Environ. Sci. Technol. 2006, 40, 5770-5777
On-Board Generation of a Highly Volatile Starting Fuel to Reduce Automobile Cold-Start Emissions MARCUS D. ASHFORD* AND RONALD D. MATTHEWS Department of Mechanical Engineering, The University of Texas at Austin, Austin, Texas 78712
The on-board distillation system (OBDS) was developed to extract, from gasoline, a high-volatility fuel for exclusive use during the starting and warm-up periods. The use of OBDS distillate fuel results in much improved mixture preparation, allowing combinations of air/fuel ratio and ignition timing that are not possible with gasoline, even with a fully warm engine. The volatility of the distillate is a function of the parent fuel volatility; however, the variability in distillate quality can be diminished via manipulation of the OBDS operating conditions. Thus, it is possible to develop aggressive starting calibrations that are relatively immune to variations in pump gasoline volatility. The key benefits provided by the OBDS fuel relative to standard gasoline were found to be (1) improved mixture preparation allowing a 70% reduction of cranking fuel requirements, elimination of air-fuel mixture enrichment during the warm-up period, and significant extension of warm-up ignition timing retard; (2) a 57% decrease in catalyst light-off time, (3) emissions reductions over the FTP drive cycle of 81% for regulated hydrocarbons (NMOG); (4) emissions index (NMOG) approaching that of SULEV/PZEV vehicles; and (5) an apparent 1% increase in fuel economy over the FTP drive cycle.
Introduction The transportation sector is a significant source of hydrocarbon (HC) pollutants. According to the U.S. Environmental Protection Agency (EPA), mobile sources accounted for 44% of all U.S. hydrocarbon emissions in 2002 (1). Hydrocarbon emissions are of particular concern because (a) many emitted HC species are acutely toxic or carcinogenic; and (b) hydrocarbons are precursors to the formation of photochemical smog. Accordingly, regulatory pressure has mounted on the transportation sector to reduce HC emissions. Emissions of hydrocarbons are highest immediately after starting a cold engine. In modern vehicles, 60-95% of all HC emissions occur during the first 90 s after a cold-start (2-5). The primary reasons for this are twofold: low/unknown fuel volatility and poor catalytic converter efficiency. The relatively low volatility of gasoline (10-30% vaporizes upon injection at 20 °C (2, 4, 5)) requires the injection of considerably more fuel (typically 8-15 times) than the stoichiometric amount in order to generate a reliably ignitable fuel/air mixture. Furthermore, the volatility of the fuel is not known a priori by the engine controller (the powertrain control module, PCM), so the PCM is often calibrated to command fueling * Corresponding author
[email protected]. 5770
9
phone:
(205)
348-4703;
e-mail:
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 18, 2006
rates based upon the expectation of the worst-case volatility fuel. This results in an air/fuel mixture that is overly fuel-rich for most starting conditions and a large amount of liquid fuel that will enter the combustion chamber during the first several cycles. Much of the excess fuel exits the engine unburned or only partially combusted. A conventional threeway catalytic converter will not reach ”light-off” temperature (corresponding to 50% HC conversion efficiency) for 30-40 s or more during the FTP drive cycle (2, 6, 7). Ironically, the period of highest engine-out HC emissions coincides with the period of lowest catalyst efficiency, making the combination of low fuel volatility with poor catalysis the primary cause of high HC emissions during the cold-start and warmup period. Traditional methods to reduce cold-start emissions have focused on increasing catalyst efficiency and/or improving engine system design to allow the use of less fuel-rich mixtures. Varying degrees of success have been realized by enhancing fuel vaporization by avoiding wall wetting (2, 8) or via increased turbulence (2, 3, 8). Special injectors have been used to heat (2, 9) or agitate (10, 11) the fuel. Tighter control of the air/fuel ratio (12) has shown modest emissions improvement. The predominant method to reduce cold-start HC emissions has been the improvement of catalyst efficiency, with some manufacturers resorting to exotic zeolyte adsorbers to store HCs for later release after catalyst conversion efficiency increases (6, 7, 12). Hydrocarbon adsorbers can be found in the most strictly regulated markets (e.g., California SULEV/PZEV certification) (3, 8, 13). Comparatively few attempts have been made to address the root cause of the problem: the inherently low volatility of gasoline. Speciation of exhaust gases during startup revealed that the combustible mixture