Fine-grained identification of internal characteristic zones within strike-slip faults via the XGBoost algorithm
-
摘要:
受制于走滑断裂内部强烈的非均质性、储集类型多样及流体分布复杂的影响，走滑断裂内部裂缝带、破碎带和溶蚀带之间的测井响应复杂多变，为有效利用成像及常规测井资料识别走滑断裂内部3特征带造成了困难。引入XGboost算法建立模型，以提高对走滑断裂内部3特征带的识别精度。分析了走滑断裂内部3特征带的测井响应特征，优选敏感测井曲线构建基于均值及方差的特征向量空间集，采用极端梯度提升算法，建立了走滑断裂溶蚀带、破碎带和裂缝带的XGBoost回归预测模型，并通过多分类评价指标对XGBoost模型的关键参数进行调优，提高了走滑断裂内部特征带的识别精度。利用构建的XGBoost模型对研究区走滑断裂内部特征带进行了识别，其中总样本数234个，识别正确样本208个，识别正确率达88.89%；预测结果表明在走滑断裂内部特征带中，裂缝带分布范围最广，破碎带其次，溶蚀带最窄，这与实际走滑断裂内部特征带的分布范围相符。基于XGBoost算法的走滑断裂内部特征带识别模型能够有效地识别裂缝带、破碎带和溶蚀带，从而有助于对走滑断裂内部尺度更小的溶蚀孔洞及裂缝储集空间的分布进行更为有效的分析，对走滑断裂内部结构的精细刻画有一定借鉴意义。
Abstract:Because of the strong heterogeneity of fault zones, diverse reservoir types, and complex fluid distributions, the logging responses between damage, fault breccias , and dissolution zones within strike-slip faults are complex and variable, making it difficult to identify the three characteristic zones effectively inside strike-slip faults using imaging and conventional logging data.
Objective The extreme gradient boosting (XGBoost) algorithm is introduced to establish a model to improve the identification accuracy of the three characteristic zones within strike-slip faults.
Methods The logging response characteristics of the three characteristic zones within strike-slip faults are analyzed, and the sensitive logging curves are selected to construct a feature vector space set based on the mean and variance. The XGBoost algorithm is applied to establish XGBoost regression prediction models for the dissolution, breccias, and damage zones of strike-slip faults. The key parameters of the XGBoost model are optimized through multiclass evaluation indicators to improve the identification accuracy of the characteristic zones within strike-slip faults.
Results The constructed XGBoost model was used to identify the internal characteristic zones of strike-slip faults in the study area, with a total of 234 samples; 208 samples were correctly identified, resulting in an identification accuracy of 88.89%. The prediction results reveal that, within the internal characteristic zones of strike-slip faults, the damage zone has the widest distribution, followed by the breccias zone, and the dissolution zone is the narrowest, which is consistent with the actual distribution of the internal characteristic zones of strike-slip faults.
Conclusion The identification model of internal characteristic zones within strike-slip faults based on the XGBoost algorithm can be used to effectively identify the damage, breccias, and dissolution zones, thereby supporting more effective analysis of the distribution of small-scale dissolution cavities and fracture reservoir spaces inside strike-slip faults, and providing reference information for the accurate characterization of the internal structure of strike-slip faults.
-
图 1 川中高石梯−磨溪地区走滑断裂展布图
Figure 1. Distribution of strike-slip faults in the central part of Sichuan Gaoshiti-Moxi region
图 2 走滑断裂“三带”地质分布示意图
Figure 2. Geological distribution of the "Three Zones" of strike-slip fault
图 3 单井“三带”测井响应特征图
Figure 3. Logging response characteristics of the "Three Zones" in a single well
图 4 研究区溶蚀带、破碎带、裂缝带测井响应曲线特征
Figure 4. Logging response curve characteristics of the dissolution, fault breccias and fracture (damage) zones in the study area
图 6 模型参数分析变化趋势图
Wmin. 叶子点中最小样本权重和；S. 生成每棵树的随机样本采样的比例；γ. 最小损失函数下降值；下同
