Intelligent Fluid Identification Based on the AdaBoost Machine Learning Algorithm for Reservoirs in Daniudi Gas Field
-
摘要: 大牛地气田储层复杂，矿物组分多样、储集空间复杂、非均质性强，导致流体识别困难。为提高该气田复杂储层流体识别的准确率和解释效率，以广泛发育的低阻气藏为主要研究对象，采用Adaboost机器学习算法，分别以逻辑分类、决策树等主流智能算法作为弱分类器，集成了4类强分类器模型。基于低阻气藏成因机理分析优化了模型输入参数，基于常规测井和试油、试采资料进行了参数优选，并将上述模型应用到6口实际井资料中。结果显示，其中以决策树为弱分类器集成的强分类器取得了最佳识别效果，流体识别准确率达到86.5%，F1得分达到86.6%。研究结果表明，该方法可作为低阻气藏常规测井资料识别流体的有效手段，为流体评价提供了新思路。
-
关键词:
- 复杂储层 /
- 流体识别 /
- 机器学习 /
- 智能识别；大牛地气田
Abstract: Complex reservoirs in Daniudi Gas Field are characterized by diverse mineral components, complex reservoir space, and strong heterogeneity, which make fluid identification difficult. To improve the accuracy rate and interpretation efficiency of fluid identification in complex reservoirs, Daniudi Gas Field, with its extensively developed low-resistance gas reservoirs, was taken as the main research object. Then, four strong classifier models were formed by the Adaboost machine learning algorithm with mainstream intelligent algorithms (such as logical classification and decision tree) as weak classifiers. The input parameters of the model were optimized based on the analysis of the genesis mechanism of the low-resistance gas reservoir, the parameters were optimized on the basis of conventional well logging, oil testing and production testing data, etc. The above model was applied to the data of 6 actual wells. The results showed that the strong classifier achieved the best identification effect by using the decision tree algorithm as the weak classifier, with the fluid identification accuracy of 86.5% and the F1 value up to 86.6%. The results indicates that this method is effective for identifying fluid with conventional logging data for low-resistance gas reservoirs, and providing new ideas for fluid evaluation. -
-
![]()
图 2 大牛地气田上古生界流体识别交会图
Figure 2. Cross plot for fluid identification of the Upper Paleozoic in Daniudi Gas Field
![]()
图 3 大牛地气田上古生界束缚水饱和度和孔隙度交会图
Figure 3. Cross plot of irreducible water saturation and porosity of the Upper Paleozoic in Daniudi Gas Field
![]()
图 5 最优智能模型流体识别效果
Figure 5. Fluid identification results from the optimal intelligent model
表 1 大牛地气田上古生界典型流体参数
Table 1 Typical fluid parameters of the Upper Paleozoic in Daniudi Gas Field
井名 层位 井段/m 孔隙度, % 电阻率/（Ω·m） 电阻增大率 产水量/m3 无阻流量/
（104m3·d–1）解释结论 X1 盒3 2219.6～2226.0 10.79 14.40 0.96 6.20 0 水层 X2 盒3 2605.4～2617.0 9.72 18.50 1.23 0.50 0 水层 X3 盒1 2526.1～2534.0 11.43 30.40 2.02 0 2.70 低阻气层 盒1 2535.5～2541.8 10.39 28.10 1.87 山2 2543.9～2556.1 11.75 26.80 1.79 X4 山2 2741.8～2752.0 11.12 37.50 2.53 0 2.06 低阻气层 X5 盒3 2697.1～2703.6 8.58 324.30 21.62 0 16.51 中高阻气层 X6 山2 2462.0～2478.4 6.69 66.89 4.46 0 0.96 中高阻气层 X7 山2 2486.0～2497.0 8.50 80.49 5.37 X8 山1 2852.1～2858.9 9.64 17.92 1.19 6.34 2.50 含气水层 X9 盒3 2472.3～2479.3 11.92 16.33 1.09 3.72 微 含气水层 
