基于神经网络的地铁牵引整流器控制算法研究  

Research of Control Algorithm of Traction Power Supply Rectifiers in Subway Based on Neural Network

作　　者：杜荣茂[1] 李向葵 景利学[1] 

机构地区：[1]兰州交通大学自动化与电气工程学院,甘肃兰州730070 [2]甘肃省高等学校招生办公室,甘肃兰州730030

出　　处：《铁道机车车辆》2012年第2期58-63,103,共7页Railway Locomotive & Car

基　　金：甘肃省教育厅硕导项目(20872)

摘　　要：提出了一种利用神经网络对传统的PID控制器进行改造并应用于PWM单位功率因数整流器的控制方法,选择SVPWM控制方法对三相电压型PWM整流器进行控制,电压环采用基于神经网络自适应调整参数的复合型单神经元PI控制器。仿真结果表明,该PI控制器可以在线调整PI参数,快速跟踪整流器的变化,能有效地改善由于系统结构和参数变化而导致的控制效果不稳定的状况,适用于地铁需要频繁的起动、停车的条件。A control method using neural network to improve the traditional PID controller for PWM rectifier with unity power factor is presented.In this paper,the space-vector PWM is used to control the rectifiers,and an adaptive compound single-neuron PI controller based on neural network is used in the voltage loop to adjust the parameters.Simulation results show that the controller built in the paper is able to adjust the PI parameters online,trace the changing process of the rectifier quickly,improve the stability of system effectively when the structure and parameters of the system is changed,and can meet the conditions of frequent start and stop,which is necessary for subway.

关 键 词：三相电压型PWM整流器 SVPWM 单神经元PID 

分 类 号：U264.3[机械工程—车辆工程]