Hi all,

I am modelling two 3D simulations in which a hotter (another is cooler) retangular block (boundary condition as "wall")is placed in outdoor air.

As the case is simplified into pure convection case, so I define another rectangular domain, which closes the heated block, with 6 faces as "pressure" boundary condition.

My question is, for this outdoor environment where the domain only consists of air, which face should I use as " pressure inlet" or "pressure outlet"?

Thank you very much

I don´t know if you are interested in the flow of the natural convection itself. If you are not one posibility is just to calculate the solid and take a convection - bc with the free stream temperature of the air and a htc between 5 and 10 W/m²K.

If you want to simulate the convection: i think it does not matter if you take pressure inlet or outlet ( i would prefere an inlet opposite an outlet) but mention that you generate single cells in your system ( special in the corners ) that have more than 1 bc ( i mean outlet on top and outlet or inlet on another surface ) I would try to avoid that. Gernot

Deae Gernot,

Yes, I want to simulate this convection.

So you mean that I have to specify those 6 faces of the domain with "pressure inlet" and "pressure outlet" opposite to each other? I don't quite undertand your explanation.

As the heated block is placed at the centre of the 3D domain, I use the "pressure outlet" at the top surface of the domain, and other 5 faces with "pressure inlet".

Is it OK? If not, can you suggest me how to deal with the case? Thanks

what I wanted to tell you if you do it like you want to then you´ll get 8 cells ( one in ervery corner ) that have 3 boundary conditions ( one on every outside ). Also they are the same bc this can cause problems i don´t think that you´ll get good convergence. I don´t know how many cells you have used until now but i think the better solution is to make the area that large that you can make one inlet and one outlet opposite an the rest walls. hope that helps

Dear Gernot,

Yes, I am facing the convergence problem so I do wonder my boundary condition settings.

The "cell" you refer, do you mean that I should first divide the large domain volume into smaller ones, and then I assign those boundary conditions of smaller volumes one by one? Yet I don't quite understand your concept.

Thanks very much

Hi,

Have you considered using pressure outlet boundary conditions on all faces? This being unsteady problem, you may be interested in time accurate solutions. Then, you will be able to take care of air incoming from your outer rectangular block boundaries with proper settings for backflow. Also, what about the "far field conditions"?

Also, you probably need to fix the distance between inner solid and the outer enclosure so that you can capture the flow field without un-due effects from boundary conditions.

Gernot is probably mentioning the corner cells at the inter-section of two faces of outer block. Is it necessary to pay special attention to such cells whose two faces may have separate boundary conditions?

Swarup.