Measurements of In-Use Emissions from Modern Vehicles Using an ...

Environ. Sci. Technol. 2007, 41, 6554-6561

Measurements of In-Use Emissions from Modern Vehicles Using an On-Board Measurement System JOHN F. COLLINS,* PAUL SHEPHERD, THOMAS D. DURBIN, JAMES LENTS, JOSEPH NORBECK, AND MATTHEW BARTH College of Engineering-Center for Environmental Research and Technology, University of California, Riverside, California 92521

Emissions from “low emitting” modern vehicles were measured on-road using a Fourier transform infrared (FTIR) on-board emissions measurement system. Twenty vehicles were tested on road and on a chassis dynamometer. A subset of four vehicles was tested on a test track as well as on the dynamometer. Comparison of on-board measurements with laboratory measurements while operating on the dynamometer showed agreement within measurement and test to test variability. Comparison of dynamometer measurements with test track measurements showed some larger differences attributable to track test conditions. On-road and dynamometer tests were conducted on the remaining 16 vehicles, with the on-road testing including freeway, arterial, and residential streets. The on-road testing showed that most of the low emitting vehicles under most operating conditions are operating below certification levels. Most vehicles reached a hot stabilized condition within 60 to 100 s. Hot running emissions were on average very low once the catalyst lights off. For NMHC, the majority of the “certification” emissions occur during the start-up, especially for PZEVs. NOx and CO also showed a high fraction of “certification” emissions during start-up, but also showed emission spikes under hot running conditions, especially during transients.

Introduction The measurement of emissions using on-board, portable emissions monitoring systems (PEMS) under in-use driving conditions is a key element in estimating vehicle emissions and inventories. The EPA has put considerable emphasis on PEMS in its research programs (1-3), in development of its Mobile Vehicle Emission Modeling System (MOVES) (4, 5) and in the regulatory process (6). Automobile manufacturers have developed and utilized on-board measurements systems in an effort to better understand in-use performance (7-9). Several instrument manufacturers and other private companies have developed on-board systems for commercial use (10-13). Other researchers have used on-board systems for emissions measurements for inventory development, for developing emissions models, and for other applications (1416). The significance of PEMS is that by measuring emissions during real world driving, they eliminate reliance on the artificial driving cycles necessary for use during laboratory dynamometer measurements. * Corresponding author phone: (951) 781-5791, fax: (951) 7815790, e-mail: [email protected]. 6554

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 18, 2007

The College of Engineering Center for Environmental Research and Technology (CE-CERT) at the University of California, Riverside, has conducted a multiyear study of the emission characteristics of a new generation of super-clean gasoline-fueled vehicles and their potential impact on future urban air quality. These new vehicles are 98% to 99% cleaner than catalyst equipped vehicles produced in the mid 1980s. For purposes of this study, this class of vehicles was designated as “Extremely Low Emitting Vehicles” or “ELEVs,” and the associated study was consequently designated as the “Study of Extremely Low Emitting Vehicles” or the “SELEV” study. This program included several elements including the measurement of tailpipe emissions under onroad conditions, development of a modal emissions model to predict emissions from ELEVs, and modeling of emissions inventories and atmospheric concentrations for the potential impact of ELEVs. The results for the modeling and other portions of the study have been reported elsewhere (17-19). The focus of this paper is on the evaluation of an FTIR-based on-board emissions measurement system, emissions from in-use ELEVs over standard cycles, and in-use measurements of ELEVs over typical on-road highway, arterial, and residential road driving, as well as a characterization of their start emissions.

