LES->BFS
 

Hi all,

I am working on to develop a LES code, with the implementation over, I am trying to validate the same with Backward facing step using Moin's DNS paper.

Although I am new to this, I have used the following for my Boundary conditions, for the Inlet, I have used a fully developed channel flow imposed with Random fluctuations using Random guassian numbers. For the lower wall, I have been using No-slip BC and top the no-stress condition. And not to forget time marching is through Crank-nicolson scheme.

My question is when I run the simulation for some time, I could c the vortex gettting formed behind the step, but after sometime, it breaks and then gets washed away..and again the same thing keeps repeating...due to this the velocity profile seems to alter near the lower wall and keeps changing due to the formation and breakage...although I am taking the time averaged values..yet all that I am getting is a flat plate profile.

I am not sure whether this phenomena is correct or not? any help in this regard is highly appreciated.

Thanks in advance,

Glen

Re: LES->BFS
 

Did you try to run the code with different method ? Also try solving the laminar version of the problem and make sure that the code works fine. Then proceed with the turbulent analysis.

-H

Re: LES->BFS
 

HI Harish,

I forgot to mention, the code works well with, laminar and also k-e model in the steady state, as LES demands unsteadiness, I have included the same, and also the revelant changes.

Thanks,

Glen

Re: LES->BFS
 

What numerical scheme has Moin used in his paper... Crank Nicholson is second order in time and space.Has someone got interesting results with second order scheme ? . If not may be try to change to fourth order runga kutta in time and compact or DRP scheme in space.

-H

Re: LES->BFS
 

I am currently doing a BFS and I am using the Realizable k-e model with non-equilibrium wall functions - from the initial results it would seem that I capture the primary and secondary recirculation (lets not call it a vortex for now), shear layer and reattachment point.

Going to the wind tunnel in the next couple of weeks to try and validate.