摘要
在HHT方法中,边界处理问题是经验模态分解过程中的核心问题,处理不好会引起数据端部的“飞翼”,进而“污染”整个数据,使分解的固有模态函数失效。解决此问题有两种方法:在数据两端加子波和在数据两端各添加两个极值点。对仿真信号的处理证明这两种方法都是有效的,而后一种方法效果更好。利用后一种方法,对1996年丽江7.0级地震的一条强余震加速度记录进行了分析,加深了对强地震动特性的了解。
Boundary-processing is crucial to EMD process in HHT, whose failure may cause the invalidity of the decomposed IMF due to boundary fluctuation and the whole data invalidity. We propose two methods to solve this problem, one of which is the addition of wavelet at the data's ends, and the other is the addition of extreme points at the data's ends. The two methods prove to be effective by the processing of emulation signals, and the latter is more effective. We use the proposed methods to analyze an acceleration record of the strong aftershock in the Lijiang Ms7. 0 earthquake in 1996, and therefore achieving a better understanding of strong motion characteristics.
出处
《地震研究》
CSCD
北大核心
2007年第1期54-58,共5页
Journal of Seismological Research
基金
云南省"十五"重点项目--昆明高(超高)层建筑中近场强震响应研究(项目编号:[2002]54号02-02)
关键词
经验模态分解
端点飞翼
三次样条
丽江地震
Empirical Mode Decomposition, boundary fluctuation, spline, Lijiang earthquake