Experimental Section Vehicles and Procedures. Two sets of vehicles were tested for this study. The first set of four vehicles was used extensively for evaluation and validation of the on-board system, as well as some preliminary on-road measurements. A larger set of 16 vehicles were then tested as the main test fleet of the in-use measurements. Both sets of vehicles were tested over dynamometer test cycles and over the road, with additional test track runs also performed on the initial four vehicles. The first four vehicles included one low emitting vehicle (LEV), one ultralow emitting vehicle (ULEV), and two super-ultralow emitting vehicles (SULEVs). The set of 16 vehicles included 13 ULEVs and 3 Partial Zero Emitting Vehicles (PZEVs). Table 1 identifies the first four vehicles by make and model year but lists the remaining vehicles anonymously. The order of this list does not correspond to the vehicle identification numbering used in the figures. The test protocol included the following components: Dynamometer Test Procedures. The vehicles were tested in CE-CERT’s Vehicle Emissions Research Laboratory (VERL) or in California Air Resources Board’s (CARB) Haagen-Smit Laboratory in El Monte, CA. The initial four vehicles were also tested on a test track at the Honda Proving Grounds. Proving ground tests were on a test track not a dynamometer, but the driver used a driver’s aid to follow standard driving cycles. The dynamometer test cycles for all vehicles included a Federal Test Procedure (FTP) (6), a US06 (6), and the NCHRP99a, which is a cycle that has been used for the development of the CE-CERT’s Comprehensive Modal Emissions Model (CMEM) (20). On the first four vehicles, additional dynamometer tests were also performed over the CARB Unified Cycle (21, 22). On-Road Test Procedures. Each vehicle was tested over an on-road course that includes approximately 15.9 miles of freeway driving, 8 miles of arterial driving, and 2.2 miles of residential driving. On a given day, the on-road course was driven beginning at 7:00 AM, at 11:00 AM, and 5:00 PM. The daily test was repeated on 3 days. The same route was driven every time, but the traffic varied from congested to free flowing and the driver stayed with normal traffic flow. The vehicle carried one driver, and the measurement system. 10.1021/es0627850 CCC: $37.00

 2007 American Chemical Society Published on Web 08/17/2007

TABLE 1. Test Vehicles vehicle ID

year

make

model

odometer

LEV01 ULEV01 SULEV01 SULEV02 ULEV ULEV ULEV ULEV ULEV ULEV ULEV ULEV ULEV ULEV ULEV ULEV ULEV PZEV PZEV PZEV

2001 2001 2000 2001 2002 2002 2001 2002 2002 2003 2001 2002 2002 2002 2002 2002 2003 2003 2003 2003

Chevrolet Honda Honda Nissan Acura Buick Ford Ford Honda Honda Mazda Mitsubishi Mitsubishi Nissan Saturn Toyota Toyota Honda Honda Toyota

Malibu Accord LX Accord EX-L Sentra CA 3.2TL Regal Focus Mustang Civic Civic Hybrid Prote´ ge´ Galant Lancer Altima L200 Camry LE Corolla Accord EX Civic Hybrid Camry LE

11324 300 7000 3863 32344 21184 35089 23894 26632 13700 27114 22350 13300 13747 14888 13098 21835 7731 1502 2600

The resulting vehicle weight exceeded the certification equivalent test weight (ETW) by 200 to 400 pounds. The vehicles were soaked indoors. Temperatures were maintained to within cold-soak temperature limits specified by the Code of Federal Regulations for the FTP (6). Soak time ranged from 12 to 36 h for the morning test, was about 3 h for the midday test, and was about 5 h for the afternoon test. Test Track Procedures. A series of tests were conducted on a test track over standard cycles following a driver’s aid at the Honda Proving Center California. The track is a 7-mile oval with simulated new and aged freeway sections, grades of less than one-half percent, and a minimum turning radius of approximately one mile. For cold-start testing, the vehicles were soaked overnight, outdoors at the test track. Overnight

soak temperatures outdoors were uncontrolled and ranged from near freezing to the mid sixties depending on day and season. The weight of the driver, passenger, and equipment resulted in a vehicle weight that exceeded the standard inertial test weights by 200-400 pounds depending on the vehicle. The combination of extra weight, uncontrolled soak temperatures, slight grades, and light winds increased both emissions and test to test variability in emissions compared to dynamometer testing. Emission Measurement System. The on-board measurements were made with a Fourier transform infrared (FTIR) spectrometer operating at a 0.5 wave number resolution over an 8.3 m path length in a heated gas cell. The FTIR samples the raw exhaust through a heated sample extraction and drying system that controls temperature, pressure, flow rate, and moisture content to provide a stable FTIR background. Classical least-squares (CLS) is used to quantify the FTIR spectral measurements. Non-methane hydrocarbons (NMHC) are a mixture of many different molecules without uniquely defined absorption lines in the infrared. NMHC was calibrated using continuum regions of the spectrum against a blended mixture of hydrocarbons extending from C2 through C6 and was verified by comparison of FTIR measurements with FID measurements of dried vehicle exhaust. Exhaust flow rate was calculated using OBDII system measurements and calibrated by comparison of on-board CO2 mass emissions with laboratory bench CO2 mass emissions. For more detail on the measurement system, see Truex et al. (14). Fuel. Testing for the first four vehicles used CA Phase 2 fuel available at retail stations. For the second round of testing, fuel was obtained exclusively from Union 76 (ConocoPhillips). ConocoPhillips was an early distributor of gasoline with ethanol, and at the time of testing, this fuel was expected to contain ethanol oxygenate rather than MTBE oxygenate, but was not expected to yet meet the California Phase 3 gasoline caps on sulfur.

