Ball bearing fault classification and diagnosis techniques: a review

In the early stages detection of faults in antifriction bearing in health condition monitoring of bearing is a very cumbersome task. The defects in bearing unless detected in time may lead to downtime of setup. The main cause of rolling element defects are following ; inappropriate design of the bearing or inappropriate manufacturing or mounting, bearing races misalignment, uneven diameter of rollers, improper lubrication, wear, fatigue, overloading. The objective of this paper is to present a detailed overview of the different detection techniques used for measuring rolling bearing defects. This paper presents an overview on various detection techniques used for measuring ball bearing defects; they may be majorly classified as vibration measurement, acoustic measurement, temperature measurement. Vibration and AE signal analysis techniques like time, frequency, time-frequency domain, fast and accurate techniques like ANN are focused study areas. Effectiveness of three ANN classifier MLP, RBF, PNN are also studied. Latest topic of interest like KNN, CNN, AEN, ML, DL are also reviewed.