General Mechanical

General Mechanical

Preventing ANSYS from skipping commands after non-convergence in Batch-Mode

    • FordPrefect
      Subscriber
      Dear all,nWhile solving in batch mode, the program terminates the analysis run if a load step does not converge and jumps to the next finish-command in the APDL-script ignoring every other command following the last solve-command. The output file of the batch run says:n *** NOTE *** CP = 1689.297 TIME= 16:12:43n Solution not converged - Skipping the remaining commands to the FINISH command.nIs there a way to change this behaviour, so that ANSYS doesn?t jump to the FINISH-command but issues all commands following the solve command regardless if the load step converged or not?nBecause if the solve-command is nested in a *DOWHILE-loop even the *ENDDO is ignored. If there is a second DO-loop later in the scirpt, the *ENDDO of this second DO-loop ends the *DOWHILE-loop and the second DO-loop remains active. nFor Example:n*DOWHILE,LS_checkn ! set-up LS with NCNV,2n solven ! some more commandsn*ENDDOnFINISH n/POST1ni = 1n*DO,i,1,10,1n ! input post-processingn*ENDDOnI hope this makes sense to anyone. Thanks for your support.nAll the best nAndyn
    • FordPrefect
      Subscriber
      BTW: Using the arc-length method with ARTRM,L seems to be a workaround. However, the ALM is not always the method of choice... n
    • FordPrefect
      Subscriber
      The obvious workaround is using the RESTART function after each load step. The RESTART-function allows you to put a FINISH command and a subsequent check on the load step status within the DO-loop. However, this workaround works only with the Newton-Raphson-method. The arc-length method does not allow restarts. n
Viewing 2 reply threads
  • You must be logged in to reply to this topic.