结合地质导向作业的实时智能地层识别预测模型
作者:
作者单位:

1.成都理工大学环境与土木工程学院,四川 成都 610059;2.地质灾害防治与地质环境保护国家重点实验室(成都理工大学),四川 成都 610059

作者简介:

通讯作者:

中图分类号:

P634

基金项目:

川庆钻探钻采工程技术研究院项目“地质导向辅助决策技术研究”(编号:CQCDLG-2020-06)


Integration of real-time intelligent stratigraphic identification and prediction model for geologically-guided field operation
Author:
Affiliation:

1.College of Environment and Civil Engineering, Chengdu University of Technology,Chengdu Sichuan 610059, China;2.State Key Laboratory of Geohazard Prevention and Geoenvironment Protection;(Chengdu University of Technology), Chengdu Sichuan 610059, China

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    地质导向是当前钻探智能化的核心组成部分,但当前地质导向中关键的地层识别还缺乏算法模型支撑。本文采用深度学习算法,基于四川威远区块实际采集的井身轨迹参数数据、测井数据,建立了能够实时预测钻遇地层的算法体系。该体系首先通过时序预测算法完成对后续井身轨迹的预测,其次基于该预测值实现对地层伽马值的预测,最后根据地层伽马值完成对钻遇地层的类型识别。为提高整个算法体系预测准确性,各环节采用多种算法进行对比,结果证明采用循环神经网络训练的实时预测,随机森林预测伽马值回归,支持向量机预测地层分类的训练效果最好,预测和识别精度分别达到了0.81、0.83、81.8%。地质导向过程中对实时钻遇地层的预判,为现场提供数据支撑以及辅助判断的手段,帮助实际工程更可靠有效地实现地质目标。

    Abstract:

    This paper employs deep learning algorithms based on actual wellbore trajectory parameter data and well logging data collected from the Weiyuan block in Sichuan. An algorithmic framework capable of real-time predicting encountered strata during drilling is established. Firstly, this framework utilizes a time-series prediction algorithm to forecast the subsequent wellbore trajectory. Subsequently, based on this prediction, it achieves the forecast of strata gamma values. Finally, the identification of encountered strata types is accomplished by utilizing the predicted strata gamma values. To enhance the predictive accuracy of the entire algorithmic framework, various algorithms are employed and compared at each stage. The results demonstrate that real-time prediction using recurrent neural networks, gamma value regression prediction through random forests, and strata classification prediction using support vector machines yield the best training outcomes. The predictive accuracies reach 0.81, 0.83 and 81.8%, respectively. The real-time anticipation of encountered strata during the geologically guided process provides on-site data support and auxiliary judgment means. This contributes to the more reliable and effective achievement of geological objectives in practical engineering.

    参考文献
    相似文献
    引证文献
引用本文

陈嘉豪,李谦,曾小龙,等.结合地质导向作业的实时智能地层识别预测模型[J].钻探工程,2023,50(S1):135-142.
chenjiahao, LiQian, ZengXiaolong, et al. Integration of real-time intelligent stratigraphic identification and prediction model for geologically-guided field operation[J]. Drilling Engineering, 2023,50(S1):135-142.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-05-30
  • 最后修改日期:2023-08-09
  • 录用日期:2023-08-10
  • 在线发布日期: 2023-10-21
  • 出版日期: