Automatic Methodology for Multiple Fault Detection in Induction Motor Under Periodic Low-Frequency Fluctuating Load Based on Stray Flux Signals

Induction Motors are widely used in industrial facilities to perform energy conversion from electrical to mechanical. Hence, the detection of faults in these machines may lead to reduced losses and avoid unwanted stoppages. On the other hand, fault identification is often difficult since most of the processes are operating under non-stationary conditions produced by fluctuating loads, among others. Thereby, this work lies on the proposal of an automatic diagnosis methodology for detecting the occurrence of multiple faults in an induction motor that operates under the effect of a fluctuating load; the proposal is based on the processing of stray flux signals and on their characterization through a meaningful set of statistical features and then, a feature learning procedure is performed by Self-Organizing Maps to achieve a 2d grid modeling that preserves the most relevant topological properties for each assessed condition. Subsequently, the linear discriminant analysis is considered to carry out the dimensionality reduction of the previously modeled grids; all considered conditions are projected into a 2d space and finally, the identification of five different conditions in the induction motor is achieved by a proposed Neural Network classifier. Moreover, in the article, the analysis of the stator current signatures is also carried out for validation purposes. The proposed method is evaluated with experimental data. The obtained results make this proposal suitable for being applied in the monitoring of industrial applications that operate under fluctuating load conditions.