[andrew-c]

I am trying to simulate a very simple case of molecular diffusion. I have a thin rectangular domain filled with water vapor and open at the left end (pressure-outlet, with gauge pressure 0. Note that I found the same results with a pressure-inlet of gauge pressure 0, as expected). The right wall has a raised temperature, and all other walls are adiabatic. Gravity is disabled (note that this is not a simulation of natural convection).

I've enabled the species model, and I am using incompressible-ideal-gas for the density of the water vapor, along with a constant mass diffusivity and kinetic-theory for thermal diffusivity.

I expect mass flow to occur with heating, due to density changes creating a concentration gradient. However, when I use the pressure-based solver, this does not occur. Densities (and concentrations) drop on the heated end, but no mass flow occurs. Even stranger, total system mass decreases without any outflow, which appears to me as a violation of mass continuity.

On the other hand, the density-based solver shows small amounts of mass flow due to these density changes (orders of e-2 to e-4 m/s). The total system mass decreases, but the mass flowrate oscillates from positive to negative. I also find this odd because I expect primarily inflow, due to density decreasing within the domain, but I may just be thinking about it wrong.

Any ideas why the pressure-based solver is not showing these diffusion effects? If there's anything I may have missed in my setup, that may be useful to know as well.

Thanks!