CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   SIMPLE for compressible flows (http://www.cfd-online.com/Forums/main/9703-simple-compressible-flows.html)

Luuk August 16, 2005 08:34

SIMPLE for compressible flows
 
Hello,

I need some help with the following. I'm implementing the SIMPLE method for compressible flow calculations in an incompressible finite volume code. I'm following the approach described by Ferziger and Peric in Computational Methods for Fluid Dynamics. I'm a bit confused by the following:

I have 3 variables: pressure (p), density (rho) and temperature (T). I solve a transport equation for T, so it is known. In the SIMPLE method, I calculate a pressure correction (p') and a density correction (rho'). To calculate the (new) p and rho, I have 3 possible equations: equation of state (p=rho*R*T), p_new=p+p' and rho_new= rho+rho'.

Summarizing: Option 1: p_new=p+p' and rho_new=rho+rho' ; Option 2: p_new=p+p' and rho=p_new/RT ; Option 3: rho_new=rho+rho' and p=rho_new*R*T.

What is now the best option (in terms of stability and accuracy)?

Thanks in advance for any help,

Luuk

ag August 16, 2005 08:55

Re: SIMPLE for compressible flows
 
In my (limited) experience I have found option 2 to be the most robust.

George August 20, 2005 08:45

Re: SIMPLE for compressible flows
 
Hi Luuk! I have exactly the same question. I have tried the 3 options and I get very different convergence rates. Also the results that I get are better with the first option, while with the second option are very strange, although I agree with Ag that the second option looks more robust. I am on holidays at the moment and I hope when I return to discuss this matter again. Bye


All times are GMT -4. The time now is 03:09.