I am trying to set up a 3D CFD scheme for thermal and flow modelling in Python using the finite volume method. The first concern is to build the geometry and an accompanying mesh that is efficient for the purpose (resembling something like the 3rd picture below).

enter image description here

For Python, there are many libraries for geometry and mesh creation (the last two are for optimization and formatting respectively):

Somewhat overwhelmed, I tried the MeshPy library since its description mentions finite element simulations specifically. However, there does not appear to be a lot of documentation with the libraries. I was unable to construct a mesh that could be fed to a solver (such as fipy) - also inhibiting the required properties or types for each material; in order to specify for example the initial temperature, thermal conductivity, specific heat and density.

I would like to inquire for a working example of a grid generation that could be fed to a CFD solver, from which ever library is preferred, preferably such that different grid types can be created as seen on the image above, and such that the mesh generator is compatible with STEP format. As geometry a simple box, sphere or cylinder is abundant.

  • 2
    $\begingroup$ Are you specifically interested in structure quadrilateral (resp. hexahedral) grids? All pictures you show are structured. Does the generated mesh have to carry the structure information with it? Note that finite elements that you mentioned are often done on unstructured triangular (resp. tetrahedral) meshes. Mesh generation is hard. Does it have to be done in Python? What about starting in the GUI of those softwares like Gmsh first? Mesh generation is hard. More so if there are additional requirements on that mesh. $\endgroup$ Apr 7, 2022 at 13:23


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.