estimation of energy dissipation in non-aerated flow regimes over stepped spillways using advanced soft computing techniques

Stepped spillways have garnered significant attention due to their high efficiency in dissipating flow energy, primarily attributed to the presence of steps that enhance turbulence and energy loss. a stepped spillway consists of a sequence of vertical drops extending from the crest at the upstream end to the stilling basin at the downstream. under high discharge conditions, the flow regime transitions into non-aerated skimming flow, characterized by substantial energy levels that necessitate careful management. accurate estimation of energy dissipation is essential for the safe and economical design of downstream energy dissipators. in this study, 154 experimental data points from physical models of stepped spillways were utilized, encompassing a broad range of hydraulic conditions by varying parameters such as the drop number, spillway slope, number of steps, critical depth-to-step height ratio, and froude number. to predict the energy dissipation, several soft computing techniques were applied, including artificial neural networks (ann), adaptive neuro-fuzzy inference systems (anfis), and support vector regression (svr). the models' predictive capabilities were assessed using key statistical performance metrics, including the coefficient of determination ( ), root mean square error (rmse), and mean absolute error (mae). comparative analysis of the results demonstrated that the ann model exhibited superior accuracy over the other models, achieving , rmse, and mae values of 0.99, 0.96, and 0.67, respectively. the findings underscore the potential of soft computing models, particularly ann, as powerful predictive tools in hydraulic engineering applications. the proposed modeling approach offers an effective means for estimating energy dissipation in stepped spillways, facilitating optimized and cost-effective design of hydraulic structures.