Prediction and analysis of foundation pit support deformation for ultra-large dock chambers based on CNN-LSTM-Attention model
-
摘要: 某超大型船坞坞室基坑采用锚碇式组合钢板桩围护体系。为研究基坑变形规律，对坞墙承台、组合钢板桩、锚碇墙等围护结构进行沉降、水平位移和受力监测，分析其变化趋势、原因及时空关联性。结果表明，采用无内支撑组合钢板桩的船坞坞室围护设计，受软土特性和降水影响，其围护结构顶部沉降和水平位移、组合钢板桩深层水平位移在开挖过程中变形较大，各监测项目数据变化具有较强的时空关联效应。建立了一种可以提取监测时序数据关键特征的CNN-LSTM-Attention组合神经网络模型，以不同监测项目的监测数据作为模型输入，经过CNN神经网络卷积运算、LSTM神经网络回归预测、Attention注意力权重分配、全连接层等操作输出预测值，并与实测数据对比分析。根据试验结果，引入注意力机制的CNN- LSTM-Attention的组合模型，比LSTM模型和BiLSTM模型预测精度更高，且在相邻监测点的预测中，该模型适用性良好，验证了监测点数据变化趋势具有较强的空间相关性。Abstract: The foundation pit of an ultra-large dock chamber adopts the anchor type composite steel sheet pile retaining system. To study the deformation law of foundation pit, the settlement, horizontal displacement and stress of retaining structures such as dock wall bearing platform, composite steel sheet pile and anchor wall were monitored, and their change trend, causes and temporal and spatial correlation were analyzed. The results show that the top settlement and horizontal displacement of the retaining structure and the deep horizontal displacement of the composite steel sheet pile are large in the excavation process due to the influence of soft soil characteristics and precipitation, and the data change of each monitoring item has a strong spatio-temporal correlation effect. A CNN-LSTM-Attention combined neural network model which can extract the key features of monitoring time series data is established. The monitoring data of different monitoring items are taken as the model input, and the predicted values are output through CNN neural network convolution operation, LSTM neural network regression prediction, Attention weight distribution, full connection layer and other operations, and compared with the measured data. According to the experimental results, the combined model of CNN-LSTM-Attention with attention mechanism has higher prediction accuracy than LSTM model and BiLSTM model. In the prediction of adjacent monitoring points, the model has good applicability, which verifies that the change trend of monitoring points’ data has strong spatial correlation.
-
表 1 土层主要物理力学参数
层号 土层名称 重度$ \gamma $/(kN·m−3) 内摩擦角$ \varphi $/(°) 黏聚力c/kPa 渗透系数k/(cm·s−1) ①1-1d 吹（冲）填土 17.6 29.5 4 5.0×10−4 ①1-2a 灰褐色淤泥质粉质黏土夹粉性土 17.9 21 11 8.0×10−5 ①1-2b 灰黄色黏质粉土 18.6 26.5 8 4.0×10−4 ①1-2c 灰褐色淤泥质粉质黏土夹粉性土 17.9 18.5 12 5.0×10−5 ②1 灰黄色粉质黏土 18.4 20.5 19 3.0×10−6 ②2 灰色淤泥质粉质黏土夹黏质粉土 18.1 20.5 13 2.0×10−5 ②3-1 灰色砂质粉土 18.7 27 8 5.0×10−4 ②3-3 灰色黏质粉土夹淤泥质粉质黏土 18.1 23.5 10 7.0×10−4 ④ 灰色淤泥质黏土 16.9 11.0 12 3.0×10−7 ⑤1-1 灰色黏土 17.3 13.0 14 5.0×10−7 ⑤1-2 灰色黏土 17.8 16.0 16 5.0×10−6 ⑤3-1 灰色粉质黏土夹粉性土 18.2 20.5 20 5.0×10−6 ⑤3-2a 灰色粉质黏土夹粉性土 18.4 22.5 21 5.0×10−6 ⑤3-2b 灰色粉质黏土夹粉性土 18.9 30.5 6 5.0×10−6 
下载: 导出CSV
表 2 监测相关技术要求
监测项目 点位间距 设计报警值 监测频率 承台顶竖向位移 20 m ±50 mm 1次/d 承台顶水平位移 ±250 mm 1次/d 锚碇墙顶竖向位移 20 m ±100 mm 1次/d 锚碇墙顶水平位移 ±250 mm 1次/d 钢板桩深层水平位移 40 m ±250 mm 1次/d 锚拉杆受力 40 m 1200 kN 1次/d
下载: 导出CSV
表 3 监测点处施工过程及日期
位置 工况施工时段（年-月-日—年-月-日） 基坑开挖 垫层及底板施工 底板完成及结构施工 W40，W41 2022-11-05—
