基于数据驱动的岩石微型桩水平极限承载力预测

Prediction of Horizontal Ultimate Bearing Capacity of Rock Micro-Piles Based on Data-Driven Technology

摘要

结合数值模拟和多种数据驱动预测模型能够提高山区岩石微型桩水平极限承载力的预测精度。基于正交试验设计,建立了桩长、桩半径和覆土层厚度等变量的三维有限元模型,生成了微型桩在不同服役工况下的承载力数据集。利用BP神经网络、支持向量机(SVM)、极限梯度提升(XGBoost)和随机森林(RF)模型分别构建了承载力预测模型,并通过可决系数(R²)、平均绝对误差(MAE)和均方根误差(RMSE)等评价指标对模型性能进行对比分析。研究结果表明:随机森林(RF)模型在预测精度和泛化能力方面表现最佳,其中可决系数<i>R</i>²达到0.999 5,平均绝对误差MAE为1.2 kN,均方根误差RMSE约为0.002 6 kN。而BP神经网络、SVM和XGBoost模型虽整体精度略低,但能有效反映承载力变化的整体趋势,可作为辅助工具,用于岩石微型桩水平极限承载力的初步预测。

Abstract

To improve the prediction accuracy of the horizontal ultimate bearing capacity of rock micro-piles in mountainous areas, a comprehensive research method combining numerical simulation and multiple data-driven prediction models is proposed. Based on the orthogonal experimental design, the three-dimensional finite element model with variables such as pile length, pile radius, and overburden soil thickness are established,and a dataset of bearing capacity of micro-piles under different service conditions is generated. Using the BP neural network, support vector machine (SVM), extreme gradient boosting (XGBoost), and random forest (RF), the prediction models for bearing capacity were constructed, and the model performance is compared and analyzed by the evaluation metrics including the coefficient of determination (R²), mean absolute error (MAE), and root mean square error (RMSE). The results show that the random forest model achieves the best prediction accuracy and generalization ability, with R² reaching 0.999 5, MAE of 1.2 kN, and RMSE of approximately 0.002 6 kN. Although the BP neural network, SVM, and XGBoost models have slightly lower overall accuracy, they can effectively capture the overall trend of bearing capacity variation and can serve as auxiliary tools for preliminary prediction of the horizontal ultimate bearing capacity of rock micro-piles.

关键词

微型桩 / 有限元 / 机器学习预测 / 预测模型 / 数据驱动

Key words

Micro-piles / Finite element / Machine learning prediction / Prediction model / Data-driven

参考文献

[1] 薛正元, 余亮, 杨明, 等. 山区特高压输电线路岩石锚杆基础的应用及造价分析[J]. 武汉大学学报(工学版), 2022, 55(S1): 142-147.

[2] 沈茂丁, 孟建, 崔少东. 微型桩在山区油气管道防护中的变形机制模型试验研究[J]. 自然灾害学报, 2023, 32(6): 208-219.

[3] 冯忠居, 张亮, 张聪, 等. 地震作用下变截面钢管混凝土单桩动力响应[J]. 河北大学学报(自然科学版), 2024, 44(2): 122-130.

[4] 张德华, 郭腾, 崔隽. 岩层设承力盘的嵌岩桩抗拔静载试验及计算方法研究[J]. 土木工程学报, 2021, 54(8): 120-128.

[5] 叶青, 郭伟, 任宇晓, 等. 基于内聚力模型的嵌岩双斜桩水平承载特性分析[J]. 岩土工程学报, 2024, 46(S1): 233-238.

[6] 唐孟雄, 凌造, 刘春林, 等. 桩端型式对嵌岩随钻跟管桩承载性能的影响[J]. 岩石力学与工程学报, 2022, 41(S1): 3053-3062.

[7] Wang Lei , Pu Qianhui , Chen Guangpeng , et al. Dynamic responses characteristics of bedrock and overburden layer slope with anchored frame piles based on shaking table test[J]. Scientific Reports, 2025, 15: 4574.

[8] Zhang Junbao , Hu Yulie , Wang Yiteng , et al. Sidewall roughness measurement and bearing performance simulation of rock—socketed piles based on laser scanning point cloud[J]. Applied Sciences, 2025, 15: 889.

[9] Yan Nan , Hao Zengming , Bai Xiaoyu , et al. Load—bearing characteristics and optimization design for rock—socketed bored piles in sandy silty clay in coastal areas[J]. Soil Dynamics and Earthquake Engineering, 2025, 190: 109207.

[10] 王敬德, 张新春, 刘增炜, 等. 山区机械化成孔的大孔径硬岩嵌岩桩承载能力[J]. 河北大学学报(自然科学版), 2024, 44(6): 561-570.

[11] 张鹏远, 吴宝峰. 基于PSO算法优化BP神经网络的打入桩承载力预测模型研究[J]. 水利科学与寒区工程, 2024, 7(11): 62-65.

[12] Nguyen T , Ly K D , Shiau J , et al. Optimizing load—displacement prediction for bored piles with the 3mSOS algorithm and neural networks[J]. Ocean Engineering, 2024, 304: 117758.

[13] 杜轲, 吴文贤, 林志鹏, 等. 基于物理驱动支持向量机方法的地震作用下结构动力响应求解[J]. 振动与冲击, 2025, 44(3): 284-290.

[14] 苏国韶, 张研, 燕柳斌. CFG桩复合地基承载力预测的高斯过程模型[J]. 计算机工程与应用, 2011, 47(4): 236-238.

[15] 杜双奎. 试验优化设计与统计分析[M]. 北京:科学出版社, 2020: 143-145.

[16] 季雨坤, 王钦科, 赵国良, 等. 斜坡上嵌岩抗拔桩竖向承载变形特性模型试验及数值模拟[J]. 岩土力学, 2023, 44(6): 1604-1614.

[17] 李克先, 雷刚, 李健, 等. 土岩组合地层深基坑桩撑体系变形及受力分析[J]. 科学技术与工程, 2021, 21(1): 310-317.

[18] 白晓宇, 刁浩杰, 银吉超, 等. 大直径嵌岩灌注桩承载特性试验与有限元模拟[J]. 重庆交通大学学报(自然科学版), 2024, 43(9): 25-33.

[19] Zhang Xinchun , Gu Lirong , Yin Xiaodi , et al. Mechanical behavior and failure prediction of cylindrical lithium—ion batteries under mechanical abuse using data—driven machine learning[J]. Journal of Applied Mechanics—Transactions of the ASME, 2025, 92: 021003.

[20] Ai Yuhao , Song Bin , Wu Shaocheng , et al. Diagnosis of reverse—connection defects in high—voltage cable cross—bonded grounding system based on ARO—SVM[J]. Sensors, 2025, 25: 590.

[21] Dashtgoli S D , Dehnad H M , Mobinipour A S , et al. Performance comparison of machine learning algorithms for maximum displacement prediction in soldier pile wall excavation[J]. Underground Space, 2024, 16: 301-313.

[22] Yaychi B M , Esmaeili F M . Estimating axial bearing capacity of driven piles using tuned random forest frameworks[J]. Geotechnical and Geological Engineering, 2024, 42: 7813-7834.

基金资助

河北省专业学位教学案例(库)项目“输电线路工程案例建设”(KCJSZ2024125)