12. J modeling#
12.1. Characteristics of modeling#
The behavior tested is MONOCRISTAL, in C_ PLAN and in 3D.
12.2. Tested sizes and results#
C_ PLAN modeling:
Variances (%) |
T_Pa |
T_rot |
\(\text{N1}\) |
|
|
|
V44_p |
0 |
0 |
0 |
0.009 |
0.008 |
0 |
VMIS |
0.022 |
0 |
0 |
0.022 |
0.004 |
0 |
TRACE |
0.024 |
0 |
0 |
0.008 |
0.018 |
0 |
3D modeling:
Variances (%) |
T_Pa |
T_rot |
\(\text{N1}\) |
|
|
|
V13_p |
0 |
0 |
0 |
0.07 |
0.02 |
0 |
VMIS |
0 |
0.13 |
0.018 |
0 |
||
TRACE |
0 |
0 |
0 |
0.4 |
0.1 |
0 |
3D tangent matrix:
Discrepancies |
\(\text{N25}\) |
\(\text{Max}(\text{Ktgte}-\text{Kpert})\) |
0.024 |
Note:
The precision in the C_ PLAN case is lower than in the 3D case for the « \(\text{Pa}\) » load case. This is because the De Borst algorithm only leads to the exact solution after a relatively large number of iterations. To avoid increasing the time CPU of this test, the iterations (and the stopping criterion) of the De Borst method are here taken by default.
The other values are satisfactory (good convergence, and no robustness problems even for long periods of time).
In particular, this test makes it possible to validate the rotation for this anisotropic behavior.