Please wait a minute...
Journal of ZheJiang University (Engineering Science)  2025, Vol. 59 Issue (6): 1233-1240    DOI: 10.3785/j.issn.1008-973X.2025.06.014
    
SVM prediction method for displacement of high-speed railway piers caused by deep foundation pit excavation
Xuming SONG1(),Xiaolong LI1,Mian TANG1,Tianliang WANG2,Lijuan CHENG3
1. School of Civil Engineering, Central South University, Changsha 410075, China
2. Henan Communications Planning and Design Institute Co. Ltd, Zhengzhou 451450, China
3. Hunan Province Communication Planning, Survey and Design Institute Co. Ltd, Changsha 410200, China
Download: HTML     PDF(963KB) HTML
Export: BibTeX | EndNote (RIS)      

Abstract  

A three-dimensional finite element model considering the influence of groundwater on soil and bridges was established based on a deep foundation pit excavation project, in order to study the impact of additional displacements of high-speed railway bridge piers caused by adjacent foundation pit excavation on railway operation safety. The single factor sensitivity of additional displacements of high-speed railway bridge piers was analyzed. The Box-Behnken Design (BBD) experimental design method combined with the support vector machine algorithm (SVM) was used to establish a displacement prediction model for the top of high-speed railway bridge piers. By combining the Monte Carlo method, 107 sampling calculations were performed on the parameters to obtain the reliable probability of additional displacements at the pier top. The research results showed that the change in the distance between the foundation pit and the high-speed railway bridge pier had the greatest impact on the horizontal and vertical displacements at the pier top. Among the eight different combinations of hyperparameters of the SVM model, the maximum error between the prediction values of the optimal model and the finite element calculation values was within 6%, indicating that the optimal model could replace the finite element for calculation. Under the limit of 2 mm lateral displacement at the pier top, the reliability probability of lateral additional displacement at pier top was 33.12% when the distance between the background engineering foundation pit and the bridge pier was 35 m; when the distance increased to 39 m, the reliability probability of lateral additional displacement at pier top reached 99.68%. The analysis method used can avoid uncertain evaluation results caused by large discretization of soil layer mechanical parameters, providing reference for safety evaluation of similar projects.



Key wordshigh-speed railway      deep foundation pit      additional displacement of pier top      support vector machine (SVM)      reliability     
Received: 15 April 2024      Published: 30 May 2025
CLC:  U 24  
Fund:  国家自然科学基金资助项目(52078486).
Cite this article:

Xuming SONG,Xiaolong LI,Mian TANG,Tianliang WANG,Lijuan CHENG. SVM prediction method for displacement of high-speed railway piers caused by deep foundation pit excavation. Journal of ZheJiang University (Engineering Science), 2025, 59(6): 1233-1240.

URL:

https://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2025.06.014     OR     https://www.zjujournals.com/eng/Y2025/V59/I6/1233


深基坑开挖致高铁桥墩位移的SVM预测方法

为了研究邻近基坑开挖引起的高铁桥梁墩顶附加位移对铁路运营安全的影响,依托某深基坑开挖工程,建立考虑地下水影响的土体-桥梁三维有限元模型. 分析高铁桥墩附加位移的单因素敏感性. 采用Box-Behnken design(BBD)试验设计方法结合支持向量机算法(SVM)建立高铁桥墩墩顶位移预测模型,结合蒙特卡洛法,对参数进行107次抽样计算,得到墩顶附加位移的可靠概率. 研究结果表明:基坑与高铁桥墩距离的变化对墩顶横向位移和竖向位移的影响最大. 在8组不同超参数组合的SVM模型中,最优模型的预测值与有限元计算值的最大误差小于6%,最优模型可代替有限元进行计算. 在墩顶横向位移为2 mm的限值下,背景工程基坑与桥墩距离为35 m时,墩顶横向附加位移的可靠概率为33.12%;当基坑与桥墩距离增加到39 m时,墩顶横向附加位移的可靠概率为99.68%. 所采用的分析方法可以削减因土层力学参数离散性大而产生的评估结果不确定性,为类似工程的安全评估提供参考.


