«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《地球科学与环境学报》[ISSN:1672-6561/CN:61-1423/P]

卷:

第46卷

期数:

2024年第01期

页码:

127-142

栏目:

沉积地质与油气勘探

出版日期:

2024-01-15

文章信息/Info

Title:

Lithology Identification in Optuna-BiGRU Logging Based on Enhanced Multi-head Attention Mechanism

文章编号:

1672-6561(2024)01-0127-16

作者:

王婷婷¹; 2; 王振豪¹; 李 方¹; 赵万春³; 4*

(1. 东北石油大学 电气信息工程学院,黑龙江 大庆 163318; 2. 东北石油大学 黑龙江省网络化与智能控制重点实验室,黑龙江 大庆 163318; 3. 东北石油大学 非常规油气研究院,黑龙江 大庆 163318; 4. 东北石油大学 陆相页岩油气成藏及高效开发教育部重点实验室,黑龙江 大庆 163318)

Author(s):

WANG Ting-ting¹; 2;  WANG Zhen-hao¹;  LI Fang¹;  ZHAO Wan-chun³; 4*

(1. School of Electrical & Information Engineering, Northeast Petroleum University, Daqing 163318, Heilongjiang, China; 2. Heilongjiang Provincial Key Laboratory of Network and Intelligent Control, Northeast Petroleum University, Daqing 163318, Heilongjiang, China; 3. Institute of Unconventional Oil & Gas, Northeast Petroleum University, Daqing 163318, Heilongjiang, China; 4. Key Laboratory of Continental Shale Hydrocarbon Accumulation and Efficient Development of Ministry of Education, Northeast Petroleum University, Daqing 163318, Heilongjiang, China)

关键词:

岩性识别;  深度学习;  BiGRU;  增强多头注意力机制;  小波包自适应阈值;  超参数优化

Keywords:

lithology identification;  deep learning;  BiGRU;  enhancing multi-head attention mechanism;  wavelet packet adaptive threshold;  hyperparameter optimization

分类号:

P618.13; P631.8⁺1

DOI:

10.19814/j.jese.2023.07011

文献标志码:

A

摘要:

测井岩性识别是油气勘探开发中至关重要的内容。针对现有算法模型在处理测井曲线数据时,无法有效捕获曲线内部深层关联和深度方向关系、拟合能力较弱、难以准确提取关键特征、噪声干扰以及模型超参数调优过程复杂困难等问题,提出了一种通过Optuna超参数优化双向门循环单元(Optuna-BiGRU)结合增强多头注意力机制(EMHA)的测井岩性识别模型——Optuna-BiGRU-EMHA模型。该模型引入残差机制和层归一化以改进多头注意力机制模块,并结合双向门循环单元(BiGRU)解决了处理测井数据时的问题,同时使用Optuna超参数优化框架和小波包自适应阈值方法分别解决了超参数调优和噪声干扰问题。首先通过交会图分析和敏感性箱线图分析选取自然伽马、深感应电阻率、中子-密度孔隙度、平均中子-密度孔隙度和岩性密度5个特征参数的测井数据,通过小波包自适应阈值方法对数据进行去噪,并将测井数据分割成数据块,然后利用Optuna框架优化BiGRU-EMHA模型超参数,最后通过实验对比K-近邻算法(KNN)、随机森林(RF)、极端梯度提升算法(XGBoost)、长短期记忆(LSTM)神经网络、BiGRU、双向长短期记忆(BiLSTM)神经网络、BiGRU-MHA、Optuna-BiGRU-EMHA等8种模型在测井岩性识别中的精度。结果表明:Optuna-BiGRU-EMHA模型识别准确率达到80%,相对于传统机器学习模型和深度学习模型,综合岩性识别准确率分别提高15.94%～23.14%和3.93%～15.94%,该模型为常规测井岩性识别提供了坚实的理论支持。

Abstract:

