CLASSIFICATION OF FAULT DIAGNOSIS IN A GEAR WHEEL BY USED PROBABILISTIC NEURAL NETWORK, FAST FOURIER TRANSFORM AND PRINCIPAL COMPONENT ANALYSIS

This paper presents the results of an experimental application of artificial neural network as a classifier of the degree of cracking of a tooth root in a gear wheel. The neural classifier was based on the artificial neural network of Probabilistic Neural Network type (PNN). The input data for the classifier was in a form of matrix composed of statistical measures, obtained from fast Fourier transform (FFT) and principal component analysis (PCA). The identified model of toothed gear transmission, operating in a circulating power system, served for generation of the teaching and testing set applied for the experiment.