Reply To: Strange Divergence in natural circulation simulation simulation using VOF

[Rob]

Yes. In CFD the maths will conserve mass. The domain size is fixed. If density is temperature dependent where is the volume going to go?  In theory you can set temperature dependent liquid density & ideal gas but that'll still cause problems as you try and boil the liquid: Fluent doesn't allow temperature & pressure dependent density other than idea or real gas. Try patching a layer of vapour using ideal gas at the top of the domain & see how it behaves.