Heat transfer analysis of non-Newtonian hybrid nanofluid flow over a stretching surface using an artificial neural network

The present study investigates the role of Gold-Fe 2 O 3 -Fe 3 O 4 nanoparticles mixed in blood with magnetohydrodynamics Casson fluid flow through a porous stretching surface along with Heat source/sink. In the present study considered Gold-Fe 2 O 3 -Fe 3 O 4 are nanoparticles along with blood as a base fluid. The governing highly nonlinear PDEs are converted into ODEs with the help of suitable self-similarity variables. The highly nonlinear ODEs are solved using the BVP4c numerical method in MATLAB software. The outcomes of the main active physical parameters of energy and velocity profiles are explained by graphs and tables. The thermal profile exhibits increasing behaviour when the thermal radiation parameter varies. The velocity graph is decreasing with the variations of the magnetic field parameter. The artificial neural networks (ANNs) model is highly forecasting, with a minimal mean squared error of 6.0494 × 10⁻⁷ and a regression coefficient of R 2 = 1, indicating close to perfect consistency with numerical data. Comparative evaluation with previous literature demonstrates good consistency, guaranteeing the correctness and dependability of the current approach. Consistent with earlier research, the results validate the model's correctness and provide light on how to optimize biomedical fluid structures based on nanoparticles. Overall, the thermal efficiency and flow characteristics of blood-based mixed nanofluids are used in various sectors like biomedical science, targeted medication administration, and thermal therapy.