Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

General questions about Transient Acoustics and Solid Interaction

Please login with a confirmed email address before reporting spam

Hello fellow Comsol users,

I would very much like to ask a few relatively simple questions regarding transient acoustic models with some regarding transient solid interaction.

In essence my models are looking at how a membrane found in animal ear's behaves to incoming sound pressure, I am only looking to model the membrane not the ear itself. As you can imagine the membranes are very thin (around 1um in thickness and 2000um in diameter). I am doing the models in 3D and am looking at how the membrane deforms in the presence of a sound pressure wave.

I have adapted the Gauss Pulse model to a 3D one. I am modelling the source of the sound as a point Gauss Pulse.

My questions?

1)
Is there a better way to model an incoming sound field? The problem is that I need to place the Gauss Pulse relatively far from the membrane such that the incoming wave has a relatively flat wave front as it comes into contact with the membrane. This however creates relatively large models.

Is there perhaps a plate that I could use as the source of the pulse?

2)
Does anyone know of any other models which I could look at? There are not that many models in Comsol that look at Acoustic-Solid interaction.

3)
Any thoughts or suggestions on how to make the model more efficient.

I have included an image of the model so make things a little clearer.

Thanks very much, I hope there is some discussion to build on the knowledge base on acoustics in Comsol.

Rob


12 Replies Last Post 15.10.2015, 15:41 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13.01.2012, 12:33 GMT-5
Any ideas anyone?

I thought it might be nice to have a bit more discussion about the transient analysis on structural interaction.

Any comments would be really welcome.

Thanks,

Rob
Any ideas anyone? I thought it might be nice to have a bit more discussion about the transient analysis on structural interaction. Any comments would be really welcome. Thanks, Rob

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13.01.2012, 23:02 GMT-5
Hi,

I can give you suggestions on the different ways to give your source input.
Don't really know how efficient they will be.

1) Using a flow point source you can give your own pulse input .
For eg. S=4*sin(2*pi*100000[Hz]*t)*(t<=0.000001)
In the equation dS/dt is used.Here you have a pulse of frequency 1MHz and time period=1u sec.

2) Using the pressure boundary condition. Since its in 3D you can select a surface as the boundary
and give your pressure pulse as above in the p0 field.

3) Having a background field. Use the cylindrical, plane or spherical wave boundary condition and using the incident pressure field mention the direction of incidence by specifying your propagation constants.

I'm not sure of how effective they are comparatively, just started working on acoustics some time back.

There are couple of models that could help in understanding, a 4.1 version is the hollow cylinder .
I'm not exactly sure if i've helped you.

Regards
Glenston
Hi, I can give you suggestions on the different ways to give your source input. Don't really know how efficient they will be. 1) Using a flow point source you can give your own pulse input . For eg. S=4*sin(2*pi*100000[Hz]*t)*(t

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14.01.2012, 04:04 GMT-5
Hi

I do not have acoustics, unfortunately, but as general remark, your pulse 1) is not quite differentiable so you might have some solver issues, you could multiply it by a pulse operator to smoothen it, or try a 1+cos() over a period, or a gaussian type smoothing

--
Good luck
Ivar
Hi I do not have acoustics, unfortunately, but as general remark, your pulse 1) is not quite differentiable so you might have some solver issues, you could multiply it by a pulse operator to smoothen it, or try a 1+cos() over a period, or a gaussian type smoothing -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14.01.2012, 05:23 GMT-5
Thank you both for your replies!

I am going away for the weekend and will be away from Comsol for a few days, but I am very much looking forward to trying these out.

Thanks again, I will upload some models which I find to be handy.

Again, thank you

Rob
Thank you both for your replies! I am going away for the weekend and will be away from Comsol for a few days, but I am very much looking forward to trying these out. Thanks again, I will upload some models which I find to be handy. Again, thank you Rob

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.01.2012, 05:44 GMT-5
Hello fellow Comsol-ers!

After a nice weekend of relaxing I am again ready to pull my hair out!

Thank you Glenston and Ivar. Very happy to start a discussion on the topic.

I have a few follow up comments...

Glenston, thank you for your three suggestions. I think your first is the most suitable for my problem.

Ivar, I am not sure what you mean by "is not quite differentiable" but I do believe I have come into some solver issues you mentioned.

I am not sure how you apply smoothing but shall have a read of it now.

Again, thank you both. I will play with these models more and let you know what I find.

