Implementation of an Adaptive Neuro-Fuzzy Inference System Technique for Evaluating Hydraulic Performance of Porous Gabion Type Spillways
摘要
Porous gabion spillways offer several advantages, making them significant for sustainable water management and environmental conservation. In this study, based on 540 laboratory test runs conducted under controlled experimental conditions, the dataset of flow over a gabion stepped spillway was assessed using the soft computing technique Adaptive Neuro Fuzzy Inference System (ANFIS), which relates the input and output variables. The ANFIS model achieved very high predictive performance with R2 = 0.985 (training) and 0.976 (testing), with minimal RMSE values ranging from 0.017 to 0.024 and a mean relative error of only 1.25%. Predicted and observed energy dissipation values showed strong agreement in terms of mean, range and variability. These findings confirm that ANFIS is a reliable and robust tool for accurate prediction and optimization of spillway energy dissipation, which can be practically applied in the design and performance enhancement of efficient and eco-friendly hydraulic structures.