Enhanced Modeling of Aircraft Taxiway Noise, Volume 1: Scoping (2009)

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.

D-1 Appendix D EDMS Model ing of Airside Operat ions D.1 . In t roduct ion EDMSD1,D2 is a combined emissions and dispersion model for assessing air quality at airports. The model was developed by the Federal Aviation Administration (FAA) in cooperation with the United States Air Force (USAF). The model is used to produce an inventory of emissions generated by sources on and around the airport or air base, and to calculate pollutant concentrations in these environments. Within EDMS, taxiway emissions are modeled for on-ground, non-runway operations of an aircraft. These operations are discretely defined along with the other modes as indicated below: • Taxi Out (& Idle) • Takeoff • Climb Out • Approach • Taxi In (& Idle) Included within the taxiing operations are idling times where the aircraft is not moving but the engines are still on. The EDMS user is required to specify the number of operations for these modes. This can be accomplished in one of two ways. The Landing and Takeoff (LTO) cycles can be specified as a whole for each aircraft-engine in which case the number of departures and arrivals are identical (i.e., LTO cycles = departures = arrivals), or the departures and arrivals can be specified separately (i.e., number of departures and arrivals are different). Often, quarter-hour profiles are used to distribute the operations across the 96 quarter hours in a day, the 7 days in a week, and the 12 months in a year. This appendix outlines a set of methods used to model aircraft operations in EDMS 5.x. Since aircraft engine emissions depend on the thrust setting and runtime, accurate inventory estimation requires modeling of both. EDMS has two options to model times-in-mode of aircraft operations. Moreover, EDMS has two modes of operations: inventory and dispersion, with significantly different data requirements. This appendix aims to briefly describe modeling options and data requirements without going into specifics of the supporting algorithms. There are two basic modes in modeling of airport operations: Emissions Inventory and Dispersion modeling Samples studies are: dispersion modeling for elevated mobile sources are provided in reference D.13 and an estimate of aircraft emissions for future traffic scenarios may be found in reference D.14. Both modes are further elaborated below. D.2 . Emiss ions Inventory An emissions inventory is a report on cumulative assessment of pollutants generated by all active emission sources included in the study (e.g., aircraft, APU, GSE). To perform an emissions inventory, the user needs to identify emission sources and the annual activity for each of the sources. Moreover, emission factors are required if the user opts to create user-defined sources (e.g., aircraft, APU, GSE). EDMS calculates the total annual pollutant emissions for each of the identified sources and presents it in both a summarized report and a detailed report. The following pollutants are consideredD1:

D-2 • CO2 (carbon dioxide) for aircraft only, • CO (carbon monoxide), • THC (total hydrocarbons) for aircraft and APUs only, • NMHC (non-methane hydrocarbons), • VOC (volatile organic compounds), • TOG (total organic compounds), • NOx (nitrogen oxides), • SOx (sulfur oxides), • PM-10 (particulate matter, 10 microns) • PM-2.5 (particulate matter, 2.5 microns), and • 395 Speciated Hydrocarbons (44 HAPs, and 351 non-toxic compounds). As it could be noted from the description above, if EDMS is being used to run an emissions only analysis, the user is not required to provide neither detailed layout of the airport nor detailed weather data. Thus, in this mode EDMS does not model individual aircraft movements. To calculate the total annual pollutant emissions EDMS uses a (pseudo)schedule and input information on the amount of time each aircraft spends in each mode of operation (portion of a landing and takeoff (LTO) cycle). EDMS divides LTO cycles into six phases: approach, taxi-in, startup, taxi-out, takeoff, and climb-out. Out of these, approach, takeoff and climb-out are airborne phases. The landing roll part of the approach segment is incorporated into the taxi-in time. The six modes of operation involved in the LTO cycle can be described as followsD1: 1. Approach: The airborne segment of an aircraft’s arrival extending from the start of the flight profile (or the mixing height, whichever is lower) to touchdown on the runway. 2. Taxi In: The landing ground roll segment (from touchdown to the runway exit) of an arriving aircraft, including reverse thrust, and the taxiing from the runway exit to a gate. 3. Startup: Aircraft main engine startup occurs at the gate. This methodology is only applied to aircraft with ICAO certified engines. All other aircraft will not have startup emissions. Aircraft main engine startup produces only HC, VOC, NMHC, and TOG emissions. A detailed speciated hydrocarbons profile does not exist for main engine startup emissions. 4. Taxi Out: The taxiing from the gate to a runway end. 5. Takeoff: The portion from the start of the ground roll on the runway, through wheels off, and the airborne portion of the ascent up to cutback during which the aircraft operates at maximum thrust. 6. Climb Out: The portion from engine cutback to the end of the flight profile (or the mixing height, whichever is lower). Generally, there are two options for determining times in mode for LTO cycle: • Performance Based (SAE AIR 1845)D12 and • ICAO/USEPA DefaultD3 Performance based modeling uses the specific airframe and engine characteristics along with weather data to model each flight dynamically resulting in non-constant times-in-mode. ICAO/USEPA defaults are standardized values read from a table (Table D.1). The user can modify these times if necessary