includes only the more volatile constituents of the fuel (5, 14). The less volatile fraction of the liquid fuel, upon entering the cylinder, is protected from combustion by entering flame quench zones in fuel films on in-cylinder surfaces (2, 6, 8) and inside crevices (2, 15), which are largest in cold engine conditions (2, 11). Much of this unburned fuel will vaporize as cylinder pressure falls during exhaust blowdown and evacuation (15); Stanglmaier et al. predicted that only hydrocarbon components more volatile than hexane are likely to emerge while in-cylinder conditions are conducive to oxidation (16). Thus, the lowvolatility species contribute disproportionately to hydrocarbon emissions during the cold-start and warm-up periods. The on-board distillation system (OBDS) was initially conceived to address the cold-start problems inherent in E85-fueled vehicles. The OBDS separated the most volatile fractions of gasoline from E85 for use as a starting fuel. A vehicle featuring this system was entered into the Department of Energy-sponsored 2000 Ethanol Vehicle Challenge, where the entry received top marks for starting, driveability, and emissions, even besting stock gasoline vehicles in some lowtemperature starting tests (17). The present paper discusses the development of a version of OBDS patented in 2000 specifically designed for gasoline vehicles (18). Complete detail of OBDS operating modes were presented by Kane et al. (17), but salient details will be discussed here. A schematic of the OBDS is shown in Figure 1 in distillate generation mode. Fuel returning from the engine to the main fuel tank is intercepted and routed into the OBDS where the start-up fuel is extracted by distillation and stored in a dedicated tank. The heavier gasoline fraction is routed back to the main fuel tank. The heat source for the distillation is engine coolant. Thus, distillate is best made when the engine is fully warmed. 10.1021/es051950t CCC: $33.50
2006 American Chemical Society Published on Web 08/11/2006
FIGURE 1. Simplified depiction of the OBDS layout in fuel generation mode.
Materials and Methods This OBDS was installed on a 2001 Lincoln Navigator equipped with a 5.4 L 32-valve V-8 engine originally calibrated to meet federal Tier I emissions standards. Close-coupled catalytic converter temperatures were measured both in the inlet and approximately 25 mm into the catalyst brick. Widerange lambda sensors were mounted near the locations of the stock oxygen sensors, immediately downstream of the exhaust manifolds. At the time testing began, the test vehicle had accumulated approximately 8000 miles. Experiments Performed. The first phase of the test program consisted of analyses of the distillate produced by the OBDS and the resultant effect on the composition of the fuel remaining in the main fuel tank. In the second phase of this work, the fuel and ignition timing limits of the distillate were explored in cold-start and warm-up conditions to determine a suitable ignition/fuel strategy. Emissions testing over the FTP drive cycle verified the effectiveness of the most promising calibration. Fuel Characterization. Detailed distillation and compositional analyses were performed of the parent fuels, the distillate fuels generated, and the residual fuels left behind. Residual fuel refers to the bulk fuel remaining in the main fuel tank. Except where noted, the parent fuel used for this development program was EEE certification gasoline, supplied by Haltermann Products. The same certification fuel was used for all baseline testing. Distillation curves, the primary metric by which OBDS distillate quality was judged, were generated per ASTM D86 using a Koehler Instruments K45000 distillation apparatus, and separately by Southwest Research Institute (SwRI) under contract. SwRI also performed detailed compositional analyses (ASTM D5134) and determined research and motor octane numbers (ASTM D2699M/D2700M) of the test fuels. Calibration Development and Validation. To develop PCM calibrations that exploited the properties of the distillate fuel, cold start tests were performed which mimicked the first 20 s of the FTP drive cycle. The goals of these tests were to discover the ability of the distillate fuel to extend the limits of engine combustion under lean fueling and/or retarded ignition timing. “Lean fuel” calibrations explored the leanlimits of cranking and warm-up, with no changes to stock ignition timing. In “Retarded ignition” calibrations, ignition timing changes were added to the lean fuel calibrations to explore the extent to which warm-up spark retard could be applied to bring about faster catalyst light-off when using the volatile OBDS fuel. These tests were performed at the General Motors Foundation Automotive Research Laboratory at The University of Texas.