2023-03-102022-03-11—
2023-04-102023-04-11—
2023-06-10W42，W43 2022-10-09—
2023-02-172023-02-18—
2023-03-152023-03-16—
2023-05-12W45，W46 2022-09-30—
2022-12-162022-12-17—
2023-01-162023-01-17—
2023-03-06W47，W48 2022-08-25—
2022-10-152022-10-16—
2022-11-182022-11-19—
2023-02-05
下载: 导出CSV
表 4 不同预测模型RMSE值
点号及监测项目 LSTM RMSE BiLSTM RMSE CNN-LSTM-Attention RMSE 训练集 测试集 训练集 测试集 训练集 测试集 W48沉降/mm 2.25 0.98 2.45 2.01 1.54 0.65 W48水平位移/mm 2.86 3.97 2.27 1.73 1.74 0.47 QX25深层水平位移/mm 9.08 2.65 7.45 3.14 6.23 1.11 ML6锚杆拉力/kN 13.44 17.28 12.33 14.72 11.79 13.83
下载: 导出CSV
-
[1] 陈飞仰. 钢板桩在软土基坑工程中的应用研究[J]. 山西建筑, 2024, 50(1): 99-101,120. (CHEN F Y. Research on the application of steel sheet piles in soft soil foundation pit engineering[J]. Shanxi Architecture, 2024, 50(1): 99-101,120. (in Chinese)CHEN F Y. Research on the application of steel sheet piles in soft soil foundation pit engineering[J]. Shanxi Architecture, 2024, 50（1）: 99-101,120. (in Chinese) [2] 侯亚彬, 陈 玉, 王新胜, 等. 山地城市某地铁深基坑监测与特性分析[J]. 测绘通报, 2020(S1): 91-95. (HOU Y B, CHEN Y, WANG X S, et al. Monitoring and characteristic analysis of asubway deep foundation pit in the mountainous city[J]. Bulletin of Surveying and Mapping, 2020(S1): 91-95. (in Chinese)HOU Y B, CHEN Y, WANG X S, et al. Monitoring and characteristic analysis of asubway deep foundation pit in the mountainous city[J]. Bulletin of Surveying and Mapping, 2020（S1）: 91-95. (in Chinese) [3] 刘志刚. 软土地区地铁深基坑监测分析及控制措施研究[J]. 公路, 2019, 64(10): 239-244. (LIU Z G. Research on monitoring analysis and control measures of deep foundation pit of metro in soft soil area[J]. Highway, 2019, 64(10): 239-244. (in Chinese)LIU Z G. Research on monitoring analysis and control measures of deep foundation pit of metro in soft soil area[J]. Highway, 2019, 64(10): 239-244. (in Chinese) [4] 王卫东. 软土深基坑变形及环境影响分析方法与控制技术[J]. 岩土工程学报, 2024, 46(1): 1-25. (WANG W D. Analytical methods and controlling techniques for deformation and environmental influence of deep excavations in soft soils[J]. Chinese Journal of Geotechnical Engineering, 2024, 46(1): 1-25. (in Chinese) doi: 10.11779/CJGE20231146WANG W D. Analytical methods and controlling techniques for deformation and environmental influence of deep excavations in soft soils[J]. Chinese Journal of Geotechnical Engineering, 2024, 46（1）: 1-25. (in Chinese) doi: 10.11779/CJGE20231146 [5] 李金华, 宋 涛, 朱海西, 等. 南宁地铁2号线车站深基坑开挖变形规律研究[J]. 施工技术, 2017, 46(19): 32-35. (LI J H, SONG T, ZHU H X, et al. Research on deformation law of station deep foundation excavation in Nanning subway line 2[J]. Construction Technology, 2017, 46(19): 32-35. (in Chinese)LI J H, SONG T, ZHU H X, et al. Research on deformation law of station deep foundation excavation in Nanning subway line 2[J]. Construction Technology, 2017, 46（19）: 32-35. (in Chinese) [6] 倪小东, 王 琛, 唐栋华, 等. 软土地区深基坑超大变形预警及诱因分析[J]. 