122 al133 al112
al211 al222 al233 al212
var_aux Name:
evi11 evi22 evi33 evi12
Default Output:
eto11 eto22 eto33 eto12
sig11 sig22 sig33 sig12
evcum
evi11 evi22 evi33 evi12
============================================
In fact what happens here is the behavior is put together by the different “bricks,”
with each brick naming its own variables. Note that the potential line input was
**potential associated ev and that the last key ev was used as a pre-fix for that potentials variables. Thus one has evcum for the plasticity (viscoplasticity in this case) multiplier,
and evi## as components of the inelastic strain tensor. The variable construction in this
model can be represented as shown below:
<BEHAVIOR>
(gen_evp)
VI:
VA:
FL:
GR:
eel
sig
eto
<POTENTIAL>
VI:
VA:
FL:
GR:
<FLOW>
evcum
evi
<ISOTROPIC>
<KINEMATIC>
al1
VI:
Potential was
named ev
9.8
<KINEMATIC>
al2
VI:
Z-set — Non-linear material
& structure analysis suite
Material variables
Naming conventions:
The dynamic nature of the material assembly notwithstanding, some conventions are given in
the documentation and reflected in the default naming of variables. A small summary table
of some common names follows.
CODE
eto
DESCRIPTION
total strain
ETO
“material strain” for finite strain. This is the integration
of the corotational strain measure – not the logarithmic
strain, etc.
sig
Cauchy stress
eel
elastic strain
eth
thermal strain
evcum
cumulated, monotonically increasing scalar measure of viscoplastic strain. Actually the integration of the viscoplastic
multiplier.
epcum
cumulated plasticity equivalent of evcum. The naming difference is usually only symbolic.
f
porosity for porous materials.
Z-set — Non-linear material
& structure analysis suite
9.9
Material variables
9.10
Z-set — Non-linear material
& structure analysis suite
Material file
Material file
The material and global-scope coefficients to be used are defined in a separate file henceforth described as the material file. The material file is a standard ASCII text file of form
similar to the main input file. Most instances where a material filename needs to be specified
allows an optional integer value for the instance of behavior in that file, the default instance
being the first. So, when a material file is opened, the program will search for the n-th occurrence of the keyword ***behavior, skipping all other data contained before. Frequently for
example and verification problems, the material file is given in the same physical file as the
referring input file for compactness and file management purposes3 .
Regardless of how or where the material file is located, the material file is a separate
entity from the other input commands, with an independent command hierarchy starting at
the ***-level. The general structure of this file is the following:
***behavior BEHAVIOR [ modifiers ]
**functions
list of function declarations
...
**behavior sub-procedures
...
**coefficient
coefficient name COEFFICIENT
...
**save_coefficients
coef-names
**plane_stress
***return
***behavior begins the definition of a material law. The options which follow are of course
specific to each material model.
**coefficient indicates the definition of intrinsic coefficients which are not specific to material laws4 . These coefficients are applicable to all the material models. The allowable
global coefficients are summarized below:
masvol
DESCRIPTION
volumetric mass of the material
capacity
volumetric heat capacity(ρCp )
CODE
3
Note that we frequently find this the advisable approach for production runs as well, as it avoids confusion
or possible version conflicts when numerous runs are to be made.
4
Older definitions with the ***coefficient command outside the ***behavior and ***return commands
are no longer allowed.
Z-set — Non-linear material
& structure analysis suite
9.11
Material file
Units should be coherent within the problem definition. For example, if using SI system,
masvol is in kg · m−3 and capacity in J · m−3 · K −1 , but if using the (mm,MP a,s) unit
system, masvol is in ton · mm−3 and capacity in J · mm−3 · K −1 .
**plane stress indicates that a plane stress behavior is to be used (σ33 = 0). This is not
available with all material behaviors. Zébulon plane stress must not use this option.
**save coefficients list coefficient names which are to be saved as output variables. New
var-aux variables are created, and therefore increase the total storage for each material
integration point. This option is extremely useful when the variation of coefficients is
important, such as in coupled problems.
Example:
Another example of material file input follows, showing some tabular coefficient input:
***behavior linear_elastic
**elasticity isotropic
young temperature humidity
200000.e0 100.
0.
100000.0
"
1.
100000.0
200.
0.
50000.0
"
1.
poisson 0.3
**thermal_strain
alpha temperature
1.e-6 0.
1.e-6 1000.
ref_temperature 0.0
**coefficient
masvol 7.e-9
***return
9.12
Z-set — Non-linear material
& structure analysis suite
<BEHAVIOR>
<BEHAVIOR>
Description:
This class of objects provides the basic building block for material models. Each object type
tries to cover as broad a range of behavior as possible, using the idea of sub-model objects to
increase the possible combinations.
Syntax:
***behavior BEHAVIOR [modifier]
**-level commands
**coefficient
coefficient list
***lagrange_modifier type
We have broken the behavior models up into three classifications reflecting the following
three chapters of this book. The first is for models which make up significant frameworks
for treating broad ranges of characteristics, the details of which are fixed by selecting from
a broad range of options and sub-components. The second group of models are “secondary
models” in that they are coded in a specific fashion, usually as a prototype stage on their
way to being incorporated into the general class materials. The third “Other” classification is
for material models fitting a particular application which in general is not suitable for Z-mat
interfaces with standard mechanical codes (e.g. debonding or spring behaviors specific to
Zebulon).
Finite Strain:
The behaviors are normally defined using small strain assumptions. These models may be
transformed to finite deformations / rotations using one of the behavior modifiers described on
page 14.3. The hyper elastic behavior models are however formulated specially with total
Lagrangian assumptions and must therefore be used with the appropriate total Lagrangian
elements.
Z-set — Non-linear material
& structure analysis suite
9.13
<BEHAVIOR>
9.14
Z-set — Non-linear material
& structure analysis suite
Chapter 10
Material Models
Z-set — Non-linear material
& structure analysis suite
10.1
10.2
Z-set — Non-linear material
& structure analysis suite
***behavior linear elastic
***behavior linear elastic
Description:
This behavior class provides classical linear elastic behavior with optional thermal deformation. The behavior understands two “blocks” from which it is defined.
Syntax:
***behavior linear_elastic [ modifier ]
**elasticity ELASTICTY
[ **thermal_strain THERMAL STRAIN ]
By default there is no thermal dilatation in the model.
Example:
Two examples of linear elastic behavior follow:
***behavior linear_elastic
**elasticity isotropic
young 200000.
poisson 0.30
***return
***behavior linear_elastic
**thermal_strain anisotropic
codila1 1.0e-06
codila2 2.0e-06
codila3 3.0e-06
**elasticity orthotropic
y1111 350000.
y2222 280000.
y3333 280000.
y1122 150000.
y2233 120000.
y3311 150000.
y1212 180000.
y2323 180000.
y3131 180000.
***return
Z-set — Non-linear material
& structure analysis suite
10.3
***behavior damage elastic
***behavior damage elasticity
Description:
This is a simple behavior for elastic energy based damage. It is duplicated by the gen_evp
**damage option with type *elastic. This behavior is given as an example of a simple user
model programmed without ZebFront. The source is available in the developer handbook.
σ = (1 − D)Del : (to − zeth)
1
Ȳ = el : Del : el
2
h
p
p i
D = α max Ȳ − Y0
Syntax:
***behavior damage_elasticity
**elasticity <ELASTICITY>
**Y0
COEFFICIENT
**alpha COEFFICIENT
Stored Variables:
The grad variable is the gradient of temperature, and the flux is the heat flux.
prefix
size
description
default
eto
sig
y max
damage
T-2
T-2
S
S
total (small deformation) strain
Cauchy stress
Ȳ
damage D
yes
yes
no
yes
Example:
***behavior damage_elasticity
**elasticity isotropic
young 200000.
poisson 0.0
**Y0
20.0
**alpha 0.001
***return
10.4
Z-set — Non-linear material
& structure analysis suite
***behavior hyper elastic
***behavior hyper elastic
Description:
This behavior handles all (simply) hyperelastic material laws. For cases of hyper-viscoelastic
or hyper-viscoplastic models please see those corresponding behaviors. Note that the syntax
for hyperelasticity in Z-mat has changed significantly with the 8.3 release, and the (un-mixed)
hyperelastic laws are now available for use with the different Z-mat interfaces. All the previous
individual behaviors for each potential form have been deprecated, though a compatibility
syntax is still provided.
The hyperelastic potential is defined by a material component for this task, which is
shared between the other laws employing such a potential. There are several main classes of
hyperelastic “rules” which can be inserted in this behavior depending on their assumptions
and integrated rules. The following type are allowed:
• default HYPERELASTIC LAW objects.
• isotropic HYPERELASTIC LAW objects which are somewhat more specific and where certain properties of the potential tangent can be made.
• mixed hyperelastic models which are necessary for use with the Zebulon hyperelastic
element, but cannot be used with Z-mat interfaces. These will be deprecated in the
next version when the incompressible element is changed.
Syntax:
***behavior hyper_elastic
[ **thermal_strain <THERMAL_STRAIN> ]
[ **hyperelasticity <HYPERELASTIC_LAW> ]
[ **mixed_hyperelasticity <MIXED_HYPERELASTIC_LAW> ]
[ **isotropic_hyperelasticity <ISOTROPIC_HYPERELASTIC_LAW> ]
[ **model_coefficients ]
...
Example:
***behavior hyper_elastic
**hyperelasticity arruda_boyce
**model_coefficients
mu
0.893
lambda 9.0
d
0.1
***return
Z-set — Non-linear material
& structure analysis suite
10.5
***behavior linear viscoel
***behavior linear viscoelastic
Description:
This behavior defines a generalized linear viscoelastic Maxwell model. The model defines the
stress, σ, to the strain by the following relation:
Z t
Z t
σ(t) =
2G(t − τ )ė(τ ) dτ + 1
K(t − τ ) Trace ˙ dτ
0
0
with e the deviator of the strain tensor . The terms G and K are relaxation functions
defined by Prony series:
P α
= G∞ − (G∞ − G0 )Ψ1 (τ ) Ψ1 (τ ) = i=n
ωi exp(−τ /τi )
Pi=1
i=nβ
= K∞ − (K∞ − K0 )Ψ2 (τ ) Ψ2 (τ ) = i=1 ωi exp(−τ /τi )
G(τ )
K(τ )
(1)
G0 and G∞ are shear modulus coefficients
K0 and K∞ are bulk modulus coefficients
Remark:
The sum of the coefficients ω for modulus terms must equal one.
The implementation of the model is in differential form with the internal variables α and
β. The state equations are written in the following form:
σ=
i=n
Xα
i=1
i=nβ
Xi +
X
Yi + 2G∞ e + K∞ Trace 1
i=1
with
Yi
Xi = −2(G∞ − G0 )ωi (e − αi ) 1 ≤ i ≤ nα
= −3(K∞ − K0 )ωi (Trace()/3 − βi )1 1 ≤ i ≤ nβ
(2)
The evolution equations for the internal variables are:
β̇
α̇i = τ1i (e − αi ) 1 ≤ i ≤ nα
= τ1i (Trace ∼/3 − βi ) 1 ≤ i ≤ nβ
(3)
It is necessary to define a single time the coefficients K0 , K∞ , G0 and G∞ using the key
words **K0, **K_inf, **G0 and **G_inf respectively. One may then define an arbitrary
number of the variables α and β using the key words omega and tau. omega is a constant
coefficient.
Syntax:
The syntax for this behavior model is the following:
10.6
Z-set — Non-linear material
& structure analysis suite
***behavior linear viscoel
**K0
**K_inf
**G0
**G_inf
**shear
**volumic
COEFFICIENT
COEFFICIENT
COEFFICIENT
COEFFICIENT
...
...
Options **shear represent the shear mechanisms and are defined by the following form:
**shear
tau COEFFICIENT
omega COEFFICIENT
where the coefficient tau corresponds to the material constant τ and omega to ω as in
equation 1.
Similarly, the option *volumic represents the volumetric mechanisms using the same
coefficient / material constant naming:
**volumic
tau COEFFICIENT
omega COEFFICIENT
Stored Variables:
The internal variables stored for this model are the total strain (code etoxx), the tensorial
variables α
(code alpha#xx) and the βi variables (code beta#).
∼i
prefix
size
description
default
eto
sig
alpha#
beta#
T-2
T-2
T-2
S
total (small deformation) strain
Cauchy stress
α variable
β variable
yes
yes
no
no
The code names will replace the # symbol with the sequential number of that variable type
as given by the order of declaration, and the xx symbols will be replaced with the tensorial
components. The default saving of variables in the output files are only eto and sig. Specify
other saved variables in the .inp file.
Example:
***behavior linear_viscoelastic
**K0
42261.904761
**K_inf 13500.0
**G0
29098.360655
**G_inf 0.
**shear
tau
0.4321660
omega 0.2324006
**shear
tau
9.070154
Z-set — Non-linear material
& structure analysis suite
10.7
***behavior linear viscoel
omega 0.1891879
**shear
tau
27.61690
omega 0.2665674
**shear
tau
102.8596
omega 0.3118441
**volumic
tau 0.1000000E-01
omega 0.29800651
**volumic
tau
0.3096638
omega 0.8500050E-01
**volumic
tau
0.2696395
omega 0.4522469E-01
**volumic
tau
6.517014
omega 0.5717688
***return
10.8
Z-set — Non-linear material
& structure analysis suite
***behavior viscoelastic s
***behavior viscoelastic spectral
Description:
This behavior defines a spectral viscoelastic model. The model defines the stress, σ, to the
strain by the following differential equations system:
σ = C0 (T ) : ( − a − th )
th = α(T − T0 )
˙a = g(σ)
i=n
Xt
Ẋi
i=1
Ẋi =
1
−1
(µi (T ) ∗ g(σ)CR
(T ) : σ − Xi )
τi
C0 is the elastic matrix. Note that S0 used latter is defined by : S0 = C0−1 .
g(σ) is the viscous non linear function
τi and µi define the spectrum. Each mechanism Xi is associated with a relaxation time τi
weighted by µi . The whole strain family (Xi ) describes a gaussian continuous spectrum.
Note:
For the moment, the thermal strains are not taken into account.
Syntax:
The material file structure for the viscoelastic spectral model consists of an elasticity
object, the definition of spectrum, of the viscous effects tensor and of the assymptote. The
syntax for this behavior model is the following:
***behavior viscoelastic_spectral [ modifier ]
**elasticity
<ELASTICITY>
**spectrum
...
**viscous_effects
...
**reversible_asymptote ...
Options **spectrum represent the spectrum which defined µi and τi (see next figure).
µi =
1
i − nc 2
exp(−(
) )
no ∗ π
n0
The spectrum is defined by the following form:
Z-set — Non-linear material
& structure analysis suite
10.9
***behavior viscoelastic s
µ
n0
ln( τ )
**spectrum
[*limit double]
[*n1 double]
[*n2 double]
*nt interger
*nc COEFFICIENT
*n0 COEFFICIENT
See the figure for the meaning of nt, nc and n0. n1 and n2 can be read in the material file
(*n1 and *n2) or calculated such as µ(n1) > limit and µ(n2) < limit.
Note:
By default limit is equal to 1.e-4 and nt to 30.
The option *viscous effects defined the viscous effects tensor CR .
**viscous_effects
*lrt COEFFICIENT
*lrc COEFFICIENT
CR is defined such as :
for i equal 1 to 3 :
CR (i, i) = lrt ∗ S0 (i, i)
for i equal 4 to 6 :
CR (i, i) = lrc ∗ S0 (i, i)
and if i6=j for i and j equal 1 to 3 (isotropic condition):
CR (i, j) = lrn ∗ S0 (i, j)
with lrn = (lrt ∗ S0 (1, 1) − lrc ∗ S0 (2, 2))/S0 (1, 2).
Finally the option *reversible asymptote is used to define the non linear funtion g(σ).
10.10
Z-set — Non-linear material
& structure analysis suite
***behavior viscoelastic s
**reversible_asymptote
*beta COEFFICIENT/
*p COEFFICIENT/
p
−1
g(σ) = 1 + beta( (σ : CR
: σ))p
Remark:
beta=0 is equivalent to a classical linear assymptote.
Note:
Another definition of the asymptote (**plastic asymptote) is also implemented but not
documented here because still under development.
Stored Variables:
The internal variables stored for this model are the total strain (code etoxx), the tensorial
variables Xi (code kip#xx).
prefix
size
description
default
eto
sig
eel
ean
kip#
T-2
T-2
T-2
T-2
T-2
total (small deformation) strain
Cauchy stress
elastic strain
anelastic strain
Xi variable
yes
yes
yes
yes
no
The code names will replace the # symbol with a sequential number from 1. to nt.
The default saving of variables in the output files are only those marked by yes in the
previous table.
Example:
***behavior viscoelastic_spectral
**elasticity isotropic
young 2800.
poisson 0.3
**spectrum
*n1 -30.
*n2 30.
*nt 50
*nc 7.
*n0 3.
**viscous_effects
*lrt 0.6
*lrc 0.6
**reversible_asymptote
*beta 1.
*p
1.
***return
Z-set — Non-linear material
& structure analysis suite
10.11
***behavior hyperviscoe
***behavior hyperviscoelastic
Description:
This behavior is a general Hyper-viscoelastic form with both shear and volumetric recovery
components. The model is based on the discussions from Simo and Hughs [Simo98]. The
model uses the same hyperelastic potential components as in the hyper_elastic behavior.
Also any number of shear and volumetric terms can be added to the model.
Syntax:
***behavior hyperviscoelastic
[ **thermal_strain <THERMAL_STRAIN> ]
[ **hyperelasticity <HYPERELASTIC_LAW> ]
[ **hyperelasticity <ISOTROPIC_HYPERELASTIC_LAW> ]
[ **model_coefficients ]
...
possible **-level commands for hyper law
Example:
The following example comes from the validation tests. More examples can be found in
test/Viscoelastic_test/INP
***behavior hyperviscoelastic
**hyperelasticity logarithmic
**elasticity isotropic
young
2.8125
poisson 0.40625
**shear
tau
1.0
omega 0.4
**shear
tau
10.0
omega 0.4
**volumic
tau
3.
omega 0.5
***return
10.12
Z-set — Non-linear material
& structure analysis suite
***behavior gen evp
***behavior gen evp
Description:
This material model is a generalized implementation of elastic, elastic-plastic or viscoplastic,
and multi potential constitutive equations. The model is constructed entirely using object
“bricks.” For the moment, the model will be constructed using an elasticity object, a number of “potentials” and interactions between the potentials. Potentials represent inelastic
dissipations which describe the evolution of independent inelastic deformation mechanisms.
Hardening mechanisms are modeled with objects within the individual potentials (type of
hardening depending on the type of potential) and are thus not yet specified. This behavior
is essentially a manager of sub-model objects, and will be discussed in general terms about
the permissible variables.
The model’s internal variables are determined by the sub-objects which have been selected
by the user. The general storage form includes variables which are “global” to the material
laws, and therefore form relations with the imposed (observable) variables or apply to all the
potentials. Each potential may additionally contain “parameter” variables which do not have
associated forces, and “hardening” variables which do have associated forces. The distinction
concerns the form of interaction which is possible. The total variables may therefore be
envisioned as:
[el . . .][p1 h1 ][p2 h2 ] . . . [pn hn ]
where the first bracketed term represents the “global” model variables and each additional
represents the potential mechanisms chosen. The first term is noted to assume a linear elastic
small deformation law1 . Again, the pi and hi variables are defined by the chosen potentials.
The model additionally stores auxiliary variables used as secondary output data. These
variables take the following form:
[. . .][1 . . .][2 . . .] . . . [n . . .]
The first bracketed term is again for the “global” auxiliary variables and the following bracketed terms indicates each potential’s
P variables. In the event of more than one potential, the
first total inelastic strain, in =
i will be stored in addition to each inelastic deformation
component.
If interactions are present, the hardening variable evolutions will take the form:
ḣi = λ̇i mi (σ, vi , Hi ) − ω̇ i (Hi )
where λ̇i is the plasticity or viscoplasticity multiplier for the i-th potential, vi is the cumulated
inelastic strain equivalent, and Hi are the associated forces for the potential i. Note that the
evolution equations will normally only be written in terms of the associated force and thus
one can immediately extend these models to include state coupling2 .
State coupling is given through symmetric interaction matrices M:
Hi = Mij hj
1
2
extensions which change the el variable will be given in the next release of ZéBuLoN
exceptions are available
Z-set — Non-linear material
& structure analysis suite
10.13
***behavior gen evp
Syntax:
The material file structure for the gen evp model consists of an elasticity object, an optional
thermal strain object, an optional arbitrary number of potentials (without restriction on
types), and a number of optional interaction objects:
***behavior gen_evp [ modifier ]
**elasticity
<ELASTICITY>
[ **global_output
[ **damage
<DAMAGE>
[ **localization
<LOCALIZATION>
[ **global_function <GLOBAL_FUNCTION>
[ **thermal_strain
<THERMAL_STRAIN>
[ **conductivity
<CONDUCTIVITY>
[ **potential
<POTENTIAL> [name]
...
[ **interaction
<INTERACTION>
]
...
]
]
]
]
]
]
]
The compatibility of objects with the other objects, and with the integration method will
be investigated during the running of the problem. Because of the dynamic nature of these
models however, it is often difficult to make any verification as to the physical meaning of
a particular model combination. It is therefore strongly advised to observe the material
behavior on a single element. This allows experimentation of the integration method and
selection of the output variables without performing costly full scale calculations.
Multi-potential models are primarily used for cases of time independent plasticity in combination with viscoplastic deformation, or to assemble multiple crystalline deformation systems.
By default, there is no thermal strain, no inelastic deformation or any interactions.
The optional names (name) given after each potential type are used as a means to specify
individual potentials (for interactions), and in construction of the output variable names.
Normally, it is advised to use ep for a plastic potential’s name, and ev for a viscoplastic one.
Unless the option **global output is given, the internal variables are stored in their local
material frame instead of the global one.
Stored Variables:
The stored variables for this model are the following:
prefix
size
description
default
eto
sig
eel
ein
enmi
T-2
T-2
T-2
T-2
T-2
total (small deformation) strain
Cauchy stress
elastic strain
total inelastic strain tensor
potential named nm inelastic strain
tensor
yes
yes
no
yes
yes
The variable ein is only stored in the event of multiple potentials. The separate inelastic
strain tensors for each potential, and their hardening variable names will be given for each
separate potential type.
Because the behavior does not know any specifics of the potentials or the user supplied
names, and the applications of the same potential object may differ according to the rest
of the behavior options, the variable names are vaguely specified at this moment. Name
10.14
Z-set — Non-linear material
& structure analysis suite
***behavior gen evp
verifications are strongly advised with the -v command line switch or using the **verbose
output option.
Example:
The first example is for a viscoplastic behavior with two kinematic hardening variables and a
Von Mises criterion (see the following sections for the sub-model syntax):
***behavior gen_evp
**elasticity isotropic
young 200000.
poisson 0.30
**potential gen_evp ev
*criterion mises
*flow norton
n 7.0
K 400.
*kinematic nonlinear
C 15000.0
D 300.0
*kinematic nonlinear
C 6000.0
D 100.0
*isotropic constant
R0 130.0
***return
Different cases using this behavior are described more fully in the example handbook.
This behavior is designed to work normally with all the integration methods, and gives
the best tangent matrix possible for a given model. The use of specialized options may
however limit the use to a certain integration, or otherwise. The user of this behavior is
therefore advised to try the theta method method integration on a volume element, and
adjust solution parameters according to the messages output if any3 .
3
This process will be automated for the best combination in later versions of the code
Z-set — Non-linear material
& structure analysis suite
10.15
***behavior reduced plast
***behavior reduced plastic
Description:
The reduced plastic behavior is an alternate formulation of the gen_evp behavior which
provides a much more efficient implicit integration. The behavior reduces the integration
variables to a single tensor and a scalar per deformation potential. Because the integration
must assume certain forms for the potentials and hardening variables, this behavior uses a
sub-set of the gen_evp options. In particular, there is no possibility for state interactions, and
the number of potentials implemented is reduced. Domain modifying options such as damage
or localization is nor currently implemented in this framework either.
Thermal deformations, variable coefficients, all the elasticity models, all the flow laws,
and the majority of kinematic hardening models are implemented. There is no limitation
on the number or mixing of kinematic models which are entered per potential, as long as
they support the reduced integration. Note that certain laws break distinctly the reduced
integration and are not implemented (i.e. Ziegler).
There is currently a limitation that the isotropic hardening be a function of the cumulated
multiplier. This means that the internal variable versions of isotropic model are not supported.
Note that non-macro potential models such as the mono-crystals are not implemented.
Syntax:
***behavior reduced_plastic [ modifier ]
**elasticity
<ELASTICITY>
[ **thermal_strain <THERMAL_STRAIN>
]
[ **potential
<POTENTIAL> [name] ]
...
Stored Variables:
The stored variables for this model are the following:
prefix
size
description
default
eto
sig
eel
ein
T-2
T-2
T-2
T-2
total (small deformation) strain
Cauchy stress
elastic strain
total inelastic strain tensor
yes
yes
no
yes
The variable ein is only stored in the event of multiple potentials. The separate inelastic
strain tensors for each potential, and their hardening variable names will be given for each
separate potential type.
This behavior must be used with a θ-method of type A.
Example:
An example the input for which can be found in /test/Kinematic/M AT /.. is given below.
***behavior reduced_plastic
**elasticity isotropic
young 185000.
poisson 0.3
**potential gen_evp ev
10.16
Z-set — Non-linear material
& structure analysis suite
***behavior reduced plast
*store_all
*flow norton
K .05
n 2.0
*kinematic nonlinear_with_crit
C
600000.
D
1000.
omega 0.50
m1
1.
m2
1.
eta
1.e-12
*isotropic constant
R0 400.
***return
Z-set — Non-linear material
& structure analysis suite
10.17
***behavior porous plastic
***behavior porous plastic
Description:
This material model is used for damage and densification of porous materials for which the
flow rule is associated4 . The model will be fabricated based on an assemblage of various
objects to model the elasticity, criterion, flow, and various hardening (isotropic and kinematic)
and nucleation options. The model supports multiple potentials, viscoplasticity, anisotropy,
and combined isotropic and kinematic hardening. Many other features are also available for
modifying the behavior including models for nucleation of porosity, fine tuning the numerical
implementation, and studying the onset of bifurcation instabilities.
Please note that the kinematic hardening for porous plasticity is fundamentally different
from what is used for non-porous unified viscoplastic models such as in gen evp [BessXX]. The
purpose of kinematic hardening is also not so much to simulate cyclic behavior, but rather to
provide control of the yield surface curvature.
Many porous plastic analysis problems involve finite strain. Please refer to the documentation for <MODIFIER> at page 14.5 for the specifics of different corotational finite strain
methods. Note that frequently we use the no J option where the det(F) volumetric adjustment of the Cauchy stress is ignored. The reason to do this is because the volumetric change
effects on the stress are essentially resolved by the porous potentials.
A common theme in porous plasticity is that a matrix stress σ? is solved for based on the
macroscopic nominal stress σ, current plastic strain p, and current porosity f .
σ
σ∗
The means of localizing to the effective matrix stress is therefore given by a potential function
φ which must always be satisfied:
φ(σ
, f, σ? ) = 0
∼
Since the matrix stress σ? is the “real material,” porous plasticity uses the measure σ? − R as
the overstress condition with R being the current isotropic yield radius. The general porous
viscoplastic material is therefore described by the following plastic multiplier:
ṗ = v̇(σ? − R)
with v̇ being any of the Z-mat flow laws given under <FLOW> (see page 13.44). The evolution
of plastic strain is found via an equivalence of plastic work between the matrix stress/plastic
4
This behavior is programmed only for the θ-method. No plane stress is available. The coefficients can
depend on porosity, notably of which is the f ∗ for the Gurson potential.
10.18
Z-set — Non-linear material
& structure analysis suite
***behavior porous plastic
strain and the macroscopic stress and strain (of the composite porous medium):
(1 − f )σ? ṗ = ∼˙p : σ
= (1 − f )λ̇
∼
∂σ?
:σ
∼
∂σ
∼
The evolution of porosity is given via conservation of mass given that the plastic strain has a
volumetric part due to the pore growth or compaction:
f˙ = (1 − f ) Trace ∼˙p
In the case of multiple criterions Hp and Hf are used to specify interaction between potential
so that:
pie =
X
fi =
X
fti
X
Hpij (pj )
j
Hfij (fgj + fnj )
j
=
j
Hfij (fgj + fnj + fncl
)
j
. . . continued
Z-set — Non-linear material
& structure analysis suite
10.19
***behavior porous plastic
Syntax:
The whole-behavior keyword summary is given below:
***behavior porous_plastic
**thermal_strain
**elasticity
**porous_potential
*porous_criterion
*flow
*isotropic_hardening
*strain_nucleation
*kinematic
*shear_anisotropy
**porous_potential
**Hf
**Hp
**broken_behavior
**adiabatic_heating
**no_C_trick
**save_D
**save_L
**bifurcation_D
**bifurcation_L
**perturbation
**additional_var_aux
[modifier]
<THERMAL_STRAIN>
<ELASTICITY>
[name]
<POROUS_CRITERION>
<FLOW>
<ISOTROPIC>
<STRAIN_NUCLEATION>
<POROUS_KINEMATIC>
<CRITERION>
...
<SMATRIX>
<SMATRIX>
<ELASTICITY>
The following sub-options control the global behavior operation, while the rest of the specific
model will be determined by the dynamic components chosen, most important of which is the
porous potential.
**additional var aux allows requesting that additional auxiliary variables be added to
the model output. These can be chosen from triax, p1, p2, p3. These variables are
closely coupled to the behavior so can be used for additional coefficient dependencies.
**adiabatic heating includes adiabatic heating via plastic work, and temperature is
included as a state variable named T.
**broken behavior is used to enter an elasticity matrix to be used as the behavior after
failure has been reached. The measure of “breaking” is determined by the models
potentials (there could be different criteria for several potentials together).
**Hf Coupling term for porosity given above. The matrix is read in as a series of real
(floating point) values to fill the matrix. Note these are not general coefficients, but
fixed values. The matrix is read Hf11 , Hf12 , . . . , Hf1N , Hf21 , . . . , HfN N with N the number
of porous potentials.
**Hp Interaction matrix for plastic strain influence between potentials. The matrix is read
Hp11 , Hp12 , . . . , Hp1N , Hp21 , . . . , HpN N with N the number of porous potentials.
10.20
Z-set — Non-linear material
& structure analysis suite
***behavior porous plastic
**porous potential define a potential which is part of the model. This command is
detailed separately on page 10.23. Any number of potentials (greater than or equal to
1) can be entered with repeated uses of this command.
Stored Variables:
The stored variables for this model are the following:
prefix
size
description
default
eto
sig
eel
broken
seq
p
fg
fn
fncl
f
ft
pe
Xmic
Xmac
alpha
X
T-2
T-2
T-2
S
S
S
S
S
S
S
S
S
T-2
T-2
T-2
T-2
Total (small deformation) strain
Cauchy stress
Elastic strain
if broken
effective flow stress σ ∗
equivalent plastic strain
growth porosity
nucleation porosity
effective crack like nucleation porosity
void volume fraction fg + fn
total effective porosity fg + fn + fncl
effective plastic strain
microscopic back stress
macroscopic back stress
kinematic hardening internal variable
sum of all macroscopic back stresses
yes
yes
no
yes
yes
yes
yes
yes
yes
yes
yes
yes
no
no
no
no
Z-set — Non-linear material
& structure analysis suite
10.21
***behavior porous plastic
Example:
As discussed above, the porous plastic material behavior is a very broad code, and encompasses
many options and features. The following is a simple “whole” example to give an indication
of overall syntax. The user will please refer to the different applicable porous potentials for
further examples and details.
***behavior porous_plastic lagrange_rotate_no_J
**elasticity isotropic
young 210000. poisson 0.3
**porous_potential
*porous_criterion gurson
fs = f
q1 1.5
q2 1.
*shear_anisotropy mises
*isotropic_hardening constant
R0
T
200. -50.
200.
0.
150. 100.
100. 500.
*flow norton K .01 n 5.
**adiabatic_heating .9
**coefficient
capacity 3.6
***return
10.22
Z-set — Non-linear material
& structure analysis suite
***behavior porous plastic
**porous potential
**porous potential
Description:
*porous criterion
*shear anisotropy
*flow
*isotropic
*strain nucleation
*kinematic
Z-set — Non-linear material
& structure analysis suite
10.23
***behavior mechanical ste
***behavior mechanical step phase
Description:
This behavior is given to model materials which undergo a a phase change
Syntax:
***behavior mechanical_step_phase
**integer_steps
**parameter phase_id
**phase Oxide 0
*file step.mat 2
*rotation <rot definition>
**phase Alu 1 *file Alu.mat
**return
Example:
***behavior mechanical_step_phase
**integer_steps
**parameter phase_id
**phase Oxide 0
*file step.mat 2
*rotation <rot definition>
**phase Alu 1 *file Alu.mat
**return
10.24
%
%
%
%
%
%
specify integer step values to be used
the parameter defining the step phases
use this phase if phase_id = 0 (int value)
This file, 2nd ***behavior instance
a possible rotation defintion
use mat defined in Alu.mat for phase_id = 1
Z-set — Non-linear material
& structure analysis suite
***behavior umat
***behavior umat
Description:
This behavior is used to run a UMAT function within Z-set. It can be used to run Z-mat
using the ABAQUS interface within Zebulon, or can be used to run a “real” Fortran UMAT
within Z-set, which is perhaps very useful to be able to run the UMAT within the simulation
module.
Syntax:
***behavior umat
**constants num-coef
**finite_strain
**skip_param
**cmname mat-name
**umat_file fname
**depvar num-sdv
**model_coef
C## value
Example:
The following example shows a umat material definition, which is linked in fact to the Z-mat
file on the right. By default (in the absence of a **umat file definition) the umat behavior
uses a Z-mat one.
***behavior umat
**cmname visc3_zmat.mat
**depvar 2
**coefficient
masvol 8.1e-9
***return
***suppress_temperature
***material
*file visc3_zmat.mat
*integration theta_method_a 0.5 1.e-9 120
***external_storage
*file oo.store
*vars
evi22 evcum
*buffer_size 3
***behavior gen_evp
**elasticity isotropic
young 260000.
poisson 0.3
**potential gen_evp ev
*criterion mises
*flow norton
n 7.0
K 400.
*kinematic nonlinear
C 30000.0
D 500.0
*isotropic constant
Z-set — Non-linear material
& structure analysis suite
x1
10.25
***behavior umat
R0 130.0
***return
10.26
Z-set — Non-linear material
& structure analysis suite
Chapter 11
Secondary Models
Z-set — Non-linear material
& structure analysis suite
11.1
11.2
Z-set — Non-linear material
& structure analysis suite
***behavior aging
***behavior aging
Description:
This behavior is a viscoplastic model intended for modeling the thermally activated aging
behavior of aluminum for applications such as cast Al cylinder heads in TMF loading. The
model includes classical combined nonlinear isotropic-kinematic hardening (only 2 kinematic
terms are now allowed). The model is programmed for both Runge-Kutta and theta-method
implicit integration, and for both general FEA and simulation modes.
The model has a standard additive strain decomposition with elastic, viscoplastic and
thermal strain parts.
˙ el
= ˙ to − ˙ th − ˙ vp
The stress is computed from linear elasticity depending on the elastic matrix component
selected:
σ
= Del : el
There is an integrated aging parameter ζ which will cause aging effects on the hardening.
This parameter evolves from 0 (unaged) to 1 (fully aged) with a saturating nonlinear form.
ζ∞ − ζ
ζ̇
=
τ
and the coefficient τ is the time constant for saturation of the aging. We expect that the two
coefficients ζ∞ and τ are functions of temperature to be able to handle overheating effects.
The hardening parameters are computed with aging effects as follows:
X1
= 23 C1 α1
X2
= (1 − ζ) 32 C2 α2
R
= R0 + Q(1 − e−bλ ) + (1 − ζ)R0∗
The remainder of the model is classical viscoplasticity:
f
λ̇
˙ vp
α̇i
= J(s − X1 − X2 ) − R
n
f
=
K
= λ̇n
h
= λ̇ n −
3 Di
2 Ci X i
n
∂f
∂σ
i
Note the aging effects are in the recall term for X2 .
. . . continued
Z-set — Non-linear material
& structure analysis suite
11.3
***behavior aging
Syntax:
The basic input syntax here is:
***behavior aging_theta
[ **elasticity
<ELASTICITY> ]
[ **thermal_strain <THERMAL_STRAIN> ]
**model_coef
...
The following coefficients are available:
K,n viscoplastic Norton law coefficients for λ̇ = hf /Kin
C1,D1 nonlinear Armstrong-Frederick kinematic hardening. C1 is the kinematic modulus
and D1 is the saturation rate. The saturation back stress is occurs at C1/D1.
C2,D2 nonlinear kinematic coefficients which have the aging effects applied to them.
R0, Q, b nonlinear isotropic hardening (or softening with Q < 0) having the same meaning
as the nonlinear isotropic model.
R0 star isotropic aging effect on the yield radius.
Tau,a inf the two aging variable coefficients.
alpha optional thermal strain expansion coefficient.
Stored Variables:
The following variables are stored with this model:
prefix
size
description
default
eto
sig
eel
evi
eme
evcum
T-2
T-2
T-2
T-2
T-2
S
yes
yes
yes
yes
yes
yes
age
alpha(i)
S
T-2
total strain
Cauchy stress
elastic strain tensor
viscoplastic strain tensor
mechanical strain tensor
cumulated viscoplastic strain magnitude
aging variable
kinematic back strain, i=1,2
yes
yes
Note:
Early releases of this model misspell the model name as ageing
Example:
There is an example file in the test database directory ZebFront test/INP named
aging*.inp
These tests are normally deprecated compared to the equivalent but more general capability in the tests Simulator test/INP aging tmf*.
