Enhancing strip footing bearing capacity on soil slopes with geogrid reinforcement: insights from finite element analysis and machine learning

This study investigates the influence of a horizontal geogrid layer reinforcement on the ultimate bearing capacity (UBC) of a strip footing under static and seismic loading conditions resting on a c-phi soil slope subjected to uniformly distributed load using Finite Element Limit Analysis (FELA). The Bearing Capacity Ratio (BCR), which is the ratio of the UBC of the strip footing with and without reinforcement in soil slope, quantifies the influence of the reinforcement on the UBC of the footing. A parametric analysis is carried out to evaluate the influence of the vertical depth of geogrid placement (u/B), offset distance of the footing (x/B), slope inclination (beta), and the horizontal seismic acceleration coefficient (kh) on the BCR. A reinforcing layer beneath the foundation can improve the UBC by up to 1.8 times under static and seismic loading, and this improvement is governed by u/B, x/B, beta, and kh values. The optimal depth for maximizing this capacity ranges as u/B = 0.5-0.75. The position of the footing on the slope (x/B) also impacts the effectiveness of the geogrid reinforcement. The study identified x/B and u/B as crucial factors influencing the footing-slope system's failure mechanisms. This investigation developed various advanced machine-learning models to predict BCR values, including Fine Tree regression, Linear Regression, Linear SVM, and Narrow Neural Networks. The Narrow Neural Network outperformed the others, achieving a lower Root Mean Squared Error (RMSE) of 0.0334 and a higher R-2 value of 0.9538, indicating its strong predictive power compared to the other models.