Please wait a minute...
浙江大学学报(工学版)  2025, Vol. 59 Issue (4): 706-716    DOI: 10.3785/j.issn.1008-973X.2025.04.006
土木与建筑工程     
基于多模型Stacking融合的基坑测斜时序预测
胡比澜1,2(),王洋洋1,张永强1,*()
1. 浙江大学 建筑工程学院,浙江 杭州 310058
2. 国家电投集团云南国际电力投资有限公司,云南 昆明 650228
Time series prediction of horizontal displacement in foundation pits based on Stacking multi-model
Bilan HU1,2(),Yangyang WANG1,Yongqiang ZHANG1,*()
1. College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, China
2. SPIC Yunnan International Power Investment Limited Company, Kunming 650228, China
 全文: PDF(2018 KB)   HTML
摘要:

为了准确预测基坑倾斜变形,提出基于极致梯度提升(XGBoost)、长短期记忆(LSTM)和线性回归(LR)的堆叠多变量预测模型. 利用XGBoost集成学习的优越性和双层LSTM算法预测传统基坑变形的准确度,提升模型的预测精度和泛化能力. 在数据预处理阶段,引入K最近邻(KNN)插补算法增加可有效利用的数据总量,使用深度学习模型Informer的时间信息处理方式,改善传统算法中有监督学习忽略时间序列数据不同时间间隔的问题. 以杭州某在建基坑为工程案例,插补616条缺失数据,将时间信息转为3列时间点特征信息,使用所提模型进行基坑变形预测分析. 已有实测数据验证表明,所提模型在预测基坑最大测斜位移及该位移点处深度时的训练精度和泛化能力相比双层LSTM模型及XGBoost模型均有较大提升,使用时间点特征的XGBoost模型比LSTM模型更适合预测对时间因素敏感的指标.

关键词: 时间序列分析基坑测斜双层LSTM极致梯度提升(XGBoost)堆叠算法    
Abstract:

In order to accurately predict the lateral deformation of a foundation pit, a multivariable stacking prediction model based on extreme gradient boosting (XGBoost), long short-term memory (LSTM) and linear regression (LR) was proposed. By using the XGBoost’s advantage of ensemble learning and the accuracy of the two-layer LSTM algorithm in the traditional foundation pit deformation prediction, the prediction accuracy and the generalization ability of the model were improved. In the data pre-processing stage, the K-nearest neighbors (KNN) interpolation algorithm was introduced to increase the total amount of data that can be effectively utilized, and the processing method of time information in the deep learning model Informer was used to deal with the problem of ignoring the different time intervals of time series data by supervised learning in the traditional algorithm. Taking a foundation pit under construction in Hangzhou as a practical engineering case, 616 missing data were interpolated, the time information was converted into three columns of time point feature information, and the proposed model was used for foundation pit deformation analysis. Existing measured data verified that the training accuracy and the generalization ability of the proposed model were greatly improved compared with both the two-layer LSTM model and the XGBoost model when predicting the maximum slope displacement of the foundation pit and the depth of the displacement point. The XGBoost model, which used time-point features, was more suitable for predicting time-sensitive indicators than the LSTM model.

Key words: time series analysis    foundation pits lateral deformation    two-layer LSTM    extreme gradient boosting (XGBoost)    Stacking algorithm
收稿日期: 2024-02-06 出版日期: 2025-04-25
CLC:  TU 433  
通讯作者: 张永强     E-mail: hubilan@zju.edu.cn;cyqzhang@zju.edu.cn
作者简介: 胡比澜(1997—),女,硕士生,从事基坑变形预测研究. orcid.org/0009-0006-7060-2965. E-mail:hubilan@zju.edu.cn
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章  
胡比澜
王洋洋
张永强

引用本文:

胡比澜,王洋洋,张永强. 基于多模型Stacking融合的基坑测斜时序预测[J]. 浙江大学学报(工学版), 2025, 59(4): 706-716.