11.4
Z-set — Non-linear material
& structure analysis suite
***behavior aniso damage
***behavior aniso damage
Description:
This model implements an anisotropic damage model for fibrous ceramic composites1 . The
damage is described by a number of scalar and tensorial variables as input by the user. The
effect of closure strains and non-symmetric tension-compression behavior is handled as well.
The stress is calculated using the following definition of effective modulus:
X
Ceff : ( − c )
σ = C0 : −
where C0 is the undamaged elasticity tensor, Ceff is the damage modification of the elasticity,
and c is a user loadable closure strain. Each scalar and tensorial variable will contribute a
term to the summation of Ceff .
Syntax:
***behavior aniso_damage [modifier ]
**C0 ELASTICITY
[ **thermal_strain THERMAL STRAIN ]
[ **damage ANISOTROPIC DAMAGE]
[ **closure ]
[
t1 ... t6 ]
[ **scalar_interaction (full | identity)
]
Stored Variables:
prefix
size
description
default
eto
sig
d(i)
D
T-2
T-2
S
T-2
strain tensor
Cauchy stress
scalar damage variables
tensor damage variable
yes
yes
yes
yes
Example:
The following example shows the input format for a material with two scalar variables:
***behavior aniso_damage
**closure 1.e-3 1.e-3 1.e-3
**c0 orthotropic
c11
2.5662e11
c22
2.5662e11
c33
2.6887e11
c12
6.0439e10
c13
8.2435e10
c23
8.2435e10
c44
8.5e9
c55
8.5e9
c66
8.5e9
.5e-3 .5e-3 .5e-3
% 11 22 33 12 23 31
**damage scalar
*eta 1.0
1
the model is coded in Aniso damage.z in the source dir zZfrontBehavior
Z-set — Non-linear material
& structure analysis suite
11.5
***behavior aniso damage
*n
(1. 0. 0.)
*K_coeffs 1. 0. 0. 0. 0. 0. 0.7 0.0 0.7
*g g1
Y0
1.69e4
Yc
1.6e5
delta_c
0.6
**damage scalar
*n
(0. 1. 0.)
*K_coeffs 0. 1. 0. 0. 0. 0. 0.7 0.7 0.0
*g g1
delta_c 0.6
Y0 1.69e4
Yc 1.6e5
**damage scalar
*n
(0. 0. 1.)
*K_coeffs 0. 0. 1. 0. 0. 0. 0.0 0.7 0.7
*g g1
delta_c
0.6
Y0
1.69e4
Yc
1.6e5
**damage elastic_tensorial
*Q diagonal
Q1 1.0
*Qd diagonal
Qd1 1.0
*K_coeffs
1. 1. 1. 0.7 0.7 0.7 0.5 0.5 0.5
*g g1
delta_c
0.6
Y0
0.0
Yc
1.6e5
***return
11.6
Z-set — Non-linear material
& structure analysis suite
***behavior becker needleman
***behavior becker needleman
Description:
This model is a direct implementation of the viscoplastic porous damage model given by
R. Becker and A. Needleman “Effect of Yield Surface Curvature on Necking and Failure in
Porous Plastic Solids,” J. Appl. Mech. v53, 491-498 (1986).
B0
B
=σ−X
φ
q2 B : 1
3 B0 : B0
2
∗
− 1 − q1 f ∗
+ 2q1 f cosh
=
2
2 σF
2σF
=U:B
n=
∂φ
∂σ
Where f ∗ is a modification of the porosity f 2 so that if f < fc f ∗ = f and otherwise
f ∗ = fc + ((1/q1 − fc )/(ff − fc )) ∗ (f − fc )
∆λ =
(1 − f )σF ∆p
B:n
ρ
0
g (p)(B : n)2
∂φ
(1 − b)
+ (1 − f )(1 − g(p)) (1 : n)
[B : n]−1
1 − b + bg(p)
(1 − f )σF
∂f
=
Rate equations
˙ el
= ˙ to − λ̇n − ˙ th
σF
(1 − b)σo + bσo g(p)
ṗ
= o
ḟ
= λ̇(1 − f )(1 : n)
α̇
= λ̇ρB
1/m
Syntax:
2
In the porous plastic behavior f ∗ is calculated from f using the coefficient mechanism
Z-set — Non-linear material
& structure analysis suite
11.7
***behavior becker needleman
***behavior becker_needleman modifier
**elasticity <ELASTICITY>
[ **thermal_strain <THERMAL_STRAIN>
**model_coef
coefs
q1, q2, f_c, f_f, e_dot0, b, sig0, m, eps0, C
Example:
Here is an example using coefficients from the original paper.
***behavior becker_needleman
**elasticity isotropic
young
1865.67
poisson 0.3
**model_coef
q1
2.38
q2
0.748
eps0
0.0125
e_dot0 1.e-3 % same as loading rate
sig0
1.0
b
1.0
% 1==isotropic, 0==kinematic
f_c
0.15
f_f
0.25
m
0.02 % 0.002
C
1.
***return
11.8
Z-set — Non-linear material
& structure analysis suite
***behavior cast iron
***behavior cast iron
Description:
This behavior is a combined damage-viscoplasticity model to simulate the nonlinear and
fatigue behavior of cast iron materials. It allows significant flexibility in terms of the combination of damage and plastic mechanisms, hardening models, and driving force for damage.
There are quite a number of coefficients to manipulate however. A description of how to
attack this problem is given in the Theory manual.
This model uses a scalar damage variable which is the summation of a brittle mode based
on maximum principal stress values, and a fatigue damage model like that described by
[Lema85b]. The scalar damage value is however applied to an anisotropic closure treatment
utilizing 2 separate closure/opening criteria. The use of a scalar variable is an approximation
assuming close to cyclically proportional strain paths. The closure point is based on a strain
criterion which shifts into compression as damage accumulates, and the opening criterion
is a sole function of the principle stress being positive. Both terms are smoothed using an
interpolation function with adjustable width. Note that the opening and closing combined
with the general difficulties of damage based models can lead to convergence difficulties. The
damage effects here are not really meant for predicting true failure, but rather are needed to
predict the LCF hysteresis loops. Other methods of cast iron modeling are available using
the cast_iron yield criterion and non-symmetric kinematic hardening model.
Syntax:
The basic input syntax here is:
***behavior cast_iron
[ **thermal_strain <THERMAL_STRAIN> ]
[ **elasticity
<ELASTICITY> ]
[ **isotropic
<ISOTROPIC_HARDENING> ]
**model_coef
...
With the following coefficients available:
K,n viscoplastic Norton law coefficients for λ̇ = hf /Kin
Q1,D1 nonlinear Armstrong-Frederick kinematic hardening. Q1 is the saturation stress level
and D1 is the saturation rate. Any number of these terms may be added, but both
coefficients must be entered always, and the numbering is sequential from 1 to N.
dmax,dmax b,dmax f damage limits preventing total failure.
e 0, e, E Brittle mode damage.
a, b, A, r Simple cumulated plastic strain fatigue mode damage.
delta e, delta s Controls for the width of closure in strain dimensions (δe of 0.004 recommended as a start) and damage “opening” in stress dimensions (δs of 25 recommended).
delta d Shift parameter for progressively compressive closure as damage increases. The
closure point is δd × d
coupled Set to 1.0 in order to have damage coupling to the kinematic hardening.
Z-set — Non-linear material
& structure analysis suite
11.9
***behavior cast iron
Stored Variables:
prefix
size
description
default
eto
sig
eel
eps me
eps th
evi
evcum
df
db
Dsum
alpha(i)
T-2
T-2
T-2
T-2
T-2
T-2
S
S
S
S
T-2
strain tensor
Cauchy stress
elastic strain
mechanical strain
thermal strain
viscoplastic strain
inelastic strain equivalent
fatigue damage
brittle damage
total effective damage
kinematic hardening variable
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
Example:
The following is a short example material file for gray iron.
***behavior cast_iron
**elasticity isotropic
young
130000.
poisson 0.26
% yield
**isotropic constant
R0 150.
**model_coef
coupled 1.0
% viscoplastic
K 400.
n 4.
% kinematic
Q1 300.0
D1 80.
% brittle damage
e
1.0 % exponent
e_0 2.0 % criterion
E 1000. % approx critical stress
% progressive closure
delta_d
0.006 % shift in strain closure according to d
delta_e
0.008 % width of strain transition
***return
11.10
Z-set — Non-linear material
& structure analysis suite
***behavior bodner partom
***behavior bodner partom
Description:
This behavior is an implementation of the classical model due to Bodner and Partom. The
model is viscoplastic and incorporates scalar and tensorial hardening variables. There is no
initial yield radius thereby allowing inelastic deformation at very low stress levels over long
periods of time. The state and evolution equations are the following:
√
Z = Zi + Z0 + β : u
u = σ/ σ : σ
2 n Z
2
Dp2 = D0 exp −
3J2
p
S=U:σ
λ2 = Dp2 /J2
˙ vi = λ2 S
p
Wp = σ : ˙ vi = λ2 S : σ
˙ el
Żi
β̇
J2 = 21 S : S
= ˙ to − ˙ vi
= m1 (Z1 − Zi − Z0 )Wp
= m2 (Z3 u − β)Wp
(1)
Syntax:
***behavior bodner_partom
[ **thermal_strain <THERMAL_STRAIN> ]
[ **elasticity
<ELASTICITY> ]
**model_coef
Coefficient names are n, Z0, Z2, D0, Z1, Z3, m1, m2, A1, A2, r1, r2
Stored Variables:
prefix
size
description
default
eto
sig
evi
Zi
beta
p
Ztot
T-2
T-2
T-2
S
T-2
S
S
total strain
Cauchy stress
inelastic strain tensor
isotropic drag stress
kinematic variable
inelastic strain equivalent
Sum of Z parts
yes
yes
yes
yes
yes
yes
yes
Z-set — Non-linear material
& structure analysis suite
11.11
***behavior bodner partom
Example:
The following is a simple example of the Bodner-partom material using room temperature
coefs for HASTELLOY-X as given by Rowley and Thornton J. Eng. Mat. Tech. 118, 19-27
(1996).
***behavior bodner_partom
**elasticity isotropic
young 196.6e3
poisson 0.33
**model_coef
n
1.0
Z0 1860.
Z2 1860.
D0 10000.0
Z1 2390.
Z3 603.
m1 0.139
m2 3.49
A1 1.0e-9
A2 1.0e-9
r1 1.
r2 1.
***return
11.12
Z-set — Non-linear material
& structure analysis suite
***behavior finite strain crystal
***behavior finite strain crystal
Description:
This behavior is a simple implementation of a finite strain formulation of a single crystal3
This behavior is included as an example of ZebFront programming, and the source can be
found in the developers manual. This model only allows one crystal orientation to be input,
and works for Runge-Kutta integration only.
This model works with the assumption of multiplicative plasticity, where the deformation
gradient F is broken into an elastic part Fe and a plastic part Fp .
Fe = FF−1
p ;
The elastic deformation gradient is separated into a rotation component and a stress component.
Fe = Re Ue
and the stress is calculated using the logarithmic strain measure from the elastic stretch.
σ = Del : log(Fe )
There is a yield criterion for each slip system.
fi = mi : σ − Cαi − R(gi )
which is used to calculate the evolution.
X
Ḟp =
v̇(fi )mi Fp
i
γi is the integration of v̇(fi ).
α̇i = γ̇i sign(mi : σ − Cαi ) − D ∗ αi ∗ γ̇i
3
Only Runge-Kutta integration is implemented.
Z-set — Non-linear material
& structure analysis suite
11.13
***behavior finite strain crystal
Syntax:
***behavior finite_strain_crystal
**elasticity <ELASTICITY>
**flow
<FLOW>
**isotropic
<ISOTROPIC_HARDENING>
**orientation <CRYSTAL_ORIENTATION>
**model_coef
C COEFFICIENT
D COEFFICIENT
Stored Variables:
11.14
prefix
size
description
default
F
sig
Fp
gamma#
alpha#
crss#
UT-2
T-2
UT-2
V
V
V
deformation gradient
total Cauchy stress
plastic deformation gradient
resolved shear strains
back strains on slip system
current resolved shear stress
yes
yes
yes
yes
yes
yes
Z-set — Non-linear material
& structure analysis suite
***behavior matmod
***behavior matmod
Description:
This model is an implementation of the MATMOD equations due to Miller [Mill76]. The model
accounts for isotropic/kinematic hardening and presents a particular form of temperature
dependence in the material coefficients.
(
˙ in = Bθ
θ0
0
"
sinh Af1
= exp
J(σ − R)
D
i
nh
Q
ln
− 0.6kT
m
θ0
3/2 #)n
0.6Tm
T
3 s−R
2 J(σ − R)
o
+1
= exp −Q/kT )
for T ≤ 0.6Tm
for T ≥ 0.6Tm
(2)
Ṙ = H1 (˙ in − Bθ0 [sinh(A1 |R|)]n )
R
|R|
n
Ḋ = H2 |˙ in | C2 + |R| − (A2 /A1 )D3 − H2 C2 Bθ0 sinh(A2 D3 )
Syntax:
***behavior matmod
[ **thermal_strain <THERMAL_STRAIN> ]
[ **elasticity
<ELASTICITY> ]
**model_coef
Stored Variables:
prefix
size
description
default
eto
sig
ein
D
R
T-2
T-2
T-2
S
T-2
total strain
Cauchy stress
inelastic strain tensor
isotropic drag stress
kinematic rest stress
yes
yes
yes
yes
yes
Example:
The following is a simple example of the MATMOD material corresponding to the material
in Miller’s original paper:
%
% As given by Mill76 (Tests matmod#)
%
***behavior matmod
Z-set — Non-linear material
& structure analysis suite
11.15
***behavior matmod
**elasticity
young
temperature
1.93e5 23.0
1.55e5 538.0
poisson 0.3
***model_coef
A1
0.108
% MPa^-1
A2
2.27e-7
% MPa^-1
B
1.e15
% sec^-1
C2
0.69
% MPa
H1
1930
% MPa
H2
100
%
n
5.8
%
Q
91000.0
% cal/mol
Tm
1800.0
% Degree K
***return
11.16
Z-set — Non-linear material
& structure analysis suite
***behavior matmod z
***behavior matmod z
Description:
This model is an implementation of the MATMOD-Z equations due to Miller [Mill76]. The
model accounts for isotropic/kinematic hardening and presents a particular form of temperature dependence in the material coefficients.
Syntax:
***behavior matmod
[ **thermal_strain <THERMAL_STRAIN> ]
**model_coef
Stored Variables:
prefix
size
description
default
eto
sig
lam vp,
lam ir
R
Fdef
Fsol
evp
eir
T-2
T-2
S
total strain
Cauchy stress
plastic multipliers
yes
yes
yes
viscoplastic strain
irradiation strain
yes
yes
yes
yes
yes
S
S
S
T-2
T-2
Example:
The following is a simple example of the MATMOD-Z material corresponding to the material
in Miller’s original paper:
%
% Zircaloy model from Oldsberg, Miller and Lucas STP 681 (1978).
%
***behavior matmod_z
**model_coef
E function 9.65e4 - (9.65e4-5.51e4)*(temperature-300.0)/700.0;
poisson 0.3
F 0.5
G 0.25
H 0.75
A 500.0
A1 3240.0
A2 1.78e8
A4 8.89e-13
B 5.0e12
C1 0.143
C2 1.0e-10
Fsa -1.815e-7
Z-set — Non-linear material
& structure analysis suite
11.17
***behavior matmod z
Fsb 4.99e-8
Fsm2 2.32e-7
Fsm3 3.25e-6
H1 0.0070 % dR
H2 6.64e-4 % dFdef
H3 2.0e-26 % dFdef irr
k 1.9859
kp 91.0
kq 3.0
n 4.5
p 0.5
Qs 60000.0
Q4 8400.0
Tt 930.0
Z2 1.0e23
Z3 1.0e12
beta2 6.0
beta3 3.0
% phi = flux
phi 0.0
11.18
Z-set — Non-linear material
& structure analysis suite
***behavior memory
***behavior memory
Description:
This behavior is a ZebFront implementation of a strain range memory isotropic hardening
model described by Lemaitre and Chaboche [Lema91]. The model has 2 non-linear kinematic
hardening components (coefficients C1, D1, C2, and D2), and a Norton type flow law (coefficients K and n).
Q = Q0 + (Qsat − Q0 ) exp(−2µq)
R
= R0 + bQr
(3)
n∗ = 12 (vi − z)/J(vi − z)
η
= n : n∗
q̇
= η λ̇
(4)
ṙ = λ̇ [1 − (R − R0 )/Q]
ż = 2(n∗ : vi )n∗
Syntax:
***behavior memory
**elasticity <ELASTICITY>
**model_coef
coefs
Example:
***behavior memory
**elasticity isotropic
young 260000.
poisson 0.3
**model_coef
n
7.0
K
100.
C1
30000.0
D1
80.
C2
100000.
D2
1200.
R0
150.0
b
150.
Q0
400.
Qsat -350.
mu
100.
***return
Z-set — Non-linear material
& structure analysis suite
11.19
***behavior non associated
***behavior non associated
Description:
The non_associated behavior is a ZebFront behavior used as an example for non-associated
deformation with kinematic hardening variables, which can have deviatoric and spherical
components4 . The source for this model can be found in the developer manual.
The model uses elasticity, criterion, flow, and isotropic classes. An interesting use of this
model is with a non-associated criterion such as linear_drucker_prager. There is however
no such requirement, so the model could be used with von Mises as well.
Any number of kinematic hardening variables are possible. For kinematic i, the back stress
X is calculated as follows:
Xi
Xsi = Csi αsi
Xdi = Cdi αdi
= Xdi + δXsi
(5)
Noting that different back stress moduli are available for the two components.
The evolution of back stress is also separated in two components:
Dsi
α̇si = λ̇ Tr(n) −
Xs
Csi i
α̇di
Ddi
0
= λ̇ n −
Xdi
Cdi
Note here that n is the criterions normal and not ∂∂f
σ.
Syntax:
The syntax is in standard ZebFront format, with a number of standard sub-classes.
***behavior non_associated
**elasticity <ELASTICITY>
**flow
<FLOW>
**isotropic
<ISOTROPIC_HARDENING>
**criterion
<CRITERION>
**model_coef
Cd1 COEFFICIENT
Cs1 COEFFICIENT
Dd1 COEFFICIENT
Ds1 COEFFICIENT
..
4
11.20
there is only a Runge-Kutta implementation in Z8.0 – this model will be implemented in gen evp in Z8.1
Z-set — Non-linear material
& structure analysis suite
***behavior non associated
Stored Variables:
prefix
size
description
default
F
sig
Fp
gamma#
alpha#
crss#
UT-2
T-2
UT-2
V
V
V
deformation gradient
total Cauchy stress
plastic deformation gradient
resolved shear strains
back strains on slip system
current resolved shear stress
yes
yes
yes
yes
yes
yes
Example:
***behavior drucker_prager
**elasticity isotropic
young
2.25
poisson 0.125
**isotropic constant
R0
0.0011547
**flow norton
K 1.e-6
n 2.
**criterion linear_drucker_prager
friction_angle 20.0
dilatation_angle 20.0
K
.9
**model_coef
Cd1 1.e-1
Dd1 0.0
Cs1 1.e-1
Ds1 0.0
***return
Z-set — Non-linear material
& structure analysis suite
11.21
***behavior visco aniso
***behavior visco aniso damage
Description:
This model implements a model of viscoplasticity with fully anisotropic coefficients and damage. The model accepts a number of mixed damage components based on scalar or tensorial
variables, and any number of kinematic variables. The later must be chosen from a sub-set of
the KINEMATIC objects described later in this chapter. The compatible models are indicated
in the later section.
Syntax:
***behavior visco_aniso_damage [modifier ]
[ **thermal_strain THERMAL STRAIN
**C0
ELASTICITY
**H0
COEFFICIENT MATRIX
**flow
FLOW
**isotropic ISOTROPIC
**kinematic KINEMATIC
...
[ **effective_stress
]
[ **effective_operator ]
[ **damage
ANISOTROPIC DAMAGE
...
[ **closure ]
[
t11 t22 t33 t12 t23 t31 ]
[ **scalar_interaction (full | identity) ]
Stored Variables:
prefix
size
description
default
eto
sig
delta(i)
d
Y
evi
evcum
alpha(i)
T-2
T-2
S
T-2
T-2
T-2
S
T-2
strain tensor
Cauchy stress
scalar damage variables
tensor damage variable
driving force for tensorial damage
inelastic strain tensor
inelastic strain equivalent
kinematic hardening variable
yes
yes
yes
yes
yes
yes
yes
yes
Example:
***behavior visco_aniso_damage
**effective_stress
**C0 orthotropic
c11 154844.0
c22 154844.0
c33 203236.0
c12 54844.0
11.22
Z-set — Non-linear material
& structure analysis suite
***behavior visco aniso
c23 49357.0
c31 49357.0
c44 50000.0 % 0.5(b11-b12)
c55 50000.0
c66 50000.0
**H0 orthotropic
c11 0.5
c22 0.5
c33 0.15
c12 -0.3
c23 -0.1
c31 -0.1
c44 0.4
c55 0.35
c66 0.35
**flow norton
n 12.0
K 250.0
**isotropic nonlinear
R0 240.0
Q
-70.0
b
180.0
**kinematic nonlinear X1
*C orthotropic
c11 15000.00 c12 0.0 c23 0.0 c31 0.0
c22 15000.00
c33 50000.00
c44
7500.00 % 0.5(b11-b12)
c55
7500.00
c66
7500.00
**kinematic nonlinear X2
*C orthotropic
c11 30000.00 c12 0.0 c23 0.0 c31 0.0
c22 30000.00
c33 100000.00
c44 15000.00 % 0.5(b11-b12)
c55 15000.00
c66 15000.00
*D diagonal 180.0
***return
Z-set — Non-linear material
& structure analysis suite
11.23
***behavior visco aniso
11.24
Z-set — Non-linear material
& structure analysis suite
Chapter 12
Other Models
Z-set — Non-linear material
& structure analysis suite
12.1
12.2
Z-set — Non-linear material
& structure analysis suite
***behavior coefficient di
***behavior coefficient diffusion
Description:
Diffusion behavior which uses the COEFFICIENT to model variations in D with respect to the
concentration.
J~ = D(C)∇C
Syntax:
***behavior coefficient_diffusion
D COEFFICIENT
Stored Variables:
prefix
size
description
default
dC
J
C
V
V
S
gradient of concentration
flux of concentration
the concentration
yes
yes
yes
Example:
***behavior coefficient_diffusion
D
C
5.50339E-22 -1000.
5.50339E-22
0.
4.02370E-20
0.29
8.39143E-20
0.40
1.02221E-18
0.55
2.56706E-18
0.59
2.56706E-18
0.66
2.56706E-18
1000.
***return
Z-set — Non-linear material
& structure analysis suite
12.3
***behavior needleman debonding
***behavior needleman debonding
Description:
This behavior1 is used for the special problem of interface debonding. See the command
**create_interface_elements (and similar) in the Z-set user manual on how to insert
cohesive elements in the mesh. The Needleman model2 is described through a scalar variable
λ which characterizes the relative crack opening:
s
huN i 2
k~uT k 2
+
,
(1)
λ=
δN
δT
where hxi = x if x > 0 and hxi = 0 if x ≤ 0. With respect to the interface normal ~n,
~uN = (~u · ~n) ~n ≡ uN ~n and ~uT ≡ ~u − ~uN denote, respectively, the normal and shear opening
displacements and δN and δT the corresponding maximum allowable values of their norms.
The damage variable λmax , which is the maximum value of λ reached up until the current
instant, increases from 0 (no damage) to 1 (for a broken
element). The normal and shear
components of the cohesive traction T~ , i.e. T~N = T~ · ~n ~n ≡ TN ~n and T~T = T~ − T~N , are
defined by
TN =
uN
27
F (λmax ) , T~T = α ~uδTT F (λmax ) , F (λ) = σmax (1 − λ)2 ,
δN
4
(2)
with α a constant representing the relative magnitude of kT~T k with respect to TN , and σmax
the maximum stress allowable by the element. For the compressive case, where uN < 0, the
normal component of the traction is modified to
TN = αc
uN
F (0),
δN
(3)
with αc a penalization factor. In the literature, αc usually is at least 10α. Figure 1 illustrates
the typical response of the cohesive zone model under specific loads, for the parameters as
given in the example.
Syntax:
***behavior needleman_debonding
sigmax σmax
deltan δn
deltat δt
alpha
α
alphac αc
[no_penetration]
. . . continued
1
this behavior is Z-set specific, and therefore does not apply for Z-mat for other codes
Needleman A., “A continuum model for void nucleation by inclusion debonding”, J. of Applied Mechanics,
54 (1987), pp. 525-531.
2
12.4
Z-set — Non-linear material
& structure analysis suite
***behavior needleman debonding
16.0
125
14.0
100
12.0
75
50
T [MPa]
u [µm]
10.0
8.0
6.0
4.0
25
0
-25
-50
2.0
-75
0.0
-100
-2.0
0.0
1.0
2.0
3.0
4.0
-125
0.0
5.0
1.0
2.0
time [s]
3.0
4.0
5.0
time [s]
1.0
125
100
0.8
75
50
λmax
T [MPa]
0.6
0.4
25
0
-25
-50
0.2
-75
-100
0.0
0.0
1.0
2.0
3.0
4.0
time [s]
5.0
-125
-2.0
0.0
2.0
4.0
6.0 8.0 10.0 12.0 14.0 16.0
u [µm]
Figure 1: Example in two dimensions of the evolution of the cohesive traction as a function
of the opening displacement, for two different loading cases: uN (t) with uT (t) = 0 (thick red
curves), and uT (t) with uN (t) = 0 (thin green curves). Top left: applied load u(t). Note: for
2 ≤ t ≤ 4 s, the applied loading becomes negative (but the response uN remains 0 because
of an implicit non-penetration condition). Top right: response T (t). Bottom left: λmax (t).
Bottom right: u(t) vs T (t).
If no_penetration is specified, a broken element continues to prevent penetration.
Example:
***behavior needleman_debonding
sigmax 100.
deltan
1.e-5
deltat
1.e-5
alpha
1.
alphac
1.e3
***return
Z-set — Non-linear material
& structure analysis suite
12.5
***behavior crisfield debonding
***behavior crisfield debonding
Description:
This behavior3 is used for the special problem of interface debonding. See the command
**create_interface_elements (and similar) in the Z-set user manual on how to insert
cohesive elements in the mesh. The Crisfield model4 is described through a scalar variable λ,
which characterizes the relative crack opening:
r
2 2
u0T
u0N
1 hκi
huN i
k~
uT k
, κ=
=1−
,
(4)
+
− 1, η = 1 −
λ=
u0N
u0T
η 1 + hκi
δN
δT
where hxi = x if x > 0 and hxi = 0 if x ≤ 0. With respect to the interface normal ~n,
~uN = (~u · ~n) ~n ≡ uN ~n and ~uT ≡ ~u − u~N denote, respectively, the normal and shear opening
displacements and δN and δT the corresponding maximum allowable values of their norms.
The parameters u0N and u0T denote, respectively, the opening displacements corresponding
to the maximum cohesive traction of the normal and shear components. The damage variable
λmax , which is the maximum value of λ reached up until the current instant, increases from 0
and uδ0T
have to be the
(no damage) to 1 (for a broken element). In this model the ratios uδ0N
N
T
same. The normal and shear components of the cohesive traction T~ , i.e. T~N = T~ · ~n ~n ≡
TN ~n and T~T = T~ − T~N , are defined by:
TN =
uN
T
F (λmax ), T~T = α u~u0T
F (λmax ), F (λ) = σmax (1 − λ) ,
u0N
(5)
with α a constant representing the relative magnitude of kT~T k with respect to TN , and σmax
the maximum stress allowable by the element. For the compressive case, where uN < 0, the
normal component of the traction is modified to
TN
= αc
uN
F (0),
u0N
(6)
with αc a penalization factor. In the literature, αc usually is at least 10α. Figure 5 illustrates
the typical response of the cohesive zone model under specific loads, for the parameters as
given in the example.
Syntax:
***behavior crisfield_debonding
sigmax σmax
u0n
U0N
u0t
U0T
deltan δn
deltat δt
alpha
α
alphac αc
. . . continued
3
this behavior is Z-set specific, and therefore does not apply for Z-mat for other codes
Alfano G. and Crisfield M. A., ”Finite element interface models for the delamination analysis of laminated
composites: mechanical and computational issues.”, Int. J. Numer. Meth. Engng. 50 (2001), 1701-1736.
4
12.6
Z-set — Non-linear material
& structure analysis suite
***behavior crisfield debonding
16.0
125
14.0
100
12.0
75
50
T [MPa]
u [µm]
10.0
8.0
6.0
4.0
25
0
-25
-50
2.0
-75
0.0
-100
-2.0
0.0
1.0
2.0
3.0
4.0
-125
0.0
5.0
1.0
2.0
time [s]
3.0
4.0
5.0
time [s]
1.0
125
100
0.8
75
50
λmax
T [MPa]
0.6
0.4
25
0
-25
-50
0.2
-75
-100
0.0
0.0
1.0
2.0
3.0
4.0
time [s]
5.0
-125
-2.0
0.0
2.0
4.0
6.0 8.0 10.0 12.0 14.0 16.0
u [µm]
Figure 2: Example in two dimensions of the evolution of the cohesive traction as a function
of the opening displacement, for two different loading cases: uN (t) with uT (t) = 0 (thick red
curves), and uT (t) with uN (t) = 0 (thin green curves). Top left: applied load u(t). Note: for
2 ≤ t ≤ 4 s, the applied loading becomes negative (but the response uN remains 0 because
of an implicit non-penetration condition). Top right: response T (t). Bottom left: λmax (t).
Bottom right: u(t) vs T (t).
Example:
***behavior crisfield_debonding
sigmax 100.
u0n
1.e-6
u0t
1.e-6
deltan
1.e-5
deltat
1.e-5
alpha
1.
alphac
1.e3
***return
Z-set — Non-linear material
& structure analysis suite
12.7
***behavior chaboche debonding
***behavior chaboche debonding
Description:
This behavior5 is used for the special problem of interface debonding. See the command
**create_interface_elements (and similar) in the Z-set user manual on how to insert
cohesive elements in the mesh. The Chaboche model6 is described through a scalar variable
λ which characterizes the relative crack opening:
s
huN i 2
k~uT k 2
+
,
(7)
λ=
δN
δT
where hxi = x if x > 0 and hxi = 0 if x ≤ 0. With respect to the interface normal ~n,
uN = (~u · ~n) ~n ≡ uN ~n and ~uT ≡ ~u − ~uN denote, respectively, the normal and shear opening
displacements and δN and δT the corresponding maximum allowable values of their norms.
The damage variable λmax , which is the maximum value of λ reached up until the current
instant, increases from 0 (no damage) to 1 (for a broken
element). The normal and shear
~
~
~
components of the cohesive traction T , i.e. TN = T · ~n ~n ≡ TN ~n and T~T = T~ − T~N , are
defined by:
TN =
1
27
uN
F (λmax ) , T~T = α ~uδTT F (λmax ) , F (λ) = Kσmax λ n −1 (1 − λ)
δN
4
(8)
with α a constant representing the relative magnitude of kT~T k with respect to TN , σmax the
maximum stress allowable by the element and K and n model parameters. When λ <1.e-8,
the finite values of the cohesive traction and the consistent matrix are guaranteed by the
parameters df_0 and dfdl_0. For the compressive case, where uN < 0, the normal component
is modified to
uN
F (0),
(9)
TN = αc
δN
with αc a penalization factor. In the literature, αc usually is at least 10α. Figures 3 illustrates
the typical response of the cohesive zone model under specific loads, with the parameters as
given in the example.
Syntax:
***behavior chaboche_debonding
sigmax σmax
deltan δn
deltat δt
alpha
α
alphac αc
n
n
K
K
df_0
value
dfdl_0 value
. . . continued
5
this behavior is Z-set specific, and therefore does not apply for Z-mat for other codes
Caliez M., ”Approche locale pour la simulation de l’écaillage des barrières thermiques EBPVD”, Ph.D.
thesis, Ecole Nationale Supérieure des Mines de Paris, 2001
6
12.8
Z-set — Non-linear material
& structure analysis suite
***behavior chaboche debonding
16.0
125
14.0
100
12.0
75
50
T [MPa]
u [µm]
10.0
8.0
6.0
4.0
25
0
-25
-50
2.0
-75
0.0
-100
-2.0
0.0
1.0
2.0
3.0
4.0
-125
0.0
5.0
1.0
2.0
time [s]
3.0
4.0
5.0
time [s]
1.0
125
100
0.8
75
50
λmax
T [MPa]
0.6
0.4
25
0
-25
-50
0.2
-75
-100
0.0
0.0
1.0
2.0
3.0
4.0
time [s]
5.0
-125
-2.0
0.0
2.0
4.0
6.0 8.0 10.0 12.0 14.0 16.0
u [µm]
Figure 3: Example in two dimensions of the evolution of the cohesive traction as a function
of the opening displacement, for two different loading cases: uN (t) with uT (t) = 0 (thick red
curves), and uT (t) with uN (t) = 0 (thin green curves). Top left: applied load u(t). Note: for
2 ≤ t ≤ 4 s, the applied loading becomes negative (but the response uN remains 0 because
of an implicit non-penetration condition). Top right: response T (t). Bottom left: λmax (t).
Bottom right: u(t) vs T (t).
Example:
***behavior chaboche_debonding
sigmax
100.
deltan
1.e-5
deltat
1.e-5
alpha
1.
alphac
1.e3
n
22.6
K
0.18
df_0
44260000.
dfdl_0
0.
***return
Z-set — Non-linear material
& structure analysis suite
12.9
***behavior generalized debonding
***behavior generalized debonding
Description:
This behavior7 is used for the special problem of interface debonding. See the command
**create_interface_elements (and similar) in the Z-set user manual on how to insert
cohesive elements in the mesh. The Generalized model8 is based on the consideration of the
characterization of interfacial properties and permits to use three different shapes for the
cohesive law (bi-linear, tri-linear and trapezoidal). This model is described through a scalar
variable λ, which characterizes the relative crack opening:
∗
ασ δ0 hδf − δi
δ − ασ δ0
hδ − δ ∗ i
hδ − δ ∗ i
hδ − δ0 i
+
1+
,
(10)
λ= 1−
(δ − δ ∗ )
δ (δ ∗ − δ0 )
(δ − δ ∗ )
δ (δ ∗ − δf )
where hxi = x if x > 0 and hxi = 0 if x ≤ 0. The parameter δ is the relative displacement
defined as
q
δ = huN i2 + k~uT k2 .
(11)
With respect to the interface normal ~n, ~uN = uN ~n and ~uT ≡ ~u − u~N denote, respectively,
the normal and shear opening displacements. The parameters δ0 and δf are respectively the
relative displacement associated with the interfacial strength τ0 and the one attained when
the energy release rate is equal to the fracture toughness GC . It should be noted that the
evolution of τ0 (resp. GC ), as a function of the mixed-mode ratio, is defined by the stress
criterion (resp. the propagation law). Finally, the parameter δ ∗ is associated with the stress τ ∗
attained at the end of the first part of the damage process (Figure 4). These both parameters
depend on the shape of the cohesive law which is defined using two shape parameters : ασ
and αδ determined by
ασ =
∗
τ∗
0)
.
, αδ = (δδ −δ
( f −δ0 )
τ0
(12)
Figure 4: Shape of the tri-linear model for a constant mixed-mode ratio in the (δ, T ) plane .
The damage variable λmax , which is the maximum value of λ reached up until the current
instant, increases from 0 (no damage) to 1 (for a broken element). The normal and shear
7
this behavior is Z-set specific, and therefore does not apply for Z-mat for other codes
Vandellos T., Huchette C. and Carrere N., ”Proposition of a framework for the development of a cohesive
zone model adapted to Carbon-Fiber Reinforced Plastic laminated composites.”, Comp. Struct. 105 (2013),
199-206.
8
12.10
Z-set — Non-linear material
& structure analysis suite
***behavior generalized debonding
components of the cohesive traction T~ , i.e. T~N = T~ · ~n ~n ≡ TN ~n and T~T = T~ − T~N , are
defined by:
TN = KuN (1 − λ) , T~T = K~uT (1 − λ) ,
(13)
where K is the interfacial stiffness. For the compressive case, where uN < 0, the normal
component of the traction is modified to
TN
= αc KuN ,
(14)
with αc a penalization factor. Figure 5 illustrates the typical response of the cohesive zone
model under specific loads, for the parameters as given in the example.
Remarks:
• the parameter αδ is not required with the bi-linear shape;
• the parameter ασ has to be defined only with the tri-linear law;
• two stress criteria (puissance and reinforcement) and three propagation laws (puissance,
benzeggagh and vandellos) are available;
• the option *turon is required for applying the modification proposed by Turon9 .
. . . continued
9
Turon A., Camanho P.P., Costa J. and Renart J., ”Accurate simulation of delamination growth under
mixed-mode loading using cohesive elements: Definition of interlaminar strengths and elastic stiffness.”, Comp.
Struct. 92 (2010), 1857-1864.
Z-set — Non-linear material
& structure analysis suite
12.11
***behavior generalized debonding
Syntax:
***behavior generalized_cohesive_zone
*strength
Zt
Zt
Sc
Sc
*toughness
G1c
GIC
G2c
GIIC
*stiffness
K
K
alphac
αc
*propagation propagation law (puissance or benzeggagh or vandellos)
n
n
*initiation stress criterion (puissance or reinforcement)
*softening
shape of the cohesive law (bi linear or tri linear or trapezoidal)
alpha_delta αδ
alpha_sigma ασ
*turon
(optional)
Example:
***behavior generalized_cohesive_zone
*strength
Zt
100.
Sc
100.
*toughness
G1c
0.0005
G2c
0.0005
*stiffness
K
1.e8
alphac
1.e3
*propagation benzeggagh
n 1.142
*initiation puissance
*softening tri_linear
alpha_delta 0.4
alpha_sigma 0.8
***return
. . . continued
12.12
Z-set — Non-linear material
& structure analysis suite
***behavior generalized debonding
Figure 5: Example in two dimensions of the evolution of the cohesive traction as a function
of the opening displacement, for two different loading cases: uN (t) with uT (t) = 0 (thick red
curves), and uT (t) with uN (t) = 0 (thin green curves). Top left: applied load u(t). Note: for
2 ≤ t ≤ 4 s, the applied loading becomes negative (but the response uN remains 0 because
of an implicit non-penetration condition). Top right: response T (t). Bottom left: λmax (t).
Bottom right: u(t) vs T (t).
Z-set — Non-linear material
& structure analysis suite
12.13
***behavior diffusion
***behavior diffusion
Description:
Diffusion behavior is analogous to the stationary thermal behaviors.
J~ = D∇C
Syntax:
***behavior diffusion
**inter_phase_diffusion val
**phase
D COEFFICIENT
comp_max val
comp_min val
Stored Variables:
prefix
size
description
default
dC
J
C
phase
Deff
V
V
S
S
S
gradient of concentration
flux of concentration
the concentration
phase id (in input order)
effective D at point
yes
yes
yes
yes
yes
Example:
***behavior diffusion
**inter_phase_diffusion 8.e-10
**phase oxide
D 1.e-7
comp_min 1200.0
**phase metal
D 5.e-9
comp_max 400.
***return
12.14
Z-set — Non-linear material
& structure analysis suite
***behavior linear spring
***behavior linear spring
Description:
This behavior class is used to specify linear response for spring or truss elements (element
formulation linear_spring or spr).
F = kU
Where F is the spring tension, and U is the axial displacement of the spring. Note that
k can vary with any external parameter, but not in relation to U .
Syntax:
***behavior linear_spring
k coef
By default there is no thermal dilatation in the model.
Example:
A simple spring definition:
***behavior linear_spring
k 1.e4
***return
Z-set — Non-linear material
& structure analysis suite
12.15
***behavior thermal
***behavior thermal
Description:
This is the basic thermal behavior. The conductivity is determined using a CONDUCTIVITY
object. Transient thermal problems require that a coefficient capacity be defined as well
(global behavior coefficient, see page 9.11).
The behavior calculates the heat flux from the gradient in temperature:
~q = k∇T
The enthalpy term is
Z T
dH =
Cp (τ )dτ
T0
where the integration will be carried out numerically if the heat capacity depends on the
temperature.
Syntax:
***behavior thermal
**conductivity CONDUCTIVITY
**coefficient
capacity val
Note that capacity and the coefficients in the conductivity object can be a variable coefficients (e.g. tables, functions, etc) of the temperature variable. This gives the model the
possibility of including considerable non-linear effects.
Stored Variables:
The grad variable is the gradient of temperature, and the flux is the heat flux.
prefix
size
dT
V
q
V
temperature S
description
default
gradient of temperature
heat flux
the temperature at the Gauss point
yes
yes
yes
Example:
Two examples of thermal behavior follow:
***behavior thermal
**conductivity isotropic
k 48.822
**coefficient
capacity 4.8168e6
***return
***behavior thermal
**conductivity isotropic
k temperature
0.1
0.0
0.6
500.0
***return
12.16
Z-set — Non-linear material
& structure analysis suite
***behavior variable friction
***behavior variable friction
Description:
This behavior10 is used to specify a variable friction in certain contact models (see the command ***contact **zone *variable_friction in the Z-set user manual). The behavior
specifies a friction coefficient µ that may depend on the current total relative sliding us of
the contact surfaces, and/or the cumulative relative sliding ucum . This is analogous to the
role played by the current inelastic deformation and the cumulative plastic deformation in
isotropic and kinematic hardening in conventional material models. The analogy is as follows:
the total friction coefficient µ is written as µ = µ0 + µR (ucum ) + µX (ucum , us ), where µ0 plays
the role analogous to the elastic limit, µR to the isotropic hardening and µX to the kinematic
hardening.
Syntax:
***behavior variable_friction
*isotropic constant | linear | nonlinear
fric1 µ0
[ fric3 Q ]
[ b1
b1 ]
[ betai βi ]
[ *kinematic linear | nonlinear ]
[ fric2 C ]
[ b
b ]
[ betak βk ]
Possible choices are:
*isotropic constant takes one parameter µ0 .
*isotropic linear takes two parameters µ0 and Q to give µR = Qucum .
*isotropic nonlinear
takes
four parameters µ0 , Q, b1 and βi to give µR =
Q 1. − exp[−(b1 ucum )βi ] .
*kinematic linear takes one parameter C to give µX = C|us |.
*kinematic nonlinear
three parameters C, b and βk , to give µX
takes
|us |
β
k
C 1. − exp[−(bucum ) ] umax . Here umax is...
=
Currently, the behavior parameters cannot depend on any external parameter (such as
temperature). Also note that not all contact models support a variable friction coefficient.
If so, it is indicated in the description of each contact model.
Example:
10
this behavior is Z-set specific, and therefore does not apply for Z-mat for other codes
Z-set — Non-linear material
& structure analysis suite
12.17
***behavior variable friction
***behavior variable_friction
*kinematic nonlinear
fric2 3.
b 4.e-2
betak 1.8
*isotropic nonlinear
fric1 0.18
fric3 3.
b1 5.e-2
***return
12.18
Z-set — Non-linear material
& structure analysis suite
Chapter 13
Material Components
Z-set — Non-linear material
& structure analysis suite
13.1
13.2
Z-set — Non-linear material
& structure analysis suite
<ANISOTROPIC DAMAGE>
<ANISOTROPIC DAMAGE>
Description:
These behavior objects are specific to the elastic and viscoplastic anisotropic damage models.
They allow different types of damage variable to be used easily, and in combination. All the
coefficients and functions entered here are local to the each damage object, and must therefore
be input for all applicable objects even if the coefficients have the same meaning.
Syntax:
**damage [ type ]
[ *use_e_bar
]
[ *dont_use_e_bar
]
[ *eta <COEFFICIENT> ] % default 1.0
[ *K_coeffs V1 V2 V3 V4 V5 V6 V7 V8 ]
[ *g
<G_FUNCTION> ]
where type may be from the following types:
scalar
DESCRIPTION
scalar variables with fixed direction
elastic tensorial
tensorial variable
rate tensorial
tensorial variable
CODE
*use e bar Makes the calculation of Y or Y use the tensor in place of el .
*dont use e bar
default.
use el to calculate Y or Y in place of . This option is currently the
*K coeffs coefficients are entered to construct the tensor K use the compliance matrix
(S = D−1
el ) and are as follows:


