Does anybody have the experience to model "Stack Effect" (hot air rises and exit the domain from the chimney) using Fluent? The temperature is between 300 -1000k. I am using Incompressible ideal gas law to model the dependence of gas density on temperature, but the convergence is a big problem. As FLuent User Manual says, Boussinesq approxmiaton only applies in the condition of small variance of temperature. So does anybogy give me suggestions or references on modeling chimney effect? Thanks a lot.

What are you using for inlet and outlet boundary conditions? For a simple case, you can use pressure boundary conditions with the difference in pressure equal to the density difference in the stack and the stack height.

Thanks. My case is cool gas (300k) flow into the domain through two pipes (one on the top of the domain, one is the bottom), and the domain wall is hot with fixed temperature, 1000k). There is a opening on the ceiling of the domain. So my inlet bc is velocity inlet with 300k, and outlet bc is pressure outlet with 0psig.

So would the flow field be different from a real case if I use the pressure difference rather than density gradient, as you suggested?

By the way, do I need to set the radiation boundary conditions for the wall?

Thanks.

I guess you should change the density of air to 'Bouyancy' from the drop-down list (in the materials panel) instead of Incomp. ideal. Than you can adjust the thermal expansion coefficient of the air in the same panel a little bit below... (For air it is 0,033[1/K] if I remember correctly, but check it in a database!!) Also in the 'solver' panel click in the 'High bouyancy effect' I hope I could help you

Thank a lot Laci.

Are you talking about the item "Density" of Material panel? But I didn't find such a option of "Bouyancy". There are opitions of: constant, ideal gas, incompressible-ideal-gas, etc.

I tried both Boussinesq and incompressible-ideal-gas, but they are hard to convergent to my desidied level, even 1.e-3 for continuity.

You are right, it must be the Boussinesq. For the better convergence you should decrease the Energy coefficient from 1 to 0.8 On the 'solver' panel you can also click the 'high bouyancy effect' (Sorry, if I don't write the exactly places and names, but I cannot reach the software right now ) If you do these, you can have lower residals, however sometimes it is even not enough. In that case I would 'adopt' the grid respect to the temperature gradient. I hope ot will work. Bests.

Hi Laci, thank you very much for your suggetion. But my case is really hard to be convergent with buoyancy effect (continuity index can only decrease to 1.e-2). Do you have any other comments on that?

I appreciate your time.

[vivek]

hi guys...i juss wanna knw how to analysis the buoyancy effect of fluid(air) inside a solar engine in 3d.....plz give me som suggetions abt boundary conditions and the inputs for the analysis

[udayrg]

Hi

Feidao Li


Well, I would suggest to model pr. inlet and pr. outlet bc conditions and for density use Ideal gas equation and set up the gravity you certainly see the flow developed and you can set up relaxation factors for energy 0.7 or 0.8 and momentum 0.6 or so......

Are you modeling the domain around the stack ?

I am quite confident you would get continuity to 1e-3 range at least.

Uday

[xavi]

Hi Feidao Li,
I recommend you to create an extra volume control and impose pressure inlet for the inlet and pressure outlet for the outlet, and introduce the heat source in the domain (increase in temperature), using the incompressible ideal gas law or the compressible should be all right. As you said Boussinesq is for small temperature differences.Decreasing under relaxation in the solver may help as well. Best luck

[hook]

Hi, Feidao Li,
Why not try the ideal gas? It will get better results.