Personal tools
You are here: Home Forum Other... (new forum, problems, etc.) Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

Up to Other... (new forum, problems, etc.)
  • Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

    Posted by John Smith at March 23. 2008
    I get this message frequently when running Salome (smesh?) 3.2.6 under CAELinux (PCLinuxOS).
    "Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()"
    I'm using .stp STEP models imported from CATIA V4/V5 as the base geometry.
    Mesh algorithms Tet(netgen)/Triangle(mefisto)/Wire discretisation/segments around vertex


    Any suggestions what may be the cause?  I've tried search engine googling and didn't find anything helpful.

    Thanks!
    John

    • Re: Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

      Posted by gunawan at May 20. 2010

      Please Help..!

       

      Could You help me to mesh my s'tep file, i was imported from solidwork n autocad

       

      my crane hook file is here :  http://www.4shared.com/file/5NdgLrjv/crane_hook_Gunawan.html

       

      I was try to mesh but I got the problem like the model is not solid.

       

      i hope u can help me to mesh my cranhook in step.

       

       

      thank's before

       

      Gunawan

       

    • Re: Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

      Posted by Ulf Moritz at July 08. 2008
      Hey John,

      did you solved this Problem? I'm running the Salomé-Version 9.2.9 of CAELinux and receive the same Problem. If you know about the reason, please tell me, where the source of the Problem is located!

      Thank you!

      ulf
      • Re: Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

        Posted by John Smith at July 08. 2008
        I'm still solving this myself.  I have to go back through this whenever I have a new CAD specialist send me models (some training) - I don't do the CAD work, only the FEA and Engineering analysis.

        This error seems to be caused by surface construction tools in Catia - some features were made from surfaces rather than solid geometry (like a cutting plane instead of subtracting a block).  One recently failing Catia V4 geometry model was "fixed" by pulling it into SolidWorks, fixing edges and changing to a solid entity and then exported from SolidWorks into a step file.  Outputting an iges file from Catia V4/V5 also causes this problem.

        To test if this is the problem with your model... try subtracting a cylinder or a block through some flat (non-contoured/complex) section in your part.  If Salome throws an error during cutting - not allowing completing the operation - then the model is suspect.  If it cuts, but you can see two "sheets" of the original model, then the model conversion is suspect.  You should see a nice solid section around the hole that when you "explode to faces" has a face you can select.  If it's just a model of surfaces you won't be able to generate that face where you cut the model to select it.

        If you can get the model to mesh (but mesh fails), go to the option of "controls" (I think, I don't have my Caelinux box running now) and click on "show free edges" (it's a toggle button).  Then click on the geometry line in the left tree selection window (something other than the mesh model) so the mesh grid gets "hidden" you should be able to rotate the model and see many small red lines outlining free edges on the dark blue part.  Some will be very small so you can only see in this manner (mesh hides them).  If these free edges are small and in unimportant regions you can go back to geometry and take very small "bites" to cut them out or add small blocks or patches over them.  Try a couple of different mesh densities and try to minimize the number of error edges. I've found most of them localized around complex curves (corner of a cube with radii on all edges).

        The other thing I've tried is to cut a model up into many smaller chunks and try meshing each piece separately.  I had one I cut in half, then one half into six pieces, then the problem piece into six more pieces, and found the CAD model had a surface construct of some sort.

        The best solution is to get a new Catia model created with solid geometry and not surfaces.  Then only work with step files (iges may work but I've had the best luck with step files so far, iges inherently strips out the solids and presents surfaces from the research I've done).
        .
        Hope that helps!   I haven't solved it - I just have found some ways to test and check. If you get any closer to solving this then let me know!
        • Re: Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

          Posted by Ulf Moritz at July 08. 2008
          Hi John,

          thank you very much for your elaborate answer!

          Yes, right now i'm just working with .stp-files, because i also made better experiences with it than with .igs.

          I really can create a mash. But at some locations in the geometry Salome has problems. You said, when i found by the highlighted "free edges" this critical locations, i should go back to the geometry part of it and fix them there. I didn't understood, what you meant completely. Sorry for the perhaps little bit stupid question. But do you think, when i have some "vacuum" , i should just put in this free space just some material? I think the prolbem is the face, and even if you fill this space the faces will be the same, i think. Did i understood you right?

          And one more fundamental question. Sorry about this: where can i observe, if my CAD programm (Autodesk Inventor in my case) creates solids or surfaces? I created my things by extruding and think, that it normely should create solids. But also in this case i'm not sure, if i'm right.

          Thank you for your help!

          Best regards!

          ulf
          • Re: Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

            Posted by JMB at August 24. 2008
            Ulf Moritz wrote:
            Hi John,

            thank you very much for your elaborate answer!

            Yes, right now i'm just working with .stp-files, because i also made better experiences with it than with .igs.

            I really can create a mash. But at some locations in the geometry Salome has problems. You said, when i found by the highlighted "free edges" this critical locations, i should go back to the geometry part of it and fix them there. I didn't understood, what you meant completely. Sorry for the perhaps little bit stupid question. But do you think, when i have some "vacuum" , i should just put in this free space just some material? I think the prolbem is the face, and even if you fill this space the faces will be the same, i think. Did i understood you right?

            And one more fundamental question. Sorry about this: where can i observe, if my CAD programm (Autodesk Inventor in my case) creates solids or surfaces? I created my things by extruding and think, that it normely should create solids. But also in this case i'm not sure, if i'm right.

            Thank you for your help!

            Best regards!

            ulf

            Hello,

            I have faced numerous problems with STEP files exported by CAD designers who do not realize the effects (or havoc) created by CAD models that look fine in appearance (shaded model looks great!), etc but do not mesh at all.  John's answers have aided me in resolving many issues, especially using the Display & Publish options of the window that shows up when meshing has failed.  Thank you John!

            As he mentioned creating 3D models using NO surface modeling procedures in one way to ensure meshing success.  Another tip is to use dimensions that are accurate to the 6th or 8th decimal place when using mating geometry.  Thirdly, avoiding the creation of very narrow or very small surfaces to faithfully make a model may be irrelevant to the analysis process.  An example to this last point is creating accurate fillet radii in regions where stresses are low, is un-necessary faithfulness to geometric accuracy.  Such faithfulness may result in small patches or surfaces that subsequently fail in the meshing process.

            Also at times the "Repair -> Shape Processing" tool has aided me, at other times it has worsened the problem.  Experiment with it and its underlying values.  I have not found a "this works everytime" type of solution yet!

            As for the your question about AutoCAD and its faithfulness to produce good mesh worthy solid models, I do not have the answer.  Sorry I have not tried it, but may play with it sometime to test it.

            Regards,
            JMB


            • Re: Tetrahedronn (Netgen) ... Exception in Ng GenerateVolumeMesh()

              Posted by John Smith at August 25. 2008
              JMB,

              I have learned recently... there is an option to "turn radii/fillets off" in Catia - I had a problem part that my Cad Designer turned them off and resent the step file and I could run the rest of the analysis fine.

              I've also used the Repair->shape processing successfully (if the model is close).

              I'll have to try out the tolerances out 6-8 decimal places as you stated, it looks promising (and fits with the Repair->shape processing improvements - might be some relationship there that could be exploited).

              Glad I was able to help!


Quick reply
URLs will be automatically hyperlinked. Basic HTML tags are OK.
Powered by Ploneboard
History
Activate by Site Administrator on Sep 02, 2009 12:02 PM
Document Actions