Distribution and Prediction of Boundary Shear in Diverging Compound Channels

Measurement of bed shear stress is always a challenging task for engineers. In river engineering, bed shear is a fundamental variable and is important in estimating flow resistance and sediment transport. In this study, experiments are carried out in diverging compound channel with smooth bed (perspex sheet) and rough bed (Gravel) conditions to determine the effect of roughness. The shear velocity is estimated from universal logarithmic law. The effect of geometry and roughness on Von-Karman constant, eddy viscosity coefficient, friction factor is studied. The mass conservation and momentum conservation equations are used to derive apparent shear forces at interface of main channel and floodplain. A genetic algorithm model is developed to predict percentage of shear force (%Sfp) carried by sub-sections. To perform better with less and unseen data K-Fold cross-validation technique is used. The model is compared with available models in literature and it is observed that developed model gave better predictions with low MAPE.