关键词: 高速铁路,  深基坑,  墩顶附加位移,  支持向量机(SVM),  可靠度 
Fig.1 Schematic diagram of optimal hyperplane for SVM and SVR
Fig.2 SVM algorithm prediction and reliability analysis process
Fig.3 Layout plan of construction site
Fig.4 Finite element model after excavation of foundation pit
代号参数名称基准值
1基坑距离/m35
2围护桩等效板厚/m0.832
3素填土内摩擦角/(°)10.00
4素填土黏聚力/kPa12
5素填土泊松比0.31
6素填土卸载模量/MPa35.0
7粉质黏土内摩擦角/(°)18.80
8粉质黏土黏聚力/kPa29.6
9粉质黏土泊松比0.3
10粉质黏土卸载模量/MPa43.4
11强风化泥质砂岩内摩擦角/(°)35.61
12强风化泥质砂岩黏聚力/kPa71.26
13强风化泥质砂岩泊松比0.26
14强风化泥质砂岩卸载模量/MPa175.7
15中风化泥质砂岩内摩擦角/(°)35.02
16中风化泥质砂岩黏聚力/kPa84.34
17中风化泥质砂岩泊松比0.24
18中风化泥质砂岩卸载模量/MPa261.6
Tab.1 Model parameter reference value
参数
代号
横向
位移
竖向
位移
参数
代号
横向
位移
竖向
位移
13.29321.5331100.60350.3548
20.08770.0260110.02180.0344
30.01070.0252120.05780.0453
40.03220.0335130.13140.2072
50.03390.0696140.24950.1854
60.04800.0461150.04090.0755
70.02280.0579160.05040.0889
80.08200.0855170.25680.5057
90.47530.3690180.12010.2642
Tab.2 Parameter sensitivity of pier top displacement
水平A/mBC/MPaDE/MPaFG/MPa
最低水平300.28328.780.238111.590.227180.48
平均水平350.30043.400.260175.700.240261.10
最高水平400.31758.020.282239.810.253341.72
Tab.3 Experimental factors and level of BBD
组号CγYZ
R2RMSE/mmR2RMSE/mm
11005.0000.99990.00320.99990.0004
2702.5000.99990.00310.99990.0004
3100.6000.99980.00600.99530.0048
4110.0890.99680.02910.98070.0097
5300.1000.99890.01760.98490.0086
620.5000.99820.02200.98720.0079
70.051.0000.87710.18170.83680.0283
80.20.0500.84250.20570.74040.0357
Tab.4 Hyperparameter design and corresponding training model regression evaluation indicators
Fig.5 Comparison of prediction results of trained model with results of finite element calculation
组号YZ
R2RMSE/mmemax/%R2RMSE/mmemax/%
10.88460.147116.070.79170.02809.27
20.98280.05686.720.82950.02537.98
30.99280.03683.200.92090.01735.48
40.99590.02792.230.93380.01584.93
50.99540.02942.470.93450.01575.06
60.98690.04954.430.94330.01465.77
Tab.5 Regression evaluation index and maximum error of prediction model
Fig.6 Comparison of predicted values of SVM model and finite element calculation results
Fig.7 Comparison of predicted values of SVM model and finite element calculation results after sample reduction
Fig.8 Relationship curve of prediction accuracy index and kernel function parameter
Fig.9 Comparison of polynomial fitting values and finite element calculation results
因素代号μCVσ
A350.0511.785
B0.3000.020.006
C43.400.125.208
D0.2600.030.0078
E175.700.1322.841
F0.2400.020.0048
G261.100.1128.721
Tab.6 Statistical characteristics of parameter indicators
[7]   雷华阳, 冯双喜, 万勇峰, 等 基坑开挖对既有临近滩涂铁路路基影响规律及安全措施研究[J]. 天津大学学报: 自然科学与工程技术版, 2019, 52 (9): 969- 978
LEI Huayang, FENG Shuangxi, WAN Yongfeng, et al Influence law and safety measures of foundation pit excavation on existing railway subgrade in tidal flat areas[J]. Journal of Tianjin University Science and Technology, 2019, 52 (9): 969- 978
[8]   YE S, ZHAO Z, WANG D Deformation analysis and safety assessment of existing metro tunnels affected by excavation of a foundation pit[J]. Underground Space, 2021, 6 (4): 421- 431
doi: 10.1016/j.undsp.2020.06.002
[9]   LI H, ZHAO Z, DU X. Research and application of deformation prediction model for deep foundation pit based on LSTM [J]. Wireless Communications and Mobile Computing, 2022, 2022: 9407999.
[10]   车天鑫. 土体开挖对高铁桥墩附加位移的响应面方法研究[D]. 长沙: 中南大学, 2022.
CHE Tianxin. Study on response surface method of soil excavation on additional displacement of high-speed railway pier [D]. Changsha: Central South University, 2022.
[11]   SAXENA A, JAT M K, KUMAR S Sensitivity analysis and retrieval of optimum SLEUTH model parameters[J]. Geocarto International, 2022, 37 (25): 7431- 7444
doi: 10.1080/10106049.2021.1974957
