混合有源滤波器的新型电流迭代学习控制

摘要

为提高电网电能质量,针对注入式混合有源电力滤波装置(HAPF)提出一种新的电流补偿控制策略,由电网谐波电流误差信号的微分型学习律前馈控制环节和反馈迭代学习控制器构成.在电流误差的反馈迭代控制器中应用了一种带遗忘因子的PI新型迭代学习控制算法,遗忘因子的加入能够增强迭代学习控制算法的鲁棒性,且该算法区别于传统PI型迭代学习控制算法,其算法收敛性不依赖于谐波误差信号的初始控制输入.在反馈迭代控制器中应用了一种改进的Ziegler-Nichols方法实现控制器参数的优化,提高了系统的控制精度.利用模糊优化算法对前馈控制增益进行优化,以改善系统的动态响应性能.最后,对所提控制策略进行了仿真和实验验证,仿真和实验结果证明了所提控制方法的可行性,电网中的特征次谐波有明显减少.%A current compensation control is proposed for HAPF(Hybrid Active Power Filter) to improve power quality of grid,which is composed of the feed-forward link based on derivative-type learning law of harmonic current error and the feedback iterative controller. An improved Pi-type iterative learning control with forgetting factor is applied in the feedback iterative controller of current error to improve its robustness, the convergency of which is independent of its initial input, different from the conventional iterative learning algorithm. An improved Ziegler-Nichols method is applied in the feedback iterative controller to optimize its parameters for the system control accuracy improvement. Fuzzy optimization is used for the control gain of feed-forward link to improve the system response performance. Simulative and experimental results demonstrate the characteristic harmonic currents of grid decrease obviously, confirming the feasibility of proposed method.