Physics-informed online deep learning for advanced control of shield tail clearance in tunnel construction

Abstract

To more accurately estimate and control the magnitude of the shield tail clearance, a hybrid deep learning model with the integration of an online physics-informed deep neural network (online PDNN) and non-dominated sorting genetic algorithm-II (NSGA-II) is developed. The online PDNN has evolved from a deep learning framework constrained by the underlying physical mechanism of shield tail clearance measurements. The algorithm is used to forecast the shield tail clearance in tunnel boring machines (TBMs). The NSGA-II is employed to conduct the multi-objective optimization (MOO) process for shield tail clearance. The proposed method is validated in a tunnel case in China. Experimental results reveal that: (1) In comparison with some state-of-the-art algorithms, the online PDNN model demonstrates superior capability in predicting shield tail clearance above, upper-left, and upper-right, with R² scores of 0.93, 0.90, and 0.90, respectively; (2) The MOO achieves a comprehensive optimal solution, with the overall improvement percentage of shield tail clearance reaching 30.87% and a hypervolume of 32 under the 20% constraint condition, which surpasses the average performance of other MOO frameworks by 23 and 5.48%, respectively. The novelty of this research lies in coupling the constructed physical constraints and the online update mechanism into a causal analysis-oriented data-driven model, which not only enhances the model’s performance and interpretability but also realizes the control for the shield tail clearance by the integration of NSGA-II.

Graphical abstract

Keywords

physics-informed neural network / online update mechanism / multi-objective optimization / NSGA-II / shield tail clearance

References

[1] Afradi A, Ebrahimabadi A, Hallajian T, (2020). Prediction of TBM penetration rate using support vector machine. Geosaberes: Revista de Estudos Geoeducacionais, 11: 467–479

[2] Chai R, Guo Y, Zuo Z, Chen K, Shin H S, Tsourdos A, (2024). Cooperative motion planning and control for aerial-ground autonomous systems: Methods and applications. Progress in Aerospace Sciences, 146: 101005

[3] ChaiRSavvarisATsourdosAChaiS (2017). Solving multi-objective aeroassisted spacecraft trajectory optimization problems using extended NSGA-II. AIAA SPACE and Astronautics Forum and Exposition: 5193

[4] Chai R, Savvaris A, Tsourdos A, Xia Y, Chai S, (2020). Solving multiobjective constrained trajectory optimization problem by an extended evolutionary algorithm. IEEE Transactions on Cybernetics, 50( 4): 1630–1643

[5] Chai R, Tsourdos A, Savvaris A, Chai S, Xia Y, Philip Chen C, (2021). Multiobjective overtaking maneuver planning for autonomous ground vehicles. IEEE Transactions on Cybernetics, 51( 8): 4035–4049

[6] Chai R, Tsourdos A, Savvaris A, Xia Y, Chai S, (2020). Real-time reentry trajectory planning of hypersonic vehicles: A two-step strategy incorporating fuzzy multiobjective transcription and deep neural network. IEEE Transactions on Industrial Electronics, 67( 8): 6904–6915

[7] Chakraborty S, (2021). Transfer learning based multi-fidelity physics informed deep neural network. Journal of Computational Physics, 426: 109942

[8] Chen J, Liu S, Li W, Liu F, Zheng D, (2021). Design of laser vision measurement system for shield tail clearance of shield tunneling machine. Machine Tool & Hydraulics, 49: 77–80

[9] Chen J, Zhou Z, Liu F, Zheng D, (2020). Research on the visual measurement based on shield tail clearance space structure of the shield machine. Machine Tool & Hydraulics, 48: 116–121

[10] Chen K, Zhou X, Bao Z, Skibniewski M J, Fang W, (2025). Artificial intelligence in infrastructure construction: A critical review. Frontiers of Engineering Management, 12( 1): 24–38

[11] CSTNET (2021). The “Jiangcheng Pioneer” tunnel boring machine surpasses 1,000 rings while tunneling under the Yangtze River. Available at the website of finance.sina.com.cn

