CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Upwind vs High Resolution (http://www.cfd-online.com/Forums/cfx/20306-upwind-vs-high-resolution.html)

Liwau April 17, 2004 04:50

Upwind vs High Resolution
 
Dear All,

I can get a very good converged solution for a gas-solid two-phase flow by using upwind scheme. However, when I change the upwind to high resolution, no converged results come out. Could anybody please tell me how to solve this problem? Thanks.

Liwau

BAK_FLOW April 18, 2004 16:07

Re: Upwind vs High Resolution
 
Hi Liwau,

there is always a trade-off between accuracy and robustness. The upwind scheme is very robust because it adds significant numerical diffusion and increases the diagonal dominance of the equation system. Second order schemes produce more accurate results on reasonable grids but have dispersive errors which can result in over-shoots and under-shoots in the solution (commonly called solution wiggles) and difficult to converge solutions. For more on this there is a great deal written, some good sources are the texts by Ferziger and Peric, Versteeg, etc See the books online section of this site for more info.

The hi-resolution scheme is a blend of second order and first order which tries to maintain as much second order as possible. This is a desirable mix however, because the order of the scheme is a function of the solution this also adds an additional level of non-linearity.

There are a number of things you can try:

1. refine the mesh in the regions of hight gradients

2. use a fixed blend factor as high as you can say 0.9 or higher

3. smaller time steps

Good Luck,

Bak_Flow

Jane April 20, 2004 11:49

Re: Upwind vs High Resolution
 
Hi guys,

What second order scheme does CFX use? Central differencing? Power Law? QUICK?

Thanks, Jane


All times are GMT -4. The time now is 08:56.