支持向量机在致密砂岩储层孔隙度预测中应用  

Application of SVM to Porosity Prediction in Tight Sandstone Reservoir

作　　者：张静[1] 王志章[2] 赵永军[1] 王飞腾 

机构地区：[1]中国石油大学(华东)地球科学与技术学院,山东青岛266580 [2]中国石油大学(北京)地球科学学院,北京102249

出　　处：《河南科学》2015年第6期993-997,共5页Henan Science

基　　金："十二五"国家油气重大专项基金资助(2008ZX05045)

摘　　要：大牛地山2段气层组属于典型的低孔低渗致密砂岩储层,其孔隙结构复杂,非均质性强,采用传统数理统计方法解释孔隙度误差大.根据已有测井、录井和岩心资料,在进行储层特征和四性关系研究的基础上,采用能够深度挖掘测井参数和孔隙度间非线性关系的支持向量机方法,对该研究区储层孔隙度进行预测.实例证明该方法预测孔隙度和岩心分析孔隙度符合率很高,与线性多元回归相比,能有效提高致密砂岩储层测井解释模型的精度,为储层的综合评价提供了可靠的地质参数.Daniudi gas field Shan-2 belongs to typical low porosity and permeability tight sandstone reservoir. Due to the complex pore structure and strong reservoir heterogeneity,it is hard to calculate the porosity exactly by the traditional method. Based on the existing logging,well log and core data,the reservoir characteristics and“four characters of reservoir”is studied. SVM is used to find the nonlinear relationship between Logging parameters and porosity. Examples prove that the coincidence rate of prediction porosity and core analysis porosity is high by using this method. Compared with the linear regression,SVM can effectively improve the accuracy of well logging interpretation model. Reliable geological parameters are provided for the comprehensive evaluation of reservoir.

关 键 词：致密砂岩 测井参数 孔隙度预测 支持向量机 

分 类 号：TE122[石油与天然气工程—油气勘探]