Vehicle mass emissions measurements using a ...

Final version - Emission paper - Wednesday 10/1/97 - Page 1

Vehicle mass emissions measurements using a portable 5-gas exhaust analyzer and engine computer data

Michal Vojtisek-Lom and James T. Cobb, Jr. Dept. of Chemical and Petroleum Engineering, 1137 Benedum Hall, University of Pittsburgh, Pittsburgh, PA 15261; tel (412) 624-1346, fax (412) 624-9639, e-mail: [email protected] and [email protected]

ABSTRACT The University of Pittsburgh commuter vanpool operates 20 dedicated compressed natural gas vans. Data on fuel use, maintenance, oil contaminants, and tailpipe emissions on these vans are collected. A method is being developed to determine “real-world” mass emissions using an on-board low-cost fivegas non-dispersive infra-red exhaust gas analyzer, an engine diagnostic scanner and a laptop computer. The computer uses live engine diagnostic port data to compute molar exhaust flow, which, when multiplied by molecular weight and measured concentrations of exhaust gases, yields grams per second data. Grams per mile emissions are calculated from known distance and time of the trip. The $10,000 system, also capable of measuring fuel consumption, is portable, requires no modifications to the vehicle, and uses commercially available instruments. Although the system yields consistent data on an 8-mile test route, values measured during I/M 240 tests show correlations of 0.67 for NOx, 0.50 for HC and 0.68 for CO between the portable and I/M 240 instruments and high test-to-test differences. Others have found good correlation between I/M 240 and portable analyzer mass emission. The poor correlation in this study may be attributed to extremely low emissions (tenths of a percent of the instrument range) of the tested vehicles. Improving the accuracy is the focus of ongoing research.

INTRODUCTION The University of Pittsburgh commuter vanpool operates 20 dedicated compressed natural gas (CNG) 15-passenger 1996 Dodge Ram vans. Data on fuel use, maintenance, oil contaminants, and exhaust emissions on these vans are collected for a comprehensive technical evaluation of CNG vehicles. Another goal of the project is to determine the “real-world” mass exhaust emissions for a study of the impact of the vans on the air basin. The conventional method for measuring mass emissions is to drive each vehicle to a test facility, and measure emissions as the vehicle is being driven on a dynamometer following a set driving cycle. Two most common cycles are the Federal Test Protocol (FTP-75 or FTP), used for most accurate and detailed measurements, and its simplified four-minute version, I/M 240, used primarily in inspection/maintenance programs. The cost of FTP (about $1000 per test) was prohibitive for our project. I/M 240 tests are available for $20-100, but require 230 mile round-trip travel to the nearest available facility. How well the driving cycles represent the “real-world” driving is also an issue in this study. The vans, carrying 10-15 passengers each, are operated by volunteer drivers in hilly terrain and outside temperatures ranging from 0 to 90°F. Heavy loads, hills, and observed aggressive driving patterns of some drivers (wide-open


throttle accelerations) all lead, on gasoline vehicles, to higher emission levels than measured by FTP. 1 2 3 This is partly due to higher fuel consumption, partly because gasoline engines operate fuel-rich at wide-open throttle (WOT) and partly because of higher combustion chamber temperatures under such conditions. The I/M 240 cycle produces even lower emission readings than FTP, because it does not include starting and driving with a cold engine, which account for over half of the emissions on a modern gasoline-powered vehicle. Natural gas engines can operate at stochiometric air/fuel ratio both when cold (fuel does not condense) and at WOT,4 and therefore produce notably lower emissions during these conditions, as has been observed during this project. Still, previous tests on CNG vehicles show that neither FTP nor “real-world” emissions of CNG vehicles can be reliably determined on an I/M 240 cycle.5 As an alternative to dynamometer testing, the option of using an on-board system on a typical daily route has been examined in this project. Most of the existing on-board systems, such as ones developed by Sierra Research for California Air Research Board6, are mounted on dedicated research vehicles. The University’s goal, to test vehicles in daily use on their usual routes, mandated a portable system that displaces no more than one passenger, and requires no or minimal modifications to the vehicle. A portable, inexpensive system was previously used at the University of Denver7, where mass emissions of hydrocarbons (HC) and carbon monoxide (CO) were calculated from concentrations of exhaust gases measured by a NDIR four-gas exhaust analyzer on a trip of known length, and estimated fuel consumption. This method successfully predicted I/M 240 results, but has two disadvantages. First, the error in estimated fuel consumption results in the same relative error in mass emissions readings. Second, the method does not account for a varying exhaust flow, which makes it less accurate in the case where a strong correlation between concentrations of nitrogen oxides (NOx) and exhaust mass flow is observed. Both deficiencies could be eliminated by measuring exhaust mass flow. The scope of this study is the development of an inexpensive, portable on-board system that uses exhaust gas concentrations measured by a five-gas analyzer, calculated exhaust flow, and distance data to calculate grams/mile emissions on an arbitrary route. The system does not require any modifications to the vehicle, and consists of commercially available components of total cost under $10,000. This system has the advantage over the one developed for CARB by Sierra Research in that it is portable, while the CARB system uses permanently mounted instruments.3

