Hi.
I have an issue related to recent change in ABAQUS new version.
I've recently found a problem related to cotact data in ABAQUS new version, 6.8. I use fil files, but I've heard from a Simulia engineer that data file also has the same problem (I have not checked it out yet.)
When I use 6.7 (or more old version), a file can be generated well for all nodes in a predescribed set related to contact (I use Contact file keyword and monitor contact pressure, CSTRESS). However, in 6.8, some nodes could not be found in the file, and there is no consistency. For example,
NSET (range of node number): 1 ~ 48, 101 ~ 148
6.7 version
1 ~ 48, 101 ~ 148 for all increments
6.8 version
Increment 1: 1 ~ 27, 101 ~ 127
Increment 100: 1 ~ 26, 101 ~ 126
The Simulia engineer said:
-------------------------------------
If you add
*contact print, frequency=1
cstress
You will see that .dat file also only outputs the active contacting nodes. These inactive nodes have zero values in 6.7 but occupy 5 times more space in your case. Although in either version, you don't lose any useful information, for large problems, this design will save a lot of space in .fil file if the active contacting area is relatively small.
Probably, you can define a counter to count how many entries are processed in your FORTRAN program which will give you the size information. There are ways in FORTRAN to process files. Please check some FORTRAN manuals.
----------------------------------------
However, I think it is definitely a serious problem. I really don't understand why ABAQUS had to change the postprocess. This change means the contact coordinates and contact pressure have different array sizes. For example, the contact happens irregularly in an increment for a 2D problem, the file can be something like this:
------------------------------------------------------------------
6.7
Node 1 2 3 4 5 6 7 8 9 10 11 ....
CSTRESS 0 0 1 2 1 0 0 1 1 0 0 ....
COORD 0.0 1.1 2.2 3.3 4.4 5.5 6.6 7.7 8.0 9.0 10 ....
------------------------------------------------------------------
6.8
Node 1 2 3 4 5 6 7 8 9 10 11 ....
CSTRESS 1 2 1 1 1 ....
COORD 0.0 1.1 2.2 3.3 4.4 5.5 6.6 7.7 8.0 9.0 10 ....
-------------------------------------------------------------------
If so, we could not match CSTRESS with COORD without information of node numbering. I need another file supplying contacted node numbers. Actually, space problem mentioned by the engineer could not be an issue when we consider the recent hard space.
For just 2D problems, it is even not so easy to get array size information at every increment. For 3D problems, it is more complicated. If I want to draw contact pressure distributions at every increment in a 3D problem, the array size could not give proper information because the contact surface is not a line. An array size could not guarrentee unique combination. I has to know spatial information. If the no. of line is 11 and each line has 11 nodes and the contact shape can change with increment, I should know not just 1 array size at an increment, but 11 array size including whole information of the node numbers, to know the information of contact pressure distribution (or contact area). Of course, I think I can handle it with the node numbering, but it will be very complicated and I don't want to do that kind of tedious work.
I hope SIMULIA solve the problem in the next version, but desperately, the engineer doesn't understand the problem (Actually, I am still arguing about that problem...).
Thank you for the reading.
JHLee