Measurement of Gaseous HAP Emissions from Idling Aircraft as a Function of Engine and Ambient Conditions (2012)

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.

71 Introduction This section describes the development of the text matrix used to characterize the temperature dependence of HAP emissions from idling aircraft. It outlines background material and motivates the development of a new near-idle test matrix for use with on-wing aircraft engine testing. This section will describe some of the characteristics of measured fuel flow, N1 and EGT (exhaust gas temperature) during the test. Although the main report and Appendix A describe the complete results, some emissions performance will be discussed here as it relates to the engine operational states defined in the test matrix. A secondary goal of the project was to collect information on emissions performance during initial engine warm up. Those results are also discussed in this section. The emission index for hydrocarbons (and by extension HAPs) is highly dependent on the specific fuel flow rate in the vicinity of idle. A precise understanding of the dependence of the emission index on fuel flow rate is imperative for an accept- able description of the temperature dependence of hydrocarbon emission index. Thus significant emphasis has been placed on a well defined “near-idle” test matrix for the measurements described in this work. This section will examine emission index and fuel flow trends in the ICAO databank and perform a comparison with the data collected here. Finally, some prac- tical recommendations for using the test matrix described here are presented. Background on Emissions Certification for Idle The International Civil Aviation Organization (ICAO) recommends practices (ICAO 1981) for the emissions cer- tification performance of CO, NOx, Smoke Number and unburned hydrocarbons (UHC). The results are tabulated in the Emissions Databank (ICAO 2006). The certification databank sheets are the legal benchmarks for engine emissions performance and are used in regulatory evaluations. They are also used in performing global modeling, air quality research and airport inventory assessments (Pison and Menut 2004; Unal et al. 2005). Emissions of CO, NOx and UHC are compiled in units of grams per kilogram of fuel for the following named engine conditions: idle, approach, climb-out and take-off. NOx is expressed as “NO2 equivalents” while UHC is expressed as “CH4 equivalents.” Fuel flow rates for each engine condition are also tabulated for the named conditions for each engine. The Emissions Dispersion Modeling System (EDMS) uses the values tabulated in the ICAO databank in airport related inventory and modeling applications (Anderson et al. 2007; FAA 2006). Emissions inventory assessments could suffer bias due to uncertainties in the following: 1. time-in-mode 2. effect of ambient conditions on emissions, and 3. actual operational thrust levels or engine conditions. In order to address the potential bias introduced in the assumed time(s)-in-mode, it is possible to analyze specific datasets to refine these values for case studies. For example, an analysis of tabulated “out-off-on-in” times for each airport can refine the time spent in the bulk “idle” mode for a specific airport. Though knowledge of the distribution of time(s) spent in on the ground would improve an inventory assessment, it does not directly evaluate the actual operational fuel flow values noted above as #3. The effect of ambient conditions on HAP emissions near-idle engine states is the primary goal of this study. As discussed in the introduction and project motivation section, the idle phase of operations dominates the aircraft contribution to hydrocarbon emissions (Wood et al. 2008). Anecdotally, it is generally understood that most modern engines operation- ally idle on the ground at a fuel flow rate that is lower than the engine certification value (defined as 7% of rated thrust). A p p e n d i x B Development of the “Near-Idle” Test Matrix

