极端频率直接估计的新自适应陷波器方法

Extremely frequency direct estimation algorithm based on new adaptive notch filter(ANF)

针对振动工程中常见的极端频率信号(归一化频率接近于0或π两端),现有的自适应陷波器(ANF)频率估计方法存在收敛速度慢、估计精度不高、算法稳定性差的问题。为此,提出一种极端频率直接估计ANF新方法。首先,提出一种新误差函数以提升极端频率信号下ANF的收敛速度和稳定性;然后,根据ANF估计极端频率信号时偏差产生的机理,通过偏差补偿方式,降低噪声对ANF的影响,以获得近似无偏的频率估计结果,提高ANF极端频率估计精度,并对该方法进行稳态条件下的性能分析;最后,给出了ANF极端频率信号估计结果,并讨论了ANF各参数对极端频率信号估计精度的影响,给出了具体计算结果。计算表明方法的有效性与正确性。

In order to improve the frequency estimation accuracy,convergence speed and stability of adaptive notch filter（ANF）especially for the extreme frequency signal（normalized frequency close to 0orπ）which exists in vibration engineering domain,a new method based on ANF for extreme frequency signal is proposed.Firstly,the proposed new error function could improve convergence speed and stability of ANF;secondly,according to the mechanism of the deviation generated by ANF,a bias offsetting method is adopted to reduce the noise influence on the accuracy of extreme frequency estimation and to obtain the unbiased extreme frequency estimation results,where the performance analysis is also given;thirdly,the extreme frequency estimation results for different parameters of ANF are acquired and discussed.Finally,the calculated results show the effectiveness and correctness of the algorithm.