低频振荡参数Prony辨识中的数字滤波器设计

Design of Digital Filter for Recognition of Low-frequency Oscillation Parameters Using Prony Algorithm

Prony算法可有效进行电力系统低频振荡分析,但是其存在抗噪性弱的缺点。为降低PMU数据中噪声对低频振荡分析辨识精度的影响,本文提出了一种用于Prony算法的数字滤波器设计方案及幅相补偿方法:首先仿照周期信号滤波要求设计了模拟滤波器,针对模拟滤波器造成的Prony算法辨识的偏移,提出幅值及相位补偿方法;然后基于脉冲响应不变法将模拟滤波器离散化为数字滤波器;最后通过滤波后的含噪信号与理想信号对比,验证了该方法的抗噪性能。将该滤波器用于PMU数据进行算例仿真,得到了较好的辨识效果。

Prony algorithm can be used to analyze the low-frequency oscillation of power system effectively;however,its anti-noise capability is poor.To minimize the interference that noises existing in phase measurement unit(PMU)da ta impose on the recognition precision of low-frequency oscillation analysis,a digital filter design scheme which can be applied to the Prony algorithm is presented in this paper,together with its phase and amplitude compensation method.First,an analog filter is designed according to the periodic signal filtering requirement.Then,considering the deviation of recognition using Prony algorithm due to the analog filter,an amplitude and phase compensation method is pro posed;meanwhile,the analog filter is discretized to a digital filter based on the impulse response invariance method.The denoising performance of the proposed method is proved by comparison between noise signals after filtering and ide al signals.A numerical example is simulated by applying this filter to PMU data,and a better recognition effect is ob tained.