Lubricant oils are used in many technical applications,
e.g., in
automotive, refrigeration technology, and many other industries. Reliable
knowledge of thermophysical properties of such oils is essential,
but modeling all of the important properties, including density, phase
behavior, heat capacity, entropy, enthalpy, viscosity, and thermal
conductivity, remains a key challenge today. To tackle this challenge,
we propose a novel modeling approach based on treating the lubricant
oil as a quasi-pure fluid, setting up a simple set of equations for
all of the important properties of the oil, and developing a parameter
fitting procedure using a minimal set of experimental data. This simple
model set includes the Patel–Teja–Valderrama cubic equation
of state, a simple expression for the ideal gas isobaric heat capacity
as a linear function of temperature and residual entropy scaling for
viscosity and thermal conductivity. To fit some of the parameters
in this model set, two extra models are required: Raoult’s
law of boiling point elevation and the modified Rackett equation.
As a result, fewer than 20 (at least 12) experimental points are needed
to fit all 15 parameters of a pure or quasi-pure component, and one
experimental mixture bubble-point pressure is required to enable a
binary system prediction. For pure or quasi-pure components, in the
liquid phase and not in the vicinity of the critical point, this modeling
approach has an uncertainty over large temperature and pressure ranges
of less than 7% for viscosity and less than 3% for all other properties.
For binary mixtures, except for viscosity, the modeling approach still
yields good predictions for all other properties, typically within
8%.