Multi-Step Prediction of Excavation Deformation of Subway Station Based on Intelligent Algorithm

doi:10.16183/j.cnki.jsjtu.2022.419

Abstract

Abstract:

To better predict the lateral displacements of diaphragm walls during deep excavation, a long short-term memory (LSTM) multi-step prediction model is developed in this paper based on the LSTM algorithm. First, the multi-output strategy of multi-step prediction model is discussed. Then, the construction method of the LSTM multi-step prediction model is introduced in detail, and the two hyperparameters, i.e., the space and time dimensions of the model input set, are explored to improve the prediction accuracy of the model. Finally, the errors between the predicted values and the field monitoring data are analyzed based on an excavation project buried in water-rich sandy strata. The analysis results of three typical monitoring points indicate that the LSTM prediction model is characterized by solid generalization ability, and the relevant algorithm is practically helpful for improving and optimizing deformation prediction methods of deep excavation.

Key words: excavation, deformation prediction, long short-term memory (LSTM) intelligent algorithm, multi-step prediction model

CLC Number:

TU46+3

LIU Juncheng, TAN Yong, ZHANG Shengjie. Multi-Step Prediction of Excavation Deformation of Subway Station Based on Intelligent Algorithm[J]. Journal of Shanghai Jiao Tong University, 2024, 58(7): 1108-1117.

Figures/Tables 15

Tab.1

Tab.2

Multi-output strategy

数据样本	输入	输出
训练-1	[x₁ x₂ x₃ x₄]	[x₅ x₆ x₇]
训练-2	[x₂ x₃ x₄ x₅]	[x₆ x₇ x₈]
训练-3	[x₃ x₄ x₅ x₆]	[x₇ x₈ x₉]
预测	[x₆ x₇ x₈ x₉]	[ $x^10$ $x^11$ $x^12$ ]

Tab.2

Fig.1

Fig.2

Tab.3

Fig.3

Fig.4

Tab.4

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Fig.10

Fig.11

References 19

[1]	ZHENG G, YANG X, ZHOU H, et al. A simplified prediction method for evaluating tunnel displacement induced by laterally adjacent excavations[J]. Computers and Geotechnics, 2018, 95: 119-128.
[2]	ZHENG G, DU Y, CHENG X, et al. Characteristics and prediction methods for tunnel deformations induced by excavations[J]. Geomechanics and Engineering, 2017, 12(3): 361-397.
[3]	OU C Y, TENG F, LI C W. A simplified estimation of excavation-induced ground movements for adjacent building damage potential assessment[J]. Tunnelling and Underground Space Technology, 2020, 106: 103561.
[4]	HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. doi: 10.1162/neco.1997.9.8.1735 pmid: 9377276
[5]	SUTSKEVER I, VINYALS O, LE Q V. Sequence to sequence learning with neural networks[J]. Advances in Neural Information Processing Systems, 2014, 27: 3104-3112.
[6]	XIE P, ZHOU A, CHAI B. The application of long short-term memory (LSTM) method on displacement prediction of mu.pngactor-induced landslides[J]. IEEE Access, 2019, 7: 54305-54311.
[7]	ZHANG P, WU H N, CHEN R P, et al. A critical evaluation of machine learning and deep learning in shield-ground interaction prediction[J]. Tunnelling and Underground Space Technology, 2020, 106: 103593.
[8]	钱建固, 吴安海, 季军, 等. 基于小波优化LSTM-ARMA 模型的岩土工程非线性时间序列预测[J]. 同济大学学报(自然科学版), 2021, 49(8): 1107-1115.
	QIAN Jiangu, WU Anhai, JI Jun, et al. Nonlinear time series prediction of geotechnical engineering based on wavelet optimization LSTM-ARMA model[J]. Journal of Tongji University (Natural Science), 2021, 49(8): 1107-1115.
[9]	赵华菁, 张名扬, 刘维, 等. 基于神经网络算法的深基坑地连墙变形动态预测[J]. 地下空间与工程学报, 2021,17 (Sup.1):321-327.
	ZHAO Huajing, ZHANG Mingyang, LIU Wei, et al. Dynamic prediction of deformation of ground connected wall in deep foundation pit based on neural network algorithm[J]. Chinese Journal of Underground Space and Engineering, 2021, 17 (Sup.1): 321-327.
[10]	AHMED A, KHALID M. Multi-step ahead wind forecasting using nonlinear autoregressive neural networks[J]. Energy Procedia, 2017, 134: 192-204.
[11]	SHAO X, KIM C S. Multi-step short-term power consumption forecasting using multi-channel LSTM with time location considering customer behavior[J]. IEEE Access, 2020, 8: 125263-125273.
[12]	HAAS A, ROSSBERG A, SCHUFF D L, et al. Bringing the web up to speed with WebAssembly[C]//Acm Sigplan Conference on Programming Language Design & Implementation. Barcelona, Spain:ACM, 2017: 185-200.
[13]	RICE L, WONG E, KOLTER Z. Overfitting in adversarially robust deep learning[C]//International Conference on Machine Learning. Vienna, Austria: PMLR, 2020: 8093-8104.
[14]	MASSERAN N, RAZALI A M, IBRAHIM K, et al. Modeling air quality in main cities of Peninsular Malaysia by using a generalized Pareto model[J]. Environmental Monitoring and Assessment, 2016, 188(1): 1-12.
[15]	JAIS I K M, ISMAIL A R, NISA S Q. Adam optimization algorithm for wide and deep neural network[J]. Knowledge Engineering and Data Science, 2019, 2(1): 41-46.
[16]	BAI C L, ZHAO H. Generalized extended tanh-function method and its application[J]. Chaos, Solitons & Fractals, 2006, 27(4): 1026-1035.
[17]	SMITH L N. Cyclical learning rates for training neural networks[C]//2017 IEEE Winter Conference on Applications of Computer Vision (WACV). Santa Rosa, CA, USA: IEEE, 2017: 464-472.
[18]	TAN Y, WANG D. Characteristics of a large-scale deep foundation pit excavated by the central-island technique in Shanghai soft clay. I: Bottom-up construction of the central cylindrical shaft[J]. Journal of Geotechnical and Geoenvironmental Engineering, 2013, 139(11): 1875-1893.
[19]	丁勇春, 王建华, 徐中华, 等. 上海软土地区地铁车站深基坑的变形特性[J]. 上海交通大学学报, 2008, 42(11): 1871-1875.
	DING Yongchun, WANG Jianhua, XU Zhonghua, et al. Deformation characteristics of deep excavations for metro stations in Shanghai soft soil deposits[J]. Journal of Shanghai Jiao Tong University, 2008, 42(11): 1871-1875.

试验编号	S	T	n	α
1-a	1	5	32	0.001
1-b	2	5	32	0.001
1-c	3	5	32	0.001
1-d	4	5	32	0.001
1-e	5	5	32	0.001
1-f	6	5	32	0.001

试验编号	S	T	n	α
2-a	2	2	32	0.001
2-b	2	3	32	0.001
2-c	2	4	32	0.001
2-d	2	5	32	0.001
2-e	2	6	32	0.001
2-f	2	7	32	0.001