Each cold start was performed at 20 °C, after an 18-hour period during which the vehicle was maintained at 20 °C, a process known as “soaking” (20). Time-resolved tailpipe hydrocarbon emissions were measured via the UT-developed Fast-Spec (19). The Fast-Spec correlates IR absorption at 3.4 µm to the mass density of total hydrocarbons (THC) in its chamber. Measurements were made at the tailpipe exit, with great care taken to prevent the entry of condensed water into the test chamber. A series of six LA-4 drive cycle tests were conducted to evaluate low-emitting calibration strategies. FTP testing also allowed observation of OBDS characteristics during a controlled drive. Any fuel economy or emissions benefits occurring after the conclusion of OBDS fuel use would be noteworthy. The OBDS fuel and calibrations were in use for approximately the first 20 s of operation after engine start; thereafter, stock calibrations were employed, and the engine was fueled with gasoline from the main fuel tank. To ease the changeover to gasoline operation and to eliminate the chance of fuel pressure loss, the main fuel pump (gasoline) was activated for 2 seconds prior to deactivation of the OBDS fuel pump. Each drive-cycle test was performed at 20 °C after an EPA LA-4 pre-conditioning cycle and 20 °C overnight soak. Engineout and tailpipe emissions of gas-phase species were measured continuously (second-by-second) and also from bag samples. A constant volume sampling (CVS) system was in use for all testing. The calibration with the lowest HC emissions from the LA-4 test series was then tested over the full FTP drive cycle. Emissions were measured again as before, but in addition, speciation was performed to determine oxygenates and specific reactivity.
Results and Discussion Fuel Characterization. Figure 2 shows distillation curves of OBDS distillates along with the parent fuels from which they were derived. Driveability indices for the same fuels are given in Table 1. The driveability index (DI), a commonly used measure of fuel volatility, is calculated as DI ) 1.5T10 + 3T50 + T90. Where TNN is the temperature at which NN% of a fuel sample is recovered in the ASTM D86 fuel volatility test (23). Higher DI values correspond to lower volatility. Two parent fuels were used: federal certification fuel and a standard low-volatility, high-DI hot weather test fuel, both blended by Haltermann Products. This “low-volatility” fuel corresponds to the lowest volatility gasoline auto manufacturers expect to be present in the fuel tank, representing the worstcase starting fuel for which the PCM will be calibrated. Four representative batches of distillate are presented in Figure 2. Two samples were distilled from the certification fuel, and two were distilled from the low-volatility fuel. In distillation process “B”, the OBDS operating conditions were such that a higher volatility distillate was produced at the expense of longer generation time. However, the tailpipe emissions data presented herein were generated with starting fuel produced by process “A”. Two observations are immediately clear: (1) The OBDS distillates are considerably more volatile than the parent fuels, even in the case of the distillates derived from the low-volatility fuels. For example, T10 of the certification parent fuel which corresponds to T5 of the lowvolatility parent fuel occurs at approximately the same temperature as the T40 and T70 points of the least and most volatile distillates, respectively. The vaporization advantages of the distillates are particularly pronounced over the moderate to cold temperature range of importance both for FTP cold-starting and starting in cold weather. (2) Given identical operating conditions, less volatile parent fuels produced less volatile distillates. However, this variation could be mollified in practice through manipulation VOL. 40, NO. 18, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5771
FIGURE 2. Distillation curves of parent and distillate fuels at two OBDS operating regimes. These distillation curves represent two distillation processes: process “B” generates a higher volatility distillate at the expense of longer generation time. The distillates were generated from EEE certification fuel and a low-volatility fuel representative of a worst-case starting fuel.
TABLE 1. Distillation Data for Parent and Distillate Fuels sample
T10 [°C]
T50 [°C]
T90 [°C]
DI [°C]
DI [F]
cert fuel parent low vol parent cert fuel distillate A cert fuel distillate B low vol distillate A low vol distillate B
54 63 33 31 40 34
106 112 53 41 64 51
161 181 111 101 121 103
559 612 320 270 373 305
1182 1277 752 662 847 725
of the OBDS operating conditions. A comparison of the “cert fuel distillate A (CFD A)” to the “low vol distillate B” shows that it is possible to generate a distillate from low-volatility fuel (DI ) 1277) with a distillation curve similar to the startup fuel used in the drive tests. (Recall, “CFD A” is the startup fuel used in the current and previous series of drive-cycle testing.) The implication is that OBDS can reduce or eliminate
the need to calibrate for starting with high-DI fuels. Future research will include cold-start tests with the starting fuel derived from the low-volatility parent. As one would expect, the volatility of the bulk parent fuel decreases as the OBDS process removes light fractions. Figure 3 illustrates this effect, showing distillation curves of the residual fuel from four successive distillation cycles. Also shown for comparison is the distillation curve of the lowvolatility test fuel. In each cycle, 4 liters of starting fuel were distilled from a tank initially holding 100 liters of EEE certification fuel. In the low-temperature regions, the upwardbulging shape of the distillation curves indicates the removal of light-end species from the residual fuel. Moreover, by the third distillation cycle, the residual fuel has lower volatility than the “worst-case” starting fuel in the low-temperature region.
