节点文献

三维地质建模与射线正演方法研究

Study on 3-D Geological Modeling and Seismic Ray Tracing Methods

【作者】 卞爱飞

【导师】 於文辉;

【作者基本信息】 中国地质大学 , 固体地球物理学, 2007, 硕士

【摘要】 当前勘探地震学的发展趋势是地震资料采集、处理、解释一体化。一体化过程的核心是三维地质模型。随着勘探的逐步深入,在已有地质、地球物理研究成果的基础上建立从区域到局部的不同分辨率、不同属性的三维地质模型是可能的。三维地质建模为建立三维地质模型提供了必要保障。将三维地质建模与地球物理应用相结合,可以促进相关学科的共同发展。本文通过对三维建模方法和射线正演方法进行联合研究,将构造地质建模研究成果成功引入射线正演中,分别完成了基于块状模型的三维快速两点射线追踪方法和基于射线坐标系中三角网拓扑描述的波前重建射线追踪方法两项研究内容。本文研究成果总结为如下几点:(1)对三维地质建模元素及相关核心技术进行分析总结,探讨常用含断层构造三维地质建模方法的特点与不足,利用现有软件模块对两个地区的资料进行建模。(2)在建模数据成果的基础上,实现基于块状模型和射线编码的针对目标层位的快速两点射线追踪。模型各个块体内速度为常速,块体之间以界面相联系,界面用三角网描述,能够较好地描述复杂地质构造。利用射线路径编码进行针对目的层的全路径迭代射线追踪。该编码方法可以方便地定性描述位于模型不同区域的射线,并能够适应复杂观测系统下常规P波、转换波、多次波的射线追踪要求。在复杂构造区通过射线冗余检查剔除不合理的射线,保留有效波的信息。利用C++语言编制了块状模型射线计算程序,给出VSP计算实例。(3)波前重建射线追踪方法中引入地质建模中三角网拓扑的思想构建射线坐标系中的波前面,模拟地震波在目标区域的传播过程。对处于射线管内的成像网格点利用四面体剖分和线性插值方法对其进行属性插值。传播过程中射线密度降低时通过分裂波前面上的三角形并直接从源点引入新射线以保证射线密度满足给定要求。利用C++语言编制程序,给出三维低速体模型中的计算示例。(4)结合射线具体应用改进了模型空间的搜索效率。在块状模型射线正演中,通过预先建立三角网分区索引结构和三角形外接矩形框等措施提高了插值时三角形搜索的效率,保证了在实际模型数据量大的情况下射线追踪方法的实用性;点与体关系判定时利用光线碰撞快速检测模型任一点所处的块,保证了块状模型中两点射线追踪方法对复杂观测系统的适应性。在波前重建射线追踪方法中,利用射线管外接包围盒快速排除了不需要做插值的网格,利用四面体剖分确定需要做插值的点,提高了插值效率。

【Abstract】 Nowadays the development trend of exploration seismology is the integration of seismicdata acquisition, processing and interpretation. 3-D geological models are the key of this process.Along with the development of exploration and exploitation, it is possible to buildmulti-resolution geological models with different attributes from regional to local scale, based onthe previous geological and geophysical studies. Geological modeling provids necessaryguarantee for building different kind of models. The combination of geological modeling andgeophysical applications can prompt the development of related disciplines. In this article, thejoint study of 3-D geological modeling and ray tracing methods is carded out. The geologicalmodeling results are successfully introduced into seismic ray tracing studies. Two researchcontentses are completed respectively: 3D fast two-point ray tracing in block models andwavefront construction with triangulated surface in ray coordinates. The results of this paper areas follows:(1) The modeling elements and related core technology are analysised and concluded. TheCharacteristics and disadvantages of current fault modeling methods are discussed. Two real datamodeling examples are shown with geological modeling modules.(2) Based on model data from modeling process, a fast two point ray tracing algorithm inblock models is developed for geometry optimization in seismic data acquisition. The forwardmodel consists of serveral blocks, whose physical parameters are constant in it. The surface ofeach block is represented by TIN, which is suitable for complex geological models. Ray pathcode is used to discribe different types of rays and rays in different parts of the model. Wholepath iteration method is used for ray path calculations. Redundancy check is used to check thelegitimacy of traced rays in complex structrues. A ray tracing code is developed with C++.Examples of VSP ray tracing in different models are shown.(3) The topology of Tin structure is introduced into wavefront construction ray tracingstudies. Each wavefront is represented by triangle network in ray coordinates. Initiated fromsource, the wavefront can expand to the whole region of the model.The ray density on thewavefront is controlled by dynamic checking of triangle qualities. New rays are introduced bydirect ray tracing from source, which provides a higher accuracy. Kenematic and dynamicproperties of rays are interpolated on fine migration grids in ray tubes. A wavefront construction ray tracing code in C++ is developed and a low velocity model is tested.(4) Model space searching efficiency is improved in specific ray tracing applications. On raytracing for block models, regularly partition index of surface triangles and rectangle of eachtriangle are built for fast point positioning on surface. Collision detection is used for pointpositioning in block. The upon methods greatly improve ray tracing efficiency in models withlarge data sets. On wavefront construction ray tracing, rectangular box is used for fast excludingpoints from ray tubes. Tetrahedron is used for point positioning, and a linear method is used forfast interpolation of different kind of attributes in each tetrahedron.

  • 【分类号】P631.4;P628.3
  • 【被引频次】2
  • 【下载频次】650
节点文献中: 

本文链接的文献网络图示:

本文的引文网络