72 For pollutants whose emission indices increase with decreas- ing thrust, such as CO and UHC, the ICAO certification value named ‘idle’ generally underestimates total emissions because the emission index increase outweighs the fuel flow decrease. The relationship of the emission index, however with engine state can become complicated at “idle” because the engine is not performing in the optimized region of its design space. On-wing emissions testing programs have characterized the difference between minimum ground idle (no bleed air demand) and the reference definition of idle (7%) (Spicer et al. 1994, Spicer et al. 1992, Timko et al. 2008, Yelvington et al. 2007). Observations of emissions from in-use aircraft suggest that, for a significant portion of the taxiway operational phase, the fuel flow rate is greater than minimum ground idle but less than the reference idle at 7% of rated thrust (Herndon et al. 2009). The test matrix developed for this project, and described in the next section, is intended to explore a practi- cal airline/pilot operational definition of the engine states in addition to the minimum ground idle with no bleed and 7% thrust reference points. This test matrix does not attempt to define a single point as representative of the idle phase of opera- tion for inventory development purposes but rather attempts to measure the emissions characteristics at each of the defined points. Test Matrix Design Points The aircraft engine states defined in this project were chosen based on settings using the cockpit control system. The engine states defined in this test were chosen to cover each of the following; typical operational procedures; atypical procedures; and the engine emissions certification reference condition. For the majority of the engines tested, the reference condition was set using the main fan speed (N1 = 25%). When the test was being conducted in weather where it was plausible that the pilots would use wing and/or inlet anti-ice systems, these states were investigated. The cockpit controls were used to enable the cabin climate control system. In order to investigate the no-bleed air engine state two differ- ent scenarios were employed. In one, the auxiliary power unit was operated to power other aircraft functions; in the other the entire bleed air demand was directed to the opposite engine. Whenever possible, the digital flight data recorder (DFDR) was used to correct and refine the engine sensor data (e.g., fan speeds, temperatures, fuel flows) recorded during the test. For each of the ground idle operational states, the bleed air demand was modulated by alternately operating the cabin Table B-1. Test matrix used during ACRP 02-03a winter testing. Test Point Time Nominal Name Instructions Mi ns. 00 Engine start Start engine and stabilize using no (or low) bleed on E1 10 Hold for test team and apparent engine warm-up 01 2 GI:nominal bleed Bleed balanced E1 <=> E2 02 2 GI: w/de-ice (max bleed) Bleed balanced E1 <=> E2 03 2 GI: w/de-ice (max all e1) Bleed shifted E1 <== E2 04 5 Set N1=25% Nominal bleed, balanced E1 <=> E2 05 1 GI: no bleed Bleed shifted E1 ==> E2 06 3 GI: nominal bleed, without de-ice Nominal Bleed, balanced E1 <=> E2 The truck moved downwind 07 1 Set N1=25% Nominal bleed, balanced E1 <=> E2 08 5 GI: nominal bleed, without de-ice Nominal Bleed, balanced E1 <=> E2 The truck creeps into position behind E2 09 2 GI:nominal bleed Bleed shifted E1 <== E2 10 2 GI: w/de-ice (max bleed) Bleed balanced E1 <=> E2 11 2 GI: w/de-ice (max all e2) Bleed shifted E1 ==> E2 12 2 Set N1=25% Nominal bleed, balanced E1 <=> E2 13 2 GI: no bleed on e2 Bleed shifted E1 <== E2 14 Shutdown Engine shutdown or ‘safe to pull away’

73 “packs” and the aircraft wide de-ice system. When both engines were on, the bleed demand was either balanced between them or shunted to a single engine. The test matrix included ground idle states with zero bleed flow, nominal bleed flow, de-ice and maximum bleed flow on a single engine. A summary of the winter time test matrix used is tabulated below. The times spent on each defined test point were vari- able depending on which experiments were being conducted. For example when running the gas chromatographic columns, additional time was spent on each condition. In order to obtain as many cold starts as possible, test point 0 was not always scripted. The fall test (DAL 2009) was similar, but the matrix dropped the states where additional deicing demand was placed on the engine because the ambient conditions did not warrant their use. Also, the DAL 2009 test was repeated three times in series due to the probe methodology experiments described in Appendix C. Results Fuel Flows Resulting from the Test Matrix The fuel flow rate is depicted as a