Latest revision as of 11:54, 11 February 2026

Abstract

Minimizing entropy production is critically important, particularly in nanofluid flows. Applying this principle to flows with porous cavity structures helps optimize heat transfer applications and enhance system efficiency. In this study, the entropy production and heat transfer characteristics of a hybrid nanofluid composed of Al2O3-Cu particles suspended in water were investigated using machine learning. The nanofluid was analyzed in the context of convection–radiation interaction flow within a porous cavity.An artificial neural networkmodel was developed to predict the average Nusselt number, Bejan number, and entropy production as functions of the Hartmann number and inclination parameters. The Bayesian Regularization algorithm was employed to train the multilayer perceptron network model. Prediction results obtained from the model with 10 neurons in the hidden layer were compared with the target values and showed excellent agreement.The developed artificial neural network model successfully predicted the Nusselt number, Bejan number, and entropy productionwith average deviation rates of−0.007%,−0.11%, and 0.0002%, respectively.

Document