FIGURE 3. Residual fuel distillation curves as the result of successive distillation cycles. 5772
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 18, 2006
TABLE 2. OBDS Fuel Composition, Mass Fractions by Carbon Number C4 C5 C6 C7 C8 C9 C10 C11+ unknown
parent
distillate
residual
0.9 12.1 9.3 25.9 24.6 18.5 3.4 2.0 3.3
3.2 28.4 18.2 24.9 19.0 3.1 0.6 0.3 2.3
0.6 10.7 9.3 30.6 21.2 19.7 3.3 2.1 2.4
The obvious ramification of the depletion of light-ends from the residual fuel is that the main fuel supply might not be sufficiently volatile to support a cold-start should the distillate fuel somehow be depleted. Likely circumstances leading to this scenario would be numerous consecutive short trips that consume starting fuel, but do not warm the engine enough so that new distillate could be generated. This scenario was investigated via computer simulation of distillate and main fuel consumption during drives. It was determined that if a dedicated fuel delivery system is used for the starting fuel (unlike in this installation where fuel lines are shared, requiring flushing with distillate prior to cold-starts) then it is unlikely the cold-start tank would be depleted of fuel before the main fuel tank. Compositional analyses per ASTM 5134 were performed of the parent, distillate and residual fuels. These results are summarized in Table 2, which shows mass-based fuel composition with species grouped by carbon number. Lightends are designated as species C4-C6. These data were taken from a first generation distillation from fresh EEE certification fuel; the distillate and residual fuels correspond to the distillation curves “cert fuel distillate A” in Figure 2 and “residual 1” in Figure 3, respectively. The OBDS process more than doubled the mass concentration in light-ends in the distillate, relative to the parent fuel, from 22.4 to 49.8%. Moreover, there was a decrease in the mass concentration of light-ends in the residual fuel (to 20.6%), confirming our earlier observation. Noteworthy is the presence of a substantial amount of higher hydrocarbons in the distillate. In particular, there is little difference in the C7 concentrations of the parent and distillate fuels. Recalling the prediction by Stanglmaier et al. that an ideal starting fuel should consist of species C6 and lighter (16), a substantial presence of higher hydrocarbons may have a detrimental effect on the coldstarting performance of the distillate fuel. Calibration Development and Validation. As stated earlier, the goal of this research was to quantify an emissions reduction attributable to the use of OBDS fuel for FTP coldstarts. Two calibrations were developed for this research: (1) “lean fuel” with revised fuel mapping, but stock ignition scheduling, and (2) “ignition retard” which added warm-up ignition timing retard to the lean fuel calibration. Expectations were that a stoichiometric or lean air/fuel mixture during cranking would reduce engine-out HCs by eliminating the in-cylinder fuel films resulting from startup enrichment. We also anticipated further reduction of engineout HCs and significant reduction of tailpipe HCs could be accomplished by retarding combustion phasing. The OBDS fuel allowed 15-20 CA of spark retard beyond what was possible with gasoline. Throughout development, considerable effort was expended to maintain consistent operating conditions to isolate the effects of the OBDS fuel. For example, the idle air bypass calibration was manipulated to compensate for decreasing engine speeds caused by leaner air/fuel mixtures and retarded ignition timing. In fact, the extremely retarded ignition timing
necessitated electronic remote actuation of the throttle. Furthermore, calibration changes were limited to areas in which a change was specifically enabled by the change in fuel. For example, a change in idle speed, or transmission shift strategy could result in lower emissions, but since those changes could be made without regard to fuel choice, they were disallowed. Developing “maximum” calibrations for fuel, spark, and airflow for cold-starting is a difficult endeavor. The appropriate combination of fuel injection rate and ignition timing is a function of the quality of mixture preparation, which in turn, is a function of the temperatures of wetted components within the engine. The success of the nth firing cycle is dependent upon the relative successes of the preceding (n ) 1) cycles. The stock engine controller was not equipped to manage the switch from OBDS fuel to gasoline, making it necessary that all calibration modifications from stock expire before the fuel switch occurred. Nevertheless, the calibrations presented here, though relatively immature, demonstrate distinct benefits of the use of OBDS distillate. Figure 4 shows a comparison of test vehicle operating data for OBDS calibrations in relation to the stock configuration over the first 20 s of the LA-4 drive-cycle. Shown are the normalized air/fuel ratio (the excess air ratio, λ), throttle position, and ignition timing deviation from stock. Note the stock ignition timing was used for the “lean-fuel” (LF) calibration; all OBDS calibrations use the same base-fuel calibration as shown below; only for the retarded-ignition calibrations was the throttle opened before the drive began (at 20 s).
