油田智能化辅助决策技术综述

Summary of Intelligent Aided Decision Technology in Oil Field

摘要: 传统决策技术由于自身缺乏挖掘数据背后隐藏知识的能力，已无法有效从油田海量数据中获取有用信息或知识，无法对决策提供有效支持。利用智能决策技术可以增强决策开发和利用的能力，改善决策的智能化水平，并能以油田分散异构的数据为支撑进而解决油田生产问题。论文从油田复杂决策存在的困难入手，了解了油田由传统到智能决策的技术改进；着重梳理了目前主流智能决策技术，并根据技术底层原理不同归纳为由机器学习驱动、专家系统驱动、数据仓库与数据挖掘驱动以及预测算法驱动这4大类技术，深入探究了不同技术的实现原理与架构；分析了不同技术解决石油领域问题的应用场景；最后提出未来多技术交叉融合的群决策技术的发展方向，以期帮助科学选择辅助决策技术，进而提高油田综合经济效益。

Abstract: Traditional decision-making technology itself lacks the ability to mine the hidden knowledge behind the data, and has been unable to effectively obtain useful information or knowledge from the massive data of oil field, and can not provide effective support for decision-making. The use of intelligent decision-making technology can enhance the ability of decision-making development and utilization, improve the level of intelligent decision-making, and solve the problem of oilfield production with the support of dispersed heterogeneous data. This paper starts with the difficulties of complex decision-making in oil field, and understands the technical improvement from traditional to intelligent decision-making in oil field. This paper focuses on sorting out the current mainstream intelligent decision-making technologies, and according to the different underlying principles of technology, it is summarized into four categories: machine learning driven, expert system driven, data warehouse and data mining driven, and prediction algorithm driven. The implementation principles and architectures of different technologies are deeply explored. The application scenarios of different technologies to solve problems in the petroleum field are analyzed. Finally, the development direction of group decision-making technology with multi-technology cross integration in the future is proposed, in order to help scientifically select auxiliary decision-making technology and improve the comprehensive economic benefits of oil fields.