Exhaust Emissions from In-Use General Aviation Aircraft (2016)

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.

Other Parameters Affecting Emissions 45 engine exhaust temperatures. To combat this problem, dibromoethane is added to 100 LL. Dur- ing combustion, the lead reacts with the bromine, thereby forming PbBr2, which is volatile and passes through the exhaust system before it condenses. The aerosol mass spectrometer confirmed the presence of lead and lead bromide in the particle phase. Based on information found on the Internet (http://www.pipercubforum.com/marvel.htm), older low-compression piston aircraft engines at taxi/idle speeds have engine exhaust temperatures below the volatility limit of lead bromide and suffer lead fouling. Two products, Marvel Mystery Oil and Alcor TCP, show up on pilot forums as products that can minimize lead fouling. Analysis of a sample of 100 LL doped with Marvel Mystery Oil using the PTR-MS did not result in any apparent change in the amount of toluene observed. Marvel Mystery Oil contains a small amount of chlorinated hydrocarbon, which could produce lead chloride, a product that is more volatile than lead bromide. A sample of Alcor TCP was not available for analysis, but the material safety data sheet (MSDS) for this product lists toluene as a major ingredient (Table 5-3). All pilots and other aircraft operators were asked about fuel additives. All reported no such use. The research team cannot explain the high levels of toluene observed if no fuel additives were present. This discrepancy underscores the importance of determining how widespread the use of fuel additives is in the GA sector. Hydrocarbon Emissions from GA Are Primarily Unburned Fuel The exhaust gas composition of GA engines is expected to reflect the effects of operating under excess fuel (rich combustion) conditions. Because these engines operate far from their stoichio- metric fuel-to-air limit, it is logical to assume that a significant fraction of the HC measured by the flame ionization detector instrument consists of unburned fuel. In this section, the research team examines this assumption by examining the fraction of the fuel carbon that can be accounted for by the individual species that were directly measured (e.g., CH4, C2H2, C2H4, C2H6, HCHO, CH3CHO, acetone, C6H6, C7H8, C8H10, and C10H8). With the exception of toluene (C7H8), which is present as a fuel additive (previous section), all of these components are combustion byproducts. Although this analysis cannot definitively identify the HC components that were not directly measured, one can conclude through comparison with automotive piston engine exhaust studies that the unidentified portion of the exhaust is primarily composed of unburned fuel. Figure 5-8 shows histograms of the fuel carbon accounted for by the individual components measured, one representing the total and another with the total excluding the toluene contribu- tion, which is present at considerable quantities only when it is present in the fuel. Figure 5-8 illustrates several important observations. First, the measured components only account for a small fraction (10 to 20%) of the total. Second, the measured fraction is highly variable, with at least some of the variability arising from the presence of the toluene fuel additive. The influence of toluene provides direct evidence that unburned fuel exists in significant quantities within the exhaust. Automotive piston engine exhaust studies (Schauer et al. 2002) without emission control systems show that the exhaust composition is strongly correlated to the fuel composition, where Table 5-3. MSDS composition information for Alcor TCP fuel additive.

46 Exhaust Emissions from In-Use General Aviation Aircraft approximately 80% of HC exhaust consists of the same compounds present in the fuel in nearly the same distribution. The remaining 20% of the HC exhaust consists of decomposed fuel in the form of small hydrocarbons such as acetylene and partially oxidized material like formaldehyde. Tallying the five most prevalent non-fuel components, formaldehyde, acetylene, ethylene, meth- ane and acetaldehyde in Schauer’s study accounts for 12.5% of the total HC exhaust carbon, which is comparable to the results observed in Figure 5-8 for most of the aircraft studied here. This result is not surprising, given that GA and vehicle engines both employ spark-initiated piston engines. It therefore stands to reason, by comparison, that the composition of the bulk of the unidentified GA HC resembles that of AVGAS 100LL fuel. In assessing the effect of HC emissions at a GA airport, composition information is useful because not all hydrocarbon species affect local air quality and human health equally. Yet here again the research team observed significant variability in the composition of HC emissions among piston engines. Statistical tools that use the measured distribution of compositions pre- sented here can help bound speciated hydrocarbon emission burdens and enable informed deci- sions at airports. New Measures of PM to Replace “Smoke Number” The current ICAO database for large commercial gas turbine engines quantifies emissions for NOx, CO, Unburned Hydrocarbons (UHCs, equivalent to HCs elsewhere in this document), and smoke number (ICAO 2013). The latter was instituted to provide a quantity for visible smoke that could be regulated and that could be used to control the smoke particle emissions through the metric of that smoke number. PM science has progressed tremendously since the 1970s, when SN was introduced, and the international aviation community is active in develop- ing a new standard for aviation non-volatile PM (nvPM) emissions. Rather than just a measure of the visible obscuration like the SN, the new standard will report nvPM mass and number, quantities that are directly connected to environmental and human health impacts. For avia- tion PM, like many other combustion-generated PM, the particles’ sizes are small enough to be considered part of existing PM10 (PM smaller that 10 µm) and PM2.5 (PM smaller that 2.5 µm) Figure 5-8. Fraction of exhaust UHC accounted for by the individual measured components under idle power condition.

