Technique de diagnostic des défauts dans les systèmes électriques par les réseaux de neurones (machine asynchrone)

Abstract

The asynchronous machine is the most used in industry due to its robustness and low purchase or maintenance cost, but it can be exposed to many electrical or mechanical faults during its operation, which requires early detection. This has led to the use of many diagnostic methods that allow us to identify and classify faults that occur in the machine, among these techniques is the use of artificial intelligence and artificial neural networks. The objective of this work is to diagnose malfunctions of squirrel cage induction machines (broken bar fault) using artificial neural network techniques. We developed a neural network model to detect and classify defects, then we performed tests to validate the neural network model