Multi-Dimensional Modeling of Direct-Injection Diesel Spray Liquid Length and Flame Lift-off Length using CFD and Parallel Detailed Chemistry

Abstract

<div class="htmlview paragraph">Recent measurements by Siebers et al. have shown that the flame of a high pressure Diesel spray stabilizes under quiescent conditions at a location downstream of the fuel injector. The effects of various ambient and injection parameters on the flame “lift-off” length have been investigated under typical Diesel conditions in a constant-volume combustion vessel. In the present study, the experiments of Siebers et al. have been modeled using a modified version of the KIVA-3V engine simulation code. Fuel injection and spray breakup are modeled using the KH-RT model that accounts for liquid surface instabilities due to the Kelvin-Helmholtz and Rayleigh-Taylor mechanisms. Combustion is simulated using Convergent Thinking's recently developed detailed transient chemistry solver (SAGE) that allows for any number of chemical species and reactions to be modeled. While detailed chemistry is believed to be an accurate methodology for modeling Diesel combustion, in the past the extensive run times rendered it impractical. To expedite the calculations, SAGE has been implemented into KIVA using the Message-Passing Interface (MPI). This implementation allows for the chemical reactions to be simulated in parallel on multiple CPUs. An n-heptane mechanism was used to model Diesel fuel ignition and combustion.</div> <div class="htmlview paragraph">The improved KIVA-3V code was used to simulate the spray-bomb cases over a wide range of injection pressures, nozzle hole sizes, ambient temperatures, and ambient densities. In general, excellent agreement was obtained between the measurements and simulation results for liquid length and flame lift-off length.</div> <div class="htmlview paragraph">To further validate the chemical kinetics model, simulations of a heavy-duty direct injection Diesel engine were conducted. The results indicate that the detailed chemistry model is able to accurately predict ignition delay and cylinder-averaged pressure for a range of start of injection timings.</div>