csk62
Mechanical
- Jun 2, 2014
- 108
Just some observation from some analyses that I have run and I am wondering if anyone else has run across something similar.
For element birth/death in Abaqus I use the model change keyword in the interaction module. For some reason, I cannot use the element deactivation command in the first step. It must be the third or subsequent step in order to achieve convergence. Additionally, in the element deactivation step extremely small stress artifacts (on the order of 1e-12) occur, I assume because of subtle changes in the stiffness matrix. Also, as a consequence of the strain-free element birth, the deformed contour plots look crazy. I understand that this is a consequence of the procedure, and for all purposes except aesthetics, it doesn't matter, but is there any work around for the display besides using the undeformed plot option?
I have also experimented with jobs that involve mid-analysis material changes by definition of user specified field variables that "point" to specified material properties. I found this method, which some have suggested, to be inadequate. At the fringes of material interfaces there are unexpected jumps in stress fields which are unrealistic. Is anyone aware of a suitable work around?
I have been working on starting to develop suitable weld analyses against test data, and have even used the Abaqus Weld Interface which uses Python to automate multi-pass welding analyses. From weld analyses element birth/death and material definitions are important. (This is the background of these questions.)
I have been using Abaqus for a few years now and have a pretty good grip on what I would call standard analyses procedures, user subroutines etc. I just am interested to see if anyone would be willing to share their experiences with model change difficulties / tricks etc.
Thanks for your thoughts. If you read to this point, you have more patience than I do.
For element birth/death in Abaqus I use the model change keyword in the interaction module. For some reason, I cannot use the element deactivation command in the first step. It must be the third or subsequent step in order to achieve convergence. Additionally, in the element deactivation step extremely small stress artifacts (on the order of 1e-12) occur, I assume because of subtle changes in the stiffness matrix. Also, as a consequence of the strain-free element birth, the deformed contour plots look crazy. I understand that this is a consequence of the procedure, and for all purposes except aesthetics, it doesn't matter, but is there any work around for the display besides using the undeformed plot option?
I have also experimented with jobs that involve mid-analysis material changes by definition of user specified field variables that "point" to specified material properties. I found this method, which some have suggested, to be inadequate. At the fringes of material interfaces there are unexpected jumps in stress fields which are unrealistic. Is anyone aware of a suitable work around?
I have been working on starting to develop suitable weld analyses against test data, and have even used the Abaqus Weld Interface which uses Python to automate multi-pass welding analyses. From weld analyses element birth/death and material definitions are important. (This is the background of these questions.)
I have been using Abaqus for a few years now and have a pretty good grip on what I would call standard analyses procedures, user subroutines etc. I just am interested to see if anyone would be willing to share their experiences with model change difficulties / tricks etc.
Thanks for your thoughts. If you read to this point, you have more patience than I do.