0
$\begingroup$

I am analyzing vibration data collected by an accelerometer positioned in a stamp press machine. The fundamental frequency is around 10 Hz. The problem here is that when I take the FFT of the data I see harmonics of the fundamental frequency until 5000 Hz every 10 Hz (of course). I would expect some visible harmonics in low-frequency but with a decaying behavior until, at some point, it disappears. I attached 2 figures. The first one with a zoom from 0 to 1000 Hz and the other from 4000-5000 hz. Note that the Harmonics dominate the signal.

enter image description here

enter image description here

And here is a sample of the data:

enter image description here

I have being trying to convince myself that is normal but I can´t. What do you guys think? Can it be an effect of something that I should discard?

Thank you!

$\endgroup$
  • $\begingroup$ How is the accelerometer attached to the press? $\endgroup$ – A_A Jun 28 '18 at 16:12
  • $\begingroup$ Hey A_A, I don´t have this information yet. I will ask. but how the mount can influence the harmonics of my signal? $\endgroup$ – Hululu Jun 28 '18 at 16:36
1
$\begingroup$

Well, looking at your time-domain signal, it looks a bit similar to a pulse train. And the Fourier Transform of a Pulse train with distance T is a Pulse train with distance 1/T

Hence, if you see a pulse every 10Hz, this corresponds to pulses every 0.1seconds in the time domain. That's what the time-domain signal looks like.

$\endgroup$
  • $\begingroup$ Thanks for the answer Maximilian. So, do you agree that in order to analyze the "true" frequency content of my signal I will have to perform the fft in smaller chuncks of data where the impacts don´t look like impulse functions? $\endgroup$ – Hululu Jun 28 '18 at 17:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.