Entropy Generation and Heat Transfer Analysis of the Hydromagnetic Flow of Three Distinct Viscoelastic Fluids Over a Cone With Soret and Dufour Effects via Machine Learning

Abstract

Efficient heat and mass transfer is crucial in fields like energy systems and chemical processes, especially when dealing with non-Newtonian fluids, such as Casson, Maxwell, and Williamson. However, the interactions of thermal radiation, Soret, and Dufour effects in magnetohydrodynamic free convection over a vertical cone have not been thoroughly studied, nor has the impact of entropy generation on thermodynamic efficiency. This study aims to explore these interactions, focusing on how they affect heat and mass transfer and entropy generation in three types of non-Newtonian fluids. The governing equations are converted into dimensionless forms and solved using MATLAB's BVP4C solver, with results verified using an artificial neural network model. The main findings indicate that the Casson fluid has better heat transfer characteristics due to its lower viscosity at high shear rates. It was also found that magnetic fields can decrease velocity but increase the thermal and concentration boundary layers, which enhances diffusion rates. Additionally, thermal radiation, Soret, and Dufour effects significantly improve heat and mass diffusion, and the analysis of entropy generation highlights their importance for system efficiency. By combining numerical methods with machine learning, this study provides useful insights for improving heat and mass transfer in energy systems, chemical reactors, and manufacturing processes that use non-Newtonian fluids.