油田开采智能评价有用规则提取方法  

Investigation into useful rule extracting for intelligent evaluation of oilfield production

作　　者：于镝[1] 宋盛瑶 任伟建[1] 康朝海[1] 刘玉敏[1] 

机构地区：[1]东北石油大学电气信息工程学院,黑龙江大庆163318

出　　处：《东北石油大学学报》2016年第3期111-117,共7页Journal of Northeast Petroleum University

基　　金：国家青年科学基金项目(51404074);中国石油科技创新基金研究项目(2012D-5006-0205)

摘　　要：基于油田开采智能评价的有用规则是通过专家问卷调查产生的,受专家选择和实践检验的难度,以及某些专家经验规则可能与实际数据不一致影响而具有明显的局限性。采用基于区间Ⅱ型模糊集合的语言摘要方法,从油田开采数据库中产生IF-THEN规则,并根据真实度、覆盖度、可靠度和离群度等4个质量测度提取油田开采评价有用规则;进而基于平行坐标法及GUI界面建立可视化逻辑规则群。应用结果表明:该方法节省专家的时间,有助于在经验与实际数据不一致基础上完善专家经验规则。Useful rules for intelligent evaluation of oilfield production are usually generated based on expert questionnaire survey.However,there are some shortages for the survey such as difficulties of expert selection and practice test,and possible inconsistency between obtained rules and field data.So aimed at the problem of extracting the useful rules for oilfield production intelligent evaluation from the actual data of oilfield in the paper,linguistic summarization based on type-Ⅱfuzzy sets is applied to generate IF-THEN rules from oilfield production database.Then useful rules are extracted according to four quality measures including truth,coverage,reliability and outlier.Moreover,visual logic rules groups are established based on the parallel coordinates method and GUI interface.The proposed method,used to extract useful rules for oilfield production intelligent evaluation,not only saves time of experts but also helps improve rules considering the inconsistency between obtained rules and field data.

关 键 词：区间Ⅱ型模糊集合 语言摘要 规则提取 智能评价 可视化 逻辑规则群 油田开采 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]