[Archive] Mesh quality 1

Message by Emanuele Ruà:
I am trying to create a simple domain to simulate the plume of a Hall Thruster. However, I cannot create a mesh with an acceptable minimum cell quality. The domain is quite simple: the thruster and a semisphere as the external boundary, do you know if there are some best practices that I can use to get a good mesh quality?

Message by Fredrik Johansson:
Hello Emanuele. I mesh the .geo file in gmsh myself, and use both “optimize 3D” and "Optimize 3D (Netgen) sequentially before saving my mesh. That way I almost never have any mesh volumes below 0.3 quality. Is this the case for you? When starting a SPIS project, I therefore always skip the geometry editor when creating a new SPIS project.
If you’re still having problems I suggest simplifying the thruster model.

Message by Sébastien Hess:
Dear Emanuele,
I totally agree with Fredrik answer. In addition, note that I was pretty conservative when attributing a “note” to the mesh quality. If simple geometries should easily reach the “excellent” note, more complex one requiring large ranges of mesh size (usually the case for Electric propulsion) may not go over “bad”. This notation is primary here to warn the user of the importance of the mesh quality (the unawaereness of which was one of the primary cause of complaints about SPIS results in the past), so that if you can not reach better than “bad”, it does not mean that you cannot run the simulation, but rather that you have to pay attention to potential effects of the mesh quality in the results.
Note that Artenum is proposing a visual mesh inspector for SPIS (not free, but as I recall not that expensive either).
Best regards,