[12] ENZAN(2021). Tail Clearance control system [EB/OL]. Available at the website of enzan-k.com

[13] Feng S, Zhou H, Dong H, (2019). Using deep neural network with small dataset to predict material defects. Materials & Design, 162: 300–310

[14] Feng S X, Chen Z Y, Luo H, Wang S Y, Zhao Y F, Liu L P, Ling D S, Jing L J, (2021). Tunnel boring machines (TBM) performance prediction: A case study using big data and deep learning. Tunnelling and Underground Space Technology, 110: 103636

[15] Feng X, Wang P, Liu S, Wei H, Miao Y, Bu S, (2022). Mechanism and law analysis on ground settlement caused by shield excavation of small-radius curved tunnel. Rock Mechanics and Rock Engineering, 55( 6): 3473–3488

[16] Gao X, Shi M, Song X, Zhang C, Zhang H, (2019). Recurrent neural networks for real-time prediction of TBM operating parameters. Automation in Construction, 98: 225–235

[17] Han X, Zhang F, He Y, Zhong H, (2021). Research on deformation treatment and control technology of tail shield of underwater large diameter slurry shield. In: IOP Conference Series: Earth and Environmental ScienceIOP Publishing, 783: 012027

[18] HanY (2014). Selection of construction parameters for general segment of shield tunnel. Theoretical Research in Urban Construction (electronic version): 3446–3447

[19] He C, Xu C, Xiong D, (2021). Development of an automatic measurement device for double laser shield tail gap based on image recognition technology. In: IOP Conference Series: Earth and Environmental ScienceIOP Publishing, 783: 012070

[20] LinPZhangLTiongR (2023). Soil chamber pressure forecasting in tunnel construction using physics-informed deep learning. Available at SSRN 4614636

[21] Liu J, Bao Z, (2021). Construction technology of shield tunnel in extremely-soft fluid plastic stratum. Tunnel Construction, 41: 527–532

[22] Liu J, Shi C, Wang Z, Lei M, Zhao D, Cao C, (2021). Damage mechanism modelling of shield tunnel with longitudinal differential deformation based on elastoplastic damage model. Tunnelling and Underground Space Technology, 113: 103952

[23] Liu L, Zhou W, Gutierrez M, (2023). Physics-informed ensemble machine learning framework for improved prediction of tunneling-induced short-and long-term ground settlement. Sustainability, 15( 14): 11074

[24] Lou P, Li Y, Tang X, Lu S, Xiao H, Zhang Z, (2023). Influence of double-line large-slope shield tunneling on settlement of ground surface and mechanical properties of surrounding rock and segment. Alexandria Engineering Journal, 63: 645–659

[25] Mahmoodzadeh A, Mohammadi M, Hashim Ibrahim H, Nariman Abdulhamid S, Farid Hama Ali H, Mohammed Hasan A, Khishe M, Mahmud H, (2021). Machine learning forecasting models of disc cutters life of tunnel boring machine. Automation in Construction, 128: 103779

[26] Pan Y, Fu X, Zhang L, (2022). Data-driven multi-output prediction for TBM performance during tunnel excavation: An attention-based graph convolutional network approach. Automation in Construction, 141: 104386

[27] Phoon K K, Zhang W, (2023). Future of machine learning in geotechnics. Georisk: Assessment and Management of Risk for Engineered Systems and Geohazards, 17: 7–22

[28] Raissi M, Perdikaris P, Karniadakis G E, (2019). Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378: 686–707

[29] Rao C, Sun H, Liu Y, (2020). Physics-informed deep learning for incompressible laminar flows. Theoretical & Applied Mechanics Letters, 10( 3): 207–212

[30] Roy A M, Bose R, Sundararaghavan V, Arróyave R, (2023). Deep learning-accelerated computational framework based on physics informed neural network for the solution of linear elasticity. Neural Networks, 162: 472–489

[31] Shaban W M, Elbaz K, Zhou A, Shen S L, (2023). Physics-informed deep neural network for modeling the chloride diffusion in concrete. Engineering Applications of Artificial Intelligence, 125: 106691

