What is the Eigenfrequency mean in vibration?

[HunterYeoh]

Hi,

I have a simple question to ask: What is the Eigenfrequency mean in vibration? I am confused and can anyone here explain to me or provide me some useful link?

I am trying to use comsol to analyse the vibration frequency from a plate, which was mount on top of 4 camshafts. These camshafts will rotate and hence the plate sitting on top will vibrate. If anyone is familiar with comsol, Pls help me by teaching me which mode i should use.

Many Thanks.

 

[minger]

In a modal and buckling analysis, you have a problem that looks something like:

$$[K]\left\lbrace\phi_i\right\rbrace = \lambda_i [M]\left\lbrace\phi_i\right\rbrace$$

Where [M] is the structural mass matrix, [K] is the stiffness matrix. For i modes, the eigenvalues represent the the ith mode, and the eigenvector represents the "shape" of that mode.

The eigenfrequency that you're referring to is probably the ith mode.

 

[coreyh]

Eigenfrequency basically means the "characteristic" frequency. "Eigen" comes from German and means "own", so i guess "own-frequency" is more literal but sounds dumb. As for the Eigenfrequency of a camshaft in the context of vibration, i think this would refer to the rotational frequency at which you would have the largest amount of vibration, which could also be called the resonant frequency. In the design of the camshaft, this is something you want to avoid during operation. So you design it to make this frequency higher than the max speed required.
I have used comsol quite a bit, but not so much for structural mechanics. However, i do know there is a model in the library which simulates the vibrational modes of a crankshaft. This would be a good place to start. As for making a model of four rotating camshafts with a plate on top, i think this will be very challenging, if not impossible. The biggest problem is modeling the contact between the cams and the plate, which you haven't really explained in your post. As a start, i would try and make up functions to describe the forces imposed on the plate by the camshaft and then impose them as periodic boundary conditions onto the plate. Then you should be able to solve it as an eigenvalue problem. This is one of the solver options that you can select.

Good luck!

 

[Dr.D]

As coryeh says, the eigenfrequency is the "characteristic" frequency in the sense of the frequency that characterizes the system, the frequency that is natural or characteristic of the system.

 

[minger]

You dont' really need to do a full transient test with contact elements. Basically, just do a modal analysis of the plate. That gives you the numbers. Then multiply that by how many times the cam lobes "hit" the plate per revolution. For example, if the four lobes are all 90° out of phase, then multiply by four. This gives you essentialyl "exitations per revolution". From here you can convert to frequency to make it easier to compare to the results.

Anyways, the modal analysis will give you the natural frequencies, neglecting any type of damping. You then just dont' want to be there, you don't need any other type of analysis unless your design forces you to run close.

 

[FredGarvin]

You essentially have a rectangular plate that is clamped on all sides. There are methods for calculating the natural frequencies of this scenario, like the Rayleigh method. I would start there. Check out this link. It is to a book that lists a sample calculation including some varying boundary conditions.

http://books.google.com/books?id=xl...X&oi=book_result&resnum=5&ct=result#PPA553,M1

 

[HunterYeoh]

Eigenfrequency basically means the "characteristic" frequency. "Eigen" comes from German and means "own", so i guess "own-frequency" is more literal but sounds dumb. As for the Eigenfrequency of a camshaft in the context of vibration, i think this would refer to the rotational frequency at which you would have the largest amount of vibration, which could also be called the resonant frequency. In the design of the camshaft, this is something you want to avoid during operation. So you design it to make this frequency higher than the max speed required.
I have used comsol quite a bit, but not so much for structural mechanics. However, i do know there is a model in the library which simulates the vibrational modes of a crankshaft. This would be a good place to start. As for making a model of four rotating camshafts with a plate on top, i think this will be very challenging, if not impossible. The biggest problem is modeling the contact between the cams and the plate, which you haven't really explained in your post. As a start, i would try and make up functions to describe the forces imposed on the plate by the camshaft and then impose them as periodic boundary conditions onto the plate. Then you should be able to solve it as an eigenvalue problem. This is one of the solver options that you can select.

Good luck!

Hi, Thanks for the reply. Can I ask you how can I do the rotation? I am new to comsol. I try to apply the rotational force on the shaft so it can turn and i will analyses the frequency. But fail to do it. and Is there possible to see the rotation animation just like in Cosmos or Ansys? Thank you

 

[HunterYeoh]

You dont' really need to do a full transient test with contact elements. Basically, just do a modal analysis of the plate. That gives you the numbers. Then multiply that by how many times the cam lobes "hit" the plate per revolution. For example, if the four lobes are all 90° out of phase, then multiply by four. This gives you essentialyl "exitations per revolution". From here you can convert to frequency to make it easier to compare to the results.

Anyways, the modal analysis will give you the natural frequencies, neglecting any type of damping. You then just dont' want to be there, you don't need any other type of analysis unless your design forces you to run close.

Thanks for the reply. I have difficulty in applying the rotational force to the shaft. Can you tell me how to do it? thank you.

 

[HunterYeoh]

You essentially have a rectangular plate that is clamped on all sides. There are methods for calculating the natural frequencies of this scenario, like the Rayleigh method. I would start there. Check out this link. It is to a book that lists a sample calculation including some varying boundary conditions.

http://books.google.com/books?id=xl...X&oi=book_result&resnum=5&ct=result#PPA553,M1

Thank you for your reply. It is very helpful

 

[saharsiamak]

please helpe me!I'm new in comsol I want to simulate a RF filter that includes 2 IDT and some electrodes in comsol but I can't get accurate response:(i want to analyse at frequency domain with some condition boundaries but osculate with a error but at eigenfrequency analyse not be problem so I have some question:1)at eigenfrequency analyse at comsol how many responses can we get?and about frequncy domain?2)we want to get resonant frequncies of my model can I get the curve of it or admittance matrix at comsol?