岩性油气藏 ›› 2024, Vol. 36 ›› Issue (2): 65–75.doi: 10.12108/yxyqc.20240207

• 地质勘探 • 上一篇    

基于特征变量扩展的含气饱和度随机森林预测方法

桂金咏, 李胜军, 高建虎, 刘炳杨, 郭欣   

  1. 中国石油勘探开发研究院 西北分院, 兰州 730020
  • 收稿日期:2023-10-09 修回日期:2023-11-15 发布日期:2024-03-06
  • 作者简介:桂金咏（1986—），男，博士，高级工程师，主要从事油气储层地震综合预测研究。地址：（730020）甘肃省兰州市城关区雁儿湾路535号。Email：guijy@petrochina.com.cn。
  • 基金资助:
    中国石油天然气股份公司前瞻性基础性重大科技项目“复杂气藏地震识别与预测技术研究”（编号：2021DJ0606）与中国石油直属院所基础研究和战略储备技术研究基金项目“孔、渗、饱多参数联合智能化地震反演技术研究”（编号：2022D-XB01）联合资助。

A random forests prediction method for gas saturation based on feature variable extension

GUI Jinyong, LI Shengjun, GAO Jianhu, LIU Bingyang, GUO Xin   

  1. PetroChina Research Institute of Petroleum Exploration and Development-Northwest, Lanzhou 730020, China
  • Received:2023-10-09 Revised:2023-11-15 Published:2024-03-06

PDF (PC)

111

摘要： 采用数据驱动的方式，提出了一种基于随机森林机器学习算法训练出含气饱和度地震预测方法，并将该方法应用于中国西部复杂天然气藏中，分别对单井资料和二维地震资料进行了含气饱和度预测与分析。研究结果表明：①抽取井旁道纵波速度、横波速度和密度3个弹性参数叠前地震反演结果作为基本特征变量样本，引入边界合成少数类过采样技术对基本特征变量样本和对应的含气饱和度样本进行平衡化处理；利用扩展弹性阻抗结合数学变换自动生成一系列的扩展变量；再利用随机森林对特征变量进行含气饱和度预测重要性排名，并优选重要性较高的特征变量进行含气饱和度随机森林训练。②该方法大幅减少了特征变量提取和优选的人工工作量，且有效减少了信息冗余以及因含气饱和度样本不平衡导致的训练偏倚问题，有效增强了随机森林算法在含气饱和度地震预测方面的能力。③实际单井应用中预测的含气饱和度与测井解释的含气饱和度的相关系数可达0.985 5；在二维地震资料应用中，该方法比基于常规未平衡化的11个弹性参数作为随机森林输入预测出的含气饱和度精度更高。

关键词: 含气饱和度, 随机森林, 纵波速度, 横波速度, 密度, 特征变量, 不平衡数据, 机器学习, 气层预测, 地震预测

Abstract: A data-driven approach was proposed to predict gas saturation based on random forests machine learning algorithm. This method was applied to predict and analyze gas saturation in a complex natural gas reservoir in western China from single well data and two-dimensional seismic data respectively. The results show that：（1）The method extracts the pre-stack seismic inversion results of three elastic parameters from the uphole trace（compressional wave velocity，shear wave velocity，and density）from well log data as basic feature variables. It employs the boundary synthetic minority oversampling technique to balance the basic feature variables and corresponding gas saturation samples，generates a series of extended variables by combining the extended elastic impedance with mathematical transformations，and then uses random forests to rank the importance of the feature variables for gas saturation prediction，finally selects feature variables with higher importance for gas saturation random forests training.（2）This method significantly reduces the manual workload for feature variables extraction andselection，effectively reduces information redundancy and training bias caused by imbalanced gas saturation samples，and effectively enhances the capability of the random forests algorithm in predicting gas saturation. （3）In practical applications，the predicted gas saturation using this method shows a high correlation coefficient of 0.985 5 with the gas saturation by log interpretation. In the case of two-dimensional data，it achieves higher accuracy in gas saturation prediction compared to using 11 conventional unbalanced elastic parameters as inputs for random forests.

