Subsidence Prediction for Twin Tunnels Using Genetic Algorithms-Case Study: Isfahan, Iran

Today, with the increasing pace of urbanization and population growth, the demand for metro tunnels has risen significantly. These tunnels are often constructed at shallow depths and in close proximity to urban infrastructure. In such scenarios, it becomes crucial to safeguard buildings and other structures from potential damage caused by metro tunnel excavation. Therefore, studying ground movements and surface settlements induced by these tunnels is essential to ensure the safety of surface structures. In this research, based on the geotechnical conditions of the study area and data collected during the excavation of twin metro tunnels using a Tunnel Boring Machine (TBM), an empirical method was developed to estimate subsidence, maximum subsidence, and inflection points. Boltzmann and Gaussian functions were employed to derive these parameters. Furthermore, a sensitivity analysis was conducted using three-dimensional Plaxis software for the cross-section of Si-o-se-pol alongside three other soil types. A numerical relationship for predicting maximum subsidence was then proposed using a genetic algorithm. The results revealed strong alignment between the empirical and numerical approaches derived in this study. Consequently, these findings enable the accurate prediction of ground subsidence for Isfahan Metro Line 2, which is now under excavation.