[LF_05]

Hi everyone ! :)


I'm facing difficulties to properly run a simulation using the LRR Reynolds-Stress Model so maybe one of you may know how to solve my problem ?


Case :

I want to compute the flow over a thick 2D wind turbine airfoil (DU97W300) with a fully resolved boundary layer and an angle of attack of -4° (to begin with). I'm using an O-Grid mesh which works correctly with other turbulence models such as Spalart Allmaras and k-epsilon.

Since I want a fully resolved boundary layer, my y+ values are set <1 along the airfoil.

I also want to notice that I use simpleFoam in OF 4.1!


Problem :

My problem is the following : I tried many procedures to launch my LRR simulation (from a uniform field, after running potentialFoam, and after running a k-epsilon simulation) and I always face a divergence problem which seems to be related to diverging epsilon and time step continuity.

I tried to modify a lot of parameters (boundary conditions, solvers, smoothers, divergence schemes, relaxation factors, etc.) but none of theses worked at all.

That's why I'm now asking for help here since I absolutely do not know what to do to solve my issues.


I put my files in a zip file, I first ran a k-epsilon for 2000 iterations and then tried to switch to LRR but without success :(

So if any of you may know any piece of advice concerning this issue this would really be wonderful, thanks a lot in advance !


Luc :)

[LF_05]

So nobody has any clue about such an issue ? :/

[lpz456]

Originally Posted by LF_05

So nobody has any clue about such an issue ? :/

Hi predecessors, I am new to OpenFOAM. Have you solve your problem?
Recently, I also encountered the convergence problem of RSM model.
I downloaded your case and hope to get some guidance. I find that the Epsilon wall value seems too low. I think the relaxation factor can also be smaller, maybe try 0.2.
Perhaps you could try pisoFoam.I am going to use an unsteady solver to see whether I can get a steady result)
Thanks in advance!