Akademik Veri Yönetim Sistemi
Arabian Journal for Science and Engineering, 2026 (SCI-Expanded, Scopus)
The growing need for enhanced thermal regulation in both electric and internal combustion engine vehicles has driven significant research efforts toward the development of next-generation heat exchanger technologies. Motivated by the pressing need for thermally consistent and high-fidelity prediction tools in contemporary automobile systems, this study introduces a novel hybrid framework by coupling empirical results from ZnO–water nanofluid experiments with physics-informed neural networks (PINN). This approach is crucial for connecting empirical data with fundamental physical rules for accurate thermal prediction. The experimental investigations were carried out across a range of operating parameters, including nanoparticle volume concentrations from 0 to 0.1%, inlet temperatures between 40 °C and 80 °C, and volume flow rates varying from 5 to 11 LPM. Major results revealed that the highest thermal enhancement was achieved at a ZnO concentration of 0.025%, which led to an increase of up to 213% in the convective heat transfer coefficient and a 214% improvement in the Nusselt number relative to pure water, particularly under low-temperature conditions. Beyond this concentration, performance deteriorated due to agglomeration and increased viscosity. Using the experimental data, both artificial neural network (ANN) and physics-informed neural network (PINN) models were developed to predict key parameters such as the Nusselt number and heat transfer coefficient. The PINN model outperformed the ANN, yielding approximately 7.6% higher R2 for the heat transfer coefficient (0.9167 vs 0.8354) and 8.7% higher R2 for Nusselt number predictions (0.8697 vs 0.7901). These findings highlight the importance of creating a high-fidelity digital twin architecture that addresses the ‘black-box’ constraints of traditional AI, offering a comprehensive solution for real-time optimization and monitoring of advanced automobile cooling systems.