|August 26, 2009, 10:19||
determining the time step size in LES using FSM
Join Date: Apr 2009
Posts: 41Rep Power: 8
I am a running a LES on a confined impinging jet reactor(CIJR). This a brief summary of what I encoutered..
The reactor is planar and has two inlets and one outlet. It is design that prud'homme group has been studying. From our microPIV experiments we know that we were getting a flapping motion where the two jets impinge in the reactor. But when I ran my LES simulation with the smagorinsky-lilly model, I didn't see any flapping at all.
I used a second order time integration and SIMPLE scheme. I had set the convergence for continuity to be 1e-3. Later in our discussions, we realized that 1e-3 for continuity is too high and I ran a simulation with 1e-6 as the criterion for continuity. It took a lot of iterations (~75) to converge at each time step (2e-5).
I then came across a suggestion in one of the threads here, wherein they recommended using non-iterative scheme like FSM instead of iterative scheme.
I have been running that for a couple days now. There are couple of questions however, that I am not clear about. How do you decide on the time step size in FSM and when I select the FSM, in the residual monitors, I can no longer specify the converge criteria like I could for SIMPLE scheme. So, how do I know if the solution is converged in the non-iterative FSM scheme.
Any help is greatly appreciated.
Thanks a bunch in advance
|August 27, 2009, 14:07||
Join Date: Mar 2009
Blog Entries: 14Rep Power: 17
Well, the question regarding the time step does not depend on the pressure-velocity coupling but on the required time-resolution and the courant number. In my experience the LES in Fluent needs a courant number much lower than 1, the optimum being 0.1...for practical applications this could be maybe relaxed somehow up to 1.5-2. whatever was the time step you used with the simple there is no reason to change it if it satisfies the previous requirement.
The FSM, differently from the SIMPLE and the other classical P-V coupling methods, does not involve the outer iterations between the pressure and the momentum equations because the error committed in this can be made of second order in the time step which is consistent with the accuracy of the time integration method. All the parameters related to the FSM are defined in the same panel where you activate it and you should check the user's manual for a reference (i don't remember any of them)...however the solver is already optimally tuned to obtain the second order accuracy in time unless the courant number is very small (in this case the parameters are too much restrictive) or too large (in this case the parameters shoul be made more restrictive).
|Thread||Thread Starter||Forum||Replies||Last Post|
|Superlinear speedup in OpenFOAM 13||msrinath80||OpenFOAM Running, Solving & CFD||18||March 3, 2015 06:36|
|How to write k and epsilon before the abnormal end||xiuying||OpenFOAM Running, Solving & CFD||8||August 27, 2013 15:33|
|Is there a way to write the time step size, time a||may||FLUENT||6||November 22, 2009 12:52|
|Navier-Stokes time step size||Martin||Main CFD Forum||2||June 6, 2008 03:38|
|unsteady calcs in FLUENT||Sanjay Padhiar||Main CFD Forum||1||March 31, 1999 12:32|