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 The information provided may be out of date.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.

Resonant pressure sensor

Please login with a confirmed email address before reporting spam

Dear All,

Could someone share a model/tutorial showing modeling of a resonant pressure sensor? I am able to add multiple study parameters (static and eigenfrequency) but unable to link both. Any similar model also would be helpful . OR if someone could suggest me some direction from my model it would be gr8 ! Please find the model as an attachment.

Best regards,

Ashwin Simha.


7 Replies Last Post 8 févr. 2011, 04:31 UTC−5
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 1 févr. 2011, 06:22 UTC−5
Hi

that is slightly tricky in v4.0, this are among the nice corrections/improvements in v4.1 so you should contact your rep to update asap :)

Now I do not have 4.0 anymore running so I cannot easily provide you with a model, and I do not really have the time neither. But why 2 physics ?
use one solid, where you fix and load your BC for the static case

then use only one study i.e. you use the default stationary but you add also thereafter an eigenfrequency (right-click - Others)

then you generate the default solution. In v4. it will propose by default a stationary sequence, a stored solution and an eigenfrequency linearised and starting from the stored solution. In v4.0a you need to construct it by hand if I remember right, just right click and it will add the options one by one (or set up a buckling case and modify it accordingly). Slightly tedious but you learn the sequences like that ;)

Then you solve and inspect. To be sure you have some stress stiffening you can add a study with only the eigenfrequency. Do not forget that body loads ()or any right hand side loads) are not considered by the eigenfrequency solver. That is why one need to add the stationary case to store the stress in the initial values of the stress tensor

--
Good luck
Ivar
Hi that is slightly tricky in v4.0, this are among the nice corrections/improvements in v4.1 so you should contact your rep to update asap :) Now I do not have 4.0 anymore running so I cannot easily provide you with a model, and I do not really have the time neither. But why 2 physics ? use one solid, where you fix and load your BC for the static case then use only one study i.e. you use the default stationary but you add also thereafter an eigenfrequency (right-click - Others) then you generate the default solution. In v4. it will propose by default a stationary sequence, a stored solution and an eigenfrequency linearised and starting from the stored solution. In v4.0a you need to construct it by hand if I remember right, just right click and it will add the options one by one (or set up a buckling case and modify it accordingly). Slightly tedious but you learn the sequences like that ;) Then you solve and inspect. To be sure you have some stress stiffening you can add a study with only the eigenfrequency. Do not forget that body loads ()or any right hand side loads) are not considered by the eigenfrequency solver. That is why one need to add the stationary case to store the stress in the initial values of the stress tensor -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 1 févr. 2011, 23:34 UTC−5
Hi Ivar,

I have the diaphragm and the beam suspended on the upper cavity of the diaphragm in my model. The reason for defining 2 physics is that I could separate the beam and the diaphragm as two domains. For the static analysis I include both the domains (beam and diaphragm) while for the eigenfrequency analysis I use the stored solution of the static case and apply it on the domain corresponding to the beam with appropriate BCs.

On the otherhand if I do not go for 2 physics then I end up running the eigenfrequency analysis for the entire sensor geometry which in any case is not desired.

Please let me have your views on the same.

Best regards,
Ashwin.
Hi Ivar, I have the diaphragm and the beam suspended on the upper cavity of the diaphragm in my model. The reason for defining 2 physics is that I could separate the beam and the diaphragm as two domains. For the static analysis I include both the domains (beam and diaphragm) while for the eigenfrequency analysis I use the stored solution of the static case and apply it on the domain corresponding to the beam with appropriate BCs. On the otherhand if I do not go for 2 physics then I end up running the eigenfrequency analysis for the entire sensor geometry which in any case is not desired. Please let me have your views on the same. Best regards, Ashwin.

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 2 févr. 2011, 01:08 UTC−5
Hi

indeed you have some good arguments there, provided that the "rest" can truly be considered as infinitely stiff. As a first approach it's certainly a good idea, but as eigenfrequency analysis are not very demanding on PC resources (compared to other non-linear runs, I believe it's worthwhile to try a full eigenfrequency analysis, just to see and try to understand what happens. In particularly if you have a pre-stressed system.

