基于集成学习的松辽盆地砂岩型铀矿地层岩性自动识别研究
Automatic Lithology Identification of Sandstone-type Uranium Deposit in Songliao Basin Based on Ensemble Learning
-
摘要: 地层岩性的准确识别与砂岩型铀矿层的圈定密切相关，岩性组合的正确分析对于开展砂岩型铀矿的勘查与异常识别具有重要意义。本文针对传统测井岩性识别方法与机器学习类方法中存在的问题，以北方松辽盆地砂岩型铀矿为研究对象，采用两种典型的集成算法模型（XGBoost和SMOTE随机森林）开展地层岩性自动识别研究，并将识别结果与K最近邻分类算法（KNN）、梯度提升决策树算法（GBDT）等典型机器学习算法进行对比。结果表明，XGBoost和SMOTE随机森林两种集成算法模型对砂岩型铀矿地层岩性识别的准确率都在95%以上，且较KNN模型和GBDT模型的准确率有明显提高。XGBoost模型用于控制过拟合的正则项和节点分裂时支持特征多线程进行增益的计算，显著提高了运算效率，SMOTE合成少数过采样技术解决了样本数据不平衡的问题。基于集成算法的优化过程可为砂岩型铀矿岩性分类问题提供理论依据与技术支撑。Abstract: The accurate identification of stratigraphic lithology is closely related to the delineation of sandstone-type uranium deposits. In the face of complex stratigraphic structure, the correct analysis of lithology combination is of great significance to the exploration and anomaly identification of sandstone-type uranium deposits. In uranium exploration, geophysical logging data, as a bridge between the change of geophysical properties and the underground geological environment, is an effective and irreplaceable method to understand the underground rock structure and reservoir characteristics. Conventional lithology identification methods such as crossplot method, probability statistic method, cluster analysis method and conventional machine learning class method have some defects, such as low accuracy, identification efficiency and generalization ability. Ensemble learning is a method of achieving consensus in predictions by integrating significant attributes of two or more models, making the final learning framework more comprehensive than that of a single component model, reducing errors and other factors. Compared with ordinary machine learning algorithms, integrated learning algorithms have more advantages in data processing. In this paper aiming at the problems existing in traditional logging lithology identification methods and machine learning methods, the sandstone-type uranium ore in Songliao basin in north China was taken as the research object, and the original data were analyzed and pretreated. Combined with previous studies, two typical integrated algorithm models (XGBoost and SMOTE random Forest) were used to carry out automatic lithology identification of sandstone-type uranium ore in Songliao basin, and the recognition results of the two integrated algorithm models were compared with K-Nearest Neighbor (KNN), Gradient Boosting Decision Tree (GBDT) and other typical machine learning algorithm models were also compared. The results show that the accuracy of XGBoost and SMOTE stochastic forest integrated algorithm model for lithology identification of sandstone-type uranium ore is above 95%, and the accuracy of KNN model and GBDT model is significantly improved. In order to solve the problem of overfitting in operation, XGBoost algorithm model was used to control the regular term of overfitting and node splitting, and support characteristic multithreading to calculate the gain, which improves the operation efficiency and ensures the reliability of the integrated algorithm model. SMOTE synthetic minority oversampling technique solves the problem of sample data imbalance in the random forest algorithm model. The optimization process based on integrated algorithm model provides a theoretical basis for lithology classification of sandstone-type uranium deposits, and provides technical support for strategic breakthrough in uranium exploration.
-
Keywords:
- XGBoost ,
- SMOTE ,
- ensemble learning ,
- sandstone-type uranium deposit ,
- lithology identification
-
-
[1] 李子颖,秦明宽,蔡煜琦,等. 鄂尔多斯盆地砂岩型铀矿成矿作用和前景分析［J］. 铀矿地质,2020,36(1):1-13. LI Ziying, QIN Mingkuan, CAI Yuqi, et al. Analysis on the mineralization and prospect of sandstone type uranium deposit in ordos basin［J］. Uranium Geology, 2020, 36(1): 1-13(in Chinese).
[2] 操应长,姜在兴,夏斌,等. 利用测井资料识别层序地层界面的几种方法［J］. 石油大学学报(自然科学版),2003,27(2):23-26. CAO Yingchang, JIANG Zaixing, XIA Bin, et al. Several methods of identifying sequence stratigraphic interface by logging data［J］. Journal of the University of Petroleum (Natural Science Edition), 2003, 27(2): 23-26(in Chinese).