Bilan HU,Yangyang WANG,Yongqiang ZHANG. Time series prediction of horizontal displacement in foundation pits based on Stacking multi-model. Journal of ZheJiang University (Engineering Science), 2025, 59(4): 706-716.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2025.04.006        https://www.zjujournals.com/eng/CN/Y2025/V59/I4/706

图 1  堆叠算法计算流程图
图 2  极致梯度提升算法的结构示意图
图 3  长短期记忆算法的结构示意图
图 4  基坑变形时间序列预测的反向传播网络模型
图 5  时间列作为索引时的数据处理方式
图 6  时间数据预处理流程图
图 7  基坑平面布置图及测孔位置
类别参数类型取值范围
预测变量基坑最大测斜位移点处的深度$H_{u,{\mathrm{max}}} $连续具体量测、预测值
预测变量基坑最大测斜位移$u_{\mathrm{max}} $连续具体量测、预测值
预测变量基坑最大测斜变化点处的深度$H_{v,{\mathrm{max}}} $连续具体量测、预测值
预测变量基坑最大测斜变化率$v_{\mathrm{max}} $连续具体量测、预测值
特征变量测斜孔号分类0~19等20个测斜孔
特征变量坑外水位连续具体量测值
特征变量支撑轴力连续具体量测值
特征变量主动土压力连续具体计算值
特征变量被动土压力连续具体计算值
特征变量时间(月份)连续[?0.5,0.5]
特征变量时间(星期)连续[?0.5,0.5]
特征变量时间(年)连续[?0.5,0.5]
表 1  模型性能验证参数
图 8  树的最大高度和测试误差
图 9  基坑位移预测的长短期记忆网络结构
图 10  不同算法的多变量误差对比
图 11  不同算法4个预测变量的平均绝对百分比误差
图 12  不同算法的多变量预测结果
图 13  不同算法的逐点相对均方根误差对比
1 钱七虎 迎接我国城市地下空间开发高潮[J]. 岩土工程学报, 1998, 20 (1): 112- 113
QIAN Qihu Meet the upsurge of urban underground space development in China[J]. Chinese Journal of Geotechnical Engineering, 1998, 20 (1): 112- 113
2 钱七虎 中国城市地下空间开发利用的现状评价和前景展望[J]. 民防苑, 2006, (Suppl.1): 1- 5
QIAN Qihu Current situation evaluation and prospect of urban underground space development and utilization in China[J]. Civil Defence Realm, 2006, (Suppl.1): 1- 5
3 油新华, 何光尧, 王强勋, 等 我国城市地下空间利用现状及发展趋势[J]. 隧道建设(中英文), 2019, 39 (2): 173- 188
YOU Xinhua, HE Guangyao, WANG Qiangxun, et al Current status and development trend of urban underground space in China[J]. Tunnel Construction, 2019, 39 (2): 173- 188
4 余苑航, 阎波 我国超大城市地下空间开发现状及其发展趋势[J]. 地下空间与工程学报, 2021, 17 (Suppl.1): 1- 7
YU Yuanhang, BO Yan Present situation and development trend of underground space in megacity in China[J]. Chinese Journal of Underground Space and Engineering, 2021, 17 (Suppl.1): 1- 7
5 李尚明, 洪成雨, 姬凤玲, 等 深基坑的机器视觉监测与变形预测研究[J]. 地下空间与工程学报, 2023, 19 (3): 992- 1000
LI Shangming, HONG Chengyu, JI Fengling, et al Study on machine vision monitoring and deformation prediction of deep foundation pit[J]. Chinese Journal of Underground Space and Engineering, 2023, 19 (3): 992- 1000
6 曹净, 丁文云, 赵党书, 等 基于LSSVM-ARMA模型的基坑变形时间序列预测[J]. 岩土力学, 2014, 35 (Suppl.2): 579- 586
CAO Jing, DING Wenyun, ZHAO Dangshu, et al Time series forecast of foundation pit deformation based on LSSVM-ARMA model[J]. Rock and Soil Mechanics, 2014, 35 (Suppl.2): 579- 586
7 程龙飞, 袁宝远 基于前馈神经网络的基坑测斜位移校正与变形预测研究[J]. 岩石力学与工程学报, 2004, 23 (12): 2038- 2041
CHENG Longfei, YUAN Baoyuan Displacement correction and deformation forecast of foundation pit with feedforward neural network[J]. Chinese Journal of Rock Mechanics and Engineering, 2004, 23 (12): 2038- 2041
doi: 10.3321/j.issn:1000-6915.2004.12.016
8 王雨, 刘国彬, 屠传豹 基于遗传-GRN在深基坑地连墙测斜预测中的研究[J]. 岩土工程学报, 2012, 34 (Suppl.1): 167- 171
WANG Yu, LIU Guobin, TU Chuanbao Deformation prediction for deep excavations based on genetic algorithms-GRNN[J]. Chinese Journal of Geotechnical Engineering, 2012, 34 (Suppl.1): 167- 171