{
t)0 30% of stock volume stoichiometric Fuel schedule ) 0 < t < 3 10 < t < 20 λ ) 1.2
}
“Ignition retarded 2” (IR2) reflects calibration and timing improvements over “ignition retard 1” (IR1), which will be explained further in a subsequent section. During the warm-up period, the engine was much more tolerant of lean air/fuel ratio and aggressive spark retard with the OBDS fuel than with gasoline. The engine would not start and run when combining the OBDS calibrations with gasoline, even when fully warmed up ≈90 °C coolant). The extension of tolerance for lean mixtures and retarded ignition timing is most likely due to improved in-cylinder fuel distribution with the OBDS distillate fuel. Figure 5 shows second-by-second engine-out and tailpipe HC emissions over the first 90 s of FTP phase 1, for stock and retarded ignition lean calibrations. Cranking time and subjective idle quality were not degraded from stock. However, no cylinder pressure data were taken to confirm the latter assertion. With the exception of IR1, engine-out HC emissions for the first moments after start are considerably higher for the stock calibration versus the OBDS calibrations, a reflection of the difference in starting enrichment. (Recall that all OBDS calibrations are identical until their ignition timing maps diverge.) For the warm-up period from start until 20 s, engineout HCs are much lower for the LF and IR2 calibrations than stock, a result of lean (OBDS) vs rich (stock) combustion. The HC-reducing effect of aggressive spark retard is seen in a comparison of LF and IR2 calibrations. Engine-out HC’s during the period of spark retard (3-15 s after start) are considerably lower than in the lean fuel calibration, although both calibrations have identical base fuel mapping. The starting tailpipe HC emissions for the OBDS calibrations are higher than stock, despite having lower-than-stock engine-out HC emissions. Furthermore, for the period approximately 18-28 s after start, tailpipe HCs are higher than engine-out. This suggests the exhaust system selectively VOL. 40, NO. 18, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5773
FIGURE 4. Normalized air/fuel ratio (λ), throttle position, and ignition timing deviation from stock for the OBDS retarded ignition calibrations over the first 25 s of the FTP drive cycle, conducted at 20 °C. stores heavier possibly unburned hydrocarbons, a characteristic that has been indicated by other research (21, 22). A plausible explanation is that these heavier hydrocarbons condensed in the relatively cool (soon after starting) exhaust system. After the drive began these hydrocarbons were driven out, most likely by a combination of exhaust gas temperature and velocity. Immediately apparent in the IR1 trace of Figure 4 is an air/fuel ratio excursion in the range of λ > 1.8 accompanying the throttle opening. Testing indicated that lean air/fuel mixtures where λ exceeds 1.3 suffer from high hydrocarbon emissions, most likely due to incomplete or unstable combustion. Accordingly, in Figure 5, there is a sharp peak in HC emissions corresponding to this moment. The lean excursion was caused by improper compensation for the change in fuel film mass that accompanied the throttle opening and subsequent intake manifold pressure transient. That is, even for the highly volatile OBDS fuel, the increased MAP (intake manifold absolute pressure) from opening the throttle generated a liquid fuel film. The shallow slope of the distillate’s distillation curve means that variations in temperature and/or pressure result in relatively large changes in fuel film mass, when compared to gasoline. Over certain temperature ranges the vaporization character of the distillate approaches that of single component fuels. Also of note are the rich air/fuel excursion in Figure 4 and the associated HC 5774
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 18, 2006
emissions spike in Figure 5 that correspond to throttle closing for IR1, just after time ) 14 s. This is further affirmation of improper compensation for fuel film mass transients as a source of HC emissions for the IR1 test. For the IR2 calibration, the IR1 lean fueling strategy was retained. The fueling compensation for manifold pressure transients was revised, however, and throttle opening was delayed by 3 s in order to allow more heat buildup in the intake valves (aiming point for the fuel injectors). Higher valve (and consequently, fuel film) temperatures enable higher intake manifold pressure while still achieving acceptable levels of fuel vaporization. As in the IR1 calibration, the throttle was opened to increase airflow during the aggressive ignition retard period in order to compensate for lost torque and maintain stock-like engine speeds. To maintain a similar level of exhaust energy as in the IR1 calibration, the throttle opening was increased and ignition timing retarded slightly. The relative smoothness of the air/ fuel graph in Figure 4 reflects the calibration improvements. This is also manifested as a significant improvement in both engine-out and tailpipe HCs in Figure 5. Figure 6 shows the effectiveness of the OBDS distillate and associated PCM calibrations over the entirety of phase I of the FTP drive cycle. Shown are accumulated mass emissions for five tests with gasoline and the base calibration and three tests with the distillate fuel and the IR2 calibration.