下载: 导出CSV
表 2 4类储层的常规测井响应值分布
Table 2 Distribution of conventional log response eigenvalues of four types of reservoirs
储层类型 测井响应值 自然伽马/API 中子，% 密度/（g·cm–3） 声波时差（μs·m–1） 深电阻率/（Ω·m） 孔隙度，% 束缚水饱和度，% 中高阻气层 范围
均值40.7～78.6
58.915.1～17.3
16.42.30～2.48
2.41197.8～242.2
218.741.8～445.1
112.78.4～17.4
15.326.9～43.4
29.3低阻气层 范围
均值44.2～73.8
63.48.4～14.2
11.82.33～2.53
2.44214.2～263.8
242.516.7～46.1
29.16.7～12.4
10.133.7.9～50.4
38.5含气水层 范围
均值44.6～74.3
62.311.7～16.1
13.92.42～2.49
2.45222.2～249.7
236.510.7～31.8
21.16.3～15.4
11.329.2～52.4
35.8水层 范围
均值46.3～84.5
68.66.3～13.4
9.92.40～2.54
2.46205.2～262.8
238.810.5～28.2
15.85.2～14.6
9.230.25～59.5
40.9
下载: 导出CSV
表 3 4个监督模型的重要参数和最优参数值
Table 3 Important parameters and their optimal values of four supervision models
机器学习模型 优化参数 搜索范围 最优参数 逻辑回归（LR） 正则化策略
惩罚参数Ｃ11/12
0.1~1012
1.693决策树（DT） 树的最大深度
特征选择准则基尼系数／信息熵
0~10信息熵
3支持向量机（SVM） 核函数的γ参数
惩罚参数Ｃ0.1~10
0.1~1002.015
15神经网络（NN） 最大层数
最大隐层数量3~8
10~2004
100
下载: 导出CSV
表 4 不同模型的流体识别结果
Table 4 Fluid identification results from different models
试油试采结果 总层数/个 AB-LR模型 AB-DT模型 AB-SVC模型 AB-NN模型 正确样本/个 符合率, % 正确样本/个 符合率，% 正确样本/个 符合率，% 正确样本/个 符合率，% 中高阻气层 23 23 100 23 100 23 100 23 100 低阻气层 38 26 68.4 31 81.6 28 73.7 23 78.9 含气水层 28 17 60.7 23 82.1 21 75.0 30 78.6 水层 30 23 76.7 26 86.7 24 80.0 22 76.7
下载: 导出CSV
-
[1] 张海涛,郭笑锴,杨小明,等. 姬塬地区低对比度油层成因机理与流体识别方法[J]. 测井技术,2019,43(5):542–549. ZHANG Haitao, GUO Xiaokai, YANG Xiaoming, et al. Genesis mechanism and fluid identification of low contrast reservoirs in Jiyuan area[J]. Well Logging Technology, 2019, 43(5): 542–549.
[2] 张峰,罗少成,李震,等. 四川盆地茅口组岩溶缝洞型储层有效性测井评价[J]. 石油钻探技术,2020,48(6):116–122. doi: 10.11911/syztjs.2020140 ZHANG Feng, LUO Shaocheng, LI Zhen, et al. Logging evaluation on the effectiveness of karst fractured-vuggy reservoirs in the Maokou Formation, Sichuan Basin[J]. Petroleum Drilling Techni-ques, 2020, 48(6): 116–122. doi: 10.11911/syztjs.2020140
[3] 张丛秀,郝晋美,刘治恒,等. 基于测录井资料的环西—彭阳地区延安组储层流体性质识别方法研究[J]. 石油钻探技术,2020,48(5):111–119. doi: 10.11911/syztjs.2020079 ZHANG Congxiu, HAO Jinmei, LIU Zhiheng, et al. A study on the logging-based identification method for reservoir fluid properties of the Yan’an Formation in the Huanxi-Pengyang[J]. Petroleum Dril-ling Techniques, 2020, 48(5): 111–119. doi: 10.11911/syztjs.2020079
[4] 陈四平,谭判,石文睿,等. 涪陵页岩气优质储层测井综合评价方法[J]. 石油钻探技术,2020,48(4):131–138. doi: 10.11911/syztjs.2020091 CHEN Siping, TAN Pan, SHI Wenrui, et al. A comprehensive logging evaluation method for high quality shale gas reservoirs in Fuling[J]. Petroleum Drilling Techniques, 2020, 48(4): 131–138. doi: 10.11911/syztjs.2020091
[5] 李义,周全,张伟. 陆丰凹陷文昌组储层流体性质识别方法研究[J]. 海洋石油,2020,40(1):70–73. LI Yi, ZHOU Quan, ZHANG Wei. Study on identification method of reservoir fluid properties of Wenchang Formation in Lufeng Sunken[J]. Offshore Oil, 2020, 40(1): 70–73.
