7 Replies Latest reply on Nov 12, 2017 8:31 AM by Bill McEachern

    bonding face to face

    Sigma Tero

      I am trying to bond these four surfaces together so that they act as one body.  This is a simplified example for a much larger assy that I am working on and I want to verify results first.   I need the larger assy to have components bonded together via mating faces rather than just edges.

       

      surface bonding test_1.png

       

      Hand calc from the results indicates that they are acting like a leaf spring- not bonded.  My Contact Set is between faces of the surfaces (not the edges).  I've tried all sorts of combos.  Global bonding via Component Contact, manual bonding via Contact Set, non-touching faces option selected,  etc.

       

      How do I bond faces together so that they act as one?

        • Re: bonding face to face
          Seckin Uslu

          Surface modelling may not be a solution. use solid elements .

          It may takes much times

          • Re: bonding face to face
            Sigma Tero

            Solids take far too long.  Why can't this do it with surface to surface faces? The manual seems to indicate it's possible.

            • Re: bonding face to face
              Seçkin Uslu

              I have created a similar study . You should check definition of thickness.

                • Re: bonding face to face
                  Sigma Tero

                  Thanks Seckin but this replicates what I'm seeing.  The layers aren't acting as if they're bonding. 

                   

                  I never trust FEA without validation with testing or hand calcs. 

                   

                  In this case you have a beam 50mm thick x 50mm wide x 500mm long.  Peak stress based on classical hand calc methods of this rigidly supported cantilever beam should be 24.0MPa and deflection 0.381mm.  Your FEA gives a peak stress of 64.8MPa and a deflection of 3.78mm.  Plus from the picture you can see that the stress is roughly equal throughout each layer. It shouldn't be.

                   

                  mesh_1.png

                • Re: bonding face to face
                  Sigma Tero

                  Thanks for your help but my point was that it reinforced my belief- SW doesn't seem able to join two faces of surfaces to make them act as one.

                    • Re: bonding face to face
                      Bill McEachern

                      I have found that in SWX sim you need to be really careful when bonding shell faces or any other contact definition. You could, if suitable,  always use a laminated shell.

                      In abaqus you would not be allowed to do this either at it would lead to an over constraint condition with "tie" constraints (the equivalent of "bonded in SWX). Abaqus has other options to handle this situation (continuum shells) which are built for this situation.

                      However, SWX is a lot less strict on these over constraint checks as it is not an inherently NL code. To see if this will work you need to make sure that all the shells have their surface normals pointed in the same direction. Further, the distance between them should be less that the shell thickness. It may still not work but I have had some success in the past.