What's wrong with my RMS profile in LES

Jiang Ning

I am now trying to simulate some channel flow with LES in fluent, but I met some porblem, and I am relly confused about that.

What I calculate is a very simple channel flow. The domain I chose is a cube with the size of 0.09+ACo-0.03+ACo-0.045m(x+ACo-y+ACo-z), the grid is 64+ACo-64+ACo-30. Air flow in x direction. the centre velocity is 8m/s. the inlet and outlet surface that normal to x direction is periodic condition(to simulate a very very long channel), the left and right surface(normal to z direction) of the cube is also the periodic condition, the top and bottom surface of cube is wall(no slip) condition.

I use the K-E model to calculate till it convergened, the result is used as the initial condition of the LES calculation. Then I switch to LES model(SGS model is simagorinsiki model), chose the PISO and centre difference in the discretization item. The time step fixed at 1e-5s. After 5e4 time steps, the mean velocity profile is quite similar to the theortic value.

My problem is the shape of the RMS velocity is quite different from the theortic analysis result. In y direction,it shoule be like , < ) <

but it is actrully like

< ( <

I tried many way to solve this problem, but I just can't find the solution. I think my grid is fine enough and the time step is small enough. I don't know what's wrong with my simulation.

Acturally I simulated several different cases, but non of them can get the correct rms profile.

Is here anybody suceed and get good rms profile? Any advise is welcome and highly appreciated.