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

doi:10.12108/yxyqc.20190411

岩性油气藏 ›› 2019, Vol. 31 ›› Issue (4): 101–111.doi: 10.12108/yxyqc.20190411

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

刘跃杰, 刘书强, 马强, 姚宗森, 佘家朝

中国石油吐哈油田分公司勘探开发研究院, 新疆哈密 839009

收稿日期:2019-01-09 修回日期:2019-03-20 出版日期:2019-07-21 发布日期:2019-06-21
第一作者:刘跃杰(1989-),男,硕士,工程师,主要从事储层地质建模方面的研究工作。地址:(839009)新疆哈密市伊州区石油新城街道吐哈石油基地勘探开发研究院。Email:lyj_215@sina.com。
基金资助:
十三五国家重大科技专项专题"吐哈-三塘湖盆地岩性地层油气藏分布规律与目标评价"（编号：2016ZX05001003-006）和中国石油股份有限公司重大科技专项"中国石油第四次油气资源评价"（编号：2013E-050206）联合资助

Application of BP neutral network method to identification of shale lithofacies of Lucaogou Formation in Santanghu Basin

LIU Yuejie, LIU Shuqiang, MA Qiang, YAO Zongsen, SHE Jiachao

Research Institute of Exploration and Development, PetroChina Tuha Oilfield Company, Hami 839009, Xinjiang, China

Received:2019-01-09 Revised:2019-03-20 Online:2019-07-21 Published:2019-06-21

摘要/Abstract

摘要： 对于复杂岩性页岩岩相的识别，传统的建立岩相图版的方法因未充分考虑到测井数据间的相似性造成的干扰以及与岩心实验数据尺度上的差异性，导致建立的识别图版中不同类别的样本点相互重叠、界限模糊，预测偏差较大。针对该问题，以三塘湖盆地马朗凹陷芦草沟组二段为例，在对储层特征充分认识的基础上，采用了一种基于主成分分析的BP神经网络方法，首先分析研究区岩心资料并对其进行归类组合，划分出富有机质纹层相、富碳酸盐纹层相和富凝灰质纹层相3种岩相类型，以便缩小与测井数据间的尺度误差；其次建立岩相图版并提取自然伽马、声波时差、补偿密度、补偿中子、电阻率等5条对岩相变化响应较为敏感的测井曲线，分析各主成分的因子载荷地质因素并优选出3个含有大量岩相信息的主成分PC2，PC3和PC4；最后建立起岩相与测井曲线间的映射关系，同时对研究区重点井芦1井进行了验证性的岩相识别。结果表明，与传统图版识别方法相比，将主成分分析与BP神经网络相结合的岩相识别方法可有效消除测井曲线相似性带来的干扰，解决因岩心数据与测井数据尺度不同所造成的预测偏差增大的问题，使岩相识别正确率得到明显提高。该方法对页岩岩相识别较为实用，具有一定的推广应用价值。

关键词: 主成分, 页岩岩相, BP神经网络, 测井参数, 芦草沟组, 三塘湖盆地

Abstract: For the identification of shale lithofacies,the traditional method of establishing lithofacies chart does not fully take into account the interference caused by the similarity of logging data and the differences in the scale of experimental data,which results in the overlap of different types of sample points in the established identification chart,the ambiguity of boundaries and the large deviation of prediction. Aiming at this problem,taking the second member of Lucaogou Formation in Malang Sag of Santanghu Basin as an example,based on the full understanding of reservoir characteristics,a BP neutral network method based on principal component analysis was adopted. Firstly,the core data of the study area were used to classified the lithofacies into three types,such as organicrich laminar facies,carbonate-rich laminar facies and rich tuff-grain laminar facies,so as to reduce the scale error with the logging data. Secondly,the lithofacies chart was established to extract logging curves such as AC,GR, DEN,CNL,R_t and so on,which were sensitive to the response of lithofacies,the factor loading geological factors of each principal component were analyzed,and three principal components PC2,PC3,PC4 containing a large amount of lithofacies information were selected. Finally,the mapping relationship between lithofacies and logging curves was established,and the lithofacies identification of well Lu1,a key well in the study area,was carried out. The results show that compared with the traditional chart identification method,the lithofacies identification method combining principal component analysis with BP neural network can effectively eliminate the interference caused by the similarity of logging curves and reduce the error caused by the difference between the core data and the logging data,as so to improve the accuracy of lithofacies identification. This method is practical for shale lithofacies identification and has certain application value.

