Safety Risk Warning of Deep Foundation Pit Deformation Based on LSTM
-
摘要: 为了预防深基坑施工安全事故，提出了一套基于监测数据的风险预警标准，建立了基于长短时记忆网络（Long Short-Term Memory，LSTM）的深基坑变形安全风险预警模型.依托实际深基坑工程项目，将风险预警模型应用其中，对基坑各监测项目的变形量进行短期的预测，预测数据与实际数据最大误差为5.04%，最小误差为0.04%，平均相对误差为2.41%，证明该模型的预测效果良好.表明基于LSTM的深基坑变形安全风险预警模型在基坑变形预测方面有着良好的精确性和优越性，可以为基坑工程的安全性判断与风险管控提供可靠的保障.Abstract: In order to prevent deep foundation pit construction safety accidents, a set of risk warning standards based on monitoring data is proposed, and a deep foundation pit deformation safety risk warning model based on long short-term memory (LSTM) was established. Relying on the actual deep foundation pit engineering project, the risk warning model is applied to it to make short-term predictions of the deformation of each monitoring item of the foundation pit. The maximum error between the predicted data and the actual data is 5.04%, the minimum error is 0.04%, and the average relative error is 2.41%, which proves that the prediction effect of the model is good. It shows that the LSTM-based deep foundation pit deformation safety risk early warning model has good accuracy and superiority in the prediction of foundation pit deformation, and can provide a reliable guarantee for the safety judgment and risk management of foundation pit engineering.
-
图 1 基坑安全风险预警指标体系
Fig. 1. Pre-warning indicator system for safety risks of foundation pits
图 5 点位DB9-1累积变形量实测数据和LSTM模型预测数据曲线对比
Fig. 5. Comparison of the measured data of the cumulative deformation of the point DB9-1 and the predicted data curve of the LSTM model
表 1 深基坑变形安全风险预警指标体系
Table 1. Deep foundation pit deformation safety risk warning index system
目标 一级指标 二级指标（部分可选） 深基坑变形安全风险预警 主体结构 （坡）顶水平位移 墙（坡）顶竖向位移 围护墙深层水平位移 土体深层水平位移 墙（桩）体内力 支撑内力 立柱竖向位移 锚杆、土钉拉力 土层分层竖向位移 坑底隆起 土压力 孔隙水压力 周边环境 地下水位 墙后地表竖向位移 周围建（构）筑物变形 周围地下管线变形 
下载: 导出CSV
表 2 风险等级划分标准及权重系数
Table 2. Risk classification standards and weight coefficients
α值 < 0.3 0.3~0.7 0.7~0.9 0.9~1.2 > 1.2 风险等级 r 一级 二级 三级 四级 五级 权重系数γ 0.018 0.050 0.135 0.368 1.000
下载: 导出CSV
表 3 风险接受准则
Table 3. Risk acceptance criteria
等级 接受准则 控制方案 一级 可忽略 日常管理和审视 二级 可容许 需注意，加强日常管理审视 三级 可接受 引起重视，需防范，监控措施 四级 不可接受 需决策，指定控制、预警措施 五级 拒绝接受 立即停止，整改，规避或启动预案
下载: 导出CSV
表 4 所有点位LSTM预测结果
Table 4. LSTM prediction results of all points
日期 DB9-1 J3 DL-1 SW-1 B1（竖向） B1（水平） P1 LZ12 ZL2-1 2019-08-05 -26.20 -3.35 -8.68 -833.74 11.27 5.99 29.92 12.26 3 455.46 2019-08-06 -26.19 -3.64 -8.65 -831.61 11.31 5.99 29.93 12.22 3 523.61 2019-08-07 -26.19 -3.55 -8.72 -860.73 11.33 5.99 29.93 12.28 3 549.88 2019-08-08 -26.18 -3.30 -8.69 -868.82 11.31 6.23 29.93 12.29 3 538.18 2019-08-09 -26.19 -3.14 -8.72 -876.15 11.28 6.23 29.93 12.24 3 230.88 2019-08-11 -26.19 -3.09 -8.72 -882.62 11.28 6.23 29.94 12.24 4 265.69 2019-08-12 -26.18 -3.04 -8.66 -875.40 11.27 6.23 29.95 12.20 3 999.38 2019-08-13 -26.18 -3.11 -8.70 -816.85 11.29 6.23 29.95 12.18 4 015.25 2019-08-14 -26.17 -3.22 -8.75 -751.35 11.24 6.23 29.95 12.15 4 152.56 2019-08-15 -26.17 -3.46 -8.75 -799.07 11.24 6.40 29.96 12.15 4 821.04
下载: 导出CSV
表 5 DB9-1监测数据与预测值
Table 5. DB9-1 monitoring data and predicted values
日期 实测值
(mm)预测值
(mm)绝对误差(mm) 相对误差(%) 2019-08-05 -25.36 -26.20 0.84 3.31 2019-08-06 -25.94 -26.19 0.25 0.96 2019-08-07 -26.58 -26.19 0.39 1.47 2019-08-08 -26.89 -26.18 0.71 2.64 2019-08-09 -26.20 -26.19 0.01 0.04 2019-08-11 -26.20 -26.19 0.01 0.04 2019-08-12 -26.88 -26.18 0.70 2.60 2019-08-13 -26.98 -26.18 0.80 2.97 2019-08-14 -27.56 -26.17 1.39 5.04 2019-08-15 -27.56 -26.17 1.39 5.04
下载: 导出CSV
表 6 点位DB9-1实测数据与预测数据风险量对比
Table 6. Comparison of measured data and predicted data risk of DB9-1 at points
日期 实测数据 预测数据 累积变形量
