并联有源滤波器的迭代学习控制方法及其设计(英文)

Design of a Three-phase Shunt Active Power Filter with Iterative Learning Control

在系统存在非线性负载下,谐波电流呈现激烈波动并存在许多不连续点,对有源滤波器设计高性能的电流跟踪控制器是非常重要的。基于传统设计的控制器,在不连续点处性能有限,同时要求控制器增益高,但高增益控制器易引起系统的高频振荡甚至不稳定。由于在非线性负载下所需补偿的谐波电流为周期性,利用迭代学习控制的学习特性,设计三相并联有源滤波器,可以有效的避免控制器的高增益现象,同时提高电流追踪性能,大大减小了谐波总畸变率,功率因素也得以提高。实验制作的以DSP为控制CPU实现迭代学习控制器的系统中,运行结果表明本文方法效果良好。

Active power filter（APF） is commonly used to eliminate harmonic contamination and improve power factor.When the system contains nonlinear loads,the harmonic current command become drastically vibrating and contains many discontinuous points.It is important to design high performance current tacking controller for active power filters.For current controller based on linear design,its performance is limited near operating point.On the other hand,most controller designs require high gain at these discontinuous points to achieve fast tracking.But a high gain control will excite high-frequency unmodelled dynamics and cause unstable.An iterative learning controlled APF is used to improve these drawbacks.The harmonic current commands in d-q frame are periodic and a PD-type iterative learning scheme is employed.As learning control scheme adds control input of previous learning cycle to control law generated from tracking errors in present learning cycle,the controller will eventually produce desired control input to achieve accurate and fast current tracking which leads to low current total harmonic distortion（THD）and high power factor.Hence,the high gain effect in control law is prevented.Experimental results on a 2 kVA DSP-based controlled evidence these attractive features.