11 Replies Latest reply on Jul 12, 2013 11:43 AM by Jared Conway

    Define loads on reference geometry

    Wolfgang Grzesik

      Hello, I'd like to ask how i can define loads on specific reference geometry in SW 13. Actually i want to put loads on a glass plate in specific regions (circles) to simulate the loading with stamps. By now i did find the reference geometry function, but couldnt use the sketch i made upon it in the simulation. I did add a cut out picture of the model where you can see the circle, where the area load should be added on.

       

      Furthermore id like to know how i can define the number of elements within a height of a flat geometry wich is 0,2 mm thick. Lets say i want to have two elements within its height.

       

      And the third and last question: In the simulation results i want to make a cut in the laminate-geometry wich is parallel to the top glass area and can be moved in several coordinates but mainly to vary it within the z-coordinate to look on the strain and stress in different layers of the laminate.

       

      Thanks for the replies. I kinda like this software, even if the simulation part by now looks easy for me, but sometimes its lacking the right possibilities to set things up on the right place.

       

      With special greetings

        • Re: Define loads on reference geometry
          Dave Laban

          Wolfgang,

           

          For your first question, I would use the Split Line tool.  This will partition the face for specific loading, without further affecting the model.

           

          When defining the mesh, under "Mesh Parameters" you can define the element size.  Setting for 0.1mm should give you two elements through a 0.2mm thick sheet, but depending how big the panel is you may have a lot of elements to deal with.  Could this be modelled as a surface rather than a solid?

            • Re: Define loads on reference geometry
              Wolfgang Grzesik

              Hi Dave, awesome thanks for the split line tip Now it works with the specific forces. google-search couldn't help me much, even i found a youtube-video after you told me the name of the tool. (still have to change the language to english, a lot easier and better for supporting and talking about functions).

               

              Well with those mesh parameters its quite logical. I still have to accustom myself to the SW meshing. I was used to UGS NX5 Nastran and Comsol Multiphysics later on, so I'm still not used to meshing with just one type of element to choose. Still missing mapped or swept meshes and more mesh-options even this kind of meshing here is a lot more simple wich can be very nice too.

               

              I'm not sure if I should go for surface modeling. The model I want to analyze is a photovoltaics-module with several layers to mechanical stress and strain induced by mechanical loads. The layers are glass 3.2 mm, polymer1 0.4mm, silicon-cells 0.2 mm, polymer1 0.4 mm and polymer2 0.35 mm. I'm sure modeling with surfaces would decrease calculation time, even I now have 32 GB of Ram and 10 CPU cores and dont have to make things too simple. How would I do that in SW and how could I benefit from it?

              Hopyfully this questions aren't too common so I may search them easily.

               

              Thanks

                • Re: Define loads on reference geometry
                  Dave Laban

                  Glad I could help!

                   

                  The mesh limitations are my only major issue with SW Sim, hopefully in future we may get more options.

                   

                  Unfortunately due to the products we make I've never had chance to explore surface simulations beyond the tutorials so I am not sure if it is the right tool for this job.  One other way may be to define the plate material as a composite lay-up?  You'd have to wait for some of the experts on that sort of thing to chip in though.

                  • Re: Define loads on reference geometry
                    Jared Conway

                    follow the standard practices for using shell vs solids. if it is much larger than it is thick, use shells. you'll save a lot of computational time. take a look at the tutorials on shell meshing, or if you have a chance, the standard simulation class should cover all the different ways to create a shell mesh. for experienced users like you, i'd also recommend something like our mentoring service where we would work with you to convert your knowledge of a previous tool to solidworks simulation. in the end it is really just picks and clicks, the methods are all basically the same. we would also cover things like the best shell method way (sheetmetal, surface model or define by surfaces) for your specific problem.

                     

                    dave's suggestion on a composite is an intersting one if you're ok with them all acting together.

                • Re: Define loads on reference geometry
                  Wolfgang Grzesik

                  Hi, thanks for the ideas. Well all the layers are working together, theyre laminated in a heat process. So even i have a laminate, i could create midplanes for every layer, hide the 3D-bodies and they would still be connected? (would look strange to have 5 not visible connected layers).

                   

                  I didn't work with the composite tool by now. But I'll consider it.

                   

                  I'm right now rushing thru the Book FEM-Praxis with SolidWorks by Michael Brand. I guess even if there are lots of very simple models used there, it might give me some better basics with SW.

                    • Re: Define loads on reference geometry
                      Jared Conway

                      "Hi, thanks for the ideas. Well all the layers are working together, theyre laminated in a heat process. So even i have a laminate, i could create midplanes for every layer, hide the 3D-bodies and they would still be connected? (would look strange to have 5 not visible connected layers)."

                       

                      >depends on how you're defining shells. sheetmetal > yes. manually, no. you'd need to use contact sets which adds constraint equations to bond them together. this is where composites might have an advantage.

                       

                      Haven't heard of the reference you're describing but post your experience with it.