Theoretical Study of Rolling Bearing Defect Condition Monitoring Techniques

Abstract

Rolling bearing is an important part that is often used in rotating machinery.Rolling bearings faults are one of the main causes of breakdown of these machines. Therefore, the fault diagnosis of rolling bearing is very important to guarantee the production efficiency and plant safety. This document presents different techniques for rolling bearing diagnosis among that the frequency analysis include enevelope analysis using (Fast fourier tronform) FFT and Singular value decomposition (SVD) , In the category of time-domain analysis technique there are STFT, Continuous wavelet transform(CWT), Spectral kurtosis,Time domain include EMD decomposition. In addition A brief introduction of different AI algorithms is presented including the following methods: k-nearest neighbour K-NN), naive Bayes (NB), support vector machine(SVM) , artificial neural network (ANN) and fuzzy neural network. Finally deep learining methods include Convolutional neural network (CNN)and Cyclic Spectral Coherence (CSCoh).