GEP-based predictive modeling of deformational behavior of circular tunnels excavated in soft soils

Abstract

The present paper aimed to estimate the ground surface settlement induced by excavated circular tunnel in soft soils, taking into account the uncertainties of geometric and geotechnical parameters. To this end, following two-dimensional numerical simulation of the tunnel based on finite difference method in FLAC2D, the ground surface settlement profiles were determined using regression analysis (Gaussian equation). The effect of changes in tunnel geometric parameters (including buried depth C, diameter D and depth distance H of tunnel) and soil geotechnical properties (including modulus of elasticity E, undrained shear strength Su and density γ) based on three dimensionless ratios named depth ratio (C/D), soil strength ratio (γD/Su) and the soil stiffness ratio (E/Su) were studied on the ground surface settlement. The results showed that there is a relationship in the form of Smax/H = f (γD/Su, E/Su, C/D) between the parameters to determine the normalized ground surface settlement induced by tunnel. Therefore, using the gene expression programming (GEP) algorithm and forming a database containing 1000 different simulations in terms of a combination of C/D, γD/Su changes for fixed E/Su ratios (inputs) and Smax/H results (outputs), two powerful empirical equations were proposed as new empirical models to estimate the normalized maximum ground surface settlement. © 2024 Techno-Press, Ltd.