Objective Magnetically influenced bioconvective flow of ternary nanofluid induced by the expansion of curved surface by incorporating thermophoresis, Brownian motion, chemical species, activation energy and motile microbes to elucidate complex thermal fluid transport phenomena. Method ology: The mathematical model describing flow mechanism was formulated in sense of PDEs (partial differential equations), which are converted in ODEs (ordinary differential equations) by employing similar set of variables. Numerical technique by integrating the shooting method and RK-4 approach is employed to obtain the outcomes of study. Afterwards, neuro-computing model is designed to forecast Nusselt number for mono, hybrid and ternary nanoparticles comparatively. Key findings Findings of the analysis indicate that velocity of fluid intensifies by uplifting curvature factor while thermal profile goes down. Thermophoretic and Brownian diffusion factors cause the temperature of the fluid to rise but lower the associated flux. Higher curvature and activation energy factors elevate concentration distribution, whereas microbe density depreciates versus Peclet and bioconvective Lewis numbers. The MSE values obtained during training (2.79e-08, 7.63e-08, and 1.55e-07) demonstrate the model's robustness. Applications It is concluded that heat and mass transportation phenomenon is superior with the induction of ternary nanoparticles as compared to mono and hybrid, giving valuable insights for the design of improved thermal energy storage and bioconvective transference mechanism in engineering and biomedicine utilizations.

Kerosene oil-based penta-hybrid nanofluids have attracted significant attention because of their improved thermal conductivity, mechanical stability, and potential use in advanced heat transfer systems. In this study, a Physics-Informed Neural Network (PINN) Analysis of Kerosene-Based Penta-Hybrid Nanofluid Flow and Heat Transfer, is performed to understand the flow pattern and heat transfer characteristics of a nanostructured fluid loaded with several types of nanoparticles. This new approach integrates the physical soundness of governing transport equations with the deep learning’s ability to predict, for modeling nanofluid flow and heat transfer phenomena. The initial partial differential equations governing momentum and energy transfers are converted via appropriate similarity transformations into dimensionless ordinary differential equations. These are then used as the key ingredients (embedded alongside boundary conditions) of the loss function of Physics-Informed Neural Network to make the model’s output comply with physical laws. Variations in parameters leading to changes in velocity and temperature distributions are explored, and to check the correctness and trustworthiness results are compared with classical numerical solutions and previously published data. The findings indicate that the PINN approach accurately characterizes the complex flow and heat transfer features of kerosene-based penta-hybrid nanofluids. By incorporating physics-based modeling with deep learning, reliance on extensive numerical data is diminished while excellent predictive capability is preserved. This research draws attention to PINN-based methods as promising and powerful instruments for the study of high-tech nanofluid products and direct engineering of superior heat exchangers, refrigeration systems, and thermal management devices.