EXPERIMENTAL METHOD Instrumentation Measuring Exhaust Mass Flow Exhaust mass flow can be measured either directly, by using a mass flow meter, or indirectly, by computations. In selecting a method for the University’s portable system, several were considered: 1. Mounting a mass flow meter on the end of the tailpipe, and either interfacing it with a portable computer or coupling it with a data logging device. A suitable system is made by Sierra Instruments, Harrison Township, MI8. This option has been eliminated because of the costs ($2000-3000 for meter plus $1000-2000 for data logger or RS-232 data acquisition interface).


2. Installing a fuel flow meter, also with computer interface or data logging device. The exhaust flow would be then computed indirectly from the known fuel flow and air-fuel ratio, estimated based on emission data. This option was eliminated because it would require modification to the vehicle. 3. Using the data from engine computer to calculate the intake air mass flow. Live engine data can be obtained on newer vehicles by using an engine diagnostic scanner, a common instrument owned by many repair shops. Exhaust gas mass flow is then calculated using intake air flow, exhaust concentration data, and known composition of the fuel. The last method has been used, because engine diagnostic scanners • are readily available on the market, already owned by many repair shops, and collected data can be used for studying traffic patterns or driving behavior, and other purposes • are relatively inexpensive (under $2000 for scanner with RS-232 interface) • have a lower potential for being a traffic hazard than a meter with a (second) cable mounted on the end of the tailpipe Data Acquisition Instrument Setup The overall system setup is shown of Figure 1. Concentrations of hydrocarbons (HC), carbon monoxide (CO), nitrogen oxides (NOx), carbon dioxide (CO2) and oxygen (O2) in the exhaust gases are measured by a five-gas RG240 Digital Gas Analyzer made by OTC SPX (655 Eisenhower Drive, Owantonna, MN 550601171, tel. (507) 455-7000). The instrument uses a non-dispersive infra-red (NDIR) bench by Andros to measure HC, CO and CO2. Because of the wavelength at which hydrocarbons are measured, the instrument does not detect methane9, which accounts for about 90% of the hydrocarbons emitted by a CNG engine10. HC thus represents non-methane hydrocarbons. The analyzer sits on the floor of the vehicle. The sampling hose runs through a side window straight up to the gutter, is clamped to the gutter (hose sits snug in the clamp but is not deformed by it) every 3-5 feet, and goes straight down to the tailpipe, to which the probe is secured with a radiator hose clamp. Water from the analyzer is drained to a small plastic bottle located in door well or other convenient location below the analyzer. Exhaust gas outlet hose can be routed outside, but was also routed to the drain container, and vehicle was driven with open windows or forced ventilation. (The flow of gases through analyzer is relatively small.) Since the cigarette lighter outlet loses power during engine starting on some vehicles, the power is supplied to the analyzer by a cable directly clamped onto the battery. The cable has an inline fuse about 3” from the battery clamp and is routed in between the hood and the front mask, wrapped around the antenna base when applicable, then routed in a similar fashion as the sampling hose. The concentration data are continuously downloaded every 0.72 s into a laptop computer via an RS-232 serial port; this is a standard feature of the analyzer. A special memory-resident program, OTC_R, was written to capture the data and save them into a file. (Unless otherwise noted, all programs were written by the first author in Pascal.) Engine control unit live data are monitored with a Snap-On MT-2500 engine diagnostic scanner connected to a diagnostic link, located in the engine compartment or below the dashboard. (Only newer, usually 90’s vehicles have this option.) The scanner provides numerous engine parameters, out of which intake manifold absolute pressure (MAP), engine speed (RPM), intake air temperature (IAT), vehicle speed (MPH), coolant temperature (CLT) and intake mass air flow (MAF - available only on some vehicles) are of particular importance. The data are continuously downloaded in binary form via a second RS-232 serial port into the laptop computer. A slightly modified version of ScanGrafix software (optionally supplied with the scanner) was obtained from Engine Control and Monitoring (586 Weddell Dr. #2, Sunnyvale, CA 04089, tel. (408) 7343433), to capture and save 4 sets of data per second.


