Search

• A novel approach to study uplift resistance of suction caisson in NGI-ADP soils. • 2D finite element analysis for suction caisson. • Failure mechanism, respective calculated tables and charts are presented. • Machine learning approach for building empirical equation and sensitivity analysis. The paper presents a novel solution for evaluating the uplift resistance of caisson foundations in anisotropic clay using displacement finite element analysis (FEA) and artificial neural network (ANN). The commercial FEA PLAXIS code is used to examine the uplift resistance of caisson foundations with a built-in anisotropic soil model developed by Norwegian Geotechnical Institute (NGI) based on the Active-Direct Shear-Passive concept (NGI-ADP model). The uplift resistance is expressed by a dimensionless stability factor that is a function of four dimensionless design parameters, i.e., the ratio of depth to diameter, the adhesion factor at the interface between caisson and clay, the shear strength gradient ratio, and the anisotropic strength ratio. Using the produced numerical results, ANN is adopted as the data driven technique to propose an empirical equation to estimate the uplift resistance factor. Furthermore, a sensitivity study is implemented to evaluate the relative importance of the four considered parameters on the uplift resistance of the caisson. Design charts, tables and empirical equations are developed and can be used to assist in practical designs. [ABSTRACT FROM AUTHOR]