Getting model from frequency response

[sweetday]

Hi All,

I need to know how can we get a mathematical model(state space or transfer function) from frequency response?
The frequency response is resulted from FFT analysis (Force vs. Acceleration.
Thank you in advance for any reply.

sweetday

 

[IRstuff]

Is this for a class?

TTFN

 

[CESSNA1]

SWEETDAY: The FFT results in "amplitude verses frequency", not force vs. frequency, where amplitude is typically acceleration. This is obtained hy doing the FFT on a time domain plot of acceleration. If you truely have the FFT you can do an inverse FFT and get back to the time domain out to the filter frequency. The amplitude of the higher frequencies past the filter will be reduced and will probably be aliased. If you have a PSD you can do an inverse PSD and get a time domain plot that may or may not be what you want.

Regards
Dave

 

[GregLocock]

In the experimenatl modal analysis world a frequency response function, or FRF, is usually a transfer function of acceleration/force vs frequency. (velocity/force, diplacement/force and all their inverses are also used)

It is obtained from the complex spectra of the two signals in a variety of ways, known as H1 H2 H3 and HV (the latter being an engineer's joke by a Mr Harvard Vold).

H1 = (BxA*)/(AxA*)

where B and A are the original complex spectra of each channel, and * means conjugate.

Each of these H formulations has its uses, as they are snsitive in different ways to the noise in each of the data channels. Obviously H1 is more sensitive to errors in A than B for example.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[GregLocock]

Perhaps you are asking how to get a set of eigenvectors (modeshapes) froma set of FRFs.

There are many ways, get Ewins book.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.