5 Replies Latest reply on Sep 25, 2017 3:00 AM by Ananda Ganesh Madheswaran

    No Penetration contact sets don't work right

    Tom Yuschak

      I did some simple testing of no penetration contact sets in 2017 solidworks simulation professional and go disappointing results. I'm doing static analyses that start-off with some gaps (~0.25") between two different sub-assemblies. One sub-assembly is fixed and the other has forces applied. The forces should result in the two sub-assemblies coming into contact and then setting up the stress and displacement states. It just doesn't work well however. In the attached powerpoint file I simplified the heck out of the problem and it seems to show that no penetration does not work right. Any advice would be appreciated.

       

      Tom Yuschak

        • Re: No Penetration contact sets don't work right
          Bill McEachern

          Hi Tom,

          In SWX Sim pro the formulation for the contact problem is for small sliding. It is not a generalized non-linear solution and hence moving a contact of .25" on 1" blocks is not considered small displacement contact. Further, with no stiffness to restraint the blocks from infinite motion in a static analysis things go a long way as you found out. Applying soft springs that are immaterial to the problem at hand is one way to address this. I find the elastic foundation handy for problem stabilization. I use ABAQUS quite a bit where the contact capability is considered excellent. However, it is still a struggle to deal with "sloppy" contacts where a body in a static analysis has unrestrained motion available. You are always adding something and then taking it away once the problem becomes stiff - but these are general non-linear solutions where that is how things are done. In a "linear" contact analysis you only get one shot at the problem sequence so you need to have inherent stiffness to get it to solve correctly. If you have bolts with pre-loads it actually runs a sequence and you often need friction in the bolted connection to make the problem solvable. I understand that your real world problem has a space for a reason and that you want bearing contact to occur. You put them as close together as you can manage, you ensure that all the components have some stiffness so that the program can actually iterate to get to the equilibrium you want. Contact is inherently non linear and is probably the most difficult thing to pull off in FE analysis. It takes some skill and practice to get a handle on contact. Remmber it is a stiffness based method so you can't have zero stiffness in any direction or things get difficult really quickly. Swx sim is better than most and dealing with this but not always. Good luck.

          • Re: No Penetration contact sets don't work right
            Ananda Ganesh Madheswaran

            always prefer node to node contact & apply no penetration only for the contact / touching faces
            do not apply no penetration for complete parts

             

            Example 3 is the correct way of handling these type of problems..