3 Replies Latest reply on May 5, 2009 12:10 PM by Derek Bishop

    Weld beads / fillets in Simulation

    Dan Bertschi
      Hi All,
      I have a model in thin (.079") sheet metal in a tray form with stiffeners that run across the center area of the tray for strength. There will be fillet welds (not continuous) on the side walls and across the bottom.
      What is the best way to model this for analysis? Weldment part or assembly of parts? I've done enough research to know that weld fillets features (part or assy) will not meet my needs. So, I will be inserting the welds as features or parts.
      Finally, if using a single part, is it best to join the bodies, at the welds only, into a single body for analysis? Or would a multi-body part be OK with proper restraints at the welds?
      Running SW20080 SP5.0
        • Weld beads / fillets in Simulation
          Bill McEachern
          If it were me I would model it as an assembly with no weld materail or prep and everything coincendent, use global default as bonded and make sure that the compute free body forces is checked in the study properties dialog. Then (redundantly) use the find contacts sets and find all the contacts and set them all to bonded and make the contact sets (the global bonded apparently has regressed and does not provide the free body forces as it once did). Apply loads/BC and run. Use exploded view and interogate loads on welded surfaces using the free body force tool under the results folder. Once you have the loads you can use conventional weld sizing calcs to size the welds. You can then put them in and rerun the analysis if you feel it is prudent to do so.
          Good luck.
            • Weld beads / fillets in Simulation
              Thanks Bill,
              I figured I'd log in again so you can see my alter ego! I'm not a COSMOS whiz by any means, but I think I understand what you're describing. I think the prototypes are already built...imagine that...so, I may just go for the welded version. I mainly wanted to make sure the .sldasm version would be appropriate. I will model the welds as parts and constrain them to the sheet metal parts as-built. We are testing design modifications, not original designs.