function of the category of test point condition during the MDW 2009 tests (with a CFM56-7 engine) in Figure B-1. The categories in Figure B-1 are drawn from the nominal name given to each test point in the test matrix. The points have been horizontally offset (slightly) within each cate- gory in order to better depict points which are very close in fuel flow. Fan Speed, Exhaust Gas Temperature, and the Relationship with Fuel Flow In addition to fuel flow rate, two other engine character- istics, N1 and exhaust gas temperature (EGT) are indirectly modulated as a consequence of the test point conditions. N1 is related to the rotational speed of the fan in the turbofan engine. N1 is reported in units normalized by the maximum rotational speed for that engine and reported as a percent. N1 = 25% defines the 7%-idle condition for the CFM56 family of engines. Figure B2 depicts the relationship between N1, fuel flow and the effect of test conditions on exhaust gas tempera- ture. At first glance the figure suggests there are two disparate data groups that do not correlate well with fuel flow or EGT. Setting the data points with N1 greater than 23% aside for a moment, there is a general increase in EGT with fuel flow (dark reddish at lowest fuel flow, orange and yellow at higher fuel flow). Because the bleed air demand on the engine is responsible for increasing fuel flow rate and there is no con- comitant increase in the engine intake fan speed, there is relatively more combustion heat available than at lower fuel flows and the exhaust gas temperature increases. When the engine is set to perform at a specified rotational speed, 25%, the outer fan speed is ensuring that the fuel to air ratio and the pressure in the combustor are scaling accordingly. The exhaust gas temperature is cooled with the increase in by-pass air flow induced by N1. Figure B-1. Digital fight data recorded fuel flows by test condition category. Figure B-2. N1 fan speed is depicted versus fuel flow rate. The data points are colored by exhaust gas temperature. The small data points were part of the warm up test points. The circled data in blue all represent the points where de-icing technologies were enabled.

74 This figure underscores the complexities in the near-idle regime for the on-wing engine between fuel flow, N1 and EGT. Similarly the combustor pressure and fuel-to-air ratio are not related linearly when bleed air demand is introduced. The overall CO2 mixing ratio by volume at the engine exhaust plane is greater for the high bleed air demand condition than when N1 = 25%. Despite the complexity of accounting for bleed air demand in the engine conditions (between the lowest fuel flow rate—minimum engine idle, no bleed— and the greatest fuel flow rate—7% thrust or N1 = 25%), the emissions of VOCs are observed to be generally linear with absolute fuel flow rate. Warm Up Emissions During engine ground start there are three basic tempo- ral regimes to consider: pre-ignition, post-ignition/pre-idle acceleration, and post-ignition at ground idle. The pre-ignition hydrocarbon emissions are represented by evaporated fuel. Surrounding the moment of ignition there will be a mixture of fuel like hydrocarbons and some partially burned hydro- carbons. The post-ignition period will be characterized by increasing combustion efficiency and any hydrocarbon emis- sions will be associated with concomitant carbon dioxide, CO2, emissions. The examination of “warm-up” involves assessing the time constant of the phenomenon but also involves a consideration of what is influencing combustor performance during the dynamic post-ignition phase. The figure depicting a turbofan engine from the main body of the report is repeated here as Figure B-3 in order to facilitate the discussion of combustor performance. The figure is has been adapted from Kerrebrock (Kerrebrock 1977). The methodologies that predict emission combustor perfor- mance based on the pressure and temperature at the inlet to the combustor (P3 and T3 in Figure B-3) have been used to do an excellent job of predicting the gaseous emissions characteristics for turbofan engines at ground level and in flight altitudes (NEPAIR 2003, Sarli et al. 1975). Knowledge of the instantaneous combustion zone pressure and incom- ing pressure, along with fuel to air ratio can greatly improve the predictive skill of a model estimating combustor