S11 V1 S12 V4 S13 V5


S22 V2 S23 V6




S
V
33
3

H0 = 


S44 V7




sym
S55 V8
S66 V9
note that the tensors are stored as
[S1 . . . S6 ] = [S11 S22 S33 S12 S23 S31 ]
K = Del : H0 : Del
The factors for V are entered in order after the *K_coeffs keyword is entered. See the
behavior descriptions for examples.
Z-set — Non-linear material
& structure analysis suite
13.3
<ANISOTROPIC DAMAGE>
*K matrix form for the K. This input is entered as an ELASTICITY object.
Only one definition for the matrix K0i may be given with the K_coeffs or K. The coefficients
correspond directly to the associate terms marked with a subscript i.
13.4
Z-set — Non-linear material
& structure analysis suite
<ANISOTROPIC DAMAGE>
<ANISOTROPIC DAMAGE> scalar
Description:
Each scalar damage variable is defined using a “damage direction” input by the user, ~ni , and
an effective fourth order “damage effects” tensor K0i . The modulus modification induced by
a scalar damage variable i is the following:
0
Ceff
i )Ni : K0i : Ni )
i = δi (Ki − ηi h(−¯
where δi is the scalar internal damage variable. The above uses the following terms to measure
the degree of opening strain:
¯i
Ni
= (~ni ⊗ ~ni ) : ( − c )
= (~ni ⊗ ~ni ) ⊗ (~ni ⊗ ~ni )
(1)
The material coefficient η is used to describe the influence of closing on the damage
softening.
**damage scalar
std options
[ *n
<VECTOR> ]
% orientation
n the directions of the axes of damage for the isotropic variables entered as a vector.
Example:
Please see the examples in the anisotropic behaviors aniso_damage and visco_aniso_damage.
Z-set — Non-linear material
& structure analysis suite
13.5
<ANISOTROPIC DAMAGE>
<ANISOTROPIC DAMAGE> elastic tensorial
Description:
This model using a 2nd order tensorial damage variable is calculated as follows:
0
Ceff
j = Dj : Kj −
3
X
ηj h(−¯
n )Pn : (Dj : K0j ) : Pn )
n=0
where the summation over n is over the three principal directions of the strain tensor, and
Dj is a fourth order tensor constructed based on the second order damage variable dj . This
will use the following calculations based on the principal strain eigenvectors p~n :
¯n
= (~
pn ⊗ p~n ) : ( − c )
Pn = (~
pn ⊗ p~n ) ⊗ (~
pn ⊗ p~n )
(2)
and the fourth order tensors Dj are constructed their tensorial variables as:
Dj = (α − β)(1 ⊗ dj )s +
13.6
β
2
((1⊗d)s + 1⊗d)s )
Z-set — Non-linear material
& structure analysis suite
<COEFFICIENT>
<COEFFICIENT>
Description:
Coefficient objects are used to enter the values for material coefficients. Each behavior object
may have any number of coefficients to parameterize the models for different materials. The
coefficients are themselves behavior objects however, and therefore have a standard format
for entry. Coefficients also provide the means to add external parameter or local material
variable dependencies.
Creating user-dependency on the internal variables may however significantly alter the
material model, and invalidate integration methods. The only sure use for arbitrary dependencies on the integrated variables is with Runge-Kutta integration, and a time-dependent
flow law. The tangent matrix may also be altered by these dependencies. Coefficients which
are a function of the external parameters are however robustly implemented, and are valid
for all integration methods.
Syntax:
Supposing that a material model coefficient C is required, a syntax similar to the following
will be given:
C COEFFICIENT
where the term COEFFICIENT is to be replaced with a coefficient definition. The replacement syntax using the coefficient objects is given below:
C [ type ]
parameters
equivalence
DESCRIPTION
coefficient which is the value of a variable
=
same as equivalence
function
coefficient which is a function of variables (see chapter
Functions)
equivalence
coefficient which is the value of variables
step wise
step-wise tabular values
default
tabular description in terms of variables
CODE
Example:
The first example is for a tabular coefficient. This example has an equivalent stress, sigeq,
as a function of the cumulated flow, and the temperature. Note any number of variables may
be given. If the parameter values are out of range from what is given in the table, and error
will result.
sigeq epcum temperature
400.0 0.0
20.0
350.0
"
120.0
290.0
"
200.0
Z-set — Non-linear material
& structure analysis suite
13.7
<COEFFICIENT>
450.0
410.0
330.0
500.0
460.0
400.0
0.002
"
"
0.01
"
"
20.0
120.0.
200.0
20.0
120.0
200.0
The second example uses functions to describe the coefficient value. Function syntax is
described more fully in the Functions chapter. Note parameter names must be on the same
line as the function declaration.
young function 230. + 1.e-2*temperature + temperature^2.0;
The last example assumes a parameter is calculated in the material law called Bpa1v. This
example will set the coefficient value to the value of that parameter at all times.
*kinematic nonlinear
A1 40.0
Bp = Bpa1v
13.8
Z-set — Non-linear material
& structure analysis suite
<COEFFICIENT MATRIX>
<COEFFICIENT MATRIX>
Description:
This object is used to enter coefficients for a desired form of 4th order coefficient matrix. The
use of these matrices is often as an elasticity matrix, so we sometimes refer to the object as
ELASTICITY.
Syntax:
These objects take a list of coefficient declarations after the type keyword is given. There are
no explicit restrictions on the types of coefficients or their dependencies. The different types
(keywords) available are the following:
diagonal
DESCRIPTION
1 or dim coefficients
isotropic
2 coefficients
cubic
3 coefficients
transverse
5 coefficients
orthotropic
9 coefficients
anisotropic
21 coefficients
CODE
The default type depends on its application. For elasticity objects this is isotropic. The
anisotropic model represents the most general elasticity model. Coefficient names follow
the same convention to describe components of the coefficient matrix. This sets the coefficient
name for a component Cijkl to Cijkl. For example, a component y1212 is named y1212.
When using elasticity objects, special care has to be taken with respect to the nomenclature
of the components. When representing a tensor of elasticity in 6×6 form, the Voigt convention
is assumed, but with one important difference: the order of storage is
11 → 1, 22 → 2, 33 → 3, 12 → 4, 23 → 5, 31 → 6 (Zebulon)
instead of the usual
11 → 1, 22 → 2, 33 → 3, 23 → 4, 31 → 5, 12 → 6 (Voigt)
For instance, a Voigt coefficient c66 becomes c44 or y1212 in Zmat/Zebulon.
The matrix forms are given below:
isotropic The isotropic model allows several methods of entering the coefficients for
convenience. The primary method gives the Young’s modulus (E) and Poisson coefficient
(ν) which correspond to the names young and poisson respectively.
Dij =
E
νE
+
δij
1+ν
(1 − 2ν)(1 + ν)
An alternate definition defines the shear modulus (G or µ) and the bulk modulus (K or
κ). These correspond to the coefficient names G or mu, and K or kappa.
Z-set — Non-linear material
& structure analysis suite
13.9
<COEFFICIENT MATRIX>
transverse Transverse isotropy has two directions which are equivalent. There are two
ways to specify transverse elasticity parameters.
Elasticity modules
Direction 1 is the longitudinal one and directions 2 and 3 are the two transverse
ones. The elasticity is specified with 5 classical parameters : El, Et, nult, nutt
and Glt. They define the Hooke relation such as :

 


1
−nult
−nult


ε11
0
0
0
σ
 El

11
Et
Et


 






 




−nutt
1


σ
0
0
0
 ε22  


22 

Et
Et

 






 




1


 ε33  


σ
0
0
0
33 

Et

=






 





1+nutt

 ε23  


0
0 
σ23 
Et

 






 





1

 ε13  


sym
0 
σ13 
2Glt


 








1 
ε12
σ12
2Glt
An old parameter named glt (instead of Glt) is defined as :
σ12 = glt ε12
whereas Glt is defined as :
σ12 = 2 Glt ε12
This old parameter is still available for compatibility with old data sets.
Matrix coefficients
The program allows the user to specify which terms in the elastic matrix will be
set equal, with a tag input after the transverse keyword. It is best to explain this
syntax through an example:


c11 c12 c13
0
0
0






c11 c13
0
0
0








0
0
0
c


33


D=

1


(c
−
c
)
0
0


11
12


2






sym
c
0
55




c55
which is created by **elasticity transverse and the coefficients c11, c12, c13,
c33 and c55 (or the corresponding y1111 etc.).
For the moment it is only possible to work with c11 and c22 equal instead of, for
instance, equal c22 and c33.
The following relations are also given for reference in the case of c11 = c22 :
c11
c12
c13
13.10
2 )E
(1 − nνzx
x
AB
2 )E
(νxy + nνzx
x
=
AB
νzx Ex
=
B
=
Z-set — Non-linear material
& structure analysis suite
<COEFFICIENT MATRIX>
c33
c44
c55
cxy
n
(1 − νxy )Ez
B
= 21 (C11 − C12 )
= C66 = Gxy = Gxz
Ex
=
2(1 + νxy )
= Ex /Ez
A = 1 + νxy
=
2
B = 1 − νxy − 2nνzx
cubic

y1111 y1122 y1122
0
0




y
0
0
y

1111
1122



y
0
0

1111
D=


y1212
0





sym
y
1212


0
0
0
0
0
y1212




















orthotropic

0
0

y1111 y1122 y3311


y2222 y2233
0
0




0
0
y

3333
D=


y
0

1212




sym
y2323


0
0
0
0
0
y3131




















An alternative naming uses c11, c22, c33, c44, c55, c66, c12, c13, and c23 (or their
symmetric counterparts).
anisotropic

y1111 y1122 y1133




y2222 y2233




y3333

D=







sym


y1112
y2212
y3312
y1212
y1123
y2223
y3323
y1223
y2323

y1131 


y2231 



y3331 



y1231 




y2331 


y3131
Example:
**elasticity isotropic
young
200000.
poisson 0.3
**elasticity cubic
y1111 162321.0
y1122 78075.0
Z-set — Non-linear material
& structure analysis suite
13.11
<COEFFICIENT MATRIX>
y1212 110615.0
**elasticity
young
T
200000. 0.
200000. 1000.
poisson 0.3
**elasticity transverse
El
115000.
Et
8500.
Glt
4500.
nult
0.32
nutt
0.40
13.12
Z-set — Non-linear material
& structure analysis suite
<CONDUCTIVITY>
<CONDUCTIVITY>
Description:
Thermal conductivity is available in istropic and anisotropic forms. Conductivity acts much
like the elasticity matrix does in mechanical problems, except the thermal behavior accounts
for direct coefficient variations with respect to the temperature.
~q = k∇T
Syntax:
**conductivity type
...
The following forms are available:
isotropic


k 0 0
k = 0 k 0
0 0 k
where k is the only coefficient.
anisotropic


k1 0 0
k =  0 k2 0 
0 0 k3
where the coefficients k1, k2, and k3 are to be entered.
Z-set — Non-linear material
& structure analysis suite
13.13
<CRITERION>
<CRITERION>
Description:
The criterion object is used for specifying the calculation of an equivalent stress to be used
in plasticity and viscoplasticity behavior models (for each potential in a gen_evp material).
The criterion may also be used for determining the flow direction and in hardening evolution
models depending on the application.
Syntax:
The syntax to specify a criterion object will consist of giving the keyword for a particular
criterion desired, followed by a list of appropriate coefficients which are dependent on the
model chosen. The possible criterion types are:
mises
DESCRIPTION
classical von Mises criterion
2M1C
criterion 2M1C (only for potential mises 2m1c)
hill
Hill criterion
ratio
classical criterion useful as viscoplastic overstress function
karafillis boyce
Karafillis Boyce criterion
full hill
Hill criterion
tensile mises
acts as von Mises criterion
nouailhas
macroscopic model developed by ONERA
modified nouailhas
modified nouailhas model
linear drucker prager
non-associated drucker-prager criterion
anisotropic
classical energy equivalent form due to von Mises
cast iron
criterion which provides asymmetric tension/compression
bron
modes anisotropic behavior in the criterion and flow direction
unsym
non-symmetric deviatoric / hydrostatic criterion
tresca
Tresca criterion
CODE
Some expressions which will be used are:
r
3
1
J2 (σ̂) =
ŝ : ŝ
sij = σij − σii δij
2
3
where σ̂ is the effective stress which may be displaced by a kinematic back stress:
X
σ̂ = σ −
X
The true form of this stress will be determined by the potential.
Stored Variables:
Criterion objects are not allowed to have any variables.
13.14
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
<CRITERION> anisotropic
Description:
This criterion is the classical energy equivalent form due to von Mises. Here the effective
stress is calculated based on the J2 stress invariant:
q
fcr = 32 (s − X) : M : (s − X) − R
where s is the deviatoric component of stress.
**<cmd> anisotropic matrix-type
coefficients
The input matrix-type indicates one of the COEFFICIENT MATRIX type keywords, and the
coefficients will depend on the matrix chosen (e.g. diagonal, orthotropic, etc).
Example:
brief example using the anisotropic criterion is:
*criterion anisotropic orthotropic
c11 .44667
c22 .58
c33 .666
c44 0.
c55 1.7
c66 0.
c12 -.18
c23 -.4
c31 -.26667
The *criterion command indicates the application of the created criterion object in the
higher-level, so can generally be different depending on the application.
Z-set — Non-linear material
& structure analysis suite
13.15
<CRITERION>
<CRITERION> bron
Description:
The Bron criterion allows modeling of anisotropic behavior in the criterion and flow directions.
The proposed yield function is defined by an equivalent stress:
1/a
σ̄ = α(σ̄ 1 )a + (1 − α)(σ̄ 2 )a
1/bk
σ̄ k = ψ k
b1 b1 b1 1 1
ψ1 =
S2 − S31 + S31 − S11 + S11 − S21 2
2
2 2 2 3b
S12 b + S22 b + S32 b
ψ 2 = b2
2 +2
k
where Si=1−3
are the principal values of a modified stress deviator ∼sk defined as follows:
k
sk = L
:σ
∼
∼
∼
 k

