This paper presents neural network based on-line bearing fault detection system using wavelet technique. The presence of fault in a motor can be detected by analyzing its stator current signal which is the basis for the detection of bearing fault in this paper. Discrete Wavelet Transform (DWT) is used to preprocess the stator current signal followed by extraction of features based on calculation of 11 statistical parameters. Artificial Neural Network (ANN) is then trained and tested using the extracted features. The use of ANN as the optimization technique helps in reducing both computation time and error. A system is then built using Virtual Instrumentation consisting of DWT, Feature Extraction and ANN block. The result of the above proposed system is the healthiness of the motor, severity of the fault and type of fault. Various test results and conclusions are presented in this paper.
S. P. Charan, Dr. Sindhu Thampatty K.C., Preethi, P., and T. Balaram, H., “ANN Based Online Bearing Fault Detection System Using Discrete Wavelet Transform”, International Conference on Advances In Engineering And Technology - ICAET 2014. RIT, Roorkee, India, pp. 669 - 677, 2014.