CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   UPWIND scheme: boundary condition ? (http://www.cfd-online.com/Forums/main/13815-upwind-scheme-boundary-condition.html)

Lionel S. July 16, 2007 12:01

UPWIND scheme: boundary condition ?
 
Hi all,

It's said in the Fluent's manual that physical values are stored at cell centers. The values at cell faces are taken from the nearest cell center in the upwind direction.

It all seems good, but what do I do with my solver if I have a boundary condition ? If the boundary condition is in the downwind direction, I will never consider it as a value for my cell face.

I suppose that I have to take the boundary condition anyway if there is some, and otherwise take the upwind cell center. Any hint about that problem ?

Thanx !!

sudhakar July 17, 2007 05:17

Re: UPWIND scheme: boundary condition ?
 
hi Lionel,

I suggest you to refer CFD book by Malalasekera & Versteeg.. In that book, an 1D problem is solved using upwind scheme in chapter 5.. also chapter 9 explains how to implement boundary conditions in staggered grid... hope it helps..


TITAN Algorithms July 17, 2007 05:39

Re: UPWIND scheme: boundary condition ?
 
yes, that is correct. The standard first order upwind terms will not sense this. That is, for a first order stencil, the advection terms will not be affected by the BC. However, your diffusion terms, mu*DELTA([v] will sense this boundary condition. By setting outlet velocity BC, you are emulating an extraction type flow. If you have a vortex roll out the boundary, this will be consumed by the BC. As a side note, this type of BC will tend to shorten vortex bubble lengths, i.e., solve back step with short boundary and a mass-balanced outlet velocity BC and you will get a reduced recirculation length.

regards

DSS


All times are GMT -4. The time now is 01:10.