基于2种机器学习方法的页岩TOC含量评价——以川南五峰组—龙马溪组为例

doi:10.12108/yxyqc.20220113

岩性油气藏 ›› 2022, Vol. 34 ›› Issue (1): 130–138.doi: 10.12108/yxyqc.20220113

基于2种机器学习方法的页岩TOC含量评价——以川南五峰组—龙马溪组为例

杨占伟^1,2, 姜振学^1,2, 梁志凯^1,2, 吴伟³, 王军霞^1,2, 宫厚健^1,2, 李维邦^1,2, 苏展飞^1,2, 郝绵柱^1,2

1. 中国石油大学(北京)油气资源与探测国家重点实验室, 北京 102249;
2. 中国石油大学(北京)非常规油气科学技术研究院, 北京 102249;
3. 中国石油西南油气田公司页岩气研究院, 成都 610051;
4. 中国石油大学(北京)理学院, 北京 102249

收稿日期:2021-06-17 修回日期:2021-08-02 出版日期:2022-01-01 发布日期:2022-01-21
第一作者:杨占伟(1996-),男,中国石油大学(北京)在读硕士研究生,研究方向为非常规油气地质评价及地质建模。地址:(102249)北京市昌平区中国石油大学(北京)。Email:1286325171@qq.com
通信作者: 姜振学（1963—），男，博士，教授，主要从事页岩气储层评价与成藏机理相关教学与研究工作。Email：jiangzx@cup.edu.cn。
基金资助:
海相重大专项“五峰—龙马溪组富有机质页岩储层精细描述与页岩气成藏机理”（编号：2017ZX05035-02）和“海相富气页岩低阻成因及其对含气性的控制机理”（编号：42072151）联合资助

Evaluation of shale TOC content based on two machine learning methods: A case study of Wufeng-Longmaxi Formation in southern Sichuan Basin

YANG Zhanwei^1,2, JIANG Zhenxue^1,2, LIANG Zhikai^1,2, WU Wei³, WANG Junxia^1,2, GONG Houjian^1,2, LI Weibang^1,2, SU Zhanfei^1,2, HAO Mianzhu^1,2

1. State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum(Beijing), Beijing 102249, China;
2. Research Institute of Unconventional Oil and Gas Science and Technology, China University of Petroleum(Beijing), Beijing 102249, China;
3. Shale Gas Research Institute, PetroChina Southwest Oil & Gas Field Company, Chengdu 610051, China;
4. College of Science, China University of Petroleum(Beijing), Beijing 102249, China

Received:2021-06-17 Revised:2021-08-02 Online:2022-01-01 Published:2022-01-21

摘要/Abstract

摘要： 为了建立合理准确的川南五峰组—龙马溪组页岩TOC含量预测方法，以长宁、泸州等地区的测井曲线及17口井实测TOC含量数据为基础，利用主成分分析法对这些资料进行预处理，基于BP神经网络和梯度提升决策树（GBDT）方法建立2种TOC含量预测模型，并将之与传统TOC含量预测方法进行对比。结果表明： ① 2种新模型的准确度均高于传统方法，预测结果与实际值吻合度均满足要求。②与BP神经网络模型相比，GBDT预测精度更高，均方根误差仅为0.0387。利用GBDT方法所建立的TOC含量预测模型具有低成本、高效、连续等特点，能够快速准确地预测目的层TOC含量。该成果可为提高页岩油气勘探开发效率提供有效技术支撑。

关键词: 主成分分析法, BP神经网络, GBDT, 五峰组—龙马溪组, 川南

Abstract: In order to establish a reasonable and accurate prediction method of shale total organic carbon(TOC) content of Wufeng-Longmaxi Formation in southern Sichuan Basin,the principal component analysis method was used to preprocess the logging curves and the measured TOC content data of 17 wells in Changning and Luzhou areas. Two TOC content prediction models were established based on BP neural network and gradient boosting decision tree(GBDT),and compared with the traditional TOC content prediction methods. The results show that:(1) The accuracy of the two models is higher than that of the traditional methods,and the consistence between the predicted results and the actual values can meet the requirements.(2) Compared with BP neural network model, GBDT has higher prediction accuracy,and the root mean square error is only 0.0387. The TOC content prediction model established by GBDT has the characteristics of low cost,high efficiency and continuity,and can be used to predict the TOC content of the target layer quickly and accurately. This achievement can provide effective technical support for improving the efficiency of shale oil and gas exploration and development.

