基于交叉验证支持向量机储层预测方法及应用  

Reservoir Prediction Method and Its Application of Support Vector Machine Based on Cross Validation

作　　者：张军华[1] 任雄风 赵杰[1] 谭明友[2] 于正军[2] ZHANG Jun-hua;REN Xiong-feng;ZHAO Jie;TAN Ming-you;YU Zheng-jun(School of Geosicences,China University of Petroleum(East China),Qingdao 266580,China;Shengli Oilfield Geophysical Research Institute,Dongying 257022,China)

机构地区：[1]中国石油大学(华东)地球科学与技术学院,青岛266580 [2]胜利油田物探研究院,东营257022

出　　处：《科学技术与工程》2020年第13期5052-5057,共6页Science Technology and Engineering

基　　金：国家科技重大专项(2017ZX05009-001);中石化先导项目(P18051-4)。

摘　　要：东营凹陷深部储层埋深大,构造及相带变化复杂,钻遇目标层的井少,储层预测有很大的困难。以东营凹陷东部孔一段为例,将适合较小样本预测的支持向量机方法(support vector machine,SVM)应用到储层预测中。为了提高预测精度,惩罚因子选取和核函数参数训练过程中引入了交叉验证。输入样本为井点处的地震属性和储层厚度,属性通过井震关系优选,选取的是带宽、能量半时、最大振幅、均方根振幅、过零点个数和弧长等6种属性。预测结果表明,本文方法较常规的多元线性回归、不加交叉验证的SVM方法,有更高的预测精度,在深层勘探中有推广价值。The deep reservoirs in the Dongying sag have large buried depths,complex changes in structure and facies,and few wells meeting the target layer,thus,reservoir prediction is very difficult.Taking the eastern section of the Dongying sag as an example,the support vector machine(SVM)method suitable for small sample prediction to reservoir prediction was applied.To improve the prediction accuracy,cross validation was introduced into the penalty factor selection and kernel function parameter training.The input samples were the seismic attribute and reservoir thickness at the well point.The attributes were optimized by the well-seismic relationship.Six attributes of bandwidth,energy half-time,maximum amplitude,root mean square amplitude,zero-crossing point number,and arc length were selected.The prediction results show that the proposed method has higher prediction accuracy than those of conventional multiple linear regression and SVM without cross validation,and shall be promoted in deep layer exploration.

关 键 词：支持向量机 惩罚因子 核函数参数 地震属性 储层厚度 

分 类 号：P631[天文地球—地质矿产勘探]