Preprocessing

Preprocessing

Issues with Hemispherical Tip Mesh in Maxwell 3D

    • ajbroadbent
      Subscriber

      As part of my Master's Thesis, I am using ANSYS Maxwell to run a series of simulations of an array of micro "needles" to see how the electric field changes as the ceter-to-center spacing changes. The needles are a cone with a hemispherical tip, modeled using Maxwell 3D in ANSYS Electronics Desktop 2017 as 1/4 of the full array with symmetry boundaries applied. Below is a screenshot of the model at 20um spacing:


      Needle 20um Spacing 


      The issue I run into is the meshes of the needle tips along the symmetry boundaries do not generate properly when I get to a spacing of 20 um. I have tried using different mesh operations to better refine the mesh, but I still run into some sort of "hole" in the mesh. I have also searched through Maxwell3D's documentation and online, but have not found any solution. Additionally, no one at my school has any prior experience with the particular software package as we just recently got a license for it. Here is an image of the corner needle's mesh at 20um spacing with a 0.01um length-based mesh operation on the tip:


      Corner Needle 20um Spacing


      As you can see, the mesh near the bottom of the spherical region starts out fairly uniform, but once it gets close to the tip, it falls apart. The needles along the boundary sides also have a similar issue, while the 4 "internal" needles that are fully modeled have no issues with mesh generation. I am not very experienced with mesh generation and simulations, so I am not sure if the mesh is too fine, too coarse, or if there is some setting or function I need to apply in order to resolve the issue. If anyone with Maxwell3D experience has any suggestions, that would be greatly appreciated as I need results in order to finish my thesis.

    • peteroznewman
      Subscriber

      I don't have Maxwell, but I have meshing skills and am willing to look at your geometry. Save just the SpaceClaim file and attach that so I can have a closer look.  If the Attach button complains about an unsupported file type, put it in a zip file as that is supported.  So is a Workbench archive, but you would have to make one with just a Geometry cell from the Components toolbox.  What Version of SpaceClaim are you using?


       

    • ajbroadbent
      Subscriber

      I'm not sure what version of SpaceClaim, or if that is in fact what is being used as everything is self-contained within the Electronics Desktop application (Version R18.0). Nothing is being run through workbench, though I have tried to generate a model and mesh that way and import it, but was unable to find a way to import the mesh. There seems to be an option to link or import a SpaceClaim  geometry, but I am still looking into it. I am able to export the geometry as ANSYS 3D Modeler (.sm3), IGES, STEP, and a few other file types, but not SpaceClaim files. I've attached a .zip folder with my geometry in .sm3, .iges, and .step formats if that works. 

    • peteroznewman
      Subscriber

      Hey AJ,


      I used SpaceClaim 19.0 and it would not open the .sm3 file, though it did open the .step file. While the needles in the center came in as solids, the needles on the symmetry plane came in as spheres at the top and a surface on the cone. SpaceClaim seemed unable to easily zoom in on one needle. I split the sphere into quarters but I don't know how to extract a face and join that with the cone. I guess I'm done with trying to help you mesh.


      Good luck,


      Peter

    • ajbroadbent
      Subscriber

      Thanks for taking a look. I posted the same question on another forum, and someone mentioned to adjust the surface deviation and aspect ratios in the initial mesh settings. I did that and it seems to have fixed the issue, as I no longer get the type of breakdown seen in the pictures from my original post. The results so far look promising, so hopefully no other issues come up.

Viewing 4 reply threads
  • You must be logged in to reply to this topic.