Approaches to improve mixing in compression ignition engines

This thesis presents three approaches to suppress soot emissions in compression-ignition (CI) engines. First, a fuel chemistry approach is proposed. A particular class of fuels - cyclic oxygenates - is identified which is capable of significantly reducing engine-out soot emissions. By means of experiments in "closed" and optical engines, as well as on an industrial burner, two possible mechanisms are identified that could account for the observed reduction in soot: a) an extended ignition delay (ID) and b) a longer flame lift-off length (FLoL). Further analysis of the available data suggests that both mechanisms are related to the inherently low reactivity of the fuel class in question. These findings are largely in line with data found in literature. In the second approach, it is attempted to reduce soot by adopting an alternative combustion concept: early direct injection premixed charge compression ignition (EDI PCCI). In this concept, fuel is injected relatively early in the compression stroke instead of conventional, close to top-dead-center (TDC), injection schemes. While the goal of soot reduction can indeed be achieved via this approach, an important draw-back must be addressed before this concept can be considered practically viable. Due to the fact that combustion chamber temperature and pressure is relatively low early in the compression stroke, fuel impingement against the cylinder liner (wall-wetting) often occurs. Consequently, high levels of unburned hydrocarbons (UHC), oil dilution and poor efficiency are observed. Several strategies, combining a limited engine modification with dedicated air management and fueling settings, are investigated to tackle this drawback. All of these strategies, and especially their combination, resulted in significantly lower UHC emissions and improved fuel economy. Although UHC emissions are typically a tell-tale sign of wall-wetting, as mentioned earlier, the relation between these two has long been hypothetical. Therefore, computational fluid dynamics (CFD) calculations of the injection process are performed to confirm whether or not liquid fuel impingement on the combustion chamber walls is indeed reduced as a result of the aforementioned UHC reduction strategies. Combined model and experimental results indicate that, for most strategies, the measured hydrocarbon emissions and predicted spray impingement are well correlated for a conventional DI injector nozzle, supporting the earlier assumption that wall-wetting is responsible for high hydrocarbon emissions in the investigated early injection timing approach. Lastly, in the third approach, a new injector nozzle design is proposed to improve the mixing process; again with the aim of soot abatement. In conventional nozzles, fuel is injected through 5-10 holes with nominal diameters of 100-200 micrometer. From both literature and in-house experiments it is known that mixing can be improved by reducing the nozzle diameter. Unfortunately, in order to preserve the overall flow rate, the number of required holes quadratically increases with a reduction in hole diameter. Alternatively, it is proposed to not drill the holes, but to use a porous medium instead. The utilized medium is a sintered metal permeated by an interconnected network of (continuous) pores with nominal diameters of 10 micrometer. This material is machined into a nozzle like shape and subjected to atmospheric injection tests as well as to experiments in the Eindhoven High Pressure Cell. Macroscopic experimental data (e.g. shorter ignition delay, larger spray volume) suggests that mixing is indeed improved. However, more research is required, preferably in a(n) (optical) engine, to investigate the impact on (soot) emissions and overall engine performance. In addition, the issues of durability and fouling still have to be addressed