Summarizing and Interpreting Aircraft Gaseous and Particulate Emissions Data (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

12 This chapter describes the field studies, and how they were conducted, in chronological order. A tabulation of missions, dates, locations, operators, airframes, and engines is provided in Appendix A. A list of particle and gas-phase species mea- sured is provided in Appendix D. 3.1 APEX1 The Aircraft Particle Emissions eXperiment (APEX1) was the first ground-based experiment to simultaneously exam- ine gas and particle emissions from a modern commercial aircraft over the complete range of engine thrust settings. APEX1 was conducted at NASA Dryden Flight Research Center (DFRC), Edwards Air Force Base, California, between April 20-29, 2004. Particle and gas emissions from one of the NASA DC-8 aircraft’s CFM56-2C1 engines were measured as functions of engine thrust, fuel composition, plume age, and local ambient conditions. The specific objectives were to examine the impact of fuel sulfur and aromatic content on non-volatile (soot) and volatile particle formation; follow the evolution of particle characteristics and chemical composi- tion within the engine exhaust plume as it cooled and mixed with background air; examine the spatial variation of particle properties across the exhaust plume; evaluate new measure- ment and sampling techniques for characterizing aircraft particle and gas emissions; and provide a dataset for use in studies to model the impact of aircraft emissions on local air quality. APEX1 was a collaborative research effort sponsored by NASA, EPA and DOD. It brought together scientists from three NASA centers, the EPA, the U.S. Army and Air Force (USAF), three universities, engine and airframe manufacturers, and two private research corporations. During APEX1, particle and gas emissions were measured at 11 engine thrust settings for each of 3 different fuels (base, high-sulfur, and high-aromatic fuels) in samples drawn from probes located 1, 10, and 30 m (3, 33, 98 ft, respectively) downstream from the engine exhaust plane, see Figure 3. At the 1-m and 10-m (3-ft and 33-ft) sampling locations, multi- ple probe tips were used to examine the spatial variations of emissions properties across the exhaust plume. This testing matrix provided engine gas and particle emission informa- tion for more than 400 test conditions. Ambient conditions (during the testing period, the prevailing wind was from the southwest, but varied from all directions during the experi- ment period; wind speeds ranged from 0.4 to 14.3 m/sec (1.3 to 46.92 ft/sec); ambient temperature and dew point ranged from 16 to 36°C (61° to 97°F) and from −10 to −2°C (14° to 28°F), respectively; the wide ranges of ambient condi- tions impacted the engine operation and therefore the emis- sions data; some of the apparent variations in the measured data have been traced to changes in ambient conditions; ambient submicron particle concentrations measured at the testing site were typically <5 μg/m3), as well as engine temper- atures, fuel flow rates, and fan speeds, were carefully docu- mented for each of the test points examined during the experiment. APEX results represent the first and most exten- sive set of gas and particle emissions data from an in-service commercial engine wherein multiple instruments were used to quantify important species of interest. Two different engine testing matrices were used for each fuel used. The NASA test matrix was designed to investigate the effects of engine operating parameters on particle emissions. It included 11 steady-state engine thrust settings: 4, 5.5, 7, 15, 30, 40, 60, 65, 70, 85, and 100% rated thrust output. (Full take- off thrust at the high desert altitude corresponds to 93% of the rated engine thrust; henceforth, 100% will be used to de- note 93%). Except for the 100% thrust level where run-time was limited to 1.5 min, approximately 10 min were spent at each thrust setting to allow adequate time for analyzing sam- ples from each of the three downstream probes. The EPA test matrix followed the ICAO-defined LTO (landing-take off) cycle to simulate aircraft emissions at the airport, and consisted of approximately 26 min at idle (7%), 0.7 min at take off (100%), C H A P T E R 3 Primer on Field Studies

