In times of ever stricter exhaust emission regulations, the importance of
alternative combustion processes in internal combustion engines continues to
grow. One approach to create a combustion progress which produces low
CO2, soot, and methane emissions is the “High-Pressure Dual-Fuel”
(HPDF)-combustion. Here, the direct-injected methane is ignited by a small
amount of pilot-diesel and burns in a diffusive combustion mode.
This study describes the development of a three-dimensional computational fluid
dynamics (3D-CFD) model for the HPDF-combustion. A Reynolds-Averaged
Navier-Stokes (RANS) approach with k-epsilon modelling for turbulence was chosen
for the calculation of the flow field. The pilot fuel injection is implemented
by using Lagrangian Particle Methods, whereas the gas injection is a mass flow
boundary which is derived from measurements of the injector. The model is
validated using data from a fully optically accessible single-cylinder research
engine. The flow field is compared with particle image velocimetry (PIV) data
taken before the start of injection (SOI). Concerning pilot injection, a grid
convergence study is conducted and an optimization is developed to reduce
computational costs. The penetration length of the liquid fuel spray is
validated against Mie-scattering images which are taken during the
“Pilot-Diesel-only” experiments in the fully optical single-cylinder research
engine.
The ignition and combustion is modeled via detailed chemistry, which is solved
using the commercial Software CONVERGE and the SAGE chemistry solver. The flame
liftoff length of the pilot-diesel and the ignition and combustion of the
underexpanded gas jets are validated using high-speed imaging of flame
luminosity and OH* chemiluminescence. It can be shown that the used
n-heptane mechanism is capable of correctly reproducing the
trends in the ignition and combustion process.
