神经PID控制在SVPWM整流器VFDPC中的应用研究

Application and Research on Neural Network PID Control in VFDPC of SVPWM Rectifier

分析了PWM整流器虚拟磁链直接功率控制(VFDPC)的工作原理,得出了基于虚拟磁链观测的瞬时功率估算式。设计了2个高通滤波器级联构成的二阶环节代替纯积分器滤除直流分量来解决虚拟磁链观测中存在的因积分初值选取不当造成的直流偏移问题。为提高PWM整流器直流电压稳定性,在电压外环设计了神经PID控制器。仿真结果表明,与传统PID控制相比,基于神经PID控制的SVPWM整流器直接功率控制系统直流侧电压稳定、动态性能更好。

The principle of VFDPC （virtual flux direct power control） for PWM rectifier was analyzed and the instantaneous power estimating equations based on virtual flux observer was obtained. The pure integral link is replaced by a second order link composed of two high-pass filter to filtering out the DC current. The DC voltage shifts problem caused by improper selection of integral initial value in virtual flux observer is solved. A neural network PID controller was designed in the voltage loop in order to improve the DC voltage stability of PWM rectifier. The simulation results show that, the PWM rectifier VFDPC system based on neural network PID control have the advantages of more stable DC voltage and better dynamic performance in start-up phase compared with the traditional PID control system.