CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   stream function & vorticity boundary condition (http://www.cfd-online.com/Forums/main/95049-stream-function-vorticity-boundary-condition.html)

 sara2115 December 4, 2011 12:45

stream function & vorticity boundary condition

HI everyone.
I've wrote a code to simulate velocity field over a square cylinder.
I've used vorticity-stream function solver at 0 angle of attack.(not immersed)
but now I don't know how to specify outflow boundary condition for different angles of attack.
I used a rectangular as my domain.

 mettler December 5, 2011 10:10

I did this for a finite difference method (except I did flow over a step). My exit boundary condition was @x=L, Streamfunction(@x=L) = Streamfunction(@x=L-1). I did this based on assuming that the flow after the step was fully developed and not changing. My vorticity BC was the same. The results were good.

 sara2115 December 6, 2011 11:53

Quote:
 Originally Posted by mettler (Post 334688) I did this for a finite difference method (except I did flow over a step). My exit boundary condition was @x=L, Streamfunction(@x=L) = Streamfunction(@x=L-1). I did this based on assuming that the flow after the step was fully developed and not changing. My vorticity BC was the same. The results were good.
but in my case I have vortex shedding which makes it more complicated.
By the way, did free stream velocity have an angle of attack in your case? or it was simply horizontal?

 mettler December 6, 2011 12:08

it had no angle of attack, but why can't you extend the rectangular grid long enough to where the vortex shedding has dissipated and the flow is fully developed?

 sara2115 December 7, 2011 18:42

Quote:
 Originally Posted by mettler (Post 334893) it had no angle of attack, but why can't you extend the rectangular grid long enough to where the vortex shedding has dissipated and the flow is fully developed?
Dear Mettler.
Well, I did so.
but still there are lot's of problems.
for example, even if the flow behind the cylinder is fully developed, it still has its angle of attack, so I can't impose the boundary condition you used for outflow.
there are other difficulties in dealing with far field boundaries.