Gene-expression programming to predict the local scour depth at downstream of stilling basins

Abstract

This research introduces an approach to predict maximum scour depth, scour hole length, and the length of the deepest point of the scour hole at equilibrium conditions at downstream of stilling basins by using the GEP. Five non-dimensional parameters in terms of physical properties of bed sediments, stilling basin length, tail water depth, and discharge of spillway were considered as input variables to evaluate the scour hole geometry. The GEP model was developed using experimental datasets collected from literature. Results of the GEP models were compared with those obtained using ANFIS and non-linear regression analysis. Performances indicated that the proposed GEP models to characterize scour hole geometry produced more accurate results than the other methods. In addition, results of sensitivity analysis to de ne the most e ective independent parameters on scour hole geometry were reported. Finally, proper application of the proposed model has been con rmed as the GEP-based best formulation, which is a useful soft computing tool for prediction of scour hole geometry at downstream of a stilling basin.