Hi,all forum users I want use fluent to simulate the convective heat transfer enhancement by acoustic cavitation, but is badly short of experience about this aspect. The problem is simply described as follows.

The heat transfer is natural convective. The heat transfer is enhanced through radiating ultrasonic wave (sound pressure: p=pm*sin(wt), plan to realize it by UDF) to the liquid. The essential is that the liquid cavitation is indued by ultrasonic wave.

Who have some experiences about this. Any suggestion,prompt,clue and document will be welcome and appreciated.

My email address is sc_caijun@163.com

Fluent does have a liquid cavitation model. Read the relevant manual section. One of the tutorials also deals with cavitation.

The problem may be in how to generate the conditions necessary for the cavitation model to be activated. The liquid may need to be modelled as "compressible" for the pressure waves to be able to spatially propagate through the fluid. Setting the fluid as compressible may exclude the use of the cavitation model.

Alternately, you could just model a microregion with the assumption of constant local pressure and then simply impose the prescribed pressure.

I know that cavitation in fuel injectors has been modelled with fluent.

Search for "cavitation ultrasound cfd" in Google Scholar.

hello Joe, thanks for your reply an answer. Yes,I just do it like what you said. First, the steady natural convective heat transfer without cavitation was modeled. The laminar model was selected, and a steady-state solution (temperature and velocity field)was obtained and regarded as the initial conditions for the heat transfer computation with cavitation.

In the computation with cavitation, the cavitation model in Multiphase menu was and k-e turbulence model was selected. The pressure of fluid in a given region changes according to the prescribed pressure (p=pm*sin(wt)) and was inputted by UDF(DEFINE_ADJUST micro). when I started the computation, the iteration can continue if unselecting the Flow model in Solve/Controles/Solution panel, but will stop if selecting it and the following error message appeared. "Error: Floating point error: invalid number Error Object: ()." Even if the iteration can continue, the residual curves, especially the vf-phase-2 curve, fluctuated drastically and could not reach a good convergence.

I am not sure for the following problems: 1. Whether selecting the turbulence k-e model is right or not? How about the other models, such as laminar,LES etc. 2. Is it reasonalbe to unselect the Flow model in Solve/Controles/Solution panel? I think it's wrong. 3. How to obtained a good convergence?

I am longing to get useful advice and help from you.