永磁同步电机系统神经网络逆解耦控制

Neural network inverse based decoupling control for PMSM drive system

提出了一种应用于全速域范围内的表贴式永磁同步电机神经网络逆解耦控制策略,该策略能够实现电机的精确解耦控制以及良好的动静态性能.在证明电机系统可逆性的基础上,对全速域范围内的线性化解耦控制特性进行了分析,针对逆解耦控制策略过于依赖系统数学模型的缺点,提出神经网络逆解耦控制策略,并设计了基于滑模变结构的附加转速与电流控制器.以一台5.2 k W永磁同步电机为被控对象进行仿真以及实验研究,结果表明,此策略能够在全速域范围内实现电机良好的解耦控制,并能够获得较好的动静态性能.

This paper presents a neural network inverse based decoupling control strategy for SPMSM(surface permanent magnet synchronous motors) in full speed range. The strategy can achieve good decoupling control performance, and obtain good dynamic and static performance in full speed range. Based on the motor system reversible proof, the linearized decoupling control characteristics are analyzed. Aiming at the disadvantage of the inverse decoupling control strategy which is too dependent on the mathematical model, the neural network inverse based decoupling control strategy is proposed, and an additional speed and current controller based on sliding mode variable structure are designed. A 5.2 k W permanent magnet synchronous motor is used as the controlled object for simulation and experimental study, and the result show that the neural network inverse based decoupling control strategy can achieve good decoupling control performance, and can obtain good dynamic and static performance in full speed range.