D-3 through input dialogs for all aircraft or for each study aircraft. Within EDMS, a study aircraft refers to a specific group of aircraft operations performed by a single aircraft-engine combination. These operations have a set of common characteristics such as: aircraft performance and engine characteristics, are serviced by the same set of GSEs, have the same APU, etc. Table D-1. ICAO Time in mode for landing and take-off (LTO) cycle Phase of flight Operation Time [min] Take-off 0.7 Climb-out 2.2 Approach 4.0 Taxi-out 19.0 Taxi-in 7.0 Sample studies have used performance based times-in-modeD4,D7 and using both methodsD6 to obtain times in mode. A sample study in which the user modified time-in-mode data (ICAO/USEPA recommendations for taxi-in and taxi-out times are modified) is outlined in Reference D.5. For emissions inventory development, just the total time attributed to idling and taxiing operations can be used. The EDMS user can supply this information from various sources including flight schedules and tower logs. The following defaults from ICAOD3 can also be used: • Taxi Out: 19 min • Taxi In: 7 min Taxi emissions are modeled using engine-specific emissions indices and fuel flow rates corresponding to the lowest (7%) of the standard power settings from the ICAO emissions databank: • Idle (& taxi): 7% • Approach: 30% • Climb Out: 85% • Takeoff: 100% D.3 . D ispers ion Model ing EDMS dispersion modeling requires knowledge of both when and where emissions took place. Thus, when modeling aircraft operations, the user is required to use: • performance based aircraft modeling of airborne operations (SAE AIR 1845) and • Detailed modeling of aircraft surface movements (queuing/sequencing). For dispersion modeling, EDMS uses the following third party components: • AERMOD dispersion model • Two AERMOD processors: • AERMET – meteorology • AERMAP – terrain In 2000. the US Environmental Protection Agency (EPA) developed AERMODD8,D9,D10,D11 as the newest generation of short-range steady-state atmospheric dispersion models. Although it was not originally intended to model elevated mobile sources, AERMOD has been integrated into EDMSD4 since 2001. AERMOD is a plume model that is used for modeling concentrations of pollutants stemming from various sources which may be represented as ideal point, area or volume sources.

D-4 EDMS can calculate hourly emissions, and generate AERMOD input files for the following pollutants: • CO (carbon monoxide), • THC (total hydrocarbons) for aircraft and APUs only, • NMHC (non-methane hydrocarbons), • VOC (volatile organic compounds), • TOG (total organic compounds), • NOx (nitrogen oxides), • SOx (sulfur oxides), • PM-10 (particulate matter, 10 microns), and • PM-2.5 (particulate matter, 2.5 microns). The amount of data required to perform a dispersion analysis is significantly greater than the data necessary for just an emissions inventory. The additional information required for a successful dispersion analysis includes: 1. Detailed (pseudo-)schedule, 2. Aircraft performance modeling (SAE AIR 1845), 3. Airside delay and sequencing modeling, 4. Hourly weather data, 5. Placement of receptors. An aircraft delay and sequencing modeling requires: • Detailed airport layout, and • A set of airport configurations and activation method. Airside delay and sequencing model data flowD2 is depicted in Figure D-1. Figure D-1. Airside delay and sequencing model data flow Q ueuing Engine (Pseudo-)Schedule Configurations Taxi paths WWLMINET airside delay model (2) Sequencing model Airport hourly level of detail operations/delays Gates Taxiways Runways Weather Airside network model AERMET Operation level of detail (position and time)

