Application of Fuzzy Inference System for Cage Induction Motors Rotor Eccentricity Diagnostic

Abstract

This paper presents a diagnostic method of the eccentricity of the rotor of the cage induction motor, using a diagnostic inference block, based on ANFIS (adaptive-network-based fuzzy inference system). A construction and principle of work of the diagnostic inference block for a selected induction motor will be provided, in order to assess the eccentricity of the rotor. The input data for the proposed diagnostic inference block will consist of the values of two eccentricity indicators, representing the static and the dynamic eccentricity. The indicators are set for every motor on the basis of the solution of multi-harmonics mathematical model of the cage induction motor. The proposed system will assess relative eccentricity levels. At the output of the system, information regarding the sum of the relative eccentricity levels (both static and dynamic) will be available. Examples of the eccentricity diagnosis supplied by the developed diagnostic inference block are interpreted with regard to the diagnosis correctness and accuracy.