Key words: gas saturation, random forests, compressional wave velocity, shear wave velocity, density, feature variable, imbalance data, machine learning, gas layer prediction, seismic prediction

中图分类号: 

  • P618.13
[1] 胡胜福,周灿灿,李霞,等.测井饱和度解释模型的演化历程分析与思考[J]. 地球物理学进展,2017,32(5):1992-1998. HU Shengfu,ZHOU Cancan,LI Xia,et al. Analysis and prospect of logging saturation model evolutionary history[J]. Progress in Geophysics,2017,32(5):1992-1998.
[2] 何绪全,黄东,赵艾琳,等. 川中地区大安寨段页岩油气储层测井评价指标体系[J]. 岩性油气藏,2021,33(3):129-137. HE Xuquan,HUANG Dong,ZHAO Ailin,et al. Well-logging evaluation index system of shale oil and gas reservoir of Da'an-zhai member in central Sichuan Basin[J]. Lithologic Reservoirs, 2021,33(3):129-137.
[3] 刘仕友,段治川,周凡,等. 基于布谷鸟算法的储层物性参数同步反演[J]. 石油地球物理勘探,2022,57(3):638-646. LIU Shiyou,DUAN Zhichuan,ZHOU Fan,et al. Simultaneous inversion of petrophysical parameters of reservoir based on cuckoo search algorithm[J]. Oil Geophysical Prospecting,2022, 57(3):638-646.
[4] BACHRACH R. Joint estimation of porosity and saturation using stochastic rock-physics modeling[J]. Geophysics,2006,71(5):O53-O63.
[5] 胡华锋. 基于叠前道集的储层参数反演方法研究[D]. 青岛:中国石油大学(华东),2011. HU Huafeng. The study of petrophysical-properties inversion base on pre-stack seismic data[D]. Qingdao:China University of Petroleum(East China),2011.
[6] DE FIGUEIREDO L P,GRANA D,SANTOS M,et al. Bayesian seismic inversion based on rock-physics prior modeling for the joint estimation of acoustic impedance,porosity and lithofacies[J]. Journal of Computational Physics,2017,336:128-142.
[7] 刘兴业,陈小宏,李景叶,等. 基于核贝叶斯判别法的储层物性参数预测[J]. 石油学报,2016,37(7):878-886. LIU Xingye,CHEN Xiaohong,LI Jingye,et al. Reservoir physical property prediction based on kernel-Bayes discriminant method[J]. Acta Petrolei Sinica,2016,37(7):878-886.
[8] 李红兵,张佳佳,潘豪杰,等. 基于弹性阻抗的孔隙结构与物性参数非线性同步反演[J]. 中国科学:地球科学,2021,51(7):1166-1180. LI Hongbing,ZHANG Jiajia,PAN Haojie,et al. Nonlinear simultaneous inversion of pore structure and physical parameters based on elastic impedance[J]. Scientia Sinica(Terrae),2021,51(7):1166-1180.
[9] 桂金咏,高建虎,雍学善,等. 基于双相介质理论的储层参数反演方法[J]. 地球物理学报,2015,58(9):3424-3438. GUI Jinyong,GAO Jianhu,YONG Xueshan,et al. Inversion of reservoir parameters based on dual-phase media theory[J]. Chinese Journal of Geophysics,2015,58(9):3424-3438.
[10] LANG Xiaozheng,GRANA D. Bayesian linearized petrophysical AVO inversion[J]. Geophysics,2018,83(3):M1-M13.
[11] LIU Qian,DONG Ning,JI Yuxin,et al. Direct reservoir property estimation based on prestack seismic inversion[J]. Journal of Petroleum Science and Engineering,2018,171:1475-1486.
[12] 李坤,印兴耀,宗兆云. 岩石物理驱动的相约束叠前地震概率化反演方法[J]. 中国科学:地球科学,2020,50(6):832-850. LI Kun,YIN Xingyao,ZONG Zhaoyun. Facies-constrained prestack seismic probabilistic inversion driven by rock physics[J]. Scientia Sinica(Terrae),2020,50(6):832-850.
[13] FJELDSTAD T,AVSETH P,OMRE H. A one-step Bayesian inversion framework for 3D reservoir characterization based on a Gaussian mixture model:A Norwegian sea demonstration[J]. Geophysics,2021,86(2):R221-R236.
