The performance of Diesel engines is influenced by the quality of the fuel sprays, which in turn determine the combustion process and the associated formation of  pollutants. The main objective of this project is the numerical modeling of the fundamental mechanisms which govern the spray formation and the chemical reactions, with the intent to establish a reliable, predictive tool to be used in the design process of combustion systems. The development and use of computational models demands extensive verification procedures which involve stability analyses with respect to inital and boundary conditions as well as the exploration of the range of the model applicability. An essential contribution towards these validations are experimental data obtained under controlled conditions by means of high-pressure and high-temperature combustion cells.

The backbone of the simulations is KIVA-3, a computer code which solves the three-dimensional conservation equations for mass, species, momentum  and energy in combination with a  k-epsilon-based turbulence model. The sprays are described by a stochastic evolution law which considers droplet collisions, evaporation, turbulent droplet/gas interactions and droplet breakups.  Several chemical reaction models of various degree of sophistication are available to study the combustion processes.

Recent new model developments include the enhanced Taylor analogy breakup (ETAB) atomization and drop breakup model and the simplified kinetics ignition (SKI) model. In the breakup model the fuel jet atomization is modeled as a cascade of drop breakups governedby Taylor's linear drop deformation dynamics, where each breakup event is modeled after experimentally observed drop breakup mechanisms.The SKI model utilizes one transport equation for a single ignition progress variable in combination with a reduced kinetic scheme.
 

Figure 1: Contour plots of the fuel mass fraction in the centerplane of the spray for ETAB computations without (left) and with an initial drop size distribution (IDSD). The IDSD simulates the jet surface stripping near the  nozzle exit and leads to an improved fuel vapor distribution  (dnoz=0.2 mm, tinj=1.3 ms, vinj=215 m/s, pgas=50 bar, Tgas=800 K).

Figure 2: Ignition locations and ignition delays as a function of the inverse temperature (pgas=4.5 MPa)  for ETAB/SKI computations and the experiments of Reuter.  The experimental and computational fluctuations are indicated by the shaded region and the dashed lines, respectively. The green line indicates the predicted ignition locations without use of the initial drop size distribution (IDSD).