Studies on Induction Motor Roller Bearing Fault Analysis Using Soft Computing Techniques

Abstract

Bearing components in Induction Motor IM play a critical role in operational performance and reliability of the system Therefore it necessitates the development of condition monitoring and fault diagnosis system to reduce the malfunctioning of the roller bearing Vibration analysis is commonly used in the detection of roller bearing failures The fault diagnosis method comprises of pattern recognition and classification paradigms in which feature extraction is the crucial role The effective and accurate classification of roller bearing faults depends on the salient feature extraction and reducing the dimensionality In roller bearing fault diagnosis a large amount of data is collected from the operating machinery Extraction of feature is difficult as the relevant information might be submerged inside the large data pool Principal component analysis multidimensional scaling and linear discriminate analysis are used for reduction of redundant data But these feature extraction methods work effectively only in linear data with Gaussian distribution whereas vibration signal of Induction Motor is nonlinear in nature The roller bearing faults can be predicted both in time domain and in frequency domain response of the system. newline