4 Replies Latest reply on May 4, 2013 12:17 PM by Jared Conway

    SolidWorks crashing during flow simulation

    Calum Paterson



      I am running an external analysis on our Formula Student car for a thesis project, trying to develop an undertray diffuser. I have run many simulations (mostly) without issue, but there is one model which I cannot get to run. When I begin meshing, Solidworks closes.


      All of the models have identical mesh, computational domain and boundary conditions (a moving ground is set up to simulate real life), and as far as I can tell I have set up all other parameters the same.


      I have managed to get simulations to run on this model a few weeks ago, then I changed the speed of the fluid flow and can't get it to work, despite having done the same thing for the two other models I have.


      Does anyone have any ideas? I don't think its a hardware/memory issue seeing as I can run other sims just fine.





        • Re: SolidWorks crashing during flow simulation
          David Woodruff

          Calum, hard to say without seeing the files. However, I have seen this when working with multiple configurations. After cloning a project you start a run and SWX will just crash instantly! I always mesh only first and still see this happen. (2009, 2010, 2012 & 2013) What I do is clone the project, save it, close the file and SWX, reopen and then the file will run. Sometimes I have to shut down and restart. Hope this helps! Woody

          • Re: SolidWorks crashing during flow simulation
            Jared Conway

            First thing to do, clear the results directory with swx closed. See if that helps. Usually a good idea to reboot before clearing.


            If you change the flow speed back, does it work? If not, I'd setup a brand new study.


            If it works, I'd replace your fsae car with a block and see if it runs. If it does, I'd guess you have some bad geom causing the issue at the higher speed. Basically the solver is crashing out on run when it tries to mesh. Checking mesh only like Dave said might confirm this.

              • Re: SolidWorks crashing during flow simulation
                Calum Paterson

                I set up a new study which seemed to solve the problem, (although I input all of the same parameters which I had before), thanks!


                I am now having a strange issue where simulaltions are producing really low downforce values, roughly a quarter of the values produced before. I have re run exactly the same models as previously. Increasing the flow speed from 9m/s to 27m/s only leads to an increase in downforce of 40N, much lower than I was expecting, and much lower than I had previously recorded. Before, an increase in speed from 15m/s to 27m/s lead to an increase in downforce from 25N to 110N, much more close to the values I had expected.


                I am measuring downforce using surface goals across all surfaces of the car, measuring force in Y direction. This is the same method I have always used so I have no idea why the downforce values are now so different.


                Any ideas?

                  • Re: SolidWorks crashing during flow simulation
                    Jared Conway

                    When you say recorded before, do you mean physical data or do you mean from a previous sim? If you have data, then you should be double checking your sim matches the real testing conditions exactly.


                    Regarding expected valued, I'm going to go out on a whim and say this is based on hand calcs. Have you setup that exact problem in simualtion to vet your setup and flow simualtion? That would be a good first start.


                    If those both check out, it is likely the issue is the way you are reading things. In flow, force is an integral value. Highly dependent on the underlying geometry. First, make sure it is a goal. Second, make sure every external face is selected in that goal. Don't leave out faces and don't do individual elements and add them together. Both will result in bad results. For an fsae car, this will be hard if you are working with the full model. I would recommend essentialy building an envelope part of the basic shape to use for this. That will make selection easier.


                    If you still have problems, make sure you have a good mesh, consider solution adaptive, and make sure your simulation has converged. If it is, and all the above tests pass, I would guess your expectations may be too high for the design. Keep in mind, lift and drag are hard calcs. Flow will get you close and help you understand trends. You'll still need a real test to validate and tune your setup and analysis.