Predicting the Settlement of Gypseous Soil Using Artificial Intelligence Techiniques

Abstract

Abstract The problem of estimating the settlement of the shallow foundation on the gypseous soil is very complex and not fully entirely understood. many methods have been developed to predict the settlement of the isolated and strip foundations. However, methods for such prediction that have the required degree of accuracy and consistency. In this study several artificial intelligent modeling method were applied, deep neural network (DNN), artificial neural network (ANN), support vector regression (SVR), and linear regression (LR). The parameters of predict shallow footing settlement are selected carefully based on previous studies. These were footing geometry, Df/B ratio, gypseous soil properties like water content, gypsum content, dry unit weight, cohesion, angle of internal friction, and time of testing. effect of the adopted parameters on the prediction ability of the surface settlement like Df\B, footing geometry, load, and time of testing. It is significant that they have assumed effect on the prediction ability of the surface settlement of the shallow foundation. A back propagation typed neural network was used in this study, where four artificial intelligent models has been adopted in this study, deep neural network model showed the most significant performance among the other model with the least mean absolute error and mean square error which were 2.9% and 3.87%. Deep neural network model recorded the highest coefficient of efficiency and variance account. It has concluded that deep neural network model can be used to predict the settlement of the shallow foundation on a gypseous soil.