Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Apply for Membership

(no subject)
  • admin March 2009
    Hi
    I' m using engrid to mesh a simple object like the ahmed body with a tetrahedral mesh and all is OK. I' ve seen in the examples page a case with polyhedral mesh: I suppose you use polydualMesh by Openfoam. I tried to pply polyDualMesh to the mesh that I' ve created in engrid (and that pass the checkmesh control) in order to have a polyhedral mesh, but I' ve a lot of errors both during polyDualMesh creation then applying checkMesh: here is some option (or some trick) in engrid in order to have a tetrahedral mesh that is compatible with polyDualMesh ?
    Regards
    Marco
  • admin March 2009
    Marco,

    If you are using the latest snapshot -- not the 0.9 version released earlier
    -- you should find an OpenFOAM entry in the export menu. This will export
    the grid as polyhedral grid. Be warned, however, that this is an incredibly
    inefficient algorithm and it will probably crash on large grids. Just try
    it, if it works it should give you a decent polyhedral mesh. What it does is
    it keeps the prismatic boundary layer as it is and only dualises the
    tetrahedral part. Thus you don't get the problem of cells "wrapping" around
    sharp edges, like you would get with polyDualMesh. Have a look at the
    pictured from the rocket example and you'll see what I mean at the
    connection between boundary layer and farfield. The export function will be
    improved soon and there will also be a dedicated smoothing in order to get a
    nice dual mesh for the tetra part.

    For now, just try it and if it doesn't work, fell free to send me your grid
    and I'll have a look. checkMesh might still give you errors, but try running
    anyway. I had a discussion with Mattijs Janssens from OpenCFD about the
    different errors and what they mean; for the future I am quite confident
    that we can get rid it of most of these error.

    Regards,
    Oliver