17 Replies Latest reply on Aug 17, 2012 11:46 AM by Chris Bargman

    Odd simulation Results...

    Chris Bargman

      What could be causing the below results?  Why is it not coming in as a smooth transition but instead this spotted thing...

       

      Could this have somethign the do with the mesh?  It is a fairly fine mesh.  Nothing crazy on the boundry conditions.  Elastic in a couple spots nothing is fixed.

       

      I should mention that yes there is a piece of steel that sits on top of that pressing down on it.  Not shown in the picture.  But that is why you see stress there to begin with.

      untitled.bmp

        • Re: Odd simulation Results...
          Anthony Botting

          HI Chris: I cannot read the legend range on the image, but I have seen this in some cases where the color spectrum is represented over a (seemingly) small "range". For example, I once had a result with near-zero values, but the contour plot defaulted to representing the entire color spectrum over what I interpreted as numerical noise in the range of zero, resulting in the spotty appearance. The legend range was something like -1.50001E-08 to +1.50002E-08.  I am wondering if the legend on your plot has a narrow range of values such that the stress is fairly constant, but the contour plot is representing each color for very small changes in the stress. Could the range be from 500.00 to 500.01 (or something like that)?

            • Re: Odd simulation Results...
              Chris Bargman

              Good guess.  I could entirley see how that would cause something like that.  Unfortinatly... No.  The range is set from 0.0 to 600.0...

                • Re: Odd simulation Results...
                  Anthony Botting

                  Ok. I am concerned maybe the contacts from the steel piece sitting on top, are not altogether working correctly. I might try to remove or suppress the steel piece, and substitute a constant displacement fixture on that surface where contact would be, to push by some amount, if only to see if the elements can represent a more uniform gradient of stress. This might be a 'process of elimination' to see if the contacts are not working correctly.

                    • Re: Odd simulation Results...
                      Chris Bargman

                      Good thought.  I am still pretty new to this.  I tried taking the insert out of the simulation and applied the same force down onto the bottom of the cavity.  The results were just as you had expected.  A much more uniform gradient stress as it should be.  So now I know its something being caused by the geomitry... But what about the geomitry could cause that?

                        • Re: Odd simulation Results...
                          Anthony Botting

                          It would seem the contact algorithm is suspect (assuming you see a uniform stress field without the component). Since the component starts-out in-contact with this plate, and there's no real slipping going on,  you could try the advanced contact option named "node-to-node" - then remesh the components and attempt another run. You would need to edit-definition on the contact set(s), and there is supposed to be an "Advanced" button in the PropertyManager that will let you change the contact set option to "node-to-node". It might make a difference in the stress field.

                            • Re: Odd simulation Results...
                              Chris Bargman

                              Gave that a try and results seem to be the same.  I have a meeting setup with a solidworks rep tommrow morning to try to solve this along with a couple other issues I have been having.  Will update when I know more.

                              • Re: Odd simulation Results...
                                Chris Bargman

                                Okay,  Well I met with the solidworks rep this morning and he was a little stumped by this one.  But what we did discover is that I had NOT tried the "Node-to-node" You were referencing.  What I had tried was the option in the mesh "at nodes"  Reason being I do not have an advanced flyout in my menu.  No idea why and the solidworks rep had no idea either.  But maybe if I can find that option I can give it a try...

                                untitled.bmp

                                  • Re: Odd simulation Results...
                                    Anthony Botting

                                    Strange. Here is what my dialog looks like. I am running SW 2012. Did the rep get back to you on this? It should be interesting to see what is the deal, here. I also read in the "help" file that the Advanced option is supposed to default to node-to-node when the faces are initially touching. So, yours might do that anyway. I hope the rep was able to get a copy of your model and get it to tech support. This particular issue is very interesting and I hope you can share what they find out. Thank you.contactdialog.png

                                      • Re: Odd simulation Results...
                                        Chris Bargman

                                        Well I did send him my file.  I will keep this updated when and if I find a solution/ cause. 

                                         

                                        As for the advanced flyout.  I did some searching after the rep confirmed there is indeed supposed to be another flyout that I am not seeing.  Apparintly the Advanced flyout is actully an option that can be turned on/off... 

                                         

                                        Go to Simulation > Options > Default Options > Mesh > The second from the bottom checkbox is what your looking for.

                                         

                                        I have not had a chance to mess around with these settings now that I know where they are.  But as I said I will keep this going as I find out more. 

                                         

                                        A-0017.jpg

                                          • Re: Odd simulation Results...
                                            Anthony Botting

                                            Thanks, Chris. I did not know that was there! Thanks for the update. -Tony

                                              • Re: Odd simulation Results...
                                                Chris Bargman

                                                Update, 

                                                 

                                                Well I have not heard back from the solidworks guy.  He really was not very helpfull.  This forum has dont a thousand times more good than calling the company that manufactures the software.  Crazy me for thinking that they might know a thing or two about it... Ugh.  Anyway on to the important stuff.

                                                 

                                                The Results:

                                                I went ahead and tried a few more things with no change in results Finally just this morning I went ahead and just remade the geomitry.  The geomitry I recieved origionally was full of un-needed cuts and scar marks that must have been causing Solidworks alot of issues...  With the new geomitry the simulation ran in about 1/16 the time and gave results more how it would be expected.  A nice smooooth gradiant no spots nothing out of the oridinary. 

                                                 

                                                I guess if you want something done right you just have to do it yourself and cant trust others...  Let this be a note to all on just how picky Solidworks can be when it comes to geomitry.  Before asking questinons make sure your geomitry checks out!

                                                 

                                                Thanks for all the help guys I will be around

                                                  • Re: Odd simulation Results...
                                                    Anthony Botting

                                                    Thanks, Chris.

                                                    That is good to know. I have experienced many models imported to SolidWorks from other CAD programs, that are not "clean". Was your original geometry imported? I have seen some imported with open edges and splits, slivers of geometry, and small holes that did not make the model "watertight"(according to SW) . As you mentioned, the software seems to be very picky, which can be a good thing. I have also witnessed that it is very good at repairing geometry and you can even remove surfaces and replace them. Thank you for the update.

                                • Re: Odd simulation Results...
                                  Nathan Gustafson

                                  Ive had it before too where Solidworks will try to animate it to show the force pushing down on the bottom piece, this would cause the red pressure zones to spread as the load is applied. Also, depending on what you are looking for (I am assuming stress) make sure you are looking for stress in the direction you want. Solidworks will default to VonMisses Stress. You can change that by looking under your results tab in the feature tree, right clicking and selecting "edit definition". I hope this helps!

                                  • Re: Odd simulation Results...
                                    Jarred Mosdal

                                    I just attended the Simulation class.  There is an exercise in the book where we analyzed stress through two perpendicular faces.  As the mesh size was made finer and finer, the stress concentration increased.  I'm new to this too, but from what I learned there, you may have a singularity on the edge transition.  Add a small fillet and run it again.