A Tool for Identifying Stationary State in Computational Fluid Dynamics Simulations of Unsteady Lube Oil Flows

The paper presents a probability density function (PDF)-based tool that can be used to guide the identification of the stationary state of computational fluid dynamics (CFD) simulation results. Specifically, the probability density distributions of local lube oil volume fractions (VFs) are used. Applications of the tool to the time-dependent turbulent flow of lube oil in a gearbox are reported. The CFD simulations were performed using a traditional finite volume method and a particle-based method, respectively. The rotational speeds were 3000 rpm and 6535.7 rpm for the driving and the driven helical gears, and the lubricant’s temperature was 37.5°C. Besides lube oil flow behavior and VFs, also discussed are flow patterns, churning-loss predictions, and individual contributions from the two gears. The results indicate that the tool was readily applicable to the simulated turbulent lube oil flows obtained by using the two CFD methods and was instrumental in assessing whether the lube oil flow had reached the statistically stationary state in the gearbox studied. In addition, the particle-based model was found to simulate the stationary lube oil flow more efficiently in terms of computational time.