Modeling of flow characteristics beneath vertical and inclined sluice gates using artificial neural networks

Abstract

Abstract In this paper, artificial neural networks (ANNs) modeling method with back propagation algorithm was employed to investigate the flow characteristics below vertical and inclined sluice gates for both free and submerged flow conditions. Two ANN models were developed yielding two generalized equations to predict the discharge coefficient (Cd) values for both modes of flow. The model network for free flow entailed four input variables, namely, dimensionless upstream water depth, Froude number, Reynolds number, and inclination angle, whereas, the Cd value represented the only single output variable. For submerged flow ANN model, a fifth input variable was added, which is the dimensionless tailwater depth. The two ANN models were trained and validated against 420 data sets collected from previous experimental studies. The results indicated that ANNs are powerful tools for modeling flow rates below both types of sluice gates within an accuracy of ±5%.