Following your drawing, there is a problem of wavelength effects. If the frequency is that of ultrasound and the angle is just right, the sound can be absorbed fully in the cavity which acts like a blackbody in this case- regardless of the turbulence. Changing the frequency or tilt angle might saves your wave. If you have a wave viewer you could see the path.

For longer wavelenghts, the cavity and turbulence can be less disruptive and your wave can be heard at the microphone. Not forgetting that some of the wave is transmitted in the walls which you need to take care off.

In any case the turbulence effects should be random, so averaging over a longer time can be useful. regards.

This kind of problem requires a specific formulation such as DNS or LES. What have you adopted ?

(@ Filippo Maria Denaro) thanks for the comment. So, I used ansys fluent Detach Eddy simulation (LES+URANs Mix), and there were so many problems. For example, when wave propagated by using inlet sinusoidal velocity of 0.024 m/s with 1800 Hz frequency. And when the meanflow was zero, I was seeing the wave was not terminating at oulet biundary conditions i.e.,P=0 showed reflection so somehow with density based solvers the non-reflecting boundary conditions worked with using 3rd order MUSCL Scheme. But then I couldnot make a transparent inlet boundary conditions ....although I thought I should change the wave inlet into non-reflecting pressure outlet conditions during middle of simulations when the wave bounce from the turbulent flow/cavity in the middle but it couldnot work. Then i put meanflow value of 40 m/s and then i couldnot see my wave anywhere ....everything was mixed in this flow and it was really a not nice experience.

So do you think that if someone want to study acoustic wave interaction lets say 1800 Hz sinusoidal wave or Ricker Pulse, is it ok to use LES OR Detach eddy simulation method?

Or there can be some specific numerical schemes something like using some non-linear hyperbolic system of euler compressible flow equations for discontinuity motion etc...i mean it will be really hard? To simulate a small wave in high mach number flow?using CFD.

One small attempt, I found that Ansys Mechanical APDL were using Convective wave equation just to propagat planar wave via meanflow velocity using Finite element method but that solution was in frequency domain and i could just see how much wave energy is at inlet and outlet.

But still dont know have a wave behaves when hit a turbulence medium!!!in time domain

Fawad Ali

first of all, you need to be sure you are able do a correct simulation by repeating some test for acoustics.

If you have computational resources, use only DNS or a pure LES.

@Filippo Maria Denaro

Thank you so much for the comments. Yes i think i should first try to correctly simulate just an acoustic wave and further i will try LES. Many thanks for your time and comment!!!

Dear Ali,

One should always start with the most simple problem, which is the propagation of sound als low frequencies in a straight pipe. B y low frequencies I mean that only plane waves propagates. For a duct with height H this corresponds to frequencies f = W/(2 pi) such that c/(f H)>2, with c the speed of sound in the fluid. For circular pipes of radius a this implies W a/c

@ Pr. Avraham Hirschberg,

Thank you so so much for the detailed answer. Infact its a nice idea to start from large wavelength sound propagation. I will try this simple approach to have a little bit confidence on simulation results then will see meanflow effect. Thank you for the references i will check one by one.