6 Replies Latest reply on Aug 8, 2008 2:47 PM by Ben Floan

    COSMOSXpress FEA Tech question

    Le Phan
      I'm impressed how easy it is to do simple FEA Static analysis in SolidWorks, COSMOSXpress. I just have a few techincal question ?

      Is the mesh type created Tets?

      If so, are they 1st or 2nd order Tets?

      In the Stress reporting, are you using the Cntroid of the Tets, or Corners?

      Thanks,

      Le

        • COSMOSXpress FEA Tech question
          Anthony Botting
          Hi Le: it is Tetrahedron, second order. For stress, they are calculating the stress at Gaussian points inside each element. I don't know how many Gaussian points are inside a second order tetrahedron, but they are generally slightly removed from corners. They way it was explained to me, where nodes of elements are connected, they obtain the stress at Gaussian points inside each element closest to the node of interest, and average the values, then plot that value. I hope this helps a little. Regards, Anthony Botting
            • COSMOSXpress FEA Tech question
              Le Phan
              Anthony,

              I always like to do a sanity check and compare softwares as well as hand calculations.

              Great ! When calculating stresses usings Tets, you must have second order. Using regular Tets (nothing else is stiffer) can give you false lower stresses and the value can be off as much as 25%. The deflection can be too stiff, giving you less deflection resulting in about a 25-28% less. Having second order Tets accomplishes this. I agree. It looks like SolidWorks is using the Centroid method of averaging. The results I'm getting from Abaqus/HeperMesh seems to be within say 5 % of the SolidWorks/COSMOSXpress.

              Thanks You for getting back with me. !

              Le
            • COSMOSXpress FEA Tech question
              Anthony Botting
              Le: Thanks for the heads-up. I can always use data like this. I have seen good results against test data, too. I worked for a while in aerospace industry. The US Air Force sent us test data (strain, displacement, vibration frequencies) and we were able to simulate the structure response in COSMOS and matched their test data within about 10%. I have also noted that COSMOS now posts NAFEMS benchmarks (www.nafems.org) and you can see they match within a few percent. You have to look at the on-line help in COSMOSWorks Designer or above (verification problems). The math is all the same! Thanks again. It's good to see Abaqus and Hypermesh numbers are similar. Tony
              • COSMOSXpress FEA Tech question
                Anthony Botting
                Hi Ben: As I mentioned, it's in the online help, and surprisingly (or not), they put it in the Tutorials section! Look in menu COSMOSWorks/Help/COSMOSWorks Online Tutorial. In the new panel that pops up, look for header "Verification Problems". I found some NAFEMS comparisons in "Linear Dynamics". Hope this helps. (Note this is in COSMOSWorks 2008).