Key words: principal component analysis, BP neural network, gradient boosting decision tree, Wufeng-Longmaxi Formation, southern Sichuan Basin

中图分类号:

TE122

杨占伟, 姜振学, 梁志凯, 吴伟, 王军霞, 宫厚健, 李维邦, 苏展飞, 郝绵柱. 基于2种机器学习方法的页岩TOC含量评价——以川南五峰组—龙马溪组为例[J]. 岩性油气藏, 2022, 34(1): 130–138.

YANG Zhanwei, JIANG Zhenxue, LIANG Zhikai, WU Wei, WANG Junxia, GONG Houjian, LI Weibang, SU Zhanfei, HAO Mianzhu. Evaluation of shale TOC content based on two machine learning methods: A case study of Wufeng-Longmaxi Formation in southern Sichuan Basin[J]. Lithologic Reservoirs, 2022, 34(1): 130–138.

参考文献

[1] PASSEY Q R, CREANEY S, KULLA J B, et al. A practical model for organic richness from porosity and resistivity logs. AAPG Bulletin, 1990, 74(12):1777-1794.
[2] 赵万金, 高海燕, 闫国亮, 等. 基于最优化估算和贝叶斯统计的TOC预测技术.岩性油气藏, 2020, 32(1):86-93. ZHAO W J, GAO H Y, YAN G L, et al. TOC prediction technology based on optimal estimation and Bayesian statistics. Lithologic Reservoirs, 2020, 32(1):86-93.
[3] 严伟, 刘帅, 冯明刚, 等. 四川盆地丁山区块页岩气储层关键参数测井评价方法. 岩性油气藏, 2019, 31(3):95-104. YAN W, LIU S, FENG M G, et al. Well logging evaluation methods of key parameters for shale gas reservoir in Dingshan block, Sichuan Basin. Lithologic Reservoirs, 2019, 31(3):95104.
[4] 杜江民, 张小莉, 钟高润, 等. 致密油烃源岩有机碳含量测井评价方法优选及应用:以鄂尔多斯盆地延长组长7段烃源岩为例. 地球物理学进展, 2016, 31(6):2526-2533. DU J M, ZHANG X L, ZHONG G R, et al. Analysis on the optimization and application of well logs identification methods for organic carbon content in source rocks of the tight oil:Illustrated by the example of the source rocks of Chang 7 member of Yanchang Formation in Ordos Basin. Progress in Geophysics, 2016, 31(6):2526-2533.
[5] SCHMOKER J W. Determination of organic-matter content of Appalachian Devonian Shales from gamma-ray logs. AAPG Bulletin, 1981, 65(7):1285-1298.
[6] HESTER T C, SCHMOKER J W. Determination of organic content from formation-density logs, Devonian-Mississippian Woodford Shale, Anadarko Basin, Oklahoma. US Geological Survey, 1987.
[7] 陈浩, 刘洛夫, 窦文超, 等. 银额盆地居延海坳陷烃源岩TOC测井预测方法优选及应用. 地球物理学进展, 2019, 34(3):1017-1024. CHEN H, LIU L F, DOU W C, et al. Optimal selection and application of prediction methods for TOC of source rocks based on logging data in Juyanhai Depression, Yingen-Ejin Banner Basin. Progress in Geophysics, 2019, 34(3):1017-1024.
[8] 朱振宇, 刘洪, 李幼铭. ΔlogR技术在烃源岩识别中的应用与分析.地球物理学进展, 2003, 18(4):647-649. ZHU Z Y, LIU H, LI Y M. The analysis and application of ΔlogR method in the source rock's identification. Progress in Geophysics, 2003, 18(4):647-649.
[9] 王祥, 马劲风, 张新涛, 等. 一种考虑密度因素的广义ΔlogR法预测总有机碳含量:以渤中凹陷西南部陆相深层烃源岩为例. 地球物理学进展, 2020, 35(4):1471-1480. WANG X, MA J F, ZHANG X T, et al. A generalized ΔlogR method for predicting total organic carbon content considering density factor:A case study of deep continental source rocks in southwest Bozhong Sag. Progress in Geophysics, 2020, 35(4):1471-1480.
[10] 郭泽清, 孙平, 刘卫红. 利用ΔlogR技术计算柴达木盆地三湖地区第四系有机碳. 地球物理学进展, 2012, 27(2):626-633. GUO Z Q, SUN P, LIU W H. Calculation of quaternary organic carbon in Sanhu area of Qaidam Basin using ΔlogR technology. Progress in Geophysics, 2012, 27(2):626-633.
[11] 刘超, 印长海, 卢双舫.变系数ΔlogR烃源岩测井评价技术关键参数厘定方法及应用. 天然气地球科学, 2015, 26(10):1925-1931. LIU C, YIN C H, LU S F. Predicting key parameters for variablecoefficient ΔlogR logging technique and its application in source rocks evaluation. Natural Gas Geoscience, 2015, 26(10):19251931.
