10 Replies Latest reply on May 2, 2013 2:02 PM by Jared Conway

    2D Axisymmetric non-uniform loading

    Justin Strempke

      I'm wondering why there is no option to use non-uniform forces/pressures in 2D axisymmetric.  I understand they could only be in one direction (along axis of symmetry), but it would be hugely helpful for hydrostatic loading cases (cylindrical tanks)...  Is there a work-around for this?  I've tried a 1-deg solid revolution, but the circular symmetry fixture doesn't seem to apply properly (pulls away from the very center along rotation axis, and doesn't seem to capture hoop stress correctly).

        • Re: 2D Axisymmetric non-uniform loading
          Jared Conway

          This is currently a limitation that I found in the SolidWorks KB. (Customerportal.solidworks.com > KB)

           

          601473

          Add nonuniform load for 2D Simplification Studies.

           

           

          The only workaround I can think of is adding split line and adding a "discrete" non uniform load or using a quarter/half of the model with the appropriate conditions. From your description, it sounds like you might not have the right symmetry conditions applied. It should work the way you describe.

            • Re: 2D Axisymmetric non-uniform loading
              Justin Strempke

              Good to hear on the fix/update, at least that it's in there!  Answers that question

               

              For the symmetry conditions, there's three options:  Face 1, Face 2, and axis.  I have a solid that begins at the axis of rotation and extends outward (slice of a hollow cylinder), then revolved a certain degree (therefore the solid should be continuous about the axis, one would think, and the cyl. symm. preview shows).  However, when the static sim is done with a small load (2500 psi hoop stress in normal conditions), it shows 50 ksi at the outer surface, and both inside edges are pulled away from the axis.  It gives me an outward displacement of 3.9 million inches...  If I fix the bottom, the top moves out as well and shows 250 ksi stress so something isn't right.  I'm not sure where/how it could allow that or where it could be wrong.  It's almost like the force isn't being updated to the symmetrical portions or something.  I'll try it in 2012 and see what happens

               

              UPDATE:  Works fine in 2012, so it's either 2013 in general or my particular install.  Can anyone else verify this quick (model size/material irrelevant, just check for disconnection)?  I can post a model if need be, but just as simple to make one up and not sure mine aren't corrupt.

                • Re: 2D Axisymmetric non-uniform loading
                  Jared Conway

                  please post model and i'll try it and replicate my own. the only thing I can think of is that in an early SP of 2013 there was something that changed with symmetry BCs that might be affecting your results.

                   

                  best to post the model because I don't want to make a guess at what you mean with "hoop stress load".

                   

                  the only thing that you description leads me to believe is that there is some kind of rigid body mode but you would have a large displacement error or something similar and would need soft springs to actually solve. And that would result in the strange displacements and stresses.

                    • Re: 2D Axisymmetric non-uniform loading
                      Justin Strempke

                      I've gotten with my VAR on this, it seems it matters which face is defined as 1 & 2 for the separation issue, but it still doesn't compute stresses properly.

                       

                      I'm not sure what you mean by 'hoop stress load' either, never mentioned that.  I mentioned hoop stress, based on thin-walled theory with a certain radius, pressure, and wall thickness.

                       

                      Anyway, here's the part (with incorrect face selection).  The bottom surface is roller as-is.  Attached is a pic with a large revolution, showing the odd stress deformation, and separation of the symmetric edge (bottom right of image).  This is the top view of a quarter-cylinder.

                       

                      circ rev top.JPG

                      Now I want to head back to 2012 and see if I just got lucky and put in the faces correctly, or if it didn't matter.  It did show the correct stress however, unlike 2013 in any form.

                        • Re: 2D Axisymmetric non-uniform loading
                          Jared Conway

                          "sim is done with a small load (2500 psi hoop stress in normal conditions)," is what i'm talking about with respect to hoop stress load.

                           

                          sorry about the rambling below, it is basically my thought process of how i went about troubleshooting the issue. in the end it looks like there is a problem with circular symmetry. replace with on flat face in the normal direction and report the issue to your reseller to report to solidworks. this should be done with shells so you'll have to define the symmetry manually anyways.

                           

                          i opened your file and ran it in 2013 sp2.0, first thing it comes up with is a large displacement warning issue. when this comes up, the results are no good which would probably explain the issues with your displacement and stress. Also you have soft spring enabled, which will also cause issues with your results. If you try and run with large displacement (even with soft spring enabled) it also fails. This indicates either a lack of restraints causing rigid body motion or large deformation beyond what a linear static, small displacement analysis can handle. Since your load is 1 psi, I'd guess the former but just to check, I ran it with 0.00001psi, still fails. Switched over to nonlinear, have to add the circular symmetry restraints manually (no circular symmetry in nonlinear), it runs. (note, had to use curvature based mesh, probably because of the aspect ratio and pointy ends, this should really be done with shells) so it isn't necessarily the way the constraints are setup. back to static analysis with the same boundary conditions, runs without issue. switch to 1psi load, it runs. so it isn't the load, we've come down to there either being an issue with your study or the cylindrical symmetry. switch to cylindrical symmetry, failure. I went to go check if there are any rules about defining that BC, unfortunately it looks like that help topic is missing. so i went and checked the solidworks KB, these 2 SPRs look likely. Both are still open:

                           

                           

                          the next step for troubleshooting would be to try a few more models to see if the issue is specific to this geometry type/loading condition or all circular symmetry. for now, i'd use the workaround.