Using ANFIS technique in prediction of reverse flow in a solar chimney

Abstract

In solar chimney, in order to save resources in Computational Fluid Dynamics (CFD), this study aims to combine Adaptive Neuro-Fuzzy Inference System (ANFIS) to predict reverse flow and investigate the ventilation performance. The inputs to the model are heat flux in the range of 200-1000W/m2, gap of 0.04-0.25m and height of 1-1.4m, while outputs are penetration depth and mass flow rate. The results of R2 and RMSE are reasonable for training and testing data, hence the ANFIS model is validated.