نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشیار گروه مهندسی عمران، دانشگاه تبریز، تبریز، ایران
2 دانشجوی دکترای عمران (سازههای هیدرولیکی) دانشگاه تبریز، تبریز، ایران
3 دانشیار گروه مهندسی رودخانه و سواحل، پژوهشکده حفاظت خاک و آبخیزداری، تهران، ایران
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Three-dimensional pattern of flow on ogee spillway and unlimited geometric parameters changing like a change in crest shape, or modification of the approach channel owing to positional geometric qualities may change the flow characteristics on the one hand and the limited information available about hydraulic ogee spillway with curve axis on the other hand, as a major challenge, has led to the use of meta-model systems and data-oriented used by researchers. In current study Artificial Intelligence Techniques (Adaptive Neuro-Fuzzy Inference System (ANFIS) and Support Vector Machine (SVM)) were applied to predict the discharge coefficient of the ogee spillway with curve axis and compared to experimental data. For this purpose, the experimental data of the ogee spillway model with varying training wall convergence angles, , (which was made by the author) was used and regarding the different models based on the concepts of ogee spillway, examined and evaluated. The obtained results indicated that in estimating the discharge coefficient of the ogee spillway with curve axis, artificial intelligence techniques are very accurate and they show good agreement between observed and predicted values. According to the obtained results of sensitivity analysis it was observed that the ratio of the design head to critical depth ( ) have the highest impact on predicting of discharge coefficient and also the ratio of the total upstream water head to design head ( ) and the ratio of the total upstream water head to spillway height ( ) have the same and marginal impact on this term. The best evaluation of test series were observed in SVM approach with the values of R=0.966, DC=0.93, RMSE=0.06 and in ANFIS approach with the values of R=0.945, DC =0.885, RMSE=0.088, which demonstrates the high accuracy of predictions.
کلیدواژهها [English]