13 2.2 min at climb (85%), and 4 min at approach (30%) engine thrust settings. For the entire test matrix, each engine condition was repeated several times to get a measure of statistical repeata- bility and to allow adequate run time for the collection of time- integrated samples for chemical characterization. A portable weather station was erected a short distance from the test site and used to continuously monitor and record ambient wind, temperature, and pressure throughout the engine runs. Multi-port particle and gas sample rakes were designed, built, and deployed to map the spatial variations of emis- sions properties across the exhaust plume at the 1- and 10-m (3-ft and 33-ft) probe locations. As shown in Figure 4, these rakes held six traditional gas inlet (“G”) probes and six parti- cle inlet (“P”) probes that allowed introduction of dilution air just downstream of the probe tip. To provide adequate flow for filter and whole-air samplers, six additional, large-diameter gas inlet (“GG”) probes were attached to either side of the 1-m (3-ft) rake, aligned horizontally with the six, centerline- mounted gas probes. The particle and gas probes were mounted in an alternating pattern at 32-mm (1.26-in.) spacings and numbered from the top to bottom in the rake. The 1-m (3-ft) sample rake was minimally cooled with low-pressure water. At the 30-m (98-ft) location, a single probe sampled the mixed exhaust plume without further dilution. The center of the 1-m (3-ft) rake was aligned approxi- mately 77 mm (0.25 ft) to the side of the engine vent tube. Figure 3. Schematic of NASA DC-8 with sampling rakes and mobile laboratories. CFM56-2C1 Water InOut P1 G1 P2 G2 P3 G3 P4 G4 P5 G5 P6 G6 GG1 GG3 GG5 Figure 4. Orientation of sampling probes (P-particulate, G-gaseous, GG-external gaseous) with respect to the engine exit plane. Temperature probes (type-K thermocouples) and total and static pressure probes mounted on the rake were used to map the core-flow position. Particle samples collected at 1 and 10 m (3 ft and 33 ft) used the same type of probes and sample transport tubing. At the 1-m (3-ft) location, particle samples were diluted at each par- ticle probe tip with a concentric flow of dry nitrogen (N2) to suppress particle-particle interactions and the generation of new aerosol due to gas-to-particle conversion involving water and sulfuric acid. The 10-m (33-ft) samples were typically not diluted. The 30-m (98-ft) location probe was a single probe

14 that sampled the exhaust plume gases without introducing any dilution because exhaust plume was already diluted significantly with ambient air. Samples to all particle instru- ments were distributed through the sample distribution manifold. Modification of the aerosol size and composition due to various mechanisms such as inertia, thermophoresis, and diffusional effects, can occur in the sample train and are accounted for with calibration experiments (Lobo, Hagen et al. 2007). Gas samples (undiluted) were transferred through about 30 m (98 ft) of heated (177°C [350°F]) sampling line and distributed to individual instruments. 3.2 Delta Atlanta-Hartsfield Study The second of the APEX series of studies was carried out with the support of Delta Airlines at Hartsfield-Jackson Atlanta International Airport in September 2004. Mobile laboratories operated by Missouri University of Science and Technology (Missouri S&T), Aerodyne Research, Inc. (ARI), and the National Oceanic and Atmospheric Administration (NOAA) were deployed to conduct two series of measure- ments of aircraft engine-generated PM emissions. The first series was conducted at the maintenance facilities of Delta Airlines and focused on PM emissions in the vicinity of the exhaust nozzle of several different aircraft whose engines were cycled through a matrix of reproducible engine operat- ing conditions as in APEX1. The second series introduced a novel approach focusing on emissions generated under actual operational conditions. This series was conducted by placing the mobile laboratories adjacent to, and downstream of, active runways. In these latter measurements advected exhaust plumes generated by a broad mix of commercial transport aircraft taxiing and departing the airport during normal oper- ations were detected and analyzed. The Atlanta study was originally subject to nondisclo- sure agreements between the research team and Delta Airlines and, until December 2006, was referred to as the Un-Named Airline—Un-Named Airport (UNA-UNA) Study. In November 2006, the nondisclosure statement was re- scinded, permitting the public release of the data, and the study was henceforth known as the Delta-Atlanta Hartsfield Study. The Delta-Atlanta Hartsfield Study was the first opportunity to measure PM and gaseous emissions from in- service commercial transports. Dedicated engine tests on stationary aircraft took place be- tween 10:00 P.M. and 5:00 A.M. on September 21-25, 2004. The aircraft tested were selected from those scheduled to be overnight at the airport. The exhaust plumes of each aircraft were investigated using both probe sampling at the engine exhaust nozzle exit (Missouri S&T-ARI), see Figure 5, and remote sensing using LIDAR (light detection and ranging) (NOAA) at a point in the plume close to the exhaust nozzle exit, thus permitting comparisons of measurement techniques. Another objective was a study of engine-to-engine variation within the same class and, where possible, two aircraft with the