[12] 叶绍泽, 曹俊兴, 吴施楷, 等. 基于深度置信网络的总有机碳含量预测方法. 地球物理学进展, 2018, 33(6):304-311. YE S Z, CAO J X, WU S K, et al. Prediction method of total organic carbon content based on deep belief nets. Progress in Geophysics, 2018, 33(6):304-311.
[13] 杨洋, 石万忠, 张晓明, 等. 页岩岩相的测井曲线识别方法:以焦石坝地区五峰组-龙马溪组为例.岩性油气藏, 2021, 33(2):135-146. YANG Y, SHI W Z, ZHANG X M, et al. Identification method of shale lithofacies by logging curves:A case study from WufengLongmaxi Formation in Jiaoshiba area, SW China. Lithologic Reservoirs, 2021, 33(2):135-146.
[14] 刘跃杰, 刘书强, 马强, 等. BP神经网络法在三塘湖盆地芦草沟组页岩岩相识别中的应用.岩性油气藏, 2019, 31(4):101111. LIU Y J, LIU S Q, MA Q, et al. Application of BP neutral network method to identification of shale lithofacies of Lucaogou Formation in Santanghu Basin. Lithologic Reservoirs, 2019, 31(4):101-111.
[15] 朱振宇, 王贵文, 朱广宇. 人工神经网络法在烃源岩测井评价中的应用. 地球物理学进展, 2002, 17(1):137-140. ZHU Z Y, WANG G W, ZHU G Y. The application of artificial neural network to the source rock's evaluation. Progress in Geophysics, 2002, 17(1):137-140.
[16] DE S, VIKRAM V K, SENGUPTA D. Application of support vector regression analysis to estimate total organic carbon content of Cambay shale in Cambay Basin, India:A case study. Petroleum Science and Technology, 2019, 37(10):1155-1164.
[17] SHI X, WANG J, LIU G, et al. Application of extreme learning machine and neural networks in total organic carbon content prediction in organic shale with wire line logs. Journal of Natural Gas Science and Engineering, 2016, 33:687-702.
[18] RUI J, ZHANG H, REN Q, et al. TOC content prediction based on a combined Gaussian process regression model. Marine and Petroleum Geology, 2020, 118:104429.
[19] 王贵文, 朱振宇, 朱广宇. 烃源岩测井识别与评价方法研究. 石油勘探与开发, 2002, 29(4):50-52. WANG G W, ZHU Z Y, ZHU G Y. Logging identification and evaluation of Cambrian-Ordovician source rocks in syneclise of Tarim Basin. Petroleum Exploration and Development, 2002, 29(4):50-52.
[20] 蒋德鑫, 姜正龙, 张贺, 等. 烃源岩总有机碳含量测井预测模型探讨:以陆丰凹陷文昌组为例. 岩性油气藏, 2019, 31(6):112-120. JIANG D X, JIANG Z L, ZHANG H, et al. Well logging prediction models of TOC content in source rocks:A case of Wenchang Formation in Lufeng Sag. Lithologic Reservoirs, 2019, 31(6):112-120.
[21] JOHNSON L M, REZAEE R, KADKHODAIE A, et al. Geochemical property modelling of a potential shale reservoir in the Canning Basin(western Australia), using artificial neural networks and geostatistical tools. Computers & Geosciences, 2018, 120:73-81.
[22] 高乔, 王兴志, 朱逸青, 等. 川南地区龙马溪组元素地球化学特征及有机质富集主控因素.岩性油气藏, 2019, 31(4):72-84. GAO Q, WANG X Z, ZHU Y Q, et al. Elemental geochemical characteristics and main controlling factors of organic matter enrichment of Longmaxi Formation in southern Sichuan. Lithologic Reservoirs, 2019, 31(4):72-84.
[23] 刘文平, 周政, 吴娟, 等. 川南盆地长宁页岩气田五峰组-龙马溪组成藏动力学过程及其意义. 南京大学学报(自然科学), 2020, 56(3):393-404. LIU W P, ZHOU Z, WU J, et al. Hydrocarbon generation and shale gas accumulation in the Wufeng-Longmaxi formations, Changning shale-gas field, southern Sichuan Basin. Journal of Nanjing University(Natural Science), 2020, 56(3):393-404.
[24] 黄利文.基于理想点的主成分分析法在综合评价中的应用. 统计与决策, 2021, 37(10):184-188. HUANG L W. Application of principal component analysis based on ideal point in comprehensive evaluation. Statistics and Decision, 2021, 37(10):184-188.
