Fault diagnosis of rotating machinery based on wavelet transforms and neural network

摘要

This paper shows usage of wavelet transform for condition evaluation of the rotating machinery by processing a signal of instantaneous angular velocity. One or more machine revolutions are used for the state evaluation. Wavelet transformation is applied to form a feature vector which, transformed by a neural network into a fault vector, is used for the description of a rotating machinery condition. Results obtained with a 2 cylinder four-stroke piston diesel engine ČKD S110 are shown.