有源电力滤波器神经网络逆解耦控制  

Decoupling control of active power filter based on neural network inverse

作　　者：刘国海[1] 杨辰星[1] 陈兆岭[1] 

机构地区：[1]江苏大学电气信息工程学院,江苏镇江212013

出　　处：《电力系统保护与控制》2013年第5期23-29,共7页Power System Protection and Control

基　　金：江苏省高校自然科学基金资助项目(10KJB470003);江苏高校优势学科建设工程资助项目(苏政办发(2011)6号)

摘　　要：三相并联型有源电力滤波器是一个多输入、多输出的非线性系统,而且在d-q旋转坐标系下的补偿电流分量间存在着耦合关系。提出了一种采用神经网络逆实现有源电力滤波器线性化解耦控制的方法。通过静态神经网络逆控制方法,将三相并联型有源电力滤波器解耦成两个关于补偿电流分量的一阶线性子系统,再分别设计线性闭环控制器对两个补偿电流分量进行控制。仿真结果表明,采用神经网络逆解耦控制方法,系统具有良好的动静态性能,系统稳态时的谐波畸变率小于2%。Three-phase shunt active power filter is a multiple-input multiple-output nonlinear system, in which a coupling relationship exists between the compensation current components under the synchronous orthogonal d-q frame. A new decoupling control method for active power filter by employing neural network inverse is proposed. By adopting the static neural network inverse control method, the three-phase shunt active power filter is decoupled into two one-order linear subsystems, which concem the compensation current components. Then, the linear closed-loop controllers are designed to control the compensation current components respectively. Simulation results verify that by applying the proposed control method, the system has both good steady-state and dynamic performances. After the compensation, the total harmonic distortion of the power source currents in steady state is less than 2%.

关 键 词：有源电力滤波器 神经网络逆 模型线性化 解耦控制 线性闭环控制器 

分 类 号：TM76[电气工程—电力系统及自动化] TN713.8[电子电信—电路与系统]