（mm）风险量 风险等级 累积变形量
（mm）风险量 风险等级 2019-08-05 -25.36 0.90 四级 -26.2 0.93 四级 2019-08-06 -25.94 0.92 四级 -26.19 0.93 四级 2019-08-07 -26.58 0.94 四级 -26.19 0.93 四级 2019-08-08 -26.89 0.95 四级 -26.18 0.92 四级 2019-08-09 -26.2 0.92 四级 -26.19 0.92 四级 2019-08-11 -26.2 0.92 四级 -26.19 0.92 四级 2019-08-12 -26.88 0.94 四级 -26.18 0.92 四级 2019-08-13 -26.98 0.95 四级 -26.18 0.92 四级 2019-08-14 -27.56 0.97 四级 -26.17 0.92 四级 2019-08-15 -27.56 0.96 四级 -26.17 0.91 四级
下载: 导出CSV
-
Chen, Y. R., 2018. Application of Intelligent Algorithm Based on Genetic Algorithm and Extreme Learning Machine to Deformation Prediction of Foundation Pit. Tunnel Construction, 38(6): 941-947(in Chinese with English abstract). Ghaboussi, J., Sidarta, D. E., 1997. New Method of Material Modeling Using Neural Networks. 6th International Symposium On Numerical Models in Geotechnics, 393-400. Hashash, Y. M. A., Whittle, A. J., 1996. Ground Movement Prediction for Deep Excavations in Soft Clay. Journal of Geotechnical Engineering, 122(6): 474-486. https://doi.org/10.1061/(asce)0733-9410(1996)122:6(474) Li, Q., 2013. Application Research of Generalized Regression Neural Network in Deep Foundation Pit Deformation Monitoring (Dissertation). Central South University, Changsha(in Chinese with English abstract). Li, Y. J., Xue, Y. D., Yue, L., et al., 2015. Displacement Prediction of Deep Foundation Pit Based on Genetic Algorithms and BP Neural Network. Chinese Journal of Underground Space and Engineering, 11(S2): 741-749(in Chinese with English abstract). Ma, L., 2018. Study of Deformation Law of Deep Foundation Pit Based on Chaotic Progressive Prediction Model and Trend Test. Tunnel Construction, 38(6): 934-940(in Chinese with English abstract). Peng, M., 2008. Research and Development of Dynamic Risk Management Software (TRM2.0) for Shield Tunnel Construction (Dissertation). Tongji University, Shanghai(in Chinese with English abstract). Song, C. P., 2019. Improved BP Neural Network Method for Deformation Predication of Deep Excavation. Journal of Civil Engineering and Management, 36(5): 45-49, 55(in Chinese with English abstract). van Staveren, M. T., Peters, T. J. M., 2004. Matching Monitoring, Risk Allocation and Geotechnical Baseline Reports. Engineering Geology for Infrastructure Planning in Europe. Springer, Berlin, Heidelberg : 786-791. https://doi.org/10.1007/978-3-540-39918-6_87 Wang, X., Wu, J., Liu, C., et al., 2018. Exploring LSTM Based Recurrent Neural Network for Failure Time Series Prediction. Journal of Beijing University of Aeronautics and Astronautics, 44(4): 772-784(in Chinese with English abstract). Werbos, P. J., 1990. Backpropagation through Time: What It does and How to do It. Proceedings of the IEEE, 78(10): 1550-1560. https://doi.org/10.1109/5.58337 陈艳茹, 2018. 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. 隧道建设(中英文), 38(6): 941-947. https://www.cnki.com.cn/Article/CJFDTOTAL-JSSD201806012.htm 李钦, 2013. 泛化回归神经网络在深基坑变形监测中的应用研究(硕士学位论文). 长沙: 中南大学. 李彦杰, 薛亚东, 岳磊, 等, 2015. 基于遗传算法-BP神经网络的深基坑变形预测. 地下空间与工程学报, 11(增刊2): 741-749. https://www.cnki.com.cn/Article/CJFDTOTAL-BASE2015S2062.htm 马琳, 2018. 基于混沌递进预测模型与趋势检验的深基坑变形规律研究. 隧道建设(中英文), 38(6): 934-940. https://www.cnki.com.cn/Article/CJFDTOTAL-JSSD201806010.htm 彭铭. 2008. 盾构隧道施工动态风险管理软件(TRM2.0)开发研究(硕士学位论文). 上海: 同济大学. 宋楚平, 2019. 一种改进的BP神经网络深基坑变形预测方法. 土木工程与管理学报, 36(5): 45-49, 55 https://www.cnki.com.cn/Article/CJFDTOTAL-WHCJ201905009.htm 王鑫, 吴际, 刘超, 等, 2018. 基于LSTM循环神经网络的故障时间序列预测. 北京航空航天大学学报, 44(4): 772-784. https://www.cnki.com.cn/Article/CJFDTOTAL-BJHK201804015.htm -
点击查看大图
计量
- 文章访问数: 324
- HTML全文浏览量: 292
- PDF下载量: 35
- 被引次数: 0
