Analysis of Natural Convection and Radiation from a Solid Rod Under Vacuum Conditions with the Aiding of ANFIS

The present paper represents an experimental and theoretical study of the characteristics of natural convection and radiation heat transfer from a solid cylindrical rod inside a closed vessel under vacuum conditions. The effect of vacuum pressure on the heat transfer rates from the heating element by convection and radiation is clarified. The experimental results are used for Artificial Neural Network (ANN) training which is employed in the Adaptive Neuro Fuzzy Inference System (ANFIS) scheme to predict the heat transfer characteristics. The heat transfer predictions are performed for the solid rod under numerous heat fluxes for each vacuum pressure and the computed outcomes are compared with that predicted from the ANFIS. Outcomes indicated that for a constant heat flux, the convection heat transfer decrease with the reduction in pressure inside the vessel, while the radiation heat transfer will increase. Three methods of ANFIS are deemed in this work; grid partition, subtractive clustering and fuzzy C-mean clustering (FCM). However, the three methods appear an adequate fitting with the experimental data due to the high ability of these schemes at formulating input-output relations. A comparison with conventional correlation for the Nusselt number (Nu) has been implemented which proved the effectiveness and robustness of artificial intelligence in the generation of a direct relation between the effective parameters and the objective functions. Also, the main factors of ANFIS such as the number of membership functions (MFs), the number of clusters and the radius of clusters have been optimized to reach best performance.