Sunteți pe pagina 1din 4

Finite Element Analysis

12,709 members
Join
Information and settings
Discussions
Promotions
Jobs
Search
Have something to say? Join LinkedIn for free to participate in the conversation. When you join, you can comment and post your own discussions.

lalit
FEA modal analysis- what role mass participation factor play for resonance check, i.e while dealing component natural
frequencies with external excitation freq?.
lalit B.CAE, Noise & Vibration Analyst
1st query - For modal analysis of say single 3D componet, to avoid resonance condition, do we need to compare the component natural frequencies with external
excitation freq of say compressor only or we have to look the mass participation factor and coresponding mass ratio also.
2nd query- If component has natural frequency of 49 Hz (in 3rd mode) and compressor operates at 48Hz, but it has found that for 3rd mode the ratio of effective
mass to total mass in X-Y-Z direction is 0.066, 0.089, 0.067 resp. though the ratios are very less, can we call it that component is in resonance.
Note: the analysis is performed in ANSYS Mechanical.
I am seriously looking for it's answer. Please help me
Thanks
Lalit
Like
Comment (7)
Share
September 6, 2012
Comments
7 comments

Karam
Karam M.
Professor of Aeronautics and Mechanics
Natural frequencies depend on both mass and stiffness distributions in an elastic body. In your case, you should construct a Campbell diagram to check the occurrence
of resonance at the critical rotating speeds. I think you need to apply optimization theory to avoid resonance. You may look at the link:

http://www.intechopen.com/books/authors/advances-in-computational-stability-analysis/stability-dynamic-and-aeroelastic-optimization-of-functionally-graded-
composite-structures

Joshua
Joshua G.
Associate Professor of Mechanical Engineering
There are two issues to consider:

1. Resonance is defined as when the excitation frequency matches one of the natural frequencies of the structure. In this case, large amplitudes can result.

2. The "participation" factor is a measure of how much energy "gets in" to a mode, leading to larger amplitudes.

So, its possible that an excitation could match a natural frequency (i.e. resonance), but if the participation factor was 0, then no energy will get into that mode.

lalit
lalit B.
CAE, Noise & Vibration Analyst
Thanks Mr. Karam and Mr. Joshua for your valuable input.

With this one more query arises that for a same single 3D component, if its 13th mode has the ratio of effective mass to total mass as 0.18 (i.e 18%) and matches with
excitation frequency, but corresponding mode shape has 5 node, still we should consider the 13th mode as critical for resonance or we can neglect it.
Note: 3D component in this case say a small diameter tube of total length of 500m routed in 3D.

Also have big doubt in mind that a statement says "extract the mode upto mass ratio of 90%". what exactly it means in context to resonance. it is a cumulative mass
addition over the frequencies where as a resonance is a case of single frequency.

Eager to know the answers for above. waiting for the reply.
thanks

Joshua
Joshua G.
Associate Professor of Mechanical Engineering
I can't give you an answer about whether should include the mode or not, as
that's too important a question to be answered without having detailed
knowledge about the system and the excitations. However, what harm is there
in including it?

I think the guidance you were quoting means that you should include enough
modes such that the cumulative mass ratio is < or = 90%.
So you will be including a bunch of modes. Remember, that even if one mode
is "in resonance", other modes close in frequency will also contribute
substantially to the response. That is why we cant include JUST the
resonant mode. The total response is the summation of the individual modal
responses. One or more of the individual modal responses may be "large" (at
resonance or near to resonance), hence the participation factor gives
guidance about how many modes are needed to capture "most" of the modal
mass of the structure. Some of these modes may not contribute significantly
to response amplitudes and some will, but you will be "covered".

Josh

Karam
Karam M.
Professor of Aeronautics and Mechanics
Dear Sir

if you'r really worried about resonance occurrence, you should apply any reasonable optimization algorithm and consider what is called frequency windows
optimization. This may help you in separating all the important frequencies from the excitation ones and avoid large amplitudes to happen. You may need to see Eq.
(3) in the link:

http://www.ufpa.br/nicae/projetos/eletronorte_cabos/ref_pdf/Negm_2000a.pd

Blas
Blas M.
Director en IBERISA
Dear Lalit,
When the frequency of the exciting force coincides with a natural frequency of the structure, then resonant vibrations may eventually lead to failure. Not only
matching between the frequency of excitation and the natural frequencies of the structure is important, but also the MODE SHAPE is very important: the deformed
shape of mode & and the deformed linear static shape of the applied force vector should be similar. It is possible to have a situation where the stiffness force and the
inertia force are exactly equal and opposite, in this case they cancel each other, not matter their magnitud is, and the system is said to be in resonance.

The mass participation factor is usually used as a solution accuracy check to ensure that a suficient number of modes have been used in the dynamic analysis.
Generally, the number of modes considered must contribute to a total mass participation factor of at least 80% of the system mass in the direction of the base motion,
specially in base excitation problems. If there is still some doubt about the accuracy of the solution, increase the number of modes and rerun the problem and see
whether the convergence has been achieved.

In force exitation dynamic problems the integration increment or time step size (smaller than 1/10 of the last mode period) plays an important role in soluction
accuracy.

Best regards,
Blas.

Joe
Joe M.
CAE Structural Dynamics & NVH
Lalit,
You may only need to pay attention to mass participation factors in the direction (and cross rotations) in line with your compressor load direction. But ideally you
would want to run a modal frequency response analysis (ie: harmonic, sine sweep), applying the compressor load at a few critical excitation frequencies, using all
modes the first time. Its fairly easy to determine the dominant modes at those frequencies depending on your FE program. Most the FE codes have a method for
calculating dynamic modal participation factors (energy and/or individual response) for your given load at each of those frequencies, which are more telling than basic
modal effective mass. After 1 run with all the modes, then you can filter out unnecessary modes not excited by that load at those frequencies for later runs. The
dynamic energies will also tell you at what freqs the system is most resonant. You may find the energy ~48 Hz less than at other freq's (good), or some modes ~48 Hz
not excited by that particular load & direction. Unfortunately most commercial FE codes dont find it necessary to provide users with decent post-processing of useful
calculations like dynamic pfs into a nice table or chart, but the modal KE participation % calculation is quite simple: If your modes are mass-normalized to 1.0, then
for each excitation frequency, output the generalized modal response magnitudes q-mag(mode-I, f-excite), into Excel. ModalKEfrac(mode-i) = q-mag(mode-i)^2 /
sum(q-mag^2), producing an nmodes x f-excite table of modalKE fracs or %s. Its not that cumbersome and quite effective. Many FE codes also have individual
response dynamic pfs, which give modal participation (or contribution) of projected response to total response. In your case you might use the response of your load
pt. There would be an nmodes x f-excite projection fraction table for ea response, but that ones a little tougher to calculate on your own. Lastly, theres methods of
doing a combined static loads vs modes analysis on the structure, without going into frequency response, and computing modal KE% or SE% of modes to static
shapes, as a modern version of Modal Effective Mass. This is an intermediate calculation more commonly found within residual modes analysis, that can be applied to
free-free structures as well as constrained, and predicts the modes excited by each load, rather than just in the general 6 rigid body modal effective mass directions.
Unfortunately my developments in this area have been client-proprietary, but someone at your FEA client support may be able to assist.

Best regards,
-Joe M
Have something to say? Join LinkedIn for free to participate in the conversation. When you join, you can comment and post your own discussions.

S-ar putea să vă placă și