Figure 6. Trend chart of model parameter analysis changes
图 7 训练样本(a)与测试样本(b)的模型预测结果图
Figure 7. Model prediction results for training (a) and testing (b) samples
图 8 gs20(a)和gs019井(b)走滑断裂“三带”识别效果图
Figure 8. Identification results of the "Three Zones" of strike-slip faults in gs20 (a) and gs019 (b) wells
图 9 高石梯−磨溪地区走滑断裂FⅠ11 及FⅡ16附近区域“三带”分布图
Figure 9. Distribution of the "Three Zones" near strike-slip faults FⅠ11 and FⅡ16 in the Gaoshiti-Moxi region
表 1 样本数据选取表
Table 1. Sample data selection table
“三带”类型 类型代码 训练数据 预测数据 溶蚀带 1 16 5 破碎带 2 269 67 裂缝带 3 210 52 合计/组 619 495 124 整体百分比/% 79.97 20.03 
下载: 导出CSV
表 2 主要参数最优取值
Table 2. Optimal values of key parameters
参数名 最优取值 迭代次数t 700 学习率eta 0.1 最大树深度Dmax 6 叶子点中最小样本权重和Wmin 5 生成每棵树的随机样本采样的比例S 0.8 最小损失函数下降值γ 0.5
下载: 导出CSV
-
[1] 梁瀚,唐浩,冉崎,等. 四川盆地川中地区走滑断裂的分布、类型与成因[J]. 地质学报,2023,97(8):2609-2620.LIANG H,TANG H,RAN Q,et al. The distribution,type and origin of the strike-slip faults in the central Sichuan Basin[J]. Acta Geologica Sinica,2023,97(8):2609-2620.(in Chinese with English abstract [2] 管树巍,姜华,鲁雪松,等. 四川盆地中部走滑断裂系统及其控油气作用[J]. 石油学报,2022,43(11):1542-1557.GUAN S W,JIANG H,LU X S,et al. Strike-slip fault system and its control on oil & gas accumulation in central Sichuan Basin[J]. Acta Petrolei Sinica,2022,43(11):1542-1557.(in Chinese with English abstract [3] 马兵山,梁瀚,邬光辉,等. 四川盆地中部地区多期次走滑断层的形成及演化[J]. 石油勘探与开发,2023,50(2):333-345.MA B S,LIANG H,WU G H,et al. Formation and evolution of the strike-slip faults in the central Sichuan Basin,SW China[J]. Petroleum Exploration and Development,2023,50(2):333-345.(in Chinese with English abstract [4] WANG Y J,ZHANG Y Z,CAWOOD P A,et al. Early Neoproterozoic assembly and subsequent rifting in South China:Revealed from mafic and ultramafic rocks,central Jiangnan Orogen[J]. Precambrian Research,2019,331:105367. doi: 10.1016/j.precamres.2019.105367 [5] YAN C L,SHU L S,FAURE M,et al. Time constraints on the closure of the Paleo-South China Ocean and the Neoproterozoic assembly of the Yangtze and Cathaysia blocks:Insight from new detrital zircon analyses[J]. Gondwana Research,2019,73:175-189. doi: 10.1016/j.gr.2019.03.018 [6] LI S Z,ZHAO S J,LIU X,et al. Closure of the Proto-Tethys Ocean and Early Paleozoic amalgamation of microcontinental blocks in East Asia[J]. Earth-Science Reviews,2018,186:37-75. doi: 10.1016/j.earscirev.2017.01.011 [7] 李献华,李武显,何斌. 华南陆块的形成与Rodinia超大陆聚合−裂解:观察、解释与检验[J]. 矿物岩石地球化学通报,2012,31(6):543-559.LI X H,LI W X,HE B. Building of the South China Block and its relevance to assembly and breakup of Rodinia Supercontinent:Observations,interpretations and tests[J]. Bulletin of Mineralogy,Petrology and Geochemistry,2012,31(6):543-559.(in Chinese with English abstract [8] LIU Z,TAN S C,HE X H,et al. Petrogenesis of mid-Neoproterozoic (ca. 750 Ma) mafic and felsic intrusions in the Ailao Shan-Red River belt:Geochemical constraints on the paleogeographic position of the South China Block[J]. Lithosphere,2019,11(3):348-364. doi: 10.1130/L1038.1 [9] CAWOOD P A,WANG Y J,XU Y J,et al. Locating South China in Rodinia and Gondwana:A fragment of greater India lithosphere?[J]. Geology,2013,41(8):903-906. doi: 10.1130/G34395.1 [10] CAWOOD P A,STRACHAN R A,PISAREVSKY S A,et al. Linking collisional and accretionary orogens during Rodinia assembly and breakup:Implications for models of supercontinent cycles[J]. Earth and Planetary Science Letters,2016,449:118-126. doi: 10.1016/j.jpgl.2016.05.049 [11] KONOPáSEK J,JANOUŠEK V,OYHANTçABAL P,et al. Did the circum-Rodinia subduction trigger the Neoproterozoic rifting along the Congo-Kalahari Craton margin?