The only available laptop with two serial ports was a used 386-based Toshiba T3100SX. The laptop, recharged from the cigarette lighter adapter, records simultaneously both sets of data, adding system time to each set. Because neither a suitable multitasking operating system nor a source code to ScanGrafix was available, OTC_R is based on interrupt-driven routines. At the beginning of the test, OTC_R loads, reserves a portion of the memory for data, redirects serial port and keyboard routines and executes a DOS shell in which the ScanGrafix runs. During this time, OTC_R allows the user to turn on or off recording of data and add a onecharacter tag to the records. This tag is set on-the-fly by the user by pressing a single key (for example B bridge, F - freeway) and is used in later analysis of data. Upon exiting the shell, OTC_R saves the data into a file and restores the interrupts. Presently, data can be captured for up to only one hour. Testing process Before the test, the instruments are connected as described above, and the analyzer is left on sampling ambient air for 30-40 minutes. Then, the analyzer is manually zeroed, and the data link to the engine scanner activated; at this point, the ignition switch must be in the position “ON”. When both sets of data are ready to be received by the laptop, recording is initiated. When starting the engine is part of the test cycle, the engine is started at this point; otherwise, it can be started at any convenient time before, so that it does (or does not, as desired) reach the operating temperature. During the test, the system is left to operate automatically, or a designated person can tag the records, turn the recording on or off, or end one and initiate another set of data. After the test, data is saved into a file, and all instruments are disconnected. Mass emissions calculations Synchronization of data Data recorded during the test are later processed in the project’s office, using a second specially-written program, PROCESS. First, the two sets of data have to be synchronized. Exhaust gases are detected by the analyzer with a delay equal to the time it takes for an average gas molecule to travel from the engine through the exhaust system, probe and hose into the detection chamber. The delay time was measured experimentally by timed inserting and removing of the probe into and out of the exhaust stream and measuring the analyzer response, and by observing analyzer response to changes in engine operating conditions. The delay varies because the exhaust mass flow, and the time it takes the gases to flow through the exhaust system, varies with engine speed and load. An average constant delay of eight seconds is used in our calculations; this delay is subtracted from the time stamp of each record of the concentration data. The data also come at different rates - the approximate interval for engine data is 0.25 s, for gas analyzer data 0.72 s. Also, on longer runs, every 30 minutes, concentration data is not available for the two to three minutes necessary for periodic zeroing of the analyzer. Occasionally, these time intervals vary when lower transmission rate is mandated by the computer speed, or when some records are missing or unusable. A special routine has been written that uses linear interpolation to produce one set of combined concentration and engine data each second. When a large gap in the concentration data exists, zero data are substituted, and that data is marked with a “missing” tag. Intake air flow calculations On engines using a mass air flow meter, mass air flow data in grams per second are available; this reading must be divided by the molecular weight of air, 28.9 g/mol. On engines using the speed-density method,


the intake air flow is calculated using two approximations: the ideal gas law, and an assumption of constant engine volumetric efficiency, which means that the intake manifold absolute pressure is directly proportional to the flow of the air through the engine. The intake air molar flow IMF, in moles per second, is then calculated using the following formula (units listed in brackets):

Manifold absolute Engine Engine pressure [kPa] * displacement [liters] * speed [RPM] / 120 IMF [mol/s] = ------------------------------------------------------------------------------- * VEF 8.314 * (Intake air temperature [deg C] + 273.15)

