[Madeleine P. Vincent]

I am looking at calculating the drag coefficient of a simple structure - in this case a rectangular building. Flow is from left to right, directly hitting the front face of the structure and flowing over the flat roof.



The top (sky) boundary condition is some length above the top of the roof, far enough not to influence the flow over the structure.



In real life, the building sits on the ground which causes a velocity profile on the front facing wall, rising from 0 velocity at the ground.



The question is, how should this bottom BC be modeled? If it is a stick (no slip) condition, then the shape of the velocity profile will have a large influence on the forces on the building.

Modeling it as a slip / symmetry boundary condition would overcome the need to decide what the inlet profile should be, but this might not model reality well.



So it seems that there are 3 options:


1) Model the ground as a stick condition, and let the inlet flow develop into an atmospheric boundary layer. Velocity would remain 0 at the base of the structure.



2) Model the ground as a slip / symmetry condition.


3) Do not have a ground condition at the base of the structure. Rather, give an equal (and large) distance between the roof and the top symmetry condition as the distance between the floor and the bottom symmetry condition. This would sort of "suspend" the building in the flow.


Is there a consensus on which is correct?


Regards,
Madeleine.

[agd]

Seems to me that you answered your own question - the boundary layer on the ground will have an impact on the drag coefficient of the structure, so including that is important. Since you don't specify a particular flow solver, I'm not sure if this is an option, but one thing that is commonly done is to use a slip boundary near the inlet and then start the viscous boundary ahead of the body so the boundary layer can develop and you don't have to worry about the inlet profile (since that part of the wall is a slip boundary). Imagine a flat plate in a wind tunnel with the structure downstream of the plate leading edge.

[FMDenaro]

Have a look to this example of cfd in urban enviromment
https://www.researchgate.net/publica...n_environments

[LuckyTran]

1) is much more correct than 2) and 2) is much more correct than 3).

In reality the velocity at the inlet will not be uniform but also a profile. Welcome to the wonderful world of modeling where we try to figure out how much we can get away with and still get the result we want. A crappy solution is better than a perfect non-existent solution.