5 Replies Latest reply on Dec 30, 2014 10:07 PM by Jared Conway

    Stress convergence

    Boris Fishman

      Hi guys,

      I'm running a simulation of an assembly and I'm interested in finding the stress on the cylindrical faces of the bolt holes.

      Although the displacement converges, the stress at the bolt head and nut head keep rising - which means singularity.

      I used bolt connectors with a pre-load of 40 Nm.


      Can anyone have an idea of how should I deal with it?

        • Re: Stress convergence
          Seckin Uslu

          Use iso clipping and probe tools. Examine the stress area.


          After a value, stress result must be rise exaggerated.. You can accept that value.

            • Re: Stress convergence
              Boris Fishman

              Thank you for your help!


              These are the probe results for the bottom face ( where the nut is ) and the cylindrical face.

              In the first one I can see that there is a singularity, but I don't know which value I can use.

              In the second one , I think there is a convergence , because after 4 loops of h-adaptive solution  the probe results look good.


              Can I assume that the stress at the bottom face (the first graph) is about 200 MPa ( where the graph starts to peak)?



              Bottom face.pngCylindrical face.png

                • Re: Stress convergence
                  Mike Pogue

                  I would not do that. I would not take any stress from that plot and assume it's accurate. FEA is terrible at bolted joints, welds, rivets etc. I would use FEA to extract the gross assembly behavior, including stress far enough away from the joints to converge. Then feed the calculated forces at the joints into a standard bolted joint hand calc, such as NASA Technical Memorandum 106943.


                  The forces and stresses around bolted joints are well understood, but they depend heavily on things that are difficult to measure or calculate accurately. Consequently, you should fall back on hand calcs and handbook references that have factored in the large associated uncertainties.