模拟复合正交神经网络控制器在直流调速系统中的应用

Application of analog compound orthogonal neural network controller in DC speed adjusting system

在数字复合正交神经网络的基础上提出一种模拟复合正交神经网络,并用于直流双闭环调速系统中。控制器采用模拟复合正交神经网络的直接自适应控制方法,并对带有负载干扰的直流双闭环调速系统作了PI控制与神经网络控制的仿真实验。仿真结果表明,相对于常规PI控制器,该神经网络控制器具有自学习,自适应功能,速度跟踪获得了满意的控制效果。该模拟神经控制器能用于直流与交流调速系统中。

This paper covers an analog compound orthogonal neural network controller and its application to two closed- loop DC speed governing system. PI and neural network controlled double closed - loop DC speed governing system simulations are performed respectively with load disturbance. The results show that the neutral network controller has shelf- learning and self- adaptive functions, and has better control effect in speed tracking, compared with traditional PI controller. The analog neutral controller can be applied to DC and AC speed governing system.