Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Frequency response depending on the size of acoustic domain

Please login with a confirmed email address before reporting spam

Hi!

I am working on a pMUT MEMS device. It consists of piezoelectric (AlN) sandwiched between two Al electrodes with a Si3N4 passive layer. The COMSOL model consists of the pMUT device and spherical pressure acoustics water domain. My goal is to observe and optimize the resonant frequency of the system.

Water domain - pressure acoustic domain (please see attached image and .mph document):

  • PML outer layer (size is 10% of the radius of pressure acoustic domain)
  • Viscous fluid model (with default viscosity, desity and speed of sound values for water)
  • Exterior field calculation - boundary between pressure acoustic domain and PML (full integral)
  • Spherical wave propagation - boundary on the edge of PML

Study: Frequency sweep (1MHz to 10MHz, step 1kHz)

Problem: When changing the size (radius) of the pressure acoustic domain I obtain multiple coupled-oscillator-like responses. I expected coupled oscillator response, but not multiple peaks. And when increasing the size of the radius, more oscillation curves can be observed (see attached image).

My goal is to create a scale-able acoustic domain model, mainly to simulate the pressure distribution at distances > 1cm from the pMUT. I have tried to resolve this problem for the past month and have tried several different model varieties, but with no success. I do not know what am I missing out, that is why I am asking for opinion you guys, maybe someone could explain this behaviour and how to resolve it.

Thank you all in advance!



4 Replies Last Post Feb 1, 2021, 3:07 a.m. EST
Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 3 years ago Jan 24, 2021, 11:56 a.m. EST

Aleksander,

the PML needs 4-5 mesh layers and it must be a concentric mapped mesh. See the documentation for PML mesh requirements. In the PML settings you can also try the rational coordinate stretching. I sometimes got pretty different results with the different stretching methods.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Aleksander, the PML needs 4-5 mesh layers and it must be a concentric mapped mesh. See the documentation for PML mesh requirements. In the PML settings you can also try the rational coordinate stretching. I sometimes got pretty different results with the different stretching methods. Cheers Edgar

Dave Greve Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 3 years ago Jan 24, 2021, 1:09 p.m. EST

Further- I have seen recommendations that the PML width be set in terms of wavelength not as a fraction of the domain size.

lambda/2 is probably a good starting point but...documentation on this is sparse.

Further- I have seen recommendations that the PML width be set in terms of wavelength not as a fraction of the domain size. lambda/2 is probably a good starting point but...documentation on this is sparse.

Please login with a confirmed email address before reporting spam

Posted: 3 years ago Jan 25, 2021, 4:26 a.m. EST

Thank you both for your rapid response!

Edgar: You are right, there should be more layers in the mesh of the PML. I have read some blog posts and decided to go with 5 layers. I have also set the PML stretching to 'rational'. However, the problem still persists, I obtain same result for bigger acoustic domain radius.

Dave: I have checked in COMSOL blogs and here: https://www.comsol.com/forum/thread/29495/width-of-pml it is mentioned that "According to Comsol Support answer I got, it does not matter. The PML will always be "one-wavelength" thick per default, you should just take care how you mesh it. " In any case, I will try to simulate bigger and smaller PML size to see if it would play any role in the outcome.

I am more than happy to try any other solution and recommendation.

Cheers, Alex

Thank you both for your rapid response! Edgar: You are right, there should be more layers in the mesh of the PML. I have read some blog posts and decided to go with 5 layers. I have also set the PML stretching to 'rational'. However, the problem still persists, I obtain same result for bigger acoustic domain radius. Dave: I have checked in COMSOL blogs and here: https://www.comsol.com/forum/thread/29495/width-of-pml it is mentioned that "According to Comsol Support answer I got, it does not matter. The PML will always be "one-wavelength" thick per default, you should just take care how you mesh it. " In any case, I will try to simulate bigger and smaller PML size to see if it would play any role in the outcome. I am more than happy to try any other solution and recommendation. Cheers, Alex

Please login with a confirmed email address before reporting spam

Posted: 3 years ago Feb 1, 2021, 3:07 a.m. EST

I figured out the problem. There were quite some actually, so I´ll post the solution in case someone else gets stuck with similar problem.

I neglected the fact that PML settings were actually the most important in infinity-boundary simulation. I just followed the documentation that I attached to this post and the model works as it should, no deviations with the acoustic-domain size variations.

Hope it helps.

Cheers, Alex

I figured out the problem. There were quite some actually, so I´ll post the solution in case someone else gets stuck with similar problem. I neglected the fact that PML settings were actually the most important in infinity-boundary simulation. I just followed the documentation that I attached to this post and the model works as it should, no deviations with the acoustic-domain size variations. Hope it helps. Cheers, Alex

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.