(c2 + ck3 )/3
−ck3 /3
−ck2 /3
0 0 0
 −ck3 /3
(ck3 + ck1 )/3
−ck1 /3
0 0 0


k
k
k
k
 −c2 /3
−c1 /3
(c1 + c2 )/3 0 0 0 
k


L
=
k
∼

0
0
0
c
0
0
∼
4


k

0
0
0
0 c5 0 
0
0
0
0 0 ck6
∼
a, b1 , b2 and α are four material parameters that influence the shape of the yield surface
but not its anisotropy which is only controlled by ck=1−2
i=1−6 . Thereby, the yield function has 16
parameters. To ensure convexity and derivability the following conditions are required: a ≥ 1
and bk ≥ 2. No restriction applies to the cki coefficients; in particular they can be negative.
1 = L2 and a = b1 = b2 corresponds to the yield function of
The particular case where L
∼
∼
∼
∼
Karafillis and Boyce (1993) and the case where α = 1 corresponds to the yield function of
Barlat et al. (1991). Finally, when α = 1 and c1i = 1, it amounts to von Mises yield function
if b1 = 2 or 4 and to Tresca yield function if b1 = 1 or + ∞. If cki = 1, the resulting yield
function is isotropic as it only depends on the eigenvalues of σ
.
∼
Example:
**porous_potential
*porous_criterion modified_rousselier ...
*isotropic function ...
*flow plasticity
*shear_anisotropy bron
a 2.2 alpha 0.60
b1 10.3
c11 0.58 c12 1.35 c13 1.14 c14 1.23 c15 1.35 c16 1.57
b2 13.1
c21 2.07 c22 0.20 c23 0.33 c24 0.85 c25 1.31 c26 0.59
13.16
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
<CRITERION> cast iron
Description:
This criterion provides assymetric tension/compression behavior which is suggested for cast
iron modeling [Hjel94,Jose95]. In general this model would be used with behavior gen_evp
potential gen_evp2 and includes one or more non_symmetric kinematic hardening terms
(and possibly others as well).
You can have either isotropic hardening scaling both the compressive and tensile behavior
(with a given ratio relating them), or use isotropic hardening in only the tensile, or only the
compressive with extra coefficients given for the non-hardening one. Probably the calling
application requires that an R be given, so defining both Rt and Rc is not allowed.
Rc
= ratio R
Rt = R
if ratio is entered
Rc
= Rc
Rt = R
if Rc is entered
Rc
=R
Rt = Rt
if Rt is entered
The different tension and compression yield values Rc and Rt are applied to two distinct
different yield functions as follows.
F1
F2
1/2
: s + Tr(σ)(Rc − Rt )
− (Rc Rt )1/2
1/2
= 32 s : s
− Rc
=
3
2s
The transition from the tension
P yield F1 to the compression yield F2 is determined if the
trace of the effective stress σ − X is less than −Rc . See the section for the non_symmetric
direct kinematic model on page 13.41 for a demonstration of the model’s effect.
Syntax:
The basic input syntax here is:
*criterion cast_iron
[ ratio value ]
[ Rc
value ]
[ Rt
value ]
[ associated 0 | 1 ]
Only one choice of ratio Rc or Rt may be given as discribed above. If the associated is
set to 1, the F1 yield function will be used all the time. The default is associated behavior.
Example:
*criterion cast_iron
ratio 3.
associated 1
Z-set — Non-linear material
& structure analysis suite
13.17
<CRITERION>
<CRITERION> hill
Description:
The Hill criterion allows modeling of anisotropic behavior in the criterion and flow directions.
fcr
3
=
σ̂ : Mhill : σ̂
2
1/2
−R
where Mhill is a diagonal matrix of independent coefficients. These coefficients are named
hill1 . . . hillN with N being 4 for two dimensional problems, and 6 for three dimensional.
There is no problem giving the complete 6 coefficients in a 2D problem however.
Example:
**potential gen_evp ev
*flow norton
n 4.0
K 500.
*criterion hill
hill1 1.
hill2 2.
hill3 3.
hill4 1.2
hill5 1.7
hill6 0.8
...
13.18
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
<CRITERION> karafillis boyce
Description:
Also known as the K-B criterion, the karafillis-boyce criterion was developed in 1993 at MIT.
This criterion is constructed by mixing two yield functions φ1 and φ2 . As shown in the
equations below φ1 represents a yield locus located between the Von Mises yield locus and the
Tresca yield locus and φ2 varies from von Mises to a theoretical upper bound as m changes
from 2 to ∞.
φ = (1 − c) φ1 + cφ2 = 2Y m
where
φ1
φ2
= |S1 − S2 |m + |S2 − S3 |m + |S3 − S1 |m
3m
(|S1 |m + |S2 |m + |S3 |m )
= m
2 −1+1
and Si are the principal values of the isotropic plasticity equivalent(IPE) stress tensor
as defined below, Y is the average yield stress in uniaxial tension, and L is a fourth order
tensorial operator which introduces material anisotropy.
S = L (σ − B)
Syntax:
The basic input sytax here is:
*criterion karafillis_boyce
coefficients c1, c2, c3 c4, c5, c6.
Example:
*criterion karafillis_boyce
m 2.0
c 0.0
c1 1.0
c2 1.0
c3 1.0
c4 1.0
c5 1.0
c6 1.0
Z-set — Non-linear material
& structure analysis suite
13.19
<CRITERION>
<CRITERION> linear drucker prager
Description:
This criterion is a non-associated criterion (flow direction is not the same as the normal to
the yield surface ∂∂f
σ . It is commonly used for soils and plastics.
β
friction angle (general coef)
ψ
dilatation angle (general coef)
K
ratio of triaxial yield in tension to triaxial yield in compression. 0.778 ≤ K ≤ 1
S
p
q
r3
= U : σ = σ + p1
= − 13 1 : σ
1/2
= 32 S : S = 92 S · S : S
(3)
t
1
1 r3
q
1+
− 1−
=
2
K
K q3
f
= t − ptanβ − d
= 1 − 31 tanβ σc
d
= 1 + 13 tanβ σt
√
3
d
=d≡
τ (1 + K −1 )
2
d
g
= t − ptanψ
n
=
τ = shear yield (cohesion)
(4)
1 ∂g
c ∂σ
c
c
13.20
σc is the compressive yield
σt is the tensile yield
= 1 − 13 tanψ
1
c
=
1 + 31 tanψ
K
5
5
4 1/2
=
+
−
2 2K 2 K
defined in compression yield
defined in tension yield
defined in pure shear (cohesion)
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
(5)
Note:
Since this criterion is not associated, it should only be used with behaviors and criterion which
accept a different criterion function from normal definition.
Syntax:
The criterion takes a few options and parameters. The following assumes that the criterion
is selected with a **criterion command as in gen_evp behaviors.
**criterion linear_drucker_prager
friction_angle
COEFFICIENT
dilatation_angle COEFFICIENT
K
COEFFICIENT
[ use_sigma_c ]
[ use_sigma_t ]
[ use_cohesion ]
Example:
**criterion linear_drucker_prager
friction_angle 20.0
dilatation_angle 20.0
K
.9
Z-set — Non-linear material
& structure analysis suite
13.21
<CRITERION>
<CRITERION> mises
Description:
This criterion is the classical energy equivalent form due to von Mises. Here the effective
stress is calculated based on the J2 stress invariant:
0.5
3
σ̂ : σ̂
−R
fcr =
2
n=
3 σ̂
2 fcr
Example:
**potential gen_evp ev
*flow plasticity
*criterion mises
*isotropic constant R0 300.
13.22
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
<CRITERION> modified nouailhas
Description:
This criterion modifies the nouailhas model above by basing the calculation of the I’s on
the actual tensors σ − X in place of the deviator, and includes an extra term:
fcr =
3
2 (I1
+ 2a2 I2 + 2a4 I4
2
+ 3a8 I8
3
4
− (a6 I6 )
1/12
−R
2
2
2
I1 = S11
+ S22
+ S33
I2 = S11 S22 + S22 S33 + S33 S11 ;
2
2
2
I4 = S12
+ S23
+ S31
I6 = S12 ∗ S23 ∗ S31
4
4
4
I8 = S12
+ S23
+ S31
with
S=σ−X
This criterion is fully implemented, such that it will work with any normal gen_evp or
reduced_plastic potential, with any otherwise valid integration. In 2D, the 23 and 31 terms
are taken as zero.
Z-set — Non-linear material
& structure analysis suite
13.23
<CRITERION>
<CRITERION> nouailhas
Description:
This criterion is a macroscopic model developed by ONERA for simulating anisotropic deformation of single crystals. The model does not accurately represent the individual slip
system deformations, or their interactions, but could be useful for efficiently modeling the
basic anisotropic behavior of a single crystal.
The criterion is written:
1/12
3
3 0
0 2
0
0 4
+ 3a8 I8 − (a6 I6 )
−R
fcr =
2 (I1 + 2a4 I4
2
2
2
I1 = S11
+ S22
+ S33
I2 = S11 S22 + S22 S33 + S33 S11 ;
2
2
2
I4 = S12
+ S23
+ S31
4
4
4
I8 = S12
+ S23
+ S31
with
S = σ 0 − X0
The criterion is associated:
n=
∂f ∂I2
∂f ∂I4
∂f ∂I8
∂f ∂I1
+
+
+
∂I1 ∂σ
∂I2 ∂σ
∂I4 ∂σ
∂I8 ∂σ
This criterion is fully implemented, such that it will work with any normal gen_evp or
reduced_plastic potential, with any otherwise valid integration. In 2D, the 23 and 31 terms
are taken as zero.
13.24
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
<CRITERION> ratio
Description:
This is a classical type criterion which may be useful for some users seeking compatibility
with other programs. Its usefulness is as a viscoplastic overstress function (i.e. when f > 0).
fcr =
J(σ̂)
−L
R
Note that R is a radius measure passed in by the controlling behavior or potential object,
and would normally be an isotropic hardening function. The effective stress σ̂ is likewise
determined by the behavior and can include back stress influences, etc.
Syntax:
The model accepts a single optional coefficient for L (named L) in order to allow the user
the option of having simply a ratio of f over R, which could be used to apply the isotropic
hardening law as a viscosity hardening. The default value of L is 1, and the user should note
that only values of 1 and 0 make sense.
Example:
The following example is from rate crit02.inp in test/Volume test/INP.
***behavior gen_evp
**elasticity isotropic
young 2.e9
poisson 0.4
**potential gen_evp ev
*criterion ratio
L
1.0
*flow norton
A 1000.
n 2.
*isotropic by_point
sigeq evcum
60.e6 0.
1.e8
2.
1.e8
200.
***return
Z-set — Non-linear material
& structure analysis suite
13.25
<CRITERION>
<CRITERION> tensile mises
Description:
ThisP
is a criterion which acts as a von Mises yield if the trace of the effective stress (e.g.
σ − Xi ) is greater than zero. If it is less than zero, the criterion will never yield.
3
σ̂ : σ̂
fcr =
2
fcr = 0
0.5
−R
T r(σ) ≥ 0
otherwise
(6)
13.26
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
<CRITERION> tresca
Description:
The plasticity threshold in Tresca criterion is linked to the maximum shear stress and is given
by:
1
max (|σi − σj |)
2 i6=j
The complete equation of the criterion can be written as:
f = max (|σi − σj |) − σs = 0
i6=j
Z-set — Non-linear material
& structure analysis suite
13.27
<CRITERION>
<CRITERION> unsym
Description:
This criterion models the equivalent stress and deformation with deviatoric and hydrostatic
terms. The criterion is written:
fcr = (1 − a)σeq + a Tr(σ) − R
This criterion accepts a single coefficient a which must be input on a separate line from the
unsym token. This criterion is also non-associated where the flow normal is written:
n=
3 σ0 − X
2 J2 (σ − X)
with the X term being the summation of back stresses. Often this criterion is used in conjunction with the ziegler kinematic hardening options [Croi92], but this is not necessarily
so.
13.28
Z-set — Non-linear material
& structure analysis suite
<CRITERION>
<CRITERION> 2M1C
Description:
For this case, the criterion is calculated in two parts using the following form:
q
fcr = f12 + f22 − R
fi = J2 (Ai σ − Xi )
The coefficients ga1 and ga2 exist in order to define the localization parameters Ai .
Z-set — Non-linear material
& structure analysis suite
13.29
<CRYSTAL KINEMATIC>
<CRYSTAL KINEMATIC>
Description:
Because the crystal potential models takes into account the strains on the individual slip
planes, kinematic hardening represents a linear measure of a slip offset distance Xi , or backstress, on the slip line. As described in the <POTENTIAL>Pcrystal section (see page 13.80),
the yield criterion on each slip plane i is of the form |τi − j Xij | − ri , where τi is the resolved
shear stress on slip plane i and ri is the sum of all isotropic hardenings on that slip plane.
The internal variables αi , to which the backstresses are associated, are thus scalar and
distinct from the tensorial kinematic hardening laws in the macroscopic potentials. They are
related to the backstresses Xi through Xi = x0 + Cαi with x0 and C parameters having the
dimensions of a stress. Trick: a large negative offset x0 might be used to model unidirectional
slip, provided that the same positive offset is added to the corresponding isotropic hardening.
These hardening laws will apply uniquely for the monocrystal potentials. Permissible
crystalline kinematic types are with their additional parameters are
13.30
linear
α̇i = γ̇
nonlinear
α̇i = γ̇ − DXi |γ̇| /C
nonlinear phi
α̇i = γ̇φ(vi ) − D (Xi − Xbar) |γ̇| /C with
φ(vi ) = (1 − phi) + phi e−delta vi
Z-set — Non-linear material
& structure analysis suite
<CRYSTAL ORIENTATION>
<CRYSTAL ORIENTATION>
Description:
This class is used to load in particular crystal slip systems into gen_evp crystal potentials (see
page 13.80), or in seperate crystal behaviors. The crystal orientation will define the number
of slip systems in the model, as well as the number of independent interaction parameters
between these slip systems.
The default orientations are as follows. For cubic crystals the (1 0 0)-axis is oriented
horizontally to the right (along the zebulon x-axis), (0 1 0) vertically upwards (along the
Zebulon y-axis) and (0 0 1) out of the paper towards the reader (along the Zebulon z-axis).
For hexagonal crystals, the directions in the basal plane are depicted in the following figure
(after [Tome and Kocks 1985]), and the c-axis coincides with the Z-set z-axis.
β =[1210]
y =[0110]
α =[2110]
x =[2110]
γ =[1120]
The crystal orientation will provide the localization tensors for the different slip systems i in
a crystal:
τi = m i : σ
with τi the resolved shear stress on slip system i and σ the ’macroscopic’ stress tensor. For
each slip system i in the particular crystal set, the second order Schmid tensor mi is defined
as follows:
i
h
mi = 12 ~ni ⊗ ~`i ,
with ~ni the normal of slip plane i, and ~`i the corresponding slip direction.
Syntax:
<creating-command> type
[ *interaction h-coefs ]
[ *c_over_a ] c-a value
In absence of the *interaction command, only self-hardening is taken into account (i.e.
h1=1. and the other coefficients are 0.). For some hexagonal crystals the ratio of the length
of the c-axis to the length of the a-axis is needed.
. . . continued
Z-set — Non-linear material
& structure analysis suite
13.31
<CRYSTAL ORIENTATION>
The following crystal systems are available:
cubic This orientation is for the 6 FCC cubic systems with up to 3 hardening coefficients.
(
(
(
(
(
(
0
0
1
1
0
0
0
0
0
0
1
1
1)[-1 1
1)[ 1 1
0)[ 0 1
0)[ 0-1
0)[-1 0
0)[ 1 0
0
0
1
1
1
1
]
]
]
]
]
]
octahedral This type of crystal is used for the 12 FCC octahedral slip systems, with up
to 6 hardening coefficients.
( 1 1 1 )[-1 0
( 1 1 1 )[ 0 -1
( 1 1 1 )[-1 1
( 1 -1 1 )[-1 0
( 1 -1 1 )[ 0 1
( 1 -1 1 )[ 1 1
(-1 1 1 )[ 0 -1
(-1 1 1 )[ 1 1
(-1 1 1 )[ 1 0
( 1 1 -1 )[-1 1
( 1 1 -1 )[ 1 0
( 1 1 -1 )[ 0 1
1
1
0
1
1
0
1
0
1
0
1
1
]
]
]
]
]
]
]
]
]
]
]
]
basal 3 HCP basal slip systems with 2 hardening coefficients.
( 1 -2 1
( 2 -1 -1
( 1 1 -2
0 )[ 0
0 )[ 0
0 )[ 0
0
0
0
0
0
0
1 ]
1 ]
1 ]
prismatic 3 HCP second-order prismatic systems with 2 hardening coefficients.
( 1
( 0
(-1
0 -1
1 -1
1 0
0 )[ 1 -2 1
0 )[ 2 -1 -1
0 )[ 1 1 -2
0 ]
0 ]
0 ]
pyramidal0 6 pyramidal systems - requires c over a to be entered.
( 1 0 -1
( 0 1 -1
(-1 1 0
(-1 0 1
( 0 -1 1
( 1 -1 0
1
1
1
1
1
1
)[
)[
)[
)[
)[
)[
1
2
1
1
2
1
-2
-1
1
-2
-1
1
1
-1
-2
1
-1
-2
0
0
0
0
0
0
]
]
]
]
]
]
. . . continued
13.32
Z-set — Non-linear material
& structure analysis suite
<CRYSTAL ORIENTATION>
pyramidal1 12 additional pyramidal systems - requires c over a to be entered.
( 1
( 1
( 0
( 0
(-1
(-1
(-1
(-1
( 0
( 0
( 1
( 1
0
0
1
1
1
1
0
0
-1
-1
-1
-1
-1
-1
-1
-1
0
0
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
1
1
1
)[ 2
)[ 1
)[ 1
)[-1
)[-1
)[-2
)[-2
)[-1
)[-1
)[ 1
)[ 1
)[ 2
-1
1
1
2
2
1
1
-1
-1
-2
-2
-1
-1
-2
-2
-1
-1
1
1
2
2
1
1
-1
-3
-3
-3
-3
-3
-3
-3
-3
-3
-3
-3
-3
]
]
]
]
]
]
]
]
]
]
]
]
( 1 0 -1
( 0 1 -1
(-1 1 0
(-1 0 1
( 0 -1 1
( 1 -1 0
1
1
1
1
1
1
)[
)[
)[
)[
)[
)[
-2
-1
1
-2
-1
1
1
-1
-2
1
-1
-2
0
0
0
0
0
0
]
]
]
]
]
]
pyramidalPi1
1
2
1
1
2
1
pyramidalPi2
(-1 2 -1
( 2 -1 -1
( 1 1 -2
( 1 -2 1
(-2 1 1
(-1 -1 2
2
2
2
2
2
2
)[-1 2 -1 -3 ]
)[ 2 -1 -1 -3 ]
)[ 1 1 -2 -3 ]
)[ 1 -2 1 -3 ]
)[-2 1 1 -3 ]
)[-1 -1 2 -3 ]
twinning 6 twinning systems in hexagonal crystals - requires c over a to be entered.
( 1 0 -1 -1 )[ 1 0 -1
( 0 1 -1 -1 )[ 0 1 -1
(-1 1 0 -1 )[-1 1 0
(-1 0 1 -1 )[-1 0 1
( 0 -1 1 -1 )[ 0 -1 1
( 1 -1 0 -1 )[ 1 -1 0
2
2
2
2
2
2
]
]
]
]
]
]
. . . continued
Z-set — Non-linear material
& structure analysis suite
13.33
<CRYSTAL ORIENTATION>
bcc112 12 additional slip systems for bcc crystals. These also apply to Shockley partial
dislocations in fcc crystals.
( 2 1 1 )[-1 1 1 ]
( 1 2 1 )[ 1 -1 1 ]
( 1 1 2 )[ 1 1 -1 ]
(-2 1 1 )[ 1 1 1 ]
( 1 -2 1 )[ 1 1 1 ]
( 1 1 -2 )[ 1 1 1 ]
( 2 -1 1 )[ 1 1 -1 ]
( 1 2 -1 )[-1 1 1 ]
(-1 1 2 )[ 1 -1 1 ]
( 2 1 -1 )[ 1 -1 1 ]
(-1 2 1 )[ 1 1 -1 ]
(-1 2 -1 )[-1 1 1 ]
climb cfc 4 climb directions on octahedral planes:
( 1 1 1
( 1 -1 1
(-1 1 1
( 1 1 -1
)[ 1 1 1
)[ 1 -1 1
)[-1 1 1
)[ 1 1 -1
]
]
]
]
climb cubic 3 climb directions on cubic planes:
( 1
( 0
( 0
0
1
0
0 )[ 1
0 )[ 0
1 )[ 0
0
1
0
0 ]
0 ]
1 ]
climb cubic1 The first of the cubic climb directions:
( 1
0
0 )[ 1
0
0 ]
climb cubic2 The second of the cubic climb directions:
( 0
1
0 )[ 0
1
0 ]
climb cubic3 The third of the cubic climb directions:
( 0
0
1 )[ 0
0
1 ]
plane Enter in a single slip system. The first system in octahedral could be entered in as:
**potential plane ( 1.
13.34
1.
1. ) (-1. 0.
1. )
Z-set — Non-linear material
& structure analysis suite
<DAMAGE>
<DAMAGE>
Description:
This object class permits addition of damage mechanisms to a behavior assembly of type
gen evp. These models are of the type “continuum damage mechanics” (CDM), and thus
provide interaction through alteration of the elasticity modulus with damage and calculation
of plasticity with the use of an effective stress.
Damage models may be used in the gen_evp behavior with and without inelastic deformation potentials. The coupling with inelastic deformations and their hardening variables is
discussed below.
Syntax:
The damage mechanisms are added through the use of a token **damage input after the
**elasticity declaration. The general syntax is:
[
[
[
[
**damage dama type
*elastic ]
*plastic ]
*creep ]
*coupling type ]
There is a restriction currently that the *damage statement must come after the
*elasticity statement in the behavior.
The number of coefficients for each option depends of course on on the particular damage
model selected. The currently implemented types are summarized below:
DESCRIPTION
elastic damage
CODE
*elastic
*anisotropic elastic
anisotropic elastic damage with scalar and tensorial variables
*plastic
plasticity damage which depends on the rate of
inelastic deformation
*creep
time dependent damage
fatigue
cyclic damage
*elastic The coefficients here are: B0 et alpha. The damage is calculated directly at a
given time as:
t+∆t
Ymax
1
= el : Del : el
2
t
= max(Ȳ , Ymax
)
d
=
Ȳ
1/2
Z-set — Non-linear material
& structure analysis suite
1/2
α (Ymax − Yo
0
) Ymax > Yo
Ymax ≤ Yo
13.35
<DAMAGE>
*plastic This model integrates the damage as a function of the inelastic strain equivalent
in the following manner:
d˙ = ṗi (Y /S0)s0
For the i-th component of inelastic deformation. Note that ṗ is calculated as:
1/2
ṗi = 32 ˙ i : depsi
*creep Classical viscoplastic damage using the Hayhurst stress function. The damage is
calculated as:
χ(σ) = αJ0 + βJ1 + (1 − α − β)J2
The rate of damage production will be calculated as:
χ(σ) r
˙
d=
(1 − d)−k
A
with J0 the maximum principle stress, J1 the trace of the stress tensor, and J2 the
second invariant of the deviator ( 32 s : s)1/2 .
Coupling with plasticity:
Applying simply the damage mechanisms to a gen_evp plasticity model will only couple
the stress calculation in the potentials and modify the elastic modulus. For the hardening
mechanisms to be coupled to the damage rate, several additional changes must be made to
the syntax.
Addition of one or more plasticity potentials to a behavior with damage causes the following equations to be used for the inelastic strain rate:
σ̃
= Del : el
σ
= (1 − d)σ̃
˙ to
= ˙ el + (1 − d)−1 λ̇n
ḣ
= λ̇m − ω̇
The default situation interfacing damage to the potential only involves use of the effective
stress in the place of the actual (weakened) stress. This leads to the following set of evolution
equations:
f
= f (σ̃, H, p)
λ̇
= λ̇(f, p)
∂f ∂ σ̃
∂f
=
=
∂σ
∂ σ̃ ∂σ
∂φ(σ̃, H, p)
=
∂H
n
m
from which it is noted that the hardening forces H are related to the internal (strain
analogue) variables h by the same linear relationship:
Hi = Mij hj
13.36
Z-set — Non-linear material
& structure analysis suite
<DAMAGE>
Example:
***behavior gen_evp
**elasticity isotropic
young 260000.
poisson 0.3
**damage isotropic
*creep
alpha 0.75
beta
0.0
A
3000.
r
5.3
k
15.
*coupling damage_hardening
**potential gen_evp ev
*var_coefs
*flow alt_norton
n
7.0
K
2070.
K2
1600.0
K3
19.0
*kinematic nonlinear
C
inv_one_minus dv 15000.0
D
300.0
*kinematic nonlinear
C
inv_one_minus dv 6000.0
D
100.0
*isotropic constant
R0
130.
***return
Z-set — Non-linear material
& structure analysis suite
13.37
<DIRECT KINEMATIC>
<DIRECT KINEMATIC>
Description:
This object defines a class of kinematic hardening (tensorial back stresses) which can have a
more general state variable form, and possibly nonlinear relationship between the integrated
strains and the associated stresses. In the gen_evp behavior framework, these models must
be used with the following potentials only:
DESCRIPTION
Generalized viscoplastic potential 13.86
CODE
gen evp2
The kinematic models all support static recovery, and possibly have more than a 1-1
relation between the integrated variable size and the kinematic back stress size (forcibly the
active tensor size for given problem dimension).
Syntax:
The syntax consists of specifying the type of kinematic object, supplying an optional name on
the same line, and giving the appropriate coefficient definitions on the following lines. This is
summarized below where the keyword specifying a DIRECT KINEMATIC object is assumed to
be *kinematic:
*kinematic type
coefficients
[name]
Example:
A duplication of the classical Armstrong Frederick kinematic class is implemented for verification and testing purposes (Normally using the nonlinear KINEMATIC type is the model to
choose). An example use of that class is:
*kinematic armstrong_frederick
C
40000.
D
500.
13.38
Z-set — Non-linear material
& structure analysis suite
<DIRECT KINEMATIC>
<DIRECT KINEMATIC> asaro
Description:
This kinematic hardening model provides nonlinear modulus behavior forming a hysteresis
loop with slight S shape. Normally this kinematic type would be combined with other linear
and nonlinear hardenings to fine tune cyclic behavior.
1/2
:α
C tanh(D αeq )
αi
=
D αeq
αeq =
Xi
α̇i
3
2α
= λ̇n
Syntax:
The basic input syntax here is:
*kinematic asaro
[ C
coef-value
[ D
coef-value
[ M
coef-value
[ m
coef-value
name
]
]
]
]
. . . continued
Z-set — Non-linear material
& structure analysis suite
13.39
<DIRECT KINEMATIC>
Example:
The following example shows the use of this kinematic model in combination with a number
of other kinematic hardenings, with the separate terms plotted below.
***behavior gen_evp
**elasticity isotropic
young 69000.00
poisson 0.3
**potential gen_evp2 ev
*store_all
*criterion mises
*isotropic nonlinear
R0 340.0
b 900.0
Q -210.0
*kinematic nonlinear 1_
C 187500.
D 1250.0
*kinematic armstrong_frederick 2_
C 11250.0
D 150.0
*kinematic linear
3_
C 1000.0
*kinematic asaro
4_
C 56250.0
D 750.0
***return
asaro1_tm
500
400
"asaro1_tm.test" using 2:3
300
200
asaro1_tm
200
"<awk ’{print $4,$5+$6+$7+$8}’ asaro1_tm.test" using 1:2
"asaro1_tm.test" using 4:5
150
"asaro1_tm.test" using 4:6
"asaro1_tm.test" using 4:7
"asaro1_tm.test" using 4:8
100
50
100
0
0
-100
-50
-200
-100
-300
-150
-400
-500
-0.015 -0.01 -0.005
13.40
0
0.005
0.01
0.015
-200
-0.005
-0.004
-0.003
-0.002
-0.001 0 0.0010.0020.0030.0040.005
Z-set — Non-linear material
& structure analysis suite
<DIRECT KINEMATIC>
<DIRECT KINEMATIC> non symmetric
Description:
This kinematic hardening model provides asymmetric behavior between tension and compression as recommended for cast iron materials [Hjel94,Jose95].
This model takes the tensile coefficients for computing Xi if the trace of the actual stress
1 : σ is greater than the I1_trans value. The evolution is therefore:
3 D∗
X
if use alpha is set
α̇i = λ̇ n −
i
2 C∗
3 D∗
∗
Ẋi = C λ̇ n −
Xi
if use x is set
2 C∗
where the coefficients C ∗ and D∗ are either Ct ,Dt or Cc ,Dc depending on the above check on
the stress. Note that in the use_x case the evolution is not the complete derivative where
the terms involving the change in C ∗ with temperature are not taken into account.
Syntax:
The basic input syntax here is:
*kinematic non_symmetric
[ I1_trans coef-value ]
[ Ct
coef-value ]
[ Cc
coef-value ]
[ Dc
coef-value ]
[ Dc
coef-value ]
[ M
coef-value ]
[ m
coef-value ]
[ use_x | use_alpha]
Only one choice of use_x or use_alpha can be used (default is use_x). The default
value of I1_trans is 0. The coefficients M and m are the standard Norton-like static recovery
coefficients.
. . . continued
Z-set — Non-linear material
& structure analysis suite
13.41
<DIRECT KINEMATIC>
Example:
The following example shows the use of this kinematic model in combination with the
cast_iron yield criterion.
***behavior gen_evp
**elasticity isotropic
young 126000.
poisson 0.265
**potential gen_evp2 ev
*criterion cast_iron
ratio 3.
associated 1
*kinematic non_symmetric
I1_trans 50.
Ct 300000.0
Dt
5000.0
Cc 30000.0
Dc
500.0
*isotropic constant
R0 100.0
***return
200
x1
200
"cast_iron1_yield.test" using 1:2
100
"cast_iron1.test" using 2:3
100
0
0
-100
sig22
sig22
-100
-200
-200
-300
-300
-400
-400
-500
-600
-600
-500
-400
-300
-200
sig11
13.42
-100
0
100
200
-500
-0.005 -0.004 -0.003 -0.002 -0.001
0
0.001 0.002 0.003 0.004 0.005
eto22
Z-set — Non-linear material
& structure analysis suite
<ELASTICITY>
<ELASTICITY>
Description:
An ELASTICITY behavior object is actually just an alias for COEFFICIENT MATRIX (see page
13.9). The class is frequently defined in the behavior descriptions however to further accentuate its task in the scheme of the constitutive relation. Virtually every location where this
keyword is used, a hypoelastic relation is employed:
σ = Del : el
Noting however that the total internal strain el may be calculated in some manner which
extends it from the normal limitations of hypoelasticity (e.g. the finite strain material transformations).
The matrix Del is the elasticity matrix defined by the ELASTICITY class, and can take
many forms, and vary according to any external parameter.
Z-set — Non-linear material
& structure analysis suite
13.43
<FLOW>
<FLOW>
Description:
This object class defines the model for inelastic flow in plastic and viscoplastic models and
potentials.
Syntax:
The syntax to specify a flow object will consist of giving the keyword for a particular flow
desired, followed by a list of appropriate coefficients which are dependent on the model chosen.
The flow type may be chosen among the following laws:
CODE
norton
DESCRIPTION
Norton power law
hyperbolic
hyperbolic sin function applied to power law of
overstress
strain hardening
Norton type viscoplastic rate with hardening
gsell
viscoplastic model for some polymers
modified visco
viscoplastic rate using strain hardening or softnening
sellars tegart
Coupled with hardening to study variable strain rate
function
A function in terms of overstress and cumulated plastic
strain
abq strain hardening
flow class available through ABAQUS
norton exp
exponential (saturating) Norton law
double norton
two term norton law
flow sum
summation of norton terms
flow sum inv
summation of inverse norton rates
inv exp
viscoplastic according to ˙ = A exp ((p − p0 )/ασ n )
interface control
interface reaction flow
plasticity
time independent plasticity
use global function
uses a function defined elsewhere
exponential crystal
physically based exponential flow law esp. useful for
crystal deformation
The default type of flow is plasticity.
norton This law corresponds to the classical Norton creep power law. The coefficients are
chosen to normalize the stress term, and then apply the exponent:
λ̇ =< f /K >n
with f positive. The coefficients K and n must be non-zero.
13.44
Z-set — Non-linear material
& structure analysis suite
<FLOW>
norton exp This flow provides a limit stress approximating creep at low stress levels and
plasticity at high stresses:
λ̇ =< f /K >n exp(α < f /K >n+1 )
which uses the coefficient names: K, n, and alpha.
double norton This model is a two term variation of the Norton law in order to model
changes in flow mechanisms over a wide stress range:
λ̇ =< f /K >n1 + < f /K2 >n2
with the required coefficients K, n1, K2, n2.
plasticity This flow type indicates time-independent plasticity (f = 0). The use of
plasticity may limit integration to implicit only in the case of some complex models.
There are no coefficients.
inv exp This law provides an
exponential dependence on the inverse of the effective over0
stress: λ̇ = A exp − p+p
ασ n with the coefficients n, A alpha, and p0, where p0 is optional
(default 0).
interface control This flow rule is expressed:
1 k1 σ
λ̇ = 2
d 1 + dσk2m
with the coefficients k1, k2, m, d. d represents a grain size.
flow sum inv This model is similar to the flow_sum rule, but sums the inverse of the
different flow rules:
#1/2
"
X 1
λ̇ =
λ̇i
i
where the λ̇i terms are given by the various sub-laws (different than sum_flow and
flow_sum_inv).
exponential crystal This law is a physically based flow law for crystalline slip which
can be used with the crystal POTENTIAL models (actually it can be used anywhere
viscoplasticity is allowed).
"
"
n1 #n2 #
f
γ̇ = gamma0 exp −F0 RT 1 −
K
Note that the terms for this equation may be composed of function coefficient types in the event that a functional form is desired.
For example:
F0_RT function 6.8e4/(1.9868*temperature);
strain hardening This flow law is a Norton type viscoplastic rate with a hardening effect.
n
f
λ̇ =
(v + v0 )m
K
The flow law accepts coefficients K, n, m, and v 0. Because when v = 0 there will be no
flow rate in the absense of v0 (and therefore v will never become anything other than
null), the coefficient should have a non-zero value (but possibly very small).
Z-set — Non-linear material
& structure analysis suite
13.45
<FLOW>
<FLOW> function
Description:
This flow law is defined by a function which can be in terms of overstress f and cumulated
plastic strain p.
Syntax:
The flow law will accept coefficients depending on the function. For example the above flow
law will accept K,n, p as coefficients.
Example:
A function such as a norton law combined with cumulative plastic strain could be of interest:
λ̇ =
f
K
n
(1 − ep )
This would result in the following syntax:
*flow function
K 120.0
n 5.0
p 10.e-3
13.46
Z-set — Non-linear material
& structure analysis suite
<FLOW>
<FLOW> gsell
Description:
This is a viscoplastic model appropriate for some polymer materials. It includes some hardening behavior using the cumulated viscoplastic multiplier v.
m1
f
λ̇ =
K (1 − e−wv ) ehvn
Note that at v = 0 the term 1 − e−wv in the denominator will become zero, setting the
denominator to zero (infinite initial strain rate). To alleviate this numerical difficulty, the
term is modified to be 1 − e−w(v+e0 )
Example:
A simple example of purely viscoplastic material follows. Because the yield radius R0 is zero,
the material is always flowing. The use of elasticity in this case is also a pseudo-elasticity,
attempting to model a purely viscous material in a displacement based FEA model. The
higher the modulus, the more in approaches to − th .
***behavior gen_evp
**elasticity isotropic
young 100000.
poisson 0.48
**potential gen_evp ev
*isotropic constant
R0 0.0
*flow gsell
e0 1.e-5
K
40.0
w
65.
h
0.75
n
1.75
m
0.08
***return
Z-set — Non-linear material
& structure analysis suite
13.47
<FLOW>
<FLOW> hyperbolic
Description:
For this law, the hyperbolic sin function is applied to the power law of overstress, and in turn
taken to a second power. The flow is written:
λ̇ = eps0 [sinh < f /K >n ]m
with f positive. The coefficient K must be non-zero, and the coefficients n and m default to
one.
The flow law is fully implemented for Runge-Kutta or the standard theta method. It
cannot presently be applied to the reduced integration.
Because the sinh can “blow up” with large values of f /K a cutoff limit on that ratio is
applied. This can be user-adjusted by entering a real value for cutoff in the law’s coefficient
section. The default value is 10.0 which would result in an unrealistically high strain rate for
this type of model.
Syntax:
The flow law accepts coefficients eps0, K, n, and m as outlined above.
Example:
A simple example follows:
*flow hyperbolic
K
22.3
m
1.44
eps0 .202e-8
cutoff 8.
13.48
Z-set — Non-linear material
& structure analysis suite
<FLOW>
<FLOW> modified visco
Description:
This flow law is a Norton type viscoplastic rate, using a strain hardening or softening viscocity
term.
Let the following terms be defined:
n
f
K = K0 + Q(1 − e−bv )
d1 =
K
With which the flow law is written:
f
λ̇ = d1 exp αd1
K
Syntax:
The flow law accepts coefficients K0, Q, b, n and alpha.
Note:
There is another version of this flow called norton_hardening which reduces to:
n
f
λ̇ =
K0 + Q(1 − e−bv )
Z-set — Non-linear material
& structure analysis suite
13.49
<FLOW>
<FLOW> sellars tegart
Description:
This is a flow rule given by C M Sellars and W J Tegart in 1969 as a proposed rate equation for
hot forming conditions. The use of a hyperbolic sine function allows for increasing overstress
at low strain rates, which is followed by essentially a saturation level of overstress. Another
way of describing the model is viscous at low overstress levels followed by essentially rate
independent behavior for high stresses. Normally this behavior would be combined with
hardening (viscoplasticity) and not taken as a pure viscous creep law.
The rule is given as:
m
f
λ̇ = 0 sinh
σ0
Note that the Z-mat formulation utilizes a ratio of the overstress by a normalizing flow stress
σ0 . Once the overstress becomes of the order of the flow stress this law will make the transition
to rate independence.
Many uses of this equation include explicitly the activation energy term to scale the flow
viscosity according to thermally activated mechanisms. In Z-mat the activation term should
be entered in using the coefficient mechanism (either as tables or functions) for the value of
de0.
Syntax:
The flow law accepts coefficients de0, m, sigma0
Example:
An example input using the coefficient mechanism to add thermal activation is given below:
*flow sellars_tegart
de0
function 0.202e-2*exp(-401.0/(8.3144e-3*(temperature+273.0)));
sigma0 120.0
m
1.44
13.50
Z-set — Non-linear material
& structure analysis suite
<FLOW>
<FLOW> sum
Description:
P
This law provides a summation of different flow rates λ̇ = i λ̇i where the λ̇i are the rates
given by other flow objects (other than sum_flow or flow_sum_inv).
Syntax:
The syntax is the following:
*flow flow_sum
<FLOW>
<FLOW>
Similar kind of syntax applies to flow sum inv law also.
Example:
A simple example is given below which uses the Norton law as one of the flow rule.
***behavior gen_evp
**elasticity isotropic
young 260000.
poisson 0.3
**potential gen_evp ev
*criterion mises
*flow flow_sum
norton
K 140.0
n 5.0
*isotropic constant
R0 130.0
Z-set — Non-linear material
& structure analysis suite
13.51
<INTERACTION>
<INTERACTION>
Description:
This option allows addition of various variable interactions in the gen evp behavior object.
Syntax:
The syntax of the interaction types depends on the type of interaction. Generally there will be
a specification of the type of interaction and two identifying tokens which define the variables
which interact. This structure is summarized below:
**interaction [ type ] item1 item2
...
where type may be from the following types:
DESCRIPTION
Crystalline isotropic hardening interaction.
CODE
iso
default In the absence of type the default “state” interaction will be implemented as
described by [Cont89]. This interaction calculates associated forces including terms of
other internal variables. The coupling will be stated by using the (user supplied) name
of a potential and the (user supplied) name of hardening mechanisms.
**interaction P1::H P2::H <COEFFICIENT>
iso This interaction will link the isotropic variables of one class of slip systems with another
(latent strain coupling) between two mono-crystal potentials. The coupling is written:
Ri = Roi + Qi
n
X
(1 − e−bvj )hij
j=1
syntax for this type of coupling takes the names of the two potentials, and a coefficient
definition named h:
**interaction iso name1 name2
h
<COEFFICIENT>
Note: for interactions between slip systems within one potential one should use the
SLIP INTERACTION class (see page 13.92).
Example:
This is an example of the state-law default coupling to link two kinematic variables in a
time-dependent viscoplastic potential and a time-independent plasticity potential.
XX1
XX2
= 32 CX1 αX1 + 23 Cint αX2
= 32 CX2 αX2 + 23 Cint αX1
(7)
The example material definition for this type of interaction is as follows:
13.52
Z-set — Non-linear material
& structure analysis suite
<INTERACTION>
***behavior gen_evp
**elasticity isotropic
young 170000.0
poisson 0.30
**potential gen_evp ev
*flow norton
n 1.0
K 20300.
*kinematic nonlinear X1
C 2000.0
D 50.0
*isotropic constant
R0 20.0
**potential gen_evp ep
*kinematic nonlinear X2
C 25000.0
D 10.0
*isotropic constant
R0 850.0
**interaction ev::X1 ep::X2 25000.0
***return
The names X1 and X2 were given in order to have readable names for the kinematic
variables. In this syntax it is important to notice that a specification ev::X1 is different than
ev::X1 because the kinematic names are localized within each potential.
Z-set — Non-linear material
& structure analysis suite
13.53
<ISOTROPIC>
<ISOTROPIC>
Description:
This object class defines models of isotropic hardening for use in a variety of material behaviors
and potentials. Isotropic hardening causes an isotropic expansion or contraction of the yield
surface in tensorial stress space. The isotropic hardening may be a function of the simulated
multiplier which is normally equal to the cumulated plastic strain equivalent p, or in terms of
its own internal variable. The latter must be used for cases of static recovery.
The initial value R0 for the yield radius is noted here to be the onset of plasticity, and not
the engineering yield stress. It may thus be found to be significantly lower than expected.
Syntax:
The syntax to specify a isotropic hardening object will consist of giving the keyword for a
particular law desired, followed by a list of appropriate coefficients which are dependent on
the model chosen. The possible isotropic laws are the following:
constant
DESCRIPTION
initial value only (perfect plasticity).
linear
(or iso linear) linear function of p.
power law
power law of p.
by point
hardening defined by a list of experimental points in an
external datafile.
iso table
hardening defined by a table of points.
linear pp
linear hardening followed by perfect plasticity.
nonlinear
nonlinear, exponentially saturating function of p.
nonlinear v1
nonlinear with an internal variable and coefs 1.
nonlinear v2
nonlinear with an internal variable and coefs 2
(equivalent to nonlinear).
nonlinear bsi
nonlinear with kinematic interaction (for use with the
mises 2m1c potential only).
linear nonlinear
combined linear and nonlinear saturating hardening.
nonlinear sum
multiple term nonlinear hardening.
nonlinear double
multiple term nonlinear hardening. This law is rigorously
equal to nonlinear sum, but is retained for backwards compatibility. Although the name suggests that only two terms
