Identification of Shale Lithofacies by Well Logs Based on Random Forest Algorithm
-
摘要: 泥页岩岩相识别是页岩油空间分布及勘探目标预测的一项重要工作，受地层非均质性及测井信息冗余的制约，基于测井响应方程的岩相识别十分困难.本文建立了一种基于随机森林算法的岩相识别模型，使用SHAP方法量化测井参数重要性.结果表明：随机森林算法可以很好地识别泥页岩岩相，其准确率高于支持向量机、KNN和XGBoost，并且对数据集中岩相类别不均衡的分类问题更加有效；对模型识别岩相最重要的前3项测井参数是自然电位、井径和声波时差；该模型可快速识别单井岩相，再根据总孔隙度、游离烃S1、TOC等参数可确定有利岩相类型，进而确定研究区有利岩相分布，为后续“甜点”预测提供依据.Abstract: Shale lithofacies identification is an important task in the spatial distribution of shale oil and exploration target prediction, but it is difficult to identify lithofacies based on logging response equations due to the formation heterogeneity and redundancy of logging information. In this paper, a lithofacies identification model based on random forest algorithm is proposed, which uses the SHAP method to quantify the contribution of logging parameters. The results show that the random forest algorithm can identify shale lithofacies well, and its accuracy is higher than support vector machine, k-nearest neighbors and XGBoost; SP, CAL and AC contribute the most to the model's identification of lithofacies. The model can quickly identify the lithofacies of a single well, and determine the favorable lithofacies by combining total porosity, free hydrocarbon S1, TOC, etc., and then determine the distribution of favorable lithofacies in the whole area, providing a basis for subsequent "sweet spot" prediction.
-
Key words:
- random forest /
- machine learning /
- logging /
- lithofacies identification /
- shale
-
图 1 松辽盆地某凹陷A段泥页岩矿物组成分布
Fig. 1. Mineral composition distribution of shale in Formation A of a depression in Songliao Basin
图 3 不同岩相与测井参数相关矩阵图
Fig. 3. Correlation matrix of different lithofacies and logging parameters
图 4 不同模型岩相识别结果对比
Fig. 4. Comparison of predicted results from different lithofacies identification models
图 5 不同岩相识别模型混淆矩阵
混淆矩阵中行代表真实岩相，每一行中的数据表示真实岩相被预测的岩相类别分布，如a图第二行中的0.95，表示95%的实测岩相为2的样品被预测为类别2；列代表预测岩相，每一列中的数据表示实测岩相被预测为该类的比例分布
Fig. 5. Comparison of confusion matrices for different lithofacies identification models
图 6 岩相识别中测井参数重要性分析
Fig. 6. Analysis of the importance of logging parameters in lithofacies identification
图 7 X井不同页岩岩相识别模型预测结果对比
Fig. 7. Comparison of predicted results from different lithofacies identification models for Well X
图 8 松辽盆地某凹陷A段有利岩相分布
Fig. 8. Distribution of favourable lithofacies in Formation A of a depression in the Songliao Basin
表 1 松辽盆地某凹陷A段岩相类型发育特征
Table 1. Characteristics of lithofacies type development in Formation A of a depression in the Songliao Basin
序号 岩相类型 岩心观察 薄片观察 TOC(%) 岩石构造 矿物组成(%) 1 富有机质
纹层状
黏土质页岩
> 2.0 页理发育，黏土纹层夹长英质/碳酸盐纹层 黏土 > 50，
长英质 < 50，
碳酸盐 < 502 富有机质
层状
黏土质页岩
> 2.0 页理发育，黏土层夹长英质/碳酸盐层 黏土 > 50，
长英质 < 50，
碳酸盐 < 503 中等有机质
纹层状黏土质页岩
1.0~ 2.0 页理发育，黏土纹层夹长英质/碳酸盐纹层 黏土 > 50，
长英质 < 50，
碳酸盐 < 504 富有机质
纹层状混合质页岩
> 2.0 页理发育，黏土质、长英质、碳酸盐纹层互层 黏土 < 50，
长英质 < 50，
碳酸盐 < 505 富有机质
层状混合质页岩
> 2.0 页理发育，黏土质、长英质、碳酸盐交互层 黏土 < 50，
长英质 < 50，
碳酸盐 < 506 低有机质
块状灰质
泥岩
< 1.0 无页理，碳酸盐矿物颗粒均匀分布 黏土 < 50，
长英质 < 50，
碳酸盐 > 50
下载: 导出CSV
表 2 数据集中不同岩相样本分布
Table 2. The distribution of the different lithofacies samples in the dataset
岩相标签 岩相 样本数 占比（%） 1 富有机质纹层状混合质页岩 46 13.26 2 富有机质纹层状黏土质页岩 78 22.48 3 富有机质层状黏土质页岩 30 8.65 4 中等有机质纹层状黏土质页岩 26 7.49 5 富有机质层状混合质页岩 48 13.83 6 低有机质块状灰质泥岩 76 21.90 10 其他 43 12.39 统计 347
下载: 导出CSV
