Condition monitoring of brushless DC motors with non-stationary dynamic conditions

This work introduces a new multi-sensor measurement framework for condition monitoring of brushless DC motors (BLDCM) with bearings. An experimental platform for equipment health monitoring is used for producing different faults on BLDCMs and log the measurement data. This work is oriented to maximize the life-cycle of industrial machinery and prevent catastrophic failures and their environmental consequences through reliable behavior classification. A public benchmark data set containing key failure scenarios is being built based on this work. This data set will be unique with respect to other available data sets due to the different sensors used and include more extensive scenarios such as non-stationary (time varying) conditions. A BLDCM with a bearing is tested under non-stationary conditions, and the scenario for their failure is developed. Supervised learning classifiers such as back propagation neural network and support vector machine are used to identify the fault state in the equipment.