Active-Disturbance Rejection Control of Brushless DC Motor Based on BP Neural Network

Brushless DC motor speed servo system is multivariable, nonlinear and strong coupling. Its performance is easily influenced by the parameter variation, the cogging torque and the load disturbance. To solve the deficiency, the paper represents the algorithm of active-disturbance rejection control (ADRC) based on back-propagation (BP) neural network. The ADRC is independent of accurate system and its extended-state observer can estimate the disturbance of the system accurately. However, the parameters of Nonlinear Feedback (NF) in ADRC are difficult to obtain. In this paper, these parameters are self-turned by the BP neural network. The simulation results indicate that the ADRC based on BP neural network can improve the performances of the servo system in rapidity, control accuracy, adaptability and robustness.