Unlocking precision in hydraulic engineering: machine learning insights into labyrinth sluice gate discharge coefficients

Abstract

This study investigates the discharge coefficient (Cd) of labyrinth sluice gates, a modern gate design with complex flow characteristics. To accurately estimate Cd, regression techniques (linear regression and stepwise polynomial regression) and machine learning methods (gene expression programming (GEP), decision table, KStar, and M5Prime) were employed. A dataset of 187 experimental results, incorporating dimensionless variables of internal angle (θ), cycle number (N), and water depth contraction ratio (H/G), was used to train and evaluate the models. The results demonstrate the superiority of GEP in predicting Cd, achieving a coefficient of determination (R2) of 97.07% and a mean absolute percentage error of 2.87%. To assess the relative importance of each variable, a sensitivity analysis was conducted. The results revealed that the H/G has the most significant impact on Cd, followed by the internal head angle (θ). The cycle number (N) was found to have a relatively insignificant effect. These findings offer valuable insights into the design and operation of labyrinth sluice gates, contributing to improved water resource management and flood control.