Key words: principal component, shale lithofacies, BP neural network, logging parameters, Lucaogou Formation, Santanghu Basin

中图分类号:

P631.84

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.

LIU Yuejie, LIU Shuqiang, MA Qiang, YAO Zongsen, SHE Jiachao. Application of BP neutral network method to identification of shale lithofacies of Lucaogou Formation in Santanghu Basin[J]. Lithologic Reservoirs, 2019, 31(4): 101–111.

参考文献

[1] 王宏语,杨润泽,张峰,等.富含有机质泥页岩岩相表征的研究现状与趋势.地质科技情报,2018,37(2):141-148. WANG H Y,YANG R Z,ZHANG F,et al. Research progress and trend of organic-rich shale lithofacies characterization. Geological Science and Technology Information,2018,37(2):141148.
[2] 欧成华,李朝纯.页岩岩相表征及页理缝三维离散网络模型. 石油勘探与开发,2017,44(2):309-318. OU C H,LI C C. 3D discrete network modeling of shale bedding fractures based on lithofacies characterization. Petroleum Exploration and Development,2017,44(2):309-318.
[3] 张晋言. 页岩油测井评价方法及其应用. 地球物理学进展, 2012,27(3):1154-1162. ZHANG J Y. Well logging evaluation method of shale oil reservoirs and its applications. Progress in Geophysics,2012,27(3):1154-1162.
[4] 周成当,成菊安.模糊神经网络岩性识别系统.江汉石油学院学报,1993,15(4):40-44. ZHOU C D,CHENG J A. A lithology recognition system based on fuzzy neutral network. Journal of Jianghan Petroleum Institute,1993,15(4):40-44.
[5] 张洪,邹乐君,沈晓华.BP神经网络在测井岩性识别中的应用.地质与勘探,2002,38(6):63-65. ZHANG H,ZOU L J,SHEN X H. The application of BP neutral network in well lithology identification. Geology and Prospecting,2002,38(6):63-65.
[6] 罗伟平,范晓敏,陈军.利用一种有监督模糊ART人工神经网络进行测井岩性识别.吉林大学学报(地球科学版),2008, 38(增刊1):137-139. LUO W P,FAN X M,CHEN J. Using supervised fuzzy ART neutral network for lithology recognition in logging. Journal of Jilin University(Earth Science Edition),2008,38(Suppl 1):137-139.
[7] 张平,潘保芝,张莹,等.自组织神经网络在火成岩岩性识别中的应用.石油物探,2009,48(1):53-56. ZHANG P,PAN B Z,ZHANG Y,et al. Application of selforganization maps network in identifying the lithology of igneous rock. Geophysical Prospecting for Petroleum,2009,48(1):53-56.
[8] 朱怡翔,石广仁.火山岩岩性的支持向量机识别.石油学报, 2013,34(2):312-322. ZHU Y X,SHI G R. Identification of lithologic characteristics of volcanic rocks by support vector machine. Acta Petrolei Sinica, 2013,34(2):312-322.
[9] 赵忠军,黄强东,石林辉,等.基于BP神经网络算法识别苏里格气田致密砂岩储层岩性.测井技术,2015,39(3):363-367. ZHAO Z J,HUANG Q D,SHI L H,et al. Identification of lithology in tight sandstone reservoir in Sulige gas field based on BP neutral net algorithm. Well Logging Technology,2015,39(3):363-367.
[10] 胡嘉良,高玉超,余继峰,等.基于PCA-BP神经网络的非常规储层岩性识别研究. 山东科技大学学报(自然科学版), 2016,35(5):9-16. HU J L,GAO Y C,YU J F,et al. Lithology identification of unconventional reservoirs based on PCA-BP neural network. Journal of Shandong University of Science and Technology(Natural Science),2016,35(5):9-16.
[11] 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.岩性油气藏,2017,29(5):127-133. MA Z,ZHANG C L,GAO S C. Lithology identification based on principal component analysis and fuzzy recognition. Lithologic Reservoirs,2017,29(5):127-133.
[12] 刘毅,陆正元,吕晶,等.主成分分析法在泥页岩地层岩性识别中的应用.断块油气田,2017,24(3):360-363. LIU Y,LU Z Y,LYU J,et al. Application of principal component analysis method in lithology identification for shale formation. Fault-Block Oil & Gas Field,2017,24(3):360-363.
[13] 祝鹏,林承焰,吴鹏,等.基于主成分分析法的成岩相测井定量识别:以五号桩油田桩62-66块沙三下Ⅰ油组储层为例. 地球物理学进展,2015,30(5):2360-2365. ZHU P,LIN C Y,WU P,et al. Logging quantitative identification of diagenetic facies by using principal component analysis:a case of Es³ x¹ in Zhuang 62-66 area,Wu Hao-zhuang Oilfield. Progress in Geophysics,2015,30(5):2360-2365.
