The objective of this study is to present artificial intelligence (AI) technique for detection and localization of fault in induction machine fault, through a multi-winding model for the simulation of four adjacent broken bars and three-phase model for the simulation of shortcircuit between turns. In this work, it was found that the application of artificial neural networks (ANN) based on Root mean square values (RMS) plays a big role for fault detection and localization. The simulation and obtained results indicate that ANN is able to detect the faulty with high accuracy. Keywords:Induction machine, faults detection and localization, broken bars, artificial neural network (ANN), root mean square (RMS), multi winding, three-phase model = [V] − [R][I](1)

In this work, we proposed multi-winding model for the simulation of broken bars in squirrel cage asynchronous machine, this model allows to study the influence of the broken bar defects on the behavior general of machines in different operating conditions (healthy and faulty). The breaking of the most frequent bars of the rotor causes oscillations of the torque, speed, and the current, the increase of the resistance of the rotor creates the defects proportional with the number of breaks bar K .The diagnosis fault using technique of single processing based on Spectrum analysis for detection broken bar. The results of the simulation obtained allowed us to show the importance of this technique for detection broken bar.

This study aims to display fuzzy logic (FL) technique for diagnosis of fault induction machine. This allows monitoring of fuzzy information from different signals to give more accurate judgment on the health of the engine, through using a multi-winding model of induction machine for the simulation of broken bars. This model allows study the influence of defects and appear the behavior of the machine in the different modes of running conditions (healthy and fault). In this work, we focus the application of a fuzzy logic technique based on the fast Fourier transformation (FFT) by analyzing the stator current for fault detection. The results of the simulation obtained allowed us to show the importance of the fuzzy logic approach based on classification of signals for detecting the faulty.

The purpose of this study is diagnosis the fault of induction machine, through detection of broken bars by stator current analysis. Therefor we present a multi-winding model for the simulation of faults as part of the fault detection study, and test the behavior of the machine in the healthy and faulty state. In this paper it is focus the signal processing technique for detecting defects. This technique is based on continuous wavelet transform (CWT) to detect and locate defects, by using the multi-level decomposition for the detection of defects and the location of the broken bars at the rotor. The findings of this research results show the importance of this technique for the analysis of fault signatures in healthy and defective cases.