This research aims to investigate the unsteady magnetohydrodynamic (MHD) flow and heat transfer of a Casson gold-silver/water hybrid nanofluid over an inclined stretching surface. The study is specifically designed to optimise localised hyperthermia and “thermal dose” delivery in cancer ablation by modelling the synergistic effects of bimetallic nanoparticles and non-linear transport phenomena.
The mathematical model incorporates the Quadratic Boussinesq Approximation (QBA), quadratic thermal radiation (QTR) and slip velocity. Governing equations are transformed via similarity variables and solved numerically using the MATLAB bvp4c algorithm. Additionally, an artificial neural network (ANN) is developed as a high-speed surrogate model to predict Nusselt number targets based on five physical input parameters.
Results indicate that an increasing Richardson number accelerates fluid velocity while enhancing convective cooling. Conversely, higher unsteadiness and Casson parameters act as suppressing agents, significantly attenuating the momentum boundary layer. A pivotal discovery is the significant divergence between radiation models: the quadratic Rosseland approximation predicts a 43.91% increase in the local Nusselt number across the studied range, whereas linear models fail to capture this sensitivity. The neural network model demonstrates exceptional predictive reliability with a correlation coefficient exceeding 0.9997.
This work establishes a pioneering, unified framework that integrates coupled non-linearities (QBA and QTR) within a Casson hybrid nanofluid system. The combination of high-precision numerical data and a validated ANN provides an agile, computationally efficient tool for real-time thermal management in advanced biomedical and engineering applications.
