4 Replies Latest reply on Jul 20, 2018 7:19 PM by Paul Millhouse

    How to simulate simple compression of a flexible part

    Paul Millhouse

      I'm trying to make some simulations of a rather simple part under static compression. The part is shown below. To make this simulation as simple as possible I have isolated the part (no assembly) and even removed features that would make things more complicated. I just want to see purely how much this part will deform under a given loading condition.


      The problem is when I apply a load and run the static study, the part does not deform nearly as much as I would expect it to. Below is a screenshot of my simulation and loading conditions. Basically I have the bottom fixed and a purely compressive external force applied to the top surface. The load increases in a linear manner from the front to the back of the part.


      The results above are for PEEK material with an Elastic modulus of 3.9 GPa. If I change the material to something really flexible, say Natural rubber with Elastic modulus of 10,000, the results are exactly the same (max displacement 3.4e-003). I realize that 100 N is a small load, but even when I increase the force to 1000 N the displacement is not nearly as much as I'd expect, and the results are the same with soft rubber as they are with stiff plastic.

      I also have "Large displacement" selected in the test setup, which has negligible effect.


      Now I understand that SolidWorks Simulation perhaps sucks for simulating compression of flexible parts, but any ideas for how to make this simulation better? Am I doing something wrong with my setup or loading conditions? Is there some setting somewhere that will improve this simulation? Is there some other material property I need to determine?
      Finally, does anyone have any experience using ANSYS, and would that yield better results?



        • Re: How to simulate simple compression of a flexible part
          James Riddell

          When you are using a highly deformable material you will no longer be able to reasonably utilize linear elastic static assumptions.  You should probably switch to a non-linear study.

            • Re: How to simulate simple compression of a flexible part
              Paul Millhouse

              Thanks for the input. I considered that and will try that next but with these forces and dimensions I should still be well within the linear portion of the material properties.

              Should I be doing anything else with the way the force is applied? Is there a way to adjust the length of time a force is applied for?

                • Re: How to simulate simple compression of a flexible part
                  James Riddell

                  What value are you using for Poisson's ratio (since it looks like you have Young's Modulus about right)?  Since SW doesn't seem to come with PEEK properties I presume you set your own up.

                  Why did you use the large displacements switch?

                  The part is fully constrained at the bottom, correct?

                  What is the size of the part?  If you are getting deflections > approximate minimum geometry size you have stepped outside the 'linear' realm.

                  Is your 100 N force total or maximum (i.e. much greater on the edge than towards the middle of the part)?

                  Do you get the same deflection value when you change the material to steel?

                  Did you change the material in the Simulation tree (lower window) or in the Model tree (upper window)?  When you change it in the Model tree it does not transfer to the Simulation tree.

                    • Re: How to simulate simple compression of a flexible part
                      Paul Millhouse

                      Actually no this version of SWX came with PEEK already in the library unless that was a carry-over but I'm pretty sure I wiped everything and installed this version fresh because we were having some problems with the license. Poisson's ratio is 0.4. Screenshot of the material properties attached.

                      I used the large displacement switch just because it seemed to compress slightly more but honestly I have no idea what that switch is for.

                      Correct, the part has a "Fixed" fixture applied at the bottom surface. You should be able to see the constraints in the first image in the post.

                      The outside dimensions are 12mm x 15mm x 5mm high.

                      The 100 N is total on the surface. Yes it increases linearly from the front of the part to the back. The way SWX applies force "Per item" I believe that means if you were to integrate the force on the top surface it would come to 100 N total.


                      The deflection decreases when I change the material to steel. But your next comment surprised me then because I've been changing the material in the upper tree ... Okay now when I change the material in the bottom tree to Natural Rubber I get an "excessive displacement" error and the solver fails. Maybe that's what I was doing wrong. EDIT: With natural rubber the solver fails even when I decrease the load to 1 N.

                      Speaking of the lower tree, when you change the materials there some of the properties are in Red and some are in Blue. What does that mean?


                      Still even with PEEK it doesn't seem to be displacing as much as I'd think. With 100 N the max displacement is 0.0034 mm and when I make the force 1000 N the displacement goes up to 0.034 mm. Maybe that's about right?


                      Thanks again.