3 Replies Latest reply on Oct 7, 2015 10:12 AM by Radoslaw Szmid

    Error when trying to run COOL in Plastics

    Brandon Baker

      Hello!

       

      I was wondering if anyone has run into this problem in Plastics.

       

      I have a less than ideal part that I have received form a customer which includes widely varying wall thicknesses and knife edges. When I run the simulation using a virtual mold and cooling channels, it runs just fine and gives me fairly decent results. However, since the mold design involves multiple materials (copper mold inserts for heat dissipation) I had hoped to run the simulation with the mold included to get a bit more accurate results.

       

      This is where the problem occurs, when I run FLOW, PACK, and WARP I get no errors but COOL stops with this message:

       

      Melt Part:

      !! ERROR !! Cell 12, Volume = -0.464234 < 0

      ievtxs = 6 iFaces = 5

       

      To make matters more complicated, the part I am using involves imported geometry, so it is difficult to suppress features to find out where the specific error lies.

       

      1. So, why does it run fine with a virtual mold but break with the full mold?

      2. Is there anything I can do to improve the full mold meshing? (No matter what I do it shows a high number of "unmatched" elements, can this be the issue?)

      3. How can the other processes finish if COOL fails?

       

      Thanks!

        • Re: Error when trying to run COOL in Plastics
          Radoslaw Szmid

          Hello,

           

          Have you tried to follow the steps I've previously described at the end of this post?

          Technical Reference to SOLIDWORKS Plastics

           

          As for the error you're showing, I think your geometry contain a highly distorted element with a negative volume.

          There's an option insight Plastics where you could check the mesh quality and even isolate the worst elements in your geometry to better investigate the problematic area. You may then try to repair mesh by merging/moving nodes within Plastics itself or as I would suggest, clean up your CAD files as much as you can before even trying to mesh them...(delete small fillets, unnecessary features, logos etc). That takes time but preparing geometry for meshing is a must in every FEA project.

           

          1. So, why does it run fine with a virtual mold but break with the full mold?

          2. Is there anything I can do to improve the full mold meshing? (No matter what I do it shows a high number of "unmatched" elements, can this be the issue?)

          - please analyse all the steps I described in my other post. Geometry intersecting is critical in case of physically touching faces...

          Yes, unmatched elements are the main issue here. Your mesh must not contain any of them.

          3. How can the other processes finish if COOL fails?

          - If you first run a COOL analysis then the following FLOW-PACK analysis could accept (read) a mold temperature profile “filename.WL?” from the previous COOL analysis as mold temperature during FLOW-PACK simulation.

          If such results are not available, the FLOW-PACK analysis will still run base on the uniform mold temperature that you normally set in Fill settings.

           

          Regards

          Radek

            • Re: Error when trying to run COOL in Plastics
              Brandon Baker

              Actually, I have been trying to follow the steps you described before. However, I'm running into an issue with step 3. The copper inserts used in my model have somewhat complex geometry associated with them, and I don't know how to get the split line intersection to work. I've tried other tools such as the "Intersect" tool but that didn't seem to help.

               

              I believe the major problem now is simply the poor meshing between the touching faces of the inserts and the part. Since I begin with the copper parts set to "Insert Domain" the meshing behaves far more poorly than when I've had them set as "Mold Domains". I think the key is getting the split line working, but for this geometry it doesn't seem to want to. I don't have much experience using the split line intersection tool though, so I might just be doing it wrong.

               

              Below is an image of a portion of the part with the copper cooling insert shown in transparency. I'm using a multibody part with both of these bodies in line to line contact with one another.

               

              Copper Insert.JPG

                • Re: Error when trying to run COOL in Plastics
                  Radoslaw Szmid

                  Brandon,

                   

                  Well..That's what I like the most - real life geometry with real life issues here..:)

                  I've been dealing with similar examples many times and what I can tell you is that before trying anything else, I would 1'st get rid of all those small fillets from both the plastic part and the insert:

                  remove_fillets.png

                  Then you may start thinking about geometry intersecting by "imprinting" one body onto another.

                  Your goal is to get as clean geometry as possible even if you'll have to slightly change the part/insert geometry just to avoid slivers etc.

                  This will not impact your simulation results at all but will help generating a mesh that is matched between each intersecting geometries.

                  Imo, there's too much detail around those cylindrical shapes. If you absolutely need them, I would at least remove all the rounded corners if not simplify them into simple cylindrical blocks...

                   

                  You said you begin with the copper part set to "Insert Domain" - you should realize that in Plastics, the Insert Domain should be only used for classic over-molding scenarios. In your case, that cooper block is in fact a second mold but made of a different (highly conductive) material. That's why at the very beginning, you must treat both the plastic part and the copper insert as a CAVITY (Domain1). Why? Only then you'll be able to mesh them AT ONCE with a big chances of getting a compatible mesh. Of course assuming you've successfully imprinted one part into another. You might still need to play with your mesh size especially in the imprinted areas. Both too coarse and to fine may still cause some issues with node matching.

                  Then before transforming the surface mesh (perfectly matched at this stage) into volume Tetrahedral mesh, you will have to change the domain of your copper insert from Cavity to Mold (Domain 2) assuming the Mold (Domain 1) is already set to your Virtual Mold.

                   

                  One important note:

                  - please be sure that your cooper block is truly in touch with your plastic part...That is not guaranteed if you're working with the geometry that's been imported from a different CAD

                  package. It is sometimes better to build an exact or approximate copy in Solidworks.

                   

                  Keep trying and you'll get there.

                  I'm sure it can be done when looking at your example.

                   

                  Regards

                  Radek