FIGURE 1. CO measurements by on-board instrumentation and laboratory instrumentation at the ARB lab. VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6555

FIGURE 2. NOx measurements by on-board instrumentation and laboratory instrumentation at the ARB lab.

FIGURE 3. NMHC measurements by on-board instrumentation in lab vs at test track.

Results Comparisons of On-Board, Dynamometer, and Test Track Measurements. Figures 1-3 compare lab bench results on 6556

9


the dynamometer, on-board results on the dynamometer, and on-board results at the test track. Each bar in these charts represents the average of three replicate tests, with the error

FIGURE 4. FTP weighted emissions during dynamometer testing.

FIGURE 5. US06 emissions during dynamometer testing. bars representing one standard deviation of the measurements. When comparing Dyno_On-board and Dyno-Bench results, both measurements are concurrent measurements of the same tailpipe emissions with different equipment. When comparing Dyno_On-board and Track_On-board results, both data sets use the on-board measurements but are from different tests. Therefore those differences include test to test variability and track condition variability.

For most combinations of species, cycle, and phase, there is a reasonable agreement between the dynamometer bench, the dynamometer on-board, and the test track measurements. Where laboratory discrepancies appear large on a relative basis, it should be noted that these emissions are generally below SULEV levels and the absolute emission differences are small. The emissions for some of the test sequences showed larger variability than others. Examination VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6557

FIGURE 6. Start NMHC emissions, grams in 300 s.

FIGURE 7. On-road NMHC emissions, hot running g/mi. of modal emission traces shows that emissions from low emitting vehicles are sensitive to details of the driving trace, particularly during aggressive driving. A single driver on a dynamometer, performing within the tolerances specified in the Code of Federal Regulations (CFR), can generate emissions different by a factor of 2 or more from test to test, depending on species, cycle, and phase. This is evident in the relatively large error bars for CO emissions over the Unified Cycle, and the NOx emissions for the ULEV01. The use of a different driver at the track, slight grades, and varying ambient conditions combined will exaggerate the test to test variability found with a single driver under laboratory conditions. This sensitivity of total emissions to 6558

9


a few transient events probably explains the reason for most differences between track and laboratory for CO and NOx . For NMHC there is an additional factor. Cold start NMHC emissions over the Unified cycle were substantially higher for LEV01 at the track than at the laboratory and several times higher at the track for SULEV01. Those tests were conducted late fall/early winter in the morning at near freezing ambient temperatures. The cold temperatures delay catalyst light off and thus cause high cold start NMHC emission during those tests. Dynamometer Results for the Main Test Fleet. Figure 4 shows the results of FTP testing in weighted g/mile. The black bars show the LEV I ULEV and SULEV emission

FIGURE 8. Cold Start NOx emissions, grams in 300 s.

FIGURE 9. On-road NOx emissions, hot running g/mi. standards. The figure shows that there is considerable variability from vehicle to vehicle. The figure also shows that these vehicles are basically still meeting their certification standard, with a couple of borderline exceptions. The figure also shows that the PZEV vehicles have substantially lower emissions of NMHC and NOx than ULEV vehicles, as expected. Figure 5 shows the emissions measured during testing over the US06 drive cycle. The US06 cycle contains aggressive freeway driving. This cycle is one component of the Supplemental Federal Test Procedure (SFTP). Not all of the SELEV vehicles were certified to meet the SFTP standard. This figure shows the great variability that exists from vehicle to vehicle, even within a vehicle emissions class. Our prior testing studies also indicated that there is high variability from test to test

over the US06 as well for some vehicles. The variability from vehicle to vehicle and test to test shows the difficulty in attempting to extrapolate fleet emissions from certification testing for these vehicle certification classes. On-Road Results for the Main Test Fleet. The modal emission traces for the on-road tests showed that emission rates are high during the start, then fall off to very low values once the catalyst lights off. Thus, it makes sense to separate the start and hot running emissions. The start emissions are calculated as emissions that occur with the first 100 s after ignition, while the hot running emission rates were calculated starting from 300 s after ignition and continuing through the end of the test. The start data shown in the figures include the engine starts during the morning, noon, and afternoon VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6559

FIGURE 10. Start CO emissions, grams in 300 s.

