[MichaelM]

Hey guys,

I am facing a little problem modeling an acoustic signal. Generally speaking, I like to model a piezo transducer with a momentum source term in CFX. I would totally appreciate it to get some more information as long as anyone might help me out. And it would almost be an honour to get an answer from ghorrocks

However, as I already mentioned I need to model a piezoelectric transducer and its influence on a fluid flow. The idea was to generate a subdomain (for the piezo device) and add an additional source term. I defined a sinusoidal user expression which is only depend on the local time variable t. The result of this expression is given as a variable 'Px' to the source term (for the time being only in x direction, y and z are set to 0). Unfortunately the results only show a wavelike pressure distribution inside the small subdomain but the pressure (or sound) waves do not reach the outer main domain.

Does anyone know what the error might be? Or is it even right to use a momentum source term to model an acoustic signal?

Thanks a lot in advance for any help that I get

[ghorrocks]

Flattery will get you nowhere. But it is a damn good place to start

I would use a mass source point, not a momentum source. That way:
* You do not need to define a subdomain, just nominate a point.
* Your momentum source has fluid moving up and down. This will generate acoustic waves but creation and destruction of mass generates more.

And of course - the sound waves will propagate out at the acoustic velocity. Have you run the simulation long enough for the waves to get to the outside of the domain?

[MichaelM]

First of all thanks for your answer

Yes I have run the simulation long enough, so this might not be the error.

I just tried it using a mass source point. The results show pretty much nothing, but I guess this is because I set the total source to 0 (I do not want to have an additional mass flow into my system).
What exactly do you mean with "creation and destruction"? How can I have a net mass flow of 0 but generate a sinusoidal pressure distribution?

Thanks in advance

[ghorrocks]

If the mass flow rate is a sinusoidal function then the mass is coming and going constantly, but over a long period of time it has no net mass flux. It all just cancels out.

[MichaelM]

Thanks ghorrocks! You were right, using a mass source term definitely is the better approach. Although I cant really visualize the acoustic waves, the pressure distribution at a fixed point in the far field shows a sinusoidal fluctuation.

Thanks again!