[12]   刘榕, 伍英, 丁延书, 等 多塔矮塔斜拉桥结构参数敏感性分析[J]. 铁道科学与工程学报, 2018, 15 (5): 1224- 1230
LIU Rong, WU Ying, DING Yanshu, et al Analysis of structural parameters of multi-span extra-dosed cable-stayed bridge[J]. Journal of Railway Science and Engineering, 2018, 15 (5): 1224- 1230
doi: 10.3969/j.issn.1672-7029.2018.05.018
[13]   CORTES C, VAPNIK V Support-vector networks[J]. Machine Learning, 1995, 20 (3): 273- 297
[14]   李晓龙. 基于支持向量机的岩体力学参数反演及工程应用 [D]. 郑州: 郑州大学, 2009.
LI Xiaolong. Mechanical parameters inversion of rock mass with support vector machine and its engineering application [D]. Zhengzhou: Zhengzhou University, 2009.
[15]   宋旭明, 王天良, 唐冕, 等 堆载作用下高铁桥梁轨道形位变化的可靠度研究[J]. 中南大学学报自然科学版, 2022, 53 (5): 1700- 1710
[1]   邵勇, 陈从新, 鲁祖德, 等 基于机器学习的深基坑人字形支护变形预测分析[J]. 岩土力学, 2020, 41 (Suppl.2): 1- 9
SHAO Yong, CHEN Congxin, LU Zude, et al Prediction and analysis of deformation of deep foundation pit herringbone retaining support (HRS) based on machine learning[J]. Rock and Soil Mechanics, 2020, 41 (Suppl.2): 1- 9
[2]   徐长节, 李欣雨 基于人工神经网络的深基坑支护结构侧移预测[J]. 上海交通大学学报, 2024, 58 (11): 1735- 1744
XU Changjie, LI Xinyu Lateral deformation prediction of deep foundation retaining structures based on artificial neural network[J]. Journal of Shanghai Jiao Tong University, 2024, 58 (11): 1735- 1744
[3]   KUNG G T C, HSIAO E C L, SCHUSTER M, et al A neural network approach to estimating deflection of diaphragm walls caused by excavation in clays[J]. Computers and Geotechnics, 2007, 34 (5): 385- 396
doi: 10.1016/j.compgeo.2007.05.007
[4]   SONG Z, LIU S, JIANG M, et al. Research on the settlement prediction model of foundation pit based on the improved PSO-SVM model [J]. Scientific Programming, 2022, 2022: 1921378.
[5]   LI X, LIU X, LI C Z, et al Foundation pit displacement monitoring and prediction using least squares support vector machines based on multi-point measurement[J]. Structural Health Monitoring, 2019, 18 (3): 715- 724
doi: 10.1177/1475921718767935
[6]   BAO X, CHENG Z, LV C, et al Analysis of the influence of deep foundation excavation on adjacent viaduct pile foundation considering train dynamic loads[J]. Applied Sciences, 2023, 13 (3): 1572- 1591
doi: 10.3390/app13031572
[15]   SONG Xuming, WANG Tianliang, TANG Mian, et al Study on reliability of track geometry of high-speed railway bridge under surcharge load[J]. Journal of Central South University Science and Technology, 2022, 53 (5): 1700- 1710
[16]   王振武, 何关瑶 核函数选择方法研究[J]. 湖南大学学报: 自然科学版, 2018, 45 (10): 155- 160
WANG Zhenwu, HE Guanyao Research on selection method of kernel function[J]. Journal of Hunan University: Natural Sciences, 2018, 45 (10): 155- 160
[17]   CHAUDHARY M T A Sensitivity of seismic response of pile-supported, multi-span viaduct bridges to interaction between soil-foundation and structural parameters[J]. Innovative Infrastructure Solutions, 2023, 8 (6): 180- 210
doi: 10.1007/s41062-023-01145-2
[18]   林升梁, 刘志 基于RBF核函数的支持向量机参数选择[J]. 浙江工业大学学报, 2007, 35 (2): 163- 167
LIN Shengliang, LIU Zhi Parameter selection in SVM with RBF kernel function[J]. Journal of Zhejiang University of Technology, 2007, 35 (2): 163- 167
doi: 10.3969/j.issn.1006-4303.2007.02.010
[19]   CHING J, PHOON K K, CHEN C H Modeling piezocone cone penetration (CPTU) parameters of clays as a multivariate normal distribution[J]. Canadian Geotechnical Journal, 2014, 51 (1): 77- 91
doi: 10.1139/cgj-2012-0259
[20]   CHING J, PHOON K K Modeling parameters of structured clays as a multivariate normal distribution[J]. Canadian Geotechnical Journal, 2012, 49 (5): 522- 545
doi: 10.1139/t2012-015
[21]   国家铁路局. 公路与市政工程下穿高速铁路技术规程: TB 10182—2017 [S]. 北京: 中国铁道出版社.
