Discrimination between Supply Unbalance and Stator Short-Circuit of an Induction Motor Using Neural Network

Abstract

The objective of this paper is to investigate the contribution and effectiveness of the use of a Neural Network (NN) Multi-Layer Perceptron (MLP) of feed-forward type technique to discriminate between the Supply Voltage Unbalance (SVU) and the stator Inter Turn Short-Circuit (ITSC) of an induction motor. The proposed approach in this paper is based on the exploitation of any useful information existing in both the current and the voltage quantities simultaneously provided by the various instantaneous powers such as the active and the reactive powers as well as the power factor. The merits of the proposed approach using the NN technique as an effective solution to the problem of discrimination is assessed by a simulation work and then validated by an experimental work conducted by the Diagnosis Group of the LDEE Laboratory.