Machine learning classification algorithm screening for the main controlling factors of heavy oil CO2 huff and puff

Q1 Earth and Planetary Sciences Petroleum Research Pub Date : 2024-04-13 DOI:10.1016/j.ptlrs.2024.04.002
Peng-xiang Diwu , Beichen Zhao , Hangxiangpan Wang , Chao Wen , Siwei Nie , Wenjing Wei , A-qiao Li , Jingjie Xu , Fengyuan Zhang
{"title":"Machine learning classification algorithm screening for the main controlling factors of heavy oil CO2 huff and puff","authors":"Peng-xiang Diwu ,&nbsp;Beichen Zhao ,&nbsp;Hangxiangpan Wang ,&nbsp;Chao Wen ,&nbsp;Siwei Nie ,&nbsp;Wenjing Wei ,&nbsp;A-qiao Li ,&nbsp;Jingjie Xu ,&nbsp;Fengyuan Zhang","doi":"10.1016/j.ptlrs.2024.04.002","DOIUrl":null,"url":null,"abstract":"<div><div>CO<sub>2</sub> huff and puff technology can enhance the recovery of heavy oil in high-water-cut stages. However, the effectiveness of this method varies significantly under different geological and fluid conditions, which leads to a high-dimensional and small-sample (HDSS) dataset. It is difficult for conventional techniques that identify key factors that influence CO<sub>2</sub> huff and puff effects, such as fuzzy mathematics, to manage HDSS datasets, which often contain nonlinear and irremovable abnormal data. To accurately pinpoint the primary control factors for heavy oil CO<sub>2</sub> huff and puff, four machine learning classification algorithms were adopted. These algorithms were selected to align with the characteristics of HDSS datasets, taking into account algorithmic principles and an analysis of key control factors. The results demonstrated that logistic regression encounters difficulties when dealing with nonlinear data, whereas the extreme gradient boosting and gradient boosting decision tree algorithms exhibit greater sensitivity to abnormal data. By contrast, the random forest algorithm proved to be insensitive to outliers and provided a reliable ranking of factors that influence CO<sub>2</sub> huff and puff effects. The top five control factors identified were the distance between parallel wells, cumulative gas injection volume, liquid production rate of parallel wells, huff and puff timing, and heterogeneous Lorentz coefficient. These research findings not only contribute to the precise implementation of heavy oil CO<sub>2</sub> huff and puff but also offer valuable insights into selecting classification algorithms for typical HDSS data.</div></div>","PeriodicalId":19756,"journal":{"name":"Petroleum Research","volume":"9 4","pages":"Pages 541-552"},"PeriodicalIF":0.0000,"publicationDate":"2024-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Petroleum Research","FirstCategoryId":"1087","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2096249524000371","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"Earth and Planetary Sciences","Score":null,"Total":0}
引用次数: 0

Abstract

CO2 huff and puff technology can enhance the recovery of heavy oil in high-water-cut stages. However, the effectiveness of this method varies significantly under different geological and fluid conditions, which leads to a high-dimensional and small-sample (HDSS) dataset. It is difficult for conventional techniques that identify key factors that influence CO2 huff and puff effects, such as fuzzy mathematics, to manage HDSS datasets, which often contain nonlinear and irremovable abnormal data. To accurately pinpoint the primary control factors for heavy oil CO2 huff and puff, four machine learning classification algorithms were adopted. These algorithms were selected to align with the characteristics of HDSS datasets, taking into account algorithmic principles and an analysis of key control factors. The results demonstrated that logistic regression encounters difficulties when dealing with nonlinear data, whereas the extreme gradient boosting and gradient boosting decision tree algorithms exhibit greater sensitivity to abnormal data. By contrast, the random forest algorithm proved to be insensitive to outliers and provided a reliable ranking of factors that influence CO2 huff and puff effects. The top five control factors identified were the distance between parallel wells, cumulative gas injection volume, liquid production rate of parallel wells, huff and puff timing, and heterogeneous Lorentz coefficient. These research findings not only contribute to the precise implementation of heavy oil CO2 huff and puff but also offer valuable insights into selecting classification algorithms for typical HDSS data.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
机器学习分类算法筛选重油二氧化碳膨化的主要控制因素
二氧化碳膨化技术可以提高重油在高水位切割阶段的采收率。然而,在不同的地质和流体条件下,这种方法的效果差异很大,这就导致了高维小样本(HDSS)数据集。传统技术(如模糊数学)难以识别影响二氧化碳喘振效应的关键因素,也难以管理 HDSS 数据集,因为 HDSS 数据集通常包含非线性和不可移动的异常数据。为了准确定位重油二氧化碳喘振的主要控制因素,我们采用了四种机器学习分类算法。这些算法的选择符合 HDSS 数据集的特点,同时考虑了算法原理和关键控制因素分析。结果表明,逻辑回归在处理非线性数据时遇到困难,而极梯度提升和梯度提升决策树算法对异常数据表现出更高的敏感性。相比之下,随机森林算法被证明对异常值不敏感,并提供了可靠的二氧化碳膨化效应影响因素排名。排在前五位的控制因素分别是平行井之间的距离、累计注气量、平行井的产液率、喘振和膨化时间以及异质洛伦兹系数。这些研究成果不仅有助于精确实施重油二氧化碳吹浮法,还为选择典型 HDSS 数据的分类算法提供了宝贵的见解。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Petroleum Research
Petroleum Research Earth and Planetary Sciences-Geology
CiteScore
7.10
自引率
0.00%
发文量
90
审稿时长
35 weeks
期刊最新文献
Applicability of deep neural networks for lithofacies classification from conventional well logs: An integrated approach Investigation of a solid particle deposition velocity in drag reducing fluids with salinity Use of graphs to assess well safety in drilling projects and during operations by identification of available barrier elements and consolidation of barrier envelopes Sedimentary microfacies of Member 5 of Xujiahe Formation in the Dongfengchang area, Sichuan Basin Research on physical explosion crater model of high-pressure natural gas pipeline
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1