基于一维卷积神经网络的横波速度预测

doi:10.12108/yxyqc.20210412

岩性油气藏 ›› 2021, Vol. 33 ›› Issue (4): 111–120.doi: 10.12108/yxyqc.20210412

基于一维卷积神经网络的横波速度预测

马乔雨¹, 张欣², 张春雷³, 周恒¹, 武中原¹

1. 中国地质大学 (北京) 数理学院, 北京, 100083;
2. 北京师范大学统计学院, 北京, 100875;
3. 北京中地润德石油科技有限公司, 北京, 100083

收稿日期:2020-10-23 修回日期:2020-12-25 出版日期:2021-08-01 发布日期:2021-08-06
通讯作者: 张欣(1996-),女,北京师范大学在读博士研究生,研究方向为统计学习、机器学习、深度学习。Email:2249288385@qq.com。 E-mail:2249288385@qq.com
作者简介:马乔雨(1995-),男,中国地质大学(北京)在读硕士研究生,研究方向为机器学习、深度学习。地址:(100083)北京市海淀区学院路29号。Email:1598974727@qq.com
基金资助:
国家科技重大专项“鄂尔多斯盆地大型岩性地层油气藏勘探开发示范工程”（编号：2016ZX05050）资助

Shear wave velocity prediction based on one-dimensional convolutional neural network

MA Qiaoyu¹, ZHANG Xin², ZHANG Chunlei³, ZHOU Heng¹, WU Zhongyuan¹

1. School of Science, China University of Geosciences (Beijing), Beijing 100083, China;
2 School of Statistics, Beijing Normal University, Beijing 100875, China;
3. Beijing Zhongdirunde Petroleum Technology Co., Ltd., Beijing 100083, China

Received:2020-10-23 Revised:2020-12-25 Online:2021-08-01 Published:2021-08-06

摘要/Abstract

摘要： 纵横波速度是地球物理勘探中识别储层岩性、物性和储层等目标极其重要的信息，由于采集技术与成本投入的限制，横波速度资料通常较为缺乏，横波速度预测便成为岩石物理分析中亟需解决的重要问题。传统上基于理论方法和经验公式的横波速度转换方法局限性较大，常规的点对点的机器学习方法无法有效表达测井参数的空间特征，对横波速度与其它测井参数之间的内在关系的表征不够充分。为此，开展了基于一维卷积神经网络（1D-CNN）模型的横波速度预测方法研究，基于声波时差、密度、自然伽马和电阻率等16种测井参数建立深度学习回归模型，通过不同尺度卷积提取测井参数在测井深度空间上的结构性特征，并采用多层网络结构，学习横波参数与测井参数深度特征之间的关系，从而建立更为精确的预测模型。通过在苏里格气田上古生界碎屑岩储层的实际应用，验证了一维卷积神经网络模型的横波速度预测精度较高，且具有良好的泛化性。

关键词: 横波速度, 测井参数, 碎屑岩储层, 一维卷积神经网络, 深度学习, 苏里格气田

Abstract: Shear and compressional waves is extremely important information for identifying reservoir lithology, physical properties and fluids in geophysical exploration. Due to the limitation of acquisition technology and cost investment,shear wave data is usually lacking,and shear wave prediction has become an urgent need for petrophysical analysis. Traditionally,shear wave conversion methods based on theoretical models and empirical formulas have great limitations. Conventional machine learning models which is point-to-point cannot effectively express the spatial characteristics of logging parameters,and the inherent relationship between shear wave parameters and other logging parameters is not sufficiently represented. To this end,a study on shear wave velocity prediction methods based on one-dimensional convolutional neural network（1D-CNN）models was carried out. A deep learning regression model was established based on 16 logging parameters such as acoustic time difference,density,natural gamma and resistivity. The structural characteristics of logging parameters in the logging depth space was extracted by different scale convolution and a multi-layer network structure learned the relationship between shear wave parameters and logging parameters depth characteristics,thereby establishing a more accurate prediction model. Through the practical application in the Upper Paleozoic clastic reservoir of the Sulige gas field,it is verified that the one-dimensional convolutional neural network model has higher shear wave prediction accuracy.

Key words: wave velocity, logging parameters, clastic reservoir, 1D-CNN, deep learning, Sulige gas field

中图分类号:

P631.4

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.

MA Qiaoyu, ZHANG Xin, ZHANG Chunlei, ZHOU Heng, WU Zhongyuan. Shear wave velocity prediction based on one-dimensional convolutional neural network[J]. Lithologic Reservoirs, 2021, 33(4): 111–120.