中南大学学报(自然科学版), 2022, 53(6): 2245-2254. (NI X D, WANG C, TANG D H, et al. Early-warning and inducement analysis of super-large deformation of deep foundation pit on soft soil[J]. Journal of Central South University (Science and Technology), 2022, 53(6): 2245-2254. (in Chinese)NI X D, WANG C, TANG D H, et al. Early-warning and inducement analysis of super-large deformation of deep foundation pit on soft soil[J]. Journal of Central South University (Science and Technology), 2022, 53（6）: 2245-2254. (in Chinese) [7] 王 淞, 刘永超, 张建新, 等. 深厚软土地区倾斜桩组合支护基坑稳定性研究[J]. 地下空间与工程学报, 2024, 20(2): 556-565,576. (WANG S, LIU Y C, ZHANG J X, et al. Research on stability of foundation pit with inclined pile combination support in deep soft soil area[J]. Chinese Journal of Underground Space and Engineering, 2024, 20(2): 556-565,576. (in Chinese)WANG S, LIU Y C, ZHANG J X, et al. Research on stability of foundation pit with inclined pile combination support in deep soft soil area[J]. Chinese Journal of Underground Space and Engineering, 2024, 20（2）: 556-565,576. (in Chinese) [8] 李泽文, 谭 勇, 廖少明, 等. 上海超深基坑地下连续墙的空间变形特性实测分析[J]. 岩土工程学报, 2024, 46(11): 2380-2390. (LI Z W, TAN Y, LIAO S M, et al. Field test on the performance of diaphragm wall for an ultra-deep excavation in Shanghai soft ground[J]. Chinese Journal of Geotechnical Engineering, 2024, 46(11): 2380-2390. (in Chinese) doi: 10.11779/CJGE20230760LI Z W, TAN Y, LIAO S M, et al. Field test on the performance of diaphragm wall for an ultra-deep excavation in Shanghai soft ground[J]. Chinese Journal of Geotechnical Engineering, 2024, 46（11）: 2380-2390. (in Chinese) doi: 10.11779/CJGE20230760 [9] 范立军. 时空效应下双环形内支撑深基坑监测分析[J]. 工程勘察, 2023, 51(11): 9-17. (FAN L J. Monitoring and analysis of deep excavation supported by double annular arched girders under space-time effect[J]. Geotechnical Investigation & Surveying, 2023, 51(11): 9-17. (in Chinese)FAN L J. Monitoring and analysis of deep excavation supported by double annular arched girders under space-time effect[J]. Geotechnical Investigation & Surveying, 2023, 51（11）: 9-17. (in Chinese) [10] 董文宝. 某深基坑监测及变形预测模型研究[D]. 武汉: 武汉理工大学, 2013. (DONG W B. The research on a deep foundation pit monitoring and deformation forecast model[D]. Wuhan: Wuhan University of Technology, 2013. (in Chinese)DONG W B. The research on a deep foundation pit monitoring and deformation forecast model[D]. Wuhan: Wuhan University of Technology, 2013. (in Chinese) [11] 程秋实. 深圳市某深基坑施工变形监测与预测研究[D]. 长春: 吉林大学, 2022. (CHENG Q S. Research on monitoring and prediction of a deep foundation pit construction deformation in Shenzhen[D]. Changchun: Jilin University, 2022. (in Chinese)CHENG Q S. Research on monitoring and prediction of a deep foundation pit construction deformation in Shenzhen[D]. Changchun: Jilin University, 2022. (in Chinese) [12] 张立奎, 段大猷, 王佐才. 基于LSTM神经网络的多源数据融合桥梁变形重构方法[J]. 土木与环境工程学报(中英文), 2022, 44(3): 37-43. (ZHANG L K, DUAN D Y, WANG Z C. A multi-source data fusion method for bridge displacement reconstruction based on LSTM neural network[J]. Journal of Civil and Environmental Engineering, 2022, 44(3): 37-43. (in Chinese)ZHANG L K, DUAN D Y, WANG Z C. A multi-source data fusion method for bridge displacement reconstruction based on LSTM neural network[J]. Journal of Civil and Environmental Engineering, 2022, 44（3）: 37-43. (in Chinese) [13] 王洪德, 肖 杨. 