[J]. International Journal of Earth Sciences,2018,107(5):1859-1894. doi: 10.1007/s00531-017-1576-4 [12] 段金宝,梅庆华,李毕松,等. 四川盆地震旦纪−早寒武世构造−沉积演化过程[J]. 地球科学,2019,44(3):738-755.DUAN J B,MEI Q H,LI B S,et al. Sinian-Early Cambrian tectonic-sedimentary evolution in Sichuan Basin[J]. Earth Science,2019,44(3):738-755.(in Chinese with English abstract [13] BIAN Q,DENG S,LIN H X,et al. Strike-slip salt tectonics in the Shuntuoguole Low Uplift,Tarim Basin,and the significance to petroleum exploration[J]. Marine and Petroleum Geology,2022,139:105600. doi: 10.1016/j.marpetgeo.2022.105600 [14] ZHU M,LIANG Z L,WANG X,et al. Mesozoic strike-slip fault system at the margin of the Junggar Basin,NW China[J]. Journal of Structural Geology,2023,175:104950. doi: 10.1016/j.jsg.2023.104950 [15] LIANG S Z,LI Z,ZHANG W,et al. The characteristics of strike-slip faults and their control on hydrocarbon distribution in deep carbonate reservoirs of the central Sichuan Basin[J]. Frontiers in Earth Science,2023,11:1064835. doi: 10.3389/feart.2023.1064835 [16] CHENG X Y,DING W L,PAN L,et al. Geometry and kinematics characteristics of strike-slip fault zone in complex structure area:A case study from the south No. 15 strike-slip fault zone in the eastern Sichuan Basin,China[J]. Frontiers in Earth Science,2022,10:922664. doi: 10.3389/feart.2022.922664 [17] 付小东,张本健,汪泽成,等. 四川盆地中西部走滑断裂及其对油气成藏控制作用[J]. 地球科学,2023,48(6):2221-2237.FU X D,ZHANG B J,WANG Z C,et al. Strike-slip faults in central and western Sichuan Basin and their control functions on hydrocarbon accumulation[J]. Earth Science,2023,48(6):2221-2237.(in Chinese with English abstract [18] 马庆佑,沙旭光,李玉兰,等. 塔中顺托果勒区块走滑断裂特征及控油作用[J]. 石油实验地质,2012,34(2):120-124.MA Q Y,SHA X G,LI Y L,et al. Characteristics of strike-slip fault and its controlling on oil in Shuntuoguole region,middle Tarim Basin[J]. Petroleum Geology & Experiment,2012,34(2):120-124.(in Chinese with English abstract [19] 吴孔友,瞿建华,王鹤华. 准噶尔盆地大侏罗沟断层走滑特征、形成机制及控藏作用[J]. 中国石油大学学报(自然科学版),2014,38(5):41-47.WU K Y,QU J H,WANG H H. Strike-slip characteristics,forming mechanisms and controlling reservoirs of Dazhuluogou fault in Junggar Basin[J]. Journal of China University of Petroleum (Edition of Natural Science),2014,38(5):41-47.(in Chinese with English abstract [20] MA D B,WANG Z C,DUAN S F,et al. Strike-slip faults and their significance for hydrocarbon accumulation in Gaoshiti-Moxi area,Sichuan Basin,SW China[J]. Petroleum Exploration and Development,2018,45(5):851-861. doi: 10.1016/S1876-3804(18)30088-0 [21] 焦方正. 塔里木盆地顺托果勒地区北东向走滑断裂带的油气勘探意义[J]. 石油与天然气地质,2017,38(5):831-839.JIAO F Z. Significance of oil and gas exploration in NE strike-slip fault belts in Shuntuoguole area of Tarim Basin[J]. Oil & Gas Geology,2017,38(5):831-839.(in Chinese with English abstract [22] 许志琴,王勤,李忠海,等. 印度-亚洲碰撞:从挤压到走滑的构造转换[J]. 地质学报,2016,90(1):1-23. doi: 10.1111/1755-6724.12639XU Z Q,WANG Q,LI Z H,et al. Indo-Asian collision:Tectonic transition from compression to strike slip[J]. Acta Geologica Sinica,2016,90(1):1-23.(in Chinese with English abstract doi: 10.1111/1755-6724.12639 [23] 邬光辉,成丽芳,刘玉魁,等. 塔里木盆地寒武−奥陶系走滑断裂系统特征及其控油作用[J]. 