Life Cycle Cost Analysis of NATM Using Ann

This research aims to explore the application of artificial intelligence (AI) techniques, particularly Artificial Neural Networks (ANN), in tunneling methods including Tunnel Boring Machine (TBM), New Austrian Tunneling Method (NATM), and Non-Mechanized Tunneling (NMT). The performance of tunneling operations directly influences the financial success of construction projects, making accurate prediction and control of tunnel behavior critical. Traditional models for TBM tunnels often fail to incorporate all key factors such as rock mass properties, machine specifications, and weathering effects, resulting in inaccurate performance forecasts. Weathering, in particular, has been underexplored despite its significant impact on tunnel stability. NATM tunnels, increasingly employed for urban tunnels with shallow overburden, require precise prediction, monitoring, and control of ground displacements to ensure safety and structural integrity. Numerical simulations like finite element methods are widely used but face challenges in defining nonlinear material parameters accurately. AI, especially ANN, offers a promising alternative by mimicking human brain functions to analyze complex, nonlinear relationships between tunneling parameters and ground behavior. This study demonstrates the effectiveness of ANN in predicting deformation in NATM tunnels, validated using field data, with high prediction accuracy. Furthermore, shallow tunneling in densely populated areas causes ground movements affecting nearby structures. Existing computational models often fail to consider all influencing parameters, leading to unreliable settlement predictions. ANN-based approaches provide a robust solution to predict surface settlements accurately, aiding in proactive monitoring and mitigation. Results from the Karaj Urban Railway project in Iran confirm that ANN is a viable and effective technique for predicting ground response in tunneling projects.