There are currently substantial efforts in the field of engine research to find combustion regimes that will allow the compliance of light-duty diesel engines with the continuously decreasing emissions standards. Low Temperature Combustion (LTC) is a method that has been shown to reduce NOx and soot to manageable or compliant levels while maintaining acceptable fuel consumption. The type of combustion employs the use of early injection timing (~ 40o BTDC) in conjunction with large amounts of cooled recirculated exhaust gas to prolong ignition delay, leading to increased charge homogeneity. The intake charge O2 levels required for this type of combustion are typically less than 11% which require re-circulated exhaust values of >60% by mass. Though NOx and soot are substantially reduced, the low temperatures bring about poor conversion of CO and UHC during the combustion process.
Research was carried out in a light-duty high speed diesel engine running in a LTC regime to understand the fundamental behavior of CO and UHC emissions under light to moderate loads. Engine RPM, boost, exhaust backpressure and load were all held constant while injection timing sweeps were performed. The work here has been carried out at 9.5% inlet charge O2 which results in an equivalence ratio that is slightly lean of stoichiometric at this load and boost. The near stoichiometric conditions raise concern regarding the co-location the fuel with available O2. Five parameters were varied in order to resolve the response of the CO and UHC emissions in this operational regime including injector cone angle, injection rail pressure, intake mixture temperature, in cylinder swirl and fuel composition. At the high dilution levels necessary for LTC the resultant CO behavior demonstrated a local minimum around the middle of the swept injection timing range. Lowering the O2 concentration below 11% resulted in more pronounced CO “sweet spot” behavior, this is evident from the higher slopes of CO trends as injection timing was moved away from the location of CO minimum in either direction. After completion of the testing, experimental results and heat release analysis were coupled with CFD modeling to deduce the primary cause of the “sweet spot” behavior.
The CO “sweet spot” was found to be the result of an optimal fuel split between the bowl and squish volumes, altered here by moving the injection timing. UHC trends seem to correlate well to CO trends at timings advanced of the CO minimum location but not as strongly at timings retarded of the CO minimum location. It was found that at early timings a disproportionate amount of fuel was being sprayed into the squish region. This excess fuel results in elevated UHC levels due to wall/piston impingement and increased CO levels due to the delayed fuel preparation process and lack of available O2in the squish region. Timings retarded of the CO minimum location resulted in a disproportionate fraction of fuel in the bowl region as opposed to the squish region; in this case the unused O2 exists in the squish region.
As the included injector angle was narrowed the location of CO minimum was advanced due to the alteration of the fuel split between the bowl and squish regions at a given timing over the angles tested. This parameter did not greatly affect the magnitude of the CO and UHC minimum indicating that an equivalent mixing system could be obtained with a different nozzle included angle at a varied timing. Included injector angle provides movement of the minimum CO and UHC emissions through the direct variation of the fuel split to the bowl/squish region at a given timing. Modification of the injection rail pressure varies the location and magnitude of the CO and UHC minima over the injection timing range. The effects of lowering the injection rail pressure are lower jet velocities and longer injection durations in order to achieve fueling adequate for a congruent load. The lower injection pressures lead to larger droplet sizes and decreased evaporation rate. The combined effects of these two parameters are increased penetration length and impingement with the bowl lip. As the injection rail pressure is increased, the timing needed to achieve a CO minimum (optimal fuel split between the bowl/squish regions) is retarded and the magnitude of the CO and UHC minimum is reduced.
Increased intake charge temperature reduces the overall magnitude of the CO and UHC compared to its lower temperature counterparts with the exception of very early timings where the lower density of the injection environment leads to longer penetration lengths. The higher charge temperatures lead to higher in-cylinder temperatures during the combustion process; these higher in-cylinder temperatures accelerate the oxidation rate of CO and UHC as expected from the Arrhenius rate equation.
In-cylinder swirl is frequently used to enhance the mixing of reactants throughout the combustion process in diesel engines. Initial observations would indicate that it would be beneficial to the mixing of partially burned reactants with available O2 but results show that increased swirl levels lead to elevated CO and UHC emissions. Computational Fluid Dynamics (CFD) revealed that the cause of this increase was the trapping of partially burned reactants in the bottom of the bowl due to the added momentum of the increased swirl. The effect of the swirl in the near stoichiometric environment actually facilitates in-homogeneity of the mixture using a re-entrant bowl.
A region of much interest is the effect of fuel parameters on the emissions trends in the LTC regime, particularly the behavior of the phenomenon using US specification fuel of a lower cetane number as compared to a higher cetane number European fuel. The US specification fuel was found to produce reduced magnitudes of CO minima. More interestingly, the ISFC behavior between the fuels deviates as injection timings are retarded from the location of CO minimum. Cetane rating alone does not constitute a an encompassing fuel study, but rather implies that substantial gains can be achieved by varying fuel parameters in LTC. Currently a mechanism does not exist to resolve the response of varying fuel parameters however one is being explored at the UW Engine Research Center (ERC). This modeling capability in addition to continued experimental work on fuel parameter effects will be used to understand this phenomenon in greater detail.