Cryogenic fuelling of engines and fundamental droplet evaporation behaviour

Liquefied natural gas (LNG), from fossil and renewable sources, is currently in use as an alternative transportation fuel with significant environmental and economic advantages compared to petroleum fuels. However, current LNG vehicles waste the potential  thermomechanical exergy of the fuel (~1 MJ kg^-1) by vaporising the fuel before delivery to the engine. The proposed solution is to deliver the fuel to the engine in the liquid phase and harness the evaporative charge cooling to increase engine performance. This research provides a robust thermodynamic foundation for this LNG engine concept and resolves significant knowledge gaps in droplet evaporation theory that allows the validity of classical droplet evaporation models to be assessed.

Results from a dedicated thermodynamic model for spark-ignited homogeneous-charge engines, constructed as part of this research, show significantly decreased fuel consumption (-8.9%), increased brake mean effective pressure (+18.5%) and reduced NOx emissions (-51%) when liquid phase LNG is directly injected into the cylinder at near bottom-dead-centre with the intake valve closed, compared to the conventional port-injected gaseous fuel. The mechanisms for this change are a combination of decreased mixture temperatures, reduced heat losses, increased trapped mass, increased compression ratio and advanced spark timing. In some high-load instances, the additional mechanical work obtained can exceed the thermomechanical exergy of the fuel.

Understanding of droplet evaporation phenomena is crucial if a true LNG engine is to be optimised. However, current droplet evaporation models are based on a range of assumptions that can not generally be taken for granted. A new, fundamental, fully transient droplet evaporation model that accounts for the moving droplet surface has been created and implemented to show that significant deviations from the classical d^2-law can arise due to transient effects in the gas phase. As significant outputs of this research, a new method for evaluating appropriate ?film properties is proposed that significantly reduces errors compared to the 1/3 rule and a tool is proposed that quantifies errors from the widely-applied quasi-steady assumption. The methods are generic to all droplet evaporation problems so have utility far beyond the present target application. For liquid methane droplets at engine intake conditions, evaporation rate errors are reduced from 8.5% to 0.3% for the new transport property method compared to the 1/3 rule and it is shown that errors from the quasi-steady assumption are 1-4%.