CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Phoenics (
-   -   High-order discretisation schemes (

Hu April 17, 2002 14:06

High-order discretisation schemes
Could anybody share some experiences of using high-order schemes? I found it is difficult to get convergent computation. I am simulating the wind flows around bluff bodies and there is no scalar transport and heat transfer involved. Could anybody shed the light of using high-order schemes with respect to which parameters should be taken care of? Thanks.

michael malin April 24, 2002 07:11

Re: High-order discretisation schemes
Detailed recommendations can be found in the PHOENICS Journal paper: Schemes for Convection Discretisation in PHOENICS by M.R.Malin & N.P.Waterson, PHOENICS Journal, Vol.12, No.2, July 1999 (ISSN 0969-8248).

Hu April 25, 2002 20:41

Re: High-order discretisation schemes
Thank you, Mike. I've got the paper you mentioned in hand. I've tried QUICK, HQUICK, HCUS, UMIST, Van Leer 1, Van Leer 2, SMART, KOREN, CHARM, MINMOND schemes and I found the rate of convergence is quite slow and sometimes the convergence is not attainable. The QUICK scheme cannot yield a convergent solution in my case which is an investigation of airflow around a surface mounted cube. The others are fine. I used the CGGR solver and the value of SARAH was 1.0e-4. I also set ISOX=1. I just want to know if there are some 'tricks' or 'tips' that may be used to accelerate the rate of convergence when using these high-order schemes.


All times are GMT -4. The time now is 17:50.