基于聚类经验模态分解的地球天然脉冲电磁场时频与能量谱分析：以芦山Ms7．0地震为例  

Time-frequency characteristics and energy analysis of the Earth's natural pulse electromagnetic fields based on ensemble empirical mode decomposition:The Lushan M_S7.0 earthquake as an example

作　　者：郝国成[1,3] 龚婷[1] 董浩斌[2] V.G.SIBGATULIN 陈忠昌[1] Alexey KABANOV 

机构地区：[1]中国地质大学(武汉)机械与电子信息学院,湖北武汉430074 [2]中国地质大学(武汉)自动化学院,湖北武汉430074 [3]教育部三峡库区地质灾害研究中心,湖北武汉430074 [4]Environmental Center for Resources Conservation,Krasnoyarsk 660049,Russia

出　　处：《地学前缘》2015年第5期231-238,共8页Earth Science Frontiers

基　　金：湖北省自然科学基金面上项目(2014075101);中国地质大学(武汉)实验技术研究项目(SJ-201423);中央高校基本科研业务费专项(2012079104)

摘　　要：针对地球天然脉冲电磁场信号的非平稳、非线性特点,本文采用基于聚类经验模态分解(EEMD)提取信号时频特性的方法,有效获得了芦山MS7.0地震前地球天然脉冲电磁场信号的时频分布特性、瞬时能量谱、能量集中分布的频段、最大振幅对应的时频分布等特性。对比经验模态分解(EMD)的希尔伯特-黄变换(HHT)方法,EEMD有效抑制了以往EMD分解过程中所出现的模态混叠问题。文章还将EEMD和傅里叶变换、小波变换进行了对比研究。结果表明,对于非平稳的地球天然脉冲电磁场数据,采用EEMD分解的HHT方法更能反映出原始数据的多种固有特性,便于进一步了解地震前地球天然脉冲电磁场的特点。Aiming at non-stationary and nonlinear signals like the Earth＇s natural pulse electromagnetic fields, this paper adopted a clustering algorithm based on ensemble empirical mode decomposition（EEMD）, which effectively extracted the instantaneous energy spectrum, energy concentration distribution hands, time and frequency corresponding to the maximum amplitudes and other time-frequency characteristics before the Lushan Ms 7.0 earthquake. Compared with the Hilbert-Huang transform （HHT） method in the decomposition o{ empirical mode decomposition （EMD）, EEMD method can effectively suppress the mode mixing effect existing in EMD decomposition process. The paper also made a comparative study among this method and Fourier transform and wavelet analysis. The result shows that using HHT method based on EEMD decomposition to the unsteady data of the Earth＇s natural pulse electromagnetic fields is better for representing inherent characteristics of the original data, and thus EEMD method owns higher validity.

关 键 词：聚类经验模态分解 希尔伯特-黄变换 时频分析 地球天然脉冲电磁场 

分 类 号：P315.63[天文地球—地震学]