Fault diagnosis of induction motor using linear discriminant analysis

Abstract

In this paper, we propose a diagnosis algorithm to detect faults of induction motor using the linear discriminant analysis. First, after reducing the input dimension of the current value vector measured at each period by using the principal component analysis method, we extract the feature vectors for each fault using the linear discriminant analysis. And then, we will diagnosis the condition of an induction motor by using a distance measure between the predefined fault vectors and the input vector. From the various experiments under noisy conditions, we found that the proposed fault detection method could be applied to prevent a fault by diagnosing the conditions of a induction motor in real industrial applications. © Springer-Verlag Berlin Heidelberg 2005.