Diagnostic des défauts statoriques des machines Asynchrones par l'approche LS-SVM

Abstract

This work presents a diagnosis of a fault (inter-turn short circuit) using the least squares support vector machine (LS-SVM) classification algorithm. Several methods have been developed to monitor the condition of machines based on intelligent techniques such as neural networks, fuzzy logic, Kalman filter, etc. However, the use of LS-SVM for machine health monitoring and fault diagnosis is still rare. LS-SVM provides highly accurate classification for machine health monitoring and diagnosis, which provides excellent generalization performance.