FIGURE 5. Second-by-second engine-out and tailpipe HC emissions over the first 90 s of FTP phase 1, for stock, lean-fuel, and ignition retard calibrations.
FIGURE 6. Accumulated tailpipe HC emissions over phase I of the FTP drive cycle. Data from the IR1 and lean fuel calibration are shown for comparison. Two of the base tests were run at the beginning of the test program, one in the middle and two at the end. The IR2 tests were completed on consecutive days near the end of the test program. Accumulated mass emissions more accurately show the relative strengths of each calibration, as the concentration data of Figure 5 do not reflect the
differences in exhaust mass flow rate (e.g., combustion air mass flow for the IR2 calibration was three times the stock rate). The combination of OBDS starting fuel and calibrations reduced phase 1 HC emissions considerably, relative to the stock configuration. Exhaust gases from drive cycle tests of the stock and retarded ignition calibrations were speciated to determine VOL. 40, NO. 18, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5775
TABLE 3. FTP Drive Cycle Data Summarya
a
light-off (s)
stock 40.4 ( 1.1
NMOG, measured (g/mi)
0.089
specific reactivity NMOG, reactivity adjusted (g/mi)
4.06 0.105
THC (g/mi)
0.114 ( 0.009
NOx (g/mi)
0.308 ( 0.113
CO (g/mi)
0.687 ( 0.079
exhaust toxics (mg/mi)
2.6
fuel economy [mpg]
13.65 ( 0.25
OBDS 17.25 ( 0.75 -57% 0.025 -72% 2.76 0.020 -81% 0.037 ( 0.001 -68% 0.360 ( 0.001 NSD 0.094 ( 1.1 -86% 1.0 -62% 13.83 NSD
NSD indicates the difference is not statistically significant at the 95% confidence level.
FIGURE 7. Emissions Indices (EI-NMOG) for OBDS Navigator and several low-emissions vehicles, normalized by EI-NMOG of the stock Navigator. Emissions data for this comparison were supplied by the California Air Resources Board (CARB); fuel economy data were supplied by CARB and manufacturer publications. emissions of organic hydrocarbons and exhaust toxics (formaldehyde, acetaldehyde, benzene, and 1,3-butadiene). Notably, exhaust toxics were reduced 62%, and specific reactivity of the exhaust gases was 2.76 compared to 4.06 for the stock vehicle. Reactivity-adjusted NMOG emissions were reduced from 0.11 g/mile to 0.02 g/mile, achieving “apparent” ULEV-II compliance for hydrocarbons (0.04 g/mile). The catalytic converters were not aged to verify ULEV-II HC compliance, but with a 100% margin, it is reasonable to conclude ULEV-II compliance would be achieved with properly aged catalysts. A summary of the FTP results is given in Table 3. Carbon monoxide emissions over the FTP were reduced significantly (86%), not surprising given the reduction in HC emissions achieved with the elimination of over-fueling. There was no statistically significant difference in NOX emissions with the IR2 calibration vs stock. NOX emissions were marginally higher than stock during the lean-fuelled warm-up period, but NOX emissions for this vehicle were dominated by NOX breakthroughs occurring under accelera5776
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 18, 2006
tion during the first 3 min of the FTP drive cycle, rendering the warm-up period NOX increase insignificant. Based purely on the change in air/fuel ratio, an estimated 1-2% fuel economy benefit could be realized if enrichment were eliminated during the starting and warm-up periods. Our measurements over the FTP revealed a 0.98% fuel economy benefit, but too few tests were performed for this to be statistically significant at the 95% confidence level. Nevertheless, the close agreement between the theory and data yields confidence that the fuel economy benefit is real. An insightful perspective on the effectiveness of OBDS at reducing hydrocarbon emissions can be gained by comparing the HC emissions index from the IR2 calibration to those of the stock vehicle and various low emissions vehicles. The emissions index (EI) is a measure of the emissions produced with respect to the fuel consumed in the process, or
