8 Replies Latest reply on Aug 18, 2014 11:09 PM by Jared Conway

    Number of Elements Isn't Consistent, and Minimum Volume Error

    Matt Mullett

      I have a symmetric assembly of a drum dryer, and I'm trying to analyze a bolted connection. Yesterday I started with a copy of a used simulation study, added my bolt connectors and new constraints, and tried to run it. I kept getting the "Element Volume is smaller than limit for element: xxxxx". So I went to track down that element to see what the problem is, and the element numbers on the probe screen stop well below what the error screen gives me, or even the quantity of elements indicated on the solver screen. Solver screen says theres about 470,000 elements, and the probe tool stops around 290,000. Today I finally bit the bullet, started a new study, and redid all 60 bolt connectors (which is a whole other gripe), and my first run solved just fine. I fixed some bonding issues, and re-ran the simulation and ran into the exact same error and mismatch. I'm really starting to worry that I'll have to re-do 60 bolts by hand every time I want to re-run this study. I'd love some troubleshooting help with this.

        • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
          Jared Conway

          could you use a 1,2,3 sequence list to describe the workflow? i'm not sure i follow.

           

          the error basically is telling you that there are bad elements

            • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
              Jared Conway

              i would look at whether your geometry is good..etc.

              • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
                Matt Mullett

                The assembly is a horizontal dryer drum, over 18'going the old assembly to a new directory and name, I replaced the single piece track that the whole dryer drum rotates on, with an assembly consisting of two identical half rings that bolt together. I have successfully meshed and solved the two piece track assembly by itself. I set up the bolted connections between the track and the other assemblies, and then added contact constraints between the track sections, and the assemblies they bolt to. When I tried running this, I immediately got the volume warning. 

                 

                I tried refining the mesh around the bolted track section, and the global mesh because I was running it coarse for the first run, and nothing I did would let it solve. About this time I noticed some weird things happening with the assembly as far as parts being suppressed and hidden when I re-opened my saved file. So I threw away that assembly and started over yesterday. I rebuilt the assembly again, and this time I actually got it to solve once. I fixed some bonding issues, and tried to re-run it and I immediately got the volume error again. This time I suppressed everything in the study I had added; the bolts, contacts, etc and simply fixed a few faces and still got the error no matter what I did.

                 

                As far as the element numbers go, I tried hunting down the bad elements, but I can never locate the element number in the probe tool screen. In my first post, you see the element #403848 called out as the bad element, with the solver saying that there are 432,753 elements in the study. However, when I use the probe tool, and select the main assembly and list all, the element numbers stop around 30-40% below the qty of elements that the solver says I should have, so I can never track down the culprit. That's been consistent across both assemblies.

                small drum.PNG

                  • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
                    Nicholas Luyster

                    Hi Matt,

                     

                    Are you using solid Elements all the way through?  It's hard to tell from the picture but it looks as though there may be some pretty thin sections on the drum in comparison to the over all size.  It may pay off to have shells at some locations if possible.  Also, are you only running into this after you setup your bonds?  What happens when you switch from compatible to incompatible bonding?  Does this error happen during the mesh or the run process?  

                     

                    All the best,

                     

                    Nick

                    SolidWorks Simulation Training

                      • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
                        Matt Mullett

                        Hi Nicholas, I've ended up solving the problem. It had to do with the virtual walls I set up to constrain the model. I had work planes defined tangent to the track surface that I used for virtual walls, and apparently they were too close and technically interfering with the mesh. But it did only start happening after I defined contacts in my model, so I wonder if those contacts changed the way it creates the mesh and causes the interfering elements. 

                         

                        As for the shells, do you have luck using them in mixed mesh cases? I've tried using shells and sheet metal shells in a few projects, and it just leads to a huge headache. It takes 5-10 times as long to mesh, because it sits there doing nothing for 5-30 minutes (depending on how many shell parts I have) after I press the mesh button before the mesh window even pops up. So if I have any trouble getting it to mesh I have to try several times which takes forever. And then the results are never as good as solids are anyway. I try to avoid them as much as I can, but maybe I'm doing something wrong.

                      • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
                        Jared Conway

                        still not following the workflow, it sounds like it meshes once, and then you add some new features, you get an error, you eliminate those features and you still have an error?

                         

                        if that is the case, either something is changed that you aren't reversing or something wrong.

                         

                        have you contacted your reseller on this?

                         

                        but the way you describe it, it sounds like this model needs some love and some idealization. this would be things like deleting interferences, suppressing small features..etc. not only will it make your life easier but the problem will also solve way faster.

                    • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
                      Matt Mullett

                      I had already performed model prep on the assembly, and like I said, the same assembly ran several times before I encountered this error. But I think I finally narrowed down the problem. It was made harder by the fact that the problem wasn't 100% consistent and seemed to run fine once, but fail the next time without changing anything. The problem had to do with the work planes/virtual wall I created to constrain the model. The virtual walls replaced the wheels, so the whole drum would sit in a valley with the track resting on the virtual wall. Well apparently if the work plane isn't created in just the right position, it ends up interfering with the mesh, and some of the elements penetrate the wall. I had defined the work planes to be tangent to the track, but that apparently was too close for Simulation. To fix it I created new planes offset from the track by .001", and made the wall ignore gaps .001" in size. The confusing thing is that I have run studies almost identical to this in the past with no problem.

                        • Re: Number of Elements Isn't Consistent, and Minimum Volume Error
                          Jared Conway

                          failing the second time without changing anything is a software issue so i suspect you were changing something or something was changing, like you mentioned, maybe the mesh.

                           

                          but virtual walls should be coincident with the parts that will come in contact with them.

                           

                          i still suspect some work on the model will make things a lot better for you. that includes simplification and idealization. if you're looking for someone to help look through your issues and help come up with a plan of attack for your model, don't hesitate me contact us at jared@hawkridgesys.com.