[14] 单敬福,陈欣欣,赵忠军,等.利用BP神经网络法对致密砂岩气藏储集层复杂岩性的识别.地球物理学进展,2015,30(3):1257-1263. SHAN J F,CHEN X X,ZHAO Z J,et al. Identification of complex lithology for tight sandstone gas reservoirs based on BP neural net. Progress in Geophysics,2015,30(3):1257-1263.
[15] 张国英,王娜娜,张润生,等.基于主成分分析的BP神经网络在岩性识别中的应用.北京石油化工学院学报,2008,16(3):43-46. ZHANG G Y,WANG N N,ZHANG R S,et al. Application of principal component analysis and BP network in identifying lithology. Journal of Beijing Institute of Petro-Chemical Technology, 2008,16(3):43-46.
[16] 温志平,方江雄,刘军,等.自适应递阶遗传神经网络测井岩性识别方法研究.东华理工大学学报(自然科学版),2017,40(4):368-375. WEN Z P,FANG J X,LIU J,et al. The method of logging lithology identification based on adaptive hierarchical genetic neural network. Journal of East China University of Technology(Natural Science),2017,40(4):368-375.
[17] 郭小波,黄志龙,涂小仙,等.马朗凹陷芦草沟组致密储集层复杂岩性识别.新疆石油地质,2013,34(6):649-652. GUO X B,HUANG Z L,TU X X,et al. Identification and application of complex lithology of Lucaogou tight reservoir in Malang Sag,Santanghu Basin. Xinjiang Petroleum Geology, 2013,34(6):649-652.
[18] 李新宁,马强,梁辉,等.三塘湖盆地二叠系芦草沟组二段混积岩致密油地质特征及勘探潜力.石油勘探与开发,2015,42(6):763-771. LI X N,MA Q,LIANG H,et al. Geological characteristics and exploration potential of diamictite tight oil the second member of the Permain Lucaogou Formation,Santanghu Basin,NW China. Petroleum Exploration and Development,2015,42(6):763-771.
[19] 柳波,吕延防,孟元林,等.湖相纹层状细粒岩特征、成因模式及其页岩油意义:以三塘湖盆地马朗凹陷二叠系芦草沟组为例.石油勘探与开发,2015,42(5):598-607. LIU B,LYU Y F,MENG Y L,et al. Petrologic characteristics and genetic model of lacustrine lamellar fine-grained rock and its significance for shale oil exploration:a case study of Permian Lucaogou Formation in Malang Sag,Santanghu Basin,NW China. Petroleum Exploration and Development,2015,42(5):598-607.
[20] 梁世君,黄志龙,柳波,等.马朗凹陷芦草沟组页岩油形成机理与富集条件.石油学报,2012,33(4):588-594. LIANG S J,HUANG Z L,LIU B,et al. Formation mechanism and enrichment conditions of Lucaogou Formation shale oil from Malang sag,Santanghu Basin. Acta Petrolei Sinica,2012, 33(4):588-594.
[21] 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.岩性油气藏,2019,31(1):1-11. LUO Q,WU A B,WANG J L,et al. Genetic types,generation models,and exploration direction of shale gas in northern China. Lithologic Reservoirs,2019,31(1):1-11.
[22] 闫林,冉启全,高阳,等.吉木萨尔凹陷芦草沟组致密油储层溶蚀孔隙特征及成因机理.岩性油气藏,2017,29(3):27-33. YAN L,RAN Q Q,GAO Y,et al. Characteristics and formation mechanism of dissolved pores in tight oil reservoirs of Lucaogou Formation in Jimsar Sag. Lithologic Reservoirs,2017,29(3):27-33.
[23] 靳军,王剑,杨召,等.准噶尔盆地克-百断裂带石炭系内幕储层测井岩性识别.岩性油气藏,2018,30(2):85-92. JIN J,WANG J,YANG Z,et al. Welling logging identification of Carboniferous volcanic inner buried-hill reservoirs in Ke-Bai fault zone in Junggar Basin. Lithologic Reservoirs,2018,30(2):85-92.
[24] 车世琦.测井资料用于页岩岩相划分及识别:以涪陵气田五峰组-龙马溪组为例.岩性油气藏,2018,30(1):121-132. CHE S Q. Shale lithofacies identification and classification by using logging data:a case of Wufeng-Longmaxi Formation in Fuling Gas Field,Sichuan Basin. Lithologic Reservoirs,2018, 30(1):121-132.
[25] 张涛,莫修文.基于交会图与模糊聚类算法的复杂岩性识别. 吉林大学学报(地球科学版),2007,37(增刊1):109-113. ZHANG T,MO X W. Complex lithologic identification based on cross plot and fuzzy clustering algorithm. Journal of Jilin University(Earth Science Edition),2007,37(Suppl 1):109113.
[26] 阳正熙,吴堑虹,彭直兴,等.地学数据分析教程.北京:科学教育出版社,2008:127-142. YANG Z X,WU Q H,PENG Z X,et al. Geoscience data analysis tutorial. Beijing:Science Education Press,2008:127-142.
[27] MA Y Z. Lithofacies clustering using principal component analysis and neural network:application to wireline logs. Mathematical Geosciences,2011,43:401-419.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

