Evaluation Performance of Gene Expression Programming and Artificial Neural Network Models in Estimation of Hydraulic Parameters in Triangular Plan Sharp-Crested Weirs

The purpose of this study is to provide an equation for prediction of discharge coefficient (Cd) for triangular plan sharp-crested weirs. Using laboratory data, a regression equation was presented based on h/p and α parameters with a reasonable accuracy. In addition, the performance of artificial neural network (ANN), gene expression programming (GEP) models and regression models (MR-linear and MR-nonlinear) in the estimation of the h/p was investigated. The stage-discharge relationship for triangular sharp-crested weirs was investigated. The results were evaluated using statistical criteria R2, RMSE, NSE and RE%. The values of the R2, RMSE, NSE and RE% were respectively, 0.998, 0.0076, 0.997 and 1.74% for ANN, 0.983, 0.0301, 0.998 and 6.86% for GEP, 0.986, 0.0196, 0.985 and 4.09% for MR-linear and 0.987, 0.0197, 0.984 and 4.09% for MR-nonlinear. The results of statistical criteria indicated the superiority of ANN as compared to other methods. In addition, results showed that based on the angle of the triangular sharp-crested weirs, the Cd is increased from 1 to 8 % respect to the normal sharp-crested weirs with identical width (suppressed weirs). In a situation where the head on the crest of these weirs is low, they will show better performance.