云粒子群优化神经PI控制器的SHAPF谐波补偿研究  

Research of SHAPF harmonic compensation based on cloud particle swarm optimization neural network PI controller

作　　者：张佩炯[1] 苏宏升[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

出　　处：《电源技术》2013年第5期832-835,共4页Chinese Journal of Power Sources

基　　金：甘肃省自然科学基金(096RJZA092);甘肃省普通高等学校研究生导师资助项目(1004-06)

摘　　要：分析了并联混合型电力滤波器(SHAPF)的工作原理,针对传统PI控制器参数调节的缺陷,提出了一种云粒子群优化BP神经网络的PI控制方案,设计了一种神经自适应PI控制器。通过仿真及实验验证:与传统PI控制器相比,该控制器结构简单,对系统的补偿精度较高,计算速度性能好,有更好的谐波补偿效果。The principle of shunt hybrid active power filter was analyzed. For parameters adjustment defects of general PI controller, a PI control method of cloud particle swarm optimization BP neural network was proposed, and a neural network adaptive PI controller was designed. The simulation and experiment show that the control has the advantages of simple structure and higher accuracy of compensation comparing with general PI controller, calculation speed performance and the harmonic compensation effect are good.

关 键 词：神经网络 混合型电力滤波器 云粒子群 PI控制器 谐波补偿 

分 类 号：TM571[电气工程—电器]