A comprehensive approach to convolutional neural networks‐based condition monitoring of permanent magnet synchronous motor drives

The increasing complexity of modern industrial systems calls for automatic and innovative predictive maintenance techniques. As suggested by the Industry 4.0 process, this demand translates in need more-intelligent drives. Herein, use a special kind neural networks to interpret data from motor currents diagnostic purposes is described. early detection possible faults electrical allows programmed reduces risk unplanned shutdowns. innovation overall approach network training, which does not call anymore large set faulty motors. A training dataset generated using combination tuned models some augmentation techniques proposed. result comprehensive effective condition monitoring algorithm, whose hearth convolutionary trained safe cheap simulation-based dataset. details design are fully reported here. method has been implemented laboratory tested on both healthy permanent magnet synchronous generality proposed also paves way other failures application different