Ground surface settlement is the most significant restriction when constructing shallow metro station tunnels in urban areas. The umbrella arch method (UAM) is generally applied as a tunnel support method. However, UAM becomes inadequate in some soil conditions, such as loose sand or soft clay. Innovative support systems are required to safely build shallow metro station tunnels in urban areas. The objective of this research is to investigate alternative tunnel support systems and appropriate soil models to safely construct shallow twin-tube metro station tunnels. The continuous pipe arch system (CPAS), which consists of horizontal and continuous pipes along the metro station tunnels, was modeled in three dimensions (3D) using the finite element (FE) program Plaxis3D for various pipe diameters. The ground surface settlement results of the 3D models were compared with the in situ settlement measurements to validate the geotechnical parameters of the soils used in the models. It was observed that the hardening soil (HS) model was more accurate than the Mohr-Coulomb (MC) soil model. As a result of the 3D FE model analysis, maximum ground surface settlements were obtained below 50 mm when the pipe diameters of CPAS were larger than an internal diameter (ID) of 1200 mm at a cover depth of 10 m in sandy clay soil. It is revealed that CPAS with pipe diameters between ID 1200 mm and ID 2000 mm can be utilized as a tunnel support system in urban areas to construct shallow twin-tube metro station tunnels with low damage risk.

Accurate settlement forecasting is essential for preventing severe structural and infrastructure damage. This paper investigates predicting tunneling-induced ground settlements using machine learning models. Empirical methods for estimating settlements are often imprecise and site-specific. Developing novel, accurate prediction methods is critical to avoid catastrophic damage. The umbrella arch method constrains deformations for initial stability before installing primary support. This study develops machine learning models to forecast settlements solely from umbrella arch parameters, disregarding soil properties. Multilayer perceptron artificial neural networks (MLP-ANN) and support vector regression (SVR) are applied. Results demonstrate machine learning outperforms empirical methods. The MLPANN surpasses SVR, with R2 of 0.98 and 0.92, respectively. Strong correlation is observed between umbrella arch configuration and settlements. The suggested approach effectively estimates surface displacements lacking mechanical properties. Overall, this study supports machine learning, specifically MLP-ANN, as an efficient, reliable alternative to empirical methods for predicting tunneling-induced ground settlements from umbrella arch design.