Publication

Abstract : In the current study, the bearing capacity of well foundation due to surcharge and self-weight of soil is determined using finite element lower bound limit analysis. The soil is considered as cohesionless and follows Mohr–Coulomb failure criterion. The bearing capacity is determined for different sets of parameters, such as (1) the ratio of internal to external radius of the foundation, (2) the internal friction angle of soil, (3) the foundation–soil interface friction angle, and (4) the ratio of depth of foundation to external radius. The dataset obtained from numerical simulation is trained using a 4-10-1 artificial neural network (ANN) architecture, to predict the bearing capacity of well foundation in terms of dimensionless bearing capacity factor (Nw). Based on the weight and bias values obtained from the trained neural network model, a series of equations is developed to predict the bearing capacity with good accuracy. The predicted values from the prediction equations are found to be very close to the actual obtained values. The prediction equation can be used by practicing engineers for designing well foundation.