(1)

where VEF is dimensionless engine volumetric efficiency. If the Manifold absolute pressure is not available, use MAP = barometric pressure - manifold vacuum. In reality, VEF is a complex function, different for each engine, of MAP and RPM.11 A comparison of calculated versus scanner-reported mass air flow on 1996 4.2l V-6 Chevy Astro van shows most of the VEF values obtained on a 40-mile trip fall between 0.70 and 0.93, dependent mainly on MAP. For this project, a constant VEF=1.0 was assumed. This value will be used until a more representative value is be found. Since dependence of the final result on constant VEF is linear, mass emissions can be easily recalculated for a new VEF. Exhaust Flow Calculations For the following calculations, it is assumed that • hydrocarbons can be represented as propane, C3H8; (to obtain a propane-equivalent HC concentration, the HC reading must be divided by the propane equivalency factor (PEF), supplied with the analyzer; however, the error resulting from omitting this conversion is negligible) • NOx can be represented as nitric oxide NO; and • fuel can be represented by a hypothetical compound CxH yOz. For locally supplied CNG, x=0.987, y=3.970, z=0.014.12 The engine is a closed system, so that balance equations must be satisfied for the carbon, hydrogen and oxygen (listed in this order) content of intake air, fuel, dry exhaust, and water in the exhaust:

(C): MFf * x = MFe * (3*ce(HC) + ce(CO) + ce(CO2))

(2)

(H): MFf * y = 2 * MFw + 8 * ce(HC) * MFe

(3)

(O): MFf * z + 2 * IMF * 0.210 (21.0% oxygen in ambient air) = = MFw + MFe * (2*ce(O2) + ce(CO) + 2*ce(CO2) + ce(NOx))

(4)

where IMF MFf MFe

intake molar flow fuel molar flow dry exhaust molar flow


MFw molar flow of water in exhaust ce(xx) relative concentration of xx in exhaust gas Solving this system for exhaust flow yields

0.420 * x * IMF MFe = --------------------------------------------------------------------------------------------[ x * (2 * ce(O2) + ce (CO) + 2*ce (CO2) + ce (NOx) - 4* ce (HC)) - (3*ce (HC) + ce (CO) + ce (CO2)) * (z - 0.5*y) ]

(5)

These calculations are valid only when no gas is added to or removed from the exhaust gases between the engine and the sampling point. Air injection into the catalytic converter (not used on modern three-way catalytic converters) will lead to lower calculated exhaust flow, and lower emission readings. Exhaust gas recirculation (EGR) may, depending on the relative location of MAP or MAF sensors to the inlet of recirculated exhaust into the intake manifold, cause higher calculated exhaust flow, and higher emission readings. None of the tested vehicles was equipped with either the air injection or EGR. Calculating mass emissions Multiplying the relative concentration of each gas by the exhaust gas flow yields moles per second data; multiplying this by the corresponding molecular weight (NO - 30, CO - 28, propane - 44, CO2 - 44, O2 - 32 g/mol) yields grams per second data; numerically integrating these with respect to time yields grams per trip. When vehicle speed is measured as part of the engine data, converting the vehicle speed to miles per second and numerically integrating all records, excluding ones with missing emission data, with respect to time, yields miles per trip. If vehicle speed is not available, the odometer must be used to determine the length of the trip. Finally, grams per mile emissions are calculated by dividing the total mass emissions by the trip length. (Grams per kilometer emissions can be obtained by setting the scanner to report vehicle speed in km/h or by dividing the grams/mile figure by 1.609.)

RESULTS Pittsburgh test cycle During the day, vans are refueled as needed by a designated University service worker at a public CNG fueling station, operated by Equitable Gas Company. The station is located three miles (4.8 km) from the University. A test cycle has been developed to evaluate the vans using the portable emission measuring system. The refueling route has been extended to 4.1 and 3.9 miles (6.6 and 6.3 km) segments to incorporate cruise at 30-40 mph (50-65 km/h) and 50-60 mph (90-105 km/h), stop-and-go transient city driving, and climbing a steep hill. Both segments usually begin with engine start. For the first segment, the engine has cooled for 3-4 hours; for the second one the engine has been off for about 10 minutes during refueling. All twenty vans were tested on this cycle several times during May, June and July 1997. Mass emissions of HC, CO and NOx are shown on Figures 2,3,4, and mean values for each vehicle are listed in Table 1. While most of the data fits within a certain range, different for each van, considerable variance exists among the tests. The variance can be partly attributed to inconsistent testing conditions: Varying traffic patterns, inconsistent


