MIT Lincoln Laboratory Partnership for AiR Transportation Noise and Emissions Reduction MIT International Center for Air Transportation Airport Characterization for the Adaptation of Surface Congestion Management Approaches* Hamsa Balakrishnan, Harshad Khadilkar, Lanie Sandberg and Tom G. Reynolds Massachusetts Institute of Technology MIT Lincoln Laboratory *This work is sponsored by the Federal Aviation Administration under Air Force Contract #FA8721-05-C-0002. Opinions, interpretations, recommendations and conclusions are those of the author and are not necessarily endorsed by the United States Government.
Outline Motivation Framework for adapting surface congestion management approaches Airport characterization Site visits Surface visualizations Operational data analysis Algorithm development Implementation design Testing and performance evaluation 2
Motivation: Scale of Problem Surface congestion increases taxi times, fuel burn and emissions Nationally (2012 ASPM) 31M min taxi-out delay; 15M min taxi-in delay LGA (2012 ASPM) 2M min taxi-out delay; 400K min taxi-in delay 19K tons of fuel, 60K tons CO 2, 239 tons NOx, 127 tons HC PHL (2012 ASPM) 1.2M min taxi-out delay; 351K min taxi-in delay 20K tons of fuel, 63K tons CO 2, 256 tons NOx, 150 tons HC BOS (2012 ASPM) 687K min taxi-out delay, 297K min taxi-in delay 13K tons of fuel, 41K tons CO 2, 164 tons NOx, 83 tons HC Potential to mitigate these impacts through surface congestion management 3
Role of Departure Metering in Surface Congestion Management Departure metering just one element of required surface management toolset Departure metering manages pushbacks during congested periods Decreased engines-on time, fuel burn & emissions In principle, can work at any congested airport, but details of successful implementation will vary e.g., ATC facility vs. airline ramp tower Demand on Surface Taxi Routing Departure Metering Runway Assignment Max efficiency limit 1 Possible Surface Management Tools Excess congestion 2 3 4 Time Interval Sequencing & Scheduling Departure Route Assurance Airport Configuration Excess flights held until later time intervals when they can be more efficiently accommodated [A. Nakahara, 2012] 4
Examples of Departure Metering Approaches Aggregation Level Examples Field tests Key Output Airport-level N-Control (Pushback Rate Control) BOS Aggregate airport pushback rate Runway-level Q-Control (TFDM prototype) DFW Runway-specific pushback rate Airline-level Collaborative Departure Queue Management MEM, MCO Airline-specific pushback quotas Ground Metering Program JFK Aircraft-specific pushback time Spot and Runway Departure Advisor (NASA) DFW HITL simulation Aircraft-specific spot release times Aircraft-level Airport Collaborative Decision Making (ACDM) AMS, CDG, FRA, HEL, LHR Aircraft-specific target start-up approval times (TSAT) Departure Manager ATH Aircraft-specific target start-up approval times (TSAT) 5
Motivation: Need for Adaptation Prior surface congestion management efforts focused on specific airports Need to adapt approaches to multiple airports with different characteristics to gain system-wide benefits BOS LGA PHL 6
Outline Motivation Framework for adapting surface congestion management approaches Airport characterization Site visits Surface visualizations Operational data analysis Algorithm development Implementation design Testing and performance evaluation 7
Framework for Adapting Surface Congestion Management Approaches Airport Selection Airport Characterization Site visits Visualizations Operational Data Analysis Algorithm Development Refinement/ Validation Implementation Design Operational Testing & Performance Evaluation Results 8
Outline Motivation Framework for adapting surface congestion management approaches Airport characterization Site visits Surface visualizations Operational data analysis Algorithm development Implementation design Testing and performance evaluation 9
Airport Characterization: Site Visits Gain understanding of airport characteristics Physical layout Equipment levels Air carrier and fleet mix Other factors that influence throughput First-hand observations of operations Standard procedures Current challenges Expert opinions from ATC professionals Explanation of operations Answering congestion management questions Identifying potential opportunities for mitigation 10
