摘要
为进一步提高状态观测器的观测精度,改善永磁直线电机调速系统的动静态性能,提出以RBF神经网络为辨识器,以单神经元PID为控制器,设计了一种智能观测器实现对电机速度的观测。在研究常规滑模观测器算法的基础上,把定子电压与电流作为观测器的输入项,以电机模型作为参考模型,以神经网络建立的观测器模型作为可调模型,通过电流实际值与估计值的比较,实现对反电势大小的观测,进而实现对电机速度的估计。仿真实验结果表明,基于参数辨识算法的智能状态观测器与传统滑模状态观测器(SMO)相比,速度响应曲线抖振较小,控制系统的动静态性能更优,稳定性更好。
In order to further enhance the observed accuracy of the state observer, and improve the static and dynamic performance of the permanent magnet linear synchronous motor （PMLSM） speed regulating system, the intelligent state observer was designed to observe the speed of PMLSM, which the RBF neural network was as the identifier in and the single neuron PID was controller. The stator voltage and the stator current were input items in intelligent observer. The back EMF of motor was observed by comparing the current val- ue of the reference model which was the motor model and the current value of the adjustable model, the ob- server model, which was built with the neural network, then the speed of motor was estimated. The simulation results prove the intelligent state observer based on the parameter identification by comparing to the traditional SMO has smaller buffet in speed response and better static and dynamic performances and stability.
出处
《微电机》
2016年第8期87-91,共5页
Micromotors
基金
郑州轻工业学院博士科研基金支助项目(000346)
关键词
永磁直线同步电机
参数辨识
RBF神经网络
智能观测器
permanent magnet linear synchronous
parameter identification
RBF neural network
intelligent observer
