PREDICTING STABILITY FACTOR NC OF RECTANGULAR TUNNELS USING ARTIFICIAL NEURAL NETWORKS
DOI:
https://doi.org/10.61591/jslhu.27.1175Keywords:
Rectangular tunnel; Artificial neural network; Stability factor Nc; Finite element limit analysisAbstract
An artificial neural network (ANN) model is proposed in this study to predict the stability factor Nc of rectangular tunnels in undrained clay. The training dataset comprising 60 cases is extracted from the verified study of Shiau et al. (2024) using the finite element limit analysis (FELA) method, with two input parameters: the depth ratio (H/D) and the width ratio (B/D), along with four outputs: Nc for collapse and blowout scenarios at both lower and upper bounds. The ANN model is developed using custom MATLAB code and trained with the modern Adam optimization algorithm. The study focuses on two main investigation directions: the effect of the number of hidden-layer neurons and the comparison of activation functions. Results show that the GELU activation function outperforms traditional functions, and the model achieves coefficients of determination R² on the test set ranging from 0.9987 to 0.9998 for all four outputs, surpassing the MARS model. The proposed model offers a practical and efficient tool for engineers to rapidly estimate the stability factor of rectangular tunnels without the need for complex numerical simulations.
References
Daniel W. Wilson, Andrew J. Abbo, Scott W. Sloan, and Kentaro Yamamoto. 2017. Undrained stability of rectangular tunnels where shear strength increases linearly with depth. Canadian Geotechnical Journal. 54(4): 469-480.
Sloan, S. W. Geotechnical stability analysis. Géotechnique, 2013, 63(7): 531–571.
DOI: https://doi.org/10.1680/geot.12.RL.001
Wilson, D. W., Abbo, A. J., Sloan, S. W., & Lyamin, A. V. Undrained stability of a square tunnel where the shear strength increases linearly with depth. Computers and Geotechnics, 2013, 49: 314–325.
DOI: https://doi.org/10.1016/j.compgeo.2012.09.005
Abbo, A. J., Wilson, D. W., Sloan, S. W., & Lyamin, A. V. Undrained stability of wide rectangular tunnels. Computers and Geotechnics, 2014, 59: 46–59.
DOI: https://doi.org/10.1016/j.compgeo.2013.04.005
Sahoo, Jagdish Prasad, and Jyant Kumar. "Stability of a circular tunnel in presence of pseudostatic seismic body forces." Tunnelling and Underground Space Technology 42 (2014): 264-276.
Shiau, J., & Keawsawasvong, S. (2022). "Producing Undrained Stability Factors for Various Tunnel Shapes." International Journal of Geomechanics, ASCE, 22(8): 06022017.
DOI: 10.1061/(ASCE)GM.1943-5622.0002487
Shiau, J., Keawsawasvong, S., Lai, V. Q., & Shiau, K. Rectangular tunnel heading stability in three dimensions and its predictive machine learning models. Journal of Rock Mechanics and Geotechnical Engineering, 2024, 16(12): 4683–4696.
DOI: https://doi.org/10.1016/j.jrmge.2023.12.035
Nguyễn Sỹ Hùng. Nghiên cứu ảnh hưởng của hình dạng tiết diện đến sự ổn định gương hầm trong đất sét bão hòa nước. Luận án Tiến sĩ, Đại học Bách khoa TP. Hồ Chí Minh, 2021.
Lại Văn Quí. Cơ học đất nâng cao và mô hình hóa bài toán địa kỹ thuật trên phần mềm Abaqus. Nhà xuất bản Xây dựng, Hà Nội, 2024.
Đỗ Minh Ngọc. Ứng dụng thuật toán học máy trong dự báo biến dạng mặt đất do thi công hầm đô thị. Tạp chí Giao thông Vận tải, 2022, 10.
McCulloch, W. S., & Pitts, W. A logical calculus of the ideas immanent in nervous activity. Bulletin of Mathematical Biophysics, 1943, 5(4): 115–133.
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. Learning representations by back-propagating errors. Nature, 1986, 323(6088): 533–536.
