近似支持向量机的AVO类型判别  

AVO types discrimination based on a proximal support vector machine

作　　者：李文秀 文晓涛[1,2] 李天 李雷豪 刘松鸣 杨吉鑫 

机构地区：[1]成都理工大学地球物理学院,四川成都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,四川成都610059 [3]云南建投第一勘察设计有限公司,云南昆明650031

出　　处：《石油地球物理勘探》2018年第5期969-974,共6页Oil Geophysical Prospecting

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：AVO技术是储层含油气性分析的重要工具,可以定性地描述油气藏。常规储层的AVO分类主要依靠人为判别,致使判别结果不准且工作量大。本文从四类AVO曲线中提取特征参数作为训练集,把近似支持向量机方法引入AVO类型判别;再以四类含气砂岩AVO曲线形态为依据,把叠前地震资料的曲线形态特征作为输入参数,获得工区内储层的AVO类型。将该方法应用于南海碎屑岩气田的AVO类型自动识别,取得了较准确的结果,为储层的AVO类型判别提供了可靠、高效、便捷的工具。AVO is an important approach for reservoir oil and gas analysis.It can qualitatively describe oil reservoirs.The AVO conventional classification depends mainly on human discrimination so that the discrimination result is often inaccurate and the workload is heavy.In this paper,we extract feature parameters from four types of AVO curves as a training set,and introduce the proximal support vector machine method to AVO types discrimination.Based on the shape of four types of gas AVO curves,taking the morphological features of prestack seismic data as input parameters,AVO types of the reservoir in a survey area are obtained.This method is applied to the automatic identification of AVO types in a clastic-rock gas field in the South China Sea,and more accurate results are obtained.The proposed method provides a reliable and convenient tool for AVO types discrimination in reservoirs.

关 键 词：近似支持向量机 AVO类型 分类 储层分析 

分 类 号：P631[天文地球—地质矿产勘探]