Integration of an intelligent neuronal technique in anomalies detection on induction machines

This paper presents an intelligent strategy for anomaly detection in induction machines using the map SOM. It involves the most significant parameters of SOM, such as the topological structure of the map, the Kohonen learning algorithm, the activity diagram in a micro light and frequencies characteristics of anomalies. A comparative study of the anomaly detection performance was conducted under the SOM neural map and the vibration analysis method. Eventually, a simulation is made on Matlab for the SOM learning accompanied by monitoring of the experimental vibration analysis in the NDC (Non Destructive Control) Laboratory. It will be a more synthetic analysis.