[22]   潘永杰. 铁路钢桥全寿命过程可靠性分析方法研究 [D]. 北京:中国铁道科学研究院, 2011.
PAN Yongjie. Research on the life cycle reliability analysis method for steel railway bridge [D]. Beijing: China Academy of Railway Sciences, 2011.
[1] Shuang CHEN,Shihua LI,Jing SUN. Model construction and evaluation method of fuzzy reliability life of spherical hinge based on accuracy[J]. Journal of ZheJiang University (Engineering Science), 2025, 59(3): 626-634.
[2] Chuxin WANG,Yingchao WANG,Jiguang YANG,Xiamin FAN,Zheng ZHANG. Land subsidence risk zoning for high speed railway based on hesitant fuzzy linguistic model[J]. Journal of ZheJiang University (Engineering Science), 2024, 58(8): 1691-1703.
[3] Zhi MA,Yao-zhi LUO,Hui-bin GE,Hua-ping WAN,Wen-wei FU,Yan-bin SHEN. Failure probability estimation for structures based on health monitoring data and Bayesian network[J]. Journal of ZheJiang University (Engineering Science), 2023, 57(8): 1551-1561.
[4] Nan WANG,Jin-liu WANG,Cong-hong LIU. Interface opening strategy of high-speed railway station buildings in response to climate and verification by simulation[J]. Journal of ZheJiang University (Engineering Science), 2023, 57(6): 1071-1079.
[5] Xiao-hang LIU,Shan-suo ZHENG,Yu HUANG,Shu-qing DONG,Feng YANG,Jin-qi DONG. Seismic reliability analysis of substation system based on adjacency matrix[J]. Journal of ZheJiang University (Engineering Science), 2022, 56(8): 1495-1503.
[6] Teng-fei YAN,Bao-guo CHEN,Lei ZHANG,Jie-xing HE,Ye-qin ZHANG. Dynamic adjustment method of diaphragm wall supporting system in deep foundation pit and its application[J]. Journal of ZheJiang University (Engineering Science), 2022, 56(2): 356-367.
[7] Xin-ying ZHANG,Lu CHEN,Wen-hui YANG. A parallel-machine scheduling problem with time-changing effect and preventive maintenance[J]. Journal of ZheJiang University (Engineering Science), 2022, 56(2): 408-418.
[8] Wei-hang CHEN,Qiang LUO,Teng-fei WANG,Wen-sheng ZHANG,Liang-wei JIANG. Reliability analysis of post-construction settlement of DMC composite foundation and design optimization[J]. Journal of ZheJiang University (Engineering Science), 2022, 56(10): 2019-2027.
[9] Xin-zhi GAO,Zuo-jun LIU,Yan ZHANG,Ling-ling CHEN. Bicycle riding phase recognition of lower limb amputees based on GWO-SVM[J]. Journal of ZheJiang University (Engineering Science), 2021, 55(4): 648-657.
[10] Xi-ran ZHANG,Shao-kuan CHEN,Bo WANG,Shuang LIU,Zhuo WANG. Emergency allocation optimization model considering reliability of replaceable rescue[J]. Journal of ZheJiang University (Engineering Science), 2021, 55(1): 20-30.
[11] Ge-hui LIU,Shao-kuan CHEN,Hua JIN,Shuang LIU,Hong-qin PENG. Optimum imperfect inspection and maintenance scheduling model considering delay time theory[J]. Journal of ZheJiang University (Engineering Science), 2020, 54(7): 1298-1307.
[12] Li LONG,Shan-suo ZHENG,Yan ZHOU,Jin-chuan HE,Hong-li MENG,Yong-long CAI. Parallel study of seismic reliability analysis of water supply pipe network based on quasi-Monte Carlo method[J]. Journal of ZheJiang University (Engineering Science), 2020, 54(2): 241-247.
[13] Xing-bo HAN,Yong-xu XIA,Yong-dong WANG,Fei YE. Probabilistic degradation model for tunnel lining flexural capacity[J]. Journal of ZheJiang University (Engineering Science), 2019, 53(11): 2175-2184.
[14] DU Ming-yu, BAO Guan-jun, YANG Qing-hua, WANG Zhi-heng, ZHANG Li-bin. Novel method in pattern recognition of hand actions based on improved support vector machine[J]. Journal of ZheJiang University (Engineering Science), 2018, 52(7): 1239-1246.
[15] ZHANG Hang, LI Hong-shuang. Structural reliability analysis with LCVT-SVR method[J]. Journal of ZheJiang University (Engineering Science), 2018, 52(10): 2035-2042.