数据挖掘技术在气田压裂效果评价中的应用  

Application of data mining to postfracture response evaluation in gas field

作　　者：孟庆民[1,2] 张士诚[1] 王莉[2] 黄侠[2] 张俊璟[1] 

机构地区：[1]中国石油大学石油工程教育部重点实验室,北京102249 [2]中石化胜利油田分公司采油工艺研究院,山东东营257000

出　　处：《中国石油大学学报（自然科学版）》2008年第5期165-169,共5页Journal of China University of Petroleum(Edition of Natural Science)

摘　　要：通过对数据挖掘体系结构的研究,结合气田压裂施工的特点,建立了一种适合于气井压裂效果评价的技术体系。该体系数据仓库采用SQL Server 2005实现,数据挖掘算法采用决策树和神经网络模型,可以实现气井压裂的选井选层和施工参数的优化。应用结果表明,该数据挖掘体系集成度高,预测精度高,可为压裂决策提供参考。Based on data mining architecture, a technical evaluation system suitable for gas well fracturing was established by combining the fracturing treatment characteristics of the gas reservoir. Data warehouse adopts SQL Server 2005 and data mining algorithms use decision trees and neural network model. Through the comparison and analysis of the two models, the well selection and fracturing interval selection standards and parameter optimization can be implemented. This data mining architecture has high integration and accuracy which can provide reference for fracturing decision.

关 键 词：数据挖掘 数据仓库 决策树 神经网络 压裂效果 

分 类 号：TE319[石油与天然气工程—油气田开发工程] TE357.13