Seismic Bearing Capacity of Strip Footing Placed on Sand Layer Over Hoek-Brown Media using Finite Element Limit Analysis and Machine Learning Approach

Bearing capacity of the foundation may be affected by the thickness of a rigid base such as rock. Earthquake conditions have a considerable impact on the bearing capacity. In the present study, a numerical analysis based on finite element limit analysis is performed to compute the seismic bearing capacity of strip footing on a sand layer overlying level rock media. The rock media obeys the Hoek-Brown yield criterion. Based on the obtained results, a new dataset is prepared. A deep neural network (DNN), support vector machine (SVM), and Gaussian process regression (GPR) models are applied to the dataset to predict seismic bearing capacity of strip footing. It was shown that the seismic bearing capacity of strip footing increases as the thickness of sand layer decreases. The results confirmed that the DNN model is more reliable for bearing capacity prediction compared to other models. Exponential GPR model was found to give the lowest value of MSE and RMSE than the rational quadratic GPR and the DNN.