[14] AVSETH P,MUKERJI T,MAVKO G. Quantitative seismic interpretation:Applying rock physics tools to reduce interpretation risk[M]. Cambridge:Cambridge University Press,2010.
[15] 彭达,肖富森,冉崎,等. 基于KT模型流体替换的岩石物理参数反演方法[J]. 岩性油气藏,2018,30(5):82-90. PENG Da,XIAO Fusen,RAN Qi,et al. Inversion of rock physics parameters based on KT model fluid substitution[J]. Lithologic Reservoirs,2018,30(5):82-90.
[16] 桂金咏,高建虎,李胜军,等. 基于弹性参数加权统计的地震岩相预测方法[J]. 地球物理学报,2020,63(1):298-312. GUI Jinyong,GAO Jianhu,LI Shengjun,et al. The method of seismic lithofacies prediction based on weighted statistics of elastic parameters[J]. Chinese Journal of Geophysics,2020,63(1):298-312.
[17] GUI Jinyong,YIN Xingyao,GAO Jianhu,et al. Petrophysical properties prediction of deep dolomite reservoir considering pore structure[J]. Acta Geophysica,2022,70(4):1507-1518.
[18] 敖亦乐. 随机森林类算法在地球物理勘探中的相关应用技术研究[D]. 北京:中国石油大学(北京),2022. AO Yile. Research on the random forest algorithms and their applications in geophysical exploration interpretation[D]. Beijing:China University of Petroleum(Beijing),2022.
[19] BREIMAN L. Random forests[J]. Machine Learning,2001,45(1):5-32.
[20] HARRIS J R,GRUNSKY E C. Predictive lithological mapping of Canada's North using random forest classification applied to geophysical and geochemical data[J]. Computers & Geosciences,2015,80:9-25.
[21] 宋建国,高强山,李哲. 随机森林回归在地震储层预测中的应用[J]. 石油地球物理勘探,2016,51(6):1202-1211. SONG Jianguo,GAO Qiangshan,LI Zhe. Application of random forests for regression to seismic reservoir prediction[J]. Oil Geophysical Prospecting,2016,51(6):1202-1211.
[22] 王光宇,宋建国,徐飞,等. 不平衡样本集随机森林岩性预测方法[J]. 石油地球物理勘探,2021,56(4):679-687. WANG Guangyu,SONG Jianguo,XU Fei,et al. Random forests lithology prediction method for imbalanced data sets[J]. Oil Geophysical Prospecting,2021,56(4):679-687.
[23] KUHN S,CRACKNELL M J,READING A M. Lithologic mapping using random forests applied to geophysical and remotesensing data:A demonstration study from the eastern Goldfields of AustraliaLithologic mapping with RF[J]. Geophysics, 2018,83(4):B183-B193.
[24] CRACKNELL M J,READING A M. Geological mapping using remote sensing data:A comparison of five machine learning algorithms,their response to variations in the spatial distribution of training data and the use of explicit spatial information[J]. Computers & Geosciences,2014,63:22-33.
[25] CRACKNELL M J. Machine learning for geological mapping:Algorithms and applications[D]. Tasmania:University of Tasmania,2014.
[26] KIM Y,HARDISTY R,TORRES E,et al. Seismic facies classification using random forest algorithm[G]. SEG Technical Program Expanded Abstracts,2018:2161-2165.
[27] LUBO-ROBLES D,DEVEGOWDAD,JAYARAM V,et al. Quantifying the sensitivity of seismic facies classification to seismic attribute selection:An explainable machine-learning study[J]. Interpretation,2022,10(3):SE41-SE69.
[28] ALVAREZ P,BOLIVAR F,LUCA M,et al. Multi-attribute rotation scheme:A tool for reservoir property prediction from seismic inversion attributes[J]. Interpretation,2015,3(4):SAE9-SAE18.