EI(species) )
grams of species emitted kg of fuel consumed
The emissions index effectively “normalizes” mass emissions
by engine displacement and vehicle mass, two factors that tend to negatively impact traditional mass/distance emissions measurements. Figure 7 compares EI (NMOG) for the OBDS Navigator (IR2 calibration) with the following vehicles: MY2000 ULEV vehicles Honda Accord and Toyota Camry; MY2003 PZEV vehicles Honda Accord, Nissan Sentra, and Toyota Camry; and the 2003 Toyota Prius (SULEV). The OBDS Navigator has a lower EI (NMOG) than the two ULEV vehicles and is only slightly higher than that of the PZEV Sentra. The implications are that a smaller vehicle equipped with OBDS may be able to meet SULEV/PZEV tailpipe hydrocarbon requirements without having to resort to semi-exotic catalyst and engine control strategies.
(9)
(10)
(11)
(12)
Acknowledgments This research was funded by Ford Motor Company, the U.S. Department of Energy (under grant no. DE-FG04-99AL66262), and the Texas Advanced Technology Program (grant no. 003658-0810-1999). We express sincere gratitude to Haltermann Products and the personnel of Southwest Research Institute, without whose help the data could not have been acquired. We also thank Scott Bohr, Eric Curtis, Wen Dai, Joe Grahor, Dave Raleigh, and Ray Willey of Ford for their invaluable contributions. The opinions and findings presented herein do not necessarily reflect the views of Ford Motor Company, the U.S. DOE or any other Federal or State agencies.
(13)
(14)
(15) (16)
Literature Cited (1) U.S. Environmental Protection Agency, Technology Transfer Network Clearinghouse for Inventories & Emissions Factors;19702002 Average Annual Emissions, All Criteria Pollutants, January 2005; U.S. Government Printing Office: Washington, DC, 2005. (2) Ashford, M. D., An On-Board Distillation System to Reduce ColdStart Hydrocarbon Emissions from Gasoline Internal Combustion Engines, Dissertation, The University of Texas at Austin, 2004. (3) Kidokoro, T.; Hoshi, K.; Hiraku, K.; Satoya, K.; Watanabe, T.; Fujiwara, T.; Suzuki H. Development of PZEV Exhaust Emission Control System, SAE 2003-01-0817; The Society of Automotive Engineers: Warrendale, PA, 2003. (4) Santoso, H.; Cheng W. Mixture Preparation and Hydrocarbon Emissions Behaviors in the First Cycle of SI engine Cranking, SAE 2002-01-2805; The Society of Automotive Engineers: Warrendale, PA, 2002. (5) Chen, K., Fuel Effects on Driveability and Hydrocarbon Emissions of Spark-Ignition Engines During Starting and Warm-Up Processes, Massachusetts Institute of Technology Ph.D. Thesis, Cambridge, MA, 1997. (6) Lassi, U. Deactivation Correlations of Pd/Rh Three-way Catalysts Designed for Euro IV Emission Limits: Effect of Ageing Atmosphere, Temperature and Time, Academic Dissertation, Department of Process and Environmental Engineering, University of Oulu, Finland, 2003. (7) Kishi, N.; Hashimoto, H.; Fujimori, K.; Ishii, K.; Komatsuda, T. Development of the Ultra Low Heat Capacity and Highly Insulating (ULOC) Exhaust Manifold for ULEV, SAE 980937; The Society of Automotive Engineers: Warrendale, PA, 1998. (8) Li, J.; Huang, Y.; Alger, T. F.; Matthews, R. D.; Hall, M. J.; Stanglmaier, R. H.; Roberts, C. E.; Dai, W.; Anderson R. W. Liquid fuel impingement on in-cylinder surfaces as a source of
