CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM Running, Solving & CFD (
-   -   Discretization Error vs Residual Error? (

dancfd March 20, 2012 23:09

Discretization Error vs Residual Error?
Hello all,

I just ran a sensitivity study on my case to see if I could improve my accuracy by using SFCD (2nd order scheme for convection terms) vs linearUpwind (1st order accurate). All else being equal, the SFCD case had a p-residual of 5e-1 and the linearUpwind case had a p-residual of 10e-3. I expected better accuracy from the 2nd order accurate case, but the residuals appear to show that was not the case. I am running simpleFoam on a 2D airfoil. Could anyone please suggest why the residual is higher for the 2nd order scheme?

More strange behaviour: the SFCD took LESS time to execute the case, and resulted in values of Cd and Cm that were farther from the experimental data than the upwind scheme. I expected different. Can anyone explain this?



akidess March 21, 2012 03:49

I always thought linearUpwind is 2nd order upwind. Is plain "upwind" what you were intending to compare to?

dancfd March 26, 2012 20:05

Hello Anton,

Thanks for finding my error, I appreciate it.


All times are GMT -4. The time now is 20:54.