Neural networks- and neuro-fuzzy-based determination of influential parameters on energy dissipation over stepped spillways under nappe flow regime

Abstract

The present study presents an assessment of neural computing approaches capability for predicting energy dissipation over stepped spillways under nappe flow regime. Using original experimental data-set, adaptive neuro-fuzzy inference system (ANFIS) and feed forward neural network (FFNN) techniques were utilized to detect the most influential parameters on nappe flow energy dissipation. The obtained results showed that the neural computing-based techniques have reliable performance in prediction of energy dissipation over stepped spillways under nappe flow regime. Also, the overall performance of the ANFIS model was superior to the FFNN model. Nonetheless, the most influential parameters on energy dissipation were identified as critical depth, height, and number of steps, respectively.