A MULTI-FIDELITY DEEP OPERATOR FRAMEWORK WITH GENERATIVE ADVERSARIAL NETWORK FOR LOAD PREDICTION AND DISPLACEMENT RECONSTRUCTION OF TUNNEL STRUCTURES

19th WORLD CONFERENCE OF THE ASSOCIATED RESEARCH CENTRES FOR THE URBAN UNDERGROUND SPACE, Belgrade, Serbia, November 4-7, 2025. (Paper No: 7.4.197,  pp. 948-955)

 

АУТОР(И) / AUTHOR(S): Chen Xu, Zhen Liu, Ba Trung Cao, Günther Meschke, Yong Yuan, Xian Liu

 

Download Full Pdf   

DOI:  10.46793/ACUUS2025.7.4.197

САЖЕТАК / ABSTRACT:

Accurately predicting external loads and evaluating the health condition of tunnel structures are fundamental for ensuring the safety and resilience of underground infrastructure. This study presents an advanced multi-fidelity Deep Operator Network (MF-DeepONet) framework that incorporates a generative adversarial network (GAN) to address these challenges by integrating simulation-based low-fidelity data and sparse high-fidelity real-world measurements. Within this framework, the GAN is employed for inverse prediction of external loads from measured displacements, ensuring that the resulting load distributions remain consistent with simulation-informed patterns. The predicted loads are then fed into the MF-DeepONet to reconstruct the full displacement field through data fusion. The MF-DeepONet framework consists of two subnetworks: a low-fidelity network trained on data generated by a validated macro-scale numerical model to capture general deformation patterns of tunnel linings, and a high-fidelity network trained on limited real-world monitoring data to learn the correlations between field observations and simulations. A full-scale test on a quasi-rectangular shield tunnel (QRST) lining structure is conducted for validation. The results demonstrate that the proposed framework enables both reliable estimation of external loads and accurate reconstruction of displacement fields, showing strong agreement with experimental observations. This data-driven approach significantly reduces reliance on dense sensing networks and offers a robust, interpretable solution for tunnel health diagnosis and predictive maintenance in real-world engineering applications.

КЉУЧНЕ РЕЧИ / KEYWORDS:

Tunnel lining, Multi-fidelity DeepONet framework, Generative adversarial network (GAN), Limited measurements, Inverse load determination, Health evaluation

ПРОЈЕКАТ / ACKNOWLEDGEMENT:

ЛИТЕРАТУРА / REFERENCES:

