I am very much still a beginner, but I would suggest going step by step and not changing too many things at once. You say the solution doesn't make sense, this can of course mean a couple of different things. You could be encountering a physics related issue, my first suspect in such a case would be boundary conditions. Second suspect could be a physical properties issue, double check your values (I once struggled to get expected results for like a week straight, only to realise I incorrectly computed the polynomial coefficients for rho(T) in my fluid properties). If it's instead a numerics issue and you are running the simulation in simple mode, then I would also recommend playing around with under-relaxation factors. Generally, a smaller value means a longer time required to reach the final convergent state, but it's 'safer' in the sense that it's less sensitive to numerical blow-up (again, in principle, I am not confident enough to claim this is always the case). Check any other field values that are being computed and if they make sense. It could also potentially be an issue with numerical schemes. What sort of schemes are you using? 1st order schemes are very diffusive, which can (in my limited experience at least) very significantly impact your result. Try out more accurate 2nd order schemes (if you can) and observe how your result is impacted. Also make sure to check your residuals, if you notice the values are too high, try increasing, for example, the number of corrector loops (but don't overdo this, as it can lead to other issues and instabilities). Finally, it could be a mesh related issue. I'm still a beginner but the very first thing I was taught was that if your starting point is a bad mesh, no amount of playing around with settings is going to save you. Checkmesh will conveniently put together any troublesome faces or cells into sets, which you can convert to VTK and easily visualize with paraview.