Length scale separation in LES
 

Hi, All,

Can anybody suggest how to determine or ensure length scale separation in LES? or how to ensure my cut-off length scale is inside the inertial subrange of the turbulent energy spectrum?

My case is that I have a package to run LES for an oscillatory flow. Apparently, the package runs well and I have got some results.I can change/refine my mesh, but do not know if this helps. I need to justify my application of LES for such a flow. As known, LES is usually valid for high Reynolds number flow where length scale separation presents. But for an oscillatory flow, the Reynolds number seems different from a pipe flow or free stream,...

Many thanks!

Re: Length scale separation in LES
 

What is your Re number?

One check you can do is to perform an FFT (Fast Fourier Transform) of your solution. Plot the energy spectrum and make sure that you don't have any accumulation of energy at the smallest scales (your grid size). If you have an energy peak at the size of your grid then you are not resolving enough.

Be careful though, many sub-grid models (Smagorinsky for example), are often too dissipative and can hide these problems.