Longitudinal deformation response of existing tunnel to upper deep excavation based on LAAF-PINN

Abstract

Deformation of existing tunnels induced by adjacent deep excavation is a key construction concern. This paper constructed a Layer-wise locally adaptive activation functions physics-informed neural networks (LAAF-PINN) model, driven by physical laws of a two-stage theoretical model, to predict the deformation response of an existing tunnel to deep excavation. The precision of the solution is improved by an enhanced training on poorly convergent regions in the basis of initial model training. The proposed LAAF-PINN model does not require differential processing as used for traditional differential algorithms to outputs continuous longitudinal deformation response, and moreover, the model can accurately predict bending moment value without prior data training. Parametric analysis show that using the Swish adaptive activation function and learning rate decay strategy can reduce the loss value by at least 10 times compared to other strategies. Furthermore, a local enhancement training can effectively mitigate local convergence issues and enhance the prediction accuracy, which means the range of loss value of the physical law differential equations in the region with poor convergence was reduced about 25 times. The proposed method, verified by field measurements, shows the feasibility of intelligent real time deformation prediction for deep excavation in the proximity to existing tunnels.

Keywords

tunnel engineering / deformation response / LAAF-PINN / deep excavation / existing tunnel

References

[1] Chang C-t, Sun C W, Duann S W, et al. . Response of a Taipei Rapid Transit System (TRTS) tunnel to adjacent excavation [J]. Tunnelling and Underground Space Technology, 2001, 16(3): 151-158

[2] Chen R-p, Meng F-y, Li Z-c, et al. . Investigation of response of metro tunnels due to adjacent large excavation and protective measures in soft soils [J]. Tunnelling and Underground Space Technology, 2016, 58: 224-235

[3] Ali Soomro M, Ali Mangnejo D, Saand A, et al. . Influence of stress relief due to deep excavation on a brick masonry wall: 3D numerical predictions [J]. European Journal of Environmental and Civil Engineering, 2022, 26(15): 7621-7644

[4] Feng G-h, Xu C-j, Zheng M-w, et al. . Study of longitudinal deformation of existing underlying tunnel due to above excavation considering shear deformation of tunnel [J]. Journal of the China Railway Society, 2022, 44(3): 132-141(in Chinese)

[5] Feng G-h, Chen G-z, Zhang D, et al. . Analytical solution on uplift deflection of underlying existing tunnel induced by foundation pit excavation [J]. Journal of Railway Science and Engineering, 2023, 20(10): 3908-3917(in Chinese)

[6] Liang R-z, Xia T-d, Hong Y, et al. . Effects of above-crossing tunnelling on the existing shield tunnels [J]. Tunnelling and Underground Space Technology, 2016, 58: 159-176

[7] Liang R-z, Xia T-d, Huang M-s, et al. . Simplified analytical method for evaluating the effects of adjacent excavation on shield tunnel considering the shearing effect [J]. Computers and Geotechnics, 2017, 81: 167-187

[8] Wang L-x, Liang R-z, Li Z-c, et al. . Heave deformation of existing shield tunnel induced by over-crossing excavation [J]. Engineering Mechanics, 2022, 39(12): 130-140(in Chinese)

[9] Zhang Z-g, Zhang M-x, Wang W-dong. Two-stage method for analyzing effects on adjacent metro tunnels due to foundation pit excavation [J]. Rock and Soil Mechanics, 2011, 32(7): 2085-2092(in Chinese)

[10] Tanoli A Y, Yan B, Xiong Y-l, et al. . Numerical analysis on zone-divided deep excavation in soft clays using a new small strain elasto - plastic constitutive model [J]. Underground Space, 2022, 7(1): 19-36

[11] Xiao X, Li M-g, Wang J-h, et al. . Numerical evaluation of control measures for tunnel deformation induced by an oversized deep excavation [J]. Journal of Aerospace Engineering, 2018, 31(6): 04018109

[12] Zheng G, Yang X-y, Zhou H-z, et al. . A simplified prediction method for evaluating tunnel displacement induced by laterally adjacent excavations [J]. Computers and Geotechnics, 2018, 95: 119-128

[13] Du Y-m, Wang B-y, Diao Y, et al. . Centrifuge modelling of the impact of excavation with partition piles on adjacent existing tunnel [J]. Applied Sciences, 2023, 13(24): 13064

[14] Huang X, Huang H-w, Zhang D-mei. Centrifuge modelling of deep excavation over existing tunnels [J]. Proceedings of the Institution of Civil Engineers - Geotechnical Engineering, 2014, 167(1): 3-18

[15] Meng F-y, Chen R-p, Liu S-l, et al. . Centrifuge modeling of ground and tunnel responses to nearby excavation in soft clay [J]. Journal of Geotechnical and Geoenvironmental Engineering, 2021, 147(3): 04020178

[16] Yu Z-t, Wang H-y, Wang W-j, et al. . Experimental and numerical investigation on the effects of foundation pit excavation on adjacent tunnels in soft soil [J]. Mathematical Problems in Engineering, 2021, 2021(1): 5587857