are allowed, more can be added.
CODE
nonlinear recovery
function
hardening according to a user-defined function of p.
. . . continued
13.54
Z-set — Non-linear material
& structure analysis suite
<ISOTROPIC>
Only the laws with an internal variable may have static recovery and be used in state interactions. Isotropic hardening is available with positive values of the Q or H parameters (depending
on the model), while softening occurs with negative values of these coefficients. The saturation
rate coefficients b are necessarily positive. All parameters may have the normal dependencies
through the COEFFICIENT mechanism (see page 13.7).
constant This gives the radius as a single coefficient R0.
linear (or iso linear) Law with initial radius and a linear evolution depending on the
cumulated inelastic strain p (i.e. R = R0 + Hp) and using the coefficients R0 and H.
power law Gives a power-law evolution according to
R = R0 + K (e0 + p)n
with the coefficients R0, K, e0, n, and p the cumulated inelastic strain. Negative values
for e0 are set to zero. The case n < 1 with e0 = 0 gives an infinite derivative at the
onset of inelastic deformation and is therefore not allowed. Use a very small value for
e0 instead.
by point Gives the isotropic curve as input by a list of points (see the example). The first
column should be sigeq and the second the cumulated plastic strain equivalent.
iso table Gives the isotropic curve as input by a list of points directly in a table (see
example). When the current value of p depasses the last value specified in the table,
the isotropic hardening no longer changes, as if the isotropic hardening saturates.
linear pp linear hardening with an initial radius R0 and slope H up to Ru (i.e. R = R0 + Hp
for R < Ru), followed by perfect plasticy at R = Ru.
nonlinear This law gives a nonlinear saturating evolution:
R = R0 + Q 1 − e−bp
where the saturation radius will be R0 + Q for large p. The law will reach 95% of the
saturation value at p ≈ 3.0/b so that higher values of b give a more rapid saturation.
linear nonlinear linear hardening with an initial radius R0 followed by a combined linear
and nonlinear evolution with coefficients H, Q and b, according to
R = R0 + Hp + Q 1 − e−bp .
nonlinear sum or nonlinear double. This law allows fine tuning of the isotropic hardening
by use of multiple-term evolution. For instance, two terms might be combined analogous
to short and long term mechanisms. There is no upper limit for the number of terms N.
The radius is calculated as:
R = R0 +
N
X
Qi 1 − e−bi p
i
with coefficients R0, Q1, Q2, ..., QN, and b1, b2, ..., bN.
. . . continued
Z-set — Non-linear material
& structure analysis suite
13.55
<ISOTROPIC>
nonlinear 1
miso = 1 −
R − R0
Q
Example:
An example of a three-term nonlinear hardening with a very rapid initial hardening, a very
slow hardening (third term) and an intermediate softening (negative Q2). Note: these are not
necessarily reasonable values.
*isotropic nonlinear_sum
R0
50.
Q1
30. b1 2000.
Q2
-10. b2
20.
Q3
40. b3
2.
The next example shows the use of *isotropic by point. Between two data points, interpolation is linear.
**potential gen_evp ep
*criterion mises
*flow plasticity
*isotropic by_point
sigeq
epcum
130.0
0.0000
140.0
0.0001
145.0
0.0002
150.0
0.0004
160.0
0.0009
170.0
0.0017
180.0
0.0028
The following gives a point-by-point specification of an isotropic harding. Between two data
points, interpolation is linear (as with all Z-set tables). For p > 0.1 it uses the values of 0.1
because that was the last specified.
*isotropic table
0.
temperature
220. 23.0
100. 200.
0.1 temperature
300. 23.0
120. 200.
13.56
% temp table for epcum = 0.
% temp table for epcum = 0.1
Z-set — Non-linear material
& structure analysis suite
<KINEMATIC>
<KINEMATIC>
Description:
This object defines the model of kinematic hardening (tensorial back stresses) applicable
in behaviors or potential objects in the gen_evp behavior. The models generally store an
additional tensorial internal variable, and therefore may support static recovery and state
coupling without modification. The name of this internal variable will be determined normally
by the object to which it belongs (a POTENTIAL for example). Kinematic hardening acts as a
translation of the yield surface in stress space (often deviatoric).
Syntax:
The syntax consists of specifying the type of kinematic object, supplying an optional name
on the same line, and giving the appropriate coefficient definitions on the following lines.
This is summarized below where the keyword specifying a KINEMATIC object is assumed to be
*kinematic:
*kinematic type
coefficients
[nom]
where the possible substitutions for type are:
CODE
linear
DESCRIPTION
linear hardening
nonlinear
nonlinear Armstrong-Fredrick hardening with
static recovery
nonlinear ai
same as the nonlinear model but using more accurate asymptotic integration
ziegler
nonlinear hardening based on the Euclidean norm
I2
nonlinear phi
nonlinear with cumulative influence
nonlinear evrad
nonlinear with reduced ratcheting
nonlinear with crit
linear-nonlinear with criterion
aniso nonlinear
nonlinear model using coefficient matrices for the
coefficients
Kinematic models are all formulated in state variable form, where the stored (integrated)
variable is analogous to a strain translation in tensorial strain space. The “back stress”
component is then calculated using an appropriate modulus. Most of the kinematic models
use the convention for calculating the back stress:
2
X = Cα
3
with X the back stress and α the internal variable. Material softening through kinematic
variables is not thought to be physically reasonable, and so the modulus coefficients C should
always be positive.
Z-set — Non-linear material
& structure analysis suite
13.57
<KINEMATIC>
Evolution of the internal variable is written in the form:
α̇i = λ̇mkin (σ, X, p, . . . ) − ω̇ kin (X)
where m is the “hardening normal” and ω̇ kine a static recovery function.
linear The linear kinematic hardening has its internal variable evolve with the inelastic
strain:
mkin = n
The only coefficient for this model is the C modulus (units of stress). The slope in a
uniaxial test will be equal to the C value.
nonlinear The nonlinear model evolves much as the isotropic nonlinear model does (with
a slight change in coefficient meaning). The evolution is the following:
X
mkin = n − 3D
D2C
Em
J
(
2 X)
ω kin = 32 J X
M
2 (X )
(8)
The coefficients are C and D for the strain evolution, and M and m for the static recovery
term. In the absence of these latter two, there will be no static recovery calculation.
The saturation of this model occurs at C/D in a uniaxial test at a rate determined by D.
Increases in D yield faster saturating (more nonlinear) behavior.
Static recovery is seen to follow a Norton type formulation in the back stress tensorial
space. This kinematic model supports RK, TM, or Reduced TM integrations.
nonlinear phi This model installs a kinematic variable with the following evolution:
mkin = n − φ(λ)
3D
X
2C
where the function φ is:
φ(λ) = φm + (1 − φm ) exp(−ωλ)
where the material coefficients φm and ω are named phim and omega. Note that ω
defines the rapidity of saturation in the non-linear coefficient, and φm is a factor between
zero and one interpolating the D coefficient between initial and saturated values. This
kinematic model works in RK, TM, or Reduced TM.
nonlinear evrad This model limits the ratcheting effect in biaxial conditions by comparing
the flow direction with the back stress. Evolution is calculated as follows:
3D
mkin = n − η1 + 23 (1 − η)n ⊗ n
X
2C
The coefficients are defined such that uniaxial response is equivalent to the nonlinear
model given equivalent coefficients. The coefficient names are C, D, and eta. This model
works in RK, TM, or Reduced TM, but has no static recovery.
13.58
Z-set — Non-linear material
& structure analysis suite
<KINEMATIC>
ziegler This model describes a back stress with hydrostatic and deviatoric components:
mkin =
ω kin =
σ −X
D
I2 (σ −X) − C X
D
Em
I2 (X)
X
M
I2 (X)
(9)
which takes the coefficients C, D, M and m. In the absence of the latter two, there will
be no static recovery. The Ziegler model may not be used with the reduced_plastic
integration.
nonlinear with crit
3D
φ(X)X
2C
DJ(X) − ω m1
1
φ(X) =
1−ω
DJ(X)m2
mkin = n −
Example:
*kinematic nonlinear_evrad
C
40000.
D
500.
eta
0.4
Z-set — Non-linear material
& structure analysis suite
13.59
<KINEMATIC>
<KINEMATIC> aniso nonlinear
Description:
This kinematic model implements anisotropic coefficients into a KINEMATIC behavior object.
Note that this model does not include the 1.5 term in the modulus coefficients, so there are
none in the evolution. The coefficients are to be entered into COEFFICIENT MATRIX objects
(see page 13.9).
mkin = n − DC−1 X m
kX kQ
ω kin = C−1 ξQ : X
M
(10)
Any of the coefficients can of course be a function or table of the temperature, or other
parameter.
The equivalent term kXkQ above is calculated as follows:
kXkQ = [X : Q : X]1/2
Syntax:
[
[
[
[
**kinematic aniso_nonlinear
C <COEFFICIENT_MATRIX>
D <COEFFICIENT_MATRIX>
xi <COEFFICIENT>
]
Q <COEFFICIENT_MATRIX> ]
m <COEFFICIENT>
]
M <COEFFICIENT>
]
[name]
% defines recovery exists
% overrides global
% ditto
Example:
*kinematic aniso_nonlinear x1
C cubic C1111 20000.0 % 30000.0/1.5
C1122 0.0
C1212 20000.0
D cubic
D1111 500.
D1122 0.0
D1212 500.
13.60
Z-set — Non-linear material
& structure analysis suite
<LOCALIZATION>
<LOCALIZATION>
Description:
This object selects a localization model to be applied to the different potentials of a gen evp
type behavior.
Syntax:
Localization is indicated within a gen_evp behavior with an entry of the following type:
**localization type
From which the following localization types may be chosen:
DESCRIPTION
polycrystal (multi-grain) model
CODE
poly
poly
The polycrystal mode here is a generalized form of that described by Cailletaud and
Pilvin [Cail92, Cail94, Pilv94]. Several potentials of type crystal will be assembled to
form a slt/phase/ of the material, which in turn is grouped according to the orientations
and volume fractions of material. The syntax for the model is:
**localization poly
*model_coefficients
coefs
[
[
[
*grains
]
grain list ]
*file file name ]
Z-set — Non-linear material
& structure analysis suite
13.61
<LOCALIZATION>
Example:
***behavior gen_evp
**elasticity isotropic
young
200000.
poisson 0.30
**localization poly
*grains
-149.676 15.61819 154.676 1.
-149.676 15.61819 154.676 1.
-150.646 33.86400 155.646 0.2
*model_coefficients
C 50000.
D 100.
**potential octahedral
*flow norton
n 25.
K 50.
*isotropic nonlinear
R0 100.
Q
50.0
b
50.0
*kinematic linear
C
2500.
*interaction slip
h1 1.0
h2 1.0
h3 1.0
h4 1.0
h5 1.0
h6 1.0
***return
13.62
Z-set — Non-linear material
& structure analysis suite
<POROUS CRITERION>
<POROUS CRITERION>
Description:
This behavior object is used to enter the citerion for a porous plastic material. The description
of these models will use the following notation:
σ
DESCRIPTION
stress tensor
s
deviatoric stress tensor
J 2
the second invariant of deviatoric stress 12 s : s
I 1
first invariant of stress Trace σ
∼
σ?
current flow stress
CODE
Syntax:
The syntax for a porous criterion requires that a type be given, followed by whatever coefficients are allowed for the particular model. As in all of Z-mat, these coefficients are entirely
context sensitive.
*porous_criterion type
coefs
The available list of porous plastic models are summarized below.
CODE
gurson
DESCRIPTION
model to capture progressive microrupture
elliptic aniso
generalization of the elliptic criterion for general
anisotropy.
elliptic
criterion for densification of material
elliptic fiber
elliptic criterion for compaction of fibers in a powder
matrix.
fkm
NIOT
modified rousselier
NIOT
rousselier
thermodynamic criterion through free energy
cam clay
criterion to study the porosity of clay
zhang niemi
NIOT
The following pages discuss these models in details.
Z-set — Non-linear material
& structure analysis suite
13.63
<POROUS CRITERION>
<POROUS CRITERION> cam clay
Description:
This criterion is developed to study the porous behavior of clay. In literature, many modified
versions of cam-clay models have been developed recently to study the behavior of partially
saturated soils. Zmat currently inlcudes the basic Cam-Clay model.
3J2
+ (I1 /3 + pc )2 − σ?2
m2
with pc and m coefficients named pc and m.
Syntax:
*porous criterion cam clay
coefficients
...
Example:
A simple example from a test case is given below
*porous_criterion cam_clay
pc
ft
100. 0.
20.
1.
m
.5
13.64
Z-set — Non-linear material
& structure analysis suite
<POROUS CRITERION>
<POROUS CRITERION> elliptic
Description:
The elliptic criterion is a simple ellipse in P-q plane, and allows a great deal of flexibility in
adjusting the pressure sensitivity of plasticity. This model can be used for general densification
of problems, such as powder compaction and foam plasticity.
3CJ2 + F I12 − σ?2
where C and F are the coefficients named C and F.
Because the ratio between the deviatoric and pressure sensitivity is determined by the C
and F coefficients, we can get great flexibility by making these coefficients depend on the
porosity state variable f . A typical limiting case is that the material should behave in a fully
deviatoric way at the fully dense condition (when f = 0 then C=1 and D=0), and similarly
when the material is fully void there should only be pressure dependence( when f = 1 then
C=0 and D=1).
Syntax:
*porous criterion elliptic
coefficients
...
Example:
A simple example from a test case is given below
*porous_criterion elliptic
F f
0. 0.
5. 1.
C f
1. 0.
11. 1.
Z-set — Non-linear material
& structure analysis suite
13.65
<POROUS CRITERION>
<POROUS CRITERION> elliptic aniso
Description:
This porous potential allows for a very general anisotropic anisotropic porous plastic material.
Anisotropic influence on the direct stress components for the (ordinarily) trace operator for
the pressure term are replaced with a coefficient factored trace. Anisotropic effects in the
shear term can be implemented with the generally available *shear anisotropy option in
the porous plastic material behavior (which replaces the J2 term with an effective anisotropic
measure of the shear criterion).
The potential is therefore:
3C J˜22 + F I˜12 − σ?2
where C and F are the coefficients named C and F, and which can depend notably on the
porosity variable f. As stated above the J˜2 will be as supplied by the shear anisotropy (default
standard J2 measure), and the I˜ is calculated by:
I˜1 = p σ11 + q σ22 + r σ33
and p q and r are coefficients (not depending on state variables).
Syntax:
The syntax follows the typical porous potential format with the 5 coefficients C, F, p, q, and
r to be entered. Again, only C and F can depend on the porosity.
*porous criterion elliptic
coefficients
...
. . . continued
13.66
Z-set — Non-linear material
& structure analysis suite
<POROUS CRITERION>
Example:
An example test is in Sam test/INP in the input files elliptica and ellaniso.inp
***behavior porous_plastic
**elasticity orthotropic
y1111 464.
y2222 239.
y3333 239.
y1212 105.
y2323 105.
y3131 105.
y1122 172.
y2233 142.
y3311 172.
**porous_potential
*porous_criterion elliptic_aniso
C 1.
F
1.839926000000000e-03
p 1.
q
2.061266000000000e+01
r
2.330418000000000e+01
*shear_anisotropy hill
hilla
4.172356995000000e-01
hillb
2.066054000000000e+00
hillc
1.992586000000000e+00
hilld 1.
hillf 1.
hille 1.
*flow norton
% pseudo rate independant
K 0.001 n 10. % plasticity with these coefs
*isotropic_hardening nonlinear_double
R0
4.500000000000000e-01
Q1
3.729991000000000e-01
b1
3.276424000000000e+02
Q2
8.166423000000000e+03
b2 0.001
***return
Z-set — Non-linear material
& structure analysis suite
13.67
<POROUS CRITERION>
<POROUS CRITERION> fkm
Description:
This criterion is as developed by Fleck, Kuhn, and McMeeking [Flec92] for metal powder
compaction.
6J2 +
p
27J22 + 4I 2
6α
with coefficients fm ax, offset and fs named fmax, offset and fs.
Syntax:
*porous criterion fkm
coefficients
...
Example:
A simple example from a test case is given below
*porous_criterion fkm
fmax
offset
fs
13.68
Z-set — Non-linear material
& structure analysis suite
<POROUS CRITERION>
<POROUS CRITERION> gurson
Description:
This criterion was proposed by Gurson and is no doubt a widely accepted model. The model
was developed to capture the progressive microrupture through void nucleation and growth.
Since then it has been modified in various forms for implementation in a variety of fields,
ductile rupture of metals being the most applied one. The model defines the criterion as:
3J2
+ 2f ? q1 ch
σ?2
q2 I1
2σ?
− (1 + q12 f ?2 )
where f? , q1 q2 are coefficients of the model. Names for these coefficients are fs, q1 and q2.
The coefficients q1 and q2 have units of stress.
Syntax:
*porous criterion gurson
coefficients
...
prefix
size
description
default
eto
sig
T-2
T-2
total strain
Cauchy stress
yes
yes
Example:
A simple example from a test case is given below
*porous_criterion gurson
q1 1.5
q2 1.0
fs = ft
Z-set — Non-linear material
& structure analysis suite
13.69
<POROUS CRITERION>
<POROUS CRITERION> rousselier
Description:
The Rousselier model is a thermodynamics model formed by coupling the plasticity and
damage through free energy. The criterion is defined as below:
√
3J2
+ σ1 f D exp
1−f
I1
3(1 − f )σ1
− σ?
with the coefficients σ1 and D named sigma1 and D. Although the Rousselier model is very
similar to the Gurson model, a major point of difference between the two is that in Rousselier
model the damage grows under pure shear and the non zero shear deformation occurs under
pure hydrostatic pressure due to the shape of the yield function.
Syntax:
*porous criterion rousselier
coefficients
...
Example:
A simple example from a test case is given below
*porous_criterion rousselier
sigma1 200.
D
0.85
13.70
Z-set — Non-linear material
& structure analysis suite
<POROUS CRITERION>
<POROUS CRITERION> modified rousselier
Description:
√
3J2
+ σ1 f D exp
1−f
q2 I1
3(1 − f )σ1
− σ?
where q2 , D and sigma1 are coefficients named q2, D and sigma1.
Syntax:
*porous criterion modified rousselier
coefficients
...
Example:
A simple example from a test case is given below
*porous_criterion modified_rousselier
sigma1 200.
D
0.85
q2
Z-set — Non-linear material
& structure analysis suite
13.71
<POROUS CRITERION>
<POROUS CRITERION> zhang niemi
Description:
2J2 + I3
1.5(1 − χ2 )Cf
The coefficients for the criterion are λ0 , χc and fs named lambda0, chi c, fs.
Syntax:
*porous criterion zhang niemi
coefficients
...
Example:
A simple example from a test case is given below
*porous_criterion zhang_niemi
lambda0
chi_c
fs
13.72
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
<POTENTIAL>
Description:
The potential object classes define the dissipation potentials within a gen evp behavior assembly. Each potential given describes an inelastic strain mechanism, along with the hardening
variables which affect its evolution. We can generally combine an arbitrary number and mix
of the potentials in order to create complex behaviors. This combination is however not verified for physical compatibility so a particular assembly should be extensively studied under
simple volume element loadings.
Syntax:
**potential potential type
[name]
The types and number of sub-options is dependent on the potential type and is thus left
to be described with the potential types. The types available are summarized below:
gen evp
DESCRIPTION
classical plasticity using un-coupled isotropic and
kinematic hardening
non associated
need info
delobelle
special kinematic hardening evolution with multiple tensor variables
gen evp2
modified gen evp with different hardening variables
coupled recovery
kinematic recovery which couples the back stresses
associated
completely associated version of classical plasticity
z6 gen evp
un-associated interactions compatible with Z6
mises 2m1c
Complex interaction and criterion 2M1C
memory
gen evp with strain-range memory
suvic
SUVIC complex hardening evolutions
CODE
If the option name is given, the names of the inelastic deformation tensors will be constructed from the names given. By default the names will be generated automatically based
on the type and number (by order of definition) of each potential. It is generally advised to
use the names ev for the first viscoplastic potential and ep for the first plastic potential.
Z-set — Non-linear material
& structure analysis suite
13.73
<POTENTIAL>
<POTENTIAL> associated
Description:
This potential takes the same form as the gen_evp potential described above, but alters the
criterion to be associated with the hardening mechanisms input. The criterion form can be
written:
f = fcr (p, σ − ΣXi ) + Ωi (R) + ΣΩi (Xi )
where fcr is the criterion as calculated by the <CRITERION> object entered after *criterion.
There are some additional restrictions with this potential type due to numerical or physical
restrictions.
The resulting behavior includes a nonlinear yield zone radius dependence on the R isotropic
variable. A dependence also is introduced with the kinematic hardening variables on the
isotropic radius. As the kinematic back stress is increased (through monotonic straining)
the yield radius is decreased thereby increasing the Bauschinger type effect (reduced yield on
reversal).
Syntax:
The syntax understood by this potential is summarized below:
**potential associated [name]
[ *flow
<FLOW>
[ *flow
<RATE_VAR_FLOW>
[ *criterion
<CRITERION>
[ *kinematic
<KINEMATIC> [name]
[ *kinematic
<DIRECT_KINEMATIC>
[ *isotropic
<ISOTROPIC>
[ *var_coefs
[ *store_all
]
]
]
]
[name] ]
]
]
]
Example:
The example given for the gen_evp potential may be run with this potential by only changing
the potential name to associated.
***behavior gen_evp
**elasticity isotropic
young 260000.
poisson 0.3
**potential associated
*criterion mises
*flow norton_k_variable
n 7.0
*kinematic nonlinear
A1 15000.0
Bp
30.0
*isotropic constant
R0 130.0
*K non_linear_recovery
13.74
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
R0 200.
A
0.0
Rp
1.0
*parameters
A
7.8125e-3
N
1.0
vdot0
1.0e-11
sig0
1.0
m
1.0
B01
1.0
B02
1.0
***return
Z-set — Non-linear material
& structure analysis suite
13.75
<POTENTIAL>
<POTENTIAL> coupled recovery
This potential provides a slight modification on the gen_evp potential in that it couples
the back stress effect in recovery. The kinematic variables are the only ones affected in this
coupling. The evolution for each kinematic variable j will use the following form:
Ẋj = λ̇mj (σ, λ, Xj ) − ω̇ j (X)
X=
X
Xi
i
13.76
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
<POTENTIAL> delobelle
Description:
Delobelle’s potential implements a special form of kinematic hardening evolution using multiple tensorial variables1 . The implementation is based on the work described in [Delo96].
The class allows any CRITERION, FLOW, object to be used, and uses the ISOTROPIC object
in a somewhat different manner than the gen_evp based potentials. There is a limitation on
the ISOTROPIC objects which can be used, such that no additional integrated variables are
added.
The inelastic flow is controlled using a scalar CRITERION value for the “overstress:”
f = f (σ − x)
n=
∂f
∂σ
where the function f is selected with the *criterion option below. The tensor x is the
current back stress (not sum of back stresses as in other models).
The scalar flow magnitude λ̇ is a function defined in terms of f using the *flow option as
in other potential models. The inelastic strain rate tensor is also defined as in other models
using the normality principle so ˙ in = λ̇ : n.
The initial hardening slope for kinematic variables follows a flow normal which is modified
from the inelastic strain by scalar hardening and a 4th order tensorial orientation matrix:
m = 23 (R(λ) + Yθ )T1 : n
The isotropic function R(λ) (λ is the cumulated equivalent inelastic strain) is chosen using
the *isotropic option below. Note that only ISOTROPIC objects which have no integrated
variables may be used (e.g. no recovery). The T1 matrix is entered using the *T1 command
(required).
The tensorial back stress evolves with the following three coupled state variables:
α̇2 = λ̇ [m − T2 x2 ]
x2 = p2 α2
α̇1 = λ̇ [m − T2 (x1 − x2 )]
x1 = p1 α1
∂g
x=pα
∂x
The last term is a static (time based) recovery mechanism in the x variable. It uses a separate
recovery potential (CRITERION) object for g and a flow rate magnitude to find λ̇r . These are
selected with the *g_function and *recovery_flow options for g and λ̇r respectively. In
these evolutions, the matrix T2 is entered similarly to the T1 matrix, but using the *T2
command.
An additional isotropic hardening component can be added to handle supplemental hardening in the case of non-radial loading. The variable is integrated using the following evolution:
x
∂h(
ẋ)
Y˙θ = λ̇ [Yt inffθ − Yθ ]
fθ = 1 − abs
Yθ = btyθ
g(x) ∂ ẋ
α̇ = λ̇ [m − T2 (x − x1 )] − λ̇r (g)T1 :
The potential function h is a CRITERION object which can be entered using the optional
*h_function command. If it is not entered, the g function will be used.
1
explicit forms of this model are available in the ZebFront files Edf modif.z and Lma cwsr.z
Z-set — Non-linear material
& structure analysis suite
13.77
<POTENTIAL>
Hardening variables will be stored in the following order:
h = [α1 α2 α3 yθ ]
while yθ is only included when the coefficient bt is given.
Syntax:
The syntax understood by this potential is summarized below:
**potential delobelle [name]
[ *flow
<FLOW>
[ *criterion
<CRITERION>
[ *isotropic
<ISOTROPIC>
[ *T1
[ *T2
[ *g_function
[ *recovery_flow
[ *model_coef
p p1 p2 bt Yt_inf
[ *h_function
Output:
Internal variables
added
by
a
]
]
]
]
]
]
]
]
]
delobelle
potential
instance
are
the
prefix
size
description
default
pnvi
pncum
pn alpha2
pn alpha1
pn alpha
yt
T-2
S
T-2
T-2
T-2
S
inelastic strain tensor
cumulated value of the λ̇
kinematic strain variable α2
kinematic strain variable α1
kinematic strain variable α
non-radial hardening variable
yes
yes
no
no
no
no
following:
Note that the program does not normalize the value of n such that n : n = 1 which could
affect the meaning of the behavior and the influence of the anisotropic coefficients.
Compatibility/limitations:
This potential is not valid for the reduced integration behavior. It also does not implement
the copy mechanism required for use in the polycrystal. It may not be reasonable to make
kinematic state variable coupling, but it should work (in the p coefficients). This model
should be valid with the damage mechanics **damage or other similar gen_evp modifiers.
Coefficients should not be allowed to vary with VINT or VAUX variables. The use of Yθ is
not exceptionally tested.
Example:
***behavior gen_evp
**elasticity isotropic
young 80000.0
poisson 0.32
**potential delobelle ev
*flow hyperbolic
13.78
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
K 22.3
m 1.44
eps0 .202e-8
*isotropic constant
R0
99.5
*criterion anisotropic orthotropic
c11 .44667
c22 .58
c33 .666
c44 0.
c55 1.7
c66 0.
c12 -.18
c23 -.4
c31 -.26667
*T1 orthotropic
c11 .1266667 c22 .338 c33 .666
c44 0.
c55 .66
c66 0.
c12 .101
c23 -.439 c31 -.2276667
*T2 orthotropic
c11 .1021667 c22 .3235 c33 .6666666
c44 0.
c55 1.
c66 0.
c12 .1205
c23 -.444 c31 -.2226667
*g_function anisotropic orthotropic
c11 3.346000 c22 3.346 c33 .666
c44 0.
c55 1.375 c66 0.
c12 -3.012667 c23 -.333333 c31 -.3333333
*recovery_flow hyperbolic
K
331.69
n
3.569
eps0 0.00000019831932773108 % .00354/11900.*2.0/3.0
*model_coef p 11900.0 p1 1656.0 p2 325.5
***return
Z-set — Non-linear material
& structure analysis suite
13.79
<POTENTIAL>
<POTENTIAL> crystal
Description:
The crystal potential type is rather a group of similar potential classes for modeling different monocrystalline slip systems. To construct a single crystal it is normally required to
add several different potentials. For instance, a FCC crystal is often constructed with the
octahedral and cubic planes modeled in two potentials.
The model includes a localization step, where the stress is resolved into a scalar τi acting
along the slip direction in the slip plane i (this is the resolved shear stress). The localization
depends on the precise orientation of slip planes and slip directions, as described in the section
<CRYSTAL ORIENTATION> on page 13.31.
The yield criterion depends on a number of scalar kinematic back stress components Xij
(see page 13.30), which oppose τi , and scalar isotropic variables Rik modeling the change in
resolved shear stress length on plane i. Multiple kinematic components may be specified,
but only one isotropic. The other contributions to the isotropic hardening come from latent hardening, i.e. from interactions with other slip systems of the current potential (see
<SLIP INTERACTION> on page 13.92), and/or with slip systems of other crystal potentials (see
<INTERACTION> on page 13.52). The variables are shown schematically below (for the specific
case of only one kinematic back stress and all contributions to the isotropic radius contained
in R).
Resolved
Shear
Stress
τ
!"!" !"!" !"!" !"!" !"!" "" &!% &!% &!% &%
!"!"!!"!"!!"!"!!"!"!!"!"!"" &!%&!% &!%&!% Viscoplastic
&!% &%
"!"!" "!"!" 2R"!"!" "!"!" "!"!" """ &!%&!%&!% &!%&!%&!% &!%&!%&!%&!% &%&%&%&%
!""!!""!!""!!""!!""!.- "" .!- .!- &!%&!% .- &!%&!% .- &!%&!% .- &%&%
"!"!"!"!"!"!"!"!"! 0"!"!"!('(' ,!+,!+,!+ "!"!"!.-.-.--!. ,+,+,+ *)*) """ .!-.!-.!--!. .!-.!-.!--!. &!% .-.!-.!--!. &!% .-.!-.!--!. &!% .-.-.--. &%
$!##!$ $!##!$ $!##!$ (' $##$ .!-.!- X.!-.!- .!-.!- .!-.!- .!-.!- .-.$!#$!## $!#$!## $!#$!## $#$## .!-.!- .!-.!- .!-.!- .!-.!- .!-.!- .-.$!$!#$!# $!$!#$!# $!$!#$!# $$#$# .!-.!- .!-.!- .!-.!- .!-.!- System
.!-.!-.!- .-.-.- elastic
!
.
!
.
!
.
!
.
zone
$!#$!# $!#$!# $!#$!# $#$#
Viscoplastic
Syntax:
The syntax of each crystalline potential takes the following form:
**potential <CRYSTAL_ORIENTATION> [ name ]
[ *flow <FLOW>
]
[ *kinematic <CRYSTAL_KINEMATIC> ]
[ *isotropic <ISOTROPIC>
]
[ *interaction <SLIP_INTERACTION> ]
[ *rotate <ROTATION>
]
[ *store_all
]
. . . continued
13.80
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
P j
P k
• The criterion is always: |τi − xi | − ri with xi =
j Xi and ri =
k Ri , where the
term k = 0 refers to the isotropic hardening of the current potential (specified with
the *isotropic command), and the other terms k > 0 reflect the contributions due to
SLIP INTERACTION (see page 13.92), or inter-potential INTERACTION (page 13.52).
• If no *flow command is given, it is assumed to be plasticity (see page 13.44).
• All the ISOTROPIC objects listed later in this manual (page 13.54) which do not have
an integrated variable may be used for this model (i.e. no static recovery). If no
*isotropic command is given, the code will assume *isotropic constant and a value
for R0 should be given. Only one isotropic hardening object is allowed (but some of these
objects contain multiple terms).
• Multiple *kinematic objects may be given.
• Due to the anisotropy of deformation for the single crystals, it is necessary to work in
3D geometries.
• In the absence of SLIP INTERACTION (see page 13.92), h1=1. and all other values are
zero (i.e. only self-hardening).
• The axes with respect to which the slip systems are defined (and thus the slip systems
themselves) can be rotated with a *rotation command (see page 2.15 for the correct
syntax).
• The *store all command stores all local associated force variables (such as the resolved
shear stresses) as well as the internal hardening variables.
Example:
***behavior gen_evp
**elasticity cubic y1111 162321.0 y1122 78075.0 y1212 110615.0
**potential octahedral
ev
*flow norton
K
100.0
n
10.0
*kinematic nonlinear_phi
C
20995.0 D
1105.0
phi 0.0
delta 0.0
Xbar 23.8
*isotropic nonlinear
R0 382.0
Q
7.93
b 2420.0
*interaction slip
h1 1.0
h2
1.1
h3 1.3
h4 1.5
h5
1.7
h6 1.9
**potential cubic cu
*flow norton
K 100.0
n
10.0
*kinematic nonlinear_phi
C 50000.0
D
1000.0
phi
0.0
delta 0.0
Xbar -6.0
*isotropic nonlinear
R0 382.0
Z-set — Non-linear material
& structure analysis suite
13.81
<POTENTIAL>
Q
-5.6
b
2429.0
*interaction slip
h1 1.0
h2 1.2
h3 1.1
**interaction iso ev cu
h
0.2
***return
13.82
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
<POTENTIAL> gen evp
Description:
The potential object of type gen evp serves as the basic type for classical plasticity and
viscoplasticity models with both isotropic variable and an arbitrary number of kinematic
hardening variables (c.f. [Lema85]). The potential will accept a wide variety of criterion types,
associated and non-associated, as well as a variety of flow rules (plastic and viscoplastic). The
dissipation potential for this model is written generally:
Ωi = fcr (p, σ − ΣXi ) + Ωr (R) + ΣΩαi (Xi )
If there are hardening variables, they will be stored in the following order:
h = [α1 α2 . . . αn r]
where the tensorial variables αi are the kinematic internal variables (analogue to an offset
strain), and r is an internal variable modeling the isotropic expansion or contraction of the
yield domain (analogue to an equivalent strain).
Syntax:
The syntax understood by this potential is summarized below:
**potential gen_evp [name]
[ *flow
<FLOW>
[ *criterion
<CRITERION>
[ *kinematic
<KINEMATIC> [name]
[ *isotropic
<ISOTROPIC>
[ *var_coefs
[ *store_all
]
]
]
]
]
]
The option *store_all is used to make all associated force variables to be stored as well
as the internal hardening variables. This allows one to observe directly the effective back
stresses or isotropic radius even in the case of coupling.
Other statements may be made about this model:
• The final form of the hardening will be determined by the options *kinematic and
*isotropic which are given by the user. If there is incompatibility with one of the
hardening mechanisms with this potential, an error message will be output as the invalid
calculation is attempted. Static recovery is allowed in both the isotropic and kinematic
variables.
• The type of flow law, λ̇, and criterion, f , will be determined by the options *flow and
*criterion. Default values for these options are plasticity (time independent flow)
and mises respectively.
• The flow direction is determined by the criterion chosen, which is not necessarily associated.
• The names of the internal variables will be dependent on the choices given by the user.
If no name is given for the potential, the potential name (henceforth referred to as pn
will be “e#” with # being the sequential number of the potential in question (i.e. 1
Z-set — Non-linear material
& structure analysis suite
13.83
<POTENTIAL>
for the first potential, etc). In the absence of names for the kinematic variables (kn),
default names will be of the form: “pna_#v” with # being the sequential number of the
kinematic variable in the potential’s kinematic list.
With these comments, the internal variables added by a gen evp potential instance are
the following2 :
prefix
size
description
default
pnvi
pncum
T-2
S
yes
yes
kn
T-2
inelastic strain tensor
cumulated value of the multiplier
λ̇
kinematic strain variable
no
Note that the cumulated value of the λ̇ multiplier is only sometimes equal to the equivalent
cumulated inelastic strain (for Von Mises criterion for example, but not for Hill). The current
version
of the potential does not allow for calculation of the true equivalent for cases where
R
λ̇ 6= p.
Example:
An example viscoplastic model with Norton flow, Hill type criterion and kinematic hardening
is given here as an example using the gen_evp potential. The dissipation potential (flow term
only) and criterion may be written as:
Ω
2
= [(σ − X) : M : (σ − X)]1/2 − R + 3D
2C X : X + (R − Ro ) /2Q
1/2
f
= [(σ − X) : M : (σ − X)] − R
(11)
which leads to the evolution rules:
˙ vp
∂f
2 M:(σ −X)
= λ̇ ∂∂Ω
σ = λ̇ ∂ σ = λ̇ 3
σeq
(12)
with σeq = [(σ − X) : M : (σ − X)]1/2 .
***behavior gen_evp
**elasticity isotropic
young 260000.0
poisson 0.3
**potential gen_evp ev
*flow norton
n 7.0
K 400.
*criterion hill
hilla 1.
hilld 1.
hillb 2.
hille 1.
hillc 3.
hillf 1.
*kinematic nonlinear x1
2
13.84
This of course means in addition to those created by other potentials and the gen evp behavior itself.
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
C 30000.0
m 1.0
D 500.0
M 20000.0
*isotropic constant
R0 130.0
***return
Z-set — Non-linear material
& structure analysis suite
13.85
<POTENTIAL>
<POTENTIAL> gen evp2
Description:
The gen_evp2 potential is a modified version of gen_evp with different assumptions of hardening and variables. The potential allows use of DIRECT KINEMATIC objects and RATE VAR FLOW
laws which can both have arbitrary numbers of integrated variables, and the kinematic law
does not limit the calculation of Xi as a linear modulus function of αi .
Syntax:
The syntax understood by this potential is summarized below:
**potential gen_evp [name]
[ *flow
<FLOW>
[ *flow
<RATE_VAR_FLOW>
[ *criterion
<CRITERION>
[ *kinematic
<KINEMATIC> [name]
[ *kinematic
<DIRECT_KINEMATIC>
[ *isotropic
<ISOTROPIC>
[ *var_coefs
[ *store_all
]
]
]
]
[name] ]
]
]
]
Example:
This potential is nominally the same as the original gen_evp potential, except for allowing
the rate flow type and the direct kinematic types.
***behavior gen_evp
**elasticity isotropic
young 260000.0
poisson 0.3
**potential gen_evp2 ev
*flow norton
n 7.0
K 400.
*criterion hill
hilla 1.
hilld 1.
hillb 2.
hille 1.
hillc 3.
hillf 1.
*kinematic armstrong_frederick
C 30000.0
m 1.0
D 500.0
M 20000.0
*isotropic constant
R0 130.0
***return
13.86
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
<POTENTIAL> suvic
Description:
This is the SUVIC model which applies well to ductile rock material such as sodium chloride
materials over long term creep loading. There are isotropic and kinematic hardenings with
static recovery, as well as evolution in the creep law and changing saturation values for the
hardening variables.
λ̇ = A
f (σ − B) − R − R0
K + K0
˙ v = λ̇n
n=
n
∂f
∂σ
Bi
J(Bi ) − Bi00 q
α̇i = λ̇ n − 0 −
Bi
MB i
R
kRk − R00 p
ṙ = λ̇ 1 − 0 −
R
MR
K
kKk − K 00 p
k̇ = λ̇ 1 − 0 −
K
MK
Bi
R
K
K0
Mi = C(A1i /A2i )1/q
Mr = C(A3 /A4 )1/p
1/u
h Mk = C(A
im 5 /A6 )
0 ln(λ̇/v )
B0i = Bi0
0
h
im
R0 = R00 ln(λ̇/v0 )
h
i
= J(σ) ln(λ̇/v0 ) − B 0 − R0 (λ̇/A)−1/n
= A1i αi
= A3 r
= A5 k
Z-set — Non-linear material
& structure analysis suite
(13)
13.87
<POTENTIAL>
Example:
***behavior gen_evp
**elasticity isotropic
young
18900.
poisson 0.25
**potential salt ev
*criterion mises
*flow norton_k_variable
n 4.0
*kinematic nonlinear
% short range
A1 6400.0
Bp equivalence Bpa0v
*kinematic nonlinear
% long range
A1 40.0
Bp = Bpa1v
*isotropic non_linear_recovery
A
1800.0
R0
0.0
Rp = Rp
*K non_linear_recovery
R0
0.04
A
260.0
Rp = Kp
*parameters
A
6.00e-7
N
4.0
% must be duplicated!
vdot0
1.e-13
sig0
4.63
R0
1.38
B01
0.83
B02
1.25
m
0.81
***return
13.88
Z-set — Non-linear material
& structure analysis suite
<POTENTIAL>
<POTENTIAL> 2M1C
Description:
The potential of type mises 2m1c exists for the particular case where there are two flow
mechanisms which act under a single criterion. This model also allows interaction between
the isotropic hardening variable and the kinematic back stresses. The model is particularly
useful for accurate modeling of ratcheting phenomenon [Cail95].
0.5
f = f12 + f22
Syntax:
**potential 2M1C [name]
*criterion
mises_2m1c
[ A1 <COEFFICIENT> ]
[ A2 <COEFFICIENT> ]
[ *flow
<FLOW>
[ *kinematic
<KINEMATIC>
[ *isotropic
<ISOTROPIC>
[ *coefficient
[ C12
<COEFFICIENT>
[ a
<COEFFICIENT>
[ beta <COEFFICIENT>
]
[name] ]
]
]
]
]
]
• The model requires giving mises 2m1c as a criterion type. This criterion will accept
coefficients A1 and A2 (scalar) to simulate a localization process in each of the two
mechanisms. In this case the stress equivalent terms in the criterion will be calculated
