Modify viscosity in stabilization tau

[hepap]

Description of the problem

• For phase change problem, the viscosity used in the tau computation for the stabilization in the solid used the numerical solid viscosity. Since this solid viscosity has no physical meaning, it was weird to used it, and the linear solver seemed to converge slower.

Description of the solution

• Introduce new attributes in NS scratch data kinematic_viscosity_for_stabilization and dynamic_viscosity_for_stabilization. Both are computed by the rheology model. For now, except in the phase change, it returns the current viscosity. However, in the phase change rheology model, it returns the liquid viscosity.

How Has This Been Tested?

• The melting cavity case was used to insure that the results stay the same and to show better convergence of the linear solver.

• [x] heat_transfer_phase_change_constrain_solid_domain

• [x] heat_transfer_vof_phase_change_constrain_solid_domain

Those two tests were changed to have a more representative behaviour.

Future changes

• Test the viscosity scale in VOF and non-newtonian cases to see if it brings some gains

[blaisb]

@hepap can you rebase? This will then be ready for merge

[hepap]

@blaisb it's rebased!

[blaisb]

@hepap something went wrong in your rebase I think