Mixing By Natural Convection Processes

Greg Perkins

Hi,

I'm looking at a process involving surface reactions (combustion/gasification) on walls in large reactors. The mixing and fluid flow in the reactor may be governed by natural convection which is most likely time-periodic.

Unfortunately, I don't think I can simulate the transient behaviour due to computational limitations - its 3D, >10m size^3 and I have a minimum of 7 species and the residence time in the reactor is large.

In order to approximate this situation I've been looking at using a steady state model, and modifying diffusion parameters to "appproximate" the mixing effect of the natural convection.

Has anybody tried this? How did you go about it? Was it successful etc???

I imagine that in the limit of very high diffusion I'll get a CSTR type model from the CFD, while with the standard steady model I'll get a lot less mixing. Probably the natural convection mixing and mass transport - if I could simulate it - would lie between these two extremes. So I thought I'd look at the extremes and see what I get.

Anybody got any ideas, suggestions or inisghts which might be helpful?

Anything appreciated

Greg