[picpic]

Dear Foamer:

Currently, I'm working on wind simulation for building geometries. I set a cylindrical domain in order to create simulations for different wind directions (e.g. 10 degrees as an angle step). There are several ways I can achieve this goal, I have several ways to achieve this goal, but they all seem to be too complicated, and I would like to ask how Foamer with experience deals with this kind of problem.

Here is one of my plans:
Split the sides of the cylindrical computational domain into 36 patches, setting 18 for the inlet and 18 for the outlet each time.

However, due to a lack of experience, I am not sure if this is feasible and what kind of trouble it will cause subsequently, and I would like to ask if anyone has done something similar.

Any suggestion would be really really helpful.

Thanks a lot
Picpic

[Alczem]

Hey!


I think you can use the inletOutlet condition on your cylinder, with the appropriate vector for the inletValue field. I remember using it a few months ago. Alternatively, the freestreamVelocity condition might work I guess. Good luck, and let us know

[picpic]

Thanks for the suggestion. In fact, I still need to define the interpolation of the velocity in the vertical direction, because the velocity in the z-direction is increasing from small to large, I am not sure if this method is suitable, anyway I will try it, thanks a lot!

[Alczem]

You might want to look into setExprBoundaryFields utility to define a velocity profile based on an equation and dependent of the Z coordinate.

[Aeronautics El. K.]

Wouldn't it be easier if you had your domain and boundaries fixed and rotated your geometry instead?

[picpic]

Originally Posted by Aeronautics El. K.

Wouldn't it be easier if you had your domain and boundaries fixed and rotated your geometry instead?

Yes, that seems quite straightforward, but it means I need to create a set of meshes for each rotation, which is complicated for multiple simulations. I would like everything to be automated.

[picpic]

Originally Posted by Alczem

You might want to look into setExprBoundaryFields utility to define a velocity profile based on an equation and dependent of the Z coordinate.

Yeah true. There are many ways to define a wind profile for me, however for me the hardest part is changing the wind direction, tomorrow I will put up my boundary conditions, hoping someone(maybe me) can give the best solution.

[picpic]

Hi Foamers!

I think in some way I solved my problem. I set the sides of the computational domain to the same Patch and afterwards used groovyBC to solve the definition of the wind profile and the control of the wind direction. The initial results look not bad, although they produce some numerical diffusion at the boundary.

Better suggestions are welcome!

pic