ANN-approach for ITSC Fault Diagnosis of Induction Motor

摘要

Induction Motors (IMs) constitute the largest proportion of electrical machines in industry. This type of motor experiences different types of failures, such as stator inter-turn short circuit or stator windings (ITSC), broken bar and static or dynamic eccentricity. Detection of stator faults in their early stage is of great importance since they propagate rapidly and may cause further damage to the motor or persons. This paper presents an artificial neural network (ANN) based-approach to detect and isolate ITSC fault in three-phase IM. A feed forward multilayer-perceptron ANN trained by back propagation algorithm is applied. The fault detection and isolation (FDI) process is based on monitoring the three-phase instantaneous power average (IP) using the stator currents and voltages. The required data for training and testing the ANN is generated from a three-phase IM analytical model under different ITSC fault and load torque levels. An experimental data, provided by a test-bed, validates the ANN. The obtained results demonstrate the effectiveness of the proposed method.