[3] RUDMAN A J, LANKSTON R W. Stratigraphic correlation of well logs by computer techniques［J］. AAPG Bulletin, 1973, 57(13): 577-588. [4] JOSHI D, PATIDAR A K, MISHRA A, et al. Prediction of sonic log and correlation of lithology by comparing geophysical well log data using machine learning principles［J］. Geo Journal, 2021, 87(6): 1-22. [5] 杨兴,李娜,赵旭. 砂岩型铀矿勘查中密度测井曲线的校正研究及应用［J］. 新疆地质,2021,39(3):497-501. YANG Xing, LI Na, ZHAO Xu. Study and application of density log correction in sandstone type uranium exploration［J］. Geology of Xinjiang, 2021, 39(3): 497-501(in Chinese).
[6] 杨怀杰,乔宝强,冯延强. 小口径砂岩型铀矿测井资料处理解释系统设计［J］. 铀矿地质,2019,35(3):170-174. YANG Huaijie, QIAO Baoqiang, FENG Yanqiang. Design of logging data processing and interpretation system for small diameter sandstone type uranium deposit［J］. Uranium Geology, 2019, 35(3): 170-174(in Chinese).
[7] 邓小卫,赵军辉,李继安,等. 可地浸砂岩型铀矿测井资料自动化处理解释系统的开发研究［J］. 铀矿地质,2007,23(6):363-370. DENG Xiaowei, ZHAO Junhui, LI Ji’an, et al. Research on the automatic processing and interpretation system of logging data of the leachable sandstone type uranium mine［J］. Uranium Geology, 2007, 23(6): 363-370(in Chinese).
[8] HOSSAIN T M, WATADA J, AZIZ I A, et al. Lithology prediction using well logs: A granular computing approach［J］. International Journal of Innovative Computing, Information and Control, 2021, 17(1): 225-244. [9] 徐建国,吴显礼,王卫国. 自组织竞争人工神经网络在砂岩型铀矿测井数据解释中的应用［J］. 世界核地质科学,2008,25(2):114-118. XU Jianguo, WU Xianli, WANG Weiguo.Application of selforganizing competitive artificial neural network in logging data interpretation of sandstone type uranium mine［J］. World Nuclear Geology, 2008, 25(2): 114-118(in Chinese).
[10] 张成勇,聂逢君,刘庆成,等. 二连盆地巴彦乌拉地区砂岩型铀矿目的层电测井曲线响应分析［J］. 铀矿地质,2010,26(2):101-107,119. ZHANG Chengyong, NIE Fengjun, LIU Qing-cheng, et al. Analysis on electric logging curve of target layer of sandstone type uranium deposit in Bayanwula area, Erlian basin［J］. Uranium Geology, 2010, 26(2): 101-107, 119(in Chinese).
[11] 陈钢花,余杰,张孝珍. 基于小波时频分析的测井层序地层划分方法［J］. 新疆石油地质,2007,28(3):355-358. CHEN Ganghua, YU Jie, ZHANG Xiaozhen. Logging sequence stratigraphic division method based on wavelet time-frequency analysis［J］. Xinjiang Petroleum Geology, 2007, 28(3): 355-358(in Chinese).
[12] 潘保芝,李舟波,付有升,等. 测井资料在松辽盆地火成岩岩性识别和储层评价中的应用［J］. 石油物探,2009,48(1):48-52,56. PAN Baozhi, LI Zhoubo, FU Yousheng, et al. Application of logging data in lithology identification and reservoir evaluation of igneous rock in Songliao Basin［J］. Geophysical Prospecting for Petroleum, 2009, 48(1): 48-52, 56(in Chinese).
[13] 吴灿灿,李壮福. 基于BP神经网络的测井相分析及沉积相识别［J］. 煤田地质与勘探,2012,40(1):68-71. WU Cancan, LI Zhuangfu. Logging facies analysis and sedimentary facies recognition based on BP neural network［J］. Coal Geology and Exploration, 2012, 40(1): 68-71(in Chinese).
[14] TIMUR M, RASSUL Y, AMIRGALIYEV Y. Machine learning algorithms for classification geology data from well logging［C］∥14th International Conference on Electronics Computer and Computation (ICECCO). Kaskelen: ［s. n.］, 2018. [15] 杨怀杰,何中波,吴仙明. 基于测井曲线的小波分析识别地层岩性界面——以三江盆地钻孔为例［J］. 铀矿地质,2021,37(2):262-268. YANG Huaijie, HE Zhongbo, WU Xianming. Identification of stratigraphic lithologic interface based on wavelet analysis of logging curve: A case study of borehole in Sanjiang basin［J］. Uranium Geology, 2021, 37(2): 262-268(in Chinese).
