7 Replies Latest reply on Oct 2, 2008 4:56 PM by Vince Adams

    Stress dosn't converge

    Jim Wisén
      When I try to check if the stress converge with smaller element size I get the following results:

      Element size Stress- fillet Stress fly
      0,5 995,42 3798,9
      1 1165,7 2881
      2 1724,6 1277,73
      3 1706,3 1120
      4 1670,6 1182

      I can't find any singularities that could cause this.
      Any suggestions how I should troubleshoot?

      Edit: Attachments works now
        • Stress dosn't converge
          Vince Adams
          Jim, can you post some pictures of the problem area(s)?
            • Stress dosn't converge
              Jim Wisén
              I have added pictures now. Let me know if you need any explanations.
                • Stress dosn't converge
                  Vince Adams
                  Hi Jim, I'm still not seeing the problem. It looks like a small enough model. Can you post it or send it to me directly?

                  Only send the SLDPRT, not the CWR file with all the results. I'll re-run it here.

                  Vince
                    • Stress dosn't converge
                      Jim Wisén
                      Thanks for your answer.

                      The assembly is quit big so I'm not able to send it to you.

                      Could the problem be because the two other components on top have different mesh size, see picture? The pressure is applied on the top glass.
                        • Stress dosn't converge
                          Vince Adams
                          Is the stress diverging across entire features or are you isolating a vertex that you know doesn't move when the mesh size changes?

                          It might help if you pinpointed the location of diverging stress on both the results JPG and the model JPG.

                          Are you bonding all these parts together? If parts are bonded with incompatible (dissimilar) meshes, the bonding element could be reaching past your expected interface to a node out on the sheet metal that could result in an inadvertent singularity. Can't tell from the images you provided if this is the problem.

                          Can you try making split lines at all bonded interfaces in the area of concern to ensure the bond doesn't reach outside of the mate?

                          Another option is to run the sheet metal part by itself with an applied load similarly distributed to the assembly intreraction. The results will be different but if this converges whereas the assy doesn't, that points to the bond itself as being the likely culprit. If it diverges by itself, then there is something in the part.

                          -- Vince
                            • Stress dosn't converge
                              Jim Wisén
                              I have only studied a small area (two faces circled in one picture below) since thats only where the stress is near yield stress.

                              The new pictures shows where the stress diverge.

                              I have changed to the same mesh size where different parts are bonded.

                              I've added split lines at all bonded areas with no difference then before.

                              If a have about 180000 nodes or more I get large displacement error with the glass, no matter which part I have mesh control on. Perhaps this have nothing to do with the other problem, but still anoying when i search for convergence.