空间信息场约束下的随机建模方法研究

发布时间：2017-12-26 16:21

本文关键词：空间信息场约束下的随机建模方法研究　出处：《电子科技大学》2016年硕士论文　论文类型：学位论文

【摘要】：随机建模作为油藏描述的核心技术,贯穿油田的整个生命周期。其目的是根据已知井信息,应用地质统计学,推测复杂的地下结构可能的分布特征,产生一系列等概率、高精度的三维储层模型供选择。随机建模主要分为两个部分:变差函数的计算,即通过已知井之间的信息去推测未知的地质变量的空间变化特征;随机模拟,即在变差函数的指导下通过空间内插的方法去模拟生成未知的地质模型可能的样子。传统的随机建模中只利用了已知的井数据,但是井数据普遍存在井网密度小、分布不均匀且分布范围远小于研究区域的问题,对于庞大的整个地质体而言,仅仅是管见所及,缺乏整体的三维空间信息,导致建模的结果难免会是狭隘片面的。针对这些问题,本文充分利用地震资料覆盖面广、横向采集密度大、包含大量空间信息的优点,弥补了井间信息的不足,从变差函数的构建和随机模拟两方面着手,提出了一种空间信息场约束的随机建模方法。本文的主要贡献如下:1.针对由单一的井数据得到的变差函数缺乏空间信息,不能有效地反映地质变量的横向变异性,本文提出了一种空间信息场约束下的非参数各向异性变差函数构建方法。该方法综合运用井数据和地震数据,通过地震反演获取空间信息场,对反演得到的波阻抗体按照角度和距离进行均匀抽样,计算所有水平切片的各个方向的变差函数,然后对变差函数进行融合,得到最终的非参数各向异性变差函数。该方法的优点在于,利用地震数据横向连续性好、空间信息丰富的优势,有效地提高了变差函数的可靠性,并且计算过程简单方便,采用“非参数”避免了传统方法中繁琐的套合步骤。2.针对仅仅依靠井数据得到的随机建模结果只是再现了宏观的统计规律,会不同程度地偏离地质真实结果,本文提出了一种空间信息场约束下的随机建模方法。该方法的主要特点是用空间信息对建模过程进行一定的约束,具体为以上文提出的非参数各向异性变差函数为指导,充分展现了地质变量的空间相关性,同时在随机模拟的过程中创造性地加入了地震反演的步骤,不再以水平切片为单位逐层进行模拟,而是以地震道为单位逐道进行模拟,一边模拟一边反演,将反演之后的地震道加入已知数据中继续模拟、反演,反复利用空间信息,建立一个高可靠性的模型。
[Abstract]:Random modeling, as the core technology of reservoir description, runs through the whole life cycle of the oil field. Its purpose is to apply geostatistics to predict the possible distribution characteristics of complex underground structures based on known well information, and to generate a series of equal probability and high accuracy 3D reservoir models. Stochastic modeling is mainly divided into two parts: the calculation of variogram, the spatial variation is well known by the information to infer the unknown geological variables; stochastic simulation, geological model in the variogram guided by spatial interpolation method to simulate the unknown might look like. Only the well known data using stochastic modeling in traditional data exists but the well spacing density, uneven distribution and the distribution range is far less than the study area, for the entire geological body is huge, only limited outlook, three-dimensional spatial information lack of overall, the result will inevitably be narrow and one-sided modeling the. To solve these problems, this paper makes full use of the advantages of wide coverage, the horizontal seismic data acquisition density, contains a large number of spatial information, to make up for the lack of information between wells, two aspects from the construction and stochastic simulation of the variation function, put forward the stochastic modeling method of spatial information field constraints. The main contributions of this paper are as follows: 1., aiming at the lack of spatial information obtained from single well data, it can not effectively reflect the lateral variability of geological variables. In this paper, a method of constructing nonparametric anisotropic variogram constrained by spatial information field is proposed. This method utilizes well data and seismic data, the seismic inversion to obtain spatial information field, the wave impedance inversion was carried out in accordance with the angle and distance of uniform sampling, to calculate the direction of all horizontal slice variogram, then the variogram fusion, non parameter anisotropic final variogram. The advantage of the method is that the reliability of variogram is effectively improved by the advantage of good lateral continuity and abundant spatial information, and the calculation process is simple and convenient. The "non parameter" is adopted to avoid the cumbersome application of traditional methods. 2., for stochastic modeling results only relying on well data, they only reproduce the macroscopical statistical laws, and deviate from the real geological results in varying degrees. In this paper, a stochastic modeling method constrained by spatial information field is proposed. The method is characterized by certain constraints on the modeling process with spatial information, non parametric anisotropic concrete above proposed variation function as a guide, fully demonstrated the spatial correlation of geological variables, at the same time in the process of stochastic simulation in a creative way to join the seismic inversion steps, not to slice as a unit layer by layer simulation, but by seismic units by channel simulation, simulation of side side inversion, the known data to simulation, inversion of seismic trace inversion after the repeated use of spatial information, establish a high reliability model.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TE319

【相似文献】