新疆石油地质,2011,32(3):239-243.WU G H,CHENG L F,LIU Y K,et al. Strike-slip fault system of the Cambrian-Ordovician and its oil-controlling effect in Tarim Basin[J]. Xinjiang Petroleum Geology,2011,32(3):239-243.(in Chinese with English abstract [24] DENG S,LI H L,ZHANG Z P,et al. Structural characterization of intracratonic strike-slip faults in the central Tarim Basin[J]. AAPG Bulletin,2019,103(1):109-137. doi: 10.1306/06071817354 [25] WU G H,MA B S,HAN J F,et al. Origin and growth mechanisms of strike-slip faults in the central Tarim Cratonic Basin,NW China[J]. Petroleum Exploration and Development,2021,48(3):595-607. doi: 10.1016/S1876-3804(21)60048-4 [26] 卢志强,马乃拜,尼加提,等. 顺北油田深层断溶体内部结构物探表征研究[J]. 内蒙古石油化工,2022,48(6):115-120.LU Z Q,MA N B,NI J T,et al. Research and application of fault-karst characterization technology in Shunbei oilfield[J]. Inner Mongolia Petrochemical Industry,2022,48(6):115-120.(in Chinese with English abstract [27] 吴丰,罗莹莹,李昱翰,等. 四川盆地公山庙油田大安寨段湖相灰岩−页岩裂缝特征与测井识别[J]. 地质科技通报,2022,41(5):55-67.WU F,LUO Y Y,LI Y H,et al. Fracture characteristics and logging identification of lacustrine limestone-shale reservoirs in Da'anzhai Member,Gongshanmiao oilfield,Sichuan Basin[J]. Bulletin of Geological Science and Technology,2022,41(5):55-67.(in Chinese with English abstract [28] 王震,文欢,邓光校,等. 塔河油田碳酸盐岩断溶体刻画技术研究与应用[J]. 石油物探,2019,58(1):149-154.WANG Z,WEN H,DENG G X,et al. Fault-Karst characterization technology in the Tahe oilfield,China[J]. Geophysical Prospecting for Petroleum,2019,58(1):149-154.(in Chinese with English abstract [29] CHEN T Q,GUESTRIN C. XGBoost:A scalable tree boosting system[J]. ArXiv e-Prints,2016:arXiv:1603.02754. [30] JOHNSON R,ZHANG T. Learning nonlinear functions using regularized greedy forest[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2014,36(5):942-954. doi: 10.1109/TPAMI.2013.159 [31] FRIEDMAN J H. Stochastic gradient boosting[J]. Computational Statistics & Data Analysis,2002,38(4):367-378. [32] 闫星宇,顾汉明,肖逸飞,等. XGBoost算法在致密砂岩气储层测井解释中的应用[J]. 石油地球物理勘探,2019,54(2):447-455.YAN X Y,GU H M,XIAO Y F,et al. XGBoost algorithm applied in the interpretation of tight-sand gas reservoir on well logging data[J]. Oil Geophysical Prospecting,2019,54(2):447-455.(in Chinese with English abstract [33] 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[J]. 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[J]. Lithologic Reservoirs,2020,32(4):98-106.(in Chinese with English abstract [34] 丁阳阳,赵军龙,李兆明,等. 基于XGBoost算法的煤体结构测井识别技术研究[J]. 地球物理学进展,2022,37(3):998-1006.DING Y Y,ZHAO J L,LI Z M,et al. Research on logging recognition technology of coal structure based on XGBoost algorithm[J]. Progress in Geophysics,2022,37(3):998-1006.(in Chinese with English abstract [35] 谷宇峰,张道勇,鲍志东. 利用混合模型CRBM-PSO-XGBoost识别致密砂岩储层岩性[J]. 石油与天然气地质,2021,42(5):1210-1222.GU Y F,ZHANG D Y,BAO Z D. Lithology identification in tight sandstone reservoirs using CRBM-PSO-XGBoost[J]. Oil & Gas Geology,2021,42(5):1210-1222.(in Chinese with English abstract [36] 赵军,王伟明,杨志冬,等. 基于DBN模型的火山岩岩性识别方法:以车排子地区为例[J]. 地球物理学进展,2022,37(1):328-337. doi: 10.6038/pg2022EE0360ZHAO J,WANG W M,YANG Z D,et al. Lithology identification method of volcanic rocks based on DBN model:An case in Chepaizi area[J]. Progress in Geophysics,2022,37(1):328-337.(in Chinese with English abstract doi: 10.6038/pg2022EE0360 [37] 汪泽成,姜华,王铜山,等. 四川盆地桐湾期古地貌特征及成藏意义[J]. 