9 洪宇超, 钱建固, 叶源新, 等 基于时空关联特征的CNN-LSTM模型在基坑工程变形预测中的应用[J]. 岩土工程学报, 2021, 43 (Suppl.2): 108- 111
HONG Yuchao, QIAN Jiangu, YE Yuanxin, et al Application of CNN-LSTM model based on spatiotemporal correlation characteristics in deformation prediction of excavation engineering[J]. Chinese Journal of Geotechnical Engineering, 2021, 43 (Suppl.2): 108- 111
10 陈艳茹 基于遗传算法和极限学习机的智能算法在基坑变形预测中的应用[J]. 隧道建设(中英文), 2018, 38 (6): 941- 947
CHEN Yanru Application of intelligent algorithm based on genetic algorithm and extreme learning machine to deformation prediction of foundation pit[J]. Tunnel Construction, 2018, 38 (6): 941- 947
11 李彦杰, 薛亚东, 岳磊, 等 基于遗传算法-BP神经网络的深基坑变形预测[J]. 地下空间与工程学报, 2015, 11 (Suppl.2): 741- 749
LI Yanjie, XUE Yadong, YUE Lei, et al Displacement prediction of deep foundation pit based on genetic algorithms and BP neural network[J]. Chinese Journal of Underground Space and Engineering, 2015, 11 (Suppl.2): 741- 749
12 刘锦, 李峰辉, 刘秀秀 优化GA-BP神经网络模型及基坑变形预测[J]. 隧道建设(中英文), 2021, 41 (10): 1733- 1739
LIU Jin, LI Fenghui, LIU Xiuxiu Optimized genetic algorithm-back propagation neural network model and its application in foundation pit deformation prediction[J]. Tunnel Construction, 2021, 41 (10): 1733- 1739
13 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
14 LI X, PAN Y, ZHANG L, et al Dynamic and explainable deep learning-based risk prediction on adjacent building induced by deep excavation[J]. Tunnelling and Underground Space Technology, 2023, 140: 105243
doi: 10.1016/j.tust.2023.105243
15 DING Q, GUO C, FAN X, et al Multi-source monitoring data helps revealing and quantifying the excavation-induced deterioration of rock mass[J]. Engineering Geology, 2023, 325: 107281
doi: 10.1016/j.enggeo.2023.107281
16 秦胜伍, 张延庆, 张领帅, 等 基于Stacking模型融合的深基坑地面沉降预测[J]. 吉林大学学报: 地球科学版, 2021, 51 (5): 1316- 1323
QIN Shengwu, ZHANG Yanqing, ZHANG Lingshuai, et al Prediction of ground settlement around deep foundation pit based on Stacking model fusion[J]. Journal of Jilin University: Earth Science Edition, 2021, 51 (5): 1316- 1323
17 XU C, CAO B T, YUAN Y, et al A multi-fidelity deep operator network (DeepONet) for fusing simulation and monitoring data: application to real-time settlement prediction during tunnel construction[J]. Engineering Applications of Artificial Intelligence, 2024, 133: 108156
doi: 10.1016/j.engappai.2024.108156
18 袁金荣, 赵福勇 基坑变形预测的时间序列分析[J]. 土木工程学报, 2001, 34 (6): 55- 59
YUAN Jinrong, ZHAO Fuyong Predicting deformation of foundation pit using ANN[J]. China Civil Engineering Journal, 2001, 34 (6): 55- 59
doi: 10.3321/j.issn:1000-131X.2001.06.011
19 ZHOU H , ZHANG S , PENG J , et al. Informer: beyond efficient transformer for long sequence time-Series forecasting [C]// Proceedings of the Thirty-Fifth AAAI Conference on Artificial Intelligence . [S.l.]: AAAI, 2021: 11106–11115.
20 陈振宇, 刘金波, 李晨, 等 基于LSTM与XGBoost组合模型的超短期电力负荷预测[J]. 电网技术, 2020, 44 (2): 614- 620
CHEN Zhenyu, LIU Jinbo, LI Chen, et al Ultra short-term power load forecasting based on combined LSTM-XGBoost model[J]. Power System Technology, 2020, 44 (2): 614- 620