Other Parameters Affecting Emissions 47 regulations, and any likely future regulations with a smaller cutoff (for particles smaller than 1 or 0.1 µm, for instance). ICAO’s Committee on Aviation Environmental Protection (CAEP) is developing a regulatory standard for non-volatile PM (nvPM) number and mass-based emissions from civil aviation air- craft engines to replace the standard of smoke number measurement. The standardized sampling and measurement method that will be used for this future regulation has been defined in the Aerospace Information Report (AIR) 6241, developed by the Society of Automotive Engineers (SAE) Aircraft Exhaust Emissions Measurement Committee (SAE E31). This standard method will become normative once it is converted into a certification document, which will then be used by engine manufacturers in certifying aircraft engines for nvPM emissions. The system defined in AIR6241 is designed to operate in parallel with existing sampling systems for gaseous emissions and smoke certification defined in ICAO Annex 16. The system specifications in AIR6241 build on the work conducted in previous research to evaluate sampling and measurement methods for aircraft engine nvPM emissions measurements. The primary measurement instruments in the AIR6241 systems report nvPM number and mass-based emissions. The nvPM number can be measured using an AVL particle counter (APC), which includes a volatile particle remover (VPR) consisting of a two-stage dilution with a rotary diluter and a catalytic stripper, and an n-butanol-based condensation particle counter (CPC) TSI 3790E, which has a 50% cutoff diameter, D50, at 10 nm. During the research, an APC owned by United Technologies Research Center (UTRC) was deployed to each of the three field measure- ment campaigns. The reported number emissions of non-volatile PM (nvPMn) were determined based on the APC measurements in compliance with the AIR6241 recommendations. For nvPM mass measurements, two real-time, high-resolution instruments that satisfied the performance specifications were recommended: the Artium Laser Induced Incandescence LII-300 and the AVL Micro Soot Sensor (MSS). However for this research, due to the lack of an available MSS or LII instrument, the research team used the conventional filter-based multi-angle absorp- tion photometer (MAAP) to measure the nvPM mass emissions. The MAAP instrument has been widely applied to determine nvPM mass for many field measurements on commercial aircraft engine emissions [e.g., Experiment to Characterize Aircraft Volatile Aerosol and Trace-Species Emissions (EXCAVATE), Aircraft Particle Emissions eXperiment (APEX) I-III, and Alternative Aviation Fuel Experiment (AAFEX) I-II]. ACRP Research Report 164 reports these nvPM quantities of mass and number, but goes fur- ther in reporting the total mass and number, as well. The total quantities represent the sum of the nvPM quantities plus the volatile contributions to mass and number, respectively. The total mass and number become important once the exhaust mixes with cooler ambient air and vola- tile PM condenses and will certainly be the PM quantities emitted to the atmosphere. PM Volatility Is High for Piston Engines A typical GA piston engine shows a particle signature where high-volatility particles dominate. This trend is particularly obvious for particulate matter number (PMn) where the total particle count (volatile + non-volatile) exceeds the non-volatile count by an order of magnitude or greater. Figure 5-9 compares the emissions burden per LTO cycle (times-in-mode as shown in Table 3-1) for several engine families. These trends agree with the research team’s understand- ing of piston engine operation. Total PM mass (tPMm) is reported by the EEPS, however, this value is not appropriate for direct comparison to non-volatile PM mass (nvPMm) reported by the MAAP. The EEPS reports mass by calculation from the measured size distribution, assuming a particle density of 1.0. The

