Prediction of Scour at a Side-Weir with GEP, ANN and Regression Models

Side-weir is known as a lateral intake structure, which is widely used in irrigation, land drainage, and urban sewerage system by flow diversion device. Local scour in-/volves the removal of material around piers, abutments, side-weir, spurs, and embankments. Clear-water scour depth based on four dimensional parameters: approach flow velocity (V (1)/ V (c) ), water head ratio (h (1) - p)/h (1), side-weir length (L/b) and sediment size (d (50)/p). The equilibrium depth of scour increases by the increase of the dimensionless parameters of approach flow velocity, water head ratio, side-weir length and sediment size. This study presents artificial neural network (ANN) and gene expression programming (GEP) models, which is an algorithm based on genetic algorithms and genetic programming, for prediction of the clear-water scour depth at side-weir. The explicit formulations of the GEP models are developed. The GEP-based formulation and ANN approach are compared with experimental results, multiple linear/nonlinear regressions (MLR/MNLR). The performance of GEP is found in slightly more influential than ANN approach and MNLR for predicting the clear-water scour depth at side-weir.