An innovative virtual fuel model for predicting the thermophysical properties of supercritical RP-3 aviation kerosene

Accurate prediction of the thermophysical properties of endothermic hydrocarbon fuels is essential for designing regenerative cooling systems in hypersonic vehicles, as it directly affects the evaluation of waste-heat conversion and energy utilization efficiency. However, accurate prediction methods for supercritical RP-3 aviation kerosene, a representative endothermic fuel, remain limited. To address this issue, this study modifies the property calculation approach and proposes a virtual fuel model for the first time, which treats RP-3 as a pseudo-pure fluid with parameters optimized using a genetic algorithm. The results demonstrate that the improved method is broadly applicable to various single-component hydrocarbon fuels. For n-decane, n-dodecane, and JP-10, deviations in property prediction are reduced by up to 91.41%. For RP-3, the virtual fuel model outperforms six existing surrogate models, reducing the maximum deviation by 97.95%. Under various pressures, prediction deviations remain below 4% for density (below 700 K), 6% for isobaric heat capacity (below 700 K), 0.4% for thermal conductivity (below 513 K), and 4% for dynamic viscosity (below 700 K), validating the accuracy over a wide temperature range. Numerical simulations further demonstrate that the model reliably captures flow and heat transfer characteristics under strong buoyancy, with predicted heat transfer enhancement and deterioration closely matching experimental data. Furthermore, the virtual fuel model exhibits strong generalization capability and is further validated for stability and scalability, demonstrating its applicability to other multi-component hydrocarbon fuels. The proposed property prediction method offers robust support for regenerative cooling investigations and provides a novel approach for studying hydrocarbon fuels.