基于树结构Parzen估计器优化后两层Stacking模型的岩石脆性指数预测

doi:10.3969/j.issn.1672-7703.2025.02.010

摘要/Abstract

摘要： 目前岩石脆性指数的评价方法众多，主要基于矿物组分或岩石力学性质开展评价，但多数评价指标获取费用高昂、耗时长。采用机器学习的手段，提出一种基于Stacking集成学习思想的岩石脆性指数预测方法，并行训练梯度提升决策树模型（GBDT）、随机森林模型（RF）、朴素决策树模型（DT）、支持向量回归模型（SVR）以及LightGBM模型等，并加以树结构Parzen估计器对各模型进行超参数调优后，串行使用XGBoost模型对基模型训练结果进行融合，从而实现各参数的快速寻优和岩石脆性指数的预测。结果表明，基于树结构Parzen估计器优化后的两层Stacking模型预测结果与使用的基模型预测结果相比具有明显优势，其可释方差得分（EVS）最高达到0.97，决定系数（R²）最高达到0.967，在同样的数据集表现中，该模型平均绝对误差（MAE）和均方根误差（RMSE）均最小，表明该模型能够在有监督学习的技术背景下较好地拟合岩石脆性指数的变化规律，验证了其在预测岩石脆性指数方面具有一定的实用价值。

关键词: 岩石脆性指数, Stacking模型, 集成学习, 树结构Parzen估计器

Abstract: Currently, there are numerous methods for evaluating rock brittleness index, mainly based on mineral composition or rock mechanical properties evaluation, but most of these evaluation indicators are costly to obtain and time-consuming. By utilizing machine learning techniques, a rock brittleness index prediction method based on the Stacking ensemble learning concept is proposed. This method involves parallel training of Gradient Boosting Decision Tree model (GBDT), Random Forest model (RF), Naive Decision Tree model (DT),Support Vector Regression model (SVR), and LightGBM model. After hyperparameter tuning using a tree-structured Parzen estimator for each model, the XGBoost model is sequentially used to merge the training results of the base models to achieve rapid parameter optimization and prediction of rock brittleness index. The results indicate that the two-layer Stacking model optimized with the tree-structured Parzen estimator shows significant advantages compared to the predictions of the base models. The explained variance score (EVS) reaches up to 0.97, and the coefficient of determination (R²) reaches a maximum of 0.967. In the same dataset performance comparison, this model exhibits the lowest Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), implying that this model can effectively capture the variation patterns of rock brittleness index under the supervised learning framework. This verifies its practical value in predicting rock brittleness index.

Key words: Rock brittleness index, Stacking model, Ensemble learning, Tree-structured Parzen Estimator

中图分类号:

TE122.1

王芷含, 温韬. 基于树结构Parzen估计器优化后两层Stacking模型的岩石脆性指数预测[J]. 中国石油勘探, 2025, 30(2): 130-147.

Wang Zhihan, Wen Tao. Prediction of Brittleness Index using Two-layer Stacking Model Optimized by Tree-structured Parzen Estimator[J]. China Petroleum Exploration, 2025, 30(2): 130-147.

