Modeling energy dissipation over stepped spillways using machine learning approaches

Numerical Simulation for Estimating Energy Dissipation over Different Types of Stepped Spillways and Evaluate the Performance by Artificial Neural Network

“…Based on the use of both FLUENT and V-Flow to modeling the unstable twodimensional flow over the stepped spillway, results obtained by the FLUENT program for the energy dissipation for the stepped spillway showed a reasonable agreement in results [7][8][9]. Also, artificial neural network (ANN) and the techniques of gene expression used for modeling the experimental data of nappe and skimming flow over stepped spillways, so the results obtained from these techniques were at a reliable level in the prediction of the energy dissipation [10]. Laboratory experiments used to find the optimal height of the step and the best slope for energy dissipation through the application of several different models and the different flow systems represented by the skimming, transition and nappe flow by the calculation of energy dissipation of various models [11].…”

Discussion of “Energy Dissipation Prediction for Stepped Spillway Based on Genetic Algorithm–Support Vector Regression” by Lei Jiang, Mingjun Diao, Hongcheng Xue, and Haomiao Sun

“…Chinnarasri and Wongwises (2006), and Salmasi and Özger (2014) already gave the flow regime information for every data point. For the data points from Roushangar et al (2014), and Sholichin et al (2016), their corresponding flow regimes are unknown, and therefore they were classified into nappe, transition, and skimming flow data according to the empirical equations about the upper limit of nappe flow and the lower limit of skimming flow (Chanson 2001)…”

“…In this discussion, 217 data points captured from Chinnarasri and Wongwises (2006), Roushangar et al (2014), Salmasi and Özger (2014), and Sholichin et al (2016) were utilized to feed GA-SVR models, and three statistical indexes were applied to evaluate the models' accuracies, namely, the root-mean-square error (RMSE), squared correlation coefficient (R 2 ), and mean relative deviation (MRD), expressions for which are given as Eqs. (7)-(9) in the original paper.…”

Closure to “Energy Dissipation Prediction for Stepped Spillway Based on Genetic Algorithm–Support Vector Regression” by Lei Jiang, Mingjun Diao, Hongcheng Xue, and Haomiao Sun

“…Bahrami asked why the authors did not compare the results with the methodology used by Salmasi and Özger (2014). The original paper compared the performance of the adaptive neuro-fuzzy inference system model (ANFIS) (Salmasi and Özger 2014;Roushangar et al 2017), multivariate adaptive regression splines model (MARS) (Parsaie et al 2016), and back-propagation neural networks (BPNN) (Roushangar et al 2014) in the "Discussion" section. The discusser also pointed out that the data cannot sufficiently represent the conditions in nappe and transition flow regimes.…”