[17] Zhang D-m, Xie X-c, Li Z-l, et al. . Simplified analysis method for predicting the influence of deep excavation on existing tunnels [J]. Computers and Geotechnics, 2020, 121: 103477

[18] Zhang D-m, Shen Y-m, Huang Z-k, et al. . Auto machine learning-based modelling and prediction of excavation-induced tunnel displacement [J]. Journal of Rock Mechanics and Geotechnical Engineering, 2022, 14(4): 1100-1114

[19] Jong S C, Ong D E L, Oh E. State-of-the-art review of geotechnical-driven artificial intelligence techniques in underground soil-structure interaction [J]. Tunnelling and Underground Space Technology, 2021, 113: 103946

[20] Zhang W-g, Zhang R-h, Wu C-z, et al. . State-of-the-art review of soft computing applications in underground excavations [J]. Geoscience Frontiers, 2020, 11(4): 1095-1106

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

[22] Li X-t, Zhu J-song. Identification of bridge influence line and multiple-vehicle loads based on physics-informed neural networks [J]. Structural Health Monitoring, 2025, 24(2): 1167-1186

[23] Guo X-y, Fang S-gen. A physics-informed autoencoder based cable force identification framework for longspan bridges [J]. Structures, 2024, 60: 105906

[24] Cao X-c, Cai Y-f, Li Y-c, et al. . Intelligent vehicle trajectory tracking control based on physics-informed neural network dynamics model [J]. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 2025, 239(7): 2315-2331

[25] Zhang Z-l, Pan Q-j, Yang Z-h, et al. . Physics-informed deep learning method for predicting tunnelling-induced ground deformations [J]. Acta Geotechnica, 2023, 18(9): 4957-4972

[26] Wang G-k, Shan Y, Detmann B, et al. . Physics-Informed Neural Network (PINN) model for predicting subgrade settlement induced by shield tunnelling beneath an existing railway subgrade [J]. Transportation Geotechnics, 2024, 49: 101409

[27] Cai Q-p, Elbaz K, Guo X-y, et al. . Physics-informed deep learning and analytical patterns for predicting deformations of existing tunnels induced by new tunnelling [J]. Computers and Geotechnics, 2025, 187: 107451

[28] Wang G, Fang Q, Wang J, et al. . Estimation of load for tunnel lining in elastic soil using physics-informed neural network [J]. Computer-Aided Civil and Infrastructure Engineering, 2024, 39(17): 2701-2718

[29] Raissi M, Wang Z-c, Triantafyllou M S, et al. . Deep learning of vortex-induced vibrations [J]. Journal of Fluid Mechanics, 2019, 861: 119-137

[30] Jagtap A D, Kawaguchi K, Em Karniadakis G. Locally adaptive activation functions with slope recovery for deep and physics-informed neural networks [J]. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2020, 476(2239): 20200334

[31] Jagtap A D, Kawaguchi K, Karniadakis G E. Adaptive activation functions accelerate convergence in deep and physics-informed neural networks [J]. Journal of Computational Physics, 2020, 404: 109136

[32] Zhang Z-g, Huang M-s, Xu C, et al. . Simplified solution for tunnel-soil-pile interaction in Pasternak’s foundation model [J]. Tunnelling and Underground Space Technology, 2018, 78: 146-158

[33] Li K-p, Chen S-h, Pei R-p, et al. . Theoretical study on diaphragm wall and surface deformation due to foundation excavation based on three-parameter Kerr model [J]. Sustainability, 2024, 16(6): 2295

[34] Feng G-h, Chen Q-s, Xu C-j, et al. . Improved theoretical solutions for estimating the tunnel response induced by overlying excavation [J]. Sustainability, 2023, 15(3): 2589

[35] Chortis F, Kavvadas M. Three-dimensional numerical analyses of perpendicular tunnel intersections [J]. Geotechnical and Geological Engineering, 2021, 39(3): 1771-1793

[36] Zhao K, Lu Y-j, Wang Y-z, et al. . Investigations on the spatial end effect of a subsea shield tunnel and the aseismic measures [J]. Journal of Vibration and Shock, 2022, 41(16): 33-42(in Chinese)

[37] Li J-ping. Numerical analysis of the influence of foundation pit excavation unloading on the underlying subway tunnel [J]. Chinese Journal of Underground Space and Engineering, 2009, 5(S1): 1345-13481360. (in Chinese)

[38] Shiba Y, Kawashima K, Obinata N, et al. . An evaluation method of longitudinal stiffness of shield tunnel linings for application to seismic response analyses [J]. Doboku Gakkai Ronbunshu, 1988, 1988(398): 319-327

[39] Wu H-n, Shen S-l, Liao S-m, et al. . Longitudinal structural modelling of shield tunnels considering shearing dislocation between segmental rings [J]. Tunnelling and Underground Space Technology, 2015, 50: 317-323

RIGHTS & PERMISSIONS

Central South University