基于近似支持向量机的流体识别因子融合  

Fluid identification factors fusion based on proximal support vector machine

作　　者：李文秀 文晓涛[1,2] 李天 刘松鸣 李雷豪 杨吉鑫 LI Wen-xiu;WEN Xiao-tao;LI Tian;LIU Song-ming;LI Lei-hao;YANG Ji-xin(Geophysical Institute,Chengdu University of Technology,Chengdu 610059,China;State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation,Chengdu University of Technology,Chengdu 610059,China;Yunnan Construction Investment First Investigation and Design Company Limited,Kunming 650031,China)

机构地区：[1]成都理工大学地球物理学院,成都610059 [2]成都理工大学油气藏地质及开发工程国家重点实验室,成都610059 [3]云南建投第一勘察设计有限公司,昆明650031

出　　处：《地球物理学进展》2020年第1期139-144,共6页Progress in Geophysics

基　　金：国家自然科学基金项目“深层碳酸盐岩储层流体地震预测理论与方法”(U1562111)和“基于频变信息的流体识别及流体可动性预测”(41774142)联合资助.

摘　　要：通过叠前反演获得的单参数或组合参数都有一定的流体识别能力,但如何将多种流体识别因子有效融合是目前进行流体识别的一个难题.利用人工参与进行流体性质的综合解释是目前流体识别因子融合的主要途径,但这种方法人为干扰较大,不确定性强.鉴于此,本文提出了一种基于近似支持向量机的流体识别方法.该方法首先以实际工区测井资料为依据,优选出对工区内储层所含流体特征敏感的流体识别因子作为输入参数,然后通过近似支持向量机进行流体性质的判别,实例证明该方法的识别结果客观准确,是一种可靠的流体识别方法.The single parameter or combination parameters obtained by pre-stack inversion have fluid recognition ability,but how to effectively fusion various fluid identification factors is a difficult problem in fluid identification.The comprehensive interpretation of fluid properties by means of artificial participation is the main approach to the fusion of fluid identification factors,however,this method has large human disturbance and strong uncertainty.In view of this,this paper proposes a fluid identification method based on proximal support vector machine.Firstly,the fluid identification factors which are sensitive to the fluid characteristics in the reservoir are selected as the input parameters based on the logging data of the actual oil and gas field.Then,the fluid properties are identified by proximal support vector machine.It is proved that the method is objective and accurate,and it is a reliable method for fluid recognition.

关 键 词：近似支持向量机 流体识别 地震属性 储层分析 

分 类 号：P631[天文地球—地质矿产勘探]