Boundary Shear Stress Distribution in Straight Compound Channel Flow Using Artificial Neural Network

Abstract

Boundary shear stress distribution of a compound channel is generally influenced by the geometric, roughness, and hydraulic parameters. Experiments are performed on both homogeneous and nonhomogeneous compound channels to study the dependency of variables on the boundary shear distribution. This study proposes an artificial neural network (ANN) model for the prediction of boundary shear stress distribution in straight compound channels. The most influential parameters such as width ratio, relative flow depth, aspect ratio, Reynolds number, and Froude number are considered as input parameters. A large number of experimental data sets comprising wide ranges of width ratio, relative flow depth, roughness ratio, Reynolds number, Froude number, bed slope, and aspect ratio with the present experimental data series are used for both training and validation of the model. Previous models can provide good results only for specific ranges of independent parameters, whereas back-propagation neural network (BPNN) models are capable of performing well for global ranges of independent parameters. This is because BPNN is able to perform nonlinear mapping between the dependent and independent variables during the training. The efficacy of the models is verified with the standard statistical error analysis using the global data sets.