Numerical analyses of EGR techniques in a turbocharged spark-ignition engine

Exhaust Gas Recycle (EGR) is an interesting technique, which is more and more adopted in spark-ignition engines in order to reduce both NOx emissions and fuel consumption in certain engine operating points. A fraction of exhaust gases recycled to the intake produces a significant charge dilution which decreases the combustion chamber temperature and, as a consequence, the NO formation rates. Furthermore, charge dilution allows reducing pumping losses at part load and heat losses through the walls increasing the engine efficiency. In this paper, numerous numerical analyses have been carried out in order to widely estimate the potential of exhaust gas recycle in a downsized, turbocharged spark-ignition engine. Many engine operating points have been examined, at different speed and load. Major attention has been paid to the portion of the engine map characterized by low rotational speed and WOT operation. Thus, the most severe conditions for knock onset have been deeply investigated. The combustion process has been modeled by means of a three-dimensional computational code. In particular, an in-cylinder analysis has been carried out to evaluate the effects of charge dilution on flame propagation, knock resistance and exhaust gas temperature distribution. Furthermore, adopting EGR, main engine control parameters (fuel to air ratio, spark advance and boost pressure) have been recalculated and the overall performances have been computed by a one-dimensional model. In this case, a simplified combustion model, tuned by means of 3-D computations, has been utilized. The models here presented have been validated by several comparisons with experimentally obtained results. The optimization of engine parameters, together with the evaluation of knock risks, allowed recovering, at some extent, the engine performance in absence of exhaust recycle. At the same time, reductions in both fuel consumption and pollutant emissions have been obtained. (C) 2012 Elsevier Ltd. All rights reserved.