Lithology identification in well logging plays a crucial role in oil and gas exploration and development. However, existing algorithmic models face challenges in effectively capturing deep-seated correlations and depth-related relationships within log curve data. They also exhibit weak fitting capabilities, struggle with accurate feature extraction, contend with noise interference, and face complexities in fine-tuning model hyperparameters. To address these issues, a model for lithology identification that combines Optuna hyperparameter optimization with a bidirectional gate cycle unit(BiGRU)and enhanced multi-head attention(EMHA)was proposed. The model is called Optuna-BiGRU-EMHA. The approach introduces residual mechanisms and layer normalization to enhance the multi-head attention module, effectively resolving issues encountered when processing well-logging data. Additionally, it leverages the Optuna hyperparameter optimization framework and wavelet packet adaptive threshold method to address hyperparameter fine-tuning and noise interference. The study begins by selecting well-logging data of five characteristic parameters, including natural gamma, deep induction resistivity, neutron-density porosity, average neutron-density porosity and lithology density, through cross-plot and sensitivity analyses. The data is then denoised using the wavelet packet adaptive threshold method and segmented into data blocks. Subsequently, the Optuna framework is employed to optimize the hyperparameters of the BiGRU-EMHA model. Finally, experimental comparisons are conducted to assess the accuracy of eight models in lithology identification based on well logging data. The eight models are KNN, RF, XGBoost, LSTM, BiGRU, BiLSTM, BiGRU-MHA, and Optuna-BiGRU-EMHA. The results show that the Optuna-BiGRU-EMHA model achieves an identification accuracy of 80%. Compared with traditional machine learning and deep learning models, there is a substantial improvement in comprehensive lithology identification accuracy by 15.94% to 23.14%, and 3.93% to 15.94%. The Optuna-BiGRU-EMHA model provides a robust theoretical foundation for conventional well-logging lithology identification in the domain of oil and gas exploration and development.

参考文献/References:

