Impact of Fuel Properties on Partially Premixed Combustion

Abstract: Compression ignited engines generally have higher efficiency than spark ignition engines. However, the most common combustion concept in compression ignited engines, conventional diesel combustion, struggles with high levels of particulate matter and NOx emissions. Therefore, new combustion strategies are used in compression ignited engines to reduce engine-out particulate matter and NOx emissions. One such promising strategy, Partially Premixed Combustion (PPC), is analyzed in this thesis. Independently of its properties, if a combustion concept should be possible to put in production in a near future it is important that it can utilize the available fuel on the market. Hence, it is important to understand how PPC responds to fuel properties, especially research octane number value and ignition quality. The objective of this work is to gain knowledge of fuel effects on the combustion and engine-out emissions in PPC. The topics are divided into three main parts. The first part investigates the impact of research octane number (RON) and fuel composition on combustion, emissions and load range for PPC. A clear interaction between RON and ignition delay was found; higher research octane number results in longer ignition delay. Also, an interaction between ignition delay and fraction of low temperature reaction was found, a prolonged ignition delay results in higher fraction of low temperature reactions. A load range comparison between four gasoline-like fuels and Swedish diesel class 1 (MK1) at a specific air-fuel ratio and exhaust gas recirculation rate was performed. A clear interaction between RON and the feasible load range was found, fuels with lower RON values could be operated at lower loads. The second part investigates the effects of engine operating parameters on combustion and emissions. As expected, the inlet oxygen concentration had a significant effect on ignition delay; reducing the inlet oxygen concentration resulted in a longer ignition delay. Also, when low inlet O2 concentration is used in combination with high RON values the effect on ignition delay is enhanced compared to the added effects of each parameter. All emissions were influenced by inlet O2 concentration; CO and HC emissions were increased with reduced inlet O2 concentration, while the trend was the opposite for NOx. A more retarded combustion phasing gave higher levels of CO and HC and lower levels of NOx. However, at lower inlet O2 concentrations the NOx values were suppressed to very low levels independently of CA50. The emission levels of smoke decreased with increasing injection pressure. The third part predicts the RON and motor octane number (MON) for oxygenated surrogate gasoline fuels using quadratic regression models. The models perform remarkably well in predicting RON and MON for any fuel blend of n-heptane, toluene, ethanol and isooctane. A similar model for Ignition delay is also presented in the thesis. By using these regression models for predicting ignition delay and RON value, it is possible to design a specific fuel to meet both emissions legislations and reducing fuel consumption.