参考文献

[1] 谢冰, 白利, 赵艾琳, 等. Sonic Scanner声波扫描测井在碳酸盐岩储层裂缝有效性评价中的应用:以四川盆地震旦系为例. 岩性油气藏,2017, 29(4):117-123. XIE B, BAI L, ZHAO A L, et al. Application of sonic scanner logging to fracture effectiveness evaluation of carbonate reservoir:A case from Sinian in Sichuan Basin. Lithologic Reservoirs, 2017, 29(4):117-123.
[2] 丁燕, 杜启振, 刘力辉, 等.基于纵横波同步联合的孔隙模量三参数AVO反演方法.岩性油气藏, 2020, 32(1):111-119. DING Y, DU Q Z, LIU L H, et al. Three-parameter AVO inversion method of pore modulus based on PP and PS wave simultaneous joint inversion. Lithologic Reservoirs, 2020, 32(1):111-119.
[3] 张艳, 高世臣, 孟婉莹等.致密砂岩储层AVO正演模拟过程中的不确定性分析.岩性油气藏, 2020, 32(6):120-128. ZHANG Y, GAO S C, MENG W Y, et al. Uncertainty analysis in AVO forward modeling for tight sandstone reservoirs simultaneous joint inversion. Lithologic Reservoirs, 2020, 32(6):120-128.
[4] CASTAGNA J P, BATZLE M L, KAN T K, et al. Offset-dependent reflectivity-theory and practice of AVO analysis Investigations in Geophysics, 1993, 8:135-171.
[5] RØGEN B, GOMMESEN L, FABRICIUS I L. Methods of velocity prediction tested for North Sea chalk:A review of fluid substitution and S estimates. Journal of Petroleum Science & Engineering, 2004,45(1-2):129-139.
[6] BERRYMAN J G. Poroelastic shear modulus dependence on pore-fluid properties arising in a model of thin isotropic layers. Geophysical Journal International, 2004, 157(1):415-425.
[7] CASTAGNA J P, BATZLE M L, EASTWOOD R L. Relationships between compressional-wave and shear wave velocities in clastic silicate rocks. Geophysics, 1985, 50:571-581.
[8] MA X H, YANG Y, WEN L, et al. Distribution and exploration direction of medium-and large-sized marine carbonate gas fields in Sichuan Basin, SW China. Petroleum Exploration and Development, 2019, 46(1):1-13.
[9] LI Y, KANG Z J, XUE Z J, et al. Theories and practices of carbonate reservoirs development in China. Petroleum Exploration and Development, 2018, 45(4):669-678.
[10] 张录录. 典型油气藏岩石物理特征分析及应用研究. 青岛:中国石油大学(华东), 2013. ZHANG L L. Analysis of rock physics characteristics of typical reservoir and applied research. Qingdao:China University of Petroleum (East China), 2013.
[11] XU J, TAN M, WANG X, et al. Predicting acoustic-wave velocities and fluid sensitivity to elastic properties in fractured carbonate formation. Interpretation, 2017, 5(1):SB69-SB80.
[12] ZHANG Y, ZHONG H R, WU Z Y, et al. Improvement of petrophysical workflow for shear wave velocity prediction based on machine learning methods for complex carbonate reservoirs. Journal of Petroleum Science and Engineering, 2020, 192:107234.
[13] 张炜, 唐宇, 余迎. 机器学习在地球物理测井中的应用进展. 测井技术, 2020(2):185. ZHANG W, TANG Y, YU Y. Application progress of machine learning in geophysical logging. Well Logging Technology, 2020(2):185.
[14] 江凯, 王守东, 胡永静, 等. 基于Boosting Tree算法的测井岩性识别模型. 测井技术, 2018, 42(4):395-400. JIANG K, WANG S D, HU Y J, et al. Lithology identification model by well logging based on boosting tree algorithm. Well Logging Technology, 2018, 42(4):395-400.
[15] CHEN T Q, GUESTRIN C. XGBoost:A scalable tree boosting system. Proceedings of the 22nd ACM sigkdd international conference on knowledge discovery and data mining. San Francisco, 2016:785-794.
[16] 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. 岩性油气藏, 2020, 32(4):98-106. SUN Y S, HUANG Y, LIANG T, et al. Identification of complex carbonate lithology by logging based on XGBoost algorithm. Lithologic Reservoirs,2020,32(4):98-106.
[17] 倪维军, 李琪, 郭文惠, 等. 基于支持向量机的页岩储层横波速度预测. 西安石油大学学报(自然科学版), 2017, 32(4):46-49. NI W J, LI Q, GUO W H, et al. Prediction of shear wave velocity in shale reservoir based on support vector machine. Journal of Xi'an Shiyou University(Natural Science), 2017, 32(4):46-49.
[18] HINTON G E, SALAKHUTDINOV R R.Reducing the dimensionality of data with neural networks. Science, 2006, 313(5786):504-507.
[19] 赵军龙, 李纲, 麻平社, 等. 神经网络在石油测井解释中的应用综述. 地球物理学进展, 2010, 25(5):1744-1751. ZHAO J L, LI G, MA P S, et al. The application of network technology to petroleum logging interpretation. Progress in Geophysics(in Chinese), 2010, 25(5):1744-1751.
[20] ESKANDARI H, REZAEE M, MOHAMMADNIA M. Application of multiple regression and artificial neural network techniques to predict shear wave velocity from wireline log data for a carbonate reservoir, South-West Iran.CSEG Recorder, 2004, 29(7):40-48.
[21] 孙宇航, 刘洋. 利用GRU神经网络预测横波速度. 石油地球物理勘探, 2020, 55(3):484-492. SUN Y H, LIU Y.Prediction of S-wave velocity based on GRU neural network. Oil Geophysical Prospecting, 2020, 55(3):484-492.
[22] SUYKENS J A K, VANDEWALLE J. Least squares support vector machine classifiers. Neural Processing Letters, 1999, 9(3):293-300.
[23] GOODFELLOW I, BENGIO Y, COURVILLE A, et al. Deep learning. Massachusetts Institute of Technology Press, Cambridge, 2016.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于一维卷积神经网络的横波速度预测