FIGURE 11. On-road CO emissions, hot running g/mi. tests. Despite the corresponding soak times of 12, 3, and 5 h, the start emissions were observed to be roughly the same for each test. The hot running emission rates are averages over the combined freeway, arterial, and residential sections. The data below are presented in terms of g/mile for comparison with vehicle certification standards. We also measured CO2 and looked at emission rates in terms of g/kg fuel. The patterns shown in the figures below remain almost the same whether expressed as mileage-based or fuel-based emission rates. While there is some variability in fuel economy among vehicles, the main reason for the difference in emissions is due to the difference in how frequently the vehicles go slightly out of near perfect emission control. 6560

9


The NMHC start and hot running emissions are shown in Figures 6 and 7. NMHC emissions were generally dominated by the start emissions. For most vehicles, the start emissions represented the majority of the emissions on the total 26 mile trip. For the PZEV vehicles, almost all the emissions occur during the start. For ULEV vehicles, the majority of the hydrocarbon emissions do occur during the start, but there are also periods of time near the beginning and end of the tests corresponding to freeway driving that show significant hydrocarbon emissions. Start and hot running NOx emissions are shown in Figures 8 and 9. NOx emissions also showed higher emissions during the cold start, but the ratio of start to hot running emissions was less than for the NMHC emissions. After the

start, the NOx emissions showed relatively long periods of low NOx emission being frequently interspersed with short spikes of high NOx emissions. The higher spikes are periods associated with freeway and aggressive driving. The trends for the running NOx emission rates for these vehicles were a function of vehicle specific power or traffic speed/ congestion categories and could be predictable using one of these approaches. The start and hot running CO emissions are shown in Figures 10 and 11. Overall, the variability for the start and hot running CO emissions was greater than that seen for NMHC or NOx emissions. CO emissions appear to be more sensitive to power demand during start conditions as well as under hot running conditions. Even PZEV vehicles can produce large CO emission spikes. These spikes occur during transients under very high power demand. CO concentrations during these events can reach several percent CO by volume. These events appear more common in PZEVs than ULEVs, and in ULEVs than in LEVs. In general, the more finely tuned the emission control system, the more susceptible it is to gross disruption. Despite CO emission spikes, the running emissions are generally well below the certification levels, which for SULEV is 1 g/mi.

(9)

(10)

(11)

(12)

(13)

(14)

Acknowledgments This work has been sponsored in part by Honda Motor Company, the US EPA, the California ARB, Chevron-Texaco, and the Manufacturers of Emission Controls Association (MECA). The contents of this paper reflect the views of the authors and do not necessarily indicate acceptance by the sponsors.

(15)

(16)

Literature Cited (1) Baldauf, R.; Fulper, C.; Nam, E.; Scarbro, C.; Kishan, S.; Sabisch, M.; Hoffman, G. Preliminary Analysis of Emission Measurements from Light-Duty Gasoline Vehicles in the Kansas City Area on Gravimetric PM2.5, Criteria Pollutants, and From “Real-World” Portable Emission Measurement Systems. Presented at the 16th CRC On-Road Vehicle Emissions Workshop, San Diego, CA, April 2006. (2) Berton, L. A. G. Real-Time On-Road Vehicle Exhaust Moluar Flowmeter and Emissions Reporting System. US Patent 6148656, 2000. (3) Spears, M. W. On-Vehicle Emissions Measurements of ThirtyTwo Light-Duty Vehicles; Proceedings of the 13th CRC OnRoad Vehicle Emissions Workshop, San Diego, CA, March 2003. (4) Nam, E. K.; Mitcham, A.; Ensfield, C. Dynamometer and OnBoard Emissions Testing of the Honda Insight and Toyota Prius; SAE Technical Paper No. 2005-01-0681; Society of Automotive Engineers: Warrendale, PA, 2005. (5) Rykowski, R. A.; Nam, E. K.; Hoffman, G. On-Road Testing and Characterization of Fuel Economy of Light-Duty Vehicles; SAE 2005-01-0677; Society of Automotive Engineers: Warrendale, PA, 2005. (6) Code of Federal Regulations 40 CFR Parts 9 and 86. Control of Emissions of Air Pollution from New Motor Vehicles: In-Use Testing for Heavy-Duty Engines and Vehicles; Final Rule. [OAR2004-0072; AMS-FRL-7922-4]. Fed. Regist. 2005, 70 (113). (7) Jetter, J.; Maeshiro, S.; Hatcho, S.; Klebba, R. Development of an On-Board Analyzer for Use on Advanced Low Emissions Vehicles; SAE Technical Paper No. 2000-01-1140; Society of Automotive Engineers: Warrendale, PA, 2000. (8) Butler, J. W.; Kornishi, T. J.; Reading, A. R.; Kotenko, T. L. Dynamometer Quality Data On-Board Vehicles for Real-World Emissions Measurements. Proceedings of the 9th CRC On-Road