[32] Shin S, Lee Y, Kim M, Park J, Lee S, Min K, (2020). Deep neural network model with Bayesian hyperparameter optimization for prediction of NO_x at transient conditions in a diesel engine. Engineering Applications of Artificial Intelligence, 94: 103761

[33] Sun L, Zhuang Q, (2016). Experimental study on the measurement device of shield tail gap. Modern Tunnelling Technology, 53: 56–59

[34] Tahmassebi A, Motamedi M, Alavi A H, Gandomi A H, (2022). An explainable prediction framework for engineering problems: Case studies in reinforced concrete members modeling. Engineering Computations, 39( 2): 609–626

[35] Tang L, Kong X, Ling X, Zhao Y, Tang W, Zhang Y, (2022). Deviation correction strategy for the earth pressure balance shield based on shield-soil interactions. Frontiers of Mechanical Engineering, 17( 2): 20

[36] Verma S, Pant M, Snasel V, (2021). A comprehensive review on NSGA-II for multi-objective combinatorial optimization problems. IEEE Access: Practical Innovations, Open Solutions, 9: 57757–57791

[37] VMT(2021). Automatic Tailskin Clearance Measurement System SLuM Ultra. Retrieved from Available at the website of enzank.com

[38] WMG(2021). Infrastructure dragon! The “Wuhan-made” super-large diameter tunnel boring machine “Jiangcheng Pioneer” rolled off the production line. Available at the website of baijiahao.baidu.com

[39] Xu Z, Wang W, Lin P, Nie L, Wu J, Li Z, (2021). Hard-rock TBM jamming subject to adverse geological conditions: Influencing factor, hazard mode and a case study of Gaoligongshan Tunnel. Tunnelling and Underground Space Technology, 108: 103683

[40] Yang H, (2021). Computer vision shield interval automatic detection device development. Building Technology, 5: 24–27

[41] Yavas M S, Gao Z, Mekaoui N, Saito T, (2023). A machine learning-based hybrid seismic analysis of a lead rubber bearing isolated building specimen. Soil Dynamics and Earthquake Engineering, 174: 108217

[42] Yin X, Liu Q, Huang X, Pan Y, (2022). Perception model of surrounding rock geological conditions based on TBM operational big data and combined unsupervised-supervised learning. Tunnelling and Underground Space Technology, 120: 104285

[43] Yu W, (2022). Analysis of shield tail clearance control. Journal of East China Jiaotong University, 39: 47–53 (in Chinese)

[44] Zhang L, Guo J, Fu X, Tiong R L K, Zhang P, (2024a). Digital twin enabled real-time advanced control of TBM operation using deep learning methods. Automation in Construction, 158: 105240

[45] Zhang L, Lin P, (2021). Multi-objective optimization for limiting tunnel-induced damages considering uncertainties. Reliability Engineering & System Safety, 216: 107945

[46] ZhangLWangYFuXSongXLinP (2024b). Geological risk prediction under uncertainty in tunnel excavation using online learning and hidden Markov model. Frontiers of Engineering Management, 1–20

[47] Zhang P, Li H, Ha Q P, Yin Z-Y, Chen R-P, (2020a). Reinforcement learning based optimizer for improvement of predicting tunneling-induced ground responses. Advanced Engineering Informatics, 45: 101097

[48] Zhang P, Wu H N, Chen R P, Dai T, Meng F Y, Wang H B, (2020b). A critical evaluation of machine learning and deep learning in shield-ground interaction prediction. Tunnelling and Underground Space Technology, 106: 103593

[49] Zhang Z, Pan Q, Yang Z, Yang X, (2023). Physics-informed deep learning method for predicting tunnelling-induced ground deformations. Acta Geotechnica, 18( 9): 4957–4972

[50] Zhou H, Wang H, Zeng W, (2018). Smart construction site in mega construction projects: A case study on island tunneling project of Hong Kong−Zhuhai−Macao Bridge. Frontiers of Engineering Management, 5( 1): 78–87

RIGHTS & PERMISSIONS

Higher Education Press