CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   URANS & SAS-SST: Computational Cost (

siw May 31, 2012 05:33

URANS & SAS-SST: Computational Cost

I've just run a 3-D cube in a wall bounded domain at Re = 40,000 with SAS-SST. The purpose of the simulation was to see if a SAS-SST run could be set-up and run on my computer successfully. The results look reasonable.

I'm now repeating the simulation with URANS inc. SST (everything else being unchanged) for comparsion purposes.

I've read some papers about SAS-SST and as far as I can see there's only an additional term in the omega-equation, called QSAS.

So I assume that the computational expense (memory required and run time) would only be slightly greater for SAS-SST compared to URANS due to the QSAS term rather than being a significant rise in computational cost. Would that be a correct assumption? Do my other CFDer's use SAS-SST, since DES and LES are expensive for industrial tasks?


PS. How can I get the eddy viscosity ratio in CFD-Post to colour the Q-criterion vortex core regions for showing the turbulence structure? I cannot find the turbulent eddy viscosity in CFD-Post or an equation to work it out.

ghorrocks May 31, 2012 06:01

The expense of SAS and related models depends on the flow. The smaller the turbulent structures (ie higher Re) the harder these models get compared to RANS. RANS tends to improve in accuracy as Re gets higher.

siw May 31, 2012 12:13

Thanks Glenn. Do you use SAS-SST much?

Also any hints with the eddy viscosity ratio query?

ghorrocks May 31, 2012 18:58


Do you use SAS-SST much?
Not for a very long time. Most of my models are micro-scale now so the flow is laminar.

Turb viscosity is a variable in the results file by default for turbulent simulations. You probably have to see the extra variables (under the "...." button) to see them.

All times are GMT -4. The time now is 15:00.