Hi,

is it possible to use a symmetry boundary for a plane channel LES calculation on the sides or does it influence the development of the flow too much? Would be nice, if someone has an explanation.

Flo

A symmetry plane in this case would essentially act like an inviscid wall. So, the velocity perturbations normal to this boundary will be zero and there will be a region near it where these perturbations ramp up to the expected value. I suppose if the length of your domain in this direction is big enough so that you see a flat region in the boundary-normal velocity perturbation rms and sample for statistics there then you are ok. Why not use periodicity?

Thanks for the explanation! Flo

You are artificially restricting your flow in the spanwise direction. You have essentially halved the width of your channel. I don't think this is a good idea. In a real flow, the turbulent eddies would be free to cross this plane, so by restricting them you are altering the physics.