perfor- mance. The combustor efficiency will be related to P3 and T3. In an engine that has been off for a long enough time, the engine components will be in thermal equilibrium with ambient. When the engine is started, the core components, the turbine (approximately where EGT is measured) and the compressor stack will be relatively cold. While the EGT measurement is likely to be convolved with a time constant associated with the EGT harness hardware and the surrounding turbo-machinery, it nevertheless is a diagnostic of the warm-up time of the core components. Thus it may not be utterly implausible to use the EGT warm up time as a proxy of how long the compressor stack requires to “warm up” as well. The extent of heat transfer between the incoming air (at T2) and the air entering the combustor (T3) will be driven by the temperature gradient between T2 and the temperature of the compressor and the residence time in this region. In the condition of a cold start at idle speeds, the “warm-up” phenomenon (higher emissions of hydrocarbons and CO initially) is plausibly being influenced by the heat transfer occurring in the flow prior to entry into the combustor. In Figure B-4 a subset of the warm up data examines the time required to establish an equilibrium temperature within whatever time constant is associated with the EGT temperature measurement. The result depicted in Figure B-4 suggests that the exhaust gas temperature probably requires less than three minutes to reach ~90% of its steady state value. In the MDW 2009 winter test, an effort was made to characterize the post-ignition emissions to empirically address the question of “warm-up.” Whenever possible, the mobile laboratory was positioned ~40 m downwind of the engine during engine start. This was only done when the engines were known to be off for at least two hours prior to the start. The order of the test matrix initially precluded measurement of the engine starts for each of engine 1 and 2. We found we had sufficient flexibility with the sampling scheme, however, to attempt two additional near start observations. The HCHO emission index as a function of time follow- ing ignition is depicted in Figure B-5. The time offset has been computed from the flight data recorder information and the time coded notes taken within the mobile laboratory. The chemical information in the emissions profile at other engine-state changes has been used to refine the estimate of the time offset between the flight data recorder and truck time. Conservatively, the absolute time since engine start is certain to five seconds. Figure B-5 suggests that the CFM56-7B24 engines at the temperatures of this test (-7°C, -2°C) have two characteris- tic times. The first, initial rapid emissions change takes place within 20 to 60 seconds where the emission index drops considerably. The secondary change in emission index is Figure B-3. Schematic of a turbofan engine.

75 Figure B-4. This figure depicts the engine measured exhaust gas temperature as a function of the approximate elapsed time since engine start. The triangles, squares and circles denote different aircraft tests. The dark blue points reflect lower N1 rotational speeds for two of the starts than the lighter blue points. Figure B-5. The HCHO emission index is depicted as a function of the post-ignition warm up time. Although the relative times are precise, the absolute time following ignition for any of these curves is uncertain by five seconds. The solid lines represent fits of the data to a exponential decay. The approximate time constant associated with each curve is noted in the legend.

76 longer and characterized by a time constant of approximately 2 to 3 minutes. The data suggests that following cold start, the emission index is approximately doubled compared to warm operation for less than a minute. A simplistic approach to gauge the effect this has on inventory modeling of a 13-minute idle time [engine-on; taxi-out] is to add approximately one minute worth of additional idle emissions due to post-ignition warm up. The weaknesses in this argument are that no study has systematically looked at the reproducibility of this observa- tion or the time needed to return to cold start conditions. Furthermore, the initial ten seconds following ignition are not well represented in this dataset. The warm up aspect of the study was secondary to the overall