Transient settlement estimation of shallow foundation under eccentrically inclined static arid cyclic load on granular soil using artificial intelligence techniques

The present study focuses on estimating the transient response of shallow strip footing on granular soil using soft computing techniques. A shallow foundation is numerically modelled using Beam on Nonlinear Winkler Foundation model. Then the footing is subjected to a combination of allowable static and cyclic load depending on the ultimate bearing capacity of the footing. The eccentricity and angle of load inclination of static load are varied to simulate more practical conditions. The cyclic load is rectangular pulse load. One cycle of rectangular pulse load is applied to observe the immediate response of the foundation, referred as the transient response. Apart from the loading parameters, three granular soils of three different relative densities (D,= 35%, 51% and 69%) are considered. Based on numerical simulation of 1728 conditions, soft computing models are developed using five techniques, viz. Neural Networks, Support Vector Machines, Multivariate Adaptive Regression Splines, Adaptive Neuro Fuzzy Interface System and Multi Gene Genetic Programming. It is found that the static load on the foundation is the most important parameter controlling the transient response of the footing.