9th AIAA Aviation Technology, Integration, and Operations Conference (ATIO)
and
Air 21 - 23 September 2009, Hilton Head, South Carolina

AIAA 2009-6959

Measuring Vertical Flight Path Efficiency in the National Airspace System Monica S. Alcabin1, Robert W. Schwab2, Charlie Soncrant3, Kwok-On Tong4, and Susan S. Cheng5 Boeing Commercial Airplanes, Seattle, WA 98124-2207

The current economic downturn and concern for global climate change has caused the aviation industry to seek ways to reduce operating costs, fuel usage, and CO2 emissions. The report, Aviation and the Global Atmosphere published by the 1991 Intergovernmental Panel on Climate Change, estimated that improvements in Air Traffic Management could improve overall fuel efficiency by 6 – 12%.1 However, in order to define and prioritize the steps that need to be taken to reduce operating costs, fuel usage, and CO2 emissions, it is necessary to understand exactly where the inefficiencies are in the current system and the contribution, by phase of flight, compared to that estimate of 6 – 12%. This paper distinguishes between current system inefficiencies due to flight delays, as well as vertical path and lateral path inefficiencies. In this paper we propose a methodology that was developed at The Boeing Company over the past year to (1) measure the vertical path inefficiency of sampled flights in the National Airspace System (NAS), (2) extrapolate the inefficiencies of the sampled flights to an annual level based on the schedule published in the Official Airline Guide, and (3) use this annualization to estimate potential benefits of proposed new technologies that support closer conformance to optimum vertical profiles. An initial estimate of the excess fuel burned due to vertical flight path inefficiency in the NAS represents 3% of the total fuel burned by domestic commercial carriers in 2006. Improvements to the methodology are proposed in addition to a discussion of the evaluation of benefits by phase of flight that will be the focus of future work.

I. Introduction

T

HE Boeing Company’s decision to provide new avionics technologies such as Required Navigation Performance (RNP), Global Positioning Landing System (GLS), Automatic Dependent Surveillance – Broadcast (ADS-B) and Cockpit Display of Traffic Information (CDTI), as well as airlines’ decisions to equip new and existing aircraft with these technologies are dependent on closing the business case – in other words, the benefits must outweigh the costs and risks. The benefits come from operational changes that provide reduced flight delays and increased flight efficiency. The combination of reduced delays and increased efficiency result in reduced fuel consumption and reduced environmental impact, in the form of reduced CO2 emissions. Costs are incurred by airlines and air navigation service providers in the form of changes to the airplane and to air traffic management infrastructure and procedures. These changes must occur together in order to recover the investment. The risks are due to the uncertainty of these changes being implemented in a timely fashion to take advantage of the benefits. The benefits provide an “opportunity cost” for the solution set of avionics and ground infrastructure. Excess fuel burn in the air traffic management system is primarily characterized by flight delay costs and flight efficiency costs. Flight delays occur when an airport or airspace resource (runway, gate, taxiway, or airspace sector) has greater demand than the available capacity. Flight delays tend to grow exponentially with increased levels of traffic. Flight efficiency is measured as the increased flight time, distance, and fuel compared to an “ideal” flight trajectory. Flight efficiency may be related to traffic or to other environmental or planning factors, but usually increases more-or-less linearly with the level of traffic. Figure 1 presents the relationships between delay, efficiency, and the associated time and fuel costs described above. Delays are accommodated by gate-holding, runway-holding, 1

Associate Technical Fellow, Avionics/Air Traffic Management, PO Box 3707, MC 07-25, Senior Member. Retired Technical Fellow, Avionics/Air Traffic Management, PO Box 3707, MC 07-25. 3 Retired Boeing Analyst, Avionics/Air Traffic Management, PO Box 3707, MC 07-25. 4 Principal Engineer, Avionics/Air Traffic Management, PO Box 3707, MC 07-25, Senior Member. 5 Engineer, Avionics/Air Traffic Management, PO Box 3707, MC 07-25, Member. 1 American Institute of Aeronautics and Astronautics 2