[1] SHEN C B,ASANTE-OKYERE S,YEVENYO ZIGGAH Y,et al.Group Method of Data Handling(GMDH)Lithology Identification Based on Wavelet Analy-sis and Dimensionality Reduction as Well Log Data Pre-processing Techniques[J].Energies,2019,12(8):1509.
[2] 张 涛,李艳萍,刘晓宇,等.基于自适应粒子群优化最小二乘支持向量机的深层变质岩测井岩性识别[J].地球物理学进展,2023,38(1):382-392.
ZHANG Tao,LI Yan-ping,LIU Xiao-yu,et al.Lithology Interpretation of Deep Metamorphic Rocks with Well Logging Based on APSO-LSSVM Algori-thm[J].Progress in Geophysics,2023,38(1):382-392.
[3] 马陇飞,萧汉敏,陶敬伟,等.基于梯度提升决策树算法的岩性智能分类方法[J].油气地质与采收率,2022,29(1):21-29.
MA Long-fei,XIAO Han-min,TAO Jing-wei,et al.Intelligent Lithology Classification Method Based on GBDT Algorithm[J].Petroleum Geology and Reco-very Efficiency,2022,29(1):21-29.
[4] 刘 昊,朱丹丹,陈 冬,等.基于聚类算法的岩性预分类方法研究[C]∥西安石油大学.2018 IPPTC国际石油石化技术会议论文集.西安:西安石油大学,2018:387-396.
LIU Hao,ZHU Dan-dan,CHEN Dong,et al.Resear-ch on Lithology Pre-classification Method Based on Clustering Algorithm[C]∥Xi'an Shiyou University.Proceedings of 2018 IPPTC International Petroleum and Petrochemical Technology Conference.Xi'an:Xi'an Shiyou University,2018:387-396.
[5] ZHOU L,ZHONG F Y,YAN J C,et al.Prestack Inversion Identification of Organic Reef Gas Reservoirs of Permian Changxing Formation in Damaoping Area,Sichuan Basin,SW China[J].Petroleum Exploration and Development,2020,47(1):89-100.
[6] SALIM A M A,PAN H P,LUO M,et al.Integrated Log Interpretation in the Chinese Continental Scientific Drilling Main Hole(Eastern China):Lithology and Mineralization[J].Journal of Applied Sciences,2008,8(20):3593-3602.
[7] SAGGAF M M,NEBRIJA E L.Estimation of Litho-logies and Depositional Facies from Wire-line Logs[J].AAPG Bulletin,2000,84(10):1633-1646.
[8] 陈玉林,李戈理,杨智新,等.基于KNN算法识别合水地区长7储层岩性岩相[J].测井技术,2020,44(2):182-185.
CHEN Yu-lin,LI Ge-li,YANG Zhi-xin,et al.Identification of Lithology and Lithofacies of Chang-7 Reservoir in Heshui Area by KNN Algorithm[J].Well Logging Technology,2020,44(2):182-185.
[9] 徐 晗,姚孔轩,程丹仪,等.基于非开挖随钻检测系统与随机森林的地层岩性识别[J].地质科技通报,2021,40(5):272-280.
XU Han,YAO Kong-xuan,CHENG Dan-yi,et al.Stratigraphic Lithology Identification Based on No-dig Logging While Drilling System and Random Forest[J].Bulletin of Geological Science and Technology,2021,40(5):272-280.
[10] 冯 瑞,杨丽萍,侯成磊,等.基于随机森林的陕西省西安市近地表气温估算[J].地球科学与环境学报,2022,44(1):102-113.
FENG Rui,YANG Li-ping,HOU Cheng-lei,et al.Estimation of Near-surface Air Temperature in Xi'an City of Shaanxi Province,China Based on Random Forest[J].Journal of Earth Sciences and Environment,2022,44(1):102-113.
[11] 孙予舒,黄 芸,梁 婷,等.基于XGBoost算法的复杂碳酸盐岩岩性测井识别[J].岩性油气藏,2020,32(4):98-106.
SUN Yu-shu,HUANG Yun,LIANG Ting,et al.Identification of Complex Carbonate Lithology by Logging Based on XGBoost Algorithm[J].Lithologic Reservoirs,2020,32(4):98-106.
[12] 罗仁泽,庹娟娟,倪华玲,等.基于改进集成学习的测井岩性识别方法研究[J].石油物探,2023,62(2):212-224.
LUO Ren-ze,TUO Juan-juan,NI Hua-ling,et al.Logging Lithology Identification Method Based on Improved Ensemble Learning[J].Geophysical Pro-specting for Petroleum,2023,62(2):212-224.
[13] 陈玉敏,魏 阳,常政威,等.基于遥感数据和XGBoost算法的31个城市NO₂、CO₂浓度比率变化特征[J].地球科学与环境学报,2023,45(6):1355-1367.
CHEN Yu-min,WEI Yang,CHANG Zheng-wei,et al.Variation Characteristics of Concentration Ratio of Nitrogen Dioxide and Carbon Dioxide in 31 Cities,China Based on Remote Sensing Data and XGBoost Algorithm[J].Journal of Earth Sciences and Environment,2023,45(6):1355-1367.
[14] 杨 笑,王志章,周子勇,等.基于参数优化AdaBoost算法的酸性火山岩岩性分类[J].石油学报,2019,40(4):457-467.
YANG Xiao,WANG Zhi-zhang,ZHOU Zi-yong,et al.Lithology Classification of Acidic Volcanic Rocks Ba-sed on Parameter-optimized AdaBoost Algorithm[J].Acta Petrolei Sinica,2019,40(4):457-467.
[15] AMIRGALIEV E,ISABAEV Z,ISKAKOV S,et al.Recognition of Rocks at Uranium Deposits by Using a Few Methods of Machine Learning[C]∥RHEE S Y,PARK J Y,INOUE A.Soft Computing in Machine Learning.Daejeon:PaiChai University,2014:33-40.
[16] YANG H J,PAN H P,MA H L,et al.Performance of the Synergetic Wavelet Transform and Modified K-means Clustering in Lithology Classification Using Nuclear Log[J].Journal of Petroleum Science and Engineering,2016,144:1-9.
