0f70a19266
Fp matters, not Fp^-1
...
mathematically absolutely equivalent, but numerically not. Sometikes makes a
huge difference in convergence behavior, even though abs(det(Fp)-1) is
in the order of 1e-15
2020-02-13 14:48:32 +01:00
fc82ec9b1b
tiny deviations from det(Fp) = 1 cause convergence problems
2020-02-13 09:18:49 +01:00
b861ad11c1
tiny differences between orientation conversions result in more
...
iterations
Seen so far only for plasticityDetectChanges test (confirmed for dislotwin and
phenopowerlaw).
The max difference between the entries of the rotation matrix from
orientation0%asMatrix() and eu2om(Eulers)/eu2om(orientation0%asEulers) is
1e-15. This is the ratio of km/atom radius! Still, the number of
iterations is consistently higher.
Results are the same. I believe this is a strange coincidence where one
particular orientation causes problems. The current version recovers almost the 'good' behavior
of math_EulerToR(Eulers)
2020-02-12 20:49:02 +01:00
8770613e9c
better readable
2020-02-12 06:26:22 +01:00
ab475b7c6b
need 'error return'
...
revert from change in 5b72110d
2020-02-12 06:02:37 +01:00
08174a119f
consistent name
2020-02-11 17:50:07 +01:00
5b72110d0a
simplified
2020-02-11 17:47:48 +01:00
e212f91fac
print statements prevents reading code
...
first re-structure and clean, than re-implement where useful
2020-02-11 17:41:30 +01:00
6463fcdabd
consistent names
2020-02-11 17:36:43 +01:00
c2c84d698f
Merge remote-tracking branch 'origin/development' into MiscImprovements
2020-02-11 17:11:43 +01:00
e932b386b9
Merge branch 'development' into MiscImprovements
2020-02-07 17:11:01 +01:00
5d4d1dcf9a
all nonlocal parts are fully explicit
...
i.e. they are based on converged (partioned0) states
2020-02-07 12:41:01 +01:00
f854dc27e9
explicit dotState for nonlocal
...
all flux related quantities are calculated based on the converged
quantities
2020-02-07 12:23:22 +01:00
47109b903b
Merge branch 'plasticity-submodule' into 'development'
...
Plasticity submodule
See merge request damask/DAMASK!120
2020-02-06 10:13:48 +01:00
3f96c12e06
avoid code duplication
2020-01-29 11:09:57 +01:00
9c7f6811a2
use default string length
2020-01-26 12:17:59 +01:00
e532641015
dependency on element not needed for homogeneous meshes
2020-01-25 09:24:42 +01:00
cad6fc7843
Merge branch 'development' into plasticity-submodule
2020-01-13 20:57:12 +01:00
4ebd89c040
shape is known (no need for automatic allocation)
2020-01-03 13:47:04 +01:00
34af10fac1
using default string length
2019-12-21 12:39:54 +01:00
f0d3b29b82
Merge branch 'development' into MiscImprovements
2019-12-21 06:53:56 +01:00
4b6388fbb2
always use HDF5 output
2019-12-18 20:05:51 +01:00
9b67ead62f
removed postResults completely
2019-12-11 00:10:02 +01:00
07ebd8d1b3
only damage/thermal 'homogenization' postResults is currently needed
2019-12-09 05:48:37 +01:00
acc252ea5b
thermal/damage constitutive (i.e. source) results are not tested
2019-12-09 05:38:15 +01:00
ab1f0dc16b
submodules allow inter-module communication
2019-12-04 22:31:15 +01:00
c36a5bdfbb
bugfix: calculation of size of postResults was wrong
2019-12-02 23:06:28 +01:00
c7d675a1e9
not needed anymore
2019-11-30 16:03:18 +01:00
97474e05a8
not needed anymore
2019-11-24 11:01:04 +01:00
e20477099e
no crystallite output
2019-11-24 09:46:46 +01:00
50b48b8bf7
IP neighbourhood deprecated
...
trivial for grid and, hence, not written out.
Test for marc/abaqus/mesh would make more sense
2019-11-24 07:54:34 +01:00
872c85112f
bugfix, Lp not available anymore for crystallite
2019-10-20 14:02:12 +02:00
f93336b072
postResults for isotropic not needed anymore
2019-10-20 13:21:51 +02:00
11993a3ad1
phasing out crystallite output
...
some outputs are still needed for a few remaining tests.
2019-10-20 12:03:08 +02:00
603973e963
correct output for multipe integration points
2019-10-19 19:55:00 +02:00
de912c22b4
Merge branch 'development' into HDF5-spectral-displacements
2019-10-18 12:54:12 +02:00
e85e13b380
[skip ci] only corrections to comments made
2019-10-11 15:21:29 +02:00
8fd9341e39
not needed
2019-09-27 14:23:19 -07:00
77011a5dba
this fix restores the previous behavior of dislotwin
...
very strange behavior in plasticity_detect changes for dislotwin. This
fix restores the old behavior with respect to number of iterations.
Using orientation%fromMatrix() results in much more iterations and an
increase in runtime for the test from approx 6 min to 40 min. Results
still match in the end. Also, crystallite_Fp0 differs by approx 1e-15
only between the two methods.
I assume that something is wrong with either dislotwin or the
state/stress integration
2019-09-22 07:46:30 -07:00
53283d5c01
using newer interface
2019-09-20 18:20:33 -07:00
ad83c8541d
same names as in python
2019-09-20 17:18:09 -07:00
792dda866d
rotation class has consistent set of conversions
...
crytallite_oriention0 was essentially a copy of material_EulerAngles
2019-09-19 19:42:28 -07:00
be0d961954
cleaning
2019-09-19 13:40:03 -07:00
cf37f8d405
added some lines to help while debugging
2019-09-06 15:45:49 -04:00
5ff4664b6d
polishing
2019-06-30 22:09:51 -07:00
0c52262e4a
white space adjustments
2019-06-15 15:44:15 +02:00
57a0d33293
consistent names
2019-06-15 14:33:20 +02:00
7a76740c31
using new names
2019-06-14 08:51:23 +02:00
0f531d5dee
clearer name
2019-06-14 08:42:12 +02:00
9dfe71aa06
better readable
2019-06-10 09:42:23 +02:00
Martin Diehl