Combination of Computational Fluid Dynamics, Adaptive Neuro-Fuzzy Inference System, and Genetic Algorithm for Predicting Discharge Coefficient of Rectangular Side Orifices

Abstract

Side orifices are used to divide and adjust flow into aeration ponds, sedimentation reservoirs, flocculation units, and other hydraulic and environmental areas. In this study, the discharge coefficients of side orifices are estimated using the adaptive neuro-fuzzy inference system (ANFIS) and a hybrid of ANFIS and a genetic algorithm (ANFIS-GA). The genetic algorithm is used to optimize the membership function of ANFIS. To predict the discharge coefficient, the ratio of the main channel width to the side orifice length (B∶L), the ratio of the side orifice height to its length (W∶L), the ratio of the flow depth in the main channel to the side orifice length (Ym∶L) and the Froude number (F) are considered. Eleven different models are introduced for each of the ANFIS and ANFIS-GA models to calculate the discharge coefficient. The side orifice discharge is simulated using computational fluid dynamics (CFD). To model the flow field turbulence, the standard κ-ϵ and renormalization-group (RNG) κ-ϵ turbulence models are used. According to the CFD model results, the RNG κ-ϵ turbulence model simulates the flow field turbulence with more accuracy. By analyzing the results of the ANFIS, ANFIS-GA and CFD models, the ANFIS-GA model is introduced as the best model in terms of B∶L, W∶L, Ym∶L, and F.