4 Replies Latest reply on Dec 8, 2015 12:08 PM by Danny Bouldin

    Can you show True Max Stress in a part with singularity stresses

    Danny Bouldin

      When you use the bolt function in SImulation, the stresses around the bolt hole are fictitiously high. Similar when you bonded contact, the "edges" of the contact again sometimes produce singularity type stresses. Is there a way to make Simulation ignore these areas to get a better plot of the true max stress whether principal or shear? I know you can use "probe" to manually spot check. But you can easily miss what the true max stress by spot checking especially in 3d parts.

        • Re: Can you show True Max Stress in a part with singularity stresses
          Richard Wehmeyer

          Max stress will not be a point value but rather the valus near your discontinuity.  No one point will be your max stress.  FEA cannot practically give you an infoIn/numberOut answer like a hand calc will.  When doing an FEA analysis you need to use judgement.

           

          Once you have a good mesh size in the area of intrest - ignore the first set of elements.  When the value of the second set stays the same(1-3%) your mesh is the correct size.  Probe 5-10 of these elements and use a value that describes them the best.

           

          To get a feel for the numbers the best thing to do is run several FEA's on objects that you have a hand calc for and verify.  hole in plate, tensile sample, pressurised ring, ect.  This should always be done on a FEA anyway, find a spot that you can h-calc on and make sure it is close.

           

          E.T.A.  To make the FEA show better on charts, always reset the min and max.  I tend to set mine 0-yield and a second at 0-acceptable.  This way the highest stresses are indistinguishable from singularities on the chart.

            • Re: Can you show True Max Stress in a part with singularity stresses
              Danny Bouldin

              I appreciate your comments. In FEA packages such as FEMAP or PATRAN, you can set up groups and hide the elements that make up the discontinuity such as a hole or where you attach MPCs or other places that cause fictitious stresses. Hiding these areas allows you to see the load path/ stress in the individual parts.

               

              I have seen maximum stresses all over the place not just close to the discontinuity. I am not comfortable just probing the model cave man style to try and hunt and peck to find the max stress. With NASTRAN, you can output a results file that has tabulated stresses for every node and/or element. You can export the results into a spreadsheet and search max stress. Nodes and elements are identified so it is relatively easy to pinpoint the area of your model where the max stress is located. Since Simulation has to calculate stress from node to node, the maximum will occur at a node, and should be able to be shown depending on whether the average between nodes is checked. 

               

              As far as hand calcs, I always verify my model and stresses. Hand calcs are conservative and you have to make assumptions. My particular problem has multiple load paths and to calculate the stress/ bolt loads, you have to assume what the bypass stress(es) will be.

               

              The main question I had occurs no matter what mesh size you use, no matter how good your model is, no matter how many hand calcs you do. I could go to the trouble of making separate solids for each hole and use glue contact to hold them in place. I could then specify which solids to show the stress on each stress plot. But there has to be a better way to show the stress on parts outside of discontinuities.