Shear wave velocity prediction based on one-dimensional convolutional neural network

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]	李志远, 杨仁超, 张吉, 王一, 杨特波, 董亮. 天然气扩散散失率定量评价——以苏里格气田苏X区块为例[J]. 岩性油气藏, 2021, 33(4): 76-84.
[2]	龙盛芳, 王玉善, 李国良, 段传丽, 邵映明, 何咏梅, 陈凌云, 焦煦. 苏里格气田苏49区块盒8下亚段致密储层非均质性特征[J]. 岩性油气藏, 2021, 33(2): 59-69.
[3]	侯科锋, 李进步, 张吉, 王龙, 田敏. 苏里格致密砂岩气藏未动用储量评价及开发对策[J]. 岩性油气藏, 2020, 32(4): 115-125.
[4]	田清华, 刘俊, 张晨, 王文胜, 黄丹. 苏里格气田下古生界储层特征及主控因素[J]. 岩性油气藏, 2020, 32(2): 33-42.
[5]	柳娜, 周兆华, 任大忠, 南珺祥, 刘登科, 杜堃. 致密砂岩气藏可动流体分布特征及其控制因素——以苏里格气田西区盒8段与山1段为例[J]. 岩性油气藏, 2019, 31(6): 14-25.
[6]	冯强汉, 阳生国, 熊哲, 高航, 张佳超, 杨懿, 杨振. 苏里格气田西部S48区气水分布特征[J]. 岩性油气藏, 2019, 31(5): 61-69.
[7]	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.
[8]	崔连可, 单敬福, 李浮萍, 崔璐, 种健. 基于稀疏井网条件下的古辫状河道心滩砂体估算——以苏里格气田苏X区块为例[J]. 岩性油气藏, 2018, 30(1): 155-164.
[9]	马峥, 张春雷, 高世臣. 主成分分析与模糊识别在岩性识别中的应用[J]. 岩性油气藏, 2017, 29(5): 127-133.
[10]	杨特波, 王继平, 王一, 付斌, 薛雯, 郝骞. 基于地质知识库的致密砂岩气藏储层建模——以苏里格气田苏X区块为例[J]. 岩性油气藏, 2017, 29(4): 138-145.
[11]	魏新善, 胡爱平, 赵会涛, 康锐, 石晓英, 刘晓鹏. 致密砂岩气地质认识新进展[J]. 岩性油气藏, 2017, 29(1): 11-20.
[12]	韩光明，潘光超，付琛，罗琪，邵远，汪锐. 含气储层及盖层速度变化对地震响应和AVO 类型的影响[J]. 岩性油气藏, 2016, 28(2): 107-113.
[13]	孙卫锋，张吉，马志欣,郝骞，石林辉,张志刚 . 苏里格气田水平井随钻地质导向技术及应用[J]. 岩性油气藏, 2015, 27(6): 132-137.
[14]	刘俊州，孙赞东，刘正涛，孙永洋，董宁，夏红敏. 叠前同时反演在碳酸盐岩储层流体识别中的应用[J]. 岩性油气藏, 2015, 27(1): 102-107.
[15]	张琴，朱筱敏，刘畅，季汉成，陈祥. 周口坳陷谭庄凹陷下白垩统低孔低渗碎屑岩储层特征及成因分析[J]. 岩性油气藏, 2014, 26(4): 44-49.