Relationship between structures Natural Frequency and Contact Mechanics 5

[yaston4]

I understand that techniques such as FEA can be used to the mode shapes and natural frequencies of structures. I am wondering if finding out the mode shapes and natural frequencies can help in providing any information or inclination of the the materials ability to cope with contact conditions? I understand that contact conditions depending on the properties and geometry of both (or more) bodies in contact. However, I am think purely on based on comparing the same structure with different material properties. I would really appreciate any discussion and input on this. Thanks.

 

[dhengr]

Yaston4:
Why don’t you explain to us how you think the two different analyses might be remotely related? Obviously, if you had sufficient vibration so as to eliminate contact for periods/intervals of time you would have some sort of a peening action rather than constant contact. If you had dynamic actions which caused the two bodies to move w.r.t. each other, at the contact point, you would have wear, fretting and the like. Otherwise, it seems to me that they are two pretty distinct concerns and analyses. I don’t get the connection you are imagining.

 

[drawoh]

yaston4,

Are the two structures touching each other or are they clamped together?

Is your FEA capable of understanding contact forces in compression, but not tension? Can it deal with separation of the two structure?

--
JHG

 

[IRstuff]

Generally, loose contact is bad. Each object has its own natural frequency, and if they're touching but not bolted together, they can sometimes wind up resonanting in a bad way.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[electricpete]

In theory, if you have a modal model, you can use it to predict response to applied force at any frequency. If you are trying to figure out the deformation under static load, that would be the response at zero frequency, so you could find it. But you would be going to a lot of extra work to develop dynamic solution from mass, stiffness, and damping characteristics, when all you need is static solution from stiffness characteristics. I may be missing your point.

=====================================
(2B)+(2B)' ?

 

[yaston4]

Hi all,

I am just brainstorming here and trying to think outside the box. All you comments and input is very much appreciated. I will try to give an example. Lets say we have a structure with a very complex range of material properties i.e. each element of the FE model has assigned a different material property (Youngs modulus, Poisson's ratio and density). If then it becomes very difficult to create an assembly of and conduct a contact analysis then how else could we assess the structures performance under contact in relation to the same structure with different material properties. As suggested by electricpete we could compare the stiffness of each structure.

dhengr: Sorry, I do not have a sensible answer to your question.
drawoh: The components would not be clamped, but touching e.g. a ball in socket joint. In terms of contact forces, it would be conducted using the contact algorithms in Abaqus.

I really enjoyed reading your kind comments and input, thank-you again.

 

[drawoh]

yaston4,

If all the components in your structure have different Young's Moduli and Poisson's ratios, you just punch the values into your software. Your fabricator is going to hate you. I see no problems doing analysis.

A ball and socket joint is fully clamped in a truss. If a beam is free to rotate, I don't know how FEA can analyse it.

--
JHG

 

[GregLocock]

If the system has non linear constraints such as contact then it does not have natural frequencies. It does still exhibit resonant behaviour, but the frequencies of maximum response amplitude will depend on the amount of force used, and where it is applied. If your FEA program says it can give you a table of natural frequencies and mode shapes even with contact present then burn it and write your own.

There are ways around this - you could consider a 50% duty cycle contact as being a bit like a compliantly mounted pin joint, but that would be very application, and excitation specific. Generally contacts create 3rd, 5th etc harmonics which of course a simple k/m analysis can't even begin to consider.





Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[yaston4]

Thanks drawoh and GregLocock, some really important points in your posts and lots to think about...

 

[spongebob007]

You can't solve for mode frequencies and mode shapes with contact conditions. The mathematical algorithms behind eigenvalue solvers require that the stifness matrix be constant. Basically no nonlinearities can exist. If you set up an FEA model with contact and run a modal analysis, most codes will simply ignore the contact conditions. You can't even use modal methods of transient analysis because again, these methods require that the eigenvectors are orthogonal and linearly independent. Nonlinear dynamic problems must be solved using direct integration. Basically the entire FEA model is solved at every time step.

Many years ago I learned of a way of faking contact by using constraint equations between the nodes on the two contacting surfaces. I don't remember too much about the details thoug since I have never used it again since then,

However, if your contact problem involves static loads you can simply run a static analysis with contact.

 

[yaston4]

Thanks spongebob007, your post is very helpful.