参考文献

[1] Meng F, Wong L N Y, Zhou H. Rock brittleness indices and their applications to different fields of rock engineering: a review[J]. Journal of Rock Mechanics and Geotechnical Engineering, 2021,13(1):221-247.
[2] 江锚，张丽平，周俊，等. 海上低渗油藏体积压裂可行性研究[J]. 非常规油气，2024,11(3):130-138.
Jiang Mao, Zhang Liping, Zhou Jun, et al . Feasibility study of volume fracturing in offshore low-permeability reservoir[J].Unconventional Oil & Gas, 2024,11(3):130-138.
[3] 刘震，张军华，于正军，等. 非常规储层脆性研究进展及展望[J]. 石油地球物理勘探，2023,58(6):1499-1507.
Liu Zhen, Zhang Junhua, Yu Zhengjun, et al . Progress and prospects of brittleness research in unconventional reservoirs[J].Oil Geophysical Prospecting, 2023,58(6):1499-1507.
[4] 李钜源. 东营凹陷泥页岩矿物组成及脆度分析[J]. 沉积学报，2013,31(4):616-620.
Li Juyan. Analysis on mineral components and frangibility of shales in Dongying Depression[J]. Acta Sedimentologiga Sinica,2013,31(4):616-620.
[5] 曹丹平，韩金鑫，肖竣夫，等. 弹性特征约束下的矿物成分页岩脆性评价方法研究[J]. 地球物理学报，2023,66(11):4781-4791.
Cao Danping, Han Jinxin, Xiao Junfu, et al . Method for evaluating the brittleness of shale minerals under the constraints of elastic characteristics[J]. Chinese Journal of Geophysics,2023,66(11):4781-4791.
[6] Rickman R, Mullen M, Petre E, et al . A practical use of shale petrophysics for stimulation design optimization: all shale plays sre not clones of the barnett shale[J]. Society of Petroleum Engineers, 2008:47-50.
[7] 李坪东，苏幽雅，邵晓岩，等. 盐池地区延长组长8 段致密油储层脆性指数测井评价[J]. 石油地质与工程，2024,38(1):6-12.
Li Pingdong, Su Youya, Shao Xiaoyan, et al . Logging evaluation of brittleness index of tight oil sandstone in Chang 8 member of Yanchang Formation in Yanchi area[J]. Petroleum Geology and Engineering, 2024,38(1):6-12.
[8] Andreev G E. Brittle failure of rock materials[M]. Boca Raton:CRC Press, 1995.
[9] Hucka V, Das B. Brittleness determination of rocks by different methods[J]. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 1974,11(10):389-392.
[10] Khandelwal M, Armaghani D J. Prediction of drillability of rocks with strength properties using a hybrid GA-ANN technique [J]. Geotechnical and Geological Engineering, 2015, 34(2):605-620.
[11] Momeni E, Jahed Armaghani D, Hajihassani M, et al . Prediction of uniaxial compressive strength of rock samples using hybrid particle swarm optimization-based artificial neural networks[J]. Measurement, 2015,60:50-63.
[12] 杜睿山，李宏杰，孟令东，等. 基于PSO-BiLSTM 的储层岩石脆性指数预测[J]. 海南大学学报( 自然科学版)，2023,41(3):260-267.
Du Ruishan, Li Hongjie, Meng Lingdong, et al . Prediction of reservoir rock brittleness index based on PSO-BiLSTM[J].Natural Science Journal of Hainan University, 2023,41(3):260-267.
[13] 张磊，陶千惠，叶婧，等. 考虑分位数回归XGBoost-Stacking 模型的短期惯量区间预测 [J/OL]. 电力系统自动化：1-11.
Zhang Lei, Tao Qianhui, Ye Jing, et al . Short-term Inertia Interval Prediction of Power System Based on XGBoost-Stacking Fusion Model Considering Quantile Regression [J/OL].Automation of Electric Power Systems, 2024,1-11.
[14] 郑颖颖，李鑫，陈延旭，等. 基于Stacking 多模型融合的极端天气短期风电功率预测方法[J]. 高电压技术，2024,50(9):1-12.
Zheng Yingying, Li Xin, Chen Yanxu, et al . Short-term wind power forecasting method in extreme weather based on stacking Multi-model Fusion[J]. High Voltage Engineering,2024,50(9):1-12.
[15] Liu Z, Li D, Liu Y, et al . Prediction of uniaxial compressive strength of rock based on lithology using stacking models[J].Rock Mechanics Bulletin, 2023,2(4):1-5.
[16] 黄遂，孙润，徐琪超，等. 不同温度条件下花岗岩的点荷载强度特征试验研究[J]. 工程与试验，2023,63(2):10-14.
Huang Sui, Sun Run, Xu Qichao, et al . Experimental Study on Point Load Strength Characteristics of Granite under Different Temperature Conditions[J]. Engineering & Test,2023,63(2):10-14
[17] 李家棒，王军成，王斌，等. 纵横波联合勘探在ZY 核电厂选址中的应用[J]. 地质学刊，2024,48(4):1-9.
Li Jiabang, Wang Juncheng, Wang Bin, et al . Application of joint compressional and shear wave exploration in ZY nuclear power plant site selection[J]. Journal of Geology, 2024,48(4):1-9.
[18] 范祥，林杭，熊威，等. 吸水率和吸水时间对红砂岩施密特硬度的影响[J]. 中国矿业大学学报，2015,44(2):233-240.
Fan Xiang, Lin Hang, Xiong Wei, et al . Influence of water absorption rate and time on the Schmidt hardness of red sandstone[J]. Journal of China University of Mining & Technology, 2015,44(2):233-240.
[19] 周洁. 一种深孔内壁邵氏硬度测量装置[J]. 上海电气技术，2022,15(4):52-55,65.
Zhou Jie. A deep-home inner wall shore hardness measuring device[J]. Journal of Shanghai Electric Technology, 2022,15(4):52-55,65.
[20] 郭亮，郭子雪，贾洪涛，等. 基于皮尔逊相关系数与SVM 的居民窃电识别[J]. 河北大学学报( 自然科学版)，2023,43(4):357-363.
Guo Liang, Guo Zixue, Jia Hongtao, et al . Residents electric larceny detection based on Pearson correlation coefficient and SVM[J]. Journal of Hebei University(Natural Science Edition),2023,43(4):357-363.
[21] 罗敏，杨劲锋，俞蕙，等. 基于树结构Parzen 估计器优化集成学习的短期负荷预测方法[J]. 上海交通大学学报，2024,58(6):819-825.
Luo Min, Yang Jinfeng, Yu Hui, et al . TPE-Based Boosting Short-Trem Load Forecasting Method [J]. Journal of Shanghai Jiao Tong University, 2024,58(6):819-825.
[22] 高嘉鑫，张伟，高铭. 基于梯度提升决策树的材料计算时间预测模型[J]. 软件导刊，2024,23(3):15-20.
Gai Jiaxin, Zhang Wei, Gao Ming. Material calculation time prediction model based on gradient boosting decision trees[J].Software Guide, 2024,23(3):15-20.
[23] 车志宏，吕峰. 基于随机森林的集成算法研究[J]. 电脑编程技巧与维护，2024,(5):48-50,80.
Che Zhihong, Lv Feng. Research on ensemble algorithm based on Random Forest[J]. Programming Skills & Maintenance,2024,(5):48-50,80.
[24] 李强，杨林，李超凡，等. 基于SVR 和电化学阻抗谱的锂电池内部温度在线估计[J]. 电源技术，2024,48(9):1738-1746.
Li Qiang, Yang Lin, Li Chaofan, et al . Online estimation of internal temperature of lithium battery based on SVR and electrochemical impedance spectroscopy[J]. Chinese Journal of Power Sources, 2024,48(9):1738-1746.
[25] 田仁飞，李山，刘涛，等. 基于XGBoost 算法的VP/VS 预测及其在储层检测中的应用[J]. 石油地球物理勘探，2024,59(4):653-663.
Tian Renfei, Li Shan, Liu Tao, et al . vP/vS prediction based on XGBoost algorithm and its application in reservoir detection[J]. Oil Geophysical Prospecting, 2024,59(4):653-663.