Development of semi-empirical soot emission model for a CI engine

Abstract

Soot is one of the main harmful emissions of diesel engines that is mainly generated in the reacting fuel jet of diesel injection. Over 99% of the engine-out soot can be filtered by a diesel particulate filter (DPF). However, when the soot load of the DPF is high, a regeneration process that oxidizes the accumulated soot reduces fuel economy. A real-time soot estimation model can contribute to real-time feedback soot control under transient conditions to minimize the engine-out soot emission and frequency of DPF regeneration.A zero-dimensional engine-out soot estimation model for a diesel engine is developed in this study. The semi-empirical soot model considers both the formation and oxidation of soot. In the model, soot formation was correlated with the cross-sectional average equivalence ratio at the lift-off length of the fuel spray. The equivalence ratio at the lift-off length is an indicator of how much air and vaporized fuel are mixed as the fuel reaches the reaction zone. The mass of the injected fuel and combustion duration were also correlated with soot formation. The Nagle and Strickland-Constable mechanism, which calculates the soot oxidation rate was correlated with the soot oxidation in this study. The results of the soot estimation showed an R2 of 0.901 and root mean square error of 10.8 mg/m3 for steady-state experimental cases. The engine-out soot model was also combined with the in-cylinder pressure model proposed by the authors, and validated through the transient Worldwide Harmonized Light Vehicles Test Cycle (WLTC) mode. The estimates agreed with the measured soot, with an accumulated soot error of approximately 6% during the WLTC, even without using an in-cylinder pressure sensor. The soot model developed in this study can help minimize tailpipe-out soot emissions and improve fuel economy by influencing the real-time feedback control during transient and frequent DPF regeneration.