表 3 随机森林算法参数调优
Table 3. Parameters tuning for random forest
参数 搜索范围 步长 最优值 迭代次数 100~500 2 120 最大树深度 1~15 1 10 内部节点再划分所需最小样本数 1~50 1 4 叶子结点最小样本数 1~20 1 1
下载: 导出CSV
表 4 不同模型的岩相识别F1-score
Table 4. F1-score of different lithofacies identification models
岩相标签 岩相 KNN SVM XGBoost RF 1 富有机质纹层状混合质页岩 0.88 0.85 0.91 0.88 2 富有机质纹层状黏土质页岩 0.91 0.80 0.93 0.95 3 富有机质层状黏土质页岩 0.78 0.84 0.82 0.91 4 中等有机质纹层状黏土质页岩 0.63 0.82 0.78 0.82 5 富有机质层状混合质页岩 0.86 0.70 1.00 1.00 6 低有机质块状灰质泥岩 0.79 0.82 0.87 0.89 10 其他 0.67 0.81 0.86 0.86
下载: 导出CSV
-
[1] Al-Mudhafar, W. J., 2015. Integrating Component Analysis & Classification Techniques for Comparative Prediction of Continuous & Discrete Lithofacies Distributions. In: Offshore Technology Conference. SPE, Houston. [2] Al-Mudhafar, W. J., Al Lawe, E. M., Noshi, C. I., 2019. Clustering Analysis for Improved Characterization of Carbonate Reservoirs in a Southern Iraqi Oil Field. In: Offshore Technology Conference. SPE, Houston. [3] Atchley, S. C., Crass, B. T., Prince, K. C., 2021. The Prediction of Organic-Rich Reservoir Facies within the Late Pennsylvanian Cline Shale (Also Known as Wolfcamp D), Midland Basin, Texas. AAPG Bulletin, 105(1): 29-52. https://doi.org/10.1306/07272020010 [4] Bhattacharya, S., Carr, T. R., Pal, M., 2016. Comparison of Supervised and Unsupervised Approaches for Mudstone Lithofacies Classification: Case Studies from the Bakken and Mahantango-Marcellus Shale, USA. Journal of Natural Gas Science and Engineering, 33: 1119-1133. https://doi.org/10.1016/j.jngse.2016.04.055 [5] Biau, G., Scornet, E., 2016. A Random Forest Guided Tour. Test, 25(2): 197-227. https://doi.org/10.1007/s11749-016-0481-7 [6] Breiman, L., 2001. Random Forests. Machine Learning, 45(1): 5-32. https://doi.org/10.1023/A:1010933404324 [7] Feng, R. H., 2021. Improving Uncertainty Analysis in Well Log Classification by Machine Learning with a Scaling Algorithm. Journal of Petroleum Science and Engineering, 196: 107995. https://doi.org/10.1016/j.petrol.2020.107995 [8] Gifford, C. M., Agah, A., 2010. Collaborative Multi-Agent Rock Facies Classification from Wireline Well Log Data. Engineering Applications of Artificial Intelligence, 23(7): 1158-1172. https://doi.org/10.1016/j.engappai.2010.02.004 [9] Goral, J., Walton, I., Andrew, M., et al., 2019. Pore System Characterization of Organic-Rich Shales Using Nanoscale-Resolution 3D Imaging. Fuel, 258: 116049. https://doi.org/10.1016/j.fuel.2019.116049 [10] Hackley, P. C., Fishman, N., Wu, T., et al., 2016. Organic Petrology and Geochemistry of Mudrocks from the Lacustrine Lucaogou Formation, Santanghu Basin, Northwest China: Application to Lake Basin Evolution. International Journal of Coal Geology, 168: 20-34. https://doi.org/10.1016/j.coal.2016.05.011 [11] Hu, S. Y., Zhao, W. Z., Hou, L. H., et al., 2020. Development Potential and Technical Strategy of Continental Shale Oil in China. Petroleum Exploration and Development, 47(4): 819-828 (in Chinese with English abstract). [12] Li, B. Y., Pang, X. Q., Dong, Y. X., et al., 2019a. Lithofacies and Pore Characterization in an Argillaceous-Siliceous-Calcareous Shale System: A Case Study of the Shahejie Formation in Nanpu Sag, Bohai Bay Basin, China. Journal of Petroleum Science and Engineering, 173: 