Copyright © 2009 by The Boeing Company. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

lateral path vectoring or speed controls. To avoid CAPACITY EFFICIENCY double-counting in the estimate of operational NAS Baseline inefficiencies we distinguish between excess lateral Airport Sector Flight distance due to vectoring and lateral distance due to Capacity Capacity Profiles Model Model inefficient lateral paths. Lateral inefficiency has been the subject of recent studies and is not included in this National analysis2-5. The focus of the work presented in this Lateral Vertical Flow Inefficiency Inefficiency Model paper is specifically on vertical flight efficiency, Model Model which is represented by the flow diagram on the far Delay Calibration right in Figure 1. Level-off Increased Path Delay Aircraft are designed and onboard flight systems Radar Vectors Fuel Burn Distance Costs & Airborne Holding are developed with the goal of minimizing the Taxi & Departure Delays Gate Holds combined costs of flight fuel and time by selecting an “optimal” vertical path for a known lateral path. Flight Time Costs Fuel Costs and Increased Emissions management systems assist flight crews in the Figure 1. Relationship between flight delays, selection of speed/altitude commands that provide the lateral, and vertical inefficiencies in the current best flight efficiency for a given set of airline and Air NAS. Traffic Control (ATC) constraints. This paper provides a description of the methodology developed to calibrate the current vertical flight path inefficiency in the NAS by sampling the measured vertical profile of actual flights and extrapolating the sampled flights to an annualized operational estimate for all scheduled flights.

II. Technical Approach Sampling the NAS to estimate vertical flight path inefficiencies begins with a realization that airport and airspace operations exhibit complex behavior with many operating modes. Initial discussions on sampling suggested selecting a day’s (or a year’s) sampling of flights. However, such an approach would have been a resource intensive task and difficult to justify in these economically constrained times. An alternate strategy was to find a way to dissect the problem such that representative flights could be sampled and aggregated to the full NAS level. The problem remained to analyze enough flights to provide a representative sample with credible results. After some initial flight analysis and categorization, a sampling methodology was hypothesized that appeared feasible and credible. The technical approach and results are presented here so that others that have access to more extensive data and automated techniques will be motivated to continue the analysis, add to, and refine the results. A. Airport Categorization The US Federal Aviation Administration (FAA) has focused much of its analysis efforts on the 35 Operational Evolution Plan (OEP) airports.6 Therefore, the technical approach selected begins by categorizing those 35 airports into high, medium, or low complexity based on terminal airspace and operations factors. Nearby airports, runway configurations, and level and type of traffic demand were taken into consideration. Subject matter experts experienced in airport and terminal procedures and operations were consulted in the categorization of airports. Figure 2 presents a map of the continental US showing the categorization of the 35 OEP airports into high, medium, and low complexity; red represents high, yellow represents medium, and blue represents low complexity.

High (Chicago) Medium (Denver)

"Seattle #Portland

Low (Other)

"Minneapolis O'Hare

$ Midway

"Salt Lake City

$San Francisco

"Denver

"Las Vegas # Honolulu

"St Louis

$Los Angeles San# Diego

#Memphis

"

Phoenix

Boston

$ La Guardia $ $Kennedy $ Newark$ Cleveland $ $ Philadelphia Pittsburgh Baltimore $Reagan $ $ Dulles

Detroit

"Cincinnati

"Charlotte

$Atlanta

"Dallas/Fort Worth

$Houston Bush

Orlando Tampa # # Fort Lauderdale

" "

Miami

Figure 2. The 35 OEP airports are categorized into high, medium, and low complexity factors.