as:
fi = J(Ai s − Xij ) i = 1, 2
with j kinematic hardenings in each mechanism.
• In the event that the isotropic hardening variable is coupled to the kinematic back stress
(type nonlinear bsi), the radius will be calculated with kinematic interaction as:
1
R(α1 , α2 ) = Ro − kb(α1 + α2 ) : (α1 + α2 ) + Qr
3
A corresponding isotropic interaction is introduced into the kinematic hardening variable:
2 o
Xi = [Cij
+ k(1 − br)]αj
3
i, j = 1, N
where i are the mechanisms and j are the kinematic variables in each mechanism.
• The coefficients a and b indicate that we desire the calculation of the coefficients C
for the kinematic hardening and C12 (kinematic interaction) for special forms of the
interaction matrix (e.g. zero determinant)
Z-set — Non-linear material
& structure analysis suite
13.89
<POTENTIAL>
<POTENTIAL> z6 gen evp
This potential gives hardening variable evolution in terms of the un-coupled internal state
variables and not as a function of the associated forces. The criterion still includes the coupled
terms, of course. This type of interaction was given in versions 6 and below. In the case of
no interaction this potential is identical to the gen_evp potential.
Example:
2
2
XX1 = CX1 αX1 + Cint αX2
3
3
2
2
XX2 = CX2 αX2 + Cint αX1
3
3
but the evolution of the internal variables has no coupling. The evolution equation for a
nonlinear kinematic hardening variable is changed to:
mkin = n − Dα
13.90
Z-set — Non-linear material
& structure analysis suite
<SINTERING STRAIN>
<SINTERING STRAIN>
Description:
The mechanisms of sintering may be represented by a sintering strain s .
Syntax:
type
[ anisotropic
a1 COEFFICIENT
a2 COEFFICIENT
a3 COEFFICIENT
]
model coefficient COEFFICIENT
...
The replacements available for type are the following
DESCRIPTION
CODE
non_linear
f˙ = −A(f − f∞ )n
coefficients : A, n, f∞
noms : A, n, f_inf
The option anisotropic permits to enter a non-isotropic strain with the use of the three
coefficients a1, a2 and a3. The strain is then defines as:
˙11 = a1 ˙s
˙22 = a2 ˙s
˙33 = a3 ˙s
we must then have:
a1 + a2 + a3 = 3
Z-set — Non-linear material
& structure analysis suite
13.91
<SLIP INTERACTION>
<SLIP INTERACTION>
Description:
The SLIP INTERACTION object is used to model the inter-system hardening of the isotropic
variable for single crystal models. In the absence of this option, only self-hardening is
taken into account (h1=1. and all other interaction parameters equal 0.). For each crystal potential, the maximum number of interaction parameters is given in the section on
<CRYSTAL ORIENTATION> (page 13.31). As an example, the hardening matrix describing the
inter-system hardening for octahedral slip systems is shown at the end of this section.
Note: hardening between different crystal potentials (for instance between cubic and octahedral systems) can be taken into account through the INTERACTION object (see page 13.52).
Syntax:
The following models are implemented for slip interaction in single crystals.
DESCRIPTION
standard Zebulon form.
CODE
slip
slip This is the “standard” Zebulon format for full slip interactions described by [Meri91].
The coupled radius Ri0 for system i is
Ri0
= R0 +
n
X
hik (Rk − R0 ) ,
k=1
where Rk is the radius of slip system k as calculated from the *isotropic variable
option (e.g. R0 + Q(1 − e−bv )).
Bd
Ba
Bc
Db
Dc
Da
Ab
Ad
Ac
Cb
Ca
Cd
13.92
h1
h2
h2
h4
h5
h5
h5
h6
h3
h5
h3
h6
Bd
h2
h1
h2
h5
h3
h6
h4
h5
h5
h5
h6
h3
Ba
h2
h2
h1
h5
h6
h3
h5
h3
h6
h4
h5
h5
Bc
h4
h5
h5
h1
h2
h2
h6
h5
h3
h6
h3
h5
Db
h5
h3
h6
h2
h1
h2
h3
h5
h6
h5
h5
h4
Dc
h5
h6
h3
h2
h2
h1
h5
h4
h5
h3
h6
h5
Da
h5
h4
h5
h6
h3
h5
h1
h2
h2
h6
h5
h3
Ab
h6
h5
h3
h5
h5
h4
h2
h1
h2
h3
h5
h6
Ad
h3
h5
h6
h3
h6
h5
h2
h2
h1
h5
h4
h5
Ac
h5
h5
h4
h6
h5
h3
h6
h3
h5
h1
h2
h2
Cb
h3
h6
h5
h3
h5
h6
h5
h5
h4
h2
h1
h2
Ca
h6
h3
h5
h5
h4
h5
h3
h6
h5
h2
h2
h1
Cd
Bd
Ba
Bc
Db
Dc
Da
Ab
Ad
Ac
Cb
Ca
Cd
(111)[101]
(111)[011]
(111)[110]
(111)[101]
(111)[011]
(111)[110]
(111)[011]
(111)[110]
(111)[101]
(111)[110]
(111)[101]
(111)[011]
Z-set — Non-linear material
& structure analysis suite
<STRAIN NUCLEATION>
<STRAIN NUCLEATION>
Description:
This option allows application of different methods from which the nucleation of porosity may
occur. Syntax:
The nucleation mechanisms are defined as follows:
[ crack_like ]
type
[ yield COEFFICIENT ]
[ porosity_yield COEFFICIENT ]
model coefficient COEFFICIENT
...
crack like keyword indicating that the mechanisms of porosity creation takes the form of
cracking or “loss of material.”
yield allows to define a value of plastic strain below which there is no nucleation.
porosity yield
nucleation.
Defines a value for the trace of the stress tensor below which there is no
The law of nucleation is given by:
f˙n = A(. . . )λ̇
where λ̇ is the plastic multiplier. The form of the function A is variable according to the type
of nucleation chosen. The table below summarizes the possible types of nucleation function:
gaussian
DESCRIPTION
˙
fn = Aλ̇
2 fn
1 (p−n )
˙
√
fn = 2πs exp − 2
λ̇
sn
exponential
f˙n =
CODE
constant
n
B
0
exp − p0 λ̇
The corresponding syntaxes are:
constant
A COEFFICIENT % A
gaussian
en COEFFICIENT
fn COEFFICIENT
sn COEFFICIENT
exponential
B COEFFICIENT
e0 COEFFICIENT
Z-set — Non-linear material
& structure analysis suite
%
%
%
n
fn
σn
%
%
B
0
13.93
<THERMAL STRAIN>
<THERMAL STRAIN>
Description:
This object is used to define the calculation of a volumetric thermal strain.
Syntax:
The syntax for this object requires only input of the model coefficients (standard form). The
types of thermal strain available are the following:
isotropic
DESCRIPTION
Isotropic thermal dilatation
anisotropic
Anisotropic thermal dilatation
CODE
The default type is isotropic.
The coefficients for the thermal strains are “secants.” That is to say that the thermal
deformation th is given by the relation:
th i = αi (T, . . . )(T − Tref )
where αi (T, . . . ) is the secant dilatation coefficient, T is the temperature, and Tref is a reference
temperature (zero by default).
NOTE : unexperienced users usually get confused about the exact signification of Tref .
Let us consider, to clarify everything, the calculation of the incremental thermal strain at a
given time t (it is assumed that the computation starts at t = 0 without thermal strain):
∆ε0→t
th = α(T (t)) (T (t) − Tref ) − α(T (t = 0)) (T (t = 0) − Tref )
In the previous equation, the dilatation coefficient α(T ) is supposed to be measured using
Tref as the reference temperature. Do not make any confusion between the initial time, which
is the time at the beginning of the computation (usually t = 0), and the reference temperature:
Tref is not related to any particular time; it is just the temperature used as ”reference“ by
the people who did the experiment to measure α.
To re-define the reference temperature, a coefficient ref_temperature3 must be input.
The method of calculation is schematically shown below:
ε
∆ε=α(Τ−Τ ref )
α
Tref
3
13.94
temperature
a standard coefficient
Z-set — Non-linear material
& structure analysis suite
<THERMAL STRAIN>
The coefficients required for the thermal deformation models are given below. All coefficient types and dependencies may be entered for the thermal dilatation coefficients.
isotropic The isotropic model has only one coefficient for the dilatational secant which
is named alpha.
anisotropic The anisotropic thermal strain object calculates different thermal strains
along each axis in the material coordinate frame. The dilatational coefficients names
are alpha1, alpha2, and alpha3 to the three material axes.
Example:
**thermal_strain isotropic
alpha
temperature
10.e-6
25.
12.e-6
500.
ref_temperature 100.
Z-set — Non-linear material
& structure analysis suite
13.95
<THERMAL STRAIN>
13.96
Z-set — Non-linear material
& structure analysis suite
Chapter 14
Modifiers
Z-set — Non-linear material
& structure analysis suite
14.1
14.2
Z-set — Non-linear material
& structure analysis suite
<MODIFIER>
<MODIFIER>
Description:
This class is used as a “behavior” wrapper in order to modify the integration method or
the formulation in some way. Primarily, these modifiers are used to transform the domain
of application for the behavior by providing some translation of the primal-dual variable
combination, such as is found for corotational large strain modifications.
CODE
lagrange polar
DESCRIPTION
corotational finite strain equivalent to a GreenNaghdi stress rate formulation (polar decomposition of
the deformation gradient)
lagrange rotate
corotational modification for updated Lagrangian finite
strain using an integrated rotation
lagrange polar no J
polar corotational formulation without a density
correction
lagrange rotate no J
integrated corotational formulation without a
density correction
bifurcation
bifurcation of mechanical solution analysis
auto step
activates automatic sub-stepping
theta method a integration
runge jacobian
mixed integration method associating explicit
(runge kutta) integration of behavior differential equations with a final Jacobian call (as in theta method a)
for the computation of the consistent tangent matrix
runge rollover
replacement of the previous one, that switches to to
mixed integration only when the initial theta method a
integration fails
plane stress
external wrapper enforcing the plane-stress condition
for material behaviors that don’t handle this case
internally
perturbation
activates calculation of a tangent matrix obtained by
perturbation of the loading strain increment tensor
for
implicit
Global convergence may be inhibited by the addition of the finite-strain options, especially
corotational frame ones. In this case, it is advised to use small increment steps or automatic
time stepping options in the global calculation sequences.
Large deformation methods:
Large deformation formulations in Z-mat are available for the majority of material behaviors1
based on corotational transformation of the stress-strain problem into an “equivalent material referential.” Corotational formulations are implemented because they are applicable to
1
excepting behaviors with explicitly different treatment of finite strain
Z-set — Non-linear material
& structure analysis suite
14.3
<MODIFIER>
material models with tensorial internal variables and anisotropic response without modifying the local evolution rules [Lade80]. Description of the models uses the following standard
expressions:
D
=
1
2 (L
L
= ḞF−1
T
+L )
Ω = 12 (L − LT )
(1)
where F is the deformation gradient and L the rate of deformation. Both these tensors are
non-symmetric. The stretch rate D and rotation rate Ω are symmetric and non-symmetric
respectively. These methods require elements of the type updated lagrangian (specified with
the option ***mesh in the .inp file).
14.4
Z-set — Non-linear material
& structure analysis suite
<MODIFIER>
<MODIFIER> lagrange polar
The lagrange polar modifier is a corotational formulation for finite strain leading to an
equivalent Green-Naghdi stress rate in elastic problems [Gree65, Nagt82]. The stretch rate
tensor is transformed into a local strain rate measure through the following expression:
ė = RT DR
where the rotation tensor R is found by the polar decomposition of the deformation gradient:
F = RU
with R being a pure rotation and U a pure stretch tensor.
Using the transformed “material” strain rate e, the behavior may be evaluated as in small
deformation. An conjugate stress results from the material behavior integration, S which is
transformed to the global Cauchy stress as:
σ = det−1 (F) R S RT
Z-set — Non-linear material
& structure analysis suite
14.5
<MODIFIER>
<MODIFIER> lagrange rotate
The lagrange rotate modifier is a corotational finite strain formulation based on an
integrated rotation tensor [Lade80]. This method yields an equivalent stress-strain response to
a Jaumann-rate formulation but allows tensorial internal variables and anisotropic relations
to be used as was in the polar decomposition case [Lade80].
The principle of this method is to always evaluate the material within a local “material”
referential. Passing from the global to material referential is made using a rotation tensor in
the following relations:
Q̇
= ΩQ
ėtot = QT DQ
Qo = 1
(local strain)
σ
=
det−1 (F) Q S QT
(2)
This method requires additional storage at each Gauss point for the non-symmetric rotation tensor, and the local material strain. The name of strain used in the material law is
therefore changed to ETO## in place of eto##. This strain is also observed to be a logarithmic
strain.
For the two methods, the nonlinear material tangent Dnl describing the ∂∆S/∂∆e relationship is modified for the additional large strain non-linearities as follows:
D0nl = det−1 (F) ADnl AT
Aijkl = Qik Qjl
which relates ∂∆σ/∂∆tD. The polar decomposition method will of course use R in the place
of Q in the above.
One may note an improved convergence for conditions where the material behavior calculation of Dnl is best (i.e. implicit θ-method integration with θ = 1). The calculation of
the stress is also noted to depend on the determinant of the total value of the deformation
gradient (F from time t0 to the current time). Nonlinearity in this term (especially in trial
solutions during iterations) may limit the convergence of the method.
14.6
Z-set — Non-linear material
& structure analysis suite
<MODIFIER>
<MODIFIER> plane stress
Description:
The plane stress modifier allows any material behavior to be run in plane stress conditions
without extra terms being added to the model. This is at the cost of extra local iterations
but with the benefit that extremely complex cases can be treated effectively (e.g. multi-mats,
runge-kutta models, multi-potential deformations, multiplicative finite strain, etc).
Syntax:
The following options are available after the **plane_stress_controls heading.
*skip first The code attempts to predict the plane strain effect ahead of time, which in
some cases can push the nonlinear computation in the wrong direction. Using this flag
prevents that step.
*eps default from theta method or 10−9
*iter default from theta method or 200.
Example:
The modifier order is important for this object to be used properly. Some typical examples
follow:
***behavior gen_evp plane_stress
***behavior gen_evp plane_stress auto_step
***behavior gen_evp lagrange_rotate auto_step plane_stress
Z-set — Non-linear material
& structure analysis suite
14.7
<MODIFIER>
<MODIFIER> auto step
Description:
The auto step modifier applies a “local” method of substepping for any material behavior,
in an attempt to pass divergences and improve accuracy without cutting back the global time
step.
*divergence div [ num ] takes parameters div (real) for the division factor to make when
a divergence occurs, and num (int) for the maximum number of successive divergences
allowed (default 10).
*security [ factor ] set the maximum amount the step can enlarge (e.g. 1.2).
*min dtime takes a real value to set the minimum dtime before giving up and calling for
a global divergence.
*use last tangent use the last computed tangent matrix instead of averaging the sub-step
tangents.
*re solve re-solve the step initializing with the sub-step result increment, possibly leading
to better global convergence.
*forceit force sub-stepping, primarily for testing the process.
*limit takes a list of integrated variable names (VINT) and maximum change values to
ensure that any sub-step will not have any greater increment than this one.
*full step jacobian use the resulting computed full-step variable increment to compute
a new tangent matrix in place of the averaged sub-step tangents.
Example:
The following is a typical example.
***behavior gen_evp auto_step
**elasticity isotropic
young 170000.0
poisson 0.30
**plane_stress
**potential gen_evp ev
*flow norton
n 6.0
K 800.
*kinematic nonlinear X1
C 3000.0
D 50.0
*kinematic nonlinear X2
C 3000.0
D 50.0
*isotropic constant
R0 10000.
14.8
Z-set — Non-linear material
& structure analysis suite
<MODIFIER>
***stepping_controls
*divergence 2. 50
*min_dtime
1.e-8
*limit evcum 0.01
***return
Note:
For finite strain applications this will only work in the order as in the following example:
***behavior gen_evp lagrange_rotate auto_step
Z-set — Non-linear material
& structure analysis suite
14.9
<MODIFIER>
<MODIFIER> runge jacobian
Description:
This modifier activates a mixed explicit-implicit integration scheme performed in 2 steps.
(i) integration of the set of material integrated variables Vi using an explicit runge-kutta
method:
Vit+∆t = runge kutta(t, ∆t, ∆, Vit )
with t the time at the beginning of the step, ∆t the time increment, and ∆∼ the strain
increment loading.
(ii) once known Vit+∆t calculated during step (i), perform a single implicit iteration to obtain
an evaluation of the consistent tangent matrix ∂∆σ
∂∆
∂∆σ
= calc grad f (t, ∆t, ∆, Vit+∆t , ∆Vi )
∂∆
where calc grad f () is the Jacobian calculation function called at each iteration of the
implicit theta_method_a integration method.
This scheme then combines the benefits of the 2 base local integration methods:
• automatic sub-stepping with error control (runge_kutta) efficient for stiff set of differential equations,
• good-quality consistent tangent matrix as provided usually by one-step implicit methods.
and can be used when theta_method_a fails to converge, or in replacement of the previously
described auto_step modifier.
Syntax:
***material
*integration runge_kutta ...
***behavior behavior name runge_jacobian
[ ***controls
*theta theta ]
***return
where the optional command:
*theta theta allows to specify the θ value used for the tangent-matrix calculation of step
(ii). θ = 1 correspond to the fully implicit case (default), while θ = 0.5 may be used to
trigger semi-implicit integration.
Example:
***behavior gen_evp runge_jacobian
***behavior gen_evp runge_jacobian lagrange_rotate
14.10
Z-set — Non-linear material
& structure analysis suite
<MODIFIER>
<MODIFIER> runge rollover
Description:
This definition is an alternative to runge_jacobian (see page 14.10), where the mixed scheme
described previously is only attempted when a single-step theta_method_a implicit integration fails to converge.
Syntax:
***material
*integration theta_method_a ...
***behavior behavior name runge_rollover
[ ***runge_rollover
*runge eps norm iter nit min_dtime dtmin ]
***return
where the optional command:
*runge can be used to enter non-default values for the runge_kutta integration method
parameters (see page 2.13). Default values are eps=norm=1.e-3, and no control on
the maximum number of runge-kutta steps (nit=∞), or the minimum step size during
sub-stepping (dtmin=1.e-19).
Example:
***behavior gen_evp runge_rollover
***behavior gen_evp runge_rollover lagrange_rotate
Z-set — Non-linear material
& structure analysis suite
14.11
<MODIFIER>
<MODIFIER> perturbation
Description:
∂∆σ
This modifier triggers the calculation of a tangent matrix
obtained by perturbation
∂∆
of each component of the input strain increment tensor ∆ (or ∆F for finite-strain models).
This particular scheme may be used for material behaviors that don’t implement a proper calculation of a full consistent tangent operator (runge-kutta only ZebFront models for example).
In this case, global convergence of FE equilibrium iterations may be greatly improved. However, the increase of global time-step allowed by this option, is somewhat counter-balanced
by an increase of local integration cost: for small-strain 3D formulations, one additional integration is needed for each perturbation of the 6 individual components of the symmetric ∆
tensor.
Syntax:
***behavior behavior name perturbation
[ ***perturbation_controls
[ *perturb pert ]
[ *symmetrize ]
[ *one_sided | two_sided ]
]
where the previous optional command are described hereafter:
*perturb can be used to set the value of the perturbation parameter
(default value is pert=1.e-9)
*symmetrize this-command activates symmetrization of the tangent operator
(default is no)
*one sided
*two sided those commands control the way the input strain tensor is perturbated to
calculate each column of the tangent operator. In the two_sided mode (default), two
integrations are needed to compute each column of the tangent (perturbation of each i
component by a value of +/-pert), while the one_sided option only needs one (perturbation by a value of +pert in this case).
Example:
***behavior norton perturbation
...
***return
14.12
Z-set — Non-linear material
& structure analysis suite
<MODIFIER>
<MODIFIER> explicit
Description:
This behavior modifier is used with ABAQUS/Explicit to provide the total integrated corotational strain. One may equally use the lagrange_polar and lagrange_rotate modifiers,
but these duplicate many calculations already done by ABAQUS before entry into VUMAT,
and therefore are costly.
Z-set — Non-linear material
& structure analysis suite
14.13
<MODIFIER>
<MODIFIER> bifurcation
Description:
This behavior is used to make bifurcation analysis. At each iteration, it find the solution of
the following minimization problem :
min ~n : C
:
~
n
∼
~
n∈R3
∼
C
is usually the consistent tangent operator (which means that the solution to the previous
∼
∼
problem depends on the time discretization), but may also be the tangent one. The solution
vanish to zero when localization occurs. It works with any mechanical behavior, and add the
following auxiliary variables :
theta
phi
det_min
the first angle of ~n
the second angle of ~n (available only in 3D)
the minimum of the determinant.
Note that the coefficient of the underlying real behavior may depend on these auxiliary variables.
It is possible to specify
***bifurcation_controls
**non_linear_bif
to take into account non linear geometry which induces new terms in the minimization, coming
from the Jaumann derivative.
14.14
Z-set — Non-linear material
& structure analysis suite
Chapter 15
Model Simulation
Z-set — Non-linear material
& structure analysis suite
15.1
15.2
Z-set — Non-linear material
& structure analysis suite
Model Simulation
Model Simulation
Description:
As part of the Z-set package there is a model simulation mode, which can be applied to
any differential model as supplied by the user, or used to simulate material behavior for a
volume element based on one of the standard behaviors, or a ZebFront behavior (see the
developers handbook for a description of the latter). To allow rapid prototyping of models,
ZebFront modes are available in either a pure model definition, or (with small additions) to
user behaviors defined in this manner.
For mechanical material behaviors, some special functionality is provided. This includes
generalized mixed-mode loading, and the visualization of yield surfaces. Note that curve output of a finite element calculation may be obtained using the **curve command (in Zebulon),
and then fed into the simulator for loading to obtain more detailed output or yield surface
visualization.
Syntax:
% Zrun -S problem ←More Information:
For more information on simulations, please refer to the following manual sections, and example files.
• Tests in the directory $Z7PATH/test/Simulator_test where there are many examples
of input files exercising all the simulator options.
• Examples handbook and corresponding example files. There is some information showing basic use (the Simulation chapter), and many more examples in the Material
Behaviors chapter. In fact, most of the behavior outputs are generated with the simulation module.
• Developer handbook to see how to make simulation compatible models including
ZebFront.
Z-set — Non-linear material
& structure analysis suite
15.3
****simulate
****simulate
Description:
This command marks the beginning of a simulator calculation definition. The Z-set program
in simulation mode (-S switch) will search this command and interpret all the sub-commands
until the next **** level command is reached. More than one simulation block may be
included in the input file, which are accessed using the -N <num> command line option.
Syntax:
The simulation accepts the following syntax for the problem definition:
****simulate
***test test name
...
***model
model definition
***solver solver-type name
solver options
****return
The definition of a ***test is used to specify particular tests to simulate. The tests will be
run through in the order of their appearance, using the default output file prefix of test name.
The test test name is required.
In each test, different loading conditions, output, models, etc may be defined. See the
following pages for these commands and their syntax.
Options ***model and ***solver are available1 as a means to change the default model
definition and the default solver method. It can be a great convenience to define these at
the ***-level in simulations where there are many tests of the same material. The command
syntax for ***model is the same as in the ***test section, as found on page 15.13 as is the
command ***solver found on page 15.15. Note that this includes the *-level.
Example:
The following is an example of a complete simulation file with a single test definition. See the
following pages for descriptions of the individual commands.
****simulate
***test funny
% define a test name which defines the default output
**load
% start the loading section
*segment 100
% 100 outputs per load segment below
time
sig11 eto22 sig33 eto12 sig23 sig31
0.0
0.
0.
0.
0.
0.
0.
1.0
0.
1.e-3 0.
0.
0.
0.
2.0
0.
1.e-3 0.
.5e-3 0.
0.
3.0
0.
1.e-3 0.
-.5e-3 0.
0.
4.0
0.
-1.e-3 0.
-.5e-3 0.
0.
5.0
0.
-1.e-3 0.
.5e-3 0.
0.
**model
% standard model definition with
*file funny.mat
% integration, file name, rotation
*integration runge_kutta 1.e-3
% and other options
1
15.4
Introduced in 8.2
Z-set — Non-linear material
& structure analysis suite
****simulate
**yield_surface yield0.test
*degrees
5.
*eps
1.e-12
*component sig22 sig12
*time
5.0
**output
time eto22 sig22 eto12 sig12
****return
Z-set — Non-linear material
& structure analysis suite
% make a yield surface output
% the default ascii file for output
% is funny.test (from above ***test def).
15.5
****simulate
***test
***test
Description:
This command marks an individual test condition to simulate. Loadings, output, and model
specification will be contained within this option. The output file name will be determined
by the test name given after the command. Loadings and outputs may be repeated in the
input sequence to create complex cases.
The model definition must be compatible with the solver type chosen. Many simple
simulation models are only compatible with the Runge-Kutta integration, with other FEAspecific material models with implicit integration only must be used with the iterative Newton
solver. If the model is not compatible, a run-time error message will be issued.
Syntax:
The definition of a test case within the simulation will follow the syntax below:
***test test name
**load
**model type
**output
**solver type
**time_ini
**time_end
**var_mat_ini
**yield_surface
**error_map
**constant_parameter
**simulation plugin
Each of these commands takes at least some extra input. The different sub-commands are
all described fully in the following sections. Use of at least one **load command is required,
although multiple entries are possible.
The model must be defined as well, either through the use of a defined default behavior
using the ***model keyword before the test instance (see page 15.4. For the explicit solver
(more accurate), this may be a simple ZebFront model of type SIMUL_MODEL or a finite element
behavior properly modified for the simulation mode2 . The syntax for a FEM behavior does
not take a behavior type after the **model command. The model name will be found in the
material file after the command ***behavior. Specialty simulation models do require the
type following **model in the test definition.
. . . continuedExample:
An example test definition for use with a finite element behavior was given in the previous
section (page 15.4). A simulation model (1D simulation only) example follows:
****simulate
***test relax12
**time_ini -5.0
**load
*file relax12.load 2
2
presently (July 1997) all the gen evp models are implemented as are some ZebFront FEM models; others
are in development
15.6
Z-set — Non-linear material
& structure analysis suite
****simulate
***test
**model ddi
*file relax12.mat
*integration runge_kutta 2.e-3 2.e-3
**output time eto sig evcum epcum
****return
Z-set — Non-linear material
& structure analysis suite
15.7
****simulate
***test
**constant parameter
**constant parameter
Description:
This option is used to set parameters that are used in the material behavior (temperature for
instance). It replaces the ***parameter bloc in a standard computation, and, as its name
suggests, only allows constant parameter values.
Syntax:
**constant_parameter param1 value1 [param2 value2 ...]
Example:
**constant_parameter
temperature 200.
depth
60.
15.8
Z-set — Non-linear material
& structure analysis suite
****simulate
***test
**error plot
**error plot
Description:
This option is used to get error maps for integration algorithms Under non-proportional
loading.
Syntax:
The error maps to output are defined in the following manner:
**error_map output name
*component comp1 comp2
*direction1 dir
*direction2 dir
*make_contour
*number
num
*scale
factor
*time
times
*component
*direction1
*direction2
*make contour
*number
*scale
*time
Example:
**error_map shear_error.test
*time
20.
% after loading to the yield surface
*scale
0.1
% ten percent strain
*make_contour
% look at with Zmaster shear_error
Z-set — Non-linear material
& structure analysis suite
15.9
****simulate
***test
**load
**load
Description:
This command defines a segment of loading for the simulation. The load commands are
executed in the order they appear, and there can be any number of loads for the problem.
Blocks of loading defined seperately are usefull to change control, or apply cycles in between
single segements of loading.
The command looses some coherence in an attempt to give additional options, and due
to the fundimental difference in the loading for fem behaviors and the generalized simulator
models.
Syntax:
Simulation loading is given with the following syntax:
**load [ cycle ] ncycle
*rate var rate for dval
*rate var rate stop_at break var dval dtime
*segment rate [num ]
*segment [num ]
loading table
*file filename num
The cycle keyword indicates that the loading is to be repeated a given number of times.
This may also be useful in the rate loading case for switching between rates (see example
following).
num The number of output points between each given loading segment in the loading table.
ncycle The number of times the loading segment will be run in the event of the cycle
keyword.
loading table a tabular form which describes the loading for the simulation. The table should
have time as the first column, followed by an appropriate list of imposed variables.
For tensor loading with fem behaviors, the size of the behavior (dimension) will be
determined from the number of components given. Thus giving 4 variables in small
deformation will give 2D behavior, while 6 gives 3D. Note that some models like the
crystal potentials require 3D and thus 6 componnets to be given. External parameters
are imposed at this level as well, using the prefixed nameing param:var name. Note:
initial values are required.
*rate this option is used for rate loading in the generalized simulator only (i.e. not
for fem behaviors). The command allows simple rate loading to be defined for a given
duration (with the for keyword), or with breakpoints which are to be determined (with
stop_at).
*segment a segment of loading. The optional keyword rate indicates that values are in
rate form.
*file indicates that the loading table will be given in the external ASCII file with name
given by filename. The syntax is exactly that desribed by loading table above. Note if
you use FEM output from a **curve option the leading pound sign must be removed
from the variable list.
15.10
Z-set — Non-linear material
& structure analysis suite
****simulate
***test
**load
. . . continuedExample:
An example with multiple load segments and cycles follows. Note that the second load segment
uses a time variable offset by the ending of the first segment.
****simulate
***test biaxe
**load
*segment time sig11 sig22
sig33 sig12
0.0 0.0
0.0
0.0
0.0
1.
0.
100.0
0.
0.0
**load cycle 4
*segment
time sig11 sig22
sig33 eto12
0.0 0.0
100.0
0.0
0.0
1.0 0.0
100.0
0.0
0.02
3.0 0.0
100.0
0.0 -0.02
4.0 0.0
100.0
0.0
0.0
**model
*file biaxe.sim
*integration runge_kutta 1.e-3 1.e-3
**output time eto22 sig22 eto12 sig12 evcum
****return
The second example imposes an external parameter temperature which can be used in the
behavior through variable coefficients or a parametric strain definition.
****simulate
***test genevp_variable_temp
**load
*segment time sig11 sig22
sig33 sig12 param:temperature
0.0 0.0
0.0
0.0
0.0
0.0
1.0 0.0
210.0
0.
0.0
0.0
2.0 0.0
210.0
0.
0.0
100.0
**model
*file genevp_variable_temp.sim
*integration runge_kutta 1.e-2
**output time eto22 epcum sig22 temperature R0 C sig::mises
****return
. . . continuedAnother useful loading employes the rate form to convienantly give a pre-loading,
followed by an anternating strain rate:
****simulate
***test rate_loading
**load
*segment 5
time
sig11 eto22 sig33 sig12
0.0
0.
0.
0.
0.
1000.
0.
0.01
0.
0.
**load cycle 8
*segment rate 5
time
sig11 eto22 sig33 sig12
4.
0.
0.004 0.
0.
Z-set — Non-linear material
& structure analysis suite
15.11
****simulate
***test
**load
400.
0.
0.004 0.
0.
**model
*file rate_loading.sim
*integration runge_kutta 5.e-2
**output time eto22 sig22
****return
15.12
Z-set — Non-linear material
& structure analysis suite
****simulate
***test
**model
**model
Description:
This command defines the model to be simulated in the current test.
Syntax:
One specifies the simulation one time per test using the following syntax:
**model [ type ]
*coefficients
coef list (constant)
*file name
*integration INTEGRATION
*initial_value
var name value
If the optional model name type is given after the **model command, the model will be of
the generalized simulation type (not fem compatible). In this case, the model coefficients will
be read below the *coefficients command, or in a file with the *coefficients command
in it (terminated with *return). Otherwise the *file command must be used to specify the
external file name with the fem compatible behavior (standard behavior file as described in
the chapter “Material File”).
*file gives the file name for a behavior type model. This can be one of the standard
finite element behaviors, or a custom ZebFront behavior. Note that the filename may
be the same as the input file name (including the suffix). The file will be scanned for
the ***behavior keyword. We normally suggest putting the behavior definition at the
end of the input file.
*integration defines the integration method. This command is equivilent to that for a
FEM material. Currently only the explicit integration (i.e. Runge-Kutta) methods are
supported by the simulator.
*rotation defines a material rotation which will be affected on the gradient/flux combination (for tensorial variables) before and after the loading steps. Note that the rotation
defines the transition from global to material as in the finite element method.
. . . continuedNote:
In the event of rotation and tensorial integrated variables, the integrated variable values which
get output are in the material coordiante system. This will change in the future sometime,
so tests and examples which depend on that will change.
Z-set — Non-linear material
& structure analysis suite
15.13
****simulate
***test
**output
**output
Description:
This option is used to create an ASCII file with the output values from the simulation. Output
is in column format, with a heading of the variable names requested commented with the #
character.
Several output files may be specified, while the default file name is the problem name
(without suffix) appeded with .test.
Syntax:
The yield surfaces to output are defined in the following manner:
**output [ filename | variable list ]
[ *file filename variable list ]
[ *sido ]
[ *precision digits ]
[ *small epsilon ]
[ *frequency options ]
[ *limit_output_var function; ]
variable list a list of variables to be output. The available variables consist of the var-int
and var-aux variables listed when the program is run with the -v command line switch,
and the imposed external parameters.
*file used to specify an output filename.
*sido output compatible with the sidolo program and graphical utilities.
*precision gives the number of significant digits (3 by default) with which the variables
are written.
*small specifies a small value epsilon (1.0e-9 by default) below which the (absolute) output
value is considered zero. Note: this command is not related to the global parameter
Solver.SmallForTest, which plays a similar role for the test output of the finite element
solver.
*frequency similar to the ***output **frequency command of the finite element solver
(see the Z-set user manual).
*limit output var defines a function of one or more output variables (or of time) which
inhibits output if its function value is smaller than 0.5.
Example:
**output example.test
time sig11 sig22 evcum
**output evcum damage
**output
*file strain.test time eto11 eto22 eto33 eto12
*file stress.test time sig11 sig22 sig33 sig12
15.14
Z-set — Non-linear material
& structure analysis suite
****simulate
***test
**solver
**solver
Description:
This command defines the solution method for the simulation. The default solver is an explicit
Runge-Kutta method. In some cases, the Runge-Kutta integration may not be available for
a material model, or greater efficiency can be achieved with use of the implicit integration.
However, use of an implicit solution allows the possibility of divergence. Which approach to
use varies greatly with the problem at hand, and some experimentation and experience may
be required to find the best choice.
Many of the convergence control options below follow the same syntax as in the Zebulon
FEA solver. In fact, the solution parallels very closely the RVE element, but gives improved
efficiency and convenience with the other simulation options.
The current solvers available are as follows3 .
explicit
DESCRIPTION
default explicit solution
newton
iterative solution.
CODE
Explicit method:
The solution of mixed loading in the explicit solver requires some additional equations for the
model derivative function (the function returning rates for all state variables as a function
of the current state). This discussion applies primarily to the mechanical solution procedure,
while other simulation models would have a similar formulation requirement.
Fundamentally, because the solver needs to support a mixed combination of stress and
strain rates, we need to write the complete equation for the rate of change in stress:
σ̇ = Del : [˙ to − ˙ in − ˙ th − . . .] + Ḋel : el
(1)
For rate dependant behaviors, ˙ in , ˙ th , and el are known, while for rate independent behaviors
the inelastic strain rate is a function of the total strain rate, normally of the form:
˙ in = λ̇n =
n : Del : ˙ to
n : Del : n + H
(2)
This term, while not greatly complicating the solution, may not be handled in the simulation
solution for a particular model4 . In that event a rate dependant flow law must be used, for
example using a norton law with very small viscosity to approximate rate independence.
The basic solution proceedure rearranges the mixed equations to first solve for the total
strain rates, and then resubstituting them into the basic rate equation to find dsig. For
example, a 2D solution could take the form:
 