48 Exhaust Emissions from In-Use General Aviation Aircraft EEPS measures total particle number, which ensures that black carbon particles are weighted to a larger size due to coating by volatiles. Given that the uncertainty in the EEPS counting efficiency increases at larger particle size, and the larger particles contribute most to mass, comparison to the MAAP is, at best, approximate. PM Size Is Small, < 20nm, for Piston Engines Piston engines emit particles as lognormal particle size distributions (PSD). A TSI engine exhaust particle sizer (EEPS) was used to measure a PSD at 1Hz so that all engine operating conditions could be observed as well as the transition between operating points. The difference between ambient and engine PSDs is obvious as shown in Figure 5-10, where the ambient signature is shown on the left and the engine on the right. The yellow line in the left-hand plot represents the lower bound of instrument detection (“zero”) and the red line on the right-hand plot represents the upper bound (“saturation”). Besides the relatively random size distribution, the ambient concentration on the left is < 4E3 cm-3, whereas the engine emission of primary particles is sig- nificantly larger (>1E7 cm-3) and lognormal. Figure 5-9. Comparison of total and non-volatile particulate matter number (tPMn vs. nvPMn). Figure 5-10. Typical ambient (left) and piston engine (right) particle size distributions.

Other Parameters Affecting Emissions 49 For almost all of the engines in the measurement database from this project the emitted par- ticles were observed to have a geometric mean diameter (GMD) that is small (<20 nm); there- fore, most of particles are in the nucleation mode (i.e., newly formed particles). Particles in this size range were found to dominate the emitted PSD at all engine operating conditions. However, during transitions from low to high or high to low power, a soot mode (20 – 100 nm) was often observed (see Figure 5-11). Because the soot mode contributes to the engine PM mass emis- sion more significantly than the nucleation mode because of particle size, the presence of the soot mode was also indicating lower overall combustion efficiency. In that respect, the overall width of the PSD, even in the nucleation mode, was reflecting the ratio of fuel-to-air mixture the engine was burning. Although the GMD is somewhat invariant vs. engine power, the particle size concentration increases as engine power increases as shown in Figure 5-12. The number concentration data in this figure is not corrected for dilution in the plume, but wind conditions were steady during this test series. This response of increasing PM emissions with increasing power was expected Figure 5-11. Typical engine PSD at a stable operating power (left) and the PSD during the transition to another stable point (right). Figure 5-12. Typical variation in the particle concentration distribution vs. engine operating condition (blue 5 total particle concentration, red 5 non-volatile particles only).

50 Exhaust Emissions from In-Use General Aviation Aircraft because the combustion process was the same for each stroke of the piston engine, unlike a gas turbine where the fuel air mixing is modulated vs. engine power. The piston engine increases power by simply increasing the frequency of the piston movement. The red curve in Figure 5-12 is a measurement of non-volatile particles only as made by the APC. To use the EEPS (total particle number) and the APC (non-volatile only) data to assess the partitioning of solid and volatile particles, the APC must be corrected for sample conditioning losses from the catalytic stripper and other losses. These losses were about a factor of 3 and were applied to the APC data in the plot. Differences between the two curves represent the volatile particle concentration, which is seen to be significant. The large fraction of volatile particles is due to the lower combustion efficiency of the piston engine compared to a gas turbine and sup- ports richer operating conditions. The lower efficiency is supported by the higher levels of carbon monoxide and unburned hydrocarbons previously noted. Another contributor of particles to the nucleation mode could come from lead in the AVGAS used in these engines. The lead is emitted as lead bromide (PbBr2) in the combustion process of GA piston engines. PbBr2 is a volatile species that can contribute to the formation of new particles in engine exhaust. GA Turbofan Engines The research team measured PM emissions in mass and number as well as particle size distri- butions from engine exhaust plumes from two gas turbine aircraft engines: a CF34-3A1 turbofan engine made by GE Aviation and a TPE331-6-252B turboprop engine, initially developed by Garrett AiResearch and made by Honeywell at present. The research team observed both nucleation and soot particles from the EEPS measurements for the CF34-3A1 engine from engine idle to take-off (see Figure 5-13). Soot mode is dominant Figure 5-13. Particle size distributions for the CF34-3A1 turbofan engine. Individual test points are shown in different colors.