Sample Site Visit Observations: LGA Insights into: Physical tower layout ATC positions and relative locations Equipment availability Standard operating practices RACD ASDE-X DSP DSP IDS RACD ASDE-X GC1 Flight strip movement RACD Class B Airspace Control Sequencer (runway crossing) Cab coordinator RACD Clearance Delivery/ Flight Data Stairs TMC ETMS/ TSD Harmony DSP1 DSP2 TMA ITWS RAPT/ IDRP METAR 11
Sample Site Visit Observations: LGA Typical taxi routes & surface congestion issues Arrivals Departures Nominal departure taxi route via B and P Extended departure taxi route to queue aircraft during periods of high demand or with re-routes Queues observed to form short of taxiway GG (hand-off point between GCs) Nominal arrival taxi route: depart 22, taxi via B and A Single aircraft push-back fully blocks alley-way Single aircraft push-back can block arrival taxi route 12
Airport Characterization: Surface Visualizations Use airport surveillance data archives (e.g., ASDE-X) Allows detailed observations for a range of airport operating conditions beyond those seen on site visits Surface procedures across configurations Standard taxi routes Runway entry, exit and crossing locations Aircraft holding/queuing locations Dynamics of demand over extended time intervals At gate At terminal At runway Dynamics of interactions between arrivals and departures 13
Sample Surface Visualization: LGA 22 13 14
Sample Surface Visualization: LGA 22 13 Holding Area Arrival/Departure Interactions Standard Taxi Routes Departure Queues 15
Sample Surface Visualization: PHL 27R 27 L 16
Sample Surface Visualization: BOS 22L, 27 22R 17
Airport Characterization: Operational Data Analysis 22L, 27 22R, 22L Historical data from ASPM and ASDE-X Quantification of airport characteristics & performance Runway configuration breakdown Traffic demand Queue sizes Taxi time Airline mix BOS Runway Configuration Usage; 6/1/11-8/31/11 33L 27 22L, 22R 15R 27 33L 47% 47% 4R, 4L 9, 4R Number of Aircraft/Time (mins) 40 35 30 25 20 15 10 5 0 BOS Surface Metrics (22L,27 22R,22L); 6/1/11-8/31/11 Number of Active Departures Queue Size Taxi Time 6 8 10 12 14 16 18 20 22 24 Local Time (hrs) 18
Operational Data Analysis: Runway Configuration Use Congestion management needs to be tailored to dominant runway configurations BOS: two dominant configurations LGA: multiple configurations PHL: one dominant configuration LGA Runway Configuration Usage; 6/1/11-8/31/11 4 4 31 31 4 13 PHL Runway Configuration Usage; 6/1/11-8/31/11 27L 27L 27R 27R 9R 9R 9L 9L 4 31 13 4, 13 12% 31 4 9R 9L 26% 17% 37% 77% 17% 27R 27L 22 13 22 31 22 22 19
Operational Data Analysis: Airline Mix Congestion management implementation may vary significantly with airline mix American Continental Delta JetBlue Southwest United PHL: dominant carrier BOS/LGA: mixed operators BOS Aircraft Operations by Airline 12% 16% 8% 12% USAirways 23% Other 13% AirTran Air Canada Cape Air Continental LGA Aircraft Operations by Airline Delta American 18% JetBlue Southwest United PHL Aircraft Operations by Airline Southwest United Chautauqua Delta 8% American Other 10% UPS 19% Other 8% 23% 8% USAirways Shuttle America Spirit Chautauqua AirTran Air Canada 68% USAirways All data from 6/1/11-31/8/11 20
Operational Data Analysis: Traffic Demand Characteristics of airport traffic for dominant configurations Departure demand Queue size Taxi time Instrumental in tuning congestion management control variables and strategies Number of Aircraft/Time (mins) Number of Aircraft/Time (mins) 40 35 30 25 20 15 10 5 0 40 35 30 25 20 15 10 5 PHL Surface Metrics (27R 27L); 6/1/11-8/31/11 6 8 10 12 14 16 18 20 22 24 Local Time (hrs) LGA Surface Metrics (22 13); 6/1/11-8/31/11 Number of Active Departures Queue Size Taxi Time Number of Active Departures Queue Size Taxi Time 0 6 8 10 12 14 16 18 20 22 24 Local Time (hrs) 21
Operational Data Analysis: PHL Traffic Demand 22
Operational Data Analysis: Throughput Saturation Differences between runway configurations at an airport Departure rate Saturation point PHL 27R 27L PHL 9R 9L 23
