变步长自适应算法在有源滤波器谐波检测中的应用

对已有改进自适应算法性能进行分析,并尝试将其应用于有源滤波器谐波检测中。实验表明这些变步长算法的性能极易受到系统中噪声信号的干扰,改进效果不明显。该文提出了一种新的变步长自适应谐波检测算法。算法采用归一化的处理方法,以误差信号e(n)在待检测信号中所占比率K(n)作为自适应回馈量,并通过K(n)的相干平均估计来调节算法的步长。其优越性在于:即使在有较大噪声干扰的情况下,谐波检测过程也能保证既具有较快的动态响应速度,又保持较高的检测精度,而不像传统算法那样必需在两者性能上进行折中选择。仿真和实验结果证明了理论分析的有效性。文章还对该算法的收敛性和稳定性进行了分析,并就算法中参数的选取提供了参考标准。

A variable Step Size Adaptive Algorithm Applied to Harmonic Detection of Active Filter

A number of conventional time-varying step size algorithms have been tried on the active filter,and experiments indicate that their performance is highly sensitive to the noise disturbance.The paper presents a novel adaptive harmonic detecting algorithm using the coherence average estimation of K(n)to control the updating of the step size,where K(n)is the proportion of the error signal e(n)in the total signal S(n).It can provide a fast convergence speed and ensure a small detecting error even under low SNR conditions.Theoretical analysis is verified by simulation and experiment results.The convergence speed and steady state performance of the algorithm are analyzed,and one method to choose the parameters of the algorithm is provided.