[17] JOSHI D,PATIDAR A K,MISHRA A,et al.Prediction of Sonic Log and Correlation of Lithology by Comparing Geophysical Well Log Data Using Machine Learning Principles[J].GeoJournal,2021,DOI:10.1007/s10708-021-10502-6.
[18] DENG C X,PAN H P,FANG S N,et al.Support Vector Machine as an Alternative Method for Lithology Classification of Crystalline Rocks[J].Journal of Geophysics and Engineering,2017,14(2):341-349.
[19] 张树义,王 波,马尽文.基于深度卷积自编码器的岩性分类与识别[J].信号处理,2023,39(1):11-19.
ZHANG Shu-yi,WANG Bo,MA Jin-wen.Deep Convolutional Auto-encoder Based on Lithologic Classification and Recognition[J].Journal of Signal Processing,2023,39(1):11-19.
[20] HE M,GU H M,WAN H.Log Interpretation for Lithology and Fluid Identification Using Deep Neural Network Combined with MAHAKIL in a Tight Sandstone Reservoir[J].Journal of Petroleum Science and Engineering,2020,194:107498.
[21] IMAMVERDIYEV Y,SUKHOSTAT L.Lithological Facies Classification Using Deep Convolutional Neural Network[J].Journal of Petroleum Science and Engineering,2019,174:216-228.
[22] 蔡浩杰,韩海辉,张雨莲,等.基于地形特征融合的卷积神经网络滑坡识别[J].地球科学与环境学报,2022,44(3):568-579.
CAI Hao-jie,HAN Hai-hui,ZHANG Yu-lian,et al.Convolutional Neural Network Landslide Recognition Based on Terrain Feature Fusion[J].Journal of Earth Sciences and Environment,2022,44(3):568-579.
[23] LIN J,LI H,LIU N H,et al.Automatic Lithology Identification by Applying LSTM to Logging Data:A Case Study in X Tight Rock Reservoirs[J].IEEE Geoscience and Remote Sensing Letters,2021,18(8):1361-1365.
[24] 熊玄辰,曹俊兴,周 鹏,等.基于双向长短期记忆神经网络的岩相预测方法[J].成都理工大学学报(自然科学版),2021,48(2):226-234.
XIONG Xuan-chen,CAO Jun-xing,ZHOU Peng,et al.Lithofacies Prediction Method Based on Bidirectional Long Short Memory Neural Network[J].Journal of Chengdu University of Technology(Science & Technology Edition),2021,48(2):226-234.
[25] 陈钢花,张寓侠,王 军,等.双向长短时记忆神经网络在滩坝砂储层岩性识别中的应用[J].测井技术,2023,47(3):319-325.
CHEN Gang-hua,ZHANG Yu-xia,WANG Jun,et al.Application of BiLSTM in Lithology Identification of Beach-bar Sand Reservoir[J].Well Logging Techno-logy,2023,47(3):319-325.
[26] 王庆凯.基于长短期记忆网络和时空序列模型的岩性识别方法研究[D].秦皇岛:燕山大学,2022.
WANG Qing-kai.Research on Lithology Identification Method Based on LSTM and Spatio-temporal Sequence Model[D].Qinhuangdao:Yanshan University,2022.
[27] 罗 群,吴安彬,王井伶,等.中国北方页岩气成因类型、成气模式与勘探方向[J].岩性油气藏,2019,31(1):1-11.
LUO Qun,WU An-bin,WANG Jing-ling,et al.Gene-tic Types,Generation Models,and Exploration Direction of Shale Gas in Northern China[J].Lithologic Reservoirs,2019,31(1):1-11.
[28] SINGH H,SEOL Y,MYSHAKIN E M.Automated Well-log Processing and Lithology Classification by Identifying Optimal Features Through Unsupervised and Supervised Machine-learning Algorithms[J].SPE Journal,2020,25(5):2778-2800.
[29] KUSUMAPUTRI F H,ARIFIN A S.Anomaly Detection Based on NSL-KDD Using XGBoost with Optuna Tuning[C]∥ICBIR.2022 7th International Confe-rence on Business and Industrial Research(ICBIR).Bangkok:ICBIR,2022:586-591.
[30] AKIBA T,SANO S,YANASE T,et al.Optuna:A Next-generation Hyperparameter Optimization Fra-mework[C]∥TEREDESAI A,KUMAR V.KDD'19: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.New York:Association for Computing Machinery,2019:2623-2631.
[31] YU W N,KIM I Y,MECHEFSKE C.Analysis of Dif-ferent RNN Autoencoder Variants for Time Series Classification and Machine Prognostics[J].Mechanical Systems and Signal Processing,2021,149:107322.
[32] ZHANG J S,JIANG Y C,WU S M,et al.Prediction of Remaining Useful Life Based on Bidirectional Gated Recurrent Unit with Temporal Self-attention Me-chanism[J].Reliability Engineering & System Safety,2022,221:108297.
[33] 孙英健.深度学习时序分析在邻接盲井数据集石油储层识别中的应用[D].秦皇岛:燕山大学,2022.
SUN Ying-jian.Deep Learning Time-series Analysis for Petroleum Reservoir Identification in Adjacent Blind Well Datasets[D].Qinhuangdao:Yanshan University,2022.
[34] 陈云天.基于机器学习的测井曲线补全与生成研究[D].北京:北京大学,2020.
CHEN Yun-tian.Research on Well Log Completion and Generation Based on Machine Learning[D].Beijing:Peking University,2020.

