Q. Why does an unconverged nonlinear buckling analysis produce the error message,"Error opening split file= file.rst**" ?
It only happens in /POST1 when you try to read in the last successfully converged substep saved to the rst file.
It is not a large model and no file splitting has actually occurred.
A. This was a code defect in 9.0 and 10.0 (Defect report 44059). It is fixed at 11.0. Problem was that once ANSYS could not bisect anymore, rather than exiting gracefully at that point we went onto the output pass. Since we had hit an error, we only wrote the displacements to the rst file and skipped the element results. We then went through the normal "abort" process where we wrote the 99999 unconverged solution (both displacements and stresses). If you went to postprocess this rst file, the last solution (99999) was fine; SET,PREV reads incomplete information that subsequently triggers the misleading error message.
The code fix is to make sure we exit gracefully in this case.
|
Q. Why does an unconverged nonlinear buckling analysis produce the error message,"Error opening split file= file.rst**" ? It only happens in /POST1 when you try to read in the last successfully converged substep saved to the rst file. It is not a large model and no file splitting has actually occurred. A. This was a code defect in 9.0 and 10.0 (Defect report 44059). It is fixed at 11.0. Problem was that once ANSYS could not bisect anymore, rather than exiting gracefully at that point we went onto the output pass. Since we had hit an error, we only wrote the displacements to the rst file and skipped the element results. We then went through the normal "abort" process where we wrote the 99999 unconverged solution (both displacements and stresses). If you went to postprocess this rst file, the last solution (99999) was fine; SET,PREV reads incomplete information that subsequently triggers the misleading error message. The code fix is to make sure we exit gracefully in this case. |
||
|