2 Replies Latest reply on Nov 12, 2011 12:12 AM by Aaron Moncur

    How to Apply Bolt Connectors to Shell Elements

    Aaron Moncur

      I am running an analysis on a large-ish assembly, the components of which are all thin (overall size of largest component is about 40" x 40" x 10" and 3/16" thick...other components are 12 gauge sheet metal).  When I apply bolt connectors to the sheet metal bodies, the visual icon looks as it should (with the blue "rings' on the selected edges of the bolt holes:

      Assembly, Solid Bolt Connector.JPG


      However, when I mesh the model (and treat the bracket as sheet metal), SW creates a shell element out of the sheet metal bracket the surface of which is the middle of the bracket.  In other words, the circle edge I originally selected for the bolt head no longer exists, as the shell surface containing that edge is now offset by half the material thickness away from the originally selected edge/face:

      Assembly, Shell Bolt Connector.JPG

      Assembly, Shell Bolt Connector, Down.JPG

      You can see in the image above that the visual indicator for the bolt head selection is offset from the shell element.  When I run the analysis, I get an error associated with the bolt connections.  I have tried doing Edit Definition > Offset in the sheet metal part to line up the shell with the originally selected edge/face for the bolt connector, but when I remesh it still shows the same mid surface created for the shell.  The offset issue may not even be the cause for the analysis failing, but the error message relating to bolt connectors makes me think it is.  Any thoughts?

        • Re: How to Apply Bolt Connectors to Shell Elements
          Bill McEachern

          First rule of shells in SWX Sim: make your own surfaces and delete the solid bodies. Then you do not need to know the secret decoder ring secrets on how you map the solid body attachements to the autoextracted shells. Even if you did know the secrets of the  decoder ring it probably wouldn't help in this case. Skipping the auto extraction process means you can attach the bolt connector directly to the shell edge and avoid your issues. I can not recall if bolt connectors are supported with shells or not but if they are this is how to get it done with out errors. SWX should just abandon this whole automatic thing - it is a total cock up invented by people who do not use this stuff ever excpet in a demo for which it looks quite spiffy. They should just auto extract the sufaces, hide the solid bodies and exclude them from the analysis - then you would be forced to pick geometry on the actual shell surfaces that the elements are mapped into. I just do all this manually and it is pretty staight forward and realtively painless but work non the less.


          Why are you using bolt connectors? Are you trying to determine the external loads on the bolts, stresses near the bolts in the adherents or are you interested in other areas and just think you shoudl use olt connectors because bolts are used in reality?

            • Re: How to Apply Bolt Connectors to Shell Elements
              Aaron Moncur

              Thanks for the reply, Bill.  The SW help files state the following:


              "You can define bolts through a mixed stack of solids, shells and sheet metal bodies."


              So, in theory, I should be able to use a bolt connector with shell elements.  I'm using bolt connectors because, among other things, I am interested in the stress regions around the bolt holes. 


              Thanks for the suggestion to create my own shells.  One question: What if I have one part (a sheet metal plate) that, before shelled, has bolt connections referencing edges on opposite sides of the plate?  Then, when I shell, only one side of that plate remains, and the bolt connection that referenced the opposite side (the side that is no longer there after shelling) is left without a reference edge.  How can I get around that?


              I just ran this thing as a solid body and it burned for 12 hours then told me large discplacements were encountered.  Agh!  I need to understand how to do this with shells in hopes of reducing the processing time (for reference, there were 1.2 million dofs, and I still didn't even have more than a single "element width" across my wall thicknesses.