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.

effect of temperature on the eigenfrequency

Please login with a confirmed email address before reporting spam

HI EVERYBODY

I'm new in Comsol and i would now how physics are used for simulate the effect of temperature on the eigenfrequency

the model is given

thanks for your help


5 Replies Last Post 29 mag 2012, 14:47 GMT-4
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 11 mag 2012, 05:51 GMT-4
Hi

it's not obvious to understand which influence you are looking for, your solid MEMS structure is in some air I see (you should not define air as "solid" as the corresponding material properties are not there, or you define something reasonnable: E=1/3 ambient pressure, nu= 0 and rho = 1 kg/m^2.

Eigenfrequency changes comes from i.e. stress buildup in your material due to thermal expansion, material property variations due to the temperature change, and in your case coupling with air that also has its "material" properties changing with the temperature

The easiest us to use coupled physics, pre-cooked by COMSOL, such as FSI and other combinations as you need T, u, U, p, or you need to mix up sourself.

I would still propose to start "simple" just in structural or TS thermal stress, forget first the air, get the model vibrating within the solid physics and see the effect of T, then you could consider the air load simply from the amplitude and air density as a surface virtual mass load.

Then you make a separate FSI model and see how the frequency changes with the air density (withot any T, or just a T in the air for controlling the air properties.

Then You should be ready for a full model, mixing everything, if you really mean this is worth it, as I suspect you will have enough data from the two first models to deduce correctly the "full" model

A model is a simplified representation of reality, it's good ingineering practices to your your physics and engineering skill to get to a complete results from two or more partial ones, withut setting up the "large" model

--
Good luck
Ivar
Hi it's not obvious to understand which influence you are looking for, your solid MEMS structure is in some air I see (you should not define air as "solid" as the corresponding material properties are not there, or you define something reasonnable: E=1/3 ambient pressure, nu= 0 and rho = 1 kg/m^2. Eigenfrequency changes comes from i.e. stress buildup in your material due to thermal expansion, material property variations due to the temperature change, and in your case coupling with air that also has its "material" properties changing with the temperature The easiest us to use coupled physics, pre-cooked by COMSOL, such as FSI and other combinations as you need T, u, U, p, or you need to mix up sourself. I would still propose to start "simple" just in structural or TS thermal stress, forget first the air, get the model vibrating within the solid physics and see the effect of T, then you could consider the air load simply from the amplitude and air density as a surface virtual mass load. Then you make a separate FSI model and see how the frequency changes with the air density (withot any T, or just a T in the air for controlling the air properties. Then You should be ready for a full model, mixing everything, if you really mean this is worth it, as I suspect you will have enough data from the two first models to deduce correctly the "full" model A model is a simplified representation of reality, it's good ingineering practices to your your physics and engineering skill to get to a complete results from two or more partial ones, withut setting up the "large" model -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11 mag 2012, 16:26 GMT-4
Thank you ivar

can you send me your idea of simulation in the some model???

cordially
Thank you ivar can you send me your idea of simulation in the some model??? cordially

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 12 mag 2012, 12:06 GMT-4
Hi

you have everthing, start without considering the big air box (you can use a delete entity node in the geometry to disable it, without destroying it so its easy to get it back), make an eigenfrequency scan with only Solid physics, then, make a new study with a stationary and an eigenfrequency in two steps (under the same study), and change the thermal reference temperature in the physics tab by perhaps 10°C w.r.t the initial temperature (or the opposite) then solve for both, by default COMSOL will solve the stress stationary case to preload your beam, then linearise that value and solve the eigenfrequency case. now you should see eigenfrequencies that have changed w.r.t the first case.

Then you need to think over how to add the "air load", and finally how to get the air load vary also with temperature

--
Good luck
Ivar
Hi you have everthing, start without considering the big air box (you can use a delete entity node in the geometry to disable it, without destroying it so its easy to get it back), make an eigenfrequency scan with only Solid physics, then, make a new study with a stationary and an eigenfrequency in two steps (under the same study), and change the thermal reference temperature in the physics tab by perhaps 10°C w.r.t the initial temperature (or the opposite) then solve for both, by default COMSOL will solve the stress stationary case to preload your beam, then linearise that value and solve the eigenfrequency case. now you should see eigenfrequencies that have changed w.r.t the first case. Then you need to think over how to add the "air load", and finally how to get the air load vary also with temperature -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 mag 2012, 13:52 GMT-4
Hi

I found this problem quite interesting. What I did it was replacing (or programing the functions) of all the parameters and properties in the materials that change with the temperature and may have influence in the resonance of the system. In the case of solids I had to replaced the Young's modulus functions (bulk modules in the case of fluids) some other properties may affect the system but depends pretty much what the "thing" does, remember that, if involved , many electrical and magnetic properties will change with temperature affecting the resonance of the model. I found that Viscosity, Speed of Sound, Thermal Expansion, Stress, etc. may affect a system in unexpected ways and is necessary to have certain previous information about the model to know what of those parameter may require to be modified with the change in the temperature. You may find that some of them will not have effect in the behavior of the system but very likely you will need to fix all the others first and move each one of them one by one to find out which ones require to be simulated ( and if your model is actually simulating them or being affected by them). Hope this ideas are useful 4 u.

Regards
Carlos
Hi I found this problem quite interesting. What I did it was replacing (or programing the functions) of all the parameters and properties in the materials that change with the temperature and may have influence in the resonance of the system. In the case of solids I had to replaced the Young's modulus functions (bulk modules in the case of fluids) some other properties may affect the system but depends pretty much what the "thing" does, remember that, if involved , many electrical and magnetic properties will change with temperature affecting the resonance of the model. I found that Viscosity, Speed of Sound, Thermal Expansion, Stress, etc. may affect a system in unexpected ways and is necessary to have certain previous information about the model to know what of those parameter may require to be modified with the change in the temperature. You may find that some of them will not have effect in the behavior of the system but very likely you will need to fix all the others first and move each one of them one by one to find out which ones require to be simulated ( and if your model is actually simulating them or being affected by them). Hope this ideas are useful 4 u. Regards Carlos

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 29 mag 2012, 14:47 GMT-4
Hi

I fully agree, it was also one of the outcomes of a multi-physics forum I participated to a months time ago:
We were several people there to state that the true limitation today, in multiphyiscs simulations, is the limited knowledge of our material models, and the multi-parameter dependence of most material parameters, which are not, or not well enough, known and available, today.

So material scientists, out there, plenty of work for you ;)

--
Good luck
Ivar
Hi I fully agree, it was also one of the outcomes of a multi-physics forum I participated to a months time ago: We were several people there to state that the true limitation today, in multiphyiscs simulations, is the limited knowledge of our material models, and the multi-parameter dependence of most material parameters, which are not, or not well enough, known and available, today. So material scientists, out there, plenty of work for you ;) -- Good luck Ivar

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.