(17)

(18)

(19)

(20)

(21)

(22)

Vehicle Emissions Workshop; San Diego, CA, April 1999; CRC: Atpharetta, GA, 1999. Gierczak, C.; Korniski, T.; Wallington, T. J.; Butler, J. W. Laboratory Evaluation of the SEMTECH-G Portable Emissions Measurement System (PEMS) for Gasoline Vehicles; SAE Technical Paper No. 2006-01-1081; Society of Automotive Engineers: Warrendale, PA, 2006. Kreso, A.; Ensfield, C. Performance of In-Use Equipment During Transcontinental Excursion and Use of the Real-World Data to Calibrate Vehicle to Ensure Compliance with Emissions Standards. Proceedings of the 13th CRC On-Road Vehicle Emissions Workshop; San Diego, CA, March 2003. Akard, M.; Nakamura, H.; Aoki, S.; Kihara, N.; Adachi, M. Performance Results and Design Consideration for a New InUse Testing Instrument; SAE Technical Paper No. 2005-01-3606; Society of Automotive Engineers: Warrendale, PA, 2005. Weaver, C. S.; Balam-Almanza, M. V. Development of the ‘RAVEM’ Ride-Along Vehicle Emission Measurement System for Gaseous and Particulate Emissions; SAE Technical Paper No. 2001-013644; Society of Automotive Engineers: Warrendale, PA, 2001. Vohtisek,-Lom, M.; Allsop, J. E. Development of Heavy-Duty Diesel Portable On-Board Mass Exhaust Emissions Monitoring System with NOx, CO2, and Qualitative PM Capabilities; SAE Technical Paper No. 2001-01-3641; Society of Automotive Engineers: Warrendale, PA. 2001. Truex, T. J.; Collins, J. F.; Jetter, J. J.; Knight, B.; Hayashi, T.; Kishi, N.; Suzuki, N. Measurement of Ambient Roadway and Vehicle Exhaust Emissions - An assessment of Instrument Capability and Initial On-Road Test Results with an Advanced Low Emission Vehicle; SAE Technical Paper No. 2000-01-1142; Society of Automotive Engineers: Warrendale, PA, 2000. Frey, C. H.; Unal, A.; Rouphail, M. N.; Colyar, J. D. On-Road Measurement of Vehicle Tailpipe Emissions Using a Portable Instrument. J. Air Waste Manage. Assoc. 2003, Vol. 53, p. 9921002. Gautam, M.; Riddle, W. C.; Thompson, G. J.; Carder, D. K.; Clark, N. N.; Lyons, D. W. Measurements of Brake-specific NOx Emissions using Zirconia Sensors for In-Use, On-Board HeavyDuty Vehicle Applications; SAE Technical Paper No. 2002-011755; Society of Automotive Engineers: Warrendale, PA, 2002. Barth, M.; Collins, J.; Scora, G.; Davis, N.; Norbeck, J. Measuring and Modeling Emissions from Extremely Low Emitting Vehicles; TRB Paper No. 06-1526, Transportation Research Board: Washington, DC, 2006. Collins, J.; Davis, N.; Lents, J.; Malcolm, C.; Norbeck, J.; Omary, M.; Tonnesen, G.; Younglove, T. Report on the Study of Extremely Low Emission Vehicles: Year 3; Final Report by the University of California, CE-CERT to program sponsors, University of California: 2004. Davis, N. In-Use PZEV Emission Rates and the Potential Impact of these Vehicle on Future Air Quality. Proceedings of the 13th CRC On-Road Vehicle Emissions Workshop; San Diego, CA, March 2004; CRC: Atpharetta, GA, 2004. Barth, M.; An, F.; Younglove, T.; Levine, C.; Scora, G.; Ross, M.; Wenzel, T. The Development of a Comprehensive Modal Emissions Model; Final report submitted to the National Cooperative Highway Research Program, November 1999, 255 pp. 1999. California Air Resources Board (1999). California Exhaust Emission Standards And Test Procedures for 2001 and Subsequent Model Passenger Cars, Light-Duty Trucks, And Medium-Duty Vehicles; Rulemaking adopted August 5, 1999; amended July 30, 2002. California Air Resources Board: Sacramento, CA. California Air Resources Board (undated). The Federal Test Procedure & Unified Cycle; California Air Resources Board: Sacramento, CA; Emissions Inventory Series Volume 1, Issue 9.

Received for review November 22, 2006. Revised manuscript received May 31, 2007. Accepted June 14, 2007. ES0627850

VOL. 41, NO. 18, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6561