driving habits among the drivers, and frequent equipment and software malfunctions in the initial part of the project (some tests beginning with the engine idling at its operating temperature). It was, however, observed that the same vehicle exhibits different emission levels under seemingly the same conditions. I/M 240 cycle To obtain mass emission data that can be compared with data on other vehicles maintained by the Alternative Fuels Data Center of NREL13, all twenty vans were tested using I/M 240 cycle in July and August 1997. Tests were performed at the Ohio E-Check station in Auburn Hills, OH, one of the closest and least busy public emission inspection stations operated in the Cleveland region by Envirotest under a contract with the Ohio EPA. The station uses a flame ionization detector (FID) for HC, non-dispersive infra-red (NDIR) units for CO and CO2, and a chemiluminiscence unit for NOx concentrations measurements. A special arrangement was made with Envirotest to receive readings in grams per mile rather than simply a “pass” certificate. The vans were driven, one or two at a time, 110 miles to Auburn Hills, and tested 5-40 minutes after arriving at the station. Two consecutive tests were run on each van with a 2-3 minutes gap between them; at least 5 minutes before the first run and between the runs, the engine was left idling; the records show a consistent coolant temperature of 87-93°C at the beginning of the cycle. To evaluate the portable system, on half of the I/M 240 tests, emissions were also simultaneously measured using the portable system. The complete set of “official” (stationary) and “portable” (measured by the portable system) I/M 240 data is listed in Table 1 and plotted in Figures 2, 3 and 4. The results show a considerable difference in emission data (1) among the vans, (2) between the first and second tests of each van, and (3) between readings from the station and the portable instruments. On the average, the first test of each van produced almost twice the amount of HC and NOx than the second one. Both tests gave comparable amounts of CO. Station readings generally are almost twice as high as those from the portable system. The coefficients of correlation between emissions measured by the portable and the stationary systems are 0.67 for NOx, 0.50 for HC and 0.68 for CO. This is comparable to correlation between stationary system results for first and second tests, which are 0.76 for NOx, 0.49 for HC and 0.57 for CO. Correlations between the Pittsburgh cycle and both “stationary” and “portable” I/M 240 data were also calculated and are listed in Table 2.

DISCUSSION Evaluation of the I/M 240 test It is assumed that the difference between an I/M 240 test and “real-world” emissions, observed by others, can be attributed primarily to the driving cycle, and the instruments of the I/M 240 test facility, which are regularly calibrated, should be capable of accurately determining the emissions of a typical gasolinepowered vehicle on the I/M 240 cycle. While the variation of the results on the I/M 240 cycle in measuring the emissions on the twenty CNG vans lies possibly in inconsistent vehicle performance, it is possible that the accuracy of the stationary instruments is insufficient for accurately measuring emissions of vehicles like CNG vans, which have very low concentration of pollutants in the exhaust. The test drivers have reported that at least on one occasion, the computer prompted the driver to start the engine, although it was already running. A station technician attributed this to the emission levels being below the detectability level of the I/M 240 instrumentation. Since the already low concentration gases (single or double-digit ppm values for HC and NOx and hundredths of percent of CO) are further diluted in the constant volume sampler before reaching the


instrument, it is possible that the accuracy of the instruments is low in this range. An alternative fuel emissions expert from West Virginia University has expressed the same opinion.14

Evaluation of the portable system Instrument accuracy for NGVs also appears to be a major deficiency of the portable system. While the upper end of the measurement range of the OTC analyzer is 10,000 ppm, 10% and 4,000 ppm for HC, CO and NOx, respectively, the usually observed values for the CNG vans are 5-20 ppm HC, 0.00 or 0.01% CO, and 5-50 (with excursions up to 500) ppm NOx. According to the manufacturer, the accuracy of the analyzer at these low ranges is not well known.9 The accuracy of measuring exhaust mass flow was also estimated. Equations 2, 3 and 4 were solved for fuel flow, and fuel consumption was calculated in a similar way as mass emissions, assuming molecular weight of natural gas 16.3 g/mol and 1 gallon being equivalent to 2.57 kg (5.66 lb) of CNG. Alternatively, on lowemission vehicles, the fuel flow can be approximated from mass emissions of CO2, assuming 1 mole of fuel produces 1 mole of CO2:

