Discussion Forum

Helene,

the incident wave gets reflected from the surface. So what you see is a superposition of the incident wave and the reflected wave, building a standing wave in front of the reflector. Could this explain the observation?

Cheers Edgar

Thank you for your reply. Yes, the total pressure equals the incident plus reflected wave. What I don't understand is how the resulting pressure could be more than doubled (more than twice the incident pressure)? If the surface is rigid (which is the case), the total pressure should be exactly twice the incident pressure, which is not the case.

How big is the deviation? It could be a matter of mesh resolution or discretization.

I guess we agree that this should not happen. The deviation is not big, the total pressure's magnitude reaches between 7 to 8 dB around 2kHz (which is not a lot more than 6dB, corresponding to doubling pressure). I also observe that the deviation increases with the size of my object. Your intuition may be true. In order to define the mesh of my 3D object, I followed the recommendations of the Comsol tutorial for the scattering on a sphere: the mesh is custom with max. and min. element size being defined by min(0.3,lambda0/4), with lambda0=c0/f0 (f0 the maximum computation frequency and c0 the sound celerity). Do you have better recommendations regarding the mesh discretization? Thank you for your help.

I would recommend to perform a mesh convergence study. If you are using linear discretization you might consider to try quadratic. 8 dB is too much to my taste. You might also check if there may be some resonance in your system.

The surface of my object is rigid, i.e. the pressure is 0 at the boundaries. Could the incident plane wave make the rigid object resonate? Could you please specify how I could verify there is no resonance in my system? Thank you very much.

If you have an acoustic-structure coupling active then the solid might resonate. If it is set up rigid it will not move. Rigid is different from solid. Also the acoustic domain might resonate, depending on the setup, think of a flute. You could run an eigenmode or eigenfrequency study to check for resonance. If an eigenmode is close to your range of excitation frequencies you can do a frequency scan around this mode.