Vibration-Based Fault Diagnosis of a Bevel and Spur Gearbox Using Continuous Wavelet Transform and Adaptive Neuro-Fuzzy Inference System

Abstract

This Chapter is based on a research study aimed at identifying faults in two different types of gearboxes under nonstationary conditions. The method proposed for fault diagnosis employs the independent angular resampling technique for processing the raw gearbox vibration signatures along with a hybrid intelligent classifier, namely, ANFIS, for fault diagnosis. With wavelet coefficients being input directly to an intelligent classifier for diagnosing gear faults, promising results have been reported in the recent past. However, considering that the computation burden associated with the implementation of an ANFIS classifier is dependent to a large extent on the network inputs, the angular domain averaged wavelet amplitude maps are segmented, each segment representing 60° of pinion rotation. Coefficients extracted from fragmented scalograms are then fed to a hybrid intelligent classifier, namely, ANFIS, for fault identification. The research outcome indicates reasonably good gear fault diagnostic potential in the case of the bevel gearbox. The methodology proposed is then extended to spur gearbox fault diagnosis with promising results. © Springer Nature Singapore Pte Ltd 2020.