In the oilfield water flooding process, reservoir properties, micro-pore structure and heterogeneity of dynamic changes will occur. Through the integrated use of multi-disciplinary theoretical approach, the use of computer tools to build a geological model to study the different parameters of the reservoir development environment changes and patterns, to predict the distribution of remaining oil reservoir, to improve the effects of oil development has a very important geological significance. Artificial neural network technology is highly adaptive and self-learning ability, through its strong nonlinear mapping, can precisely establish reservoir parameters and the nonlinear model between logging response. In the geological model, in the history of reservoir data, based on neural network technology , a large number of macro-reservoir data analysis, learning and training, selection of representative reservoir parameters, expressed the well point with reservoir parameters time evolution, and thus predict the distribution of remaining oil.%在油田注水开发过程中,储层物性、微观孔隙结构和非均质性都会发生动态变化.通过综合应用多学科的理论方法,利用计算机手段来构建地质模型,研究不同开发环境中储层参数的变化和规律,对预测储层剩余油的分布规律、提高油田开发效果具有十分重要的地质意义.人工神经网络技术具有极强的自适应和自学习能力,其通过很强的非线性映射,能够精确地建立储层参数与测井响应之间的非线性模型.在地质模型中,历史储层资料的基础上,采用神经网络技术,对大量宏观储层数据进行分析、学习与训练,选取具有代表性的储层参数,表示出各井点储层参数随时间的演变规律,进而有效预测剩余油的分布.
展开▼