2570 grams / gallon equivalent Miles per gallon = ----------------------------------------------------------------------grams/mile CO2 * (16.3 g/molFUEL / 44 g/molCO2)

(6)

Since it was determined that this approximation introduces a relative error less than 1%, mass emissions of CO2 were used to determine the fuel consumption. The Pittsburgh cycle mean calculated gasoline-equivalent fuel consumption was 11.3±0.7 mpg (20.8±0.8 l/100 km), compared to long-term mean fuel consumption (of loaded vans) 10.4±0.6 mpg (22.6±1.3 l/100 km) according to fueling records. Measured and observed fuel consumption on long trips (16-18 mpg or 13.1-14.7 l/100 km) also appear consistent. It can be therefore assumed that the portable system is reasonably accurate in measuring real-world mass emissions of CO2. Recommendations and implications To further examine the accuracy of the portable system, a similar I/M 240 comparison as described will be done on gasoline-powered vehicles with higher pollution levels to evaluate the portable system in the normal operating range of the analyzers in order to address deficiencies other than analyzer accuracy. The accuracy of five-gas NDIR analyzers at low concentrations is also the subject of a study being conducted at West Virginia University.14 If the precision of the portable system on NGVs cannot be substantially improved, it can still be used even on low-emission vehicles, such as CNG vans involved in this project. Each measurement system has a systematic and a random error; the desired accuracy can be achieved by a sufficient number of measurements of a sufficient length to substantially reduce the random error, and by adjustments to compensate for the systematic error, which can be determined experimentally by comparative measurements against high-accuracy instruments at other facilities, such as the Emission Testing Lab at West Virginia University. If comparisons ultimately show that the portable system can successfully measure mass emissions on a “regular” vehicle, it will offer an inexpensive and sufficiently accurate way of measuring “real-world”


emissions. Attempts will be made to use a different engine diagnostic scanner, so that data from both the engine and the analyzer can be processed in real-time by a single program. This will also remove the one-hour time limit, so that the system can operate for an extended length of time. The system can then simply be mounted on a vehicle, left there while the vehicle is driven in its usual manner, and, at the end of the day, disconnected. Even with the improvements, the portable system described here is not expected to be as accurate in measuring mass emissions as special instrumented vehicles or emission testing labs. Its purpose is to offer a less expensive, more accessible test method on a more representative driving cycle, at the cost of lower accuracy. It should be noted that the data on which this paper is based has been only very recently obtained. Further analysis of existing and collection of additional data needs to be accomplished before a comprehensive conclusion on the portable system can be drawn. CONCLUSIONS 1. A portable, on-board system was developed to measure mass emissions on newer light-duty vehicles, with the intended use of measuring "real-world" mass emissions on compressed natural gas (CNG) vans. 2. Twenty CNG vans were tested several times using this system on an 8-mile test track in Pittsburgh. The data are of consistent magnitude, but considerable variance exists. 3. The data from two consecutive I/M 240 tests at a public emission inspection station on each van also exhibit considerable, inconsistent test-to-test variance, with the first test producing almost twice the HC and NOx emissions of the second test. 4. The portable system was used simultaneously with the station's system on 19 out of the 40 I/M 240 tests. The correlation coefficients between results of the two systems are 0.67, 0.50 and 0.68 for NOx, HC and CO. The portable system readings are consistently lower than ones from the station's instruments. The I/M 240 grams per mile emissions, as measured by the portable system, were also generally lower than those on the eight-mile test track in Pittsburgh. 5. Comparison of actual fuel consumption with that calculated using data from the portable system shows that the portable system measures CO2 mass emissions with reasonable accuracy. 6. The variation of data can be attributed primarily to the low accuracy of both station and portable gas analyzers in the low range of concentrations, exhibited by the CNG vans. Further development will be given to the portable system to find ways to reduce this variance. 7. The portable, on-board system provides an accessible, inexpensive alternative to the stationary dynamometer tests, which, in addition, use driving cycles that do not always represent "real-world" situations. 8. The portable, on-board system provides a more portable, less expensive alternative to research vehicles on which instruments are permanently mounted. 9. Because the portable, on-board system calculates "real time" exhaust mass, it provides a more accurate alternative to other portable systems using solely a five-gas analyzer to predict solely I/M 240 emissions. ACKNOWLEDGMENTS This project is being conducted under the sponsorship of the Pennsylvania Department of Transportation, PADOT Agreement No. 116025A, Supplement No. 1. Thanks also go to Jon Nordman, OTC SPX corporation, for the loan of a five-gas analyzer; Mike Cecere, University of Pittsburgh Motor Pool, for the loan of an engine diagnostic scanner and technical assistance; Joe Artman, John Morgan and Kathleen Miller, University of Pittsburgh Parking & Transportation Dept., for organization and planning; Ed Hutkay and Fred Kendra of Envirotest Systems for arranging for the I/M 240 tests; Paul Guenther and Donald Stedman, University of Denver, for information on on-board, portable emission measuring systems; Al Eaton and Ron Patrick, Engine Control and Monitoring, Sunnyvale, CA, for assistance with acquisition of data from Snap-On