test goals, but these results could be used to further refine assessments of “warm-up” emissions as well as design better tests for future work. Future Recommendation The digital flight data recorder data (DFDR) when testing an on-wing aircraft is extremely useful when analyzing the emissions at the low fuel flows tested in this project. Although all effort is made to coordinate the test matrix between the research team and the in-cockpit engine drivers, the DFDR dataset can corroborate the actual conditions and generally explain emissions performance. Data can be collected during the initial warm up period, how- ever it should not be regarded as a stable engine condition. In these experiments, some effort was made to emulate a synthetic push-back, engine start, brief run up to achieve “break-away” thrust, followed by a wait at ground idle with a nominal bleed air demand typical of the taxi out phase per our airline partners standard procedures. In addition, the amount and variability of pre-ignition (fuel-like) hydrocarbon emissions is not well addressed by emission measurements. Such pre-ignition emis- sions might be better estimated using statistical analysis of the time between the start of fuel flow and the time of actual emission in concert with fuel analysis of the species emitted. References Cited in Appendix B Anderson, C., S. Augustine, D. Embt, T. Thrasher, and J. Plante. 2007. Federal Aviation Administration, Office of Environment and Energy. Emissions and Dispersion Modeling System (EDMS) User’s Manual. FAA, AAE-07-01, Revision 3. FAA. 2006. Emissions Dispersion Modeling System (EDMS). www.faa. gov/about/office_org/headquarters_offices/aep/models/edms_ model/. Herndon, S. C., E. C. Wood, M. J. Northway, R. C. Miake-Lye, L. Thornhill, A. Beyersdorf, B. E. Anderson, R. Dowlin, W. Dodds, and W. B. Knighton. 2009. Aircraft Hydrocarbon Emissions at Oakland Inter- national Airport. Environ Sci. Technol. 43: 1730–1736. ICAO. 1981. International standards and recommended practices, envi- ronmental protection. Annex 16 to the Convention on International Civil Aviation, Volume II, aircraft engine emissions. 1st ed. Montreal, Canada: International Civil Aviation Organization. ICAO. 2006. International Civil Aviation Organization Aircraft Engine Emissions Databank. www.caa.co.uk/docs/702/introduction- 05102004.pdf. Kerrebrock, J. L. 1977. Aircraft Engines and Gas Turbines. Cambridge, Massachusetts and London, England: The MIT Press. NEPAIR. 2003. Development of the technical basis for a New Emission Parameter covering the whole AIRcraft operation. NEPAIR Final Technical Report, WP4/WPR/01 (G4RD-CT-2000-00182). Pison, I., and L. Menut. 2004. Quantification of the impact of aircraft traffic emissions on tropospheric ozone over Paris area. Atmos. Environ. 38: 971–983. Sarli, V. J., D. C. Eiler, and R. L. Marshall. 1975. Effects of operating variables on gaseous emissions. Paper presented at the Air Pollution Control Association Specialty Conference on Air Pollution Measure- ment Accuracy as it relates to Regulation Compliance, New Orleans. Spicer, C. W., M. W. Holdren, R. M. Riggin, and T. F. Lyon. 1994. Chemical composition and photochemical reactivity of exhaust from aircraft turbine engines. Ann. Geophysicae 12: 944–955. Spicer, C. W., M. W. Holdren, D. L. Smith, D. P. Hughes, and M. D. Smith. 1992. Chemical composition of exhaust from aircraft turbine engines. Journal of Engineering for Gas Turbines and Power 114 (1): 111–117. Timko, M. T., S. C. Herndon, E. C. Wood, T. B. Onasch, M. J. Northway, J. T. Jayne, M. Canagaratna, R. C. Miake-Lye, and W. B. Knighton. 2008. Gas Turbine Emissions Part 1. Hydrocarbons and Nitrogen Oxides. ASME J Eng Gas Turbines, submitted. Unal, A., Y. Hu, M. E. Chang, M. T. Odman, and A. G. Russel. 2005. Airport related emissions and impacts on air quality: Application to the Atlanta international Airport. Atmospheric Environ. 39: 5787–5798. Wood, E., S. Herndon, R. C. Miake-Lye, D. D. Nelson Jr., and M. Seeley. 2008. ACRP Report 7: Aircraft and Airport-Related Hazardous Air Pollutants: Research Needs and Analysis. Washington, DC: Trans- portation Research Board. Yelvington, P. E., S. C. Herndon, J. C. Wormhoudt, J. T. Jayne, R. C. Miake-Lye, W. B. Knighton, and C. C. Wey. 2007. Chemical Specia- tion of Hydrocarbon Emissions from a Commercial Aircraft Engine. Journal of Propulsion and Power 23: 912–918.