深基坑开挖对邻近地埋管线扰动影响分析[J]. 辽宁工程技术大学学报(自然学报科学版), 2018, 37(5): 769-775. (WANG H D, XIAO Y. Influence of soil disturbance on adjacent buried pipelines due to excavation of deep foundation pit[J]. Journal of Liaoning Technical University (Natural Science), 2018, 37(5): 769-775. (in Chinese)WANG H D, XIAO Y. Influence of soil disturbance on adjacent buried pipelines due to excavation of deep foundation pit[J]. Journal of Liaoning Technical University (Natural Science), 2018, 37（5）: 769-775. (in Chinese) [14] 刘 栋. 深度学习框架下时空关联的基坑变形预测方法研究[D]. 泰安: 山东农业大学, 2022. (LIU D. Research on prediction method of foundation pit deformation based on spatio-temporal correlation under deep learning framework[D]. Tai’an: Shandong Agricultural University, 2022. (in Chinese)LIU D. Research on prediction method of foundation pit deformation based on spatio-temporal correlation under deep learning framework[D]. Tai’an: Shandong Agricultural University, 2022. (in Chinese) [15] 张 旭. 基于循环神经网络的时间序列预测方法研究[D]. 南京: 南京大学, 2019. (ZHANG X. Time series forecasting based on recurrent neural networks[D]. Nanjing: Nanjing University, 2019. (in Chinese)ZHANG X. Time series forecasting based on recurrent neural networks[D]. Nanjing: Nanjing University, 2019. (in Chinese) [16] 周兰庭, 邓思源, 柳志坤, 等. 基于SSA-LSTM-GF的混凝土坝变形预测方法[J]. 河海大学学报(自然科学版), 2023, 51(2): 73-80,149. (ZHOU L T, DONG S Y, LIU Z K, et al. Deformation prediction method of concrete dam based on SSA-LSTM-GF[J]. Journal of Hohai University (Natural Sciences), 2023, 51(2): 73-80,149. (in Chinese)ZHOU L T, DONG S Y, LIU Z K, et al. Deformation prediction method of concrete dam based on SSA-LSTM-GF[J]. Journal of Hohai University (Natural Sciences), 2023, 51（2）: 73-80,149. (in Chinese) [17] 高墨通, 杨维芳, 刘祖昱, 等. 结合卷积神经网络和注意力机制的LSTM采空区地表沉降预测方法[J]. 测绘通报, 2024(6): 53-58,170. (GAO M T, YANG W F, LIU Z Y, et al. LSTM goaf surface subsidence prediction method combining convolutional neural network and attention mechanism[J]. Bulletin of Surveying and Mapping, 2024(6): 53-58,170. (in Chinese)GAO M T, YANG W F, LIU Z Y, et al. LSTM goaf surface subsidence prediction method combining convolutional neural network and attention mechanism[J]. Bulletin of Surveying and Mapping, 2024（6）: 53-58,170. (in Chinese) [18] 李 鑫, 魏冠军, 张德龙. 联合PS-InSAR技术与多变量LSTM神经网络的高铁路基冻胀形变预测研究[J]. 测绘通报, 2024(1): 58-64. (LI X, WEI G J, ZHANG D L. Research on frost heave deformation prediction of high railway foundation combined with PS-InSAR technology and multivariable LSTM neural network[J]. Bulletin of Surveying and Mapping, 2024(1): 58-64. (in Chinese)LI X, WEI G J, ZHANG D L. Research on frost heave deformation prediction of high railway foundation combined with PS-InSAR technology and multivariable LSTM neural network[J]. Bulletin of Surveying and Mapping, 2024（1）: 58-64. (in Chinese) -
图(19) / 表(5)
计量
- 文章访问数: 9
- HTML全文浏览量: 2
- PDF下载量: 3
- 被引次数: 0
