盲信号分离在静止式中频电源故障分析中的应用  

Application of Blind Signal Separation in Analysis of Fault in Static Medium Frequency Power Supply

作　　者：朱俊杰[1] 黄高明[2] 聂子玲[1] 马伟明[1] 吴延好[1] 

机构地区：[1]海军工程大学舰船综合电力技术国防科技重点实验室,湖北武汉430033 [2]海军工程大学电子工程学院,湖北武汉430033

出　　处：《电气传动》2013年第8期59-64,共6页Electric Drive

基　　金：国家自然科学基金(51147009;50737004;51077131)

摘　　要：针对静止式中频电源故障信号含有间谐波,传统的信号处理方法难以对其进行有效分析的问题,提出将最新的盲信号分离(BSS)运用于电源故障分析中。根据故障信号的特点,采用基于负熵的快速独立分量分析(FastI-CA)算法解决了混叠信号的盲分离问题;采用主分量分析(PCA)进行预白化处理,减小噪声对分离精度的影响,再运用二阶盲辨识(SOBI)算法分离波动电压信号,得到各分离分量的频率估计。在此基础上通过求解超定方程组估算出故障信号分量幅值。仿真结果表明,FastICA和SOBI算法对中频电源故障信号分离的有效性和准确性。The fault signals of static medium frequency power supply cannot be analyzed effectively by traditional method of signal process because of fractional harmonics.In allusion to their limitation,the blind signal separation（BSS） which separate source signals fast and accurately was proposed.The fast independent component analysis（FastICA）algorithm was used to separate the mixtures.And the principal component analysis（PCA）whitening was used to reduce the influence of noise,and improve the separate precision.Then the second order blind identification（SOBI） algorithm was used to separate the fluctuated voltage signals.Based on that,the amplitudes of fault signals were achieved by simultaneous overdetermined equations system.Simulation results verify that the FastICA and SOBI algorithms are highly effective and correct.

关 键 词：静止式中频电源 盲信号分离 FASTICA算法 SOBI算法 PCA预白化 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]