TRB 08-1311 Link-Based Emission Factors for Heavy-Duty Diesel Trucks Based on Real-World Data

Heavy-duty diesel vehicles contribute a substantial fraction of nitrogen oxides and particulate matter to on-road vehicle emission inventory. The objectives of this study are to estimate roadway link-based emission rates for heavy-duty trucks for use in emission inventory estimation, and to quantify the impact of factors affecting truck emissions. A speed-acceleration modal emissions approach is developed from a database gathered via a portable emissions measurement system for single rear axle and tandem dump trucks. Second-by-second real-world truck speed profiles on links are analyzed based on observed patterns of time distributions of speedacceleration modes. Link-based emission rates are estimated as the product of the fraction of time spent in each mode and the corresponding modal average emission rate. The sensitivity of link-based emission rates to key factors including chassis type, vehicle load and fuel type is presented. Single rear axle trucks have lower emission rates than tandems for CO2, PM, NO and HC, but higher CO emission rates. Loaded trucks have higher fuel use and emissions than unloaded trucks. Replacing diesel fuel with biodiesel fuel for heavy-duty trucks may reduce tailpipe NO exhaust emissions and will reduce emissions of PM, CO and HC. However, both fuels generate similar CO2 emissions. Benchmark comparisons for link-based emission rates show that NO emission rates increase with mean speed. However, link-based CO and HC emission rates were not as sensitive to speed variation as NO emissions. The link-based emission rates approach is recommended to couple heavy-duty vehicle emission inventory estimation with transportation demand models. Frey, Rouphail and Zhai 2 INTRODUCTION Heavy-duty diesel vehicles (HDDVs) contribute a substantial fraction of nitrogen oxides (NOx), and particulate matter (PM) released to the atmosphere (1). In 2002, heavy-duty diesel vehicles accounted for approximately 46% of NOx and 54% of PM10 of the nationwide on-road vehicle emission inventory (2). Therefore, close attention should be paid to HDDV emissions characteristics and mitigation as they relate to vehicle duty cycles. Emission measurement methods for HDDVs typically include engine and chassis dynamometer tests, tunnel studies, and remote sensing (3-6). Many engine dynamometer test cycles are based upon steady-state modal profiles that are not likely to be representative of real world vehicle activity patterns. Chassis dynamometer tests are expensive and there are few such dynamometers. Tunnel studies are limited in their ability to discriminate among specific vehicle types. With remote sensing, each measurement is only a snap shot of the vehicle activity at a single location, and thus does not characterize the entire duty cycle (7). The U.S. Environmental Protection Agency (EPA) has developed an on-road transportable diesel emissions characterization facility for measuring real-world emissions of heavy-duty combination trucks (8). Portable emissions measurement system (PEMS) and transportable HDDV emissions testing laboratories have been developed to allow a better quantification of transportation activity and operational effects on vehicle emissions, especially under real-world traffic conditions (9-11). HDDVs are greater than 8,500 pounds in gross vehicle weight (12). Emissions from HDDVs are affected by various factors including vehicle class and weight, driving cycle, and fuel type (13-14). On-road measurement of fine particle in a tunnel study showed that HDDVs emit 15-20 times the number of particles per unit mass of fuel burned compared to light-duty vehicles (15). The emission rates of volatile organic compounds and carbon monoxide (CO) from gasoline-fueled single-unit trucks can be 2.5 to 5 times higher than those of heavy-duty diesel trailer trucks. Emission rates of NOx from diesel-fueled tractor-trailer trucks can be five times higher than those of gasoline-fueled single-unit trucks (16). Increases in gross vehicle weight may also result in increases in NOx emission rates during accelerations and higher-speed steady-state operations (14). CO and PM emission rates were found to be insensitive to the vehicle weight during nearly steady-state operation, but increased with weight when vehicles were tested on transient driving cycles (1). NOx emission rates from HDDVs driving at low speeds, in simulated congested traffic, can be much higher than while cruising on the freeway (17). In addition, biodiesel fuel is gaining increasing interest as an alternative fuel for HDDVs. An evaluation of biodiesel impacts on exhaust emissions by the U.S. EPA showed that compared to petroleum diesel, B20 biodiesel (20% blend stock and 80% petroleum diesel) reduced about 10% of CO and PM emissions, and 20% of total hydrocarbon emissions, but increased NOx emissions by 2 percent (18). There are few studies about how emissions from biodiesel fueled heavy-duty are affected by real-world vehicle operation patterns and loads. The emission factor model, MOBILE6 utilizes diesel engine emissions certification data as well as a series of conversion factors to convert certification data derived from engine testing to in-use grams per mile emission factors (12). In addition, the off Federal test procedure cycle was developed to estimate excess NOx emissions produced by HDDVs that are not