Frequency correlation between two non stationary signals

Click For Summary

Discussion Overview

The discussion revolves around analyzing the correlation between torque and speed measurements from a shaft operating at variable speeds. Participants explore methods to verify the relationship between the frequency of torque fluctuations and the shaft speed, particularly in the context of non-stationary signals.

Discussion Character

  • Exploratory, Technical explanation, Debate/contested

Main Points Raised

  • One participant suggests using signal statistics functions to verify the relationship between torque fluctuations and shaft speed.
  • Another participant proposes using a different model or eliminating a variable in the experiment to address the non-stationarity of the signals.
  • A participant questions whether the speed variation is periodic, suggesting that if it is, a FFT could be applied to identify frequency peaks related to speed variation.
  • Another participant mentions the possibility of using short-time Fourier transform (STFT) or wavelet transform for time-dependent analysis of the signals.
  • A later reply clarifies that the speed variation is not periodic and expresses intent to try STFT for analysis.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the best approach for analyzing the non-stationary signals, and multiple competing views on methodologies remain present.

Contextual Notes

The discussion highlights the challenges of analyzing non-stationary signals and the dependence on the nature of the speed variation, which is confirmed to be non-periodic.

serbring
Messages
267
Reaction score
2
Hi all,

I have measured the torque and the speed of a shaft when running at a variable speed. From the measurements, torque fluctuations are in the torque signals and I have the feeling that the fluctuations frequency is the shaft speed. I need to verify this with some signal statistics functions. In case the shaft speed would have been constant, I would have computed the PSD of the torque signal and I would have compared its peak frequency with the shaft speed. But what should I do with two non stationary signals?

Thanks
 
Engineering news on Phys.org
(a) use a different model
(b) eliminate a variable in the experiment
 
Is the speed variation itself periodic?

If yes, you can still use FFT on the whole signal and you will just get another peak for the frequency of speed variation.

You could also try a short-time Fourier transform (STFT) or wavelet transform if you wish to analyze the signals as a function of time as well.
 
Hi Boneh3ad,

thanks for youre reply. The speed variation is not periodic. I may try the STFT, thanks
 
Last edited:

Similar threads

How to Measure Torque Required for motor to spin a rotating shaft?
  • · Replies 3 ·
Replies
3
Views
5K
Torque calculations: Rotating vertical shaft
  • · Replies 5 ·
Replies
5
Views
5K
Can an electric motor and gearbox control the speed of a rotating shaft?
  • · Replies 17 ·
Replies
17
Views
2K
What's the most likely cause for this carbon seal crack?
  • · Replies 8 ·
Replies
8
Views
1K
Engineering How Can MATLAB Be Used to Analyze Signal Duration and Frequencies Using DFT?
  • · Replies 1 ·
Replies
1
Views
3K
Calculating delay b/w signals using correlation
  • · Replies 10 ·
Replies
10
Views
4K
Two DC-motors acting against eachother
  • · Replies 6 ·
Replies
6
Views
4K
Undergrad Is energy really conserved?
  • · Replies 39 ·
2
Replies
39
Views
4K
Centrifugal blower ringing?
  • · Replies 22 ·
Replies
22
Views
4K
How is torque divided between multiple wheels in a system with friction?
  • · Replies 1 ·
Replies
1
Views
4K