check NaNs for calculate_dt

[jiahe23]

Description

It will be helpful to check if calculate_dt gives NaNs during the initialization, and where the NaNs comes from.

[kpamnany]

calculate_dt uses nondiffusive_courant to compute the CFL, as you've observed. We can add a more informative error message, but can you clarify what you mean by "where the NaNs come from"?

[slifer50]

@jiahe23 can you specify how you get the NaNs? I've observed NaNs in dt when using --adapt-timestep but I always figured this meant the problem had gone unstable leading to a NaN dt. Is this similar to what you're seeing?

[charleskawczynski]

calculate_dt uses nondiffusive_courant to compute the CFL, as you've observed. We can add a more informative error message, but can you clarify what you mean by "where the NaNs come from"?

My first impression was to print an error inside nondiffusive_courant, which could say what the cause of NaNs (e.g., u, soundspeed_air), but now I'm remembering that nondiffusive_courant is called at runtime when CFLs are monitored, so that might not be the best place to throw the error. I suppose we could just print an error inside calculate_dt and say something like:

dt is NaN. This can happen for two reasons:
 - A poorly constructed mesh (e.g., min nodal distance is zero / NaN).
 - Improperly constructed initial conditions (e.g., zero velocity, or zero diffusion).