How to interpret peaks in FRF?


I am a new NVH Engineer working with LMS test.lab.


These days I am working on natural frequency measurement of a component using Impact Testing. 


Sometimes I use to get adjacent peaks or abnormal peaks as shown below. I get confused whether I should choose the first peak or second in such cases.



What is the meaning of this type of peak in a FRF measurement?   What is the modal frequency?


Also, is the phase information relevant?  Can I use it to decide?


Looking forward to all replies. 


Re: How to interpret peaks in FRF?

Siemens Phenom Siemens Phenom
Siemens Phenom

Hello  ,

Those all could be possible resonances.  I would suggest plotting all the FRFs together to see if some measurements have larger peaks than others.  It's possible that the location you're displaying is not well excited at that resonance.  Also, sometimes we see small peaks when measuring the response in X or Y and exciting in Z although that does not seem the case here.  Sometimes we see this with a 
boundary condition mode (bungee, air bladder, tire or structure you hanging the component from).  By overlaying all the measurements you will hopefully gain some insight.


The phase would shift 180 degrees as you pass through a resonance.  You can also look at the driving point measurement to identify resonances if the resonances are all excited at that location.  These articles by Dr. Peter Avitabile at the University of Massachusetts Lowell are an excellent resource for the new NVH Engineer: 

Dr. Avitabile states:  "Some important characteristics of a drive point measurement are:


  • all resonances (peaks) are separated by anti-resonances
  • the phase looses 180 degrees of phase as we pass over a resonance and gains 180 degrees of phase as we pass over an anti-resonance
  • the peaks in the imaginary part of the frequency response function must all point in the same direction"

Re: How to interpret peaks in FRF?

Siemens Phenom Siemens Phenom
Siemens Phenom

Some other ideas to decide if these are real modes or not:


1. When viewing FRFs like this, that contain smallpeaks adjacent to large peaks, try switching the Y-axis to imaginary format (right click on Y axis -> Format -> Imaginary).  If there are two modes, there is a chance that the peaks will split – one going up, one going down.  This would be a strong indication that there are two modes present.




2. Do both peaks appear in consistently in multiple FRFs acquired on different locations of the structure?  Or did it appear in just one FRF collected?  If it is in just one FRF, perhaps there was a problem with that individual measurement (someone left a coffee cup on the structure that create a temporary second mode for example).


3. Is this the sum of several FRFs, or just one FRF from the structure?  If it is the sum, and FRFs were collected in different sets, than this might be the same mode just shifted in frequency. See this Multi-Run Modal article for more information.


4. Calculate a mode shape for both peaks. Are the shapes different?  Then this could be separate distinct mode.