Hello everyone, I am a graduate student in Canada and I am doing my research in simulating the air flow patterns and heat transfer in a 13-floor building using FLUENT. The calculation domain is the inside of the whole building and the calculate domain¡¯s boundary is the walls(here we can call them wall-A so that we can distinguish them from the other wall-B inside the domain). I have got the pressure distribution from FLUENT for all the wall-A and wall-B. And I have got some experimental data of pressure distributions from the wind tunnel lab for the outside walls-A (outside of the calculate domain). For wall-A, we want to define them as permeable wall and use the pressure difference to drive the air flow through the wall-A. I don¡¯t think FLUENT can solve this problem unless we enlarge the calculation main to the outside of the building and define these wall as porous wall. I want to hear your guy¡¯s ideas about this problem. Maybe I am wrong. Thanks a lot. Lisa

No need to extend your domain to the outside. you can have internal porous region in your domain. As long as you input the correct value for the porous media, it should work ok.

Thanks for your reply. It is true for the internal wall-B that I can define them as porous-jump media. But for the outside wall-A which is locating on the boundary of the domain, the boundary condition of porous-jump wall won't work. I tried before.

Use porous zone instead of porous jump there but you have to come up with porous media resistance coefficients there.

I wonder what the objective of your work is? are you looking at HVAC simulation inside the building or heat/air transfer through the walls?

Thanks for your reply. I think I didnot explain the problem clearly. In the calculation domain, all the walls were defined as thin wall. I cannot create the wall as solid zone in the geometric model because of the memory of the computer. The objective of my work is to simulate the air flow patterns caused by the fresh air in the building and heat loss from the building, not exactly the HVAC system. We have got the heat loss through the wall. The problem of the air thansfer though the wall, is caused by how to simulate the cracks in the walls, doors, windows.We cannot create the realistic cracks in the geometric model still because of the memory of the computer. Lisa

okay, That makes it difficult. I would probably make some calcualtions by hand to find the pressure on the inner walls and apply that to the model as the boundary condition. there might be better ways though. BTW, which university do u study at? I am at UNB

Thanks for your reply. I am in UWO. Seems FLUENT cannot handle these problems. but I know CFX can do it. Thank you for your time and good ideas. Lisa

Lisa, you *may* be able to use switch your A walls from walls to inlet vents, which allow you to specify a loss factor or K factor. This limits your loss to the velocity squared term (dP = K x v^2 / 2g).

Thank you very much for your suggestion. I will try it later. Then I will tell you the results. Lisa

I think you may be right Lisa, by extending your computational domain you will have more stability that way. I am under the impression that the outside of the building is under atmospheric pressure. So you can make your new extension a pressure boundary. Good luck.

Thank you. I think my computer wonnot handle the whole domain including both inside and outside the building. it will be more than 2 million nodes. Lisa

You cannot make use of any symmetry to cut your model in half.

Yes. It is true.I cannot use the symmetry if I have to expand my calculational domain because of the experimental data is not symetric. Thank you for your reply.