same engine class were studied. The airframes and engines studied are listed in Table 1. The range of engine operating conditions examined focused on the LTO cycle with additional intermediate settings. For the JT8Ds, the complete range of thrust settings was explored, but for the higher thrust engines, transient instabilities in- duced vibration in the probe stands at mid- to high thrust, and this limited the range of thrusts sampled. The probe sampling measurements by Missouri S&T focused on physical characterization measurements including particle size distribution, number- and mass-based emission indices Figure 5. Probe rake assembly used during Phase 1 of the Delta Atlanta-Hartsfield Study. Date Thrust (kN) September 22, 2004 93 September 23, 2004 93 September 23, 2004 217 September 24, 2004 258 September 24, 2004 166 September 25, 2004 Aircraft Number 908 918 134 1816 635 640 Airframe MD-88 MD-88 B767-300 B767-400ER B757-200 B757-200 Engine JT8D-219 JT8D-219 CF6-80A2 CF6-80C2B8F PW 2037 PW 2037 166 Table 1. Airframes and engines measured during the Delta-Atlanta Hartsfield Study.

15 and soluble mass fraction. ARI focused on using an aerosol mass spectrometer and related supporting instruments to quantify the composition of the particles as a function of size and thrust. Concurrent with the probe sampling, remote sens- ing was performed by NOAA using a mobile LIDAR system. Also, NOAA supplied and operated the LIDAR, which used eye-safe ultraviolet light from a laser pulsing at 10 Hz and scanned the beam up and down in a vertical plane perpen- dicular to the direction of engine exhaust. The LIDAR system was contained in a trailer positioned about 300 m (984 ft) from the aircraft. The principal wavelength for this project was 355 nm. The back scatter (or reflection) of energy from the laser by the total aerosols emitted by the aircraft engine was measured just behind the rear stabilizer of the aircraft by the LIDAR. Upon completion of the dedicated engine testing, the research groups turned their attention to measurements of aircraft emissions on the airfield at various locations near the ends of runways where takeoff operations were occurring. With the exception of the data acquired on September 26, it was not possible to collocate the LIDAR and the Missouri S&T-ARI measurement systems. Despite this limitation, for the overall project, both groups were successful in data col- lection with 344 takeoffs measured by LIDAR and more than 500 taxi and takeoff events by the Missouri S&T-ARI meas- urement system. The Missouri S&T and ARI mobile laboratories were po- sitioned (with assistance from airport operations staff) just downwind of an active runway, as shown in Figure 6. Two locations were selected to perform these measurements based on the prevailing wind direction on a given day. On September 27, 2004, the prevailing wind was from the N/NE, and Missouri S&T and ARI were collocated on the western end of the airport’s southern runways. On September 28 and 29, 2004, the prevailing wind shifted to the W/NW direction, and Missouri S&T and ARI moved to the eastern end of the south- ern runways. In this work, exhaust plumes advected in the di- rection of the sampling systems were continuously analyzed. Exhaust pollutant emission ratios relative to exhaust CO2 were determined for various gas-phase and particulate met- rics by looking at the concomitant rise in the measurement of a target pollutant above background with increased CO2. These emissions ratios were converted to fuel-based emis- sions indices using above-ambient CO2 as an internal exhaust plume tracer. The characteristics of advected plumes—plume rise and plume spread (horizontal and vertical)—were meas- ured using the LIDAR technique. These measured parameters are key variables in dispersion modeling. 3.3 JETS-APEX2 The impetus for the JETS-APEX2 study came from CARB. In late 2004, CARB had initiated discussions with the Missouri S&T Center of Excellence for Aerospace Particu- late Emissions Research (Missouri S&T COE), the Port of Oakland for Oakland International Airport (OAK), and South- west Airlines (SWA) to provide access to in-service commer- cial B737 aircraft for such measurements since SWA operates exclusively with B737s and is the major airline operating out of OAK. In the spring of 2005, Project JETS-APEX2 emerged as a study funded by multiple agencies (CARB, NASA, FAA, EPA, Missouri S&T, UCR, UCF, AEDC, GE, Boeing, SWA, OAK, and ARI) to produce the first measurements with state- of-the-art analytical equipment of speciated total organic gases (TOG) and PM from engines on typical in-use Boeing 737-type commercial aircraft. JETS-APEX2 consisted of two series of experiments similar to the Delta Atlanta-Hartsfield study. The first series focused Figure 6. Schematic of layout of mobile laboratories during the downwind study at Delta Atlanta-Hartsfield.

