基于能量的振动信号经验模态分解终止条件  

Energy-Based Stop Condition of Empirical Mode Decomposition of Vibration Signal

作　　者：胡劲松[1] 杨世锡[2] 

机构地区：[1]宁波工程学院电信学院,宁波315010 [2]浙江大学机械与能源工程学院,杭州310027

出　　处：《振动．测试与诊断》2009年第1期19-22,共4页Journal of Vibration,Measurement & Diagnosis

基　　金：国家自然科学基金资助项目(编号:50675194);宁波市自然科学基金资助项目(编号:2007A610014)

摘　　要：经验模态分解(Empirical Mode Decomposition,简称EMD)主要思想是把一个时间序列的信号分解成不同尺度的本征模函数(Intrinsic Mode Function,简称IMF)。EMD算法用残余量极值点数量小于2来终止分解,该终止条件有无关组分多和分解速度慢的缺点。针对该缺点,提出了采用残余能量小于设定值的EMD分解终止条件。通过对非线性和实例振动信号的试验研究表明,基于能量的EMD分解终止条件既可以减少分解的无关组分,又可以提高分解的速度。该研究成果能广泛地用于振动信号分析领域。The main idea of empirical mode decomposition(EMD) is separating the series data to components with different scales,say intrinsic mode function(IMF).The EMD method stops decomposition on the numbers of the residue's extrema less than 2.Using this stop condition in EMD will produce many insignificant components and the speed of decomposition is slow.To avoid the shortage of the condition,a new stop condition is proposed,that is,the residue's energy less than the predetermined stops the process of decomposition.A nonlinear simulation signal and a real vibration signal are introduced to study the new condition.The result shows that,the new stop condition could reduce the numbers of the insignificant components and can improve the speed of EMD.The result of the study can be generally used in vibration signal analysis field.

关 键 词：振动信号 经验模态分解 终止条件 能量 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置] TH113.1[自动化与计算机技术—控制科学与工程]