But I agree, full system as optional last test is reasonable

--
Good luck
Ivar
Hi indeed you have some good arguments there, provided that the "rest" can truly be considered as infinitely stiff. As a first approach it's certainly a good idea, but as eigenfrequency analysis are not very demanding on PC resources (compared to other non-linear runs, I believe it's worthwhile to try a full eigenfrequency analysis, just to see and try to understand what happens. In particularly if you have a pre-stressed system. But I agree, full system as optional last test is reasonable -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 févr. 2011, 01:45 UTC−5
Hello All,

Finally after some preliminary study I thought that creating sub model for the beam and using the displacements from global model in the eigenfrequency analysis of the beam part (not the entire sensor)may be of some help. Similar such example has been demonstrated for stress concentration analysis on a wheel rim of an automobile(in COMSOL tutorial models : wheel_rim.mphbin). It is just that in our case we need eigenfrequency of the prestressed beam. I hav made a similar model based on this but the eigenfrequency analysis gives an error 'singular matrix'. Could someone please help me fix this model. Please find the model attached for your reference.

Thanks and best regards,
Ashwin Simha.
Hello All, Finally after some preliminary study I thought that creating sub model for the beam and using the displacements from global model in the eigenfrequency analysis of the beam part (not the entire sensor)may be of some help. Similar such example has been demonstrated for stress concentration analysis on a wheel rim of an automobile(in COMSOL tutorial models : wheel_rim.mphbin). It is just that in our case we need eigenfrequency of the prestressed beam. I hav made a similar model based on this but the eigenfrequency analysis gives an error 'singular matrix'. Could someone please help me fix this model. Please find the model attached for your reference. Thanks and best regards, Ashwin Simha.


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 7 févr. 2011, 03:38 UTC−5
Hi

I'm not sure but suspect its a v4.0 versus 4.1 the version I'm using, that does not allow me to read correctly your model.

In v4.1 I would add both solver cases stationary then eigen frequency under the same study and link them in there, just as COMSOL does with a buckling study.

One thing with 4.0 it's somewhat weak in controlling the solver sequences, that rapidly get corrupted if you change them by hand too often. Most of the time one needed to delete the solver sequences and rebuild them from scratch. I suspect that several of these solver error messages in v4.0 and "a" are linked to that

--
Good luck
Ivar
Hi I'm not sure but suspect its a v4.0 versus 4.1 the version I'm using, that does not allow me to read correctly your model. In v4.1 I would add both solver cases stationary then eigen frequency under the same study and link them in there, just as COMSOL does with a buckling study. One thing with 4.0 it's somewhat weak in controlling the solver sequences, that rapidly get corrupted if you change them by hand too often. Most of the time one needed to delete the solver sequences and rebuild them from scratch. I suspect that several of these solver error messages in v4.0 and "a" are linked to that -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 févr. 2011, 03:54 UTC−5
Hi Ivar,

As suggested by you I have tried using both the solvers (eigenfrequency and static) in the same study and selected 'fully coupled '. I just followed one of the perfectly working 'stress stiffening 'examples found in this forum. But when I run an eigenffrequency analysis for different applied pressure (at the bottom of the diaphragm ) I do not see any shift in the resonant frequency. Please find attached the model for your reference. Could you please help fix this?

And a small question before I close : Since the beam is a part of the sensor structure and eigenfrequency of the beam is the desired output, do we need to use general extrusion in the static study and make use of these mesh displacements in the eigenfrequency study?

Many thanks !

Best regards,
Ashwin Simha.
Hi Ivar, As suggested by you I have tried using both the solvers (eigenfrequency and static) in the same study and selected 'fully coupled '. I just followed one of the perfectly working 'stress stiffening 'examples found in this forum. But when I run an eigenffrequency analysis for different applied pressure (at the bottom of the diaphragm ) I do not see any shift in the resonant frequency. Please find attached the model for your reference. Could you please help fix this? And a small question before I close : Since the beam is a part of the sensor structure and eigenfrequency of the beam is the desired output, do we need to use general extrusion in the static study and make use of these mesh displacements in the eigenfrequency study? Many thanks ! Best regards, Ashwin Simha.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 8 févr. 2011, 04:31 UTC−5
Hello All,

I have now used one of the working models of ' stress stiffening' and adapted to my structure. What I see is that the resonant frequency of the entire sensor structure(beam + diaphragm) varies with the applied pressure. This is a positive development. But what I would like to see is the variation of the resonant frequency of the beam alone with respect to the applied pressure. Could someone please help me to fix this problem. Please find the file attached for your reference.

Best regards,
Ashwin Simha.
Hello All, I have now used one of the working models of ' stress stiffening' and adapted to my structure. What I see is that the resonant frequency of the entire sensor structure(beam + diaphragm) varies with the applied pressure. This is a positive development. But what I would like to see is the variation of the resonant frequency of the beam alone with respect to the applied pressure. Could someone please help me to fix this problem. Please find the file attached for your reference. Best regards, Ashwin Simha.

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.