Nonlinear Thermophysical Dynamics of MHD Hybrid Nanofluids: A Multiparametric Fractional Framework Assisted By Artificial Neural Networks

Abstract

The study of non-Newtonian fluids has drawn much scholarly attention because of the wide range of applications they have in industrial processes, such as plastic, lubrication systems, and in mining operations. When combined with magnetohydrodynamic effects their importance widens to more advanced technological and biomedical applications such as magnetic resonance imaging, medical diagnostics, and hyperthermia treatment. Recent investigations have proven that the addition of nanosized particles into base fluids improves the thermal performance and the heat transfer efficiency. Nevertheless, heat transfer in realistic mechanical systems is affected by a huge number of interacting physical parameters, each with different contribution to overall behavior. In the present study, a sensitivity analysis is conducted to study the effect of some important parameters, which include internal heat generation, magnetic field strength, and Casson nanofluid properties. The governing equations are derived from fundamental conservation laws and transformed into a dimensionless form for generalization. The model is generalized by using Fourier’s and Fick’s laws. Analytical treatment is achieved using the Laplace transform technique, while the Levenberg–Marquardt algorithm is employed for numerical prediction and validation. Results demonstrate high accuracy with mean-squared error values below 10−10 especially for a fractional nanofluid Casson flow model. The Levenberg–Marquardt algorithm is utilized, with 70% of the data used for training and 15% allocated for validation and testing. The findings reveal that momentum increases with decreasing fractional parameters. Table 4 gives the validity of present work with already published work.