D-5 An emissions inventory must first be generated before dispersion can be performed, since the set of emissions that are dispersed is the same as that produced from the annual inventory. More details on algorithms implemented in EDMS can be found in Reference D2. As is outlined above, the EDMS dispersion modeling requires a large quantity of data per airport. Thus, it is suitable for local studies (a single or a few adjacent airports). EDMS is not designed for global analysis. However, the user may elect to use EDMS to calculate emissions in many airports. In order to complete a multi airport project with EDMS, some simplification of the input data may be required to reduce the modeling and run time. The System for Assessing Aviation’s Global Emissions (SAGE)D.15 is a tool aimed to predict aircraft fuel burn and emissions for commercial flights globally. D.4 . De lay and Sequence Model ing The aircraft delay and sequencing models embedded in EDMS are designed to process five years of traffic/weather data, and calculate the aircraft taxi times in a reasonable amount of time. As a part of future improvements, for the case where the user would require a more detailed simulation of airport operations, the outputs of a more detailed simulation models (e.g., TAAM/SIMMOD) could be imported into EDMS and AEDT to provide even more realistic times-in-mode. The Queuing/Sequencing model requires the user to input an airport layout. However, the level of details entered is solely the user’s choice. For example, the user may opt to approximate a taxiway with a straight line, or neglect less used runways, etc. Using capacity information, delay modeling is conducted through the use of WWLMINET, a queuing model developed by the Logistics Management Institute (LMI). The modeled delays are used to adjust the push-back time for departures and the touchdown times for arrivals. Queues are also formed as part of this delay modeling to realistically account for the line of aircraft waiting along taxiways to use a runway. Although this is done for departures, arriving aircraft are assumed to have unimpeded taxi movements to the gate (no queue formation). Using a combination of the configuration information, the operational profiles, a delay module, and a sequencing module, detailed modeling of runway and taxi-way usage can be conducted taking into account the capacity at an airport. The user needs to specify one or more points (departures/hr versus arrivals/hr) along a Pareto frontier to define the capacity at the airport. The sequencing module models the movement of aircraft along taxiways including any queues that are formed. Along with the geometry, the user can provide an average movement speed specific to each taxiway (not specific to aircraft type). The default speed is 15 knots (17.26 mph). This allows the determination of how long each aircraft spent on each taxiway segment for proper calculation and allocation of emissions for dispersion modeling. In addition, aircraft-specific (or aircraft category- specific) dispersion parameters can be properly applied to the taxiway segment. EDMS contains a sequencing module that utilizes a delay model (WWLMINET) which predicts the formation of queues on taxiways. To engage the sequencing module, the user would have to provide the following: 1. Airport capacity data (essentially through departure per hour versus arrival per hour points along a Pareto Frontier) 2. Airport operations / LTOs 3. Taxiway geometry. 4. Speeds for each taxiway segment. Using this data, the sequencing module predicts the time-in-mode for each taxiway segment which is then used to calculate emissions. At present these calculations are done internally within EDMS so

D-6 that the user only sees the final results (emissions). For the prediction of taxi noise, CSSI could potentially output the TIM values as necessary. For integration purposes within AEDT, the TIM values for each segment would be available. For emissions it is just the total time spent within each taxiway segment that enters into the calculation since results doesn't depend on whether the aircraft is sitting idle or moving on a taxiway. Internally both stationary and moving conditions are modeled using the same fuel flow and emissions indices at 7% power. But the code could be made to distinguish between idle time and movement time for each aircraft. At present, movement times are based on using the user-supplied constant speed values for each segment. No acceleration or deceleration movements are modeled. D.5 . A i rpor t (A i rs ide) De lay Model EDMS models airside operations in two steps: 1. using WWLMINET, which determines airport throughput 2. using Sequencing model, which determines for each operation actual times of reaching significant points on the airside network. The airside network is an artificial representation of the airport layout. While modeling aircraft movements, the EDMS determines the active runway configuration that is used at each hour of the year based on meteorological information and the user-specified activation parameters in order to determine the associated airport capacity at each hour of the year. This airport capacity information along with demand information from the aircraft operational profiles or schedule are provided to the WWLMINET delay model to determine the airport throughput. Additionally, the EDMS’s sequencing module adjusts the estimated gate push-back time (for departures) and estimated touchdown time (for arrivals) into actual times that are possibly delayed. The sequencing module further models the movements of aircraft along the taxiways (or taxipaths) between runways and gates for both arriving and departing aircraft. WWLMINET and its purpose in EDMS are described in the following paragraphs. WWLMINET is the airport airside queuing model developed by LMID16, D17. It models airside queuing by using a queuing network shown in Figure D-2. Figure D-2. Airside queuing in WWLMINET Figure D-2 depicts two queuing processes: one for arrivals and another for departures. The arrival and departure processes are dependent. Thus, a departure may be released only if there is an available aircraft in the reservoir (R). The purpose of the reservoir is to balance the total number of arrivals and departures over time. Arriving aircraft enter the arrival queue as Poisson process with parameter λa(t). After being processed by the arrival server, an arriving aircraft enters the taxi-in queue. Upon arrival τ R λa(t) λd(t) taxi-in taxi-out departur e µa(t) µta(t) µd(t) µtd(t)

