CFD Online Discussion Forums - LES of Turbulent Jet

Hi Foamers,

I'm trying to simulate a turbulent jet flame using LES in OpenFOAM. I've been able to generate reasonable results with first-order time advance, but every time I try to change to second-order time advance my pressure field goes wonky. If you're interested, my struggles are described in this other thread: http://www.cfd-online.com/Forums/ope...e-advance.html

After working on this for a while, I've realized that my problem occurs even if I switch to the built-in solver reactingFoam and completely disable combustion... in short: I cannot produce a stable second-order LES simulation of a turbulent jet (with Re_D =~ 22,000) even in the absence of combustion. In light of this, I'm thinking that there is some fundamental problem with my test case, perhaps related to my grid. I've tried switching to a more uniform grid... and managed to produce a test case which fails even on first-order time advance :( Clearly something is not right; the least haphazard way that I can think of for proceeding is to find a turbulent jet LES case which is known to be stable and then experiment with test cases between my current case and the known stable case to identify what I'm doing wrong.

In light of this, my questions are:

Can anyone provide/point me to an example case for full-scale [cluster, not desktop] LES of a turbulent jet?
For those of you who've run full-scale LES: do you normally begin with first-order time advance and then switch at some point, or do you use second-order time advance throughout? What kind of transients do you observe (particularly in the p field) at initialization and/or when second-order time advance is activated?