A study of the knock limits of liquefied petroleum gas (LPG) in spark-ignition engines

© 2013 Dr. Kai J. MorgantiA substantial increase in the use of alternative transport fuels is considered by some to be an important part of our response to climate change. Liquefied Petroleum Gas (LPG) is one such fuel that shows significant potential to improve the efficiency and greenhouse gas emissions of conventional spark-ignition engines. However, there is still some uncertainty as to the best use of LPG in spark-ignition engines. This uncertainty can largely be attributed to variations in the composition of LPG, along with the susceptibility of these different mixtures to so-called ‘autoignition’. Often, this undesirable form of combustion leads to potentially damaging rates of in-cylinder pressure rise, and therefore should be avoided. This requirement places an upper limit on the efficiency of the spark-ignition engine. This thesis therefore aims to develop a fundamental understanding of the mechanisms responsible for the autoignition of LPG mixtures under conditions relevant to spark-ignition engines. A comprehensive experimental study of the susceptibility of different LPG mixtures to autoignition is first presented. This utilises the standard ASTM Research and Motor methods for liquid fuels, which are adapted to enable the autoignition propensity of different LPGs to be quantified in terms of the so-called ‘Research and Motor octane numbers’ (RON and MON respectively). The engine experiments are then examined numerically using a multi-zone combustion model that incorporates detailed chemical kinetics. This model is calibrated and validated using empirical data, and is used to investigate the key kinetic pathways leading to the autoignition of different LPG mixtures. The work presented in this thesis first demonstrates that the linear blending of octane numbers appears to be reasonable for most LPGs. It is also shown that almost all LPGs have higher RONs than both standard and premium gasolines. Importantly, imposition of the existing MON and vapour pressure requirements specified in the Australian LPG fuel standard further increases this advantage over all gasolines. This suggests that the existing fuel standards unnecessarily restrict the composition of commercial LPGs, given the widespread use of LPG in retro-fitted gasoline vehicles at present. The modelling results indicate that an accurate model of LPG autoignition in a spark-ignition engine needs to include several key phenomena. In particular, careful modelling of the in-cylinder heat transfer and residual nitric oxide concentration should be performed, in addition to the flame propagation modelling and autoignition chemistry. When these key phenomena are included in the model, the autoignition of most LPG mixtures could be modelled to within 1.0 crank angle degree (0.28 ms) of experiment