7 Replies Latest reply on Dec 5, 2013 11:53 AM by Jared Conway

    Ball on cup contact with closely conforming surfaces

    Dermott McHugh

      I am trying to model contact between spherical surfaces, one concave and one convex (ball on cup contact). The difference in diameters of the two surfaces is very small, approximately 0.1 mm difference out of 30mm nominal diameter.

       

      The two surfaces are constrained to be coincident and tangent at a point, and force is applied in line with that point to the ball while the cup is fixed. I get an artifact in the contact pressure surrounding the point of initial contact that seems like it may be due to interferences of the spherical meshed surfaces. Besides that artifact, the solution seems reasonable. Has anyone else encountered this problem or have any ideas to fix it?

      Contact Artifact.png

        • Re: Ball on cup contact with closely conforming surfaces
          Bill McEachern

          It is hard to tell what one is looking at in the picture. Maybe some annotations on the picture would help to know what you are referring to. What quantity are we looking at? If this is pressure what does te vector plot look like? Contacts have noise due to the descritization so perfectly smooth is hard to come by but more lements help.

            • Re: Ball on cup contact with closely conforming surfaces
              Dermott McHugh

              I've attached some more images. All of these are contact pressure plots. As you can see my element size is quite small. The element size on the ball is twice this size. I'm also using 16pt Jacobian elements. The earlier image was perpendicular to the load line, and the circular artifact is centered on the point of contact between the ball and cup.

              Contact Artifact.pngContact Artifact Vectors.pngContact Artifact Vectors 2.png

                • Re: Ball on cup contact with closely conforming surfaces
                  Dermott McHugh

                  Another relevant point may be that the cup is much less stiff than the ball. The modulus of the cup is about 1 GPa, while the Ball is about 200 GPa.

                  • Re: Ball on cup contact with closely conforming surfaces
                    Jared Conway

                    what does interferences detection with and without coincident contact enabled show before you run the analysis? is there any interference?

                     

                    are you doing this with static or nonilnear?

                     

                    i assume you're using no penetration contact? what type?

                     

                    what does the artifact in the result cause you concern about if the rest of the solution is correct?

                     

                    and also for reference, mates do not apply in simulation. make sure that you have enough restraints to keep the model stable

                      • Re: Ball on cup contact with closely conforming surfaces
                        Dermott McHugh

                        Hi again,

                         

                        I think I found a solution to the problem, but first some answers to your questions, for anyone else finding this in the future:

                         

                        No interferences or coincidences. It doesn't register as coincident because the models only touch at one point.

                         

                        Static. No access to non-linear unfortunately.

                         

                        I'm not sure exactly what you mean by what type of no penetration, but I am using a manually chosen contact set, with 0.1 friction.

                         

                        Any artifact that I can't explain is a problem for me. It's tough to present your results on the contact when there's something clearly not realistic going on.

                         

                        And finally I am using adequate fixtures. The cup is fully fixed on the back side and the ball is constrained to move only on the load axis

                         

                        The solution ended up being checking the "Improve accuracy for no penetration contacting surfaces (slower)." It seems like that option forces a surface-surface contact instead of node-surface? In any case the solution looks much better (see below).

                         

                        I am still curious as to the reason for the artifact. My suspicion is that the circle represented where the separation between the two surfaces was so small that SW assumes they are contacting initally? If so it may accomplish that by translating nodes of either the ball or cup to be in contact with the other surface. Can anyone confirm that that is what's going on?

                        Contact Artifact SS Contact.png