What is the reason for the lack of convergence in solving thermodynamic problems?
 

What is the reason for the lack of convergence in solving thermodynamic problems in OpenFOAM?

I know the basic momets of how to stabilize a solving process, but I often get a negative temperature error when solving such problems. However, even when using standard solvers, this error occurs. As I understand it, the error appears in the transition zones, for example, when the flow enters the MRF zone. But if the flow is already in the MRF zone and does not leave it, then no convergence errors are observed.

I use a thermodynamic VOF model to describe the melting-crystallization process. I noticed that a decrease in the number of transformation enthalpy L stabilizes the solution process, but this violates the physical meaning and the crystallization process is slower than it should, which is obvious.

I used a low Courant number, high discretization of the entire mesh, increased the number of orthogonal correctors, and everything does not lead to normal stabilization of the solution.

I already read the forums about same problems. And the main problems of the all users were wrong pressure, temperature or velocity BC. I tried this and seems it doesnt help me much or I make a mistake somewhere.