Fault Diagnosis and Automatic Classification of Roller Bearings Using Time-Domain Features and Artificial Neural Network

Rolling element bearings are critical mechanical components in rotating machinery. Fault detection and diagnosis in the early stages of damage is necessary to prevent their malfunctioning and failure during operation. Vibration monitoring is the most widely used and cost-effective monitoring technique to detect, locate and distinguish faults in rolling element bearings. In this work rolling bearing fault diagnosis using time domain statistical parameters and automatic classification with the help of neural networks is done. Four bearings are identified in this work which is namely healthy bearing, bearing with Inner race defect, outer race defect and rolling element defect. Statistical features such as RMS, Peak value, Crest factor, Kurtosis and Skewness are calculated and are used as inputs to the neural network. These features are calculated from the acceleration data obtained from the above mentioned bearings. The ratio of crest factor (Cf defective vs. Cf healthy) and the ratio RMS (RMS defective vs. RMS healthy) are found. Automatic classification is done by using two different algorithms, the levenberg-marquardt back propagation and Bayesian regularization is used. It is seen that LM algorithm converges better and is more effective in the classification process.