Predictive Maintenance for Electric Motors, RPM Data, Machine Learning for Proactive Life Cycle Management

The research develops predictive maintenance as a solution which prevents unexpected electric motor breakdowns. The objective remains to develop a machine learning algorithm for Remaining Useful Life (RUL) prediction which utilizes RPM (Revolutions Per Minute) values as input data. Past RPM data training of an autoencoder allows detection of anomalies while it predicts motor performance for proactive planning of maintenance activities. Real-time monitoring and visualization becomes possible through the implementation of this solution which functions using Python and streamer application. Their predictive model connects motor deterioration patterns to create double benefits of operational savings through decreased device downtime coupled with accident failure prevention and reactive maintenance improvement to predictive maintenance. The proposed model should be expanded to include new equipment types while new sensor integration will boost accuracy levels.