相似文献/References:

[1]蔡浩杰,韩海辉,张雨莲,等.基于地形特征融合的卷积神经网络滑坡识别[J].地球科学与环境学报,2022,44(03):568.[doi:10.19814/j.jese.2021.12016]
　CAI Hao-jie,HAN Hai-hui,ZHANG Yu-lian,et al.Convolutional Neural Network Landslide Recognition Based on Terrain Feature Fusion[J].Journal of Earth Sciences and Environment,2022,44(01):568.[doi:10.19814/j.jese.2021.12016]
[2]方 巍,齐媚涵.基于深度学习的高时空分辨率降水临近预报方法[J].地球科学与环境学报,2023,45(03):706.[doi:10.19814/j.jese.2023.01010]
　FANG Wei,QI Mei-han.Precipitation Nowcasting Method with High Spatio-temporal Resolution Based on Deep Learning[J].Journal of Earth Sciences and Environment,2023,45(01):706.[doi:10.19814/j.jese.2023.01010]
[3]辛鲁斌,韩 玲*,李良志.基于多源数据融合的滑坡智能识别[J].地球科学与环境学报,2023,45(04):920.
　XIN Lu-bin,HAN Ling*,LI Liang-zhi.Landslide Intelligent Recognition Based on Multi-source Data Fusion[J].Journal of Earth Sciences and Environment,2023,45(01):920.
[4]方巍*,张霄智,齐媚涵.MEPM模型:基于深度学习的多变量厄尔尼诺-南方涛动预测模型[J].地球科学与环境学报,2024,46(03):285.[doi:10.19814/j.jese.2023.08029]
　FANG Wei*,ZHANG Xiao-zhi,QI Mei-han.MEPM: MultivariateENSOPredictionModel Based on Deep Learning[J].Journal of Earth Sciences and Environment,2024,46(01):285.[doi:10.19814/j.jese.2023.08029]

备注/Memo

备注/Memo:

收稿日期:2023-07-07; 修回日期:2023-09-18投稿网址:http:∥jese.chd.edu.cn/
基金项目:国家自然科学基金项目(52074088,52174022); 东北石油大学省杰青后备人才项目(SJQH202002); 黑龙江省博士后科研启动项目(LBH-Q21086); 黑龙江省揭榜挂帅科技攻关项目(DQYT-2022-JS-758)
作者简介:王婷婷(1982-),女,黑龙江大庆人,教授,博士研究生导师,工学博士,E-mail:wangtingting@nepu.edu.cn。
*通信作者:赵万春(1978-),男,黑龙江大庆人,教授,博士研究生导师,工学博士,E-mail:zhaowanchun@nepu.edu.cn。

常用功能

更新日期/Last Update: 2024-01-25