A statistical parameters and artificial neural networks application for rolling element bearing fault diagnosis using wavelet transform preprocessing

Vibration monitoring and analysis is a powerful and recommended tool for preventive maintenance and early detection of impending failures in rotary machine. The demand for cost efficient, reliable and safe rotating machinery requires accurate fault diagnosis, classification and prognosis systems. This work presents a study to explore the performances of bearing fault diagnosis by using wavelet neural network classifier. The features vector required for the training and testing of neural network are calculated by the implementation of a wavelet packet. The statistical features such as standard deviation, kurtosis, central moment, wavelet energy and other statistical parameters of vibration signals was run in its normal condition and abnormal operating mode were used as input to ANN classifier. The results show that the statistical parameters identified the fault categories of rolling element bearing accurately and has a good diagnosis performance.