Regarding convergence, it depends on multiple factors such as mesh quality, solution methods, and sometimes the actual physics is complex enough to take time to converge. In addition to monitoring residuals, I would recommend also monitoring mass and heat imbalances in the system.