CLASSIFICATION OF FAULTS IN GEARBOXES - PREPROCESSING ALGORITHMS AND NEURAL NETWORKS

Classical signal processing techniques when combined with pattern clas sification analysis can provide an automated fault detection procedure for machinery diagnostics. Artificial neural networks have recently b een established as a powerful method of pattern recognition. The neura l network-based fault detection approach usually requires pre-processi ng algorithms which enhance the fault features, reducing their number at the same time. Various time-invariant and time-variant signal pre-p rocessing algorithms are studied here. These include spectral analysis , time domain averaging, envelope detection, Wigner-Ville distribution s and wavelet transforms. A neural network pattern classifier with pre -processing algorithms is applied to experimental data in the form of vibration records taken from a controlled tooth fault in a pair of mes hing spur gears. The results show that faults can be detected and clas sified without errors.