地震信号谱反演方法研究

发布时间：2018-03-24 23:12

本文选题：反射系数反演　切入点：SSF算法　出处：《电子科技大学》2015年硕士论文

【摘要】：随着对岩性油气藏勘探开发的不断深入,薄层问题日益凸显,提高薄储层识别和预测的精度成为地质学家们密切关注和急需解决的问题。基于高分辨率分频技术的地震信号谱反演为提高地层资料分辨率从而提高薄层预测精度提供了新的思路。近年来,随着国内外学者对谱反演技术研究的不断深入,谱反演成为解决薄储层预测的热点问题。然而谱反演技术在国内外的研究开展不久,还存在一系列的技术问题没有达成普遍共识。本文针对地震信号谱反演方法开展了一系列研究,深入研究了谱反演相关理论,建立了一套较完整的反演算法流程,主要工作包括:1.研究了地震子波相关理论,讨论了子波提取对谱反演算法反演结果的影响。其中,重点研究了用Ricker子波模拟地震子波进行反演的情况和统计子波的提取方法,同时给出了子波提取的流程图。2.研究了地震记录及子波的时频域表示,总结出了地震数据傅里叶变换的处理流程,同时探讨了傅里叶变换中时窗宽度及滑动步长的确定等关键问题。3.研究了反射系数奇偶分量在反演中对分辨率的影响,并论述了反射系数奇偶分解方法。4.详细推导了谱反演目标函数的构建过程,论述了求解目标函数的经典方法-共轭梯度法。5.重点研究了谱反演目标函数求解中的最优化方法,可分离替代函数(SSF)算法以及正交匹配追踪(OMP)算法,给出了相关的算法流程及仿真实例,并对重要的反演参数如何取值进行了深入分析。6.完成了所有算法模块代码的编写,对多种典型的地质模型做了仿真测试,对不同工区的实际地震资料进行了反演和测试,对两种反演算法实际地震资料的反演效果进行了对比分析。
[Abstract]:With the deepening of exploration and development of lithologic reservoirs, thin layer problems become increasingly prominent. Improving the accuracy of thin reservoir identification and prediction has become a problem that geologists pay close attention to and need to be solved. Seismic signal spectral inversion based on high resolution frequency division technique is to improve the resolution of formation data and improve the precision of thin layer prediction. Degree provides a new way of thinking. In recent years, With the deepening of the research on spectral inversion technology, spectral inversion has become a hot issue to solve thin reservoir prediction. There is still a series of technical problems that have not reached a general consensus. In this paper, a series of research on seismic signal spectral inversion method is carried out, and the related theory of spectral inversion is deeply studied, and a set of complete inversion algorithm flow is established. The main work includes: 1. The correlation theory of seismic wavelet is studied, and the influence of wavelet extraction on the inversion result of spectral inversion algorithm is discussed. Among them, the inversion of seismic wavelet simulated by Ricker wavelet and the extraction method of statistical wavelet are studied emphatically. At the same time, the flowchart of wavelet extraction is given. 2. The seismic records and the representation of wavelet in time-frequency domain are studied, and the processing flow of Fourier transform of seismic data is summarized. At the same time, the key problems such as the width of time window and the determination of sliding step size in Fourier transform are discussed. 3. The effect of even-odd component of reflection coefficient on the resolution in inversion is studied. The method of even-odd decomposition of reflection coefficient. 4. The construction process of objective function of spectral inversion is derived in detail, and the classical method of solving objective function, conjugate gradient method. 5. The optimization method for solving objective function of spectral inversion is studied emphatically. The separated alternative function (SSF) algorithm and the orthogonal matching tracking (OMP) algorithm are given. The relevant algorithm flow and simulation examples are given, and the important inversion parameters are analyzed deeply. 6. The code of all the algorithm modules is written. Several typical geological models have been simulated and tested. The inversion and test of the actual seismic data in different working areas have been carried out. The inversion results of the two inversion algorithms have been compared and analyzed.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P631.4

【参考文献】