Dear All -

I'm trying to simulate 2-D convection of water in a large geometry (30m width, 6 m depth) with CFX5. Rayleigh number is quite high due to the geometry and I was wondering if anyone has experience solving such kind of flows?

Cheers,

Toni

Natural convection problems are quite interesting in that for very large Rayleigh numbers the solution tends to be transient and it is difficult to get a converged steady state solution. Don't be surprised if you run into this. Start with a relatively low Rayleigh number, say 1000 or so, and ramp it up from there.

You will probably also find that solutions at high Rayliegh numbers are very grid and numerics dependent as well. i.e., as you refine the grid the solution will keep changing because you simply keep resolving more unstable frequencies.

Dan.