新型混合有源电力滤波器的特性分析和控制方法  

Characteristic Analysis and Control Method of a Novel Hybrid Injection Active Power Filter

作　　者：徐先勇[1] 罗志坤[1] 罗安[2] 方璐[2] 吴敬兵[2] 彭潇[1] 

机构地区：[1]湖南省电力公司科学研究院长沙410007 [2]湖南大学电气与信息工程学院长沙410082

出　　处：《电工技术学报》2011年第12期113-121,共9页Transactions of China Electrotechnical Society

基　　金：湖南省科技计划项目(2010CK3017);湖南省研究生科研创新项目(CX2010B148)

摘　　要：本文针对某变电站35kV高压侧存在谐波超标、无功不足的情况,研究了一种适用于高电压等级的大功率双谐振注入式混合有源电力滤波器。在分析了其拓扑结构的基础上,讨论了该滤波器的滤波性能特性、双谐振注入支路的无功补偿特性和其分压分流特性。为了增强整个系统的稳定性,提出了谐波电流双闭环控制策略,内环以逆变器输出电流为控制目标采用传统PI控制器;外环以电网谐波电流为控制目标,针对外环控制目标为非直流量的特性,提出了基于改进PSO-BP神经网络的递推积分PI控制方法,使用改进的PSO-BP神经网络来调整控制器的比例、积分系数,具有很好的稳态和动态性能。实例分析验证了本文研究内容的有效性和正确性。Due to high harmonic current and low reactive power in 35kV bus of a substation, a kind of high-power hybrid active power filter with dual-resonance injection branch is studied, which is suited for high-voltage level. In the analysis of its topology, the filter characteristics, reactive power compensation characteristics, voltage-sharing and current-sharing characteristics of dual-resonance injection branch are discussed. In order to enhance the stability of the system, dual closed loop control strategy of harmonic is proposed. The control aim of inner-loop is the output current of inverter and its controller is traditional PI controller. The control aim of outer-loop is the harmonic current of power grid. The novel recursive integral PI control method based on improved PSO-BP neural network is proposed for the non-straight control aim. The improved PSO-BP neural network is used to adjust the proportional and integral factors, with good steady state and dynamic performance. Research case study results verify the effectiveness and correctness of the content.

关 键 词：有源电力滤波器 特性分析 双谐振支路 双闭环控制 改进PSO-BP神经网络 

分 类 号：TM711[电气工程—电力系统及自动化]