Learning Forum

[Luigi0]

Hi, 

I am trying to model acoustophoresis phenomena. My system is composed by a fluid domain surrounded by a mechanical structure and a piezoelectric transducer. The fluid domain has an inlet where the fluid (water) and particles are inserted. The piezo vibrates and pressure waves are generated in the channel.

The fluid is modelled in fluent, he solid bodies in ansys mechanical and they are connected with system coupling (transient analysis).

[Luigi0]

Do you have any advice?n

[Konstantin]

Hello,nThis actually looks to be a correct behavior. A pure plane wave is an idealization which in reality can be reproduced only in some special cases, e. g. in a long duct with an oscillating end. Since in your simulation the oscillating plate is finite and it moves in a space wider than the plate, you'll inevitably see non-planar end effects. An analytic solution can be derived for a finite oscillating plate in an open domain which will clearly illustrate non-planar nature of the solution. I believe this analytical solution can be extended even to your case where the wave reflects off the opposite wall. So in summary, nothing wrong with your simulation, and the flow field appears to be physical.n

[Luigi0]

Thank you for your reply. I would like to add some updates and know your opinion.nMy problem is that the non-planar end effects don't remain confined at the edges in the following time-steps, but they affect the wave also in the central part. I'll put some pictures that represent the same problem of the previous post but some time-steps after (from 1 to 2 in chronological order):nnnThus, I try to extend the fluid-solid interfaces to the entire length of the bottom and top surfaces of the channel (in the first picture of my previous post, the red lines that represent the fluid-solid interfaces are extended both to the inlet and outlet) and I obtain planar waves except for the outlet. nI think there is a conflict between the pressure outlet condition and the pressure variation caused by the propagation of the waves. The problem is that the perturbation is not confined at the outlet but it travels to the left (into the channel) and affects the waves propagation (pictures below, from1 to 4 in chronological order)):nnI don't know how to solve the problem, the only thing that come to my mind is to extend the channel and move away the outlet.nDo you have further advice?n

[Konstantin]

That's pretty much what you should do: leave the wall extents as they are now and move the inlet/outlet even further away. This way the test section will see the planar wave and inlet/outlet effects will be minimized.nWhat working fluid are you using? You may also consider doing Fluent only simulation with MDM to simulate the oscillating wall without systems coupling to simplify troubleshooting / tuning of the modeling approach. Once that Fluent model works per your satisfaction, you can bring back the systems coupling n

[Luigi0]

Thanks Array . The working fluid is water and I enable the compressible liquid density method. nMDM stands for moving/deforming mesh? And what should I do to simulate oscillating walls in fluent without the system coupling?n

[Konstantin]

here is relevant Ansys How to video:nn