基于EMD方法量化识别近断层脉冲型地震动

发布时间：2018-06-29 12:20

本文选题：近断层脉冲型地震动 + EMD方法　；参考：《哈尔滨工业大学》2015年硕士论文

【摘要】：受向前方向性效应或滑冲效应的影响,在近断层附近会产生脉冲型地震动。相较普通远场地震动,脉冲型地震动已经被证明会对工程结构造成更为严重的破坏。大量的学者就脉冲型地震动的特性及其对工程结构的影响展开了深入的研究,关于脉冲型地震动的量化识别更是近年来的热点研究课题。目前大多数的脉冲识别方法都只能在速度时程上提取单脉冲,较难实现对多脉冲地震动的量化识别。基于EMD(Empirical Mode Decomposition)方法,本文对脉冲型地震动的量化识别进行了详细研究,主要研究内容和结论如下:⑴分别采用窗口傅里叶变换、小波变换、Hilbert-Huang变换对地震动记录的速度时程进行时频分析,得到对应的时间-频率图。结果表明:Hilbert-Huang变换较前两种方法在刻画地震动信号的局部特性方面具有较大优势。⑵由于地震动的复杂性,在采用基于EMD方法识别脉冲型地震动之前,需要对地震动数据进行分类,为了理解使用EMD方法识别脉冲的过程,分别对单脉冲型和多脉冲型地震动进行识别。结果表明:EMD方法不仅能够识别出单脉冲地震动,还能够较好地识别出多脉冲型地震动。⑶分别利用小波分析法、能量法以及EMD方法,对91条已知脉冲型地震动和本文所选336条地震动进行识别,并对得到的结果进行了对比分析,判断EMD法识别脉冲地震动的优劣性。结果表明,对于前4阶IMF(Intrinsic Mode Function)能量比之和小于0.3的地震动,EMD方法的识别效果更好;对于前4阶IMF能量比之和大于0.3的地震动,前两种方法的识别效果较好;对于所选的大多数地震动,EMD法在识别多脉冲地震动上展现出优越性,而前两种方法更适合识别单脉冲型地震动。
[Abstract]:Under the influence of forward directional effect or slip effect, pulsed ground motion will occur near the fault. Compared with ordinary far field ground motion, pulsed ground motion has been proved to cause more serious damage to engineering structure. A large number of scholars have carried out in-depth research on the characteristics of pulsed ground motion and its influence on engineering structure, and the quantification recognition of pulsed ground motion has become a hot research topic in recent years. At present, most pulse recognition methods can only extract single pulse in velocity time history, so it is difficult to realize quantification recognition of multi-pulse ground motion. Based on the empirical Mode decomposition (EMD) method, the quantization recognition of pulsed ground motion is studied in detail in this paper. The main contents and conclusions are as follows: 1. Wavelet transform Hilbert-Huang transform is used to analyze the velocity time history of ground motion record, and the corresponding time-frequency map is obtained. The results show that the ratio Hilbert-Huang transform has a great advantage over the former two methods in describing the local characteristics of the ground motion signal. 2. Because of the complexity of the ground motion, the method based on EMD is used to identify the pulsed ground motion. It is necessary to classify the ground motion data. In order to understand the process of using the EMD method to identify the pulses, the single and multi-pulse ground motions are identified respectively. The results show that the weight EMD method can not only recognize the single pulse ground motion, but also better identify the multi-pulse ground motion .3 using wavelet analysis, energy method and EMD method, respectively. 91 known impulse ground motions and 336 ground motions selected in this paper are identified, and the results obtained are compared and analyzed to judge the merits and demerits of the EMD method in the identification of impulse ground motions. The results show that the EMD method is better for the ground motion with the sum of the first four order IMF energy ratio less than 0.3, and the first two methods for the ground motion with the sum of the first four order IMF energy ratio greater than 0.3. For most of the selected ground motions, the EMD method is superior in the identification of multi-pulse ground motions, while the former two methods are more suitable for the identification of monopulse ground motions.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P315.9

【参考文献】