Reply To: Natural convection convergence issues

[Rob]

Yes, you should get a single density for the domain (in contours turn off node values) as opposed to an average. You can patch regions with velocity/temperature/whatever later on, this step is just to get the operating density.