7 Replies Latest reply on Jun 10, 2013 12:05 PM by Jared Conway

    Advice on bolted shell assembly - Static analysis

    Ruairi Mulligan

      I'm trying to conduct an analysis on a shell/shell assembly.

      Essentially, I have 4 legs bolted to a 3mm sheet metal pressing which has a 8mm crossmember joining its sides at the middle.

      Ideally I'd like to apply an upward force of 25kN to the crossmember and be able to see if this will affect the main pressing & the bolts connecting it.

       

      I had difficulty in using a bolt connection between the shell faces so I bonded them at the bolt holes instead. (Can shell parts be bolted together?)

       

      I have fixed the based of the legs but I seem to be getting excessive displacement on the sheetmetal pressing.

      Even if I reduced the load to 0.5kN as per the attached images, I get a displacement of 7.5mm which is obviously way off.

      I've attached an animation of the displacement. (Its a  swf file so just drag it into a browser)

       

      If anyone could tell me where I'm going wrong with this or could provide any advice, I'd be very grateful.

        • Re: Advice on bolted shell assembly - Static analysis
          Jared Conway

          shell/shell bolts > should be possible:

           

          http://help.solidworks.com/2013/English/SolidWorks/cworks/r_connector_bolt.htm

           

          but some older versions didn't have this capability.

           

          but make sure to take a look at S-058593 in the solidworks KB about shell/shell bolt connectors.

           

          any error messages when running your simulation? large displacement maybe? do you ahve any solver options enabled like soft spring/inertial relief? looks like to me that you have a rigid body motion in your model. if thats the case, you probably got an error and chose "no" which would lead to bad results.

           

          have you done the pressure loading on a part on its own? if that works, then it is your assembly setup that is the problem. if the part is the problem, you'll have to validate your expectation.

            • Re: Advice on bolted shell assembly - Static analysis
              Ruairi Mulligan

              Hi Jared,

               

              Thanks for your reply.

               

              I didn't receive any error messages regarding large displacements etc.

              The solver is set to automatic and I don't have any of the options enabled.

               

              The main pressing (yellow part) has a lot of features but I didn't think it would cause such a failure.

              I've bonded the flanges that would normally be welded & used no penetration to prevent the legs from bulging into the main pressing.

               

              I tried to create a solid part but it caused a lot of trouble as well.

               

              I have tested the pressing on its own with split lines where  the 8mm crossmember would be and it still causes excessive displacement with just a total force of 50kg.

                • Re: Advice on bolted shell assembly - Static analysis
                  Ruairi Mulligan

                  I had another go at this simplifying the main pressing and also changing the load to 15000N; however, I'm now having trouble with two parts of the study (the fixtures & converging the results).

                   

                  1: I can't replicate the bolted legs when I remove them. I fixed two holes on one of the front corners and then used reference geometry on the remaining holes (I locked each hole to prevent movement in the Y-axis). Looking at the rear holes after performing the FOS study it would seem that the load is not transferring to them.


                  Front_iso_(FOS).png

                  Rear_iso_(FOS).png


                   

                  2: When trying to converge the results, I'm receiving mixed information based on the stress & displacement.

                  As I increase the mesh density I was expecting the displacement results to converge, which they seem to at one point, but with further refinement they start to change again. Any ideas why this might be happening?

                   

                  Mesh_table.png

                  • Re: Advice on bolted shell assembly - Static analysis
                    Jared Conway

                    "I have tested the pressing on its own with split lines where  the 8mm crossmember would be and it still causes excessive displacement with just a total force of 50kg."

                     

                    sounds like to me you need to run some test cases against hand calculations to set your expectations. only until then can you determine if your setup is valid.

                      • Re: Advice on bolted shell assembly - Static analysis
                        Ruairi Mulligan

                        It seems safe when compared to the hand calculations but I've only looked at the tensile stress of the pressing sides.

                        Regarding the study on the singular part with the split lines, visually the results were completely off and it was clearly evident that something was wrong with the setup there.

                         

                        When I re-analysed the assembly with the legs removed, the results looked more realistic but I'm still concerned with the two issues highlighted in my last post. I believe that they undermine the fea results altogether.

                         

                        Essentially I'd like to know where I'm going wrong with the setup etc.. and also know how to correct it for future applications.

                        It would be great if I could learn something from this analysis, and have more confidence with fea when approaching similar studies.

                          • Re: Advice on bolted shell assembly - Static analysis
                            Jared Conway

                            "It seems safe when compared to the hand calculations but I've only looked at the tensile stress of the pressing sides.

                            Regarding the study on the singular part with the split lines, visually the results were completely off and it was clearly evident that something was wrong with the setup there."

                             

                            i would spend some time getting hand calculations to match simulation so that you can get confidence that the software is solving the problem the way you expect. check the help/tutorials for the validation examples for some more insight on how the software is validated.

                             

                            beyond that, i think i'd need to see your models to better undersatnd where you might be going wrong.