The KIVA-II code is probably the most advanced CFD code for simulating diesel engine fuel spray, but the atomization sub model remains imperfect. For this reason, several break-up models exist in published literature. The poor physical understanding of liquid break-up prevents it from being modeled faithfully. In our 1997 SAE Paper, a new approach was proposed, based on a jet fully atomized when leaving the nozzle. It avoids the break-up description by fixing some fictional initial conditions for droplet velocity, average diameter and size distribution.
The originality lies in the choice of droplet injection velocity. The proposed initial velocity is roughly twice that normally used, i.e. the value found using the Bernoulli equation with a discharge coefficient of around 0.75. In this present work the intuitive empiricism of our last SAE paper is affirmed by a detailed physical analysis, which reveals aerodynamic drag to be the dominant factor in the considerable slowing down of the smallest droplets.
In experiments performed in a constant volume combustion chamber, the droplet size and velocity measures in the diesel spray collected thanks to PDA, agree well with those predicted by the code, downstream the atomizing zone. This means that the standard stochastic trajectography and evaporation procedures existing in KIVA-II, should not need to be altered.
Through three-dimensional simulations, it is shown that modeling the jet in this way, allows droplet size and velocity to be predicted satisfactorily, for gas temperatures of 200 up to 400 °C and pressures ranging from 2 to 4 MPa. Particular attention is paid to diameter distribution.