Acoustics analysis in ANSYS AIM

  • 600 Views
  • Last Post 08 May 2018
Bene04 posted this 23 April 2018

Hi all,

I want to do some acoustics modeling for my master thesis and I want to try out the ANSYS AIM because I am very new to the system and AIM looks fantastic to me. 

Does someone has experiences with AIM in modelling acoustic physics? I found a couple of acoustic tutorial on youtube but all used the ANSYS Workbench. It would be nice if someone can help me out.

Kind regards

Bene

Order By: Standard | Newest | Votes
peteroznewman posted this 23 April 2018

Hi Bene,

I am interested in acoustics and got some confidence that I was obtaining valid results when I could compare the value from a model with a textbook equation. I am also interested to see how to model physics in AIM. I am willing to help you out. What do you want to model?

Kind regards,

Peter

  • Liked by
  • Bene04
Bene04 posted this 03 May 2018

Hi Peter,

thanks for your answer and sorry for my late response!

 

I want to analyze how ultrasound affects the physics of a glass or plastic bottle with fluid in it (e.g. what vibration can occur).

For that I downloaded the acoustics act and now try to figure out how to use it with AIM.

 

Did you already make some experiences in comparable situations with acoustics act?

 

Many thanks in advance, Peter!

 

Kind regards

Benedikt

peteroznewman posted this 03 May 2018

Hi Benedikt,

My study of acoustics has been for sound travelling through air, but ANSYS acoustics supports sound travelling through water. I don't know the upper frequency limit when the acoustic equations ANSYS used become invalid. That is a good question.

I am fortunate to have medical ultrasound equipment at work and could scan things such as a black plastic bag with water and an object inside, and visualize what that object is inside the black bag.  I also know a PhD scientist working in ultrasound who I can ask easy questions to make sure I am on the right track.

I think the Acoustics ACT only supports Workbench, not AIM, but models you build in AIM can be opened in Workbench.

I recommend you find a textbook example of an ultrasound model where you have a textbook answer, then try to reproduce that result in ANSYS. Water is covered, but you will have to use a second material where you know all the properties required for acoustic modeling to make your first model.

Let me know when you have chosen that first problem.

Regards,

Peter

  • Liked by
  • Bene04
Bene04 posted this 07 May 2018

Hi Peter,

unfortunately I couldn't find any good textbook example yet for my ultrasonic "problem". Seems as it is a very rare area in ANSYS.

The first easy question that might be trivial for you is where is the difference between the "Harmonic Acoustics" and "Harmonic Response" in ANSYS? Harmonic Acoustics uses the ANSYS solver and Harmonic Response uses Mechanical APDL solver - but what to use when I want to display the influences of an external ultrasonic device to a surface/object? It is really confusing and I can't find a good hint in the help of ANSYS.

I am very thankful for your help so far and I feel sorry for being an absolute beginner.

Best regards

Benedikt

peteroznewman posted this 07 May 2018

Harmonic Acoustics is Harmonic Response but with extensions that makes it easy to do acoustical things like define a face to apply a harmonic plane wave into a medium, or to define a face where no waves get reflected to simulate an infinite domain on a finite sized solid. 

The Problem of using FEA Acoustic Simulation for Ultrasound Frequencies

I did a "back of the envelope" calculation to scope out the problem.

What is the maximum frequency, f, of ultrasound you want to use?  Typical medical ultrasound goes between 1 and 18 MHz.

Do you know the speed of sound, c,  in water?  It is about 1500 m/s and ultrasound imaging assumes 1540 m/s.

Let's calculate the wavelength of a 15 MHz sound in water:  lambda = c/f  = 1500/15e6 = 0.0001 m or 0.1 mm.

At least 6 quadratic elements are needed along one wavelength to adequately represent the wave. That means the maximum element size in the model would be 0.016 mm for 15 MHz. If you take a 10 mm cube of material, and mesh it with 0.5 mm elements, you end up with > 35000 nodes. With the Student license, the limit on nodes + elements is 32000 before it will not solve.

Even if you are on a Research license, a 10 mm cube of material with an element size of 0.1 mm, which is okay for a 2.5 MHz ultrasound, creates 4 million nodes connecting 1 million quadratic elements, so a 0.016 mm element size would create an unreasonably large model on a very tiny sample.

This is why FEA is not a good method for ultrasound. Fortunately the Boundary Element Method (BEM) can cope with ultrasound frequencies as it does not need to fill the volume with elements. Unfortunately, ANSYS does not have a BEM solver.

I recommend you select an acoustic problem with frequencies < 20 kHz through an air medium if you want to use ANSYS.

Best regards

Peter

  • Liked by
  • Bene04
Bene04 posted this 08 May 2018

 Hi Peter,

as you suggested to find a good "example" I did some researches and I found a very good example in the "workshop files" of the acoustics act which is called "Audio Speaker and Plate".

I tried to follow the guideline which is described in the document, but there are several problems which may are produced due to the fact that I am working with version 19.0 instead of 14.0 as it was in the guideline. For example I am forced to input some "Tabular data" in fields which were not existing in the old version.

Sadly the results of the workshop documents are not reproducable for me. Maybe you can help me with that? I think that this might be a good start into the materia. 

 

Btw ... the frequency my system operates is 40kHz. I think that this still fits.

 

Kind regards

Benedikt

peteroznewman posted this 08 May 2018

Hi Benedikt,

Attached is a zip file with the Release 19 workshop you describe.

Please describe your system and how it operates at 40 kHz. 

Did you do the back of the envelope calculation for element size?

What will be your simulation goals?

Attached Files

  • Liked by
  • Bene04
Bene04 posted this 08 May 2018

Hi Peter,

that is exactly the file I was using but apparently the guideline pictures are out of v14.0 and for example from sheet 16 on the struggle becomes real. It looks pretty easy to just select the geometry for the computional domain but my version of Ansys wants me to give much more information in form of tabular data (watch the attached file). And after sheet 16 much more "new" fields appear in the most recent version in comparison to v14.0.

 

Back of the envelope calculation:

f=40kHz, c(air)=330m/s => lambda=8.25mm

That means a single element is 1.375mm (=lambda/6). My plate is about 25mm x 25mm x 3mm, I should then end up with less than 35000 nodes, correct?

 

My simulation goal is to see what effects the ultrasound does have to the area behind the plate and to the plate and demonstrate what happens if it is in water or air.

 

Best regards

Benedikt

 

Attached Files

peteroznewman posted this 08 May 2018

Some volume of air or water behind the plate has to be meshed, so that counts too.

What is the source of the ultrasound? What is the shape of the transducer?

  • Liked by
  • Bene04
Close