【作者】 邢玉忠；

【导师】 樊太亮；

【作者基本信息】 中国地质大学（北京） ， 油气田开发工程， 2007， 博士

【摘要】 裂缝发育的普遍性、裂缝分布的复杂性及裂缝性储层的重要性决定了开展裂缝性油藏研究的必要性和艰巨性。对该类油藏研究的目的是建立比较客观的裂缝网络模型,提高对裂缝空间分布状况的认识,为油田勘探和开发提供符合客观实际的依据。论文以辽河油田沈229潜山油藏为例,综合各种来源不同、尺度不同的资料,通过多学科进行裂缝的描述,然后利用嵌套随机模拟的方法得到可靠的裂缝网络模型,主要研究内容及成果包括:1.利用现场资料进行裂缝的识别及裂缝表征参数描述,形成了一套比较系统和完善的技术方法。2.利用地质统计和模糊神经网络进行裂缝属性参数的统计和分析,把裂缝的非均质性和资料的不同尺度性有机的结合起来,得到了裂缝属性参数的统计和空间分布特征。3.以裂缝属性的统计分布为基础进行裂缝属性参数的模拟,把裂缝密度和多级分数维整合到离散模型中,解决了裂缝空间分布的复杂性。4.利用嵌套模拟技术把地质成因形成的大尺度裂缝和通过岩心及测井解释得到小尺度裂缝有机结合起来,建立了反映储层裂缝及其属性客观分布的裂缝网络模型。5.所建立的沈229潜山裂缝网络模型在油田开发调整和部署实践中取得了良好的效果,证明该种方法对于研究裂缝性潜山油藏是可行有效的。通过该课题的研究得到了以下主要结论:1.裂缝性储层的非均质性和复杂性决定了利用一种学科或一种方法难以解决裂缝的空间分布,把各种来源和各种尺度的数据有机结合起来是裂缝属性参数表征和统计分析的关键。2.裂缝密度分布可用于裂缝位置分布研究,而多峰圆周统计和多级分形几何则更适合于定量研究裂缝的方位和大小。3.模糊逻辑可以有效地实现对不同来源数据重要性的评价;同时可得到裂缝的空间分布及实现裂缝密度分布和多级分形值的定量化。4.综合裂缝网络模型比基于目标的随机模拟和基于网格的裂缝模型更具有先进性,其结果可以与其他地质建模结合起来作为输入参数进行数值模拟研究。更多还原

【Abstract】 It’s essential and hard to carry out the study of fractured reservoirs because of the complexity of fracture distribution and the importance of fractured reservoir. The objective of fractured modeling is build an objective fracture network to provide geological basis for the exploration and development of oilfields.Shen229 buried hill is taken as an example in this article, fracture is described with data from different origins and different dimensions, then obtained reliable discrete fracture net model through nesting stochastic simulation, the main contents and results are as follows:1. Identify fracture and describe its parameters with spot data.2. Analyze fracture properties using geological statistics and fuzzy neural network, the method resolve the reservoir inhomogeneity and data different size.3. Simulate fracture parameters based on the statistic characteristics of fracture and explain the complexity of fracture distribution integrated the fracture density and multilevel fractal dimension to discrete model.4. Simulate the distribution of large-scale and small-scale through nesting simulation, the model can reflect the actual distribution of fracture.5. The method have been used shen229 buried hill and proved it is effective in fractured reservoir modeling.And the following conclusions are obtained from this study:1. It’s important to combine the data from different origin and scale in the denotation and statistics of the characteristics parameters of fracture. This method can solve the problem of the complexity and of fractured reservoir.2. It’s shown that fracture density can denote the distribution of fracture, circular statistics and multilevel fractal can be used to study quantitatively the fracture’s orientation and scale, respectively.3. Fuzzy neural network is useful in evaluation of the importance of data from different origin, the method can quantify fracture spatial distribution.4. Integrated stochastic simulation is more advanced than simulation based only on objective or grids, the outcome can be used as input parameter of reservoir simulation.更多还原