Design of a viable homogeneous-charge compression-ignition (HCCI) engine : a computational study with detailed chemical kinetics

摘要

The homogeneous-charge compression-ignition (HCCI) engine is a novel engine technology with the potential to substantially lower emissions from automotive sources. HCCI engines use lean-premixed combustion to achieve good fuel economy and low emissions of nitrogen-oxides and particulate matter. However, experimentally these engines have demonstrated a viable operating range that is too narrow for vehicular applications. Incomplete combustion or misfire can occur under fuel-lean conditions imposing a minimum load at which the engine can operate. At high loads, HCCI engines are often extremely loud and measured cylinder pressures show strong acoustic oscillations resembling those for a knocking sparkignited engine. The goal of this research was to understand the factors limiting the HCCI range of operability and propose ways of broadening that range. An engine simulation tool was developed to model the combustion process in the engine and predict HCCI knock and incomplete combustion. Predicting HCCI engine knock is particularly important because knock limits the maximum engine torque, and this limitation is a major obstacle to commercialization. A fundamentally-based criterion was developed and shown to give good predictions of the experimental knock limit. Our engine simulation tool was then used to explore the effect of various engine design parameters and operating conditions on the HCCI viable operating range. Performance maps, which show the response of the engine during a normal driving cycle, were constructed to compare these engine designs. The simulations showed that an acceptably broad operating range can be achieved by using a low compression ratio, low octane fuel, and moderate boost pressure. An explanation of why this choice of parameters gives a broad operating window is discussed. Our prediction of the HCCI knock limit is based on the autoignition theory of knock, which asserts that local overpressures in the engine are caused by extremely rapid chemical energy release. A competing theory asserts that knock is caused by the formation of detonation waves initiated at autoignition centers ('hot-spots') in the engine. No conclusive experimental evidence exists for the detonation theory, but many numerical simulations in the literature show that detonation formation is possible; however, some of the assumptions made in these simulations warrant re-examination. In particular, the effect of curvature on small (quasispherical) hot-spots has often been overlooked. We first examined the well-studied case of gasoline spark-ignited engine knock and observed that the size of the hot-spot needed to initiate a detonation is larger than the end-gas region where knock occurs. Subsequent studies of HCCI engine knock predicted that detonations would not form regardless of the hot-spot size because of the low energy content of fuel-lean mixtures typically used in these engines. Our predictions of the HCCI viable operating range were shown to be quite sensitive to details of the ignition chemistry. Therefore, an attempt was made to build an improved chemistry model for HCCI combustion using automatic mechanism-generation software developed in our research group. Extensions to the software were made to allow chemistry model construction for engine conditions. Model predictions for n-heptane/air combustion were compared to literature data from a jet-stirred reactor and rapid-compression machine. We conclude that automatic mechanism generation gives fair predictions without the tuning of rate parameters or other efforts to improve agreement. However, some tuning of the automatically-generated chemistry models is necessary to give the accurate predictions of HCCI combustion needed for our design calculations.