σ̇1
?
?


  = Del : ˙tot2  − Del : [˙ in + ˙ th ] + Ḋel : el
(3)
?
˙tot3 
σ̇4
?
3
versions before 8.2 do not allow the explicit keyword, it is activated by the absence of a **solver entry
starting with 8.2 rate independent plasticity is handled by gen evp and some ZebFront demonstration
models
4
Z-set — Non-linear material
& structure analysis suite
15.15
****simulate
***test
**solver
From which the following equation would be extracted:
−1 Del1 1 Del1 4
σ̇1 1
˙tot11
=
=
˙tot12
Del4 1 Del4 4
σ̇1 2
(4)
Syntax:
The solver syntax accepts a number of options which determine the convergence and automatic
time stepping controls (primarily applicable to the newton solver). syntax:
**solver [ type ]
*algorithm algo-type
*automatic_time
*divergence factor
*iteration iter
*iter_optimal opt-iter
*max_dtime max-time
*min_dtime min-time
*ratio ratio
*security sec
*algorithm
*automatic time
*divergence
*iteration
*iter optimal
*max dtime
*min dtime
*ratio
*security
15.16
Z-set — Non-linear material
& structure analysis suite
****simulate
***test
**yield surface
**yield surface
Description:
This option is used to create an ASCII file with the yield surface data at a load segment start
or end time. Any number of yield surfaces may be given for a single test. The output files
contain a list of the entire flux variable values for the surface taken in order, followed by the
angle.
Syntax:
The yield surfaces to output are defined in the following manner:
**yield_surface output file name
*time
time0 [ time1 time2 ... ]
*criterion num
*potential num
*component comp1 comp2
*degrees
degrees
*angles
angle0 [ angle1 angle2 ... ]
*factor
factor
*eps
eps
*rate
[ rate0 rate1 rate2 ... ]
*find_offset
*offset
offset values
*time A list of one or more real values time0 [ time1 time2 ... ] to define the time at which
the yield surface is evaluated. The times can be any time from 0.0 onward. In the event
that multiple times are given, the surface scans will be given in the same file seperated
by two blank lines.
*criterion This gives the criterion name which will be used for the surface. They are
arbitrary values to match with zero or the rate if given.
*potential A gen_evp behavior potential which will be used for the criterion. By default
all potentials input will be used to establish the material criteria for flow.
*component Two strain names for the two flux components which are scanned5 . An example
would be to give sig11 sig12 for the components.
*degrees The sweep angle which will be used for each step in the surface. Note that
fineness of the surface depends on this, but also that the sweep center be in the center
of the surface. Being close to the surface in one point increases the number of output
points because it dominates the “point of view.”
*angles Instead of using the sweep angle to define the steps, a list of angles (in degrees)
angle0 angle1 angle2 ... can be specified. If both *degrees and *angles are specified,
the *degrees command will be ignored.
5
unfortunately, gradient values cannot yet be used to scan for strain space yields surfaces, etc.
Z-set — Non-linear material
& structure analysis suite
15.17
****simulate
***test
**yield surface
*factor A magnitude of the flux gradient which is considered large for the surface. Look
in the output file for bad value lines (marked with a comment line). If there are bad
values, increase the factor. The default value is 400.
*eps A positive real value eps specifying the degree of convergence in the surface. The
actual value depends on the strain rates as specified by the *rate command.
• If neither *eps nor *rate commands are given, then the default value is 1.e-1.
• If no *eps command is given, and the *rate command is given without specifying
any strain rate, then the default value is 1.e-1.
• If no *eps command is given while the *rate command is given with one or more
strain rates, then the default value is 1.e-6 times the average of the specified strain
rates.
• On the other hand, if the *eps command is given, but the *rate command is not,
then the actual value is eps.
• If the *eps command is given while the *rate command is given without specifying
any strain rate, then the actual value is 1.e-12 times eps.
• Finally, if the *eps command is given, and if the *rate command specifies one or
more strain rates, then the actual value is eps times the average of the strain rates.
*rate Optional rate values for an equipotential surface. If more than one value is given,
the surfaces will be placed in the same file with two blank lines between them. The
default is zero.
*find offset This command is used to find the surface when the current position is
outside of it. A large rate value will be used to hopefully expand the surface such that
the current position is inside, and an average of that surface will be used to estimate the
center. An optional real value following the command sets this large rate factor, while
the default is quite arbitrarily 1.0.
*offset Allows setting the actual offset flux variable (the estimated difference between
current position and yield center) in place of the automated method *find_offset (not
recommended). Each component of the flux tensor should be specified.
Example:
**yield_surface yield_find_offset.test
*degrees
5.0
*factor
1000.0
*find_offset
*time
0.0 50. 200. 500.
*time
1000.
*rate
0.0 1.e-9 1.e-6 1.e-3
15.18
Z-set — Non-linear material
& structure analysis suite
Chapter 16
Optimization
Z-set — Non-linear material
& structure analysis suite
16.1
16.2
Z-set — Non-linear material
& structure analysis suite
Introduction
Introduction
Description:
The version 8.0 of Z-set introduces a generalized optimization module which may be used to
identify material coefficients, optimize geometry, etc. The optimizer will modify tokenized
parameters in a different user specified ASCII files in order to minimize the combined error of
a variety of tests. Each test will generally consist of one or more shell scripts or sub-processes
used for “simulation” of the test, and a certain method of comparison with experimental,
reference data, or optimal condition. No explicit assumption is made to the nature of simulation method so these simulations may be made using the Z-set programs internally, or by
another means. The real strength of this method is one can obtain the best comprehensive
approximation to many data sets, even if they are theoretically over-constrained.
The optimizer problem and its solution are defined in an input text file, similar to that
for the other main program types. The optimizer uses a template file for definition of the
parameters to be modified.
Basic Concepts and Notations:
Optimization problems are formulated in Z-set in the standard form,
N
minimize
F(x) =
1X
wi (f (x, ti ) − y(ti ))2
2
i=1
=
by changing
such that
N
1X
2
wi (fi (x) − yi )2
(1)
i=1
x∈S
gj (x) ≤ 0, j = 1, ng
(2)
F is the cost function (scalar), g is a vector of constraints, x are the parameters to be optimized, ti tags the experiment (experiment number, time, ...), and wi is the weight associated
to experiment i. A variable x (set of parameters) such that gj (x) ≤ 0 , j = 1, ng represents
a feasible point (vice versa infeasible). Z-set’s optimizers are primarily meant for parameter
identification of material behaviors. Therefore, the default cost function F is the least-square
distance between experiments and simulations, and constraints g are used to bound and/or
to relate parameters with each other. Of course, other general type of optimization problems
can be addressed.
There are two main categories of optimization methods, local and global optimizers.
Global optimizers seek x∗ such that
F(x∗ ) < F(x), ∀x ∈ S
Local optimizers look for x∗ such that
F(x∗ ) < F(x), ∀x such that||x − x∗ || < Typically, local methods iterate from a set of variables x in the search space S to another
based on information gathered in a neighborhood of x. Zeroth order optimizers use exclusively
Z-set — Non-linear material
& structure analysis suite
16.3
Introduction
the value of F and g. First order methods additionally use gradF and gradg, second order
methods hessianF and hessiang (or an approximation of grad and hessian). Global optimizers
typically rely on pseudo-stochastic transitions in the search space in order to be able to
escape local optima (we do not consider optimizers based on enumeration of all possible local
optima). Practically, an important difference between global and local optimizers is that
global optimizers are slower to converge, but offer greater guarantees on the quality of the
solution produced. In many cases, convergence of global optimizers is so slow that a solution
cannot be found in a reasonable time.
Optimization methods can also be divided into methods that explicitly handle constraints
(2) (SQP) and the others (Levenberg-Marquardt, Simplex, evolutionary algorithm). Penalizing f is a simple way of transforming a constrained optimization problem into an unconstrained optimization problem:
minimize Fp = F(x) +
ng
X
pi max(0., gi (x))α
i=1
where p is a vector of (positive) penalty parameters. The exponent α is usually taken larger
than 1 (typically 2) for gradient based optimization methods (in order to ensure differentiability). The tricky aspect of penalization is in choosing p. If p is to small, the solution to the
optimization problem will not satisfy the constraints. On the contrary, if p is taken too large,
convergence to the optimum might be difficult. Optimizers that explicitly handle constraints
do not require that the user specifies p. Further details on optimization in mechanics can be
found in [Gür92].
Four principal optimizers are available in Z-set: Levenberg-Marquardt, simplex, SQP, and
an evolutionary algorithm. They span the different categories of optimizers mentioned earlier.
A rapid description of their principle can be found in the relevant sections.
Exceution proceedure:
% Zrun -o problem ←-
16.4
Z-set — Non-linear material
& structure analysis suite
****optimize
****optimize
Description:
This keyword marks the start of an optimization command sequence, which will be terminated
by the next ****-level command.
Syntax:
The syntax is as follows:
****optimize type
***files ...
***values ...
***shell ...
***zrun ...
***compare ...
[ ***enforce ... ]
[ ***function ... ]
[ ***convergence ... ]
[ ***evaluate ... ]
[ ***constraint ... ]
[ ***linear_constraint ...
****return
]
Different optimizer types are available by substituting keywords for type from the following:
CODE
nelder mead
DESCRIPTION
nelder-mead (aka simplex) method (slow but it gets
there)
levenberg marquardt
Levenberg-Marquardt method. (the classic for identification)
augmented lagrangian
augmented lagrangian dual method for constrained
optimization
evolution
evolutionary stochastic method, slow but escapes local
minima
Anatomy of an optimization:
All the optimization methods rely on a centralized method of evaluating the error function, and modifying the simulation using the variables being optimized. Error functions will
be computed using a weighted sum of reference-simulation comparisons specified using the
***compare commands. The variables of the optimization are specified in the ***values section, and must correspond to one or more entries in the “optimization” files listed in ***files.
These optimization files are in fact templates for real files, which alter the simulation result of
commands listed in ***shell or ***zrun entries. The optimizer thus constructs new input
files for the simulations, with the current optimization variables inserted in the appropriate
locations.
Z-set — Non-linear material
& structure analysis suite
16.5
****optimize
The figure below shows a basic diagram of the interaction between the optimizer and any
number of sub-simulations which generate the data to be compared with experimental results.
Comparisons
opt.inp
Test1
Zrun −o opt
Optimize
variables
***shell
sim1
exp1
sim2
exp2
sim3
exp3
Test4
sim4
exp4
Test5
sim5a
sim5b
sim5c
exp5a
exp5b
exp5c
Test2
Test3
Function
value
(error)
***zrun
In general (hopefully) the number of simulation-experiment comparisons will lead the
problem to be over-constrained. Its the optimizers job then to find the best fit. If certain
experiments are “important” the user will employ weights on the comparisons functions to
make them more dominant in the function evaluation. Remember that the best, must robust
results will be found with many diverse experiment conditions. Repeated experiments which
merely provide a slight variation in response should be avoided, while complex experiments
with many interacting effects should be emphasized.
16.6
Z-set — Non-linear material
& structure analysis suite
****optimize
Recommendations:
Some comments follow which are purely the personal preferences of the author, but can
probably help ahead of time with the management of typical real-life optimization projects.
The fact of life is that optimization involves many trials of different models, simulations, and
overall rapid iteration towards the solution – both by the code and by the user. Also it is
relatively easy to find oneself in a situation where, when lightly “tweaking” some parameters,
the last best solution gets lost. It is really important to maintain control of the changes in
input files, models, coefficients, etc in order to safeguard against loosing valuable time. Some
hints are therefore:
• Use a hierarchical structure for different stages in the analysis, and keep the number
of files in each directory very specific to one task. That is, if you want to modify the
initial values dramatically, or continue from a previous solution, make a new directory,
and copy the files before running. A typical directory structure could be:
BasicIdentification/
exp_files/
..
trial1/
Makefile
simulation.inp
levenberg.inp
simplex.inp
trial2/
Makefile
simulation.inp
levenberg.inp
simplex.inp
continue2/
Makefile
simulation.inp
levenberg.inp
and so on.
• A Makefile can ease the work of cleaning out a lot of secondary files, but more importantly it is a very good way to clean those files without the risk of deleting important
files by mistake. I (RF) always set up a makefile more or less as follows:
all :
clean :
Zclean -a
rm -f *.best
Remember to use tabs for the target lines. to clean the directory one then issues
make clean
Z-set — Non-linear material
& structure analysis suite
16.7
****optimize
• It’s helpful to have the experiment (reference) files is a centralized location, as shown
above, but instead of using absolute path names use relative ones. In trial2 for example
an experiment file can be accessed1 as ../exp_files/exp1.dat In this way the project
can be moved without everything breaking. To start a whole new iteration then, one
just copies the BasicIdentification directory elsewhere and the work starts.
1
16.8
this works in Win32 platforms the same
Z-set — Non-linear material
& structure analysis suite
****optimize
Example:
This example is from EXAMPLES-MAT/Identify-in738/STEP1. First is the template file for
the material definition in738.tmpl:
*coefficients
young
149650.0000
n
?n
K
?K
n2
?n2
K2
?K2
C1
?C1
D1
?D1
R0
80.0
***return
****optimize levenberg_marquardt
***convergence
perturb 0.06
lambda0 5.0
iter
60
***zrun Zrun -Q -S simulate
***files in738
***values
**auto_init_from_file levenberg.best
K
700.
min 600.
max 5000.
n
4.47
min 1.2
max 20.
K2
500.
min 200.
max 5000.
n2
35.
min 3.
max 75.
C1
290.
min 3.0
max 800.
D1
600.
min
5.
max 5000.
***compare
g_file_file tenx2.test
1 2 ../EXP/tenx2.exp
g_file_file tenx3.test
1 2 ../EXP/tenx3.exp
g_file_file tx6x2x6.test 1 2 ../EXP/tx6x2x6.exp
g_file_file tx6x3x4.test 1 2 ../EXP/tx6x3x4.exp
g_file_file cr410.test
1 2 ../EXP/cr410.exp
g_file_file frlx3.test
1 2 ../EXP/frlx3.exp
****return
Z-set — Non-linear material
& structure analysis suite
1
1
1
1
1
1
4
4
4
4
3
4
16.9
****optimize
***compare
***compare
Description:
The option ***compare constructs the function to be minimized. Four comparisons are
possible. It is always possible to indicate a relative weight adding [weight vweight] where
vweight is the relative weight.
The function to the minimized by the optimizer is given by:
F=
X
Fi
i
where Fi is the function constructed by a ***compare instruction.
Example:
In this example, x and y are function arguments, and a, b, c are optimization variables.
****optimize levenberg_marquardt
***function f1
?a*x^2. +
?b*x
+
***compare i_func_file f1 1 2 3 teri.dat
***values
a 1. min -10. max 10.
b 1. min -10. max 10.
c 1. min -10. max 10.
***convergence
****return
?c + y;
Note:
It is recommanded to normalize the functions involved in the optimization (F and the constraints).
Note:
If one exclusively wishes to minimize a function f subjected to a constraint, one could proceed
as follows :
In the “master” file, optimize.inp
****optimize <OPTIMIZER>
***function f ?a * exp(-1. * ?b);
***constraint cos(a - b);
***compare i_func_file f 1 unreachable_goal.dat
***values
a 1. min 0. max 10.
a 2. min 2. max 5.
***convergence
...
****return
16.10
Z-set — Non-linear material
& structure analysis suite
****optimize
***compare
where the file unreachable goal.dat contains a lower bound on f, for example,“ 0.” here.
The optimizer will minimize the distance between the lower bound in the reference file and f.
Note that if the optimizer is Levenberg-Marquardt, the lower bound should be near the real
minimum of the function, otherwise the Levenberg-Marquardt approximation might be poor
and convergence impaired.
Z-set — Non-linear material
& structure analysis suite
16.11
****optimize
***compare t file file
***compare t file file
Description:
This comparison is used to compare two files (names fname1 et fname2) with a time-weighted
averaging. Files are “columns” files containing doubles. For each file, one considers a relation
yi (t), i = 1, 2 interpolated from columns y1 and t1. To compare two files, one considers the
instersection of columns t1 and t2. One defines
t− = max min, min
t+ = min max, max
t∈t1 t∈t2
t∈t1
t∈t2
The function to be optimized is given by:
Z t+
1
(y2 (t) − y1 (t))2 dt
Fi =
2(t+ − t− ) t−
The optimization variables x should affect one of the two files fname1 or fname2.
Note:
Because of the time weighting of the error, the density of comparison points is not important
for the comparison function (although more points will be more accurate). Also, if the time
scale of a test changes this method can totally miss important features of the response. For
example, if a hold-time loading profile of 10s-300s-10s is applied, the emphasis will be on the
response during 300s. Any details of the 10s segments will essentially be lost.
Syntax:
t file file fname1 t1 y1 fname2 t2 y2
[weight vweight]
16.12
Z-set — Non-linear material
& structure analysis suite
****optimize
***compare t file file
**compare g file file
**compare g file file
Description:
This comparison is a generalized-file-file method, based on point densities. The user therefore
has to think about the distribution of comparison points (either in the simulation or reference
files) in order to best control the quality of error estimation. Either file (ref or simulation)
can have more points, and the points of the smaller will be interpolated between points of the
larger file. This assumes that the larger file will have points who’s linear segment will more
closely approximate the real curve.
Files must contain continous “column” data of real (floating point format) values. Lines
can be commented with % or #.
The function to be optimized is given by:
"
#2
N
1 X ysk − ȳk
Fi =
N
|ysrefk |
k
with ȳk being the interpolated value from the smaller file, and ysk the test value from the
smaller file. ysrefk is the points value in the reference file (“experiment” file), which may be
an interpolated point if that file was smaller.
Syntax:
g_file_file sim-file sx sy exp-file ex ey
[weight vweight]
sim-file is the simulation file, which the optimization variables x should be changing, expfile is the “reference” which is the goal (should not change during optimization). sx and sy
are the x and y columns of the simulation file, and ex and ey are the x and y columns of the
reference file.
Note:
One is advised to make sure that the x variables are continuously increasing, as a comparison
of multi-valued functions or data curves are not allowed. In normal use, time makes a very
useful x variable, although a control parameter such as strain could be used as well.
The measure of error may differ quite significantly from the t_file_file output, so if it
is desired to mix the two types of comparison, use of the weight values will be very useful.
Z-set — Non-linear material
& structure analysis suite
16.13
****optimize
***compare i file file
***compare i file file
Description:
allows to compare two files (names fname1 et fname2). Files are “columns” files containing
doubles. Column c1 of file fname1 is compared with column c2 of file fname2. Both files must
have the same number of lines. The function to be optimized is given by:
Fi =
N
1 X
(yc2,i − yc1,i )2
2N
i=1
where N is the number of lines. The optimization variables x should affect one of the two
files fname1 or fname2.
Syntax:
i file file fname1 c1 fname2 c2 [weight vweight]
16.14
Z-set — Non-linear material
& structure analysis suite
****optimize
***compare i func file
***compare i func file
Description:
This comparison is similar to i_file_file except that a function will be evaluated for all
the x values of the reference file, and the values will be compared directly.
The function to be optimized is given by:
Fi =
N
1 X
(f uname(t1,i . . . tn,i ) − yi )2
2N
i=1
Syntax:
The following syntax is used here:
i func file funame ti,1 . . . ti,n y flname
[weight vweight]
funame is the name of a function defined under ***function, flname is the name of a
“column” file. ti,1 . . . ti,n are the columns id’s representing the n arguments of flname (other
than optimization variables that also can be arguments of the function). y is the column of
flname representing the function values associated with the n–uplets (ti,1 . . . ti,n ).
Z-set — Non-linear material
& structure analysis suite
16.15
****optimize
***constraint
***constraint
Description:
The ***constraint option allows to create constraints that relate the different variables.
Constraints are functions that need to be less than or equal to 0. When this option is used, an
OPTIMIZER need to be called that can handle constraints (for example, levenberg marquardt
cannot do it).
Syntax:
The syntax is:
***constraint FUNCTION [ lambda0 ]
where lambda0 is an optional value for the initial lagrange multiplier associated with the
constraint.
Example:
***constraint A - b; 0.01
In this case, the optimizer will change A and b such that A − b ≤ 0..
16.16
Z-set — Non-linear material
& structure analysis suite
****optimize
***comparison constraint
***comparison constraint
Description:
The ***comparison constraint option can be used to add constraints that force the solution near some critical points of the reference curves. Note that there may be no solution
satisfying the constraint. In this case an appropriate value of the convergence parameter
constraint_tol should be used to allow some constraint violation (see the ***convergence
options of the corresponding optimization method).
Syntax:
The syntax is:
***comparison_constraint
s file file xref fname1 t1 y1 fname2 t2 y2 [ lambda0 ]
where xref is the value along the x coordinate for which a constraint is added to enforce
y1(x)=y2(x),
fname1, fname2 are the names of the two files used in the comparison,
t1, t2 are the column numbers in files fname1 and fname2 that are used to define the x
coordinates,
y1, y2 are the column numbers in files fname1 and fname2 that are used to define the y(x)
values,
lambda0 is an optional value for the initial lagrange multiplier associated with the constraint.
Example:
***comparison_constraint
s_file_file 55. fat_al2.test 1 3 fat_al2.exp 1 3 0.0001
Z-set — Non-linear material
& structure analysis suite
16.17
****optimize
***files
***files
Description:
The ***files option is used to declare files where the optimization variables are to be replaced
during the trial evaluations of the error function. The following diagram shows the function of
the template files in order that the the optimizer can alter the execution data for simulations
in a general way.
Current
Variables
(R, Load, ...)
Optimizer
file.tmpl
*R
?R
*load ?Load
algorithm
Error
test
Compare
ref
Shell
file
*R
*load
93.502
324.0129
Zrun
The program detects variables to be optimized in the .tmpl suffixed versions of the files by
searching for alphanumeric strings with the following format:
?[a-z,A-Z,0-9]* Thus,
the ? indicates that the parameter is to be optimized, and the following token string is used
to reference the variable in the solution. References are needed to give initial values, define
constraints on the variables, etc.
If the file exists with the name given with the ***files command, the initial values will
be taken from that file. Otherwise, the ***value command must be used (see page 16.21). If
a variable name is repeatedly found in several files/functions, they are treated to be the same
and therefore only one optimization unknown is created. If the initial value file is given, it
will be copied to a new file with the suffix .bak to save the values.
Syntax:
This command is generally a one-liner, where it only takes the file name bases for optimization
files.
***file f1 . . . fn
where f1. . . fn are filenames where variables are. The software searches these variables in
“template” files named f1.tmpl . . . fn.tmpl. It then creates files f1. . . fn with the actual values
of the variables.
Example:
The following statement declares that file file1 contains the data to be optimized.
***files file1
where file1.tmpl could for example be:
***behavior gen_evp
**elasticity young 260000.
**potential gen_evp ev
*flow norton
n 7.0
K ?Visco
16.18
poisson 0.3
Z-set — Non-linear material
& structure analysis suite
****optimize
***files
*isotropic constant
R0 ?RRR
***return
The lines ?Visco and ?RRR indicate that two variable values are to be substituted in their
places during the optimization.
Z-set — Non-linear material
& structure analysis suite
16.19
****optimize
***function
***function
Description:
The ***function is used to declare functions with variables to be optimized. The functions
use a similar format for declaring places for optimization variables as in the ***files option.
That is, parameters within the function syntax which are named ?[a-z,A-Z,0-9]* are taken
to be variables for the optimization algorithm. Initial values and bounds are to be given for
each function and file optimization variable using the ***value command (page 16.21).
Syntax:
The syntax is:
***function name FUNCTION
where name is the name given to the function. If a variable name is repeatedly found in
several files/functions, they are treated to be the same and therefore only one optimization
unknown is created.
Example:
The horrible function from Ackley.
****optimize sqp
***function f1 -20. * exp(-0.2*sqrt(0.5*(?x1^2+?x2^2))) exp(0.5*(cos(4.*3.14159265358979*?x1)+
cos(4.*3.14159265358979*?x2)))+
20.+2.71828182845905;
***compare i_func_file f1 1 goal_ackley
***values
x1 -1.1 min -5. max 5.
x2 -0.5 min -5. max 5.
16.20
Z-set — Non-linear material
& structure analysis suite
****optimize
***value
***value
Description:
The ***value command is used to initialize the variables to be optimized. The initial value
of the variables is used to normalize them. It is also possible to indicate min and max values
for the variable. Some algorithms need these values (an error message will be given if they
are missing).
Variables can conveniantly be taken out of the optimization by using the fixed keyword
in the variable initialization. All fixed variables will still be substituted in the tmpl files, but
they remain at the given initial value throughout the optimization.
Syntax:
The syntax for variable initialization follows. The sub-commands will be described in the
next pages of the manual.
***value
[ **init_from_file fname ]
[ **auto_init_from_file fname ]
[ **format fmt ]
vname1 [fixed ] vini1 [min vmin1] [max vmax1] [-log|+log]
...
vnameN [fixed ] viniN [min vminN] [max vmaxN] [-log|+log]
where vname the variable name, vini its initial value (also used for scaling) (double), vmin
its min value and vmax its max value. Scaling is important, particularly when the variables
that are handled are of widely differing magnitudes. Without scaling, matrices used by the
optimization algorithms are typically ill-conditionned. Scaling is automatically performed in
Z-set using vname/vini instead of vname. It is possible to have different initial and scaling
values using the **init from file command.
fixed fix the value to the initial, and remove the variable from the optimization problem.
min set the minimum allowable value for the variable.
max set the maximum allowable value for the variable.
+log|-log treat variables as being log’s of the values to be substituted in the files (the
.tmpl substitutions)2 . if +log is selected the variables ? entries in the .tmpl file will
be substituted not with the variable, but instead ex . If the option -log is chosen the
program will use −ex .
Note that the order in which the values are given is not the order in which they are handled
(thus printed) by the optimizers. For instance, when reading the results of the evolution
optimizer, care should be taken not to misinterprete the order of the variables. THE REAL
ORDER is given in the ”.msg” files, for instance.
Example:
2
this is very useful for variables which span orders of magnitude. A classical example is a creep law of the
form Aσ n where A much change orders of magnitude in response to small changes in n for similar values of σ.
Sometimes a better approach is to re-formulate, for example with (σ/K)n
Z-set — Non-linear material
& structure analysis suite
16.21
****optimize
***value
***values
**auto_init_from_file levenberg.best
C1
fixed 200. min 1.e-5 max 1.e6
f_trans fixed 1.
min 0.3
max 1.1
m
M
f_crit
16.22
0.2
2.01
0.85
min
min
min
0.05
0.01
0.0
max 20.
max 40.
max 0.95
Z-set — Non-linear material
& structure analysis suite
****optimize
***value
**auto init from file
**auto init from file
Description:
This option is used to conveniantly re-start an optimization from the last best values.
Syntax:
The syntax takes the name of an ascii file which has the format of the problem.best file. This
file is written each time a trial of the function is less than the previous best value.
**auto_init_from_file
file
Example:
Supposing the following values entry is given:
***values
**auto_init_from_file levenberg.best
K
700.
min 600.
max
n
4.47
min 1.2
max
K2
500.
min 200.
max
n2
35.
min 3.
max
C1
290.
min 3.0
max
D1
600.
min
5.
max
5000.
20.
5000.
75.
800.
5000.
An example of the levenberg.best file is:
# best objective function : 6.050340e-01
# best values :
n 6.101396e+00
K 1.314714e+03
n2 5.967421e+01
K2 5.244540e+02
C1 3.100267e+02
D1 6.302586e+02
Z-set — Non-linear material
& structure analysis suite
16.23
****optimize
***value
**init from file
**init from file
Description:
The option **init from file filename enables initiating the optimization from the point
written in filename. This can be handy in some cases where one does not wish to change the
.inp file. The initial values of the design variables are given as real numbers. The order is the
order of the variables (cf. note above, this is the order of the .msg and .tra files). When this
option is used, the syntax for the rest of the ***value is unchanged. The value that follows
the variable name is no longer the starting value (since the starting value is read in filename)
but it is used as scaling variable. Therefore, the option **init from file filename permits
separated initial and reference values.
16.24
Z-set — Non-linear material
& structure analysis suite
****optimize
***value
**format
**format
Description:
The option format controls the output format of the variables declared as ?x in the .tmpl
files. It can be important to output lots of decimals to reduce numerical noise (for example
in order to have accurate gradient calculation).
Example:
***value A 10. min 1.
**format %30.15e
**init_from_file init
b 2. min 1. max 50.
c 9.
Z-set — Non-linear material
& structure analysis suite
16.25
****optimize
***shell
***shell
Description:
The ***shell option allows to declare external (system) commands to be executed before
the evaluation of the function to be optimized. These jobs are done using a posix system call.
Use the ***zrun command if the job is a Z-set calculation (using Zrun), because it will not
launch a whole new executable.
Example:
***shell Zmat -fg test1 >& /dev/null
abaqus post job=test1 input=test1.jnl >& /dev/null
awk ’{print $3, $4/100.0}’ test1.test > result
One can also make up a script with the shell commands, for a bit more flexibility:
***shell
./RUN
#
# From EXAMPLES-ZBB/Inverse-aba-thermal/Levenberg/RUN
#
#!/bin/sh
Zmat -fg nt4xx28c
PLT nt4xx28c 2>&1 > /dev/null
cat nt4xx28c.test
exit 0
16.26
Z-set — Non-linear material
& structure analysis suite
****optimize
***zrun
***zrun
Description:
The ***zrun option runs a sub-process of Z-set to simulate some part of the model(s) as part
of the optimzation comparison scheme. All zrun statements will be run after all the ***shell
entries.
Example:
Note that for simulation, running with -Q suppresses the screen output of the simulation, so
one can watch just the optimization progression.
***zrun Zrun -S -Q test1
Z-set — Non-linear material
& structure analysis suite
16.27
****optimize nelder mead
****optimize nelder mead
Description:
The nelder-mead or simplex method ([Neld65]) is an order 0 local optimizer. The method is
numerically simple and robust, but slow for large scale problems. A flow chart of the method
is given in the figure hereafter. The simplex method starts with a regular geometric figure
called the simplex consisting of n + 1 vertices where n is the dimension of the design space.
Then, vertices are moved according to their respective values f . There are 3 main steps during
a sequential simplex search: the reflection (from the worst point to the best), the expansion
(continuation of the displacement towards the best vertex, when reflection improved f ), and
the contraction of all the vertices towards the simplex gravity center when reflection has failed.
In the figure, xh, xl, xs and xm are the highest, lowest, second lowest point, and the gravity
center, respectively. fh, fl, and fs are associated cost functions. Typically, a ≈ 1, b ≈ 2, et
0 < c < 1.
This implementation integrates an optional automatic restarting procedure in an attempt
to render the algorithm less sensitive to local minima. In this case, the new starting point
is automatically evaluated based upon previous results. Optimization constraints are also
supported, by means of a basic penalty technique.
initialize the simplex
determine xh, xs, and xm
fh, fs, fl
Reflection : xr = xm + a ( xm − xh)
fh < fl
n
fr <= fs
n
fr < fh
o
replace xh by xr
o
o
Expansion : xe = xm + b (xr − xm)
o
replace xh by xe
convergence
Contraction : xc = xm + c (xh − xm)
fc > fh
fe < fr
n
n
n
o
replace xh by xr
o
replace all x’s
replace xh by xc
stop
Syntax:
The nelder-mead optimizer has a number of adjustable parameters which can be set in the
***convergence section. These are summarized below.
tolerance Minimal relative difference between the highest and lowest vertex for convergence. This is exclusive with min. The algorithm stops when the last change in objective
function is fractionally smaller than tolerance. It is the default stopping criterion with
a value of 1.e-9.
size optimum Minimal value of the objective function below which the search stops. This
is exclusive with tolerance. It is not the default stopping criterion. Default value is
1.e-5.
size degeneration Minimal relative value below which, simplex is considered as degenerate, ie. vertices are no more linearly independent. Default value is 1.e-6.
16.28
Z-set — Non-linear material
& structure analysis suite
****optimize nelder mead
length Initial perturbation magnitude as a factor times the initial value. It is the characteristic length scale. It is used to construct the initial simplex from one starting point.
Default value is 1. It is a rather sensitive parameter, and one should play with it if the
simplex does not start the search properly.
iter Maximum function evaluations before terminating. Default value is 100 .
max restart Maximum number of restarts. Default value is 0 (ie. no restart).
reflection Value of the reflection parameter a (default value is 1.0)
expansion Value of the expansion parameter b (default value is 2.0)
contraction Value of the contraction parameter c (default value is 0.5).
lag mult step Step size used to update the penalty parameters associated to the optimization constraint: if gi > 0. then λi + = step . gi . Default value is 0.01.
constraint tol Constraint violation tolerance value, ie. constraints are verified when
gi (x) < constraint tol, and the corresponding design point x retained as feasible. Default value is 0.001.
Z-set — Non-linear material
& structure analysis suite
16.29
****optimize levenberg mar
****optimize levenberg marquardt
Description:
This optimizer ([Leve44], [Moré77]) is a robust first order method (uses gradients of the
objective function) specially meant for least-square minimization. It does not explicitly handle
constraints.
The least-square norm of the distance between simulation and experience can be written,
F =
1
(f (x) − y)T W (f (x) − y) ,
2
(3)
where f (x) is the (N × 1) vector of simulations, y is the (N × 1) vector of experiments, and
W is the positive diagonal (N × N ) weight matrix. The first and second derivatives of F(x)
are,
gradF(x) = J T W (f (x) − y)
hessianF(§) = J T W J +
N
X
(4)
Wii (fi (x) − yi )hessianfi
(5)
i=1
(6)
where J is the (n × N ) matrix composed of [gradf1 . . . gradfN ] (here n is the number of
optimization variables). The first fundamental idea in Levenberg-Marquardt is that the last
terms of hessianF(§) can be neglected, hessianF(§) ≈ J T WJ . This is true in particular when
the terms (fi (x) − yi ) are small in comparison to hessianfi , which stands when the simulation
is near the experiments and the largest eigenvalue of hessianfi is not too large (if not true,
another optimization method, sqp for example, should be used). Necessary conditions of
optimality state that the gradient of the objective function should vanish at the optimum x∗ ,
gradF(x∗ ) = 0 .
(7)
Linearization of Equation (7) about the current point xk yields,
gradF(xk ) + hessianF(§k )(§k+∞ − §k ) = 0 .
(8)
The second idea underlying Levenberg-Marquardt is to regularize the approximation of the
Hessian of the objective function by adding positive terms on the diagonal, and use this in
Equation (8). The regularized approximation to hessianF is then strictly positive definite and
Equation (8) can always be solved. The fundamental equation used to update optimization
variables is,
(JkT W Jk + λk I)(xk+1 − xk ) = − JkT W (f (xk ) − y) .
(9)
where λk is a scalar positive parameter. The principle of Levenberg-Marquardt algorithm
can also be derived from the linearized original optimization problem with a bound on the
design variables. This yields an interpretation of λ as Lagrange multiplier associated to the
move limit on the design variables. The value of λ can therefore be iteratively adjusted to
guarantee that the linearization performed is still valid and that progress in terms of the
objective function is made. The Levenberg-Marquardt algorithm proceeds as follows:
16.30
Z-set — Non-linear material
& structure analysis suite
****optimize levenberg mar
1. k = 0 , λ = λ0 , xk = x0 .
2. Solve (9) for xk+1 .
3. If ||gradF(xk+1 )|| < gmin|| or F(xk+1 ) < f min, or number of F evaluations ¿ iter, or
||xk+1 − xk || ¡ stepmini, stop.
4. If F(xk+1 ) < F(xk ), λk+1 = λk /nu, k = k + 1, goto 2.
5. If F(xk+1 ) ≥ F(xk ), λk = λk ∗ nu, goto 2.
Gradients are calculated by finite differences. Min and max bounds on the optimization
variables are taken into account by projection into the feasible domain.
Syntax:
The following adjustable parameters are allowed in the ***convergence section.
perturb Magnitude of the perturbation (in percent of the normalized design variables)
used to calculate gradient through finite differences. Default value is 0.5 percent.
lambda0 Initial value of the λ parameter which controls the algorithm. Smaller values when
closer to the solution.
lambda max Maximal value of lambda. Default is 1.e+07.
iter Maximum number of function F evaluations before terminating. Default is 100.
nu
cutting/amplification factor of the lambda parameter on reduced or increased function
value. It was originally suggested to have 10 for this parameter (default value). A factor
of 4 may be more conservative.
fmin minimum function value for convergence. Default is 0.0001.
gmin minimum norm of the gradient of the objective function for convergence. Default is
1.e-8 .
stepmini minimum step at each iteration. Default is 0.00001. There is a possible false
stopping of the optimizer if lambda is taken very large and stepmini very small (because
lambda acts as a move limit on the parameter change).
Example:
This following is a complete optimization input using this method:
****optimize levenberg_marquardt
***files plast.sim
***shell
Zrun -S plast3 2>&1 > /dev/null
***values
cl 1000. min 200. max 15000.
cnl 140000. min 50000. max 200000.
dnl 100.
min 50. max 2000.
r0 260.
min 50. max 500.
q
200.
min 20. max 500.
b
5.
min 2.
max 20.
Z-set — Non-linear material
& structure analysis suite
16.31
****optimize levenberg mar
***convergence
perturb 0.005
lambda0 5
lambda_max 1.e+08
iter 100
nu 5
fmin 0.00001
gmin 0.00001
***compare
t_file_file plast3.test 1 3 data_plast3 1 3 weight 1.
****return
16.32
Z-set — Non-linear material
& structure analysis suite
****optimize evolution
****optimize evolution
Description:
This section describes indeed the options available for the evolution type optimizer engine
is used. In that case, an evolutionary (or genetic) algorithm is used for the optimization.
Evolution methods search the space of variables using random jumps. Even though such
jumps are oriented so as to be more efficient than a completely random search, evolution still
needs around 1000 calculations of the function to get near the optimum. The advantage of
evolution is that it can escape local optima in the search space.
A useful application of the evolution method is to generate a starting point for other
optimization methods. This is because the evolution method is not sensitive to the convexity
of the problem, which permits approaching optimization problems that are numerically illconditionned, that diverge, that have discontinuous functions,. . . .
evolution is a steady state genetic algorithm that handles continuous variables. It can
handle constraints using a penalization scheme (cf. introduction of the optimization in Zset, page 16.4). There is a rule of thumb when changing the parameters of evolution: some
parameters will make the search more random, i.e. more robust but also less efficient, others
will increase the speed of convergence, but the probability of getting trapped far from the
global optimum simultaneously increases. The tradeoff is clear, and default parameters are
supposed to represent a reasonable compromise between randomness and efficiency. Based on
its experience, the user may decide to emphasize an aspect or another of evolution’s search.
population size and prob muta should be increased to increase randomness.
The following files are generated by a run of evolution:
case.evo.log
contains various data about the convergence of the run.
case.evo.gnut is an executable file that will automatically plot (with gnuplot) the main
results found in case.evo.log .
case.evo.best contains the best solution found at the end of the run.
case.evo.gid contains “good ideas”, that is to say good solutions, different from each other,
classified in two groups, feasible solutions that satisfy the constraints, and infeasible
solutions.
Parameters:
Parameters of evolution are given in the ***convergence section. They may be taken from
the following:
population size Sets the population size of the algorithm, i.e., the number of points that
are processed by the algorithm. Default value is 50.
prob Xover Sets the probability of crossover (a number between 0. and 1.). Default value
is 0.8 .
prob muta Sets the probability of mutation (a number between 0. and 1.). Default value
is 0.2 .
max nb analyse Sets max nb analyse. The search is stopped after max nb analyse calculations of the function. Default value is 700.
Z-set — Non-linear material
& structure analysis suite
16.33
****optimize evolution
max no progress Sets max no progress The search is stopped after max no progress calculations of the function without improvement. Default value is very large, making it
inactive.
no records Sets the number of outputs recorded in the file case.evo.log . This is the
number of points that will be ploted then with the command case.evo.gnut . Default
value is 30.
no feas pts res , no ifeas pts res evolution also collects points that might not be the
best overall, but might be the best of their kind. Those points are to be found at
the end of the run in the file case.evo.gid . They are divided in 2 groups, the feasible
and the infeasible points. no feas pts res sets the number of feasible points that
are collected, and no ifeas pts res the number of infeasible points. The algorithm
that collects those points make sure that they are different based on a distance metric
that can be set with the command neighborhood dist. If two points are separated
bby a distance smaller than neighborhood dist, they are considered as being close and
only onepof
will be kept in case.evo.gid . A default value of neighborhood dist is
Pthem
n
2
1/20 ×
(x
i=0 i,max − xi, min) .
lag mult v1 . . . vm , where m is the number of constraints. It enables to set the lagrange
multipliers at the beginning of the runs. Lagrange multipliers apply to each constraint
separately, and they are used here as penalties. If a constraint is difficult to satisfy,
the corresponding lagrange multiplier should be increased. Lagrange multipliers are
automatically changed by the optimizer, and they will eventually represent the real
lagrange multipliers after convergence. Default values are 1.
lag mult step sets the amount of change of the lagrange multipliers after each evaluation of
the function. The lagrange multipliers are updated at each function evaluation according
to ,
λk+1
= λki + lag_mult_step × gi
i
, i = 1, m
λk+1
= max(λk+1
, 0.)
i
i
where λki is the lagrange multiplier of the ith constraint at the kth iteration, and gi is the
value of the ith constraint of the best solution known so far. The larger lag_mult_step,
the quicker the lagrange multipliers will change. Default value is 0.1 .
16.34
Z-set — Non-linear material
& structure analysis suite
****optimize augmented lagrangian
****optimize augmented lagrangian
Description:
This optimizer is a basic implementation of the classic augmented-lagrangian method for
constrained
optimization. For problems of the form:

