Prediction of Bearing Capacity and Settlement from SPT Values Using Genetic Algorithm

Abstract

Many empirical and semi-empirical methods are existing to predict bearing capacity and settlement, but majority of them are inconsistent and not user friendly. In this work, a genetic algorithm approach is used for predicting bearing capacity and settlement of shallow foundations on C-soil, φ-soil and C-φ soil, separately, based on those input parameters which can be easily find out from simple experiments. The development and verification of the genetic models were done using a large database containing about 832 datasets from 167 soil investigation reports. The equation developed for bearing capacity and settlement thus obtained can be used in prediction of new cases that were not used for the development of the genetic model. The results of the model obtained were compared with various conventional equations available for calculating bearing capacity and settlement and was found to be superior. The correlation of predicted data with actual field measurements was determined and it was found out that the genetic algorithm approach have high degree of accuracy. Parametric study was also done to evaluate the effect of varying each of the input parameters on the corresponding output.