摘要
研究工业过程故障诊断中的信号奇异点检测问题.采用结合小波消噪的盲源分离算法提取有用的源信号,在分析李氏指数和小波变换的极大值与信号奇异点的关系基础上,分析了信号奇异点检测所用的小波尺度及阈值选择方法.实例分析和比较表明,该方法的主要优势在于它对随机噪声的降噪效果明显,而且能有效地检测出信号的奇异点.
Signal singularity detection in industrial process fault diagnosis is studied. The blind source separation algorithm combined with wavelet canceling noise is used to extract source signal. Based on the analysis of the relationship among Lipschitz exponent, wavelet transform maximum and signal singularity, the selection methods of scale and the threshold of wavelet are analyzed. The example shows that the method has better ability of noise canceling in comparing with other methods. It can also detect the characteristic singularity of the analyzed signal effectively.
出处
《控制与决策》
EI
CSCD
北大核心
2007年第9期1035-1038,共4页
Control and Decision
基金
国家自然科学基金项目(60325311
60572070)
中国博士后科学基金项目(20060400962)
辽宁省自然科学基金项目
教育部流程工业综合自动化重点实验室基金项目(PAL200503)
关键词
小波消噪
盲源分离
奇异信号
李氏指数
Wavelet canceling noise
Blind source separation
Singular signal
Lipschitz exponent