摘要
独立分量分析利用信号的高阶统计量快速准确地实现信号的分离和恢复。提出利用独立分量分析和同步滤波相结合的电压闪变检测方法。首先,采用同步解调的方法将包络信号和工频电压相乘的形式转变为包络信号和中心频率为100 Hz的双边带调幅信号的线性组合;然后依据负熵最大化的独立性准则和基于固定点的Fast-ICA算法进行盲信号分离;对分离得到的包络信号进行幅值修正,实现真实信号的估计。仿真结果表明该算法能准确分离平稳和非平稳的电压闪变信号的包络信号,且对幅值和频率的检测精度高。
ICA（Independent Component Analysis） separates and restores signal by using the high - order statistics of that signal. A method of voltage flicker detection using synchronous filtering and ICA is proposed. The muhiplicative result of envelope signal and power voltage is transformed into the signal combination of envelope signal and DSB（Double Side Band） modulated signal with 100Hz as its central frequency. With fixed-point Fast-ICA algorithm,BSS（Blind Source Separation） is then carried out according to the independency principle of maximizing Negentropy. The amplitude of the separated envelop signal is revised to obtain the estimation of real signal. Simulation results show that,this method separates the envelop signal from steady or unsteady voltage flicker with higher precision of frequency and amplitude detection.
出处
《电力自动化设备》
EI
CSCD
北大核心
2007年第11期34-37,共4页
Electric Power Automation Equipment
