Is there any member of the forum with experience in modeling heat transfer in high buoyancy driven flows? I'm looking for some advice and literature references for my Ph.D. Thesis.

Thanks in advance and best regards,

I believe you mean thermal buoyancy. I have no guess what you mean by high buoyancy driven flows. I suggest you may go thru the book titled (perhaps) 'Buoyancy Induced Flows' authored by Gebhart, Mahajan, Jaluria and [there is a fourth author - Samakia?], if you have not done so.

what is your problem ?

I'm modeling a fixed bed reactor. Flowing fluid is taken to be in supercritical conditions. There's particle-to-fluid heat transfer and important changes in the density of the fluid through the bed. Raleight number is extremly high and so is Gr·Pr, with a very low Re (< 100). I'm trying to model it as natural convection in laminar flow regime, but energy residuals don't get lower...

What is your Ra & Gr/Re² number range ? What length scale did you use to calculate your Ra number ?. How are the continuity residuals ? Try to under-relax energy (but not too much) with to begin a first order scheme.

Thomas

Hmm, maybe you can use bousinesque method or one of the others?

To do this just define variable density in the material properties of your fluid.

Density changes are very large, so I think boussinesq method is not appliable....

Numerical data: Ra: 1.02e9 Gr/Re2: 2.34e6 and particle diameter was used as lenght scale. Continuity residuals are around 5e-2. All under-relaxation factors are set to 0.1 and first order discretization schemes are applied...

Hi,

first of, under-relaxation factor should never be 0,1. Such a factor does not make your solution converging at all even if the residuals are decreasing.

Then There is something i do not get, all your number indicate that byouncy effects induced a turbulent flow.How that can be possible in a Fixed bed.Please check that value of Gr/Re². When this value exceed the unity we already expect stron byouncy effect, so i do not think a Gr/Re around 10e6 is very realistic.What material are you using ? Check the fluent documentation.

Are simulating steady or unsteady ? Unsteady will (without any doubt) give you good results but it also depend on what you are looking for in your simulation.

Thomas

Gr has a very high value. under supercritical conditions, the viscosity value is very low and so is the beta coeficient, this makes that Gr has a high value. The material is supercritical CO2.

Experimentaly it has been demonstrated that buoyancy efects over supercritical fluids are very strong, and Gr values are not far from those obtained experimentally and that I have checked.

I've checked fluent documentation an until now i've been following the procedures expresed in the user manual. Simulations are running in unsteady...