[6] 张艺,李道清,仇鹏,等. 基于岩性分类的火山岩储层流体识别方法:以克拉美丽气田石炭系火山岩为例[J]. 西安石油大学学报(自然科学版),2020,35(6):22–29. ZHANG Yi, LI Daoqing, QIU Peng, et al. Study on fluid identification method of volcanic reservoir based on lithology classification: a case study of carboniferous volcanic rocks in Kelamei Gasfield[J]. Journal of Xi’an Shiyou University(Natural Science), 2020, 35(6): 22–29.
[7] 王月莲,袁士义,宋新民,等. “无侵线法” 流体识别技术在低渗低电阻率油藏中的应用[J]. 石油勘探与开发,2005,32(3):88–90. doi: 10.3321/j.issn:1000-0747.2005.03.022 WANG Yuelian, YUAN Shiyi, SONG Xinmin, et al. Non-intrusion line method for fluid identification and its application in low permeability and low resistivity reservoirs[J]. Petroleum Exploration & Development, 2005,32(3): 88–90. doi: 10.3321/j.issn:1000-0747.2005.03.022
[8] 孙卫涛,熊繁升,曹宏,等. 致密储层复杂流体模型及其适用性分析[J]. 石油物探,2021,60(1):136–148. SUN Weitao, XIONG Fansheng, CAO Hong, et al. Analysis of complex fluid model and its applicability in tight reservoirs[J]. Geophysical Prospecting for Petroleum, 2021, 60(1): 136–148.
[9] WU P Y, JAIN V, KULKARNIET M S, et al. Machine learning-based method for automated well-log processing and interpretation[R]. SEG-2018-2996973, 2018.
[10] 程超, 李培彦, 陈雁, 等. 基于机器学习的储层测井评价研究进展[J/OL]. 地球物理学进展: 1-15. (2021-05-31)[2021-09-06]. http://kns.cnki.net/kcms/detail/11.2982.P.20210529.1525.014.html. CHENG Chao, LI Peiyan, CHEN Yan, et al. Research progress of reservoir logging evaluation based on machine learning[J/OL]. Progress in Geophysics: 1-15. (2021-05-31)[2021-09-06]. http://kns.cnki.net/kcms/detail/11.2982.P.20210529.1525.014.html.
[11] 孙挺,赵颖,杨进,等. 基于支持向量机的钻井工况实时智能识别方法[J]. 石油钻探技术,2019,47(5):28–33. doi: 10.11911/syztjs.2019033 SUN Ting, ZHAO Ying, YANG Jin, et al. Real-time intelligent identi-fication method under drilling conditions based on support vector machine[J]. Petroleum Drilling Techniques, 2019, 47(5): 28–33. doi: 10.11911/syztjs.2019033
[12] 周雪晴,张占松,朱林奇,等. 基于双向长短期记忆网络的流体高精度识别新方法[J]. 中国石油大学学报(自然科学版),2021,45(1):69–76. ZHOU Xueqing, ZHANG Zhansong, ZHU Linqi, et al. A new method for high-precision fluid identification in bidirectional long short-term memory network[J]. Journal of China University of Petro-leum(Edition of Natural Science), 2021, 45(1): 69–76.
[13] 张银德,童凯军,郑军,等. 支持向量机方法在低阻油层流体识别中的应用[J]. 石油物探,2008,47(3):306–310, 314. doi: 10.3969/j.issn.1000-1441.2008.03.017 ZHANG Yinde, TONG Kaijun, ZHENG Jun, et al. Application of support vector machine method for identifying fluid in low-resistivity oil layers[J]. Geophysical Prospecting for Petroleum, 2008, 47(3): 306–310, 314. doi: 10.3969/j.issn.1000-1441.2008.03.017
[14] ONWUCHEKWA C. Application of machine learning ideas to reservoir fluid properties estimation[R]. SPE 193461, 2018.
[15] 王少龙,杨斌,赵倩,等. BP神经网络在复杂储层流体识别中的应用[J]. 石油化工应用,2018,37(7):45–48. doi: 10.3969/j.issn.1673-5285.2018.07.011 WANG Shaolong, YANG Bin, ZHAO Qian, et al. Application of BP neural network in recognition of complex reservoir fluids[J]. Petrochemical Industry Application, 2018, 37(7): 45–48. doi: 10.3969/j.issn.1673-5285.2018.07.011
[16] 周凡, 姜洪福, 王立艳, 等. 基于阵列感应测井的支持向量机流体识别方法[J]. 中国海洋大学学报（自然科学版）, 2011, 41（增刊1）: 317–323. ZHOU Fan, JIANG Hongfu, WANG Liyan, et al. Application of array induction logging and support vector machine to fluid identification[J]. Periodical of Ocean University of China, 2011, 41(supplement1): 317–323.