Rob

ps, Glenston thank you especially for the equation, it will come in very useful.
Hello fellow Comsol-ers! After a nice weekend of relaxing I am again ready to pull my hair out! Thank you Glenston and Ivar. Very happy to start a discussion on the topic. I have a few follow up comments... Glenston, thank you for your three suggestions. I think your first is the most suitable for my problem. Ivar, I am not sure what you mean by "is not quite differentiable" but I do believe I have come into some solver issues you mentioned. I am not sure how you apply smoothing but shall have a read of it now. Again, thank you both. I will play with these models more and let you know what I find. Rob ps, Glenston thank you especially for the equation, it will come in very useful.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.01.2012, 10:19 GMT-5
Hi

it's the general issue with i.e. time solving and functions that turn on /off as Dirac pulses, at the transition you have two different derivatives, and this is a headache for the solver. If you use the rect or step or pulse function of COMSOL V4 you will see these have a Heaviside type turn on/off transition that has at least one valid derivative, so the solver can exercise its "step and through" process

--
Good luck
Ivar
Hi it's the general issue with i.e. time solving and functions that turn on /off as Dirac pulses, at the transition you have two different derivatives, and this is a headache for the solver. If you use the rect or step or pulse function of COMSOL V4 you will see these have a Heaviside type turn on/off transition that has at least one valid derivative, so the solver can exercise its "step and through" process -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.01.2012, 11:03 GMT-5
Thank you Ivar,

I see what you are referring to, my hope is to see how the model deviates from experiment and to see if there are solver issues. My hope is that there are none!

Also, for those who find this thread on the forum please have a look here...

www.comsol.de/community/exchange/191/

Its a model I have created to encapsulate what I have learned!

Rob
Thank you Ivar, I see what you are referring to, my hope is to see how the model deviates from experiment and to see if there are solver issues. My hope is that there are none! Also, for those who find this thread on the forum please have a look here... http://www.comsol.de/community/exchange/191/ Its a model I have created to encapsulate what I have learned! Rob

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.01.2012, 12:26 GMT-5
Hi Robert,

Can you please upload a 4.2 version of the model.
Its not opening on the 4.2 so i'm assuming that its the 4.2a model.

Regards
Glenston
Hi Robert, Can you please upload a 4.2 version of the model. Its not opening on the 4.2 so i'm assuming that its the 4.2a model. Regards Glenston

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 17.01.2012, 06:15 GMT-5
Hi Glenston,

Thanks for spotting that! I have now updated the model in the model exchange with a 4.2 version.

(I did not know how to save as a 4.2 file in 4.2a so I created a new one)

Rob
Hi Glenston, Thanks for spotting that! I have now updated the model in the model exchange with a 4.2 version. (I did not know how to save as a 4.2 file in 4.2a so I created a new one) Rob

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 17.01.2012, 06:15 GMT-5
Hi Glenston,

Thanks for spotting that! I have now updated the model in the model exchange with a 4.2 version.

(I did not know how to save as a 4.2 file in 4.2a so I created a new one)

Rob
Hi Glenston, Thanks for spotting that! I have now updated the model in the model exchange with a 4.2 version. (I did not know how to save as a 4.2 file in 4.2a so I created a new one) Rob

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 08.02.2012, 15:18 GMT-5
Hi Rob,

By reviewing your guys previous discussion, I think I am trying to do sth similar. I would like generate a transient pulse to see the pressure distribution in a domain and some of the interaction with silicone which is in my environment. I am kind of stuck here by finding a similar plate source. All comsol give are the point or edge source. How can I have a circular boundary source?
Could you offer some ideas?
For your project, did you do the transient acoustic interaction physics?
By the way, I am trying to download your files from comsol exchange, however, my 4.1 version cant open any of them.
Do you happen to have 4.1 version?

Thanks a lot.

Best

Xu
Hi Rob, By reviewing your guys previous discussion, I think I am trying to do sth similar. I would like generate a transient pulse to see the pressure distribution in a domain and some of the interaction with silicone which is in my environment. I am kind of stuck here by finding a similar plate source. All comsol give are the point or edge source. How can I have a circular boundary source? Could you offer some ideas? For your project, did you do the transient acoustic interaction physics? By the way, I am trying to download your files from comsol exchange, however, my 4.1 version cant open any of them. Do you happen to have 4.1 version? Thanks a lot. Best Xu

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 15.10.2015, 15:41 GMT-4
Hi Robert,

I am facing a problem similar to your.

I am simulating a horn response.
I am trying use a curve, based from excel data, as Incident Pressure Field.
I got the data from excel and it generated a graphic that I can see in the results, but I am not capable to use it exciting my model.

Do you have a suggestion to solve my problem?

Thanks,
Best regards,
Hi Robert, I am facing a problem similar to your. I am simulating a horn response. I am trying use a curve, based from excel data, as Incident Pressure Field. I got the data from excel and it generated a graphic that I can see in the results, but I am not capable to use it exciting my model. Do you have a suggestion to solve my problem? Thanks, Best regards,

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.