(17)
(18) (19)
(20) (21)
(22)
(23)
hydrocarbon emissions from direct injection gasoline engines”, J. Eng. Gas Turbines Power, 1999, 123, 3, 659-668. Zimmermann, F.; Ren, W. M.; Bright, J.; Imoehl, B. An Internally Heated Tip Injector to Reduce H. C.; Emissions During Cold Start, SAE 1999-01-0792; The Society of Automotive Engineers, Warrendale, PA, 1999. Hilbert, H. S.; Boggs, D. L.; Schechter, M. M. The Effects of Small Fuel Droplets on Cold Engine Emissions Using Ford’s Air Forced Injection System, SAE 952479; The Society of Automotive Engineers, Warrendale, PA, 1995. Fulcher, S. K.; Gajdeczko, B. F.; Felton, P. G.; Bracco, F. V. The Effects of Fuel Atomization, Vaporization, and Mixing on the Cold-Start UHC Emissions of a Contemporary S.I. Engine with Intake-Manifold Injection, SAE 952482; The Society of Automotive Engineers, Warrendale, PA, 1995. Kishi, N.; Kikuchi, S.; Suzuki, N.; Hayashi, T. Technology for Reducing Exhaust Gas Emissions in Zero Level Emission Vehicles (ZLEV), SAE 1999-01-0772; The Society of Automotive Engineers, Warrendale, PA, 1999. Oguma, H.; Koga, M.; Momoshima, S.; Nishizawa, K.; Yamamoto, S. Development of Third Generation of Gasoline P-ZEV Technology, SAE 2003-01-0816; The Society of Automotive Engineers, Warrendale, PA, 2003. Rapp, L. A.; Benson, J. D.; Burns, V. R., Gorse, R. A., Jr.; Hochhauser, A. M.; Knepper, J. C.; Koehl, Leppard, W. R.; Painter, L. J.; Reuter, R. M.; Rutherford, J. A. Effects of Fuel Properties on Mass Exhaust Emissions During Various Modes of Vehicle Operation, SAE 932726; The Society of Automotive Engineers, Warrendale, PA, 1993. Heywood, J. Internal Combustion Engine Fundamentals; McGraw-Hill: New York, 1998. Stanglmaier, R.; Roberts, C. E., Jr.; Ezekoye, O. A.; Matthews, R. D. (1997) Condensation of Fuel on Combustion Chamber Surfaces as a Mechanism for Increased HC Emissions During Cold Start, SAE Technical Paper 972884; The Society of Automotive Engineers, Warrendale, PA, 1997. Kane, E.; Mehta, D.; Frey, C. Refinement of a Dedicated E85 Silverado with Emphasis on Cold Start and Cold Driveability, SAE 2001-01-0679; The Society of Automotive Engineers, Warrendale, PA, 2001. Matthews, R. D.; Stanglmaier, R. H.; Davis, G. C.; Dai W., U.S. Patent 6,119,637, 2000. Fuss, S. P. Infrared Absorption Spectroscopy: The Absorptance of Low Molecular Weight Paraffin Hydrocarbons and the Development of a Device for Measuring Time Resolved Molecular Species Concentrations, Dissertation, The University of Texas at, 1998. United States Government, Protection of the Environment. Code of Federal Regulations, Title 40, 2003. Kaiser, E. W.; Siegl, W. O.; Baidas, L. M.; Lawson, G. P.; Cramer, C. F.; Dobbins, K. L.; Roth, P. W.; Smokovitz, M. Time-resolved measurement of speciated hydrocarbon emissions during cold start of a spark-ignited engine, SAE 940963; Society of Automotive Engineers: Warrendale, PA, 1994. Nitschke, R. G. Reactivity of SI engine exhaust under steadystate and simulated cold-start operating conditions, SAE 932704; Society of Automotive Engineers: Warrendale, PA, 1993. Standard Test Method for Distillation of Petroleum Products at Atmospheric Pressure, ASTM D86; ASTM International: West Conoshockton, PA, 2006.
Received for review September 30, 2005. Revised manuscript received March 21, 2006. Accepted July 11, 2006. ES051950T
VOL. 40, NO. 18, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5777