804-819. https://doi.org/10.1016/j.petrol.2018.10.086 [13] Li, J. B., Wang, M., Chen, Z. H., et al., 2019b. Evaluating the Total Oil Yield Using a Single Routine Rock-Eval Experiment on As-Received Shales. Journal of Analytical and Applied Pyrolysis, 144: 104707. https://doi.org/10.1016/j.jaap.2019.104707 [14] Li, J. B., Jiang, C. Q., Wang, M., et al., 2020. Adsorbed and Free Hydrocarbons in Unconventional Shale Reservoir: A New Insight from NMR T1-T2 Maps. Marine and Petroleum Geology, 116: 104311. https://doi.org/10.1016/j.marpetgeo.2020.104311 [15] Li, Q. Q., Lan, B. F., Li, G. Q., et al., 2021. Element Geochemical Characteristics and Their Geological Significance of Wufeng-Longmaxi Formation Shales in North Margin of the Central Guizhou Uplift. Earth Science, 46(9): 3172-3188 (in Chinese with English abstract). [16] Li, S. X., Zhou, X. P., Guo, Q. H., et al., 2021. Research on Evaluation Method of Movable Hydrocarbon Resources of Shale Oil in the Chang 73 Sub-Member in the Ordos Basin. Natural Gas Geoscience, 32(12): 1771-1784 (in Chinese with English abstract). [17] Lin, M. R., Xi, K. L., Cao, Y. C., et al., 2021. Petrographic Features and Diagenetic Alteration in the Shale Strata of the Permian Lucaogou Formation, Jimusar Sag, Junggar Basin. Journal of Petroleum Science and Engineering, 203: 108684. https://doi.org/10.1016/j.petrol.2021.108684 [18] Liu, B., Shi, J. X., Fu, X. F., et al., 2018. Petrological Characteristics and Shale Oil Enrichment of Lacustrine Fine-Grained Sedimentary System: A Case Study of Organic-Rich Shale in First Member of Cretaceous Qingshankou Formation in Gulong Sag, Songliao Basin, NE China. Petroleum Exploration and Development, 45(5): 828-838 (in Chinese with English abstract). [19] Liu, Z. B., Liu, G. X., Hu, Z. Q., et al., 2019. Lithofacies Types and Assemblage Features of Continental Shale Strata and Their Significance for Shale Gas Exploration: A Case Study of the Middle and Lower Jurassic Strata in the Sichuan Basin. Natural Gas Industry, 39(12): 10-21 (in Chinese with English abstract). [20] Lu, S. F., Li, J. Q., Zhang, P. F., et al., 2018. Classification of Microscopic Pore-Throats and the Grading Evaluation on Shale Oil Reservoirs. Petroleum Exploration and Development, 45(3): 436-444 (in Chinese with English abstract). [21] Lundberg, S. M., Lee, S. I., 2017. A Unified Approach to Interpreting Model Predictions. 31st Conference on Neural Information Processing Systems, Long Beach. [22] Nie, H. K., Zhang, P. X., Bian, R. K., et al., 2016. Oil Accumulation Characteristics of China Continental Shale. Earth Science Frontiers, 23(2): 55-62 (in Chinese with English abstract). [23] Su, S. Y., Jiang, Z. X., Shan, X. L., et al., 2019. Effect of Lithofacies on Shale Reservoir and Hydrocarbon Bearing Capacity in the Shahejie Formation, Zhanhua Sag, Eastern China. Journal of Petroleum Science and Engineering, 174: 1303-1308. https://doi.org/10.1016/j.petrol.2018.11.048 [24] Wang, G. C., Carr, T. R., Ju, Y. W., et al., 2014. Identifying Organic-Rich Marcellus Shale Lithofacies by Support Vector Machine Classifier in the Appalachian Basin. Computers & Geosciences, 64: 52-60. https://doi.org/10.1016/j.cageo.2013.12.002 [25] Wang, M., Ma, R., Li, J. B., et al., 2019. Occurrence Mechanism of