عنوان مقاله [English]
Ogee crested spillways having superb hydraulic properties including simplicity in design and flow passing efficacy. So far, limited research in the area of prediction and the extraction of discharge coefficient relationship is conducted. In current study two different methods for modeling the discharge coefficient of the converging ogee spillway with a curve axis by was developed and results were compared with the observed experimental values through the Genetic Expression Programming (GEP) and artificial networks (ANNs) approaches. For this purpose, the experimental data of the Germi chay ogee spillway model with varying training wall convergence angles (), was used. Based on the obtained results, applied Artificial Intelligence (AI) models have reliable performance in predicting the discharge coefficient of converging ogee spillways. Moreover, the performance of GEP model is a bit better than ANN technique with relatively low error and high correlation values. To recognize the most effective variables on the discharge coefficient, sensitivity analysis of GEP for the best model was carried out. Results showed that ratio of the design head to the critical depth (Hd/yc) and ratio of the crest length to the downstream channel width (L/Lch) are the most and least important parameters in predicting the discharge coefficient of the converging ogee spillway respectively. The best evaluation of test series were observed in GEP approach with the values of DC=0.818 and RMSE=0.089 and in ANNs approach with the values of DC=0.77 and RMSE=0.099 which demonstrates the high accuracy of predictions.