  • Ahmed, B., Qiu, Y., Abueidda, D. W., El-Sekelly, W., de Soto, B. G., Abdoun, T., & Mobasher, M. E. (2025). Physics-informed deep operator networks with stiffness-based loss functions for structural response prediction. Engineering Applications of Artificial Intelligence, 144, 110097. doi:https://doi.org/10.1016/j.engappai.2025.110097
  • Chinese Standard, G. (2002). 50010-2010,“Code for Design of Concrete Structures,”. In: Chinese Building Press, Beijing, China.
  • Colombo, L., Gattinoni, P., & Scesi, L. (2018). Stochastic modelling of groundwater flow for hazard assessment along the underground infrastructures in Milan (northern Italy). Tunnelling and Underground Space Technology, 79, 110-120. doi:https://doi.org/10.1016/j.tust.2018.05.007
  • Gao, B.-Y., Chen, R.-P., Wu, H.-N., Chen, T., & Zhang, Y. (2025). Enhancement mechanism of UHPC secondary lining on the mechanical performance at joints of shield tunnels: Full-scale experiment and simulation. Tunnelling and Underground Space Technology, 157, 106282. doi:https://doi.org/10.1016/j.tust.2024.106282
  • Garg, S., Gupta, H., & Chakraborty, S. (2022). Assessment of DeepONet for time dependent reliability analysis of dynamical systems subjected to stochastic loading. Engineering Structures, 270, 114811. doi:https://doi.org/10.1016/j.engstruct.2022.114811
  • Guo, C., Qi, J., Shi, L., & Fang, Q. (2018). Reasonable overburden thickness for underwater shield tunnel. Tunnelling and Underground Space Technology, 81, 35-40. doi:https://doi.org/10.1016/j.tust.2018.06.016
  • He, J., Koric, S., Kushwaha, S., Park, J., Abueidda, D., & Jasiuk, I. (2023). Novel DeepONet architecture to predict stresses in elastoplastic structures with variable complex geometries and loads. Computer Methods in Applied Mechanics and Engineering, 415, 116277. doi:https://doi.org/10.1016/j.cma.2023.116277
  • Huang, H., Ruan, B., Wu, X., & Qin, Y. (2024). Parameterized modeling and safety simulation of shield tunnel based on BIM-FEM automation framework. Automation in Construction, 162, 105362. doi:https://doi.org/10.1016/j.autcon.2024.105362
  • Huang, X., Liu, W., Zhang, Z., Zhuang, Q., Zhu, Y., Wang, Q., Kwok, C. y., & Wang, S. (2020). Structural behavior of segmental tunnel linings for a large stormwater storage tunnel: insight from full-scale loading tests. Tunnelling and Underground Space Technology, 99, 103376. doi:https://doi.org/10.1016/j.tust.2020.103376
  • Koric, S., Viswantah, A., Abueidda, D. W., Sobh, N. A., & Khan, K. (2024). Deep learning operator network for plastic deformation with variable loads and material properties. Engineering with Computers, 40(2), 917-929. doi:https://doi.org/10.1007/s00366-023-01822-x
  • Latif, K., Sharafat, A., & Seo, J. (2023). Digital twin-driven framework for TBM performance prediction, visualization, and monitoring through machine learning. Applied Sciences, 13(20), 11435. doi:https://doi.org/10.3390/app132011435
  • Lin, F., & Yang, X. (2023). Mechanical properties analysis of joints for prefabricated metro station structure based on Whole-process in-situ monitoring. Tunnelling and Underground Space Technology, 138, 105202. doi:https://doi.org/10.1016/j.tust.2023.105202
  • Liu, X., Hong, J., & Liu, Z. (2025). Investigations on mechanical behavior of longitudinal joints in segmental tunnel linings reinforced with epoxy bonded-bolted steel plates. Tunnelling and Underground Space Technology, 163, 106724. doi:https://doi.org/10.1016/j.tust.2025.106724
  • Liu, X., Liu, Z., Ye, Y., Bai, Y., & Zhu, Y. (2018). Mechanical behavior of quasi-rectangular segmental tunnel linings: Further insights from full-scale ring tests. Tunnelling and Underground Space Technology, 79, 304-318. doi:https://doi.org/10.1016/j.tust.2018.05.016
  • Liu, X., Liu, Z., Yuan, Y., & Zhu, Y. (2018). Quasi-Rectangular Shield Tunneling Technology in the Ningbo Rail Transit Project. In High Tech Concrete: Where Technology and Engineering Meet (pp. 2765-2773): Springer.
  • Liu, X., Ye, Y., Liu, Z., & Huang, D. (2018). Mechanical behavior of Quasi-rectangular segmental tunnel linings: First results from full-scale ring tests. Tunnelling and Underground Space Technology, 71, 440-453. doi:https://doi.org/10.1016/j.tust.2017.09.019
