【作者】 李蕾；

【导师】 曹俊兴；

【作者基本信息】 成都理工大学 ， 信号与信息处理， 2009， 硕士

【摘要】 本文介绍地震信号的时频分析方法及其在苏里格气田薄储层识别中的应用。论文的主要内容包括有:(1)介绍了短时傅立叶变换、小波变换、Wigner-Ville变换等几种常用的时频分析方法,讨论了Wigner-Ville变换中交叉项的干扰问题及其抑制方法。(2)通过模型仿真、实际信号分析应用等对上述几种时频分析方法的分辨率及应用条件进行了对比研究工作,认为平滑伪Wigner-Ville变换拥有最好的时频分辨率,同时实现了对交叉项干扰的压制。(3)介绍了Hilbert变换和Hilbert-Huang变换的基本原理,并通过实验分析对比了两种方法在提取瞬时频率时的效果,认为Hilbert变换所要求的信号必须是窄带的或者是平稳的。对于非平稳的或是含有噪声的信号,Hilbert变换求取的瞬时频率就有可能出现负值,因此用Hilbert变换对非平稳的地震信号求取的瞬时频率无实际意义;而Hilbert-Huang变换先通过经验模态分解将信号分解成若干个基本模式分量,再进行Hilbert变换,从而赋予了瞬时频率实际的意义。(4)苏里格气田是我国发现的储量最大的天然气田,构造相对简单但河道砂岩储层薄且横向变化快,在地震记录上识别困难。对苏里格气田某区块的二维地震数据进行平滑伪Wigner-Ville分布、Hilbert-Huang变换时频分析,以井资料为标定,分析研究了地震时频分析方法识别该气田薄储层的能力。实验结果表明平滑伪Wigner-Ville分布能识别厚约10m的砂岩储层,Hilbert-Huang变换能识别10~20m的砂岩储层。更多还原

【Abstract】 This thesis introduces time-frequency analysis of seismic signal and its application to thin reservoir identification in the Su Lige Gas Field. The main content of the thesis are as follow:1. General time-frequency analysis methods are introduced, including short-time Fourier transformation, wavelet transformation, Wigner-Ville distribution; and the disturb terms and suppression methods in Wigner-Ville distribution are discussed.2. It gets the result that the smooth pseudo Wigner-Ville distribution has the best time-frequency resolution and eliminate the disturb terms by comparing these methods.3.The basic theory of Hilbert transformation and Hilbert-Huang transformation are introduced; and comparing two methods in extracting instantaneous frequency gets the result that if signal is non-stationary or contains noise, instantaneous frequency may be negative and lose the practical significance in Hilbert transformation; but instantaneous frequency has the practical significance in Hilbert-Huang transformation, as signal is decomposed several Intrinsic Mode Function by empirical mode decomposition firstly, then in Hilbert transformation.4. Su Lige Gas Field is the largest reserves of natural gas fields in china. The characteristic is relatively simple structure but thin sandstone reservoirs and fast variation in horizontal so that it is difficult to be identified in seismic records. The Use of smooth pseudo Wigner-Ville distribution and Hilbert-Huang transformation analyzed 2-D seismic data of one block in Su Lige Gas Field. According as well data, the study on th??solution of seismic time-frequency analysis identify thin reservoir. The result shows that smooth pseudo Wigner-Ville distribution can identify about 10m sandstone reservoir; Hilbert-Huang transformation can identify 10~20m sand bodies.更多还原