Modelling the capacity of micropiled-raft foundation rested on soft clayey soil using an artificial neural network approach

Abstract

ABSTRACT Micropiled-raft foundation shows its ability to solve particular type of foundation problem i.e. building of medium height and weight structures over soft deep-seated clayey soil deposit. Potential of artificial neural network method was used in the study for the assessment of capacity of micropiled-raft foundation under a given settlement in soft clayey soil considering the complex non-linear load-settlement behaviour. Experiments were conducted on micropiled-raft foundation constructed with different variables within a soft clayey soil bed in the test pit in field and a database was prepared from load-settlement graphs those were utilized to develop different ANN models. It was observed from the study that Bayesian Regularization algorithm with 90–10% validation model performs the best. Sensitivity analysis was performed to determine the relative significance of different input variables and a neural interpretation diagram was prepared. An empirical equation was proposed with best fit ANN model and an example illustrated.