 M in f (x)
gi (x) ≤ 0 (i = 1, ..., m)

x ∈ <n
a dual method is used to find a saddle point of the augmented lagrangian function
L̂(x, λ, r):
n
o
(
M axλ ŵ(λ, r) = M axλ
M inx L̂(x, λ, r)
λ ∈ <m+
where L̂(x, λ, r) is defined by:
L̂(x, λ, r) = f (x) +
Pm
i=1 λi
gi (x) + r
Pm
2
i=1 gi (x)
with λi the lagrange multipliers, and r a penalty parameter that insure the differentiability
of the dual function ŵ(λ, r) near the optimum λ∗ .
The penalty parameter r must then be chosen large enough to improve convergence of
the dual problem, but not too large because the lagrange function L̂(x, λ, r) can become
ill-conditioned in this case.
For each value of the lagrange multiplier λi the unconstrained problem M inx L̂(x, λ, r) is
solved by a BFGS method. At each iteration of the BFGS method an economic line-search
procedure based upon the Goldstein rule has been implemented.
Syntax:
The following adjustable parameters are allowed in the ***convergence section.
iter Maximum number of function f evaluations before terminating. Default is 100. Note
that this doesn’t include function evaluations associated with the line-search procedure.
perturb Magnitude of the perturbation (in percent of the normalized design variables)
used to calculate gradient through finite differences. Default value is 0.5 percent.
r0 Value of the penalty parameter r. Default is 1.0.
aug conv Convergence parameter for dual problem based upon the relative variation of
design variables. Default value is 1.e-7.
bfgs conv Convergence parameter for the inner unconstrained optimization problem based
upon the relative variation of design variables. Default value is 1.e-7.
oned conv Convergence parameter for the line search procedure. Default value is 1.e-7.
gradient tol Additional convergence parameter for the inner unconstrained problem,
based upon the gradient of the lagrangian function. Default value is 1.e-7.
Z-set — Non-linear material
& structure analysis suite
16.35
****optimize augmented lagrangian
constraint tol Constraint violation tolerance value, ie. constraints are verified when
gi (x) < constraint tol, and the corresponding design point x retained as feasible. Default value is 0.001.
16.36
Z-set — Non-linear material
& structure analysis suite
****optimize single
****optimize single
Description:
single is not an optimizer, but rather a convenience to test, or re-analyze, a set of optimization parameters. It performs one single calculation of the cost function going through the
principal routines of the optimizer. It is particularly meant to be used in conjunction with
the **init from file command to choose a particular initial point. There are no parameters in the convergence section, and the reading of this algorithm will pass by entries there
without an error message.
Example:
The following is an example where the variables cl ... b are read in the ASCII file init. The
values 1000 ... 5 are in fact used for scaling only. Only one simulation is performed, and the
objective function is calculated.
****optimize single
***files plast.sim
***shell
Zrun -S -Q plast3 2>&1 > /dev/null
***values
**init_from_file init
cl 1000. min 200. max 15000.
cnl 140000. min 50000. max 200000.
dnl 100.
min 50. max 2000.
r0 260.
min 50. max 500.
q
200.
min 20. max 500.
b
5.
min 2.
max 20.
***convergence
delta 0.005
eps 1.e-08
iter 100
***compare
t_file_file plast3.test 1 3 data_plast3 1 3 weight 1.
****return
Z-set — Non-linear material
& structure analysis suite
16.37
****optimize single
16.38
Z-set — Non-linear material
& structure analysis suite
Chapter 17
Reference
Z-set — Non-linear material
& structure analysis suite
17.1
17.2
Z-set — Non-linear material
& structure analysis suite
Functions
Functions
Description:
The functions have been generalized in code versions greater than 7.1 to have a natural
equation like form, and apply more widely throughout the program. In addition to coefficient
definitions, the functions may now be used to define load waveforms, apply to the optimizer,
etc. More applications will be added as well.
The functions follow a C-like syntax, and may be defined in terms of pre-defined functions
(sin, log, etc) and named parameters of the problem. What parameters are allowed and
available depends on application.
Syntax:
Objects which are functions may now be entered using a C-like format. This must include
a semi-colon at the end of the function definition to terminate the reading of the function.
Many operators exist, which have the same meaning and order of selection as in C. Coefficients
which are functions still require the keyword function to follow the coefficient name, but are
otherwise generally given in terms of the relevant parameters.
The following operators are defined in functions:
:= == >= <= + - * / ^ > <
as are the following functions:
triangle floor asinh acosh atanh asin acos atan
cosh sinh tanh ceil sqrt neg abs sqr sin cos tan
exp ln Hx H
Pre-defined functions may take complex arguments of other expressions or functions.
triangle
floor
function used to make a sawtooth waveform with period 1 of the parameter.
largest integral value not greater than x.
H Heaviside function. Equal to 1 for x > 0 and 0 otherwise.
Hx Heaviside of its parameter times its parameter. Equal to x for x > 0 and 0 otherwise.
Example:
***function load_tab sin(6.28319*time);
***elasticity
young function 2.e6-100.*temperature -2.*temperature^2.;
poisson 0.3
Z-set — Non-linear material
& structure analysis suite
17.3
Functions
17.4
Z-set — Non-linear material
& structure analysis suite
Environment Variables
Environment Variables
The environment variables are used to personalize the Z-set program for individual users,
and maintain a self-contained project structure. The function extends as well to manage
multi-architecture sites, thereby allowing the transparent use of the program over a diverse
network. The currently used environment variables are summarized as:
Z7PATH
DESCRIPTION
path to the head Z-set’s directory
structure
Z7 MAX NB DOF
number of DOFs above which disk storage is used
Z7 TMP DIR
path for the temporary files if needed
Z7 LICENSE
path
to
a
license
file
$Z7PATH/lib/Zebulon.License
CODE
if
not
stored
in
Z7PATH variable which indicates the location of the Z-set distribution. This variable allows
easy switching to different distributions. It also is a cause of much difficulty with users
if not configured correctly.
cd /usr/local/Z8-Sept-1999
setenv Z7PATH ‘pwd‘
source lib/Z7_cshrc
Z7 MAX NB DOF This sets the maximum number of degrees of freedom before the global
matrix is stored on disk. Because disk storage is not expensive, it is advisable to set
this to a fairly low value (e.g. 5000). One can also use ***file management to control
the parameter.
Z7 TMP DIR directory where the temporary files are to be stored. Default is the problem
directory. This can also be set with ***file management.
Z7 LICENSE fully qualified path to a license file. Extremely useful when there are multiple
versions, or for running Zebulon directly off a CD-ROM where one can’t put the license
file in a read only lib dir.
Example:
In the C-shell, these variables are set using the setenv command. The variables may be
directly in the user’s .cshrc file, or in a separate file. For the later case, one may add the
line to the .cshrc file: if (-f ~/lib/Z7_vars) source ~/lib/Z7_vars Examples of variable
settings are given below:
Z-set — Non-linear material
& structure analysis suite
17.5
Environment Variables
% cat ~/lib/Z7_vars
setenv Z7_TMP_DIR /home/disk/me/Z7
setenv Z7_MAX_NB_DOF
5000
setenv Z7_PRINTER_COLOR_A3 HPDESKJET
17.6
Z-set — Non-linear material
& structure analysis suite
Chapter 18
Bibliography
Z-set — Non-linear material
& structure analysis suite
18.1
18.2
Z-set — Non-linear material
& structure analysis suite
Bäck91
T.F. Bäck. “A survey of evolution strategies.” In R.K. Belew and L.B.
Booker, editors, Proc. of the 4th International Conference on Genetic
Algorithms, volume I, pages 2–9, USA, Morgan Kauffmann (1991).
Bely00
T. Belytschko, W.K. Liu, B. Moran, Nonlinear Finite Elements for Continua and Structures, John Wiley & Sons (2000).
Bone97
J. Bonet and R.D. Wood Nonlinear Continuum Mechanics for Finite Element Analysis Cambridge University Press Cambridge, United Kingdom
(1997).
Cont89
E. Contesti and G. Cailletaud, “Description of Creep-Plasticity
Interaction with Non-Unified Constitutive Equations: Application to
Austenitic Stainless Steel,” Nuclear Engineering Design 116, 265-280
(1989).
Cail92
G. Cailletaud, “A Micromechanical Approach to Inelastic Behavior of
Metals,” Int. J. Plasticity, 8, 55-73 (1992).
Cail94
G. Cailletaud, P. Pilvin, “Utilisation de modèles polycristallins pour le
calcul par éléments finis,” Revue européenne des éléments finis, 515-542,
1994
G. Cailletaud and K. Saı̈, ”Study of Plastic/Viscoplastic Models with
various Inelastic Mechanisms,” Int. J. Plasticity, 10 (1995).
Cail95
Croi92
D. Croizet, L. Méric, M. Boussuge G. Cailletaud, “General Formulation
of a Plasticity / Viscoplasticity Algorithm in Finite Element,” Num.
Meth. Eng ‘92. ed. C. Hirsch et al, Bruxelles, 741-747 (1992).
Delo96
P. Delobelle, P. Robinet, P. Geyer, and P. Bouffioux, “A Model to Describe the Anisotropic Viscoplastic Behavior of Zircaloy-4 Tubes,” J.
Nuclear Materials 238 135-162 (1996).
Flec92
N. A. Fleck, L. T. Kuhn, and R. M. McMeeking, “Yielding of metal
powder bonded by isolated contacts. J. Mech. Phys. Solids, 40, 11391162, (1992).
Gree65
A.E. Green and P.M. Naghdi, “A General Theory of Elastic-Plastic Continuum,” Arch. Rat. Mech. Anal. 18, 251-281 (1965).
Gür92
Z. Gürdal R.T. Haftka, “Elements of Structural Optimization” Kluwer
Academic Publishers, (1992).
Hjel94
H.E. Hjelm, “Yield Surface for Grey Cast Iron Under Biaxial Stress,” J.
Eng. Mater. Tech 116 148-154 (1994).
Jose95
B.L. Josetson, U. Stigh, and H.E. Hjelm, “A Nonlinear Kinematic Hardening Model for Elastoplastic Deformations in Grey Cast Iron,” J. Eng.
Mater. Tech, 117 145-149 (1995).
Lade80
P. Ladevèze, “Sur la Théorie de la Plasticité en Grandes Déformations,”
ENS-Cachan-LMT Internal Report No. 9 (1980).
Lema85
J. Lemaitre and J-L. Chaboche, Mechanics of Solid Materials, Cambridge University press (1985).
Leve44
K. Levenberg. “A method for the solution of certain non-linear problems
in least squares.” Quaterly of Applied Mathematics, 2, 164–168 (1944).
Z-set — Non-linear material
& structure analysis suite
18.3
18.4
Mill76
A. Miller, “An Inelastic Constutive Model for Monotonic, Cyclic, and
Creep Deformation: Part I and II. J. Eng. Mater. Tech, 97-113 (1976).
Moré77
J.J. Moré. “The levenberg-marquardt algorithm: Implementation and
theory.” In G.A. Watson, editor, Numerical Analysis Proceedings, Lecture Notes in Mathematics, pages 105-116, Springer Verlag, Berlin, Germany (1977).
Nagt82
J.C. Nagtegaal, “On the Implementation of Inelastic Constitutive
Equations with Special Reference to Large Deformation Problems,”
Comp. Meth. Applied Mech. Eng., 33, 469-484 (1982).
Neld65
J.A. Nelder and R. Mead, A simplex method for function minimization.
Computer Journal., 7, 308–313 (1965).
Pilv94
P. Pilvin, “The Contribution of Micromechanical Approaches to the
Modelling of Inelastic Behavior of Polycrystals,” Fourth International
Conference on Biaxial/Multiaxial Fatigue, May 31-June 3, Paris, France
(1994).
Pilv96
P. Pilvin, Sidolo2.3, Manuel utilisateur, Centre des Matériaux, Ecole des
Mines de Paris, France (1996).
Simo98
J.C. Simo and T.J.R. Hughes, Computational Inelasticity SpringerVerlag, New York (1998).
Schit91
K. Schittkowski, QLD : A FORTRAN Code for Quadratic Programming,
User’s Guide, Mathematisches Institut, Universit ät Bayreuth, Germany,
(1986).
Zhou97
J.L. Zhou C. Lawrence and A.L. Tits, User’s Guide for CFSQP Version
2.5 Inst. for System Research TR-94-16r1, Univ. of Maryland, College
Park, MD 20742, (1997).
Z-set — Non-linear material
& structure analysis suite
Chapter 19
Index
Z-set — Non-linear material
& structure analysis suite
19.1
19.2
Z-set — Non-linear material
& structure analysis suite
Index
****optimize
***compare, 16.10
***compare i file file, 16.14
***compare i func file, 16.15
***compare t file file, 16.12
**compare g file file, 16.13
***comparison constraint, 16.17
***constraint, 16.16
***files, 16.18
***function, 16.20
***shell, 16.26
***value, 16.21
**auto init from file, 16.23
**format, 16.25
**init from file, 16.24
***zrun, 16.27
****simulate
***test, 15.6
**constant parameter, 15.8
**error plot, 15.9
**load, 15.10
**model, 15.13
**output, 15.14
**solver, 15.15
**yield surface, 15.17
***behavior porous plastic
**porous potential, 10.23
***behavior aging, 11.3
***behavior aniso damage, 11.5
***behavior becker needleman, 11.7
***behavior bodner partom, 11.11
***behavior cast iron, 11.9
***behavior chaboche debonding, 12.8
***behavior coefficient diffusion, 12.3
***behavior crisfield debonding, 12.6
***behavior damage elasticity, 10.4
***behavior diffusion, 12.14
***behavior finite strain crystal, 11.13
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
***behavior
gen evp, 10.13
generalized debonding, 12.10
hyper elastic, 10.5
hyperviscoelastic, 10.12
linear elastic, 10.3
linear spring, 12.15
linear viscoelastic, 10.6
matmod, 11.15
matmod z, 11.17
mechanical step phase, 10.24
memory, 11.19
needleman debonding, 12.4
non associated, 11.20
porous plastic, 10.18
reduced plastic, 10.16
thermal, 12.16
umat, 10.25
variable friction, 12.17
visco aniso damage, 11.22
viscoelastic spectral, 10.9
abaqus v6.env, 3.13
algorithm, 15.16
angles, 15.17
anisotropic, 13.11, 13.95
anisotropic elasticity, 13.11
anisotropic thermal strain, 13.95
armstrong frederick, 13.38
asaro, 13.39
auto init from file, 16.23
auto step, 14.8
automatic time, 6.5, 15.16
behavior, 9.13
behavior:aging theta, 11.3
behavior:veil theta, 11.3
by point, 13.55
cast iron, 13.41
class, 9.3
19.3
compaction, 10.18
compare g file file, 16.13
component, 15.9, 15.17
constant, 13.55
constant parameter, 15.8
conventions, 1.5
creep, 13.36
criterion, 15.17
cubic, 13.11
cubic elasticity, 13.11
debonding, 12.4, 12.6, 12.8, 12.10
debug, 6.4
default, 13.52
degrees, 15.17
DEPVAR, 3.9, 3.17
dimension, 6.6
direction1, 15.9
direction2, 15.9
divergence, 2.8, 14.8, 15.16
dont use e bar, 13.3
double norton, 13.45
elastic, 13.35
ELASTICITY, 13.9
Elasticity modules, 13.10
eps, 14.7, 15.18
error plot, 15.9
exponential crystal, 13.45
factor, 15.18
file, 15.10, 15.13, 15.14
find offset, 15.18
flow sum inv, 13.45
forceit, 14.8
format, 16.25
frequency, 15.14
full step jacobian, 14.8
function, 17.3
Gurson, 10.18
init from file, 16.24
initialize variable, 2.17
integration, 2.13, 15.13
Interface files
***automatic time, 2.8
***behavior, 2.9
***debug, 2.10
19.4
***external storage, 2.11
***material, 2.12
***parameter, 2.18
***save energies, 2.19
***skip cycle, 2.20
interface control, 13.45
inv exp, 13.45
iso, 13.52
iso table, 13.55
isotropic, 13.9, 13.95
isotropic elasticity, 13.9
isotropic thermal strain, 13.95
iter, 14.7
iter optimal, 15.16
iteration, 15.16
K, 13.4
K coeffs, 13.3
lagrange rotate, 14.3
limit, 2.8, 6.5, 14.8
limit output var, 15.14
linear, 13.55, 13.58
linear nonlinear, 13.55
linear pp, 13.55
load, 15.10
local debug, 6.4
make contour, 15.9
MATERIAL, 3.9
material, 6.6
matmod, 11.15
matmod z, 11.17
Matrix coefficients, 13.10
max dtime, 15.16
min dtime, 14.8, 15.16
model, 15.13
monocrystal, 13.80
needs temperature, 6.5
nonlinear, 13.55, 13.58
nonlinear 1, 13.56
nonlinear evrad, 13.58
nonlinear phi, 13.58
nonlinear sum, 13.55
nonlinear with crit, 13.59
normalization, 16.21
norton, 13.44
Z-set — Non-linear material
& structure analysis suite
norton exp, 13.45
nucleation, 13.93
number, 15.9
object, 9.3
offset, 15.18
one sided, 14.12
orthotropic, 13.11
orthotropic elasticity, 13.11
OUT, 2.10, 2.11
output, 15.14
output (simulation), 15.14
perturb, 14.12
perturbation, 14.12
plane stress, 14.7
plastic, 13.36
plasticity, 13.45, 13.73
porous potential, 10.23
potential, 15.17
power law, 13.55
precision, 15.14
rate, 15.10, 15.18
ratio, 15.16
re solve, 14.8
ref temperature, 13.94
reversible asymptote, 10.10
ROTATION, 2.15
rotation, 2.15, 15.13
runge, 14.11
runge jacobian, 14.10
runge kutta, 2.13
runge rollover, 14.11
spectrum, 10.9
storage, 3.17
strain hardening, 13.45
symmetrize, 14.12
theta, 14.10
theta method a, 2.13
time, 15.9, 15.17
transverse, 13.10
transverse elasticity, 13.10
two sided, 14.12
use e bar, 13.3
use last tangent, 14.8
USER MATERIAL, 3.9
variable friction, 12.17
variable environnement, 17.5
verbose, 2.6
viscous effects, 10.10
volumic, 10.7
yield surface, 15.17
Z-mat, 3.5
Z7 LICENSE, 17.5
Z7 MAX NB DOF, 17.5
Z7 TMP DIR, 17.5
Z7PATH, 17.5
zebaba v6.env, 3.13
ziegler, 13.59
salt, 13.87
save scalar, 6.5
save tensor, 6.5
scale, 15.9
security, 2.8, 14.8, 15.16
segment, 15.10
shear, 10.7
sido, 15.14
sintering, 13.91
skip first, 14.7
slip, 13.92
small, 15.14
solver, 15.15
Z-set — Non-linear material
& structure analysis suite
19.5
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement