Fault Diagnosis of Crown Wheel and Pinion in Differential Using Acoustic Signals with Discrete Wavelet Transform and Neural Network

This paper concentrates on a procedure which experimentally recognizes crown wheel and pinion faultsof a typical differential system using a MLPneural network. The differential conditions were considered to be normal differential, broken and worn crown wheel, broken and worn pinion. Feature vector which is one of the most significant parameters to design an appropriate neural network were extracted from analysis of acoustic signals in time-frequency domain by means of wavelet transform method. The statistical features of acoustic signals such as mean, standard deviation, variance and etc were used as input to ANN. Different neural network structures are analyzed to find the optimal neural network with regards to the number of hidden layers. The results showthat the designed system is capable of classifying records with 90.6% accuracy with one hidden layers of neurons in the neural network.