Artificial neural network response assessment of a single footing on soft soil reinforced by rigid inclusions

Abstract

This study is devoted to investigate the behavior of a single footing resting on a soft soil reinforced by four rigid inclusions (RI) using an Artificial Neural Network (ANN). Full-scale static loading tests on a single footing are considered as the reference of this study. For the design of such structures, a geotechnical investigation allowed the soil properties determination. Then 3D Finite Difference Modeling (FDM) analyses were conducted and validated on the experimental results. A total of 196 datasets are used to construct and verify the Artificial Neural Network (ANN) response model accuracy. To follow the behavior of such footing, its center settlement, global load transfer efficiency and the maximum rigid inclusions bending moments are considered. This approach allows to efficiently assess the rigid-inclusion reinforced system performance with a high degree of success. Hence it is used to compute the RI-reinforced footing system responses under various configurations. A comparative study is carried out considering four key parameters of the load transfer platform. Four design charts are proposed and can serve for a pre-design stage.