Sensorless Commutation Error Compensation of High Speed Brushless DC Motor Based on RBF Neural Network Method

摘要

This paper proposes a novel compensation method for commutation error of brushless DC motor using position sensorless control technique for aerospace applications in order to reduce the operational losses. Low-pass filter (LPF) circuit and virtual neutral point-based comparator are used to extract the back electromotive force (back-EMF) zero-crossing points (ZCPs), and the original commutation point is trigged by delaying ZCPs 30 or 90 electrical degrees according to the real-time motor speed. Basic strategy of using back-EMF difference as the feedback signal to compensate the commutation error is designed and the effects on the voltage difference of rotor speed variation under different speed conditions is analyzed and eliminated. Adaptive compensation method based on RBF neural network is proposed. The method can effectively compensate commutation errors of the high speed motor from acceleration to steady speed state. Simulation and experimental results verify the effectiveness of the proposed compensation method.