[25] 黄继盛, 姚康, 束洪春, 等. 基于主成分分析和SVM模型的配电网故障选线方法.昆明理工大学学报(自然科学版), 2021, 46(4):1-12. HUANG J S, YAO K, SHU H C, et al. Method of fault line selection in distribution network based on principal component analysis and SVM model. Journal of Kunming University of Science and Technology(Natural Science Edition), 2021, 46(4):1-12.
[26] 张继伟.基于主成分分析的页岩油有利区评价:以仪陇-平昌地区大安寨段为例.断块油气田, 2021, 28(1):28-32. ZHANG J W. Evaluation of favorable areas of shale oil based on principal component analysis:Taking Daanzhai member of Yilong-Pingchang area as an example. Fault Block Oil and Gas Field, 2021, 28(1):28-32.
[27] 周游, 张广智, 高刚, 等. 核主成分分析法在测井浊积岩岩性识别中的应用.石油地球物理勘探, 2019, 54(3):667-675. ZHOU Y, ZHANG G Z, GAO G, et al. Application of kernel principal component analysis in well logging turbidite lithology identification. Oil Geophysical Prospecting, 2019, 54(3):667675.
[28] 曲彦胜, 钟宁宁, 刘岩, 等. 烃源岩有机质丰度的测井计算方法及影响因素探讨. 岩性油气藏, 2011, 23(2):80-84. QU Y S, ZHONG N N, LIU Y, et al. Using logging methods to calculate organic matter abundance of source rocks and its influencing factors. Lithologic Reservoirs, 2011, 23(2):80-84.
[29] 周瑾, 白洪溪, 崔健, 等. 基于电磁参数的BP神经网络模型在页岩气储层预测中的应用研究.物探化探计算技术, 2020, 42(1):76-83. ZHOU J, BAI H X, CUI J, et al. Application of BP neural network model based on electromagnetic parameters in shale gas reservoir prediction. Geophysical and Geochemical Calculation Technology, 2020, 42(1):76-83.
[30] 袁颖, 谭丁, 于少将, 等. 基于贝叶斯正则化改进BP神经网络的页岩气有机碳含量预测模型.地质与勘探, 2019, 55(4):1082-1091. YUAN Y, TAN D, YU S J, et al. A Prediction model for shale gas organic carbon content based on improved BP neural network using Bayesian regularization. Geology and Exploration, 2019, 55(4):1082-1091.
[31] 朱红, 孔德群, 钱旭.基于ATD-BP神经网络的页岩气产量预测方法.科学技术与工程, 2017, 17(31):128-132. ZHU H, KONG D Q, QIAN X. Shale gas production prediction method based on adaptive threshold denoising BP Neural Network. Science Technology and Engineering, 2017, 17(31):128132.
[32] 谷宇峰, 张道勇, 鲍志东. PSO-GBDT识别致密砂岩储集层岩性研究:以姬塬油田西部长4+5段为例. 矿物岩石地球化学通报, 2021, 40(3):624-634. GU Y F, ZHANG D Y, BAO Z D. Lithology prediction of tight sandstone reservoirs using the PSO-GBDT:A case study of the Chang 4+5 members in the western Jiyuan Oilfield. Mineral and Petrochemical Bulletin, 2021, 40(3):624-634.
[33] 谷宇峰, 张道勇, 鲍志东, 等. 利用梯度提升决策树(GBDT)预测渗透率:以姬塬油田西部长4+5段致密砂岩储层为例. 地球物理学进展, 2021, 36(2):585-594. GU Y F, ZHANG D Y, BAO Z D, et al. Permeability prediction using Gradient Boosting Decision Tree(GBDT):A case study of tight sandstone reservoirs of member of Chang 4+5 in western Jiyuan Oilfield. Progress in Geophysics, 2021, 36(2):585-594.
[34] 谷宇峰, 张道勇, 鲍志东, 等.利用GBDT识别致密砂岩储层岩性.地球物理学进展, 2021, 36(5):1956-1965. GU Y F, ZHANG D Y, BAO Z D, et al. Identification of tight sandstone reservoir lithology using GBDT. Progress in Geophysics, 2021, 36(5):1956-1965.
[35] 尹超英, 邵春福, 黄兆国, 等.基于梯度决策树的多尺度建成环境对小汽车拥有的影响.吉林大学学报(工学版):1-8[202111-02].https:?doi.org/10.13229/j.cnki.jdxbgxb20200844. YIN C Y, SHAO C F, HUANG Z G, et al. Investigating the influences of multi-scale built environment on car ownership behavior based on gradient boosting decision trees. Journal of Jilin University(Engineering Edition):1-8[2021-11-02].https:?doi. org/10.13229/j.cnki.jdxbgxb20200844.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于2种机器学习方法的页岩TOC含量评价——以川南五峰组—龙马溪组为例