  • Liu, X., Zhang, Y., Bao, Y., & Song, W. (2022). Investigation of the structural effect induced by stagger joints in segmental tunnel linings: Numerical explanation via macro-level structural modeling. Tunnelling and Underground Space Technology, 120, 104284. doi:https://doi.org/10.1016/j.tust.2021.104284
  • Liu, Z., Cao, B. T., Xu, C., Liu, X., Yuan, Y., & Meschke, G. (2025). Mechanics of longitudinal joints in segmental tunnel linings: Role of connecting bolts. Tunnelling and Underground Space Technology, 161, 106601. doi:https://doi.org/10.1016/j.tust.2025.106601
  • Liu, Z., Liu, X., Alsahly, A., Yuan, Y., & Meschke, G. (2025). Mechanics of longitudinal joints in segmental tunnel linings: A semi-analytical approach. Tunnelling and Underground Space Technology, 163, 106696. doi:https://doi.org/10.1016/j.tust.2025.106696
  • Lu, L., Jin, P., Pang, G., Zhang, Z., & Karniadakis, G. E. (2021). Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators. Nature machine intelligence, 3(3), 218-229. doi:https://doi.org/10.1038/s42256-021-00302-5
  • Lu, W., Meng, L., Li, S., Xu, Y., Wang, L., Zhang, P., Dou, S., & Sun, H. (2023). Study on progressive failure behavior and mechanical properties of tunnel arch support structures. Tunnelling and Underground Space Technology, 140, 105285. doi:https://doi.org/10.1016/j.tust.2023.105285
  • Qiu, Y., Hu, X., Walton, G., He, C., He, C., & Ju, J. W. (2021). Full scale tests and a progressive failure model to simulate full mechanical behavior of concrete tunnel segmental lining joints. Tunnelling and Underground Space Technology, 110, 103834. doi:https://doi.org/10.1016/j.tust.2021.103834
  • Soga, K., Chen, Z., Hubbard, P. G., Ford, C., Taha, M. R., Heras Murcia, D., Tang, P., Glisic, B., Ozbulut, O., & Lu, N. (2025). Evaluating Contributions of Emerging Technologies to Civil Infrastructure System Resilience. II: Case Studies. ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering, 11(3), 04025032. doi:https://doi.org/10.1061/ajrua6.rueng-1495
  • Wu, Y., Crawford, J. E., & Magallanes, J. M. (2012). Performance of LS-DYNA concrete constitutive models. Paper presented at the 12th International LS-DYNA users conference.
  • Xu, C., Cao, B. T., Yuan, Y., & Meschke, G. (2023). Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under different loading scenarios. Computer Methods in Applied Mechanics and Engineering, 405, 115852. doi:https://doi.org/10.1016/j.cma.2022.115852
  • Xu, C., Cao, B. T., Yuan, Y., & Meschke, G. (2024). A multi-fidelity deep operator network (DeepONet) for fusing simulation and monitoring data: Application to real-time settlement prediction during tunnel construction. Engineering Applications of Artificial Intelligence, 133, 108156. doi:https://doi.org/10.1016/j.engappai.2024.108156
  • Xu, C., Liu, Z., Cao, B., Meschke, G., & Liu, X. (2024). Multifidelity operator learning for predicting displacement fields of tunnel linings under external loads. Paper presented at the IOP Conference Series: Earth and Environmental Science.
  • Yuan, Q., Liang, F., & Fang, Y. (2021). Numerical simulation and simplified analytical model for the longitudinal joint bending stiffness of a tunnel considering axial force. Structural Concrete, 22(6), 3368-3384. doi:https://doi.org/10.1002/suco.202000796
  • Zhang, J.-L., Yuan, Y., Liu, X., Mang, H. A., & Pichler, B. L. (2023). Quantification of the safety against groundwater ingress through longitudinal joints of segmental tunnel linings by means of convergences. Tunnelling and Underground Space Technology, 136, 105102. doi:https://doi.org/10.1016/j.tust.2023.105102
  • Zhang, W., Liu, X., Liu, Z., Zhu, Y., Huang, Y., Taerwe, L., & De Corte, W. (2022). Investigation of the pressure distributions around quasi-rectangular shield tunnels in soft soils with a shallow overburden: A field study. Tunnelling and Underground Space Technology, 130, 104742. doi:https://doi.org/10.1016/j.tust.2022.104742
  • Zhang, X., Zhu, H., Jiang, X., & Broere, W. (2024). Distributed fiber optic sensors for tunnel monitoring: A state-of-the-art review. Journal of Rock Mechanics and Geotechnical Engineering. doi:https://doi.org/10.1016/j.jrmge.2024.01.008
  • Zhao, H., Liu, X., Bao, Y., & Yuan, Y. (2017). Nonlinear simulation of tunnel linings with a simplified numerical modelling. Structural Engineering and Mechanics, 61(5), 593-603. doi:https://doi.org/10.12989/sem.2017.61.5.593