A novel approach for estimation of discharge coefficient in broad-crested weirs based on Harris Hawks Optimization algorithm

Accurate determination of the discharge coefficient play a very important role in estimating the flow discharge over the weirs. As a result, it is significant to estimate the discharge coefficient correctly. The aim of this study is simulation and estimation the discharge coefficients (C_d) over the broad-crested weirs with cross section rectangular and suppressed. Hence, numerical simulation of hydraulic characteristics of these weirs were performed by ANSYS FLUENT software and results were obtained. Then two intelligent models of ANN, GPR and hybrid both of models namely ANN-HHO, GPR-HHO were used to determine the discharge coefficients using the efficient parameters and the results of these models were compared. Assessment of the results were performed using the statistical metrics: coefficient of determination (R²), root mean square error (RMSE), mean absolute error (MAE), scatter index (SI) and Kling-Gupta efficiency (KGE) and graphical diagrams including violin plot, percent relative error (RE%) plot and probability density function (PDF) plot of residuals. It was found that hybrid artificial neural network and gaussian process regression with Harris Hawks optimization (ANN-HHO and GPR-HHO) could improve ANN and GPR models performance in estimating the C_d in broad-crested weirs. Overall, results indicated that a combination of the HHO with the ANN (ANN-HHO) model performs better than GPR-HHO model for the estimation of the C_d.