D-7 processing by the taxi-in server, arriving aircraft are delayed for a service time (τ) and released into the reservoir (R). Departing aircraft are processed by two servers as well. The departure process is driven by: a Poisson process with parameter λd(t), and the state of the reservoir (R). After being processed by the taxi-out server, departure aircraft enter the departure queue and, after processing, are released. Arrival and departure servers may be modeled as M/M/1 or M/Ek/1 queues. Taxi times (both taxi-in and taxi-out) may be modeled as M/M/1 queues only. Arrival λa(t) and departure λd(t) demands are determined directly from the (pseudo)schedule. The arrival and departure service rates are determined by taking into account the appropriate airport capacity Pareto frontier (Figure D-3) based on the hourly surface weather observations. Figure D-3. Capacity Pareto frontier The WWLMINET is used to determine, for each operation, a time bin when the operation is released as well as an average delay time acquired during each time bin. D.6 . Equipment Model ing Aircraft types are specified through a comprehensive list of equipment names developed as part of the EDMS system databases. These aircraft are then assigned internally to the smaller set of actual performance models. Similarly, the selected engines are internally assigned to a Unique Identification Number (UID) from the International Civil Aviation Organization’s (ICAO’s) emissions databank. D.7 . Model ing of Tax i Paths EDMS 5.x provides two options of modeling aircraft taxiing operations: • User-specified taxi times for each aircraft • Detailed modeling of aircraft movements along taxipaths (Queuing/Sequencing) The user-specified taxi times option may be used to calculate an annual emissions inventory. This option may not be used for dispersion analysis, thus, the user is not required to provide an airport layout. The detailed modeling of aircraft movements option may be used to calculate an annual emissions inventory and for dispersion modeling (using AERMODD.8, D.9, D.10). Thus, in addition to the aircraft schedule, the user is required to provide the following: arrivals de pa rtu re s

D-8 • Detailed airport layout (gates, taxiways, runways, etc.) • A set of taxi paths1 connecting gates to runways and runway exits to gates • Airport configurations • Hourly weather • Location of receptors Each operation (departure or arrival) is characterized, among the others, with its assigned gate and its expected (scheduled) operation time. Based on the weather input file, EDMS identifies the most appropriate airport configuration for each hour, which further identifies: 1) the airport runway capacity and 2) the aircraft runway distribution (based on the aircraft weight class). EDMS allows users to identify one (1) taxi path for each gate-runway pair (departures) and runway- exit-gate (arrivals). Therefore, each departure gets assigned a unique taxi path when the appropriate gate-runway pair is identified. Also, each arrival operation gets assigned a unique taxi path when the appropriate runway-exit-gate pair is identified. EDMS then models the movements of individual aircraft along the taxi paths. EDMS requires the user to identify: • All gates (as polygons) • All taxi ways (each as a sequence of straight lines) • All runways • All taxi paths For the dispersion modeling, it is important to allocate emissions along the taxi paths spatially and temporally. Thus, it is necessary to model the aircraft movements in detail. Moreover, because EDMS uses AERMOD for dispersion modeling which is not designed to model elevated mobile sources, a taxiway network (a set of gates, taxiways, and runways) needs to be split into a set of emission (area) sources. Emissions from a given arrival/departure operation are distributed within an individual source proportional to the amount of time spent in that source, which is a method for aggregation of aircraft operations. For atmospheric dispersion modeling, both the spatial and temporal characteristics of emissions need to be identified. Therefore, the user supplied total taxi times cannot be used, and the full pathway from the terminal to the runway needs to be specified. In order to do this, the user must define the gates, taxiways, taxi paths, and runways (including configurations). A gate represents the location where an aircraft is parked at the terminal and also serves as the emissions source location for various ground service equipment (GSE) and auxiliary power units (APU). Because of all of these emissions, gate locations are generally specified as level polygonal areas (several contiguous XYZ values) although a single point can also be specified (in which case, a volume source is internally created by EDMS). A taxiway is defined as one or more segments with XYZ points at each end that define part or all of a taxi path that connects a gate to a runway end. Typically, several taxiways are specified to allow different combinations to be used in creating taxi paths. Hence, a taxi path is defined by one gate, one runway end and one or more taxiways. Taxi paths are also differentiated between outbound (gate to runway) and inbound (runway to gate) usage. Runways are specified by providing the XYZ coordinates of each end. Runway configurations allow the user to specify weather conditions (e.g., wind direction) and times under which the runway assignments are made using distributions based on aircraft size categories (e.g., small, large, etc.). 1 EDMS 5.xD.1 defines a taxi path as an ordered list of taxiways which connects a gate to a runway (outbound) or a runway exit to a gate (inbound)