21 COVER T, HART P. Nearest neighbor pattern classification [J], IEEE Transactions Information Theory , 1967, 13(1): 21–27.
22 CHEN T, GUESTRIN C. XGBoost: a scalable tree boosting system [C]// Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining . San Francisco: ACM, 2016: 785–794.
23 史佳琪, 张建华 基于多模型融合Stacking集成学习方式的负荷预测方法[J]. 中国电机工程学报, 2019, 39 (14): 4030- 4041
SHI Jiaqi, ZHANG Jianhua Load forecasting based on multi-model by stacking ensemble learning[J]. Proceedings of the CSEE, 2019, 39 (14): 4030- 4041
24 LI L, SITU R, GAO J, et al. A hybrid model combining convolutional neural network with XGBoost for predicting social media popularity [C]// Proceedings of the 25th ACM International Conference on Multimedia . [S.l.]: ACM, 2017: 1912–1917.
25 李国杰 有关人工智能的若干认识问题[J]. 中国计算机学会通讯, 2021, 17 (7): 44- 51
LI Guojie Some cognitive problems about artificial intelligence[J]. Communications of the CCF, 2021, 17 (7): 44- 51
26 李国杰, 程学旗 大数据研究: 未来科技及经济社会发展的重大战略领域——大数据的研究现状与科学思考[J]. 中国科学院院刊, 2012, 27 (6): 647- 657
LI Guojie, CHENG Xueqi Research status and scientific thinking of big data[J]. Bulletin of Chinese Academy of Sciences, 2012, 27 (6): 647- 657
27 孙钧, 温海洋 人工智能科学在软土地下工程施工变形预测与控制中的应用实践——理论基础、方法实施、精细化智能管理(示例)[J]. 隧道建设(中英文), 2020, 40 (1): 1- 8
SUN Jun, WEN Haiyang Application of artificial intelligence science to construction deformation prediction and control of underground engineering in soft soil: cases study on theoretical foundation, method application and fine intelligent technical management[J]. Tunnel Construction, 2020, 40 (1): 1- 8
28 PAN L X, ZHANG Y R, CAO M S, et al. Parameter sensitivity analysis of geotechnical engineering system using neural network ensemble [C]// Proceedings of the 2015 International Conference on Artificial Intelligence and Industrial Engineering . Phuket: Atlantis Press, 2015: 196–199.
29 钟国强, 王浩, 张国华, 等 基于RS-MIV-ELM模型的基坑水平位移影响因素分析和预测[J]. 上海交通大学学报, 2018, 52 (11): 1508- 1515
ZHONG Guoqiang, WANG Hao, ZHANG Guohua, et al Analysis and prediction of factors affecting horizontal displacement of foundation pit based on RS-MIV-ELM model[J]. Journal of Shanghai Jiao Tong University, 2018, 52 (11): 1508- 1515
30 林宣, 徐海涛, 高旭, 等 软土地区深基坑承压水降水对地表沉降影响分析[J]. 铁道建筑技术, 2022, (4): 156- 160
LIN Xuan, XU Haitao, GAO Xu, et al Influence of confined water dewatering on surface settlement of deep foundation pit in soft soil area[J]. Railway Construction Technology, 2022, (4): 156- 160
31 浙江省建筑设计研究院, 浙江大学. 建筑基坑工程技术规程: DB33/T 1096-2014 [S]. 杭州: 浙江工商大学出版社, 2014.
32 JAIN A. XGBoost parameters tuning: a complete guide with python codes [EB/OL]. [2024−01−21]. https://www.analyticsvidhya.com/blog/2016/03/complete-guide-parameter-tuning-xgboost-with-codes-python/.
[1] 陈功,郑春华,翁献明,HAMEEDBaustani,胡鸿昊,马晓宇,柳景青. 基于多水文数据关联分析的雨水口异常诊断[J]. 浙江大学学报(工学版), 2021, 55(1): 55-61.
[2] 魏媛,冯天恒,黄平捷,侯迪波,张光新. 管网水质多指标动态关联异常检测方法[J]. 浙江大学学报(工学版), 2016, 50(7): 1402-1409.
[3] 王智磊,孙红月,刘永莉,尚岳全. 降雨与边坡地下水位关系的时间序列分析[J]. J4, 2011, 45(7): 1301-1307.