Other Parameters Affecting Emissions 51 Figure 5-14. Particle size distributions for the TPE331-6-252B turboprop engine. Individual test points are shown in different colors. at high-power condition, while nucleation mode becomes more important in number count at low power. However for the TPE331-6-252B engine, only one mode around 35 nm was observed, as demonstrated in Figure 5-14. Nucleation and mode become indistinguishable in the engine exhausts from the TPE331-6-252B turboprop engine. The non-volatile and volatile PM compo- sitions are probably internally mixed to generate an individual particle mode. PM emissions indices of the TPE331-6-252B engine are larger than those of the CF34-3A1 engine due to the lower temperature of its engine exhausts. For both gas turbine engines, medium power conditions (40-60% of thrust) yield the lowest PM emissions in number and mass. PM emission indices for the CF34-3A1 engine are shown in Figure 5-15. This observation of low emissions at cruise condition is in agreement with previous field measurements on com- mercial aircraft engines and implies that gas turbine aircraft engine performance is optimized at the cruise condition, which consumes the most fuel and provides the best energy efficiency.

52 Exhaust Emissions from In-Use General Aviation Aircraft Figure 5-15. Emission indices of nvPM in number and mass for the CF34 turbofan engine.

53 C H A P T E R 6 In this report, emission indices are listed for 47 full engine tests. A thorough analysis of trends and variability in these EIs is presented, with emphasis on the statistical comparison of the research team’s results with existing data. A sensitivity analysis shows how substituting experi- mentally determined EIs and fuel flows into EDMS/AEDT leads to differences in reported air- port emissions. Parameters affecting emissions are investigated and discussed. The inherent variability in piston engine emissions is quantified and explored. This report achieves all three major goals of ACRP Project 02-54: (1) Verify sample data sets that exist: a. Replicate measurements of several tested engines were used to perform a statistical vali- dation of existing data. b. Given the large degree of inherent variability in piston engines, most of the existing data was validated, even if it differed a lot from results from this research. c. Several invalid data points were found. The most important of these data points is the 2.3-times underestimation of the hydrocarbon emissions data for the very common Lycoming O-320 engine by the FAA-mandated software used for calculating airport emissions (EDMS/AEDT). d. The research team recommends that the Lycoming O-320 engine data in EDMS/AEDT be substituted with engine family average results from this research. This data is provided at the beginning of the Emission Index Data Tables in Appendix P. (2) Supplement the most commonly used aircraft engine data that does not exist in EDMS/ AEDT or other emission databases: a. Forty-seven unique engines were fully sampled in all engine states as a part of this research. This included coverage of 10 engines from a list of the top 20 national piston engines. Existing EDMS/AEDT databases include only eight piston engines. b. Ten new engine families that are not included in EDMS/AEDT databases were measured. Many different subtypes of engines were also measured. c. The assumption underlying this goal of supplementation is that one engine type yields one set of well-defined emission indices. This assumption is not valid due to the flexible way in which piston engines are operated and the resulting variability in their emissions. d. Repeat measurements are particularly important given this variability. All of the emis- sion indices collected as part of this project thus serve to supplement existing data and construct a database of piston engine emissions measurements. (3) Develop recommendations for determining substitution for aircraft not in existing emission databases: a. As part of the sensitivity analysis, a method for choosing an engine substitution from within the current bounds of the EDMS/AEDT software has been described and dis- played in flowchart form. Conclusions

