Random Vibration Mass Participation Factors

[jvian]

Hello,

I am trying to run a random vibration analysis in pro/mechanica and have ran into a question about mass participation factors. I understand the need for a high mass participation factor to determine global modes over local modes but I do not understand why an increased number of modes should be used for the random analysis when the frequency range of interest is so much lower than the frequencies obtained. My preliminary modal analysis produces only a limited number of modes with a low (<35%) mpf. Increasing the number of modes did increase the mpf (>80%) but then went way beyond (~3x) the desired frequency range of the psd.

If my frequency range of interest is 10 to 2500 hz, then why should modes of upwards of 4000 hz be included in the analysis? If I use the higher modes because they have or contribute to a higher mpf, doesn't that claim that the lower modes in the frequency range of interest do not participate globally and therefore are not of much concern?

I am trying to understand this and am open to all forms of help and advice.

Thank you

 

[GregLocock]

"If my frequency range of interest is 10 to 2500 hz, then why should modes of upwards of 4000 hz be included in the analysis? "

Because the residual participation of those higher frequency modes will affect responses in your frequency range of interest.

Might be worth running a hand analysis (shock horror) on a cantilever to look at the innaccuracy you get by ignoring higher modes.

Typically in a vehicle dynamics model we include modes up to 60 Hz, even though our models have an upper frequency limit of 20 Hz.

I don't know about the wisdom of using MPFs as such an important criterion. If you have guidelines suggesting that approach I suppose you could try running it both ways and see what the difference is.


Cheers

Greg Locock


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

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

 

[jvian]

Thank you Greg.

Forgive me if I sound a bit uneducated as neither my experience nor anything that learned in school involve this subject. So far most research I have performed involving random vibration in pro/mechanica emphasized the importance of mpf's. If there is another approach other than the mpf's could you point me in a direction to research it? As far as the mpf's could it be described as a form of harmonic of the higher frequencies that is what is participating or residual in the lower frequencies or visa versa?

I appreciate the help and thank you again for responding.

 

[GregLocock]

Either you do an experiment, trying it with and without the extra modes, or a hand calc as I suggested. Or you do some heavy maths.

Cheers

Greg Locock


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

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

 

[mechengdude]

jvian,

I've asked that same question before and never been able to reconcile it completely in my head. I have run a few simple analyses as a test and do feel comfortable with your approach of leaving out the frequencies outside the area of interest.

As an example I modeled a 1" diameter steel bar 12" long. Restrained it at one end. Doing a modal analysis with only two requested frequencies I got about 60% MPF. Running a modal analysis with five requested frequencies I got about 80% MPF. Inputting a random vibration curve for both cases yielded stress of approximately the same magnitude.

Hopefully someone smarter than me can answer your question completely. Good luck.