16 on PM emissions in the vicinity of the exhaust nozzle of several different aircraft whose engines were cycled through a matrix of reproducible engine operating conditions as in APEX1. The second series focused on emissions generated under actual operational conditions, conducted by placing the mobile laboratories adjacent to, and downstream of, active runways. In these latter measurements, advected exhaust plumes generated by the mix of commercial transport aircraft taxiing and departing the airport during normal operations were detected and analyzed. The first series of experiments relied heavily on experience gained in the previous APEX study where custom-designed probes and extensive support equipment were used to sample jet exhaust in the on-wing position at six thrust settings: 4%, 7%, 30%, 40%, 65% and 85%. In all, both engines of four parked 737 aircraft were tested. Particle-laden exhaust was extracted directly from the combustor/engine exhaust flow through the probe, transported through a sample train, distributed, and analyzed in each group’s suite of instrumentation. Sampling probes were located at different positions downstream of the engine exit plane: 1 m, 30 m, and 50 m (3 ft, 98 ft, and 164 ft) on the starboard side, and at 1 m (3 ft) on the port side of the aircraft. These aircraft engine emissions measurements were performed at the Ground Runup Enclosure (GRE) at OAK during August 2005 (see Figure 7). The engine types were selected to represent both old (-300 series) and new (-700 series) technologies. Real-time PM physical characterization was conducted by Missouri S&T. Size distributions from 5 nm to 1 μm were measured for all test points and associated aerosol parameters (i.e., geometric mean diameter, geometric standard deviation, total concentration, and mass and number-based emission indices were evaluated). ARI made real-time measurements of gaseous emissions using: (1) tunable infrared laser differential absorption spec- troscopy (TILDAS) based on both lead-salt diode and quantum cascade laser sources for several important trace species emissions; (2) proton-transfer reaction mass spectroscopy (PTR-MS) for HC; and (3) chemiluminescence measurement for NO. These measurements were converted to emission indices using CO2 measured with a nondispersive infrared absorption technique. Chemical composition of the particle emissions was quantified using an aerosol mass spectrometer (AMS) in concert with a multi-angle absorption photome- ter (MAAP, for black carbon mass) and particle size and number measurements. The TOG, PM, metals, and ions were collected on filter membranes by the University of California–Riverside Center for Environmental Research and Technology. Teflo filters were used to acquire PM mass, and metals and ions concen- trations. For TOG, various sampling media—including SUMMA™ canisters, 2,4-dinitrophenyl hydrazine (DNPH) Figure 7. Layout of the mobile laboratories in the GRE and probe rake assembly used in JETS-APEX2.

