Explicit Equations for Estimating Resistance to Flow in Open Channel with Moveable Bed Based on Artificial Neural Networks Procedure

Abstract

The resistance to flow in an open channel is associated with the value of the Darcy-Weisbach friction factor f. For natural channels with a movable bed, the f value depends on the grain size of the bed materials and the bedforms, such as ripple, dune, or anti-dune. The total resistance to flow is the sum of the resistance due to grain roughness and bedform. Several researchers have proposed several graphs to determine the friction factor value due to the bedforms. Still, using these graphs requires graphical interpolation, which is inconvenient and difficult to apply to the flow and sediment transport calculation. This study proposes two explicit equations, ANN models 1 and 2, to compute the friction factor due to the bedform based on artificial neural networks (ANN) procedure. The data used to build the equations were obtained by digitizing the graph proposed by Alan and Kennedy. The explicit ANN equations are in the form of a series of hyperbolic tangent functions. The resulting equations can predict the friction factor value due to bedform satisfactorily.