COMPUTATIONAL STUDY OF HIGH-PRESSURE LIQUID INJECTION PROCESS BY MEANS OF LES AND PANS APPROACHES

For internal combustion engines, the determination of combustion characteristics and subsequent emissions formation relies heavily on the fuel injection process. With the increasing demand for enhanced fuel efficiency and reduced emissions, it becomes vital to develop fundamental understanding of the physical process involved in the fuel injection process. In this study, an optimal numerical approach to predict the high-pressure liquid injection process in the context of industrial computations has been investigated. In particular, this study focuses on the respective performance of the partially averaged Navier–Stokes (PANS) and large eddy simulation models to predict turbulent igniting sprays. Both approaches are coupled with the widely used Lagrangian discrete droplet method for spray modeling. The results are validated against well-established engine combustion network (ECN) spray A case in reactive and nonreactive conditions. For reacting conditions, the flamelet-generated manifold (FGM) combustion model is employed in the present work. Comparative study and validation against experimental data have shown that the PANS turbulence model allows for coarser grids while still maintaining accurate results.