Some weaknesses and omissions
- Cats2D does not simulate radiation heat transfer in any significant way. It has a boundary condition that computes radiation to an external temperature profile specified by the user, but it does not compute view factors or heat exchange between surfaces. It implements the P1 approximation for radiation in optically thick participating media, but this model applies under a very limited range of conditions. On the other hand, Cats2D has been successfully coupled to a radiation heat transfer code, and recently I installed a basic API for coupling to other codes.
- Cats2D uses a structured mesh of quadrilaterals which restricts the range of geometric distortion that can be simulated using its sharp interface deforming mesh treatment. Cats2D does not simulate a mushy interface, nor does it include a level set representation of interface position, things that would be useful in simulating casting and mold filling. I think that implementing these features would pose only moderate difficulty.
- Cats2D has a relatively clumsy mesh generator that works well for creating simple meshes, but is laborious for complicated meshes, and Cats2D has no provision to import meshes created by other software. The code could be adapted to overcome these limitations, but I think it would be fairly difficult.
- Cats2D is a laminar flow code, with no provision for turbulence modeling. I think it would be straightforward to add a RANS-type turbulence model. However, I question its usefulness for the translationally and rotationally symmetric geometries computed by Cats2D. On the other hand, Cats2D's ultrafast solver might be useful in DNS of turbulence in a simple geometry.
- Cats2D has no general provision for specifying temperature-dependent or concentration-dependent thermophysical properties. Implementing this capability should be straightforward, however, and I expect it to get done eventually.