APPLICATION OF IMPROVED MACHINE LEARNING METHODS TOWARD BETTER ACCURACIES IN PREDICTING TBM PERFORMANCE

19th WORLD CONFERENCE OF THE ASSOCIATED RESEARCH CENTRES FOR THE URBAN UNDERGROUND SPACE, Belgrade, Serbia, November 4-7, 2025. (Paper No: 5.7.76,  pp. 844-852)

 

AUTOR(I) / AUTHOR(S): Dansheng Yao, Mengqi Zhu, Hehua Zhu

 

Download Full Pdf   

DOI:  10.46793/ACUUS2025.5.7.76

SAŽETAK / ABSTRACT:

Tunnel Boring Machines (TBMs) play a pivotal role in modern underground construction, especially for tunneling through complex geological environments. Accurate real-time classification of surrounding rock masses is critical for optimizing TBM operation, yet conventional methods relying on engineering labels often suffer from subjectivity and label noise. To address these challenges, we propose a novel semi-supervised teacher-student framework that integrates engineering expertise, class prototype learning, and a Graph Neural Network (GNN)-based self-distillation mechanism. The model leverages soft labels, generated from domain-specific knowledge, to incorporate label uncertainty while improving feature cohesion. By employing self-supervised learning and iterative refinement of class prototypes, the framework effectively minimizes label noise, enhances classification boundaries, and adapts to unseen data. Experimental results on a real-world TBM tunneling dataset demonstrate that the proposed method significantly improves classification performance, offering more reliable and robust predictions in the presence of noisy labels. This work lays the foundation for future advancements in AI-driven geological inference, providing enhanced safety, efficiency, and reliability for TBM tunneling projects.

KLJUČNE REČI / KEYWORDS:

Tunnel Boring Machine, Rock Mass Classification, Label Noise Correction, Self-Distillation, Prototype Learning

PROJEKAT / ACKNOWLEDGEMENT:

LITERATURA / REFERENCES:

  • Adeli, H. (2001). Neural Networks in Civil Engineering: 1989–2000. Computer-Aided Civil and Infrastructure Engineering, 16(2), 126–142. https://doi.org/10.1111/0885-9507.00219
  • Ambah, E., and Elmo, D. (2024, June 23). Is there a Universal Rock Mass Classification System? Presented at the 58th U.S. Rock Mechanics/Geomechanics Symposium. OnePetro. https://doi.org/10.56952/ARMA-2024-0784
  • Azad, M. A., Najeh, T., Raina, A. K., Singh, N., Ansari, A., Ali, M., … Singh, S. K. (2024). Development of correlations between various engineering rockmass classification systems using railway tunnel data in Garhwal Himalaya, India. Scientific Reports, 14(1), 10716. https://doi.org/10.1038/s41598-024-60289-y
  • Fu, K., Qiu, D., Xue, Y., Shao, T., and Lan, G. (2024). TBM tunneling strata automatic identification and working conditions decision support. Automation in Construction, 163, 105425. https://doi.org/10.1016/j.autcon.2024.105425
  • Gangrade, R., Grasmick, J., Trainor-Guitton, W., & Mooney, M. (2022). Risk-based methodology to optimize geotechnical site investigations in tunnel projects. Tunnelling and Underground Space Technology, 127, 104589.
  • Hansen, T. F., Erharter, G. H., and Marcher, T. (2024). Towards reinforcement learning—Driven TBM cutter changing policies. Automation in Construction, 165, 105505. https://doi.org/10.1016/j.autcon.2024.105505
  • Jiang, S., and Zhang, J. (2020). Real-time crack assessment using deep neural networks with wall-climbing unmanned aerial system. Computer-Aided Civil and Infrastructure Engineering, 35(6), 549–564. https://doi.org/10.1111/mice.12519
  • Li, Z., Tao, Y., Du, Y., and Wang, X. (2024). Classification and Prediction of Rock Mass Boreability Based on Daily Advancement during TBM Tunneling. Buildings, 14(7), 1893. https://doi.org/10.3390/buildings14071893
  • Mooney, M. A., Grasmick, J. G., & Gangrade, R. (2023). Methods for local big data integration to reduce geotechnical uncertainty and risk on subsurface infrastructure projects. In Expanding Underground-Knowledge and Passion to Make a Positive Impact on the World(pp. 2799-2806). CRC Press.
  • Perslev, M., Jensen, M. H., Darkner, S., Jennum, P. J., and Igel, C. (2019). U-Time: A fully convolutional network for time series segmentation applied to sleep staging. In Proceedings of the 33rd International Conference on Neural Information Processing Systems (pp. 4415–4426). Red Hook, NY, USA: Curran Associates Inc.
  • Qiu, D., Fu, K., Xue, Y., Tao, Y., Kong, F., and Bai, C. (2022). TBM Tunnel Surrounding Rock Classification Method and Real-Time Identification Model Based on Tunneling Performance. International Journal of Geomechanics, 22(6), 04022070. https://doi.org/10.1061/(ASCE)GM.1943-5622.0002379
  • Rehman, H., Ali, W., Naji, A. M., Kim, J., Abdullah, R. A., and Yoo, H. (2018). Review of Rock-Mass Rating and Tunneling Quality Index Systems for Tunnel Design: Development, Refinement, Application and Limitation. Applied Sciences, 8(8), 1250. https://doi.org/10.3390/app8081250
  • Wang, X., Zhu, H., Zhu, M., Zhang, L., and Ju, J. W. (2021). An integrated parameter prediction framework for intelligent TBM excavation in hard rock. Tunnelling and Underground Space Technology, 118, 104196. https://doi.org/10.1016/j.tust.2021.104196
  • Xiao, H., Hong, Y., Dong, L., Yan, D., Zhuang, J., Xiong, J., … Peng, C. (2024). Multi-Level Label Correction by Distilling Proximate Patterns for Semi-supervised Semantic Segmentation. IEEE Transactions on Multimedia, 26, 8077–8087. https://doi.org/10.1109/TMM.2024.3374594
  • Xue, Y.-D., Luo, W., Chen, L., Dong, H.-X., Shu, L.-S., and Zhao, L. (2023). An intelligent method for TBM surrounding rock classification based on time series segmentation of rock-machine interaction data. Tunnelling and Underground Space Technology, 140, 105317. https://doi.org/10.1016/j.tust.2023.105317
  • Yang, B., Mitelman, A., Elmo, D., and Stead, D. (2021). Why the future of rock mass classification systems requires revisiting their empirical past. Quarterly Journal of Engineering Geology and Hydrogeology, 55(1), qjegh2021-039. https://doi.org/10.1144/qjegh2021-039
  • Yu, Honggan, Tao, J., Qin, C., Xiao, D., Sun, H., and Liu, C. (2021). Rock mass type prediction for tunnel boring machine using a novel semi-supervised method. Measurement, 179, 109545. https://doi.org/10.1016/j.measurement.2021.109545
  • Yu, Hongjie, and Mooney, M. (2023). Characterizing the as-encountered ground condition with tunnel boring machine data using semi-supervised learning. Computers and Geotechnics, 154, 105159. https://doi.org/10.1016/j.compgeo.2022.105159
  • Zhang, C., Tao, M.-X., Wang, C., and Fan, J.-S. (2024). End-to-end generation of structural topology for complex architectural layouts with graph neural networks. Computer-Aided Civil and Infrastructure Engineering, 39(5), 756–775. https://doi.org/10.1111/mice.13098
  • Zhang, M., Ji, A., Zhou, C., Ding, Y., and Wang, L. (2024). Real-time prediction of TBM penetration rates using a transformer-based ensemble deep learning model. Automation in Construction, 168, 105793. https://doi.org/10.1016/j.autcon.2024.105793
  • Zhang, Q., Liu, Z., and Tan, J. (2019). Prediction of geological conditions for a tunnel boring machine using big operational data. Automation in Construction, 100, 73–83. https://doi.org/10.1016/j.autcon.2018.12.022
  • Zhu, M., Zhu, H., Gutierrez, M., Ju, J. W., Zhuang, X., and Wu, W. (2022). Predicting Tunneling-Induced Ground Collapse Based on TBM Operational Data and Geological Data. International Journal of Computational Methods, 19(08), 2142015. https://doi.org/10.1142/S0219876221420159