Detailed exhaust speciation measurements were made on an HCCI engine fueled with iso-octane over a range of fueling rates, and over a range of fuel-stratification levels. Fully premixed fueling was used for the fueling sweep. This sweep extended from a fuel/air equivalence ratio (Φ{phonetic}) of 0.28, which is sufficiently high to achieve a combustion efficiency of 96%, down to a below-idle fueling rate of Φ{phonetic} = 0.08, with a combustion efficiency of only 55%. The stratification sweep was conducted at an idle fueling rate, using an 8-hole GDI injector to vary stratification from well-mixed conditions for an early start of injection (SOI) (40°CA) to highly stratified conditions for an SOI well up the compression stroke (325°CA, 35°bTDCcompression). The engine speed was 1200 rpm. At each operating condition, exhaust samples were collected and analyzed by GC-FID for the C1 and C2 hydrocarbon (HC) species and by GC-MS for all other species except formaldehyde and acetaldehyde. These two species were analyzed using high-performance liquid chromatography. In addition, standard emissions-bench exhaust analysis equipment was used to measure total HC, CO, CO2, O2, and NOX simultaneously with the sampling for the detailed-speciation analysis. Good overall agreement was found between the emissions-bench data and total HC from the detailed measurements. Unreacted fuel, iso-octane, was by far the most prevalent HC species at all operating conditions. Numerous other HC and oxygenated HC (OHC) species were found that could be identified as breakdown products of iso-octane. Several smaller HC and OHC species were also identified. At the highest Φ{phonetic}, emissions of all species were low, except iso-octane. As Φ{phonetic} was reduced, emissions of all species increased, but the rate of increase varied substantially for the different species. Analysis showed that these differences were related to the degree of breakdown from the parent fuel and the in-cylinder location where they formed. SOI-sweep results indicated that stratification improves combustion efficiency by reducing the fuel penetration to the crevice and cylinder-wall boundary-layer regions, as well as by creating a locally richer mixture that burns hotter and more completely.
This work explores the high-load limits of HCCI for naturally aspirated operation. This is done for three fuels with various autoignition reactivity: iso-octane, PRF80, and PRF60. The experiments were conducted in a single-cylinder HCCI research engine (0.98 liter displacement), mostly with a CR = 14 piston installed, but with some tests at CR = 18. Five load-limiting factors were identified: 1) NOx-induced combustion-phasing run-away, 2) wall-heating-induced run-away, 3) EGR-induced oxygen deprivation, 4) wandering unsteady combustion, and 5) excessive exhaust NOx. These experiments at 1200 rpm show that the actual load-limiting factor is dependent on the autoignition reactivity of the fuel, the selected CA50, and in some cases, the tolerable level of NOx emissions. For iso-octane, which has the highest resistance to autoignition of the fuels tested, the NOx emissions become unacceptable at IMEPg = 473 kPa. This happens before wandering and unsteady combustion becomes an issue for IMEPg > 486 kPa. The NOx is caused by high peak-combustion temperatures resulting from the high intake temperature required for this low-reactivity fuel. Iso-octane operation with a CR = 18 piston reduces the intake-temperature requirement. Consequently, the exhaust NOx issue vanishes while the IMEPg can be increased to 520 kPa before wall-heating-induced run-away become an issue. For a very reactive fuel like PRF60, large amounts of EGR are required to control the combustion phasing. Therefore, the maximum IMEPg becomes limited at 643 kPa by the available oxygen as the EGR gases displace air. A fuel of intermediate reactivity, PRF80, exhibits the highest IMEPg for the conditions of this study - 651 kPa. For this fuel, the maximum IMEPg becomes limited by NOx-induced run-away. This happens because even small amounts of NOx recycled via residuals enhance the autoignition sufficiently to advance the ignition point. This leads to higher peak-combustion temperatures and more NOx formation, thus making a very rapid run-away situation inevitable.
The characteristics of autoignition after top-dead-center (TDC) for both single- and two-stage ignition fuels have been investigated in a homogeneous charge compression ignition (HCCI) engine. The single-stage ignition fuel was iso-octane and the two-stage ignition fuel was PRF80 (80% iso-octane and 20% n-heptane). The results show that the heat-release rate and pressure-rise rate both decrease as the combustion is retarded later into the early expansion stroke. This is an advantage for high-load HCCI operation. However, for both fuel-types, cycle-to-cycle variations of the ignition and combustion phasing increase with combustion-phasing retard. Also, the cycle-to-cycle variations are higher for iso-octane compared to PRF80. These observations can be explained by considering the magnitude of random temperature fluctuation and the temperature-rise rate just prior to thermal run-away. Furthermore, too much combustionphasing retard leads to the appearance of partial-burn or misfire cycles, but the responses of the two fuels are quite different. The different behaviors can be explained by considering the thermal and chemical state of the residual exhaust gases that are recycled from one cycle to the next. The data indicate that a partialburn cycle with iso-octane produces residuals that increase the reactivity of the following cycle. However, for the already more reactive PRF80 fuel, the partial-burn products present in the residuals do not increase the reactivity enough to overcome the retarding effect of cool residual gases.
Thermal stratification has the potential to reduce pressure-rise rates and allow increased power output for HCCI engines. This paper systematically examines how the amount of thermal stratification of the core of the charge has to be adjusted to avoid excessive knock as the engine speed and fueling rate are increased. This is accomplished by a combination of multi-zone chemical-kinetics modeling and engine experiments, using iso-octane as the fuel. The experiments show that, for a low-residual engine configuration, the pressure traces are self-similar during changes to the engine speed when CA50 is maintained by adjusting the intake temperature. Consequently, the absolute pressure-rise rate measured as bar/ms increases proportionally with the engine speed. As a result, the knocking (ringing) intensity increases drastically with engine speed, unless counteracted by some means. This paper describes how adjustments of the thermal width of the in-cylinder charge can be used to limit the ringing intensity to 5 MW/m2 as both engine speed and fueling are increased. If the thermal width can be tailored without constraints, this enables smooth operation even for combinations of high speed, high load, and combustion phasing close to TDC. Since large alterations of the thermal width of the charge are not always possible, combustion retard is considered to reduce the requirement on the thermal stratification. The results show that combustion retard carries significant potential since it amplifies the benefit of a fixed thermal width. Therefore, the thermal stratification required for operation with an acceptable knocking intensity can be decreased substantially by the use of combustion retard. This enables combinations of high engine speed and high fueling rate even for operation with the naturally occurring thermal stratification. However, very precise control of the combustion phasing will likely be required for such operation.