Statistical Neural Networks for Induction Machine Fault Diagnosis and Features Processing Based on Principal Component Analysis

Induction machine monitoring becomes a substantial industrial need due to their constraints such as mechanical and electrical defects. Thus, many fault diagnosis tools have been employed to figure out these drawbacks. The motor current signature analysis is very developed in fault processing of induction machine fault diagnosis. In this paper, our proposed technique combines the motor current analysis and the artificial intelligence (AI) for induction motors faults recognition. Particularly, we introduce the statistical neural networks as an efficient classifier among AI techniques which provides the operating modes identification. In this context, the radial basis function neural networks and the probabilistic neural network are studied in this paper for detecting faults based on motor current analysis. In selection features, we applied the principal component analysis to transform the original database into novel uncorrelated features space. (c) 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.