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 The information provided may be out of date.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.

Vibration of a plate

Please login with a confirmed email address before reporting spam

Hi,
I'm trying to model bending vibration in a plate. The problem is that I want my coundary edges to be controlled by "normal" and rotational springs. I cannot use spring-like material at the edges since I want the rotational stiffness to be independent of other properties. Is it possible to model the springs using mathematical boundary conditions, e.g. displacement being proprtional to some time derivative of the same displacement?

Regards
Christoffer

5 Replies Last Post Mar 3, 2011, 10:18 a.m. EST
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 Feb 22, 2011, 5:03 a.m. EST
Hi

I have given a few examples already on other threads, and I believe there is more in some of the models I have put on the model exchange (done for 3.5a but portable to v4)

basically if you give a boundary load value along Fx[N] = -k[N/m]*u[m] where k is your spring constant your done.

But you need to define one spring constant per direction.

Then for the rotations it's a bit trickier as you do not have directly access to the rotations in "solid" (only in shell or beam). One way is to use the RBC's (Rigid Body Constraints) but that means you are also saying your boundaries are rigid, another is to define the rotations wither with the mall deformation angle or by more advanced means as used in the RBC's. More on that in the other threads and in the examples already mentioned

Do not forget, most fields where on usually enter numerical values, can as well (or even better) be populated by equations, depending on your different variables (COMSOL's internal or your own). This is one of the unique and specific advantages of COMSOL over all other FEM programmes I know about

--
Good luck
Ivar
Hi I have given a few examples already on other threads, and I believe there is more in some of the models I have put on the model exchange (done for 3.5a but portable to v4) basically if you give a boundary load value along Fx[N] = -k[N/m]*u[m] where k is your spring constant your done. But you need to define one spring constant per direction. Then for the rotations it's a bit trickier as you do not have directly access to the rotations in "solid" (only in shell or beam). One way is to use the RBC's (Rigid Body Constraints) but that means you are also saying your boundaries are rigid, another is to define the rotations wither with the mall deformation angle or by more advanced means as used in the RBC's. More on that in the other threads and in the examples already mentioned Do not forget, most fields where on usually enter numerical values, can as well (or even better) be populated by equations, depending on your different variables (COMSOL's internal or your own). This is one of the unique and specific advantages of COMSOL over all other FEM programmes I know about -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 22, 2011, 10:52 a.m. EST
Thanks for quick answer!
I am using v4.0. I tried to model the rotational spring by adding an edge moment load

M=kR*d(w,x)

where kR is constant rotational stiffness and d(w,x) is supposed to be the x-derivative of the displacement normal to the plate.

I wrote this in the field marked with a y to the right since I guess it means moment about the y-axis. I could not find any information about how the moment is defined in the user's manual so I didn't know what sign to use. I guess that forces are defined positive in the positive direction of the axis and that the z-axis points out from the screen.

Solving for the eigenfrequencies I noticed that the moment load did not change the eigen frequency no matter the value of kR.

Any ideas of what's going wrong?
Thanks for quick answer! I am using v4.0. I tried to model the rotational spring by adding an edge moment load M=kR*d(w,x) where kR is constant rotational stiffness and d(w,x) is supposed to be the x-derivative of the displacement normal to the plate. I wrote this in the field marked with a y to the right since I guess it means moment about the y-axis. I could not find any information about how the moment is defined in the user's manual so I didn't know what sign to use. I guess that forces are defined positive in the positive direction of the axis and that the z-axis points out from the screen. Solving for the eigenfrequencies I noticed that the moment load did not change the eigen frequency no matter the value of kR. Any ideas of what's going wrong?

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 Feb 22, 2011, 11:07 a.m. EST
Hi

you should get hands on a 4.1, there are many improvements absolutely worth it !

You need to define correctly your curl for the rotations, check your formulas, it could be that in a particular set-up your equation is OK, not sure, it's certainly not a general acceptable one ;)

For the eigenfrequencies, these do never take into account external forces, "by design" so they should have no influence. If you want to add the spring values you need to add the equations for the "iomega" and densities

--
Good luck
Ivar
Hi you should get hands on a 4.1, there are many improvements absolutely worth it ! You need to define correctly your curl for the rotations, check your formulas, it could be that in a particular set-up your equation is OK, not sure, it's certainly not a general acceptable one ;) For the eigenfrequencies, these do never take into account external forces, "by design" so they should have no influence. If you want to add the spring values you need to add the equations for the "iomega" and densities -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Mar 3, 2011, 7:57 a.m. EST
Hi again
I have v4.1, sorry for the typo. I am pretty sure the eigen value solver do take edge loads reacting to displacements normal to the plate into account. I define my edge force load as

Fz = -kT*w

which seems to work fine. I want to model roatational stiffness by applying a moment proportional to the rotation angle of the plate. I approximate the angle by the first space derivative (applicable for small angles). For examaple the moment load on the left side of the plate about the y-axis at x=0:

My = kR*d(w,x)

Since I cannot find any information on moment loads in the user's guide and no definitions of moments and rotation angles, the sign and other stuff might be incorrect. However, this approach does not seem to work at all.

I need to verify my model by comparing eigen values to analytical solutions.

I do not want to use rigid body connectors since I want to stay in 2D.

How can I solve this?

/Christoffer

Hi again I have v4.1, sorry for the typo. I am pretty sure the eigen value solver do take edge loads reacting to displacements normal to the plate into account. I define my edge force load as Fz = -kT*w which seems to work fine. I want to model roatational stiffness by applying a moment proportional to the rotation angle of the plate. I approximate the angle by the first space derivative (applicable for small angles). For examaple the moment load on the left side of the plate about the y-axis at x=0: My = kR*d(w,x) Since I cannot find any information on moment loads in the user's guide and no definitions of moments and rotation angles, the sign and other stuff might be incorrect. However, this approach does not seem to work at all. I need to verify my model by comparing eigen values to analytical solutions. I do not want to use rigid body connectors since I want to stay in 2D. How can I solve this? /Christoffer

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 Mar 3, 2011, 10:18 a.m. EST
Hi

I would rather use the curl representation for the small angles

ThX 0.5*(vZ-wY)[rad] Rotation x
ThY -0.5*(uZ-wX)[rad] Rotation y
ThZ 0.5*(uY-vX)[rad] Rotation z

pls check the signs I might have got it wrong, this is for "solid" by default with the "spatial frame driven by the displacement"

I have already given a few reference books about this on the Forum, try a search

--
Good luck
Ivar
Hi I would rather use the curl representation for the small angles ThX 0.5*(vZ-wY)[rad] Rotation x ThY -0.5*(uZ-wX)[rad] Rotation y ThZ 0.5*(uY-vX)[rad] Rotation z pls check the signs I might have got it wrong, this is for "solid" by default with the "spatial frame driven by the displacement" I have already given a few reference books about this on the Forum, try a search -- 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.