Vibration Analysis Methods in Bearing Damage Detection

Abstract

The goal of vibration analysis is to extract features from the measurements x (t) or x (kT0), T0 sampling time, to be used for fault detection and diagnosis. The corresponding analysis methods operate in the time domain (autocorrelation functions, power cepstrum, crest and kurtosis factors), in the frequency domain (Fast Fourier Transform FFT, maximum entropy spectral estimation, envelope analysis) and recently in the time-frequency domain. Time-frequency analysis methods represent efficient alternatives to detect bearings damage due to non-stationary impact signals. In this case, the short-time Fourier transform (STFT) or the Wavelet transform are applied. The paper presents a comparative study of some methods used in bearing faults detection (the envelope spectrum and the wavelet transform). A monitoring tool acquires experimental data from a bearing vibration control test rig. An accelerometer captures the signal from the bearing outer ring then it is processed using PCI-4451 National Instruments data acquisition board and LabVIEW soft.