[17] 谭茂金, 白洋, 王谦, 等. 当非常规油气遇到人工智能: 多源数据驱动下非常规油气测井智能解释方法研究进展[C]//2019年油气地球物理学术年会论文集, 南京: 中国地球物理学会油气地球物理专业委员会, 2019: 615. TAN Maojin, BAI Yang, WANG Qian, et al. When unconventional oil and gas encounter artificial intelligence-the research progress of intelligent interpretation methods for unconventional oil and gas logging driven by multi-source data[C]//Proceedings of 2019 Annual Academic Conference on Oil and Gas Geophysics, Nanjing: Professional Committee of Oil and Gas Geophysics of China Geophysical Society, 2019: 615
[18] 张晓龙,任芳. 支持向量机与AdaBoost的结合算法研究[J]. 计算机应用研究,2009,26(1):77–78. ZHANG Xiaolong, REN Fang. Study on combinability of SVM and AdaBoost algorithm[J]. Applicaion Research of Computers, 2009, 26(1): 77–78.
[19] SCHAPIRE R E. The boosting approach to machine learning: an overview[M]//DENISON D D, HANSEN M H, HOLMES C C, et al. Nonlinear estimation and classification. New York: Springer, 2002: 149-171.
[20] 曹莹,苗启广,刘家辰,等. AdaBoost算法研究进展与展望[J]. 自动化学报,2013,39(6):745–758. CAO Ying, MIAO Qiguang, LIU Jiachen, et al. Advance and prospects of AdaBoost algorithm[J]. Acta Automatica Sinica, 2013, 39(6): 745–758.
[21] 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 Zhizhang, ZHOU Ziyong, et al. Lithology classification of acidic volcanic rocks based on parameter-optimized AdaBoost algorithm[J]. Acta Petrolei Sinica, 2019, 40(4): 457–467.
[22] 张郁哲,程时清,史文洋,等. 多层合采井产量劈分方法及在大牛地气田的应用[J]. 石油钻采工艺,2019,41(5):624–629. ZHANG Yuzhe, CHENG Shiqing, SHI Wenyang, et al. Commingled producing well production split method and its application in Daniudi Gasfield[J]. Oil Drilling & Production Technology, 2019, 41(5): 624–629.
[23] 郭振华,赵彦超. 大牛地气田致密砂岩气藏低阻气层成因分析[J]. 石油天然气学报,2007,29(3):246–249, 512-513. doi: 10.3969/j.issn.1000-9752.2007.03.027 GUO Zhenhua, ZHAO Yanchao. Genetic analysis on low-resistivity gas zones of tight sandstone reservoirs in Daniudi Gas Field[J]. Journal of Oil and Gas Technology, 2007, 29(3): 246–249, 512-513. doi: 10.3969/j.issn.1000-9752.2007.03.027
-
期刊类型引用(6)
1. 李朋，张艳玉，孙晓飞，李冬冬，刘洋，陈会娟. SAGD循环预热割缝筛管参数影响规律研究. 工程热物理学报. 2020(04): 940-947 . 
百度学术
2. 薛衡，黄祖熹，王贺华，安永生，刘榧，成一，何冰，刘卡. Ahdeb油田水平井控水完井及一体化耦合模型. 石油与天然气地质. 2019(02): 423-429 . 百度学术
3. 李朋，张艳玉，孙晓飞，陈会娟，刘洋. 稠油油藏双管水平井注汽井筒参数预测新模型. 特种油气藏. 2019(04): 85-90 . 百度学术
4. 李朋，张艳玉，孙晓飞，刘洋，谢孟珂，王朝，陈会娟. SAGD循环预热注汽参数影响规律数值模拟. 中南大学学报(自然科学版). 2019(11): 2896-2905 . 百度学术
5. 陈会娟，李明忠，狄勤丰，刘春苗. 多点注汽水平井井筒出流规律数值模拟. 石油学报. 2017(06): 696-704 . 百度学术
6. 徐磊. 稠油热采井防砂筛管热稳定性优化仿真分析. 北京石油化工学院学报. 2016(02): 40-45 . 百度学术
其他类型引用(5)
计量
- 文章访问数: 356
- HTML全文浏览量: 218
- PDF下载量: 79
- 被引次数: 11