石油勘探与开发,2014,41(3):305-312. doi: 10.11698/PED.2014.03.05WANG Z C,JIANG H,WANG T S,et al. Paleo-geomorphology formed during Tongwan tectonization in Sichuan Basin and its significance for hydrocarbon accumulation[J]. Petroleum Exploration and Development,2014,41(3):305-312.(in Chinese with English abstract doi: 10.11698/PED.2014.03.05 [38] 魏国齐,杨威,杜金虎,等. 四川盆地高石梯—磨溪古隆起构造特征及对特大型气田形成的控制作用[J]. 石油勘探与开发,2015,42(3):257-265.WEI G Q,YANG W,DU J H,et al. Tectonic features of Gaoshiti-Moxi paleo-uplift and its controls on the formation of a giant gas field,Sichuan Basin,SW China[J]. Petroleum Exploration and Development,2015,42(3):257-265.(in Chinese with English abstract [39] 吴丰,代槿,姚聪,等. 塔河油田奥陶系一间房组与鹰山组断溶体发育模式解剖[J]. 断块油气田,2022,29(1):33-39.WU F,DAI J,YAO C,et al. Developmental mode analysis of the fault-karst reservoir in Yijianfang Formation and Yingshan Formation of Ordovician in Tahe oilfield[J]. Fault-Block Oil & Gas Field,2022,29(1):33-39.(in Chinese with English abstract [40] 李梁,孙庭斌,陈迎楠,等. 塔里木盆地肖塘南地区走滑断裂特征与断溶体勘探研究[J]. 地质科技通报,2023,42(1):239-245.LI L,SUN T B,CHEN Y N,et al. Characteristics of strike-slip faults and exploration of fault-dissolution body in Xiaotangnan area of Tarim Basin[J]. Bulletin of Geological Science and Technology,2023,42(1):239-245.(in Chinese with English abstract [41] BREIMAN L. Random forests[J]. Machine Learning,2001,45(1):5-32. doi: 10.1023/A:1010933404324 [42] FRIEDMAN J,HASTIE T,TIBSHIRANI R. Additive logistic regression:A statistical view of boosting (With discussion and a rejoinder by the authors)[J]. The Annals of Statistics,2000,28(2):337-374. [43] FRIEDMAN J H. Greedy function approximation:A gradient boosting machine[J]. The Annals of Statistics,2001,29(5):1189-1232. doi: 10.1214/aos/1013203450 [44] LIU C,WANG W Y,WANG M,et al. An efficient instance selection algorithm to reconstruct training set for support vector machine[J]. Knowledge-Based Systems,2017,116:58-73. doi: 10.1016/j.knosys.2016.10.031 [45] 尹帅,田涛,李俊鹿,等. 鄂尔多斯盆地西南缘延长组断缝体特征及对油藏的调控作用[J]. 油气地质与采收率,2024,31(1):1-12.YIN S,TIAN T,LI J L,et al. Fault-fracture body characteristics and their effect on hydrocarbon distribution of Yanchang Formation in southwestern margin of Ordos Basin[J]. Petroleum Geology and Recovery Efficiency,2024,31(1):1-12.(in Chinese with English abstract [46] 刘粤蛟,赖富强,徐浩,等. 基于测井曲线深程度耦合的页岩岩相智能识别方法[J]. 地质科技通报,2025,44(1):308-320.LIU Y J,LAI F Q,XU H,et al. Intelligent identification methods for shale lithology based on the coupling deeply of logging curves[J]. Bulletin of Geological Science and Technology,2025,44(1):308-320.(in Chinese with English abstract [47] 张驰,潘懋,胡水清,等. 融合储层纵向信息的机器学习岩性识别方法[J]. 地质科技通报,2023,42(3):289-299.ZHANG C,PAN M,HU S Q,et al. A machine learning lithologic identification method combined with vertical reservoir information[J]. Bulletin of Geological Science and Technology,2023,42(3):289-299. (in Chinese with English abstract [48] 田明智,朱超,李森明,等. 湖相碳酸盐岩测井岩相识别技术与应用——以柴达木盆地英西地区为例[J]. 中国石油勘探,2023,28(1):135-143.TIAN M Z,ZHU C,LI S M,et al. Application of logging lithofacies identification technology of lacustrine carbonate rocks: A case study of Yingxi area, Qaidam Basin[J]. China Petroleum Exploration,2023,28(1):135-143. (in Chinese with English abstract [49] 高辉,吴振坤,柯雨,等. 基于自然图像的钻探岩心识别[J]. 煤田地质与勘探,2023,51(9):64-71.GAO H,WU Z K,KE Y,et al. Drilling core identification based on natural image[J]. Coal Geology & Exploration,2023,51(9):64-71. (in Chinese with English abstract -
下载:
点击查看大图
计量
- 文章访问数: 25
- PDF下载量: 0
- 被引次数: 0