Evaluation of shale TOC content based on two machine learning methods: A case study of Wufeng-Longmaxi Formation in southern Sichuan Basin

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 10

[1]	杨学锋, 赵圣贤, 刘勇, 刘绍军, 夏自强, 徐飞, 范存辉, 李雨桐. 四川盆地宁西地区奥陶系五峰组—志留系龙马溪组页岩气富集主控因素[J]. 岩性油气藏, 2024, 36(5): 99-110.
[2]	邱玉超, 李亚丁, 文龙, 罗冰, 姚军, 许强, 文华国, 谭秀成. 川东地区寒武系洗象池组构造特征及成藏模式[J]. 岩性油气藏, 2024, 36(5): 122-132.
[3]	程静, 闫建平, 宋东江, 廖茂杰, 郭伟, 丁明海, 罗光东, 刘延梅. 川南长宁地区奥陶系五峰组—志留系龙马溪组页岩气储层低电阻率响应特征及主控因素[J]. 岩性油气藏, 2024, 36(3): 31-39.
[4]	熊波, 朱冬雪, 方朝合, 王社教, 杜广林, 薛亚斐, 莫邵元, 辛福东. 基于BP算法的中深层同轴套管换热量预测[J]. 岩性油气藏, 2024, 36(2): 15-22.
[5]	杨跃明, 张少敏, 金涛, 明盈, 郭蕊莹, 王兴志, 韩璐媛. 川南地区二叠系龙潭组页岩储层特征及勘探潜力[J]. 岩性油气藏, 2023, 35(1): 1-11.
[6]	闫建平, 罗静超, 石学文, 钟光海, 郑马嘉, 黄毅, 唐洪明, 胡钦红. 川南泸州地区奥陶系五峰组—志留系龙马溪组页岩裂缝发育模式及意义[J]. 岩性油气藏, 2022, 34(6): 60-71.
[7]	张梦琳, 李郭琴, 何嘉, 衡德. 川西南缘天宫堂构造奥陶系五峰组—志留系龙马溪组页岩气富集主控因素[J]. 岩性油气藏, 2022, 34(2): 141-151.
[8]	李小佳, 邓宾, 刘树根, 吴娟, 周政, 焦堃. 川南宁西地区五峰组—龙马溪组多期流体活动[J]. 岩性油气藏, 2021, 33(6): 135-144.
[9]	张兵, 唐书恒, 郗兆栋, 蔺东林, 叶亚培. 湘西北地区五峰组—龙马溪组生物地层特征及勘探意义[J]. 岩性油气藏, 2021, 33(5): 11-21.
[10]	丛平, 闫建平, 井翠, 张家浩, 唐洪明, 王军, 耿斌, 王敏, 晁静. 页岩气储层可压裂性级别测井评价及展布特征——以川南X地区五峰组—龙马溪组为例[J]. 岩性油气藏, 2021, 33(3): 177-188.
[11]	杨洋, 石万忠, 张晓明, 王任, 徐笑丰, 刘俞佐, 白卢恒, 曹沈厅, 冯芊. 页岩岩相的测井曲线识别方法——以焦石坝地区五峰组-龙马溪组为例[J]. 岩性油气藏, 2021, 33(2): 135-146.
[12]	蒋德鑫, 姜正龙, 张贺, 杨舒越. 烃源岩总有机碳含量测井预测模型探讨——以陆丰凹陷文昌组为例[J]. 岩性油气藏, 2019, 31(6): 109-117.
[13]	高乔, 王兴志, 朱逸青, 赵圣贤, 张芮, 肖哲宇. 川南地区龙马溪组元素地球化学特征及有机质富集主控因素[J]. 岩性油气藏, 2019, 31(4): 72-84.
[14]	刘跃杰, 刘书强, 马强, 姚宗森, 佘家朝. BP神经网络法在三塘湖盆地芦草沟组页岩岩相识别中的应用[J]. 岩性油气藏, 2019, 31(4): 101-111.
[15]	郑珊珊, 刘洛夫, 汪洋, 罗泽华, 王曦蒙, 盛悦, 许同, 王柏寒. 川南地区五峰组—龙马溪组页岩微观孔隙结构特征及主控因素[J]. 岩性油气藏, 2019, 31(3): 55-65.