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

Application of BP neutral network method to identification of shale lithofacies of Lucaogou Formation in Santanghu 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]	孔令峰, 徐加放, 刘丁. 三塘湖盆地侏罗系西山窑组褐煤储层孔隙结构特征及脱水演化规律[J]. 岩性油气藏, 2024, 36(5): 15-24.
[2]	魏成林, 张凤奇, 江青春, 鲁雪松, 刘刚, 卫延召, 李树博, 蒋文龙. 准噶尔盆地阜康凹陷东部深层二叠系超压形成机制及演化特征[J]. 岩性油气藏, 2024, 36(5): 167-177.
[3]	徐田录, 吴承美, 张金凤, 曹爱琼, 张腾. 吉木萨尔凹陷二叠系芦草沟组页岩油储层天然裂缝特征与压裂模拟[J]. 岩性油气藏, 2024, 36(4): 35-43.
[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]	邓远, 陈轩, 覃建华, 李映艳, 何吉祥, 陶鑫, 尹太举, 高阳. 吉木萨尔凹陷二叠系芦草沟组一段沉积期古地貌特征及有利储层分布[J]. 岩性油气藏, 2024, 36(1): 136-144.
[6]	刘海磊, 朱永才, 刘龙松, 尹鹤, 王学勇, 杜小弟. 准噶尔盆地阜康断裂带上盘二叠系芦草沟组页岩油地质特征及勘探潜力[J]. 岩性油气藏, 2023, 35(4): 90-101.
[7]	柳忠泉, 赵乐强, 曾治平, 田继军, 李正强, 罗锦昌, 胡美玲. 准噶尔盆地阜康断裂带二叠系芦草沟组页岩油成藏条件[J]. 岩性油气藏, 2023, 35(3): 126-137.
[8]	张凤奇, 李宜浓, 罗菊兰, 任小锋, 张兰馨, 张芥瑜. 鄂尔多斯盆地西部奥陶系乌拉力克组页岩微观孔隙结构特征[J]. 岩性油气藏, 2022, 34(5): 50-62.
[9]	罗锦昌, 田继军, 马静辉, 闫嘉启, 梁雅菲, 胡卓浩. 吉木萨尔凹陷吉页1井区二叠系芦草沟组沉积环境及有机质富集机理[J]. 岩性油气藏, 2022, 34(5): 73-85.
[10]	张记刚, 杜猛, 陈超, 秦明, 贾宁洪, 吕伟峰, 丁振华, 向勇. 吉木萨尔凹陷二叠系芦草沟组页岩储层孔隙结构定量表征[J]. 岩性油气藏, 2022, 34(4): 89-102.
[11]	柳忠泉, 曾治平, 田继军, 李艳丽, 李万安, 张文文, 张治恒. 吉木萨尔凹陷二叠系芦草沟组“甜点”成因与分布预测[J]. 岩性油气藏, 2022, 34(3): 15-28.
[12]	肖威, 张兵, 姚永君, 王艳, 杨洪宇, 杨凯. 川东二叠系龙潭组页岩岩相特征与沉积环境[J]. 岩性油气藏, 2022, 34(2): 152-162.
[13]	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.
[14]	张文文, 韩长城, 田继军, 张治恒, 张楠, 李正强. 吉木萨尔凹陷二叠系芦草沟组层序地层划分及演化特征[J]. 岩性油气藏, 2021, 33(5): 45-58.
[15]	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.