16 Replies Latest reply on Sep 15, 2017 9:43 AM by Becky Michalski

    SAE Frame mesh errors

    Becky Michalski

      I am working on an SAE frame and am trying to create a mesh but it is sending errors because of the bends in some of the tubes being too small. The exact error that comes up is that the body (the curve) is too small to be considered a body. The slenderness ratio of the body is less that 10. These arcs have to be this small to comply to my manufacturing rules but I haven't figured out how to create the mesh. I have increased the joint sizes to try and get the software to ignore the curve and have also played with the coarse/fine settings in the mesh. Any help would be appreciated.

      Thanks Becky

        • Re: SAE Frame mesh errors
          Keith Frankie

          Can you post a screenshot, or even better the file itself (or a portion of the file, or a simplified weldment with similar geometry)?

           

          I'd guess the joint definition is the problem.

           

          The typical issue "The body is too short to be considered as a beam.  The slenderness ration of the beam is less than 10." is just a warning.  The study should still mesh and run (but of course your results would be suspect in the affected areas).

           

            • Re: SAE Frame mesh errors
              Becky Michalski

              Here is the file I have been working with. I have done some simplified versions of it to try and find the problem and it will mesh so long as none of the arcs are present.

                • Re: SAE Frame mesh errors
                  Matt Peneguy

                  Right off the bat, I found something that may be causing some problems:

                  That pipe is not properly trimmed.  That will definitely cause some problems.

                    • Re: SAE Frame mesh errors
                      Matt Peneguy

                      Something you can do is choose multiple groups in each of your weldments. If your frame is made from all the same pipe size it could even be just one weldment.  it is much easier to work with weldments this way than adding all of those Trim/Extends.

                      This does a couple of things for you.  First off it gives you control over the end conditions at each location:

                      By changing the Group and Trim Order you can change the end conditions:

                      Another thing it does for you is it makes your feature tree much smaller.  I'm sure there are other advantages. 

                      Now, a question I've got that maybe somebody smarter than me can answer, would it be advantageous to have separate weldments for each section of the frame?  For instance, I show above the bottom of the frame.  Maybe you'd just want that as one weldment because you want the bottom welded together first.  Would there be an advantage to having it that way in the feature tree versus the entire frame as a single weldment?

                        • Re: SAE Frame mesh errors
                          Becky Michalski

                          Thanks for the suggestion Matt. I tried what you suggested on a small section of the lower rail just to test it out and i got farther than I had before (It actually meshed) but it turned the bar into 2 parts no matter what plot (FOS, deformation, etc). Ive done other sample parts and had the same problem.

                        • Re: SAE Frame mesh errors
                          Bill McEachern

                          I find the beam mesh works best with no trimming. Triming just causes problems as the ends of the beam do not terminate at the same location.

                            • Re: SAE Frame mesh errors
                              Matt Peneguy

                              Thanks for clarifying that.  I haven't actually run a weldment through a FEA simulation.

                              How do you handle weldments that you want to run through FEA?  I don't see a way to disable the joints in weldments.  Do you have to do a path and profile sweep for every piece in that frame Becky Michalski posted?  That seems rather labor intensive if that is the case.

                                • Re: SAE Frame mesh errors
                                  Bill McEachern

                                  Just don't trim them when you make them - turn it off, exactly how I can't recall but it in there. Trim them after you have it confirmed via FEA.

                                    • Re: SAE Frame mesh errors
                                      Matt Peneguy

                                      I found it.  I was looking to turn it on the box that pops up when you select a joint.

                                      • Re: SAE Frame mesh errors
                                        Matt Peneguy

                                        Looks like Becky Michalski may have to create a simplified sketch that eliminates short pieces such as this one:

                                        Is that correct Bill McEachern?

                                          • Re: SAE Frame mesh errors
                                            Keith Frankie

                                            In this model you can use the 'filter' at the top of the feature tree to search for all the trim commands, then shift-select all of them and suppress them.

                                             

                                            What you're looking for is to get the CAD geometry and joint calculation settings correct, such that there is one and only one purple 'joint' ball at each actual joint, and there are no yellow 'endpoint' balls where there should actually be joints.

                                             

                                            I'd roll back in the tree till there were just one or two beams, and try to mesh that.  Then roll forward slowly, adding features.  If it meshes, great, if not stop and fix it.

                                             

                                            Incremental work is essential for FEA - I've seen it called the 'crawl walk run' method.

                                             

                                            Using this method I can see the very first beam, pipe065(7), has trouble. 

                                             

                                            Indeed, Matt, it seems like those short sections of curved beam are having trouble.  They do seem like an important part of a weldment, though, that hopefully SW could handle.

                                              • Re: SAE Frame mesh errors
                                                Matt Peneguy

                                                Keith Frankie,

                                                You bring up a great point that new users, (Such as I assume Becky Michalski, may be) may not be aware of.  The roll back bar is a very useful, underutilized tool, not only for FEA, but for many other things.

                                                Becky, one of the things that you should take away from this thread that will help you going forward in SWX, is to add the rollback bar to your bag of tricks.  If it's not already in that bag.

                                              • Re: SAE Frame mesh errors
                                                Bill McEachern

                                                A short beam is not really a problem unless it is not oriented the right way - on imported parts with no underlying sketch if it really short it can get confused on the beam axis direction. For the case shown though I would, firstly get rid of the radius that drives the curved bit  -it looks like it will get buried in the mass of weld at the joint anyway. Otherwise I would play with the search radius and see if the other node can come up so there are ones at both ends of the curved section. Unless the curve is big I would not include them. If it is in the order of the tube diameter I would ignore them and make them straight. This is a beam analysis - it for picking members -not for analyzing joints.

                                                  • Re: SAE Frame mesh errors
                                                    Matt Peneguy

                                                    That was what I was thinking.  But, I'm not an FEA expert.  My thinking would go along the lines (no pun intended) of creating a separate 3D sketch that eliminates all of these bends:

                                                    It appears to me that these pieces would be bent, but the bends being eliminated would only make your FEA output more conservative.  Correct?

                                                      • Re: SAE Frame mesh errors
                                                        Bill McEachern

                                                        I would not say "conservative" - that would require some more thought about exactly what's going one. Everything is an approximation including an FEA analysis. Having an objective for the analysis is key. What are you trying to figure out? And, for a beam analysis you are trying to figure out if you have members or sections that are going to work. It isn't about the joints as then don't possess enough detail to sort that out in a beam analysis. The curves in the beams that are by  themselves, at the top in the image, should not be an issue but the ones at multiple member intersections they are not helping.