Gear fault detection using wavelets and artificial neural network

A procedure is presented for fault detection of gears through wavelet transforms and artificial neural network (ANN). The time domain vibration signals of a rotating machine with normal and defective gears are processed through discrete wavelet transform to decompose in terms of low-frequency and high-frequency components. The features extracted from the decomposed signals are used as inputs to an ANN based fault detection approach. The ANN is trained using back-propagation algorithm with a subset of the experimental data for known machine conditions and tested using the remaining set of data. The procedure is illustrated through the experimental vibration data of a pump with and without gear fault. The roles of different vibration signals and their characteristic features in the fault detection process are investigated. The selection of features relevant to machine conditions leads to faster training requiring far less iterations.