4 Replies Latest reply on Aug 19, 2014 8:41 AM by Dan Bovinich

    Joints Are Forming Inconsistantly on Beam

    Dan Bovinich

      Hello,

      Previously, I used the SolidWorks freebie FEA version. Then my boss sent me to training so I could use Simulation. I just got personal training on Simulation "beam elements" and am starting to use it. I learned both Beam, Solid Body and mixed element methods. We build large weldment frames (stages) where I work and if I try to use "Solids" it take entirely too long to run the analysis (Yes, I have a powerful SW computer). So we use the default "Beams" method.

       

      I used the Beam element method on my first frame with great results. It was really fast solving it even if I made the mesh smaller than the default "fine" setting. On my second frame, I am having a lot of difficulty with it. When I create a new study, Simulation doesn't pick up all of the joints. The frame is perfectly symmetrical, one side of it has the joints perfect, the symmetrically opposite side completely misses them. It also misses one beam in center. Consequently, no matter what I do, when it runs the analysis the beams without the joint fly out into space considerably and gives bad results. I need good results because the Simulation report goes to our P.E. so he can stamp it, government officials, and our customer.

       

      Here is a pic of the frame (it is shown upside down):

      Missing_1.jpg

      I have tried everything I can with SW Simulation help files, Right Mouse Clicking, Googling, deleting study and making it new. Sometimes when I make a new study, the joints may change a little, in other words, the beam next to the one shown below may have the joint, but its neighbor will not.

       

      I have gone back into the original model and deselected the beam members, then adding them back again in the weldment. None of the beams were "mirrored" when I created them. They are created from a sketch. In other words, I didn't create half a frame and mirror it.

       

      I've tried to create a bonded  "contact set", but it gives me this error (see pic):

      Bonded_Contact_Error.png

       

      So my question is, "Can you create a joint if Simulation doesn't automatically pick it up during the creation of the new study?"

       

      I have sent this problem to my VAR, but he hasn't replied back yet.

      I am on SW 2013 & 2013 Simulation. XI Computer, Windows 7 Pro, Intel, 16 Gig RAM.

       

      I'm stuck.

       

      Thanks for any consideration you may give me. If you want the file, I can send it to you or post here.

       

      Dan Bovinich

      Rochester Hills, MI

        • Re: Joints Are Forming Inconsistantly on Beam
          Chris Michalski

          Have you tried adjusting the distance value that SW uses to determine separate joints versus combining them?  If the mesh is changing just slightly such that the joints that do/do not exist are actually being combined into the joint next to them it would look like a joint is disappearing instead of being incorrectly assumed.  Make sense?

          (I'm not a beam simulation guru so off hand I can't recall the exact term for this distance)

          • Re: Joints Are Forming Inconsistantly on Beam
            Jared Conway

            your question > can you create a joint that is missing > no

            my question > if this is symmetric, why not run this as a symmetric analysis and save yourself time in modeling and analysis?

            and as for the problem, likely due to the way the model is built, you need to look at that and most likely bug specific to some modeling workflow. delete the member, recreate it, and don't trim or anything. might fix it.

            or 2014 may fix it

              • Re: Joints Are Forming Inconsistantly on Beam
                Dan Bovinich

                Hi Jared,

                 

                Well, I installed SW2014 and it works perfectly. Sorry to bother everyone. If this was a year or so ago, I wouldn't be happy with Simulation.

                 

                I have heard that I could do a symmetrical analysis, so maybe I'll try that, just in case I get a big model in the future. The model shown in the pic above does a re-mesh and run in less than 5 minutes. Thank you for the help,

                 

                Dan Bovinich