Investigation of Density, Viscosity, and Surface Tension\nat 293.15 K along with the Flash Point, Freezing Point, and Refractive\nIndex at 293.15 K, and the Distillation Curve of Hydrocarbon Fuels:\nExperimental Measurements and Interrelationship bet

Understanding the underlying relationship\nbetween the oil composition\nand properties and predicting the properties of kerosene and gasoline\nare crucial for its application. Density, viscosity, surface tension,\nfreezing point, viscosity, and flash point are fundamental physical\nproperties, but their relationships with compositional characteristics\nremain unclear. In this work, artificial neural networks and Shapley\nadditive explanation analysis are applied to link the six properties\nwith the distillation curve and refractive index, which describe the\nheavy, light, and average oil composition characteristics. The most\nimportant features for the density, viscosity, surface tension, flash\npoint, and freezing point are the refractive index, 10% boiling temperature,\n70% boiling temperature, initial boiling temperature, and 50% boiling\ntemperature, respectively. Linear regression fits the density, viscosity,\nand flash point, while nonlinear regression fits the surface tension\nand freezing point. This article establishes rapid prediction models\nfor these properties and provides a reference for revising the mixing\nrules and improving fuel properties.