6 Replies Latest reply on Sep 17, 2010 4:27 PM by Paul Jeffers

    shell model failure?

    Paul Jeffers

      I have modeled up a steel structure - as surfaces, started a static simulation with shell elements and basic fixed boundaries, simple applied load - it meshes fine and then refuses to run without any feedback of why it fails - it fails at the very start - does anyone know if there is another way to capture / discover failure information from the solver? Thanks

        • Re: shell model failure?
          Alin Vargatu

          Did you restrain it properly? Can you post the fileset here?

            • Re: shell model failure?
              Paul Jeffers

              yeah tried loads of differnet ways of restraining to check that it was not a bug on edge versus face restraint and initial used fully fixed.

               

              Attached is the file - i zipped the lot of files including the .gen so all the info is present. Any idea appreciatted.

                • Re: shell model failure?
                  Ryan Werner

                  Just wondering, which solver are you trying to use.  I have had this same problem using the Iterative Solver.  If you switch to the Direct Sparse it should run.  Also, if you restart SWX and reopen the file sometimes this will make the Iterative Solver work.  No idea why.

                   

                  Ryan W. 

                    • Re: shell model failure?
                      Paul Jeffers

                      Hey Ryan, I had twigged to the close it and open again and then sometimes it works on other models but this didn't like it. Just tried the other solver and ....worked fine! no idea why - thanks. Now the real question - do I trust the results!

                        • Re: shell model failure?
                          Ryan Werner

                          Well I definitely wouldn't worry about which solver you had to use to solve it.  There is a good chance that the Iterative Solver will go the way of the dodo here soon as computers get more and more RAM and computing power.  The Direct Sparse is acutally the more accurate solver it is just that in the past with less resources available to the average user it was a lot faster to use the Iterative.  Since it is a "steel structure" you are analyzing have you thought of trying the beam elements.  To do this most easily your model will need to be a weldment.  Just wondering because you can if possible you can run it using beams and then compare you results.

                           

                          Ryan W.

                            • Re: shell model failure?
                              Paul Jeffers

                              the structure is best described as a cone with a steel 1/4" plate skin and then standard columns used as ribs with the skin attached along the neutral axis of the column webs. I looked at using beams but could not see how to attach a shell a beam long it's length. The beam feature seems pretty good for space frame type structures but no so great for the monocoque variety and this is a hybrid. I will probably end up using some simple hand calcs adapated form Roarks to make sure I am in the right area. Thanks again.