BP神经网络算法在页岩气饱和度评价的应用  

Application of BP Neural Network Algorithm in Shale Gas Saturation Evaluation

作　　者：王志豪 周明顺[2] 魏新路 刘迪仁[1] 孙康 申威 赵桠松 Wang Zhihao;Zhou Mingshun;Wei Xinlu;Liu Diren;Sun Kang;Shen Wei;Zhao Yasong(College of Geophysics and Petroleum Resources,Yangtze University,Wuhan Hubei 430100,China;Exploration and Development Research Institute,PetroChina Huabei Oilfield Company,Renqiu Hebei 062500,China;Liaohe Branch of China Petroleum Logging Co.,Ltd,Panjin Liaoning 124000,China)

机构地区：[1]长江大学地球物理与石油资源学院,湖北武汉430100 [2]中石油华北油田分公司勘探开发研究院,河北任丘062500 [3]中国石油集团测井有限公司辽河分公司,辽宁盘锦124000

出　　处：《工程地球物理学报》2022年第2期216-222,共7页Chinese Journal of Engineering Geophysics

基　　金：国家重点研发计划项目子课题(编号:2018YFC060330502)。

摘　　要：页岩气储层中高含量的有机质和黏土矿物会对饱和度的计算会产生影响,常规油气储层建立的饱和度计算模型在页岩气储层饱和度评价中存在较大误差,亟需寻求新的途径与思路。本文分析总结了页岩气储层性质及其与含水饱和度的关系机理。利用井中地层密度和铀含量与岩心分析含水饱和度有较好的相关性,通过BP神经网络训练建立新的非电性测井饱和度评价模型。在以往饱和度计算模型基础上,考虑总有机碳含量对页岩储层含水饱和度的影响,建立新的含水饱和度计算公式。研究结果表明,此模型通过铀含量和地层密度计算得出的含水饱和度与岩心分析数据具有更高的吻合度。本次研究基于岩心分析数据,建立基于BP神经网络的含水饱和度计算模型,为页岩气含水饱和度的评价提供了新的方法与思路。The high content of organic matter and clay minerals in shale gas reservoir will affect the calculation of saturation.And the saturation calculation model established by conventional oil and gas reservoir has obvious errors in the evaluation of shale gas reservoir saturation,so it is urgent to seek new ways and ideas.This paper analyzes and summarizes the relationship between shale gas reservoir properties and water saturation mechanism.The formation density and uranium content in wells are well correlated with the core analysis of water saturation,and a new non-electrical logging saturation evaluation model is established through BP neural network training.On the basis of previous saturation calculation models,a new calculation formula of water saturation is established considering the influence of total organic carbon content on shale reservoir water saturation.The results show that the water saturation calculated by the model based on uranium content and formation density is in better agreement with the core analysis data.Based on the core analysis data,this study establishes a water saturation calculation model based on BP neural network,providing a new method and idea for shale gas water saturation evaluation.

关 键 词：页岩气 含水饱和度 地层密度 铀含量 BP神经网络 总有机碳含量 

分 类 号：P631.4[天文地球—地质矿产勘探]