[khaiching]

Dear CFD experts:

I am developing an unstructured code and encounter a problem whereby the flow vector is pointing in/out near the wall when buoyancy model is on.

I read from FLUENT manual that we can't set the normal pressure gradient at the wall = 0 when buoyant flow is simulated.

Can anyone please guide me on the proper method to interpolate the wall pressure? Any references to share? I need the wall pressure in order for me to evaluate the cell pressure gradient near the wall.

Please help.

Thanks a lot.

[CFDtoy]

Mirroring is not a good method when you have grav forces. Interpolation is the best way to achieve the pressure at the boundaries. Given this, use a second order gradient interpolation to connect Pcell with Pb - gives you second order accuracy.

/CFDtoy

Originally Posted by khaiching

Dear CFD experts:

I am developing an unstructured code and encounter a problem whereby the flow vector is pointing in/out near the wall when buoyancy model is on.

I read from FLUENT manual that we can't set the normal pressure gradient at the wall = 0 when buoyant flow is simulated.

Can anyone please guide me on the proper method to interpolate the wall pressure? Any references to share? I need the wall pressure in order for me to evaluate the cell pressure gradient near the wall.

Please help.

Thanks a lot.

[khaiching]

Dear CFD Toy:

Thanks for your prompt reply.

I tried to use the gradient projection method to interpolate the wall pressure, i.e.

p_wall = p_cell + grad(p_cell) * d; where d is vector pointing from the near-wall cell to the wall face. For the evaluation of grad(p_cell), I use Grauss divergence theorem.

Is it the method that you have suggested in your previous reply?

In fact, I still see some velocities pointing in/out near the wall.