The Lagrangian or Discrete Droplet Model approach to spray modeling has been developed for more than 30 years and has been used successfully for many different types of injector, in both diesel and gasoline engines.
There has been tremendous evolution in fuel injection systems during this time, for example, common rail systems and multiple injections have become commonplace and injection pressures have increased by an order of magnitude; models in STAR-CD have also evolved to be able to model accurately the sprays produced by these systems.
The characteristics of the spray are strongly dependent upon the primary breakup process. This, in turn, is determined by nozzle geometry, the engine operating condition, fuel injection rate and fuel and gas properties.
Although STAR-CD contains a number of semi-empirical correlations to help in the determination of initial conditions for Lagrangian calculations – and these have been used successfully for many years – CD-adapco has also pursued two alternative routes to introduce more fidelity into resolving the in-nozzle flow and the breakup of the liquid jet.
The first of these approaches is the ELSA (Eulerian- Lagrangian Spray Atomization) model in which the flow inside and immediately downstream of the nozzle hole is modelled as a continuous Eulerian liquid phase. This is coupled to the Lagrangian model by introducing a transition region between the Eulerian and Lagrangian regimes in which the primary breakup process is modelled via an equation for liquid surface area density. In this way, the effects of nozzle design or injection and engine operating conditions can be seen immediately.
The model has shown excellent agreement when compared to rig-based experimental data over a wide range of operating conditions and for both non-evaporating and evaporating sprays and is now being used for in-cylinder spray and combustion simulation. The figure above shows the main principles of the method.
An even more detailed and fundamental approach is to compute the in-nozzle flow and primary breakup processes using a method based on Volume-of-Fluid (VOF) and Large Eddy Simulation (LES). In principle, this technique models all of the turbulent scales that are important for disruption and breakup of the continuous liquid phase and is thereby able to directly calculate rather than model the primary breakup process itself.
By continuously capturing information about the size, velocity, position etc. of all the “blobs” of fluid generated from the liquid core, statistics of size and velocity can be assembled in a similar fashion to the way in which this would be done experimentally. The raw information can also be conditionally sampled, for example based on position, offering a comprehensive approach to providing the initial conditions for a Lagrangian calculation based on nozzle design.
Increasing injection pressures coupled with higher thermal loading have also resulted in impingement regimes where the surface temperature is above the liquid saturation temperature. New modeling has been introduced into STAR-CD to give more realistic behaviour under these high temperature conditions up to and beyond the Leidenfrost temperature.
Although general multicomponent modeling of evaporation has been available in STAR-CD for many years, more recently, 2-component fuel mixtures such as E10 (10% ethanol, 90% gasoline) or B30 (30% biodiesel, 70% diesel) have come into prominence and the evaporation and mixing of these is modeled correctly to yield a vapour field that can be used directly by the ECFM and PVM