explicitly covered by a certification test (19). However, basic emission rates are estimated from stationary dynamometer tests. Furthermore, the emission factor model does not account for the effects of truck operating weight on emissions. Emission rates for HDDVs were found to be highly dependent on vehicle operating mode (17). A vehicle activity-based study estimated emissions of HDDVs (20-21). In that case study, NOx emissions from transit buses were found to be Frey, Rouphail and Zhai 3 sensitive to acceleration, but not as much to vehicle speed, for a given acceleration range. However, the effect of vehicle load on link-level emissions could not be effectively evaluated for HDDVs under real-world driving cycles. Compatibility between vehicle activity indictors from transportation and emissions models outputs is a pre-requisite to the development of accurate estimates of regional emission inventories. Transportation models produce link-level activity data. Therefore, there is a need for link-based emission factors for HDDVs in emission inventory estimation. OBJECTIVES The goal of this study is to develop a methodology for estimating roadway link-level emission rates, and to evaluate the effects of vehicle activity and operation on those emissions. The methodology is illustrated based on heavy-duty vehicles operating using diesel and biodiesel fuels under real-world driving cycles and under different loads. In addition, emission rate differences for different chassis types are quantified. The principal objectives of this research are to: (a) estimate link-level emission rates for heavy-duty trucks; (b) quantify the effects of vehicle activity and load on truck emissions; (c) compare emissions for different chassis types; and (d) compare emissions for diesel versus biodiesel fueled trucks. DATABASE DESCRIPTION Emissions Data An OEM-2100 “Montana” PEMS was used for data collection. This OEM-2100 components, data collection capabilities, and data quality assurance protocols, as well as the study design for field data collection and basic results, are detailed in a previous TRB paper (22). Frey and Kim (22) tested two categories of dump trucks including four single rear axle trucks and four tandem trucks with engines subject to Tier 1 emission regulations. The engine displacement was 7.2 liters for single rear axle dump trucks and 10.2 liters for tandem trucks. The average weight of a typical load was approximately 7.0 tons for the single rear-axle trucks and 14.5 tons for the tandems. The load weight was comparable to the unloaded weight of the vehicle. Each vehicle was tested for one day using B20 biodiesel and one day using petroleum diesel. Measured pollutants included CO, hydrocarbons (HC), NO, opacity, and carbon dioxide (CO2). The field study was designed to test the effect of vehicle type, fuels, loading configuration and operating mode. Activity Data Vehicle activities under real world traffic conditions were investigated at the roadway link level. A link is defined as a roadway segment between two junctions. Thus, the segment between two interchanges is defined as a freeway link, and the segment between two traffic signals on a surface street is regarded as a surface street link. Second-by-second GPS coordinates of each vehicle during its trip were recorded by the PEMS and overlaid on a transportation network GIS map. HDDV speed profiles on freeways collected by Battelle in California were also used (23). Measured link speed profiles for single rear axle and tandem dump trucks on arterials were gathered by Frey and Kim (7). Frey, Rouphail and Zhai 4 METHODOLOGY The methodology employed for estimating link-based emission rates consists of: (1) developing a speed-acceleration modal approach based on PEMS data, and estimating modal emission rates; (2) estimating facility-specific real-world truck activity data at the link level; (3) estimating speedand facilityspecific average emission rates for heavy-duty trucks; and (4) quantifying the impacts of selected vehicle activity and fuel factors on truck emissions. Speed-Acceleration Modal Analysis For heavy-duty vehicles, the chassis type, load and fuel type are all factors that may affect exhaust emissions. The emission database was classified by chassis type (single versus tandem axles), load status (empty versus loaded) and fuel type (diesel versus biodiesel). For a given data subset, emissions data in each second were stratified into discrete predetermined modes according to instantaneous speed and acceleration measurements. Five acceleration ranges (high-deceleration, low deceleration, cruise, low acceleration, and high acceleration), and thirteen speed ranges (from 0 to 65 mph in 5 mph increments) were constructed from the database. Idling was categorized as a separate mode. Modal average emission rates were then estimated using second-by-second PEMS measurements. Link-based Average Emission Rates Estimation Speed profiles were categorized by facility type and link mean speed. The recorded speed profiles were classified into average link speed ranges, in increments of 5 mph, ranging from 25 to 45 mph for arterials and from 45 to 60 mph for freeways. Second-by-second speed profiles for links were subsequently stratified into discrete speed-acceleration modes. Such stratification enables the use of modal average emission rates to obtain aggregate estimates of emission rates on a link. Link emission rates are estimated as: