基于奇异谱分析的脉冲型地震动量化识别方法

发布时间：2019-02-16 13:22

【摘要】：二十世纪八十年代发生的几次大地震引发了地震工程界学者们对脉冲型地震动的广泛关注。深入研究地震时程中的脉冲特性,对工程结构抗震设防以及减轻震害都具有重要意义。与此同时,通过对地震动记录的量化识别以获得大量的脉冲数据也成为近年来的热点研究课题。现有的量化识别方法大多需要基于假定的脉冲模型进行脉冲的提取进而识别脉冲,该类方法一般不能识别多脉冲。针对现有方法存在的问题,本文提出基于奇异谱分析(SSA)的脉冲型地震动量化识别方法,主要研究内容和结论如下:(1)对比分析几种常用的时频信号分解方法,并分别运用小波变换、Hilbert-Huang变换和奇异谱分析(SSA)方法对同一条非稳态时程信号进行滤波分析。结果表明:SSA方法能更为精确的描述和提取原始信号中的真实信号成分,同时SSA方法还具有计算速率较快的优点。(2)基于地震动脉冲特性及SSA方法理论,并借鉴现有识别方法的研究思路,本文提出基于SSA的脉冲型地震动量化识别方法。具体研究内容包括:1.选定合理的地震动数据库;2.对疑似单、多脉冲地震动粗略分类,并相应地运用拟速度反应谱法或“适用零点法”分别确定近似脉冲周期;3.基于奇异谱分析法,结合Hurst指数进行时程滤波;4.基于奇异谱分析法分解提取脉冲成分,并利用傅里叶变换确定脉冲周期;5.根据提取成分奇异值所占比例给定合理的判别指标,实现单、多脉冲型地震动的有效识别。(3)对比分析小波分析法、能量法及EMD方法,并分别就单脉冲地震动记录和多脉冲地震动记录进行提取判别,阐述具体的计算过程。针对本文选取的353条地震动记录,分别应用小波分析法和能量法进行计算判别,并对比SSA法的识别结果,对问题数据进一步分析,判断SSA法识别脉冲型地震动的优劣性。结果表明,SSA法相较于小波分析法和能量法能更有效地实现多脉冲型地震动量化识别,以奇异值作为脉冲量化指标具有可靠的精度。
[Abstract]:Several large earthquakes occurred in 1980's have aroused widespread concern of seismic engineering scholars on pulsed ground motion. It is of great significance for engineering structures to study the pulse characteristics in the seismic process and to reduce the earthquake damage. At the same time, it has become a hot research topic in recent years to obtain a large number of pulse data by quantifying the ground motion records. Most of the existing quantization recognition methods need to extract the pulse based on the assumed pulse model to identify the pulse, but this kind of method can not recognize the multi-pulse. Aiming at the problems existing in the existing methods, this paper proposes a method for quantifying the ground motion based on singular spectrum analysis (SSA). The main research contents and conclusions are as follows: (1) several commonly used time-frequency signal decomposition methods are compared and analyzed. Wavelet transform, Hilbert-Huang transform and singular spectrum analysis (SSA) are used to filter the same unsteady signal. The results show that the SSA method can more accurately describe and extract the real signal components from the original signal. At the same time, the SSA method also has the advantage of fast calculation rate. (2) based on the characteristics of ground motion pulse and the theory of SSA method, Based on the research ideas of existing methods, this paper presents a quantification recognition method of impulse ground motion based on SSA. The specific research contents include: 1. Select reasonable ground motion database; 2. The approximate pulse period is determined by pseudo-velocity response spectrum method or "applicable zero point method". Time history filtering based on singular spectrum analysis and Hurst exponent. 4. The pulse component is extracted based on singular spectrum analysis, and the pulse period is determined by Fourier transform. 5. According to the proportion of extracted singular value, a reasonable discriminant index is given to realize the effective identification of single and multi-pulse ground motion. (3) the wavelet analysis, energy method and EMD method are compared and analyzed. The single pulse ground motion records and multi-pulse ground motion records are extracted and discriminated, respectively, and the concrete calculation process is expounded. According to 353 ground motion records selected in this paper, wavelet analysis method and energy method are used to calculate and discriminate, and the results of SSA method are compared, and the problem data are further analyzed to judge the merits and demerits of SSA method in the identification of impulse ground motion. The results show that the SSA method is more effective than wavelet analysis and energy method in quantization identification of multi-pulse ground motion, and the singular value is used as the pulse quantization index with reliable precision.
【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】：2017
【分类号】：P315.9

【参考文献】