弱信号的数据自适应检测方法研究01.doc

I 摘要 人们在探索研究宏观和微观世界中，常常需要检测微弱信号。常规仪表由于 存在较大的内部噪声，有用信号经常被噪声淹没，因此采用模拟的放大器和滤波 器无法有效地检测出有用信号。本文所讨论的对象是微弱的稳态周期正弦信号的 检测，由于正弦信号具有周期性，正弦信号间具有相关性和正交性，因此相关检 测方法成为在这种情形下普遍采用的方法。 本文针对稳态周期正弦信号的提取提出一种新的方法—自适应正弦滤波。它 是依据现代优化控制理论构造数字滤波器，滤波器的输入是标准的正弦函数，按 照某一准则自动调节滤波器系数，当满足某一目标的数字滤波器输出即为所提取 的微弱有用信号信息(幅值和相位)。 本文采用MATLAB就相干检测方法和自适应正弦滤波方法在微弱信号的提取 进行仿真研究。结果证明自适应正弦滤波方法有效可行，在不同的信号源频率组 成情况下，相比相关检测方法，自适应正弦滤波方法应用更加灵活，能更加快速 地提取微弱有用信号。 最后，本文将自适应正弦滤波方法应用于电力系统谐波的提取。成功地提取 了发电机定子单相接地故障前后的中性点三次谐波电压和电流信号，为基于三次 谐波阻抗的发电机定子单相接地保护新方法提供了技术支撑。 关键词：相关检测;自适应正弦滤波;谐波信号;定子单相接地保护II Abstract When exploring and researching on macroscopical and microcosmic world,the weak signal detection is needed.Much internal noise exists in analog detection instruments and the useful signal is covered by noise,so the useful weak signal can’t be detected by using amplifier and analog filter.The subject we discussed in this paper is the extraction of stable periodic sinusoidal signal.Correlation detection method is widely used in this case since sinusoidal signals is periodic and correlative and orthogonal to each other. A new method is proposed apply in the extraction of stable periodic sinusoidal signal—adaptive sinusoidal filtering.The digital filter is constructed by modern optimization control theory,the input of the filter is standard sinusoidal signal,the filter coefficients are automatically adjusted by certain rule,the output of the filter is the useful information(amplitude and phase)of weak signal when a certain goal is satisfacted. We do the simulation of correlation method and adaptive sinusoidal filtering method by MATLAB.The result of simulation proves that the adaptive sinusoidal filtering method is effective and feasible,in the case of the source signal is composed of signals with different frequency,the adaptive sinusoidal filtering method is more flexible,and can extract weak signal in more short time. At last,the adaptive sinusoidal filtering method is applied in the extraction of harmonic signal in power system