[T/THM] Heat transport in stationary flow w/ ice

Presence of ice phase at slightly positive temperatures with shallow
sigmoid function for ice phase should not affect the results much but
only for amount of ice formed. In the simulation the initial ice volume
fraction is 0.0379... . Displacement and stress are read from previous
simulation's steady-state result.

Incorrect update of phi_fr_prev causes pressure to drop significantly in
the first step, which is unexpected, since everything is in equilibrium.
Temperature load starts only after 10 steps at t=10000s.