cartridges, and thermal desorption tubes—were used. After the field campaign was completed, analysis of the DNPH cartridges and SUMMA canisters revealed anomalous CO2 concentrations, which were attributed to a leak in a subsystem of the sampler. Also, C4-C12 HC values based on the concen- trations measured from the thermal desorption tubes (TDS) were much lower than expected from APEX1 and other re- search. Since this leak introduced an unquantifiable dilution in these subsystems, the emission factors for the light HC and carbonyls could not be calculated. The second set of measurements sampled jet engine ex- haust downwind of an active taxiway and runway at OAK while the aircraft performed standard LTO. The runway tests demonstrated the potential of downwind emissions monitoring adjacent to active taxiways and runways as a means to rapidly acquire evolving aircraft PM characteris- tics from in-service commercial aircraft. Emissions were monitored during a 12-hr period of daylight aircraft opera- tions along a single runway where the downwind exhaust plumes for over 300 aircraft were sampled. An aerial view of the test venue is shown in Figure 8. Mobile laboratories from Missouri S&T and ARI were collocated downwind on the eastern end of the runway with the prevailing wind direc- tion coming from the W/NW. The Missouri S&T laboratory focused on the physical characterization of the downwind PM and measurement of CO2 (Whitefield et al. 2007). The ARI laboratory focused on characterization of PM compo- sition and measurement of CO2, and trace combustion gases (Herndon et al. 2007). Over 300 aircraft landings and departures were detected and monitored during the period from 7 A.M. to 7 P.M. on August 26, 2005. Aircraft tail numbers and operational status (i.e. taxi, takeoff, and landing) were acquired through visual observation, including video recordings. Aircraft-specific airframe and engine data were obtained by correlating these tail numbers with an FAA database. Figure 9 illustrates the distribution of aircraft types operating at OAK during the day of the tests. In all, exhaust from 15 different airframe types was captured, and approximately 63% of the aircraft were B737s. 3.4 APEX3 APEX3 was the fourth campaign in the APEX series. The main objective of APEX3 was to advance the knowl- edge of aircraft engine particle emissions. APEX3 was con- ducted at Cleveland Hopkins International Airport (CLE) from October 26 to November 8, 2005. In APEX3, as in the three previous studies, engine exhaust emissions and plume devel- opment were examined by acquiring data from the exhaust nozzle and in the near-field plume from a range of stationary commercial aircraft. A complementary study of downwind plumes during normal operations was abandoned because the prevailing winds during the scheduled sampling times did not transport the plumes to the available sampling locations. As with previous studies, APEX3 was a collaborative re- search effort and was supported by the following organizations: (1) Researchers from NASA, EPA, U.S.DOT Volpe Center, the Air Force Arnold Engineering Development Center (AEDC), Missouri University of Science and Technology (Missouri S&T), Montana State University (MSU), and Aerodyne Research, Inc. (ARI); 17 Prevailing wind Figure 8. Aerial view of the OAK test venue for downwind plume monitoring. B757, 7 A300, 5 A310, 1 A320, 16 B737, 177 Turboprop, 1 B727, 8 MD-80, 9 CRJ, 13 MD-11, 2 DC-10, 13 Learjet, 10 Figure 9. Distribution of aircraft activity as a function of airframe.

(2) Engineers from the aviation industry including Conti- nental, Express Jet, FedEx Express, General Electric, Pratt and Whitney, Rolls-Royce, and Rolls-Royce North America; and (3) Sponsors from FAA, EPA, NASA, and Cleveland Hopkins International Airport. Particulate matter and gas-phase emissions were acquired from a range of current in-service commercial aircraft engines including regional aircraft (ERJ 135/145 equipped with AE3007 engines) provided by Express Jet, passenger aircraft (B737- 300 with CFM56-3B1 engines, B757 with RB211-535E-4B engines) provided by Continental Airlines, a freight aircraft (A300-600 with PW4158 engine) provided by FedEx, and the NASA general aviation aircraft (Learjet 25 with CJ610 turbo- jet engines). Engine exhaust was sampled at three different locations in the plume, nominally 1 m (3 ft) (i.e., exhaust nozzle), 15 m (49 ft), and 30 m (98 ft) for the small aircraft (regional jet and general aviation jet), and 1 m, 30 m, and 45 m (3 ft, 98 ft, and 148 ft) for the large aircraft. 18