Evaluation of Regression-Based Soft Computing Techniques for Estimating Energy Loss in Gabion Spillways

Estimation of flow energy loss in gabion spillways can be effective in managing erosion downstream of structures, flood control, and riverbed stabilization. Therefore, in this research, using two soft computing models evolutionary polynomial regression (EPR) and multivariate adaptive regression spline (MARS), the amount of energy loss in these spillways was estimated. About 75% of the 74 laboratory data samples were used for training and the remaining 25% were used for testing the models. The dimensionless parameters of Froude number (Fr), spillway slope (S), gabion number (GN), and porosity (n) were used as input parameters. The results showed that the MARS model predicted the energy loss values by root mean square error (RMSE), mean absolute percentage error (MAPE), and correlation coefficient (CC) of 0.05, 0.017, and 0.99, respectively, which has better performance than the EPR model has. The results of the Taylor diagram also showed that the performance of MARS and EPR are satisfying, and their accuracy is very close to each other. The regression equation by the EPR model was more complex than the regression equation by the MARS model. According to the obtained results, the use of the two soft computing models in estimating energy loss in spillways is recommended.