54 Exhaust Emissions from In-Use General Aviation Aircraft b. The assumption underlying this goal of substitution is that one engine type yields one set of well-defined emission indices. This assumption is not valid, given the large variability in piston engine emissions. c. The two sensitivity analyses were performed using different methods; both use the substi- tution method described in Item a. The research team propose future research on deter- mining the best way to treat variability for GA airports, including substitution methods. GA emissions, in particular those from piston engines, present a significant challenge to air- ports and others wanting to perform inventory and air quality calculations. An understanding of the observed trends in GA emissions, combined with a characterization of the confidence intervals inherent to any calculated emissions estimate, will enable airports and policymakers to make deci- sions based on sound science and an understanding of the real-world operation of GA aircraft. Future Research Suggested topics for future research follow. • Research the best way to include the effects of variability in GA airport inventories. The concepts of variability and 95% confidence intervals are crucial for GA airports. A confidence interval consists of an upper limit and a lower limit such that one is 95% sure that the true aver- age emission falls between them. Variable data have wide confidence intervals. A GA airport inventory should have confidence intervals that reflect the range of possible emissions, given the inherent variability of its fleet’s emissions. Two possible ways to include this variability have been explored here. Monte Carlo methods that use random sampling show significant promise over standard methods using FAA-mandated tools. Other methods could be investigated. One such method might group all piston engine aircraft together and assign a single set of representative emission factors and confidence limits for the whole piston fleet. Broad horsepower subcategories could also be considered. This method could simplify airport emissions calculations by reducing the number of individual aircraft types chosen as part of a sensitivity analysis. • High-volume automated measurements for improved GA airport inventories. Each method that the research team explored to deal with aircraft variability relies on a large number of air- craft measurements. Although we have 47 full engine tests, including hundreds of individual test points, this may not be enough data to fully quantify distributions of piston engine emitters, par- ticularly those relatively rare high-emitting points. High-volume automated measurements of hundreds of aircraft would be an ideal way to expand this dataset. An automated measurement system could be set up at an airport, downwind of a taxi area and a runway. Data during normal airport operations could be collected for a matter of months, and then analyzed. The distribu- tion of piston engine emitters would then be well defined and could be used to determine rep- resentative emission values and confidence intervals. This measurement topic is complementary to the three topics about fleet characteristics, fleet use, and representative times-in-mode. • Fleet characteristics of a representative GA airport. Significant work was done in this research to construct a hypothetical GA airport that is representative of a U.S. national fleet. However, limitations in the FAA Tail Registry Database hampered this effort, particularly for small GA jets. Further research in this area would include surveys of fleet characteristics and number of operations at GA airports across the country. • Fleet use at a representative GA airport. Even with accurate knowledge of the fleet charac- teristics of GA airports, it is still important to understand how that fleet is used day-to-day. For example, flight school aircraft are expected to account for a disproportionate number of operations compared with based aircraft. How does the list of most-used aircraft differ from the list of most common aircraft constructed from the FAA tail number registry? Knowing the characteristics of the in-use fleet can help focus future measurements on the aircraft of highest importance.

Conclusions 55 • Researching real airport operations to determine times-in-mode that are more represen- tative of true GA operations. Throughout this research, the default values for time spent in taxi (taxi and idle are often clumped together) were used. However, these default times were designed for commercial airports with significantly more traffic than at many GA airports. Taxi/ idle times in particular are expected to be shorter than the EDMS/AEDT defaults of 26 minutes total. Research on the real times-in-mode of a subset of GA airports would improve the accuracy of airport emissions calculations using these newly developed emission indices. • Engine leaning practices. The fuel-to-air ratio in piston engines has a significant effect on the resulting emissions. This fuel-to-air ratio is dictated by a propeller plane’s mixer setting and pilot preference for a “rich” mixture (excess fuel) or a “lean” mixture (less excess fuel). What propor- tion of pilots routinely run full-rich in all engine states except cruise? What causes this preference? Full-rich operation significantly increases the emissions of both CO and HC (but decreases NOx). • Fuel additive use and impact. During the field measurements, aircraft exhaust and fuel sam- ples unusually high in toluene were observed. The presence of toluene significantly affects the hydrocarbon emission signatures of these aircraft. The most likely source of toluene was a fuel additive designed to reduce spark plug fouling. Further investigation of the actual use rates of fuel additives and the point at which they were added to the fuel is needed to pin down the extent of this activity. • Realistic partitioning of emissions in EDMS/AEDT. The emissions software programs EDMS and AEDT output not only results for the main emissions species (i.e., HC, CO, NOx, and PM), but also partition those species into different classes. For example, HC emissions are broken into the categories of non-methane hydrocarbons (NMHC), VOC, and total organic carbon (TOG). PM emissions are broken into size categories of PM10 and PM2.5. The auxiliary data collected during this research project could be used to verify and improve these parti- tions. This could have a significant effect on airports because certain subcategories of HC and PM emissions are of more concern to human health than others. Policy Implications of This Research One potential policy implication relates to the flexibility in piston engine operation. Hydro- carbon and carbon monoxide emission factors are highest in idle and taxi, the two power states that occur on the ground, and decrease precipitously with leaner fuel mixtures. A policy encour- aging pilots to run leaner could be researched, particularly during taxi and idle where acci- dentally stalling the engine poses no safety issue. Such a policy could reduce airport emissions of hydrocarbons and carbon monoxide, but comes with a potential increase in NOx emissions. A second policy implication relates to the large inherent variability observed in piston engine emissions. This variability must be taken into account to perform realistic assessments of an airport’s emissions. Any GA airport’s emissions will have upper confidence limits many times higher than the average, depending on the pollutant. Monte Carlo methods have great potential to shrink these confidence limits when combined with the large number of operations at an airport over the course of a week (or year). These methods depend on having access to large datasets of emissions and times in mode that are representative of the GA airport being simulated. The research team’s recommendation for dealing with this variability is to push for high-volume measurements of piston engine emissions, coupled with advanced statistical methods. Such high- volume measurements could be done simply, with unattended automated systems installed at airports, and without impact on operations. In the interim, GA airport environmental impact statements should be produced with confidence intervals, even if they are very wide. This type of result gives airport managers and policymakers the power to make informed decisions based on the true weight of evidence.

