贝叶斯判别方法在产能预测中的应用  

APPLICATION OF BAYES DISCRIMINATING METHOD IN THE DELIVERABILITY PREDICTION

作　　者：张姮妤[1] 刘江 邓廷勇[3] 

机构地区：[1]绥化学院数学与信息科学学院,黑龙江绥化152061 [2]大庆钻探工程公司测井公司,黑龙江大庆163412 [3]黑龙江八一农垦大学文理学院,黑龙江大庆163319

出　　处：《大庆石油地质与开发》2011年第4期167-170,共4页Petroleum Geology & Oilfield Development in Daqing

基　　金：绥化学院杰出青年基金项目（SJ10018）资助.

摘　　要：油层产能评价是油田勘探开发过程中的重要环节，也是测井评价技术的难题之一。在常规产能预测分析的基础上，提出按物性分类进行贝叶斯判别，使每一子类具有相近的油气渗流特性和油层污染状况，既减轻了采用测井资料难以确定表皮系数、微裂缝带来的影响，又减小了油相渗透率求取不准带来的误差。结合庆新油田储层泥质含量、孔渗变化大的特点，将油层分为纯砂岩、低含泥砂岩、高含泥或泥质砂岩3类油层，分别给出贝叶斯判别模型并按原油产量划分为4个级别，判别符合率达到80．1％，应用效果良好。Reservoir deliverability prediction is a key step in oilfield exploration and development, and a difficult issue in logging evaluation as well. Therefore, Bayes discrimination based on property classification is proposed on the foundation of conventional performance prediction analysis. Every property subclass has similar hydrocarbon seepage properties and reservoir pollution, which reduces not only the influence of uncertainty in skin factor and micro-fracture predicted by logging data, but also the error due to inaccurate calculation of oil-phase permeability. Considering the characteristics of shale content in reservoir and great variation of porosity and permeability of Qingxin Oilfield, formations are categorized into 3 types： pure sandstone, low shale content sandstone, and high shale content or argillaceous sandstone. Bayes discrimination models are established respectively and divided into 4 classes according to oil production. The matching rate of discrimination is 80.1%, which means the effect is favorable.

关 键 词：产能预测 物性分类 判别分析 低渗透储层 

分 类 号：P631[天文地球—地质矿产勘探]