Neural Fault Diagnosis Method for Voltage Source Inverter with a Neural Direct Torque Control of Induction Motor

Abstract

electrical drives in general incorporate an inverter and an induction machine. Thus, these both elements must be well considered to provide a relevant diagnosis of these electrical systems. So it is important to detect early different defects that can occur in these systems in order to find ways to allow us to monitor the operation and preventive action to avoid frequent breakdowns. The objective of this paper is to investigate the feasibility of detecting and diagnosing faults in a three-phase inverter supplying an induction motor. We present the simulation results of a neural direct torque control of (NDTC) of induction motor associating a fault diagnosis system by using the contribution of artificial intelligence. In this work, we give a detailed description of inverter switching faults with a simple method for feature extraction to study the possibility of detecting and diagnosing these defects. Detection and identification of faulty switches is realized within a few currents periods. The use of an intelligent technique improves the classification performance for one and only fault occurrence.