Lacustrine Shale Oil in the Paleogene Shahejie Formation of Jiyang Depression, Bohai Bay Basin, China. Petroleum Exploration and Development, 46(4): 789-802 (in Chinese with English abstract). [26] Wang, P., Chen, X. H., Wang, B. F., et al., 2020. An Improved Method for Lithology Identification Based on a Hidden Markov Model and Random Forests. Geophysics, 85(6): IM27-IM36. https://doi.org/10.1190/geo2020-0108.1 [27] Wang, Z. M., Jiang, Y. Q., Fu, Y. H., et al., 2022. Characterization of Pore Structure and Heterogeneity of Shale Reservoir from Wufeng Formation-Sublayers Long-11 in Western Chongqing Based on Nuclear Magnetic Resonance. Earth Science, 47(2): 490-504 (in Chinese with English abstract). [28] Wu, S. T., Zhu, R. K., Cui, J. G., et al., 2015. Characteristics of Lacustrine Shale Porosity Evolution, Triassic Chang 7 Member, Ordos Basin, NW China. Petroleum Exploration and Development, 42(2): 167-176 (in Chinese with English abstract). [29] Zeng, H. B., Wang, F. R., Luo, J., et al., 2021. Characteristics of Pore Structure of Intersalt Shale Oil Reservoir by Low Temperature Nitrogen Adsorption and High Pressure Mercury Pressure Methods in Qianjiang Sag. Bulletin of Geological Science and Technology, 40(5): 242-252 (in Chinese with English abstract). [30] Zhang, B., Mao, Z. G., Zhang, Z. Y., et al., 2021. Black Shale Formation Environment and Its Control on Shale Oil Enrichment in Triassic Chang 7 Member, Ordos Basin, NW China. Petroleum Exploration and Development, 48(6): 1127-1136 (in Chinese with English abstract). [31] 胡素云, 赵文智, 侯连华, 等, 2020. 中国陆相页岩油发展潜力与技术对策. 石油勘探与开发, 47(4): 819-828. https://www.cnki.com.cn/Article/CJFDTOTAL-SKYK202004021.htm [32] 李琪琪, 蓝宝锋, 李刚权, 等, 2021. 黔中隆起北缘五峰‒龙马溪组页岩元素地球化学特征及其地质意义. 地球科学, 46(9): 3172-3188. doi: 10.3799/dqkx.2020.354 [33] 李士祥, 周新平, 郭芪恒, 等, 2021. 鄂尔多斯盆地长73亚段页岩油可动烃资源量评价方法. 天然气地球科学, 32(12): 1771-1784. https://www.cnki.com.cn/Article/CJFDTOTAL-TDKX202112003.htm [34] 柳波, 石佳欣, 付晓飞, 等, 2018. 陆相泥页岩层系岩相特征与页岩油富集条件——以松辽盆地古龙凹陷白垩系青山口组一段富有机质泥页岩为例. 石油勘探与开发, 45(5): 828-838. https://www.cnki.com.cn/Article/CJFDTOTAL-SKYK201805009.htm [35] 刘忠宝, 刘光祥, 胡宗全, 等, 2019. 陆相页岩层系岩相类型、组合特征及其油气勘探意义——以四川盆地中下侏罗统为例. 天然气工业, 39(12): 10-21. https://www.cnki.com.cn/Article/CJFDTOTAL-TRQG201912003.htm [36] 卢双舫, 李俊乾, 张鹏飞, 等, 2018. 页岩油储集层微观孔喉分类与分级评价. 石油勘探与开发, 45(3): 436-444. https://www.cnki.com.cn/Article/CJFDTOTAL-SKYK201803009.htm [37] 聂海宽, 张培先, 边瑞康, 等, 2016. 中国陆相页岩油富集特征. 地学前缘, 23(2): 55-62. https://www.cnki.com.cn/Article/CJFDTOTAL-DXQY201602009.htm [38] 王民, 马睿, 李进步, 等, 2019. 济阳坳陷古近系沙河街组湖相页岩油赋存机理. 石油勘探与开发, 46(4): 789-802. https://www.cnki.com.cn/Article/CJFDTOTAL-SKYK201904020.htm [39] 王子萌, 蒋裕强, 付永红, 等, 2022. 基于核磁共振表征渝西地区五峰组‒龙一1亚段页岩储层孔隙结构及非均质性. 地球科学, 47(2): 490-504. doi: 10.3799/dqkx.2021.076 [40] 吴松涛, 朱如凯, 崔京钢, 等, 2015. 鄂尔多斯盆地长7湖相泥页岩孔隙演化特征. 石油勘探与开发, 42(2): 167-176. https://www.cnki.com.cn/Article/CJFDTOTAL-SKYK201502006.htm [41] 曾宏斌, 王芙蓉, 罗京, 等, 2021. 基于低温氮气吸附和高压压汞表征潜江凹陷盐间页岩油储层孔隙结构特征. 地质科技通报, 40(5): 242-252. https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202105025.htm [42] 张斌, 毛治国, 张忠义, 等, 2021. 鄂尔多斯盆地三叠系长7段黑色页岩形成环境及其对页岩油富集段的控制作用. 石油勘探与开发, 48(6): 1127-1136. https://www.cnki.com.cn/Article/CJFDTOTAL-SKYK202106006.htm -
点击查看大图
图(8) / 表(4)
计量
- 文章访问数: 136
- HTML全文浏览量: 77
- PDF下载量: 49
- 被引次数: 0
