Prediction of equilibrium scour depth in uniform non-cohesive sediments downstream of an apron using computational intelligence

Abstract

Accurate prediction of equilibrium scour depth downstream of hydraulic structures has an important role in their appropriate design. In this paper, the applicability of artificial neural networks (ANNs) and adaptive neuro-fuzzy inference system (ANFIS) models in prediction of equilibrium scour depth is studied. Multi-layer perceptron neural network structure is employed for the training of the ANN models. Experimental data of published literature are used for training and testing the models. The equilibrium scour depth is predicted as a function of five input variables; apron length, sluice gate opening, issuing velocity of jet, tailwater depth and median sediment size. The comparison of results shows that ANN model provides a better prediction of scour depth than ANFIS as well as the previous empirical relationship obtained by regression analysis. However, the results indicate that both ANN and ANFIS models can successfully predict equilibrium scour depth. Finally, based on the sensitivity analysis, it was found that d50 has the greatest effect on equilibrium scour depth.