56 A P P E N D I X A The prioritized list of engines below was used in the planning stages of the research to prioritize the engines measured. The rank of the aircraft engine is based on the FAA registry of the national GA fleet for piston engines and turbofan engines in 2014. Those rows highlighted in yellow show engines that were measured experimentally in all engine states. Engine Prioritization List Category MEP SEP & MEP MEP MEP SEP SEP SEP SEP & MEP MEP SEP SEP & MEP SEP SEP & MEP SEP SEP & MEP MEP SEP MEP SEP & MEP MEP SEP SETP SETP SETP SETP SETP SETP SETP SETP SETP SETP SETP SETP SETP SETP Engine Make Engine Family RANK CONT MOTOR TSIO 520 SERIES 1 LYCOMING O 320 SERIES 1 CONT MOTOR TSIO 360 SERIES 2 CONT MOTOR IO 470 SERIES 3 CONT MOTOR O 200 SERIES 3 LYCOMING O&VO 360 SERIES 4 CONT MOTOR O 470 SERIES 5 LYCOMING TIO 540 SERIES 5 LYCOMING IO 320 SERIES 6 CONT MOTOR O 300 SERIES 6 LYCOMING IO 360 SERIES 7 CONT MOTOR A&C65 SERIES 8 CONT MOTOR IO 550 SERIES 12 LYCOMING O 235 SERIES 13 LYCOMING IO 540 SERIES 14 CONT MOTOR GTSIO 520 SERIES 15 CONT MOTOR C145 SERIES 17 P & W R 985 SERIES 18 CONT MOTOR IO 520 SERIES 18 LYCOMING O 540 SERIES 19 CONT MOTOR C85 SERIES 19 P & W PT6A 67 SERIES 1 P & W PT6A SERIES 2 P & W PT6A 66 SERIES 3 P & W PT6A SERIES 4 P & W PT6A 42 SERIES 5 P & W PT6A 114 6 P & W PT6 SERIES 8 P & W PT6A 34 10 P & W PT6A 6 SERIES 12 P & W PT6A 60A 14 P & W PT6A 64 16 P & W PT6A 140 17 P & W PT6A 60 SERIES 18 P & W R1340 SERIES 20 FOCA Data Exists EDMS Data Exists Full Engine Tests Paral Engine Tests yes yes yes 16 4 4 yes yes yes 1 1 1 1 6 yes yes yes yes yes yes 4 4 yes 3 3 yes 1 1 yes 1 yes 1 2

57 A P P E N D I X B The test matrix reproduced here was used by the cockpit observer to direct the engine tests and note all relevant conditions. Test Matrix

