5 Replies Latest reply on May 9, 2014 10:55 PM by Jared Conway

    Mesh capturing

    Alexander Baker

      Hi all, long time user, new member;

       

      I've been having some trouble running a simulation of a high lift rear wing I'm designing. Sometimes during the solving process (usually just after travel 3 has completed), Flow will try to capture the mesh but for some reason just hangs. The bar at the bottom of the screen will just run from 0-100 very rapidly and nothing happens. I've tried leaving it overnight to find it still there the next morning.

       

      Quick look on here has given some things to try but nothing seems to work; cloning the project didnt work, restarting/clearing  solidworks data/previous results doesn't. I have run a simulation of a two element rear wing which was hit and miss (some would complete, some wouldnt). Just wondering if its something I'm doing...

       

      I've attached the files if someone is wanting to give it a go themselves.

       

      All help appreciated,

       

      John

        • Re: Mesh capturing
          Jared Conway

          i'm confused

          after 3 travels it says "capturing mesh"

          and then goes from 0-100?

           

          can you post some screenshots?

           

          info about version/service pack?

           

          how is this model different than others that worked fine?

           

          where are you storing the files and results?

           

          how many cells is the problem?

            • Re: Mesh capturing
              Jared Conway

              looked at your file

              i can almost guarantee it is getting hung up because you have level 8 mesh and calc control

              when you get to travel 3, you're creating a billion elements causing it to look like things are crashing

              you could check the ram and cpu when this is happening, if it is still climbing, that is likely the case

              this problem should really be setup in 2d first

                • Re: Mesh capturing
                  Jared Conway

                  also would recommend looking at other posts about wings and aero in flow and how big the comp domain needs to be and how good the mesh needs to be to get good accurate results

                    • Re: Mesh capturing
                      Alexander Baker

                      Thanks for all the pointers Jared,

                       

                      Unfortunately I don't have a screenshot and I can't get it to repeat right at this minute but I'll try and explain the problem again. In the solver window, the info box, I've noticed that when I get this error the travel count has just past an integer number of travels (usually something like 3.01...) at which point the progress bar at the bottom of the solver window says "mesh capturing". There is a dark grey bar which does not move (i assume this is overall progress) and then a lighter grey bar oveylayed ontop which moves from the left to right as if it were still solving but nothing happens. During this time the CPU and RAM usage is similar to that of system idle, roughly 1% cpu usage and 2gb ram usage so I know the solver is no longer working.

                       

                      The solidworks version is the 2013 academic suite which is running on a custom built pc within the university (windows 7, unknown rev no. 128gb RAM, 16 physical cores, 16 virtual cores (32 total). The results and files are all running off the local hard drive (C:).

                       

                      The cell count is usually around 100k at the point when it hangs. I usually leave the solver to just run the meshing on the auto settings (I.e. no local mesh) so I will give this a go.

                       

                      The variation between models is essentially part of a Design of Experiements I'm running which is looking to control the gap between the wing elements and the angle of attack (measured from the horizontal). There doesnt seem to be a pattern as to which ones work and which don't.

                       

                      As said before, when the solver "hangs" the cpu usage is already down to 1% and ram is also very low usage.

                       

                      I understand that I should reduce the computation area, but right now I'm interested in the resultant freestream created by the wing, I will try reducing this volume and the meshing level to see if this helps.

                        • Re: Mesh capturing
                          Jared Conway

                          Comp domain should be bigger

                           

                          But to do that effectively, go down to 2d analysis and ignore edge effects

                           

                          the problem is almost for sure related to the solution adaptation and the generation of cells

                           

                          2d will make this easier on you and your computer