the study of the performance of data- driven models to predict the scour depth caused by the aerated vertical jet

High flow discharges coming from the hydraulic structures usually carry a highvelocity jet of flow, which could have different short and longterm impacts on the river mechanics and the habitat conditions. scouring is one of the major effects of the incoming flow jet, which, once aerated, has a dynamic behavior and structure. plunge pools are hydraulic structures to prevent the severe damages of the scouring phenomena. in the present study, due to the high complexity of constructing a physical model, the effect of air entrainment on scoured hole’s depth is assessed using the artificial neural network (ann) and adaptive neurofuzzy inference system (anfis) methods. each soft computing model’s performance on the scouring is compared to a nonlinear regression method’s result using different statistical measures (rmse, me, mae). the prediction accuracy of ann, anfis, and nonlinear regression using rmse was calculated as 0.0137, 0.011, and 0.0262, respectively. this study presents a novel achievement in measuring and predicting the scoured hole’s depth as one of the most critical phenomena in hydroenvironmental science.

The Study of the Performance of Data- Driven Models to Predict the Scour Depth Caused by the Aerated Vertical Jet

High flow discharges coming from the hydraulic structures usually carry a highvelocity jet of flow, which could have different short and longterm impacts on the river mechanics and the habitat conditions. Scouring is one of the major effects of the incoming flow jet, which, once aerated, has a dynamic behavior and structure. Plunge pools are hydraulic structures to prevent the severe damages of the scouring phenomena. In the present study, due to the high complexity of constructing a physical model, the effect of air entrainment on scoured hole’s depth is assessed using the Artificial Neural Network (ANN) and Adaptive NeuroFuzzy Inference System (ANFIS) methods. Each soft computing model’s performance on the scouring is compared to a Nonlinear Regression Method’s result using different statistical measures (RMSE, ME, MAE). The prediction accuracy of ANN, ANFIS, and nonlinear regression using RMSE was calculated as 0.0137, 0.011, and 0.0262, respectively. This study presents a novel achievement in measuring and predicting the scoured hole’s depth as one of the most critical phenomena in hydroenvironmental science.