Theoretical Investigation on the Influence of Physical Parameters on Soot and NOx Engine Emissions

Abstract

CFD simulations need a certain number of parameters to calibrate both empirical and analytical models. The present investigation aims at identifying the effects of these parameters on the numerical prediction of a modified version of Kiva 3V code, which includes the use of the RNG k-ε model for turbulence, the gas/wall convective heat transfer model proposed by Han, Kelvin-Helmholtz Rayleigh-Taylor spray injection and breakup models. Ignition delay was modeled with the Shell model, whereas the laminar-turbulent characteristic time model was used for combustion. Soot formation and oxidation were calculated using Hiroyasu and Nagle and Strickland-Constable models, respectively. NOx was predicted by using the extended Zel’dovich mechanism. This study was carried out for a common-rail direct injection, small-bore Diesel engine, including the investigation of both numerical and physical parameters. Numerical parameters are intended to be variables related to breakup, turbulence, and combustion models that are adjusted according to grid resolution, engine and injection system geometry, and operating conditions. In particular, the effect of laminar and turbulent time scales, characteristic breakup length and time scales, initial turbulence kinetic energy density, initial swirl velocity profile, on engine emissions was analyzed. The investigated physical parameters were initial swirl ratio, air water content, Schmidt number for mass diffusion. All simulations were performed by changing one of the above parameters at each run and keeping approximately the same pressure and heat release rate curves. Results show that similar pressure vs. crank angle curves can be obtained with different values of these parameters but they lead to very different values of predicted emissions levels. In particular, changes of laminar and turbulent characteristic time resulted in a strong influence on NOx emissions but their effects on soot levels were minor. Mass diffusion characteristics (e.g. Schmidt number) were found to strongly affect both soot and NOx emissions. Spray parameters were found mainly to affect soot formation. Furthermore, NOx and soot emissions showed a dependence on swirl ratio and velocity profile.