Transport model divergence in CFD code

[CatarinaGarbacz]

I have mutation++ successfully coupled to SU2-NEMO already for several years.

The SU2-NEMO team has recently added ionization to the code. Simulations run successfully with the native thermo-chemistry library that uses the Gupta-Yos model for transport properties.

When running with Mutation++ (using the Chapman-enskog model) - the simulations diverges after a few iterations.

I noticed that values for thermal conductivity were extremely different using the two models (decimal values vs 10^2) and tried digging in further by running a simulation with Mutation++ however using the native gupta-yos model only for the transport. It worked well.

Besides welcoming any thoughts on why the Chapman-enskog could lead to such different values (in comparison to gupta-yos) for thermal conductivity and that being a possible cause for divergence, my main question is if there is any specific reason why Gupta-yos is not implement in Mutation++ since, given the aforementioned, this now became a possible development of interest.

@jbscoggi

Thanks!