0
$\begingroup$

I ask you to help me to interpret this figure corresponding to a diagnosis of gear failure using CWT (Haar); enter image description hereI have already worked on the analysis of the vibratory signals using the FFT; But since the signals from the bearings and the gears are non-stationary it is considered necessary to apply WT. So I started with simulated signals of gear failures to evaluate the sensitivity of this method compared to that of Fourrier

$\endgroup$
  • $\begingroup$ Thank you for your help and for the video. My problem is not to understand the principle of wavelets but especially to interpret them. For example in the figure I didn't know to indicate that the gear is defective ?! $\endgroup$ – Bouchra Jan 25 '17 at 8:23
  • $\begingroup$ Since the signal is oscillatory, Haar may not be the best choice. You could try higher-order complex wavelets, and display the phase scalogram as well. Then, you can start comparing displays for both a nicely working and a failing one. What kind on failure are you looking for? $\endgroup$ – Laurent Duval Oct 22 '17 at 11:21
0
$\begingroup$

I've been working on this area for a while. The wavelet transform is a more powerful tool for time-frequency (TF) & feature extraction since the FFT or STFT always suffers from the resolution issues in TF domain. As a starter, I think the short videos provided in MATLAB is a very good point to start with. I paste the 1st part here and you can also find the rest parts from it.

https://www.mathworks.com/videos/understanding-wavelets-part-1-what-are-wavelets-121279.html

Bests, GL

$\endgroup$
  • $\begingroup$ I think these video can help you understand this figure. Anyway, the major difference between FFT and CWT is that FFT is frequency domain based study and CWT is time-frequency domain based study. In your problem, the gear failure will lead to a frequency change/nonlinearity, and with the two coefficients in your second plot, you can localize when the failure occurs in the time domain. Detailed explanation for these two coefficients are in the videos. $\endgroup$ – Guoyi Li Jan 25 '17 at 19:21

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.