Original: https://doi.org/10.1016/j.csite.2025.106097
This study looks at improving car radiator performance using a hybrid nanofluid made of Al₂O₃ and SiO₂ nanoparticles in a 50:50 water-ethylene glycol mixture. Using detailed computer simulations (CFD) and symbolic regression, the research examined how different air speeds and coolant flow rates affect heat transfer, pressure drop, and friction.
The hybrid nanofluid improved heat transfer: the Nusselt number increased by 2 % when nanoparticle concentration rose from 0.05 % to 0.2 %. At the highest air speed of 39 m/s, the maximum heat flux reached 160 kW/m². Along the fins, heat flux increased linearly with wall temperature but also showed fluctuations due to flow separation and vortex formation.
Raising the coolant flow from 7 to 10 L/min boosted thermal performance by 40 %, with the Nusselt number peaking at about 90 in the tube center. Higher flow rates improved heat transfer but increased the friction factor by 11 % due to extra pressure drop. Still, the overall performance enhancement coefficient (PEC) remained almost constant (~1.002), showing a good balance between heat transfer improvement and added resistance.
Symbolic regression produced highly accurate formulas to predict heat flux along the fins, saving computational time and reducing the need for long CFD simulations while keeping R² values above 0.996. This approach makes radiator performance modeling faster and more cost-effective.