Classification of Wavelet Map Patterns Using Multi-Layer Neural Networks for Gear Fault Detection

A multi-layer perceptron pattern classifier is defined for wavelet map interpretation and its application is described as a tool for mechanical fault detection. As a key step, an instantaneous scale distribution is introduced for quantifying pattern features. Instead of directly inspecting complicated wavelet patterns in time¿scale domains with limited human experience and availability, automated classification of the localised features related to gear faults, therefore, can be implemented. The detail of constructing, training and testing the multi-layer perceptron based classifier has been described with application to a gearbox.