instruments; Reid Garst, Sterling Engineering Sales, Richmond, VA, for information on gas mass flow meters; and engineering students Emily Gaspich, Minh Pham, Janelle Saulter, Conrad Smith and Melanie Vida for help with driving and data collection.


Table 1 - Emissions in grams per mile - summary

Van Mean St.dev. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Pittsburgh cycle Official I/M 240 Official I/M 240 Portable I/M 240 Portable I/M 240 (mean value) Series 1 Series 2 Series 1 Series 2 NOx HC CO NOx HC CO NOx HC CO NOx HC CO NOx HC CO 0.234 0.099 0.578 0.436 0.207 0.424 0.229 0.110 0.491 0.216 0.085 0.284 0.140 0.033 0.298 0.139 0.035 0.147 0.398 0.140 0.257 0.251 0.083 0.220 0.251 0.069 0.160 0.112 0.027 0.089 0.09 0.08 0.58 0.05 0.02 0.44 0.03 0.07 0.74 0.02 0.07 0.31 0.29 0.06 0.31 1.25 0.18 0.17 0.47 0.11 0.32 0.22 0.05 0.25 0.21 0.09 0.43 0.90 0.38 0.32 0.21 0.04 0.27 0.80 0.29 0.17 0.09 0.03 0.20 0.46 0.13 0.73 0.81 0.33 0.91 0.88 0.35 0.82 0.53 0.16 0.53 0.11 0.08 0.52 0.08 0.14 0.25 0.05 0.03 0.53 0.32 0.10 0.28 0.46 0.20 0.05 0.07 0.10 0.34 0.01 0.02 0.00 0.14 0.10 0.79 0.56 0.59 0.98 0.07 0.17 0.97 0.03 0.05 0.56 0.06 0.07 0.70 0.01 0.10 0.47 0.04 0.03 0.74 0.01 0.04 0.37 0.10 0.08 0.60 0.08 0.16 0.61 0.02 0.08 0.72 0.30 0.08 0.58 0.24 0.09 0.34 0.35 0.19 0.33 0.11 0.08 0.23 0.30 0.14 0.58 1.06 0.31 0.20 0.57 0.17 0.40 0.59 0.13 0.15 0.33 0.05 0.20 0.26 0.22 0.83 0.89 0.33 0.41 0.54 0.20 0.38 0.21 0.08 0.52 0.68 0.34 0.27 0.19 0.10 0.14 0.07 0.05 0.26 0.19 0.08 0.72 0.10 0.14 0.19 0.07 0.06 0.60 0.06 0.08 0.48 0.64 0.10 0.58 0.79 0.26 0.53 0.59 0.22 0.53 0.22 0.08 0.35 0.08 0.07 0.72 0.02 0.16 0.46 0.05 0.10 0.61 0.17 0.10 0.61 0.24 0.07 0.15 0.11 0.07 0.43 0.11 0.04 0.15 0.04 0.01 0.30 0.28 0.10 0.54 0.11 0.20 0.66 0.06 0.06 0.33 0.17 0.13 0.47 0.18 0.08 0.77 0.12 0.02 0.30 0.32 0.07 0.34 0.08 0.01 0.43 0.32 0.08 0.45 0.21 0.04 0.30 0.07 0.04 0.29 0.14 0.06 0.18 0.08 0.02 0.31


