A novel sensorless control strategy for BLDC motor drives using a fuzzy logic-based neural network observer

This paper presents a novel sensorless control method for brushless DC motors using two similar fuzzy logic-based neural network observers. Firstly, two fuzzy logic observers are designed and used to estimate back-EMF space vector components in the stationary reference frame. The inputs of the observers are measured values of line to line voltages and currents, and their outputs are back-EMF space vector components in the stationary reference frame. To determine the commutation instants, a new technique based on the sign of the back-EMF space vector components is proposed. In order to evade the difficulties pertaining to implementation of the fuzzy logic observers, instead of them, two neural network observers are designed which use inputs and output pairs of the fuzzy logic observers as training data. In other words, the fuzzy logic observers is replaced by two neural network ones with the same performance which are easier to implement. This method provides an accurate sensorless operation over wide speed range. The validity of the proposed method is verified through simulations done on a prototype BLDC.