Numerical investigation on design of soil nailed slopes by limit equilibrium method

Abstract

ABSTRACT This paper introduces a numerical investigation into the stability analysis of soil nailed slopes. It establishes a simplified framework within the limit equilibrium method, focusing on the physical distribution of tension force along nails. MATLAB® was adapted for the Analysis of Soil Nailed Walls (ASNW) based on Bishop’s method, with major command loops optimized through ‘array operations’. Employing K-fold cross-validation, five regression models predicted the maximum tensile force (Tmax), revealing Gaussian process regression (GPR) as the best model based on root mean square error and r-squared metrics. To capture the amount of uncertainty in the nail load model, we have used a clustering method for estimating the measured Tmax. Self Organizing Map (SOM) was employed for clustering the nail load dataset. We randomly generated Tmax between the measured Tmax obtained by SOM and the predicted Tmax gained by the GPR model for each soil nail. We apply these loads in ASNW program to compute the Factor of Safety, and finally estimate system failure probability for 35 different nail layouts. A parametric study with a uniform nail layout explores the impact of nail length on system failure probability. Results show diminishing significance of nail length on failure probability beyond a nominal length.