Comparison of five different models in predicting the shear stress distribution in straight compound channels

Abstract

The ability of five different methods to estimate the shear stress distribution in compound channels is investigated. Methods proposed by Yang and Lim (YLM), Khodasheans and Paquier (KPM), Sterling and Knight (SKM), Zarrati et al. (ZAM) and Bonakdari et al. (BAM) are compared with experimental data. YLM and KPM did not provide reliable results as they produced higher mean absolute percentage of error (MAPE) values of 25-55%. SKM performed adequately in predicting the pattern of shear stress distribution on the main channel bed, but on a floodplain bed it predicted a constant value over the entire wetted perimeter. The SKM method outperformed YLM and KPM with 2 to 20% MAPE. The ZAM and BAM methods produced the best results for shear stress distribution in compound channels with average MAPE% of 2.67 and 5.66 respectively. Although ZAM showed more accurate results than BAM, however BAM required solving much fewer equations than ZAM and presented more accurate results than other geometric methods. Among all models, BAM is proposed as a simple and accurate model for predicting the shear stress distribution in compound channels.