Exercise $3$

In this exercise, students need to use their knowledge of signal$-$based analysis methods to perform the analysis of the bearing fault data. The fault data set consists of $4$ sets:

$-3$ baseline conditions sets

$-3$ outer race fault conditions $($single load condition$)$

$-7$ outer race fault conditions $($multiple load condition$)$

$-7$ inner race fault conditions $($multiple load condition$)$

The student should select one type of fault $($outer race or inner race$),$ and apply the learned methods from lectures L$6$ and L$7.$ If some methods are not applicable please include a short comment why. Write a report that should include: MATLAB code used, analysis waveforms, findings, and conclusion.

The data set can be found here: Data set link: www$.$mfpt$.$org$/$fault$-$data$-$sets$/$

The following guidelines can be used in solving the exercise:

$1.$ Time domain analysis

Start with time signal analysis $($Amplitude envelope, RMS $,$Mean, Peak value, Kurtosis, Impulse factor, Crest factor, Clearance factor, Standard deviation, Skewness$).$ Observe how the values change for healthy and faulty bearings under different loads.

$2.$ Frequency domain analysis

Apply frequency domain analysis methods: DTF and spectrum analysis. What type of spectrum analysis should be used: PS or PSD$?$ Is the signal spectrum discrete or continuous?

$3.$ Time$-$frequency analysis

Apply the STFT and Wavelet transform. Play with the window size of STFT$,$ and observe the impact on the results. Can STFT successfully differentiate the frequencies?

$4.$ Advanced time$-$frequency analysis

Apply the HHT on the signal. Start first with identifying the EDM. How many modes are identified?

$5.$ Conclusion

What findings were obtained in each domain, and which methods are applicable for this bearing fault analysis? Why were the signals $($healthy and faulty$)$ recorded for different load conditions?