Airport Characterization: Implications for Congestion Management BOS: Evening peak Two main configurations Mix of airlines Aggregate solution, tailored to two runway configurations, primarily necessary in evening LGA: Constant high demand Mix of airlines/configurations Aggregate solution, needed most of operating day PHL: Intermittent peak demand Dominant runway configuration Dominant airline Congestion management needed in demand peaks; potential for airline-specific solution 24
Outline Motivation Framework for adapting surface congestion management approaches Airport characterization Site visits Surface visualizations Operational data analysis Algorithm development Implementation design Testing and performance evaluation 25
Algorithm Development Algorithm concept Departure rate Airport X, Configuration Y, Condition Z Saturation throughput, T* Saturation point, N* Control point, N ctrl Traffic Metric, e.g. No. of aircraft on surface, Dep queue length, etc. Need curve characteristics for each airport/configuration Airport BOS LGA PHL Configuration (arrivals departures) Saturation point, N* (# active dep.) Saturation Throughput, T* (ac/hr) 4R, 4L 9, 4R 17 48 22L, 27 22R, 22L 13 45 22 13 11 36 31 4 15 40 22 31 18 42 4 13 15 36 27R 27L 12 48 9R 9L 20 40 26
Algorithm Development: Parametric Dependencies of Throughput Departure throughput dependencies vary by airport BOS: Arrival throughput, departure demand, departure fleet mix (props) LGA: Arrival throughput, departure demand, departure route availability PHL: Arrival throughput, departure fleet mix (props), fleet mix (Heavy aircraft), departure route availability Reliable throughput predictions are important for effective metering To avoid low runway utilization To avoid excessive surface congestion (mean, std deviation) of departure throughput/15 min BOS in 22L, 27 22R, 22L under saturation [I. Simaiakis, 2012] 27
Outline Motivation Framework for adapting surface congestion management approaches Airport characterization Site visits Surface visualizations Operational data analysis Algorithm development Implementation design Testing and performance evaluation 28
Implementation Design Airport/ATC tower operating characteristics Ramp or FAA tower-controlled pushbacks Tower layout and equipment Algorithm information input requirements Capacity and demand forecasts Algorithm execution platform Algorithm output format Algorithm execution procedures Tablet 2: Recommended push-back rate display BOS Tower Cab Tablet 1: Data input Capacity (Airport config.), Weather (VMC/IMC) Demand (Aircraft with Ground/Local Control, Expected arrivals) 29
Outline Motivation Framework for adapting surface congestion management approaches Airport characterization Site visits Surface visualizations Operational data analysis Algorithm development Implementation design Testing and performance evaluation 30
Operational Testing and Performance Evaluation Operational testing Validity and robustness under actual operational conditions Basis for refinement BOS Runway Utilization Benefits/cost assessment Compare surface congestion metrics before/after deployment Monetized benefits basis for investment analysis Airport operational efficiency Runway utilization Departure spacing BOS Departure Spacing 31
Sample Surface Visualization: BOS 22L, 27 22R during Metering (2011) 32
Summary Surface congestion management important to fuel burn/emissions reduction at many airports Existing deployments focused on specific airports: techniques needed for adaptation to more airports and operating conditions Adaptation framework proposed Airport characterization is an important first step: First-hand observations and opportunities to ask questions of ATC professionals with site visits Qualitative analysis with surface visualizations Quantitative analysis with operational data Significant (6-14%) potential benefits from departure metering BOS: 900K gallons savings of jet fuel per year LGA: Two most frequently-used configurations in VMC alone would yield 550K gallons savings of jet fuel per year, even after accounting for gate-conflicts PHL: 2.9M gallons savings of jet fuel per year 33