Abstract

This research investigates the influence of magnetohydrodynamics (MHD) on studying the heat transport in an Ethylene Glycol-Based (EG) ternary Carreau nanofluid flow along a needle-shaped geometry. Ternary hybrid nanofluids are composed of well-designed nanoparticles with enhanced thermal conductivity that significantly contribute to improvements in energy efficiency and offer useful insights for practical applications, including industrial processes and thermal management systems. The governing nonlinear partial differential equations (PDEs) are converted to nonlinear ordinary differential equations (ODEs) by using suitable similarity variables, which are further solved using the bvp4c scheme. The findings are trained for further advanced prediction using a dual hidden layer neural network, a novel artificial intelligence approach. i.e., Levenberg-Marquardt neural network (LM-NN). Smooth agreement was attained when the results were compared using both approaches. The influence of the critical parameters on the fluid velocity and temperature profile is investigated. According to the findings, the velocity profile decreases with intensification in the numeric growth in Weissenberg number and magnetic parameter, whereas higher radiation values elevate the temperature profile of ternary hybrid nanofluids. Nusselt number goes up by 16.59 % and 9.71 % with increasing radiation parameter from 0.2 to 0.9 and heat generation parameter from 0.1 to 0.3, but goes down by 18.89 % with increasing flow index parameter from 0.5 to 0.7. Also, the Skin coefficient drops by 7.49 %, 27.95 %, and 12.84 % with increasing Weissenberg number, infinite shear rate, and magnetic parameter, respectively. Moreover, due to the combined effect of ternary nanoparticles with different concentrations of volume frictions, the ternary nanofluid showed an increased heat transmission rate compared to the bi-hybrid and nanofluid. The plots of the regression analysis, training state, and performance show the accuracy of the considered machine learning model. The outcomes of the current study provide insight to improve thermal system optimization for use in advanced heat transfer technologies, biomedical devices, and energy systems.