无负载侧谐波检测的有源滤波器自适应预测算法

Adaptive forecasting algorithm without load-side harmonic detection for active power filter control

经典有源滤波器算法中通过控制直流侧电压环的算法可以省去负载和输出电流检测单元。为了增强其动态性能,给出了基于自适应预测算法的无负载侧谐波检测有源滤波器控制方法。该方法对直流侧电压差进行滑窗迭代,不断地对窗内数据进行加权平均计算。通过历史数据自适应滤波器系数在线滚动迭代,步长实时更新,结合内插值法能够很好地预测出网侧电流下一拍的状态。仿真结果证明了所提出的方法具有良好的稳态补偿效果和动态跟踪性能,基于DSP的样机实验结果再次证明了该算法的可行性与正确性。

The classical APF（Active Power Filter） algorithm applies the DC voltage control algorithm to save the detection units for load current and output current. An APF control method based on an adaptive forecasting algorithm without load-side harmonic detection is proposed to enhance its dynamic performance, which continuously calculates the weighted average of DC voltage error based on the sliding-window iterative algorithm with the forgetting factor. The coefficient of adaptive filter is renewed during the online rolling iteration,the step length is upgraded in real time,and the grid current of next step is well forecasted by the interpolation algorithm. Simulative results demonstrate the excellent steady-state compensation effect and dynamic tracking performance of the proposed method. Experimental results of a prototype based on DSP also verify the feasibility and correctness of the given algorithm.