Table 2 - Correlation among emission data

Official to portable I/M 240 Official I/M 240 First to Second run Pittsburgh cycle to official I/M 240 Pittsburgh cycle to portable I/M 240

NOx 0.67 0.76 0.66 0.52

HC 0.50 0.49 0.43 0.33

CO 0.68 0.57 0.62 0.71

Figure 1 - Instrument setup and calculations schematic

Engine control unit diagnostic connector Exhaust gases from tailpipe Live engine data EXHAUST GAS ANALYZER

ENGINE DIAGNOSTIC SCANNER

Concentration readings RS-232 ports

PORTABLE COMPUTER

Engine data Concentrations data MAP, IAT rpm, MAF

Intake air flow

mph

Exhaust flow Grams per second

Trip length & time driving profile

Grams per mile emissions


Figure 2 - Hydrocarbons (HC) emissions Pittsburgh cycle Official I/M 240

0.5 0.45

Portable I/M 240

0.4 grams/mile

0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0

1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 CNG Van #

Figure 3 - Carbon monoxide (CO) emissions

2

Pittsburgh cycle Official I/M 240 Portable I/M 240

1.8 1.6 grams/mile

1.4 1.2 1 0.8 0.6 0.4 0.2 0 0

1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 CNG Van #


Figure 4 - Nitrogen oxides (NOx) emissions Pittsburgh cycle Official I/M 240

1.4

Portable I/M 240

1.2

grams/mile

1 0.8 0.6 0.4 0.2 0 0

1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 CNG Van #


REFERENCES 1

Kelly, N.A., Groblicki, P.J.: Real-world emissions from a modern production vehicle driven in Los Angeles. Journal of the Air & Waste Management Association, v 43 n 10 Oct 1993, p. 1351-7. 2 St. Denis, M.J.; Cicero-Fernandez, P.; Winer, A.M.: Effects of in-use driving conditions and vehicle/engine operating parameters on ‘off-cycle’ events. Comparison with FTP conditions. Journal of the Air & Waste Management Association, v 44 n 1 Jan 1994, p.31-38. 3 Cicero-Fernandez, P.; Long, J.R.; Winer, A.R.: Effects of Grades and Other Loads on On-Road Emissions of Hydrocarbons and Carbon Monoxide. Journal of the Air & Waste Management Association, v 47 n 8 Aug 1997, p. 898-904. 4 Jääskeläinen, H.E.; Wallace, J.S.: Performance and Emissions of a Natural Gas-Fueled 16 Valva DOHC Four-Cylinder Engine. SAE Technical Paper Series #930380, Society of Automotive Engineers, Warrendale, PA, Mar 1993. 5 Kelly, K.J.: Correlation of I/M240 and FTP Emissions for Alternative Motor Fuels Act Test Vehicles. SAE Technical Paper Series #941901, Society of Automotive Engineers, Warrendale, PA. 6 Di Genova, F., Sierra Research, Sacramento, CA, 1997. Personal communication. 7 Guenther, P.L., Stedman, D.H., Lesko, J.M.: Prediction of IM240 mass emissions using portable exhaust analyzers. Journal of the Air & Waste Management Association, v 46 n 4 Apr 1996, p. 343-8. 8 Garst, R., Sterling Engineering Sales, Richmond, VA, 1997. Personal communications. 9 Nordman, J., OTC SPX, Owantonna, MI, 1997. Personal communications. 10 Lyons, D., Emission Testing Lab, West Virginia University, Morgantown, WV, 1997. Personal communication. 11 Heywood, J.: Internal Combustion Engine Fundamentals. McGraw-Hill, 1988. 12 Based on CNG composition, provided by: Equitable Gas Co., Pittsburgh, PA, 1997. Personal communication. 13 Alternative Fuels Data Center (AFDC), National Renewable Energy Lab, Golden, CO, 1997; WWW site http://www.afdc.nrel.gov/ used as a comprehensive reference. 14 Bata, R., Emission Testing Lab, West Virginia University, Morgantown, WV, 1997. Personal communication.