Radiative magneto-thermal modeling of tri hybrid nanofluid in blood: A Biomedical intelligent framework for...

The emergence of tri hybrid nanofluids has revolutionized thermal engineering by remarkably increasing the transport capabilities because of the synergistic effects of multiple nanoparticles. Motivated by their superior performance and optimistic biomedical applications especially in hyperthermia based cancer therapies, this study scrutinized the heat transfer performance of a magnetized radiative tri hybrid nanofluid containing copper oxide, titanium oxide, and silicon oxide as nanoparticles distributed in human blood modeled as Casson fluid. The adopted model comprises nonlinear form of radiative heat transfer with convective boundary conditions to present realistic physiological and engineering scenarios. A novel intelligent computing framework is constructed by employing unsupervised form of Artificial Neural Networks and the training process is optimized using a hybrid meta-heuristic scheme which is an amalgam of global exploration ability of Genetic Algorithms with the powerful local refinement of Sequential Quadratic Programming. By adopting appropriate transformations, the governing nonlinear partial differential equations are reduced into a couple of ordinary differential equations which are then solved by varying leading physical parameters to investigate tri hybrid nanofluid flow and thermal aptitude. It is observed that the thermal profile depicts a 14 % decline for higher values of Prandtl number which is an indication of enhanced thermal resistance because of less thermal diffusivity while a 15 % growth in temperature is detected with an uplift in the radiation parameter, both are critical for attaining effective and controlled localized heating in hyperthermia cancer treatment. All the obtained results are comprehensively depicted through graphical and tabulated datasets by comparing the reference solution obtained via the Adams numerical method. The reliability and accuracy of the designed framework are validated via numerical comparisons, statistical error estimation and comprehensive convergence analysis which confirm its robustness and suitability for biomedical heat transfer applications.