[16] 范宜仁,黄隆基,代诗华. 交会图技术在火山岩岩性与裂缝识别中的应用［J］. 测井技术,1999,23(1):53-56,64. FAN Yiren, HUANG Longji, DAI Shihua. Application of crossplot technique in lithology and fracture identification of volcanic rocks［J］. 1999, 23(1): 53-56,64(in Chinese).
[17] 徐德龙,李涛,黄宝华,等. 利用交会图法识别国外M油田岩性与流体类型的研究［J］. 地球物理学进展,2012,27(3):1123-1132. XU Delong, LI Tao, HUANG Baohua, et al. Identification of lithology and fluid type in M oilfield by cross plot［J］. Advances in Geophysics, 2012, 27(3): 1123-1132(in Chinese).
[18] 陶宏根,程日辉,赵小青,等. 海拉尔盆地火山碎屑岩的测井响应与应用［J］. 地球物理学报,2011,54(2):534-544. TAO Honggen, CHENG Rihui, ZHAO Xiaoqing, et al. Logging response and application of pyroclastic rocks in Hailaer basin［J］. Chinese Journal of Geophysics, 2011,54(2): 534-544(in Chinese).
[19] KAZ＇MIERCZUK M, JARZYNA J. Improvement of lithology and saturation determined from well logging using statistical methods［J］. Acta Geophysica, 2006, 54: 378-398 . [20] 刘子云,王向公. 利用概率统计方法判断岩性［J］. 江汉石油学院学报,1989,10(2):35-40. LIU Ziyun, WANG Xianggong. Identification of lithology by probabilistic statistical method［J］. Journal of Jianghan Petroleum Institute, 1989, 10(2): 35-40(in Chinese).
[21] 吴磊,徐怀民,季汉成. 基于交会图和多元统计法的神经网络技术在火山岩识别中的应用［J］. 石油地球物理勘探,2006,47(1):81-86. WU Lei, XU Huaimin, JI Hancheng. Application of neural network technology based on crossplot and multivariate statistical method in volcanic rock identification［J］. Petroleum Geophysical Prospecting, 2006, 47(1): 81-86(in Chinese).
[22] 王祝文,刘菁华,任莉. 基于K均值动态聚类分析的地球物理测井岩性分类方法［J］. 东华理工大学学报(自然科学版),2009,32(2):152-156. WANG Zhuwen, LIU Jinghua, REN Li. Lithology classification method of geophysical logging based on K-means dynamic clustering analysis［J］. Journal of East China University of Technology (Natural Science Edition), 2009, 32(2): 152-156(in Chinese).
[23] BOSCH D, LEDO J, QUERALT P. Fuzzy logic determination of lithologies from well log data: Application to the KTB Project Data set (Germany)［J］. Surv Geophys, 2013, 34: 413-439. [24] AlANAZI A, GATES I D. On the capability of support vector machines to classify lithology from well logs［J］. Natural Resources Research, 2010, 19: 125-139. [25] 肖昆,邹长春,邱礼泉,等. 漠河冻土区天然气水合物科学钻探MK-2孔地层岩性的测井识别［J］. 天然气工业,2013,33(5):46-50. XIAO Kun, ZOU Changchun, QIU Liquan, et al. Logging identification of formation lithology in well MK-2 of natural gas hydrate scientific drilling in Mohe permafrost area［J］. Natural Gas Industry, 2013, 33(5): 46-50(in Chinese).
[26] 牟丹,王祝文,黄玉龙,等. 基于SVM测井数据的火山岩岩性识别——以辽河盆地东部坳陷为例［J］. 地球物理学报,2015,58(5):1785-1793. MOU Dan, WANG Zhuwen, HUANG Yulong, et al. Lithology identification of volcanic rocks based on SVM logging data: A case study of the eastern depression of Liaohe basin［J］. Chinese Journal of Geophysics, 2015, 58(5): 1785-1793(in Chinese).
[27] CVETKOVIC M, VELIC J, MALVIC T. Application of neural networks in petroleum reservoir lithology and saturation prediction［J］. Geologia Croatica, 2009, 62(2): 115-121. [28] SINGH H, SEOL Y, EVGENIY M M. Automated well-log processing and lithology classification by identifying optimal features through unsupervised and supervised machine-learning algorithms［J］. SPE J, 2020, 25: 2778-2800. [29] 韩启迪,张小桐,申维. 基于梯度提升决策树(GBDT)算法的岩性识别技术［J］. 矿物岩石地球化学通报,2018,37(6):1173-1180. HAN Qidi, ZHANG Xiaotong, SHEN Wei. Lithology identification based on gradient lifting decision tree (GBDT) algorithm［J］. Bulletin of Mineralogy, Petrology and Geochemistry, 2018, 37(6): 1173-1180(in Chinese).
