Estimating variations in permeability of earth-dam body during leakage phenomenon by neural network

Leakage is one of the most important problems in earth dam construction. Lack of leakage phenomenon analysis for earth dams lead to destructive problems like increase in leakage forces, increase in pore water pressure and stability of the earth dam. In this study, leakage in two sections of an earth dam are modeled and analyzed by Finite Element Method (FEM), Multi Layer Perceptron network (MLP) and Radial Basis Function network (RBF). Predictions made by each of the methods are compared and it can be concluded that FEM prediction are not compatible with actual data, whereas sophisticated neural networks make acceptable predictions

Keywords:

Earth-dam , Leakage , Neural Network