2 Replies Latest reply on May 9, 2008 4:17 PM by 1-D88IXC

    Non-converging Gaps, Stiffness singularities, Auto Reset

      I was trying to simulate a o-ring deformation after compressed for certain distance, and this message comes out:
      Non-converging Gaps, Stiffness singularities, Auto Reset

      Anybody can explain this? Thanks in advance.
        • Non-converging Gaps, Stiffness singularities, Auto Reset
          Anthony Botting
          Hi Shaodun: For the "Stiffness singularties" message, it can be that you may have reached a 'limit' in COSMOS (assuming you adjusted the "Stiffness Singularity Elimination Factor" to 1). These messages can arise when elements get too distorted (too squished or too stretched-out for the math to work well), and the "Jacobian" goes bad and the calculations just won't work any longer. A solution is to pay more bucks (a lot more bucks) for software that actually stops at that step, and re-mesh the field with new elements that have nice aspect ratio and Jacobian, and then start the nonlinear analysis again. Some of these issues of element distortion are handled by the hyperelastic material models that COSMOS has (you probably have already tried that). I believe they can handle strain up to about 600%. For the "Non-Converging Gaps" message, that can be a consequence of the stiffness singularities. However, beware that COSMOS won't do "self-contact", but a nice workaround is to create separate parts of the O-ring and re-assemble them and you can set-up contact between the parts of the O-ring that you know will touch. I hope this helps. Regards, Tony B.
          • Non-converging Gaps, Stiffness singularities, Auto Reset

            Thanks for the timely reply, I guess I need to refine the mesh so that Jacobian ration will be maintained.
            Initially I was trying to solve this problem with less mesh as solver took very long time.
            Unlike Ansys or COSMOSM, CosmosWorks can not do a FEA just in 2D plane, I think that will be faster.