Update: loop parallelisation (instead of domain) seems to work, although less efficiently. If anyone has a better way of parallelising (with domains) using the .env settings above, please advise! :)
Thank you FEA Way,
I have no idea how to do this, zero experience with overwriting settings using the .env files. I tried: can't find where this variable is and I don't know how to modify it. What does "set to ON" mean? Do I write XPL_LINK_NOMERGE = ON? Or = True?
Abaqus could make our lives...
Hi and sorry to join the party so late. I've had my own fair share of parallelisation errors, the last one being "domain decomposition error":
I submit my model to the cluster with 5 cores, parallelisation in 5 domains, dynamic load balancing=off (because I have a job restart), mp_mode =...
Yes, already doing that (with the default linear energy dissipation). It seems to work with beam elements (huge loss of stiffness as expected) but I cannot choose element removal in CAE or in the .inp file, is this normal?
When I apply it to shell elements I still can't cause this stiffness...
Thanks FEA way,
So to achieve the stiffness degradation I would have to use a damage model. However the same as the above (i.e. no failure) happens when I implement the Hashin criterion. I don't see what is so wrong...
Hi,
I'm trying to implement material failure in its simplest form, as it is available in Abaqus/CAE at the elasticity table (fail stress). Even when I put small values (e.g. 0.2), the analysis shows my slab reaching a stress of 68 without failing. Where am I wrong? How can I correctly implement...
I need to dig into that. It looks like actual parts are excessively deforming indeed (even breaking). Connectors are broken, so no issues there. I have no material failure, so it's weird that the elements themselves would break. Maybe the solver is going a little crazy due to contacts? The only...
I have had no luck trying to contact 3DS before. Also, our cluster support receive no support for Abaqus, apparently. I will try contacting another IT department.
Thank you for the tips. If you have any other ideas please let me know..
Konsti
Hi,
There is a sudden change at that moment, as two connector break and the structure is accelerating. However, I'm surprised that this is causing an error in the explicit solver, which is supposed to be working well in such sudden changes, and if fact it has most of the time in my experience...
Hi,
Can anyone help me understand why is Abaqus issuing an "excessive rotation" error and aborting an analysis, when this analysis is using the explicit dynamic solver? You can find attached the error from the status file screenshot. I also attach the message file with a warning about the...
...of cores, turning off dynamic load balancing (since I have predefined fields in my analysis, which are not supported by DLB), and setting memory explicitly to 20GB (5 cores*4GB each).
Is there something awfully wrong that I am doing here?
Thank you in advance for your help,
Konstantinos
A bit late in this party, but I would like to note that I use connectors a lot and it took me a while to understand how to connect them to the rest of my model. What I do now is divide and mesh my elements such that I connect mesh nodes (and not points) together.
...model .inp file using python (without the initial state), then go into the .inp file and delete the parts and assembly and replace them with the *IMPORT and *IMPORT ELSET keywords. Or script the entire model directly in the .inp file, however I would need to learn that from the beginning (I...
...the connector forces and displacements jump to their values from the previous analysis instantaneously in the first analysis increment. This seems correct. Therefore I do not need to do any extra imports? Simply copying the model and importing the initial state of the *instances only* is enough?
...Edit: diving into the documentation, I don't see an explanation of "configuration" and "state" of connectors: configuration being to which nodes they are attached to? And why not being able to import both? Do I need to *not* import the material state in order to import connector state?
Konsti
...of details need to be taken care of regarding the materials, connectors, etc. In my case for example, the connector forces and displacements are *not* transferred to the new analysis by the initial state predefined field, because I can only choose part instances when I set this predefined...