[29] WHITCOMBE D N,CONNOLLY P A,REAGAN R L,et al. Extended elastic impedance for fluid and lithology prediction[J]. Geophysics,2002,67(1):63-67.
[30] NEVES A,MUSTAFA M,RUTTY M. Pseudo-gamma ray volume from extended elastic impedance inversion for gas exploration[J]. The Leading Edge,2004,23(6):536-540.
[31] ZHANG Guoyin,WANG Zhizhang,CHEN Yangkang. Deep learning for seismic lithology prediction[J]. Geophysical Journal International,2018,215(2):1368-1387.
[32] HOSSAIN T M,WATADA J,JIAN Zhiwen,et al. Missing well log data handling in complex lithology prediction:An nis apriori algorithm approach[J]. International Journal of Innovative Computing,Information and Control,2020,16(3):1077-1091.
[33] CHAWLA N V,BOWYER K W,HALL L O,et al. SMOTE:Synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research,2002,16:321-357.
[34] HAN Hui,WANG Wenyuan,MAO Binghuan. BorderlineSMOTE:A new over-sampling method in imbalanced data sets learning[M]//GOOS G,HARTMANIS J. Lecture notes in computer science. Heidelberg:Springer,2005:878-887.
[35] STROBL C,BOULESTEIX A L,ZEILEIS A,et al. Bias in random forest variable importance measures:Illustrations,sources and a solution[J]. BMC Bioinformatics,2007,8(1):1-21.
[36] GOODWAY B,CHEN Taiwen,DOWNTON J. Improved AVO fluid detection and lithology discrimination using Lamé petrophysical parameters "λρ","μρ",&"λ/μ fluid stack",from P and S inversions[G]. SEG Technical Program Expanded Abstracts, 1997:183-186.
[1] 郭海峰, 肖坤叶, 程晓东, 杜业波, 杜旭东, 倪国辉, 李贤兵, 计然. 乍得Bongor盆地花岗岩潜山裂缝型储层有效渗透率计算方法[J]. 岩性油气藏, 2023, 35(6): 117-126.
[2] 唐昱哲, 柴辉, 王红军, 张良杰, 陈鹏羽, 张文起, 蒋凌志, 潘兴明. 中亚阿姆河右岸东部地区侏罗系盐下碳酸盐岩储层特征及预测新方法[J]. 岩性油气藏, 2023, 35(6): 147-158.
[3] 许鑫, 杨午阳, 张凯, 魏新建, 张向阳, 李海山. 三维初至波旅行时层析速度反演算法优化[J]. 岩性油气藏, 2023, 35(4): 79-89.
[4] 马乔雨, 张欣, 张春雷, 周恒, 武中原. 基于一维卷积神经网络的横波速度预测[J]. 岩性油气藏, 2021, 33(4): 111-120.
[5] 武中原, 张欣, 张春雷, 王海英. 基于LSTM循环神经网络的岩性识别方法[J]. 岩性油气藏, 2021, 33(3): 120-128.
[6] 张艳, 高世臣, 孟婉莹, 成育红, 蒋思思. 致密砂岩储层AVO正演模拟过程中的不确定性分析[J]. 岩性油气藏, 2020, 32(6): 120-128.
[7] 孙予舒, 黄芸, 梁婷, 季汉成, 向鹏飞, 徐新蓉. 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.
[8] 张璐, 何峰, 陈晓智, 祝彦贺, 韩刚, 李祺鑫. 基于倾角导向滤波控制的似然属性方法在断裂识别中的定量表征[J]. 岩性油气藏, 2020, 32(2): 108-114.
[9] 宋宣毅, 刘月田, 马晶, 王俊强, 孔祥明, 任兴南. 基于灰狼算法优化的支持向量机产能预测[J]. 岩性油气藏, 2020, 32(2): 134-140.
[10] 李贤胜, 刘向君, 熊健, 李玮, 梁利喜. 层理对页岩纵波特性的影响[J]. 岩性油气藏, 2019, 31(3): 152-160.
[11] 严伟, 刘帅, 冯明刚, 张冲, 范树平. 四川盆地丁山区块页岩气储层关键参数测井评价方法[J]. 岩性油气藏, 2019, 31(3): 95-104.
[12] 田亮, 李佳玲, 焦保雷. 塔河油田12区奥陶系油藏溶洞充填机理及挖潜方向[J]. 岩性油气藏, 2018, 30(3): 52-60.
[13] 刘化清, 苏明军, 倪长宽, 洪忠, 崔向丽, 胡凯峰, 李政阳, 毛俊丽. 薄砂体预测的地震沉积学研究方法[J]. 岩性油气藏, 2018, 30(2): 1-11.
[14] 刘智颖, 章成广, 唐军, 肖承文. 裂缝对岩石电阻率的影响及其在含气饱和度计算中的应用[J]. 岩性油气藏, 2018, 30(2): 120-128.
[15] 王永生, 胡杰, 张静, 王传武, 朱波, 吴永国. 宽频勘探技术在柴北缘深层侏罗系勘探中的应用[J]. 岩性油气藏, 2017, 29(6): 101-107.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!

陇ICP备17006252号
版权所有© 2018 中国石油集团西北地质研究所有限公司《岩性油气藏》编辑部

地址：甘肃省兰州市城关区雁儿湾路535号(陇ICP备17006252号)    电话：0931-8686158;8686058;8686288    邮箱：yxyqc@petrochina.com.cn

甘公网安备 62010202002827号

本系统由北京玛格泰克科技发展有限公司设计开发