In modern automotive internal combustion engines the atomization and spraying processes of liquid fuel is an important mechanism by which the fuel is mixed effectively with the combustion chamber gas. The homogeneity of the fuel-air mixture charge and the timing to achieve it affect the characteristics of the subsequent combustion in a number of ways. Experimentalists have found that the atomization process is influenced by the fuel delivery system, specifically, the operating conditions of fuel injection and the internal geometry of the injector nozzles. Therefore, modeling researchers must develop spray models that properly take into account these effects from the injectors.
A Computational Fluid Dynamics (CFD) model is developed to simulate internal injector nozzle flows and the external-nozzle atomization and sprays in an integrated way. In light of the continuous nature of the flow within and near the nozzle, an Eulerian flow solver is developed and applied in these regions. This flow solver models compressible two-phase flows with general conservation laws of fluid dynamics. Differences in the thermodynamic states of liquid and gas phases are properly modeled with an Equation of State (EOS). The chosen numerical methods are sufficiently robust to handle pressure and density gradients up to 1000:1 and are able to cover the typical operating ranges of modern diesel and gasoline injection systems.
A phase equilibrium model is developed based on the fundamental thermodynamic laws. It is implemented into the Eulerian flow solver to predict phase change in the flows?in particular, the cavitation of liquid fuel within the fuel injector nozzle and how it influences the external flows. A number of test problems are simulated to verify the numerical methods and validate the proposed models. These include two-phase shock tube problems, a shock-interface interaction problem, a converging-diverging nozzle flow problem, and most importantly, high-pressure fuel injection problems.
The Eulerian flow solver requires the CFD mesh size to be smaller than the characteristic phase interface length scale in order to apply the continuum fluid assumption. This requirement is challenged beyond a certain distance from the nozzle where the spray is dispersed. Here the mesh size affordable by engineering calculation is too coarse for the Eulerian solver to be applied. At this stage, the Eulerian liquid phase can be transitioned into Lagrangian discrete particles to model the dispersed spray droplets. The modeling of the sub-grid droplet size at this transition stage is based on local balances between turbulent kinetic energy and surface energy.
The Eulerian flow solver coupled with the phase equilibrium model and the Eulerian-Lagrangian transition model provides an engineering CFD approach to study the effects of injector nozzle flows on the atomization and sprays.