58 Exhaust Emissions from In-Use General Aviation Aircraft Date Time (Local) Pilot Time (UTC) Tail Number Max HP Aircraft Make Max Propellor RPM Notes Aircraft Model Engine Hrs No. engines Direct Drive _______ Variable Pitch ___________ Engine Make Run-Up Times/Settings Engine Model Fuel/Additives Nominal Condition Time (Local) % of Max. Power Cockpit Notes Run Up Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Idle Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Taxi Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Idle Approach Idle Cruise Engine Parameters (enter n/a if required)

Test Matrix 59 Approach Idle Climb Out Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Approach Idle Take off Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Approach Idle Full Power Propellor RPM Throttle/Manifold Pressure or % Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Approach Idle Final Propellor RPM Throttle/Manifold Pressure or % Approach Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Idle Other Propellor RPM Throttle/Manifold Pressure or % (optional) Engine RPM Fuel Mixture % of full rich Oil Temp Engine Cyl. Head T/ EGT Oil Pressure Fuel Flow Air/Fuel Ratio Idle Run Down or Taxi Away

60 The ICAO maintains a database of turbofan engine emission indices, with the operational states defined based on a percentage of available thrust. The Swiss Federal Office of Civil Aviation (FOCA) maintains a database of piston engine emission indices, with the engine states defined by percentage of maximum propeller horsepower, inferred based on fuel flow measurements. ICAO vs. FOCA Databases A P P E N D I X C Engine State based on Landing Take-Off cycle (LTO) ICAO definion % available thrust (turbofan) FOCA definion % max propeller horse power Take-off 100 100 Climb 85 85 Cruise 65 Approach 30 45 Taxi 7 Operator’s manual - Table C-1. Comparison of operational states for turbojets (ICAO) and piston-powered propeller aircraft (FOCA).

61 Method for Calculating Emission Ratios Emission ratios are the first step in calculating emission indices from time series data. This appendix details methods of determining emission ratios. A time series is the measured concentration of a species of interest plotted as a function of time. Figure D-1 shows a selection of the many time series measured during the ACRP Project 02-54 field campaigns. The colored brackets at the bottom of the graph indicate the engine states during this test (red = idle, cyan = T/O, etc.). In this graph, many different species of interest are plotted, including combustion products (e.g., CO2 and CO) and speciated hydrocarbons and aromatics (e.g., methane (CH4) and toluene). The standard method for determining the emission ratio uses the plot of the species of interest versus total carbon (e.g., NOx vs. Total C). The time offset between the two traces is optimized and a linear fit of the data is taken. The slope (m) of this fit gives the emission ratio, while the coefficient of correlation (R2) gives an indicator of data quality. An example of this type of analysis is shown for the plume in the middle of Figure D-1, with the workup summarized in Figure D-2. Appendix E outlines the calculation of the emission index from an emission ratio. An improved algorithm for determining emission ratios has been developed to deal with non- ideal experimental data. Previously, some tests points were simply thrown out due to a poor R2. In the advanced algorithm, clean background periods are manually defined and used in determining individual emission ratios. Figure D-3 shows an example of two improved emission ratio determination methods. Figure D-3 (A) shows concentrations of a select few species of interest versus time for the “climb-out” state of the N62480 engine. Although the time traces for CO and total hydro- carbons (THC) match the time trace for total carbon (Total C), the traces for NOx and the PM mass signal (MAAP) are not as well correlated. Examining Figure D-3 (C) shows that a simple correlation of the raw data (blue circles, red fit line) yields a ratio of 4.35 ppb NOx/ppm Total C, but with a coefficient of correlation of 0.11, far below the data quality threshold of 0.75. This data point would previously have been rejected due to poor correlation. Figure D-3 (B) shows how periods of clean background are appended to either side of the time series of interest. The emission ratio is then computed via one of two methods: the cor- rected slope method or the corrected area method. The standard method is also shown for com- parison (blue circles, red fit line in Panel C). The corrected slope method [Figure D-3(C), cyan crosses and cyan fit line] yields a slope (m) of 6.6 ppb NOx/ppm Total C, and an R 2 of 0.49, still below the data quality threshold. For some data, this corrected slope method improves the fit sufficiently. The corrected area method [Figure D-3 (B), filled areas] yields a ratio of 7.5 NOx/Total C and will be the ratio used in the final determination of the emission index for this data point. A P P E N D I X D