D-9 D.8 . References D.1. FAA, Emissions and Dispersion Modeling System (EDMS) User’s Manual, 2008. D.2. FAA, Emissions and Dispersion Modeling System (EDMS) Technical Manual, 2008. D.3. ICAO, International Standards and Recommended Practice, Environmental Protection, Annex 16, Volume II – Aircraft Engine Emissions, 1993. D.4. Thrasher, T., Nguyen, A., Hall, C., Fleming, G., Roof, C., Balasubramanian, S., Grandi, F., Usdrowski, S., Dinges, E., Burleson, C., Maurice, L., Iovinelli, R., AEDT Global NOx Demonstration, 7th USA/Europe ATM R&D Seminar, Barcelona, Spain, 2007. D.5. Ohsfeldt, M., Thrasher, T., Waitz, I., Ratliff, G., Sequeira, C., Thompson, T., Graham, M., Cointin, R., Gillette, W., Gupta, M., Quantifying the Relationship between Air Traffic Management Inefficiency, Fuel Burn and Air Pollutant Emissions, 7th USA/Europe ATM R&D Seminar, Barcelona, Spain, 2007. D.6. Hall, C., Thrasher, T., Draper, J., A Validation of Aircraft Times in Mode in the Emissions and Dispersion Modeling System (EDMS) Version 4.2, Air and Waste Management Association 98th Annual Conference and Exhibition, June 2005. D.7. Hall, C., Mondoloni, S., Thrasher, T., Estimating the Impact of Reduced Thrust Takeoff on Annual NOx Emissions at Airports, Air and Waste Management Association 96th Annual Conference and Exhibition, June 2003. D.8. Cimorelli, A.J., Perry, S.G., Venkatram, A., Weil, J.C., Paine, R.J., Wilson, R.B., Lee, R.F., Peters, W.D., Brode R.W., Paumier, J.O., AERMOD: Description of Model Formulation. Prepared for Office of Air Quality Planning and Standards Emissions Monitoring and Analysis Division Research Triangle Park, North Carolina 2004. (EPA-454/R-03-004) D.9. Cimorelli, A.J., Perry, S.G., Venkatram, A., Weil, J.C., Paine, R.J., Wilson, R.B., Lee, R.F., Peters, W.D., Brode R.W., AERMOD: A dispersion model for industrial source applications. Part I: General model formulation and boundary layer characterization, Journal of applied meteorology, 2005, 44, (5), 682-693. D.10. Perry, S.G., Cimorelli, A.J., Paine, R.J, Brode R.W., Weil, J.C., Venkatram, A., Wilson, R.B., Lee, R.F., Peters, W.D., AERMOD: A dispersion model for industrial source applications. Part II: Model Performance against 17 Field Study Databases, Journal of applied meteorology, 2005, 44, (5), 694-708. D.11. Hanna S. R., Egan, B. A., Purdum, J., Wagler, J., Evaluation of the ADMS, AERMOD, and ISC3 dispersion models with the OPTEX, Duke Forest, Kincaid, Indianapolis and Lovett field datasets, International Journal of Environment and Pollution, 2001, 16, 301-314. D.12. SAE Committee A-21, Aircraft Noise, “Procedure for the Calculation of Aircraft Noise in the Vicinity of Airports,” SAE Aerospace Information Report SAE AIR 1845, Society of Automotive Engineers, 1986. D.13. Lucic, P., Hall, C., Thrasher, T., Iovinelli, R., Holsclaw, C., Peters, W., Dispersion Modeling for Aircraft: A Comparison of Area versus Volume Sources, Air and Waste Management Association 99th Annual Conference and Exhibition, June 2006. D.14. CSSI, JPDO-NextGen-TDM23: Investigation of Aviation Emissions, Air Quality Impacts, Draft Report, 2008. D.15. FAA, SAGE System for accessing Aviation’s Global Emissions, Version 1.5, Technical Manual, 2005.

D-10 D.16. Long, D., Lee, D., Johnson, J., Gaier, E., and Kostiuk, P., Modeling Air Traffic Management Technologies With a Queuing Network Model of the National Airspace System, NASA/CR-1999- 208988, Logistics Management Institute, McLean, Virginia, 1999. D.17. Stouffer, V., WWLMINET User Guide, Logistics Management Institute, McLean, Virginia, 2002.