[30] 吴灿灿,李壮福. 基于BP神经网络的测井相分析及沉积相识别［J］. 煤田地质与勘探,2012,40(1):68-71. WU Cancan, LI Zhuangfu. Logging facies analysis and sedimentary facies recognition based on BP neural network［J］. Coal Geology and Exploration, 2012, 40(1): 68-71(in Chinese)..
[31] 刘凯,邹正银,王志章,等. 基于机器学习的火山岩岩性智能识别及预测［J］. 特种油气藏,2022,29(1):38-45. LIU Kai, ZOU Zhengyin, WANG Zhizhang, et al. Intelligent identification and prediction of volcanic rock lithology based on machine learning［J］. Special Oil and Gas Reservoirs, 2022, 29(1): 38-45(in Chinese).
[32] MOHAMED I M, MOHAMED S, MAZHER I, et al. Formation lithology classification: Insights into machine learning methods［C］∥SPE Annual Technical Conference and Exhibition. ［S. l.］: ［s. n.］, 2019. [33] 李政宏,刘永福,张立强,等. 数据挖掘方法在测井岩性识别中的应用［J］. 断块油气田,2019,26(6):713-718. LI Zhenghong, LIU Yongfu, ZHANG Liqiang, et al. Application of data mining method in logging lithology identification［J］. Fault-block Oil and Gas Field, 2019, 26(6): 713-718(in Chinese).
[34] 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,32(4):98-106. SUN Yushu, HUANG Yun, LIANG Ting, et al. Lithologic log identification of complex carbonate rocks based on XGBoost algorithm［J］. Lithologic Reservoirs, 2020, 32(4): 98-106(in Chinese).
[35] 杨霞霞,苏锋,黄戌霞. 基于改进随机森林算法的不平衡数据分类方法研究［J］. 网络安全技术与应用,2020(10):70-71. YANG Xiaxia, SU Feng, HUANG Xuxia. Research on unbalanced data classification based on improved random forest algorithm［J］. Network Security Technology and Application, 2020(10): 70-71(in Chinese).
[36] FRIEDMAN, JEROME H. Greedy function approximation: A gradient boosting machine［J］. Annals of Statistics, 2001, 29(1): 1189-1232. [37] CHEN Tianqi, GUESTRIN C. XGBoost: A scalable tree boosting system［C］∥Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ［S. l.］: ［s. n.］, 2016. [38] GUI Zhou, CHEN Jianguo, WANG Chengbin. Classification of geological imbalance data based on PSA-SMOTE random forest—A case study of geochemical data in Eastern Tianshan Mountains［J］. Journal of Guilin University of Technology, 2017, 37(4): 587-593. [39] CHAWLA N, BOWYER K, HAL L, et al. SMOTE: Synthetic minority over-sampling technique［J］. Journal of Artificial Intelligence Research, 2002, 16(1): 321-357. [40] 康乾坤. 基于测井数据的砂岩型铀矿异常识别随机森林算法应用［D］. 长春:吉林大学,2020. [41] XIANG Min, QIN Pengbo, ZHANG Fengwei. Research and application of logging lithology identification for igneous reservoirs based on deep learning［J］. Journal of Applied Geophysics, 2020, 173: 1-8. -
期刊类型引用(5)
1. 江丽，张智谟，王琦玮，封志兵，张博程，任腾飞. 基于不同机器学习模型的石油测井数据岩性分类对比研究. 物探与化探. 2024(02): 489-497 . 
百度学术
2. 王小军，李敬生，李军辉，彭建亮，高翔，谢颖逸，侯蓓蓓，吴景峰，王彪. 大庆油田油气勘探进展及方向. 大庆石油地质与开发. 2024(03): 26-37 . 百度学术
3. 宋延杰，刘英杰，唐晓敏，张兆谦. 基于Stacking算法集成学习的页岩油储层总有机碳含量评价方法. 测井技术. 2024(02): 163-178 . 百度学术
4. 者娜，刘诗文，何攀，郑华，汪量子. 基于经验模态分解和随机森林的阀门泄漏模式识别方法. 原子能科学技术. 2024(S1): 141-148 . 本站查看
5. 张航盛，孙平贺，朱建新，邓盈盈，曹函，张晨，张鑫鑫，蒲英杰. 几种基于随钻参数地层识别方法的对比分析. 钻探工程. 2024(S1): 10-15 . 百度学术
其他类型引用(3)
计量
- 文章访问数: 71
- HTML全文浏览量: 0
- PDF下载量: 47
- 被引次数: 8
下载:
