【作者】 李军；

【导师】 张军华；

【作者基本信息】 中国石油大学（华东） ， 地球探测与信息技术， 2013， 硕士

【摘要】 随着石油勘探开发的深入发展，勘探目标已经转向了断块型油气藏这一类的隐蔽油气藏。断层破碎带、低序级断层的分布和形态对断块型油气藏的识别和描述起到越来越重要的作用。由于断层破碎带、低序级断层的构造都很复杂且规模都很小，很难进行地震的识别与描述。因此，系统地研究断层破碎带、低序级断层的识别描述技术，具有十分重要的理论意义和实用价值。通过调研国内外文献，详细介绍了低序级断层及断层破碎带的研究现状。结合断层破碎带及低序级断层的特点，选取了地震属性分析方法、属性融合RGB显示技术、蚁群算法进行详细的研究，并提出了一种比较新的属性提取方法——基于旋转菱形的属性提取方法进行了详细的介绍与应用。地震属性分析方法是地震解释中的一种非常成熟有效且重要的技术，地震属性的种类非常多，应用也很广阔，既能用于储层预测也可用于断层识别描述等。根据属性的特点及其用途，本文选取了其中的三瞬属性、曲率属性进行了断层识别描述应用研究；属性融合RGB显示技术将不同的地震属性融合显示在一张剖面图上，能够大大地提高解释效率和准确性，本文选取几种合适的属性进行了属性融合RGB显示，识别描述断层的效果确实很好。蚁群算法是近年来兴起的一种比较新的算法，它具有良好的鲁棒性、可结合性等优点，应用领域很广阔。本文详细研究介绍了基于灰度值差异的蚁群算法与基于局部极值的蚁群算法两种断层识别描述方法，并将这两种方法应用的实际地震资料的断层识别描述中，结果证明蚁群算法确实可以比较有效地识别描述断层。本文最后一章提出了基于旋转菱形的属性提取方法，方法的原理是：利用菱形对角线所具有的方向特性来判断特殊地质体的方向、范围等特征。选择合适大小的菱形对角线，让菱形以原点为中心，做360°的转动，每转动一个角度，选取合适的属性度量值，计算菱形范围内数据的属性值。这样，最大或最小属性，就最佳地反映了地质体的一些特性。利用该方法可以较好地识别断层，并且对断层破碎带也可以进行判别。更多还原

【Abstract】 With the gain momentum of petroleum exploration and development, the explorationtargets have turned to the fault block reservoirs which belong to the subtle reservoirs. Thedistribution and pattern of the shattered fault zones and low level faults take a more and moreeffort in fault block reservoir identification and description. It is hard to do the seismicidentification and description of the shattered fault zones and low level faults, due to theircomplex structure and small scales. Therefore, it has very important theoretical significanceand practical value for systematically researching the identification and descriptiontechnology of the shattered fault zones and low level faults.This paper introduces the research status of the shattered fault zones and low level faultsin detail, through researching literatures at home and abroad. In consideration of thecharacteristics of the shattered fault zones and low level faults, I select some methods to dodetail research, these methods are seismic attribute analysis technology, RGB attribute fusiondisplay technology and ant colony algorithm. I also propose a new method of seismicattributes extraction based on the rotating rhombus, and detailed introduce its principle andapplication.Seismic attribute analysis is a mature, effective and important technology of seismicinterpretation technologies. There are many different kinds of seismic attributes, and theapplication of seismic attributes is very wide. We can use them for reservoir prediction andfault description, and so on. According to the characteristics and applications of the attributes,this article selects the three instantaneous attributes and curvature attribute to identify anddescribe the shattered fault zones and low level faults. RGB display technology of attributefusion can greatly improve the efficiency and accuracy of seismic interpretation. And we usethis technology identifying and describing the fault, the result is really good. Ant colony algorithm is a relatively new algorithm arisen in recent years. The ant colonyalgorithm has good robustness and combination, and its application field is very broad. In thispaper we introduce two methods in detail, one is ant colony algorithm that based on the grayvalue differences, and the other is ant colony algorithm that based on the local extremum. Thetwo methods both can be used to the identification and description of fault. Then applyingthese two methods to the identification and description of the actual fault of seismic data, andthe results prove that ant colony algorithm can effectively identify and descript the fault.In the last chapter, the new method of seismic attributes extraction based on the rotatingrhombus is carried out. The method finds the direction and range of geologic body mainly bythe directivity of rotating rhombus. Firstly, we need choose the right size of the rhombus’sdiagonal. Secondly, we rotate the rhombus around360degrees in the origin, and record anattribute every angle. Then, a maximum or minimum attribute optimally reflects theanisotropic characteristics of geological body. We can better identify the fault anddiscriminate the fault fracture zone by using this method.更多还原