Intervention for always positive variables

[Niceno]

Dear all,

I noticed that we don't perform any interventions on the linear system for variables which should be always positive, like turbulent kinetic energy, its dissipation, species (a.k.a. scalars) and entries in the main diagonal of the Reynolds stress tensor.

I think it would be very good to do it, as it would be a neat way to dodge non-physical solutions, and should improve convergence too.

Cheers,

Bojan

[Niceno]

Connected to the above, we should also decided if we want temperature to be one of the always positive variables (which would be physically sound).

[mhadziabdic]

This is a good idea, we will then be able to remove a lot of these ugly tiny additions in denominators. Temperature can go into this category as well.