Magnetohydrodynamic flow of power-law ferrofluids with Fe ₃ O ₄ nanoparticles over a permeable moving wedge: a machine learning and local similarity solution

This study investigates the magnetohydrodynamic effects on the flow of a power-law ferrofluid over an expanding, permeable wedge, focusing on enhancing thermal efficiency using artificial intelligence-based analyses and local similarity techniques. Motivated by experimental evidence demonstrating non-Newtonian behavior in ethylene glycol suspensions with Fe3O4 nanoparticles, this work explores how magnetic nanoparticle-based ferrofluids augment heat transfer properties compared to conventional fluids. Governing equations were transformed into nonlinear ordinary differential equations and solved numerically using MATLAB’s bvp4c solver. Key performance indicators, including the skin friction coefficient, velocity and temperature profiles, and Nusselt number, were analyzed for various governing parameters. Additionally, artificial neural networks (ANNs) were utilized for optimized predictive modeling, leveraging a training-validation-testing split of 75%, 15%, and 10%, respectively. The results reveal that incorporating Fe3O4 nanoparticles into the base fluid significantly enhances the heat transfer rate and skin friction. Higher injection rates and magnetic parameters were shown to reduce fluid temperature but increase velocity near the wedge. These findings, validated through ANN-based predictions, demonstrate the novel potential of ferrofluids for industrial applications requiring efficient thermal management and reduced energy consumption.