3 Replies Latest reply on Apr 19, 2013 12:01 AM by Jared Conway

    Mesh Failure Diagnostic screen shows no info after mesh fails

    Joseph Sizemore

      I'm encountering a new (to me) error:  I've got a large assembly I'm preparing to run a frequency study on, and when I try to create a mesh the mesh process ends very quickly and I get the window where I can press Failure Diagnostics to open the Mesh Failure Diagnostics window.  When I open the mesh failure diagnostic window there are no parts listed in error. 

       

      When this first started happening, I simply turned all my mesh controls off and tried to start meshing from scratch.  This allowed me to see the 14 parts that failed to mesh.  So, I started applying mesh controls and/or closing gaps between these parts (all of which have been recently bonded to other parts in lieu of springs) one-at-a-time.  At this point, when I remesh I go back to the same problem as before: no parts listed in the mesh failure diagnostics box when I open it, so I have no idea if my changes are helping or not.  I realize there's probably some gaps to close, but what benefit is the diagnostic helper feature if it doesn't provide you any details on what parts are failing?  My only method at the moment is to suppress all my mesh controls and remesh from scratch to see a list of problem parts. 

       

      Has anyone encountered this problem?  (Empty mesh diagnostics list)

       

      I've been advised not to use incompatible mesh, as doing so causes a loss in accuracy with respect to stress results (which is what I'm ultimately interested in).  Is that true?

        • Re: Mesh Failure Diagnostic screen shows no info after mesh fails
          Jared Conway

          if the mesh process finishes quickly and there is no error, it doesn't surprise me there is nothing in the mesh failure diagnostics. is it possible that you're interpreting a failed mesh workflow rather than actually running into one? IE, are there parts missing from the graphics window that aren't meshed?

           

          incompatible mesh loss of acrruacy with respect to stress. where is your reference for that? have you looked at how incompatible mesh works? it could cause some issues at the surface because of the rigid beams but if your mesh is similar it would probably be ok. and if you want stress, frequency analysis isn't going to help you. frequency analysis is for modal shape and mode frequencies only.

           

          some screenshots of what you're seeing, what version you're using and the model would help here.

            • Re: Mesh Failure Diagnostic screen shows no info after mesh fails
              Joseph Sizemore

              When I do make a new mesh without mesh controls I find some number of parts that failed.  When I use the diagnostic tool to use a refinement, when I remesh using my new refinement(s), the mesher freezes and I have to kill the program.

               

              Also, I closed up some tiny gaps that were present with some previous component contacts, and when I do a fresh mesh w/o controls, the parts I corrected work, but now a new series of failed parts appears that was not present for the first mesh. 

               

              Sometimes the mesh displays with missing parts.  What does this mean?  Also, I tried to use the SW2013 ability to mesh individual parts, thinking that would help me, but when I selected one of my failed parts to mesh, I noticed it automatically added contact constraints to all the surfaces it touched.  Should I be using this method over the component contacts I've used?

                • Re: Mesh Failure Diagnostic screen shows no info after mesh fails
                  Jared Conway

                  When I do make a new mesh without mesh controls I find some number of parts that failed. When I use the diagnostic tool to use a refinement, when I remesh using my new refinement(s), the mesher freezes and I have to kill the program.

                   

                  >i'm not sure what you mean here. is there a question or is this just a statement? it sounds like to me that you're adding a lot of elements and you need to wait. if the process isn't at 0% CPU and memory isn't still moving, it is still working and you need to let it do its thing. as we've mentioned in a couple of posts, make sure you're using the appropriate type of elements.

                   

                  Also, I closed up some tiny gaps that were present with some previous component contacts, and when I do a fresh mesh w/o controls, the parts I corrected work, but now a new series of failed parts appears that was not present for the first mesh.

                   

                  >this makes sense. this is a good place to use incompatible mesh. the problem is that it can mesh the parts but when it has to match the elements on both parts, it can't and it fails at the interface. your only option here is to make sure your geometry is compatible. ie you don't have slivers or small overlaps that make it hard to match the elements and that the same element size works for both parts. you can test this theory by meshing the parst on their own with the same size. if it works at the part level but not at the assembly, it is the interface that is the problem. incompatible is your savior here.

                   

                  Sometimes the mesh displays with missing parts. What does this mean? Also, I tried to use the SW2013 ability to mesh individual parts, thinking that would help me, but when I selected one of my failed parts to mesh, I noticed it automatically added contact constraints to all the surfaces it touched. Should I be using this method over the component contacts I've used?

                   

                  >displaying mesh with missing parts means you have parts that failed. it meshed some but not all.

                   

                  the incremental mesh tool has to add a contact set between components because it needs to force incompatible mesh to allow a mix of standard and curvature mesh or in general to let you have 2 element sizes on 2 different parts, see explanation above.