B. City-Pair Selection Once airports were categorized, flights were selected for analysis. In the initial exercise, flights with origin from or destination to Chicago O’Hare International Airport (ORD) were selected to represent a high complexity traffic and airspace environment whereas flights to/from Denver International Airport (DEN) were selected to represent a medium complexity traffic and airspace environment. For a given airport, connecting origin and/or destination 2 American Institute of Aeronautics and Astronautics

airports were selected such that the collection of flights would encompass the different approach and departure procedure options (in the form of arrival corner-posts or departure gates) for each airport. Figure 3 presents a map showing the specific city-pairs used for the flight sampling. Chicago origin or destination flights are in blue; Denver origin or destination flights are in orange. These city-pairs form the basis of a network of flights for the analysis. C. Flight Sampling for Chicago and Denver Airspace charts for Chicago were analyzed and city# pairs were selected such that the departures selected for the Chicago analysis used each of the different airspace # # # departure gates and similarly, the arrivals selected for the # # # # analysis used each of the different Chicago arrival corner# # # posts. For Chicago departures, the flights selected were # bound for southeast, east, west, and southwest # # # # destinations. For Chicago arrivals, the selected flights # # arrived from the five arrival corner-posts: JANESVILLE # (from the northwest), BRADFORD (from the southwest), # ROYKO (from the southeast), KNOX (from the east), and SAYRS (from the northeast). A total of 81 flights were analyzed for Chicago – 63 arrivals and 18 departures. Figure 3. Network of flights used for the analysis. Similarly for Denver, a total of 61 flights were Flightaware.com analyzed – 41 arrivals and 20 departures. Denver Denver (DEN) Chicago (ORD) departures were selected to exercise various departure Arrivals Arrivals By Cornerpost JVL (13 flights) procedures. Denver airspace uses dual arrival cornerNW (8 flights) BDF (13 flights) posts from the four quadrants and used the four NE (11 flights) OXI (12 flights) directional arrival corner-posts: NW, SW, SE, and NE. SW (12 flights) ROY (13 flights) The specific city-pairs and number of flights selected for SE (10 flights) SAY (12 flights) By the sampling of flights for Chicago and Denver are Departures Destination Departures summarized in Figure 4. Airport ATL (5 flights) ATL (4 flights) Latitude, longitude, and ground speed (in knots) DFW (5 flights) DCA (4 flights) information was downloaded for each of the flights in DEN (5 flights) SEA (5 flights) one minute increments from the publicly available DFW (5 flights) SFO (5 flights) website, www.flightaware.com. Flights were selected Figure 4. Specific city-pairs used for the sampling randomly with respect to day of the week, time of day, of flights. airline, and weather conditions. However, an effort was made to select flights representing a range of aircraft types. Seattle/Tacoma

Minneapolis

Toronto

Boston

New York La Guardia Cleveland Philadelphia

Chicago O'Hare

Denver

Washington Reagan National

San Francisco

Charlotte

Los Angeles Int'l

Albuquerque

Phoenix

Tucson

Atlanta

Dallas/Fort Worth

Houston Bush

Miami

D. Flight Profile Analysis The latitude, longitude, and ground speed data for each of the sampled flights was converted to track distance and associated altitude for each flight. Figure 5a and, in a more expanded view, Figure 5b, present the lateral profiles and Figure 6 presents the vertical profiles for five sampled flights between San Francisco (SFO) and Chicago during June 2008. For each flight, the distance of any vertical flight path inefficiency (or level-off) at each altitude was estimated. The level-offs are seen in more detail when the vertical profiles are presented on an expanded scale as in Figure 7. The distance of each of the level-offs for each flight was recorded for both departures from and arrivals into Chicago. Departures were grouped together whereas arrivals were grouped by corner-post. The average expected value of the level-off distance for all sampled flights was then calculated for each altitude. For example, for Chicago arrivals from BRADFORD (the southeast corner-post), the level-off distance at FL 330 was recorded for each flight. The distance for all BRADFORD arrivals at FL 330 was summed and divided by the total number of arrivals across BRADFORD to estimate the average level-off penalty for that altitude. The level-off penalties (in nm) by altitude for all sampled flights arriving to and departing from Chicago is presented in Figure 8. The magnitude of the level-off penalty by altitude is important because aircraft are more efficient at higher altitudes and burn more fuel at the lower altitudes.

3 American Institute of Aeronautics and Astronautics

Ground Tracks

Ground Tracks

#MSP

#ORD #

ORD

#DSM

#

#

SFO

LNK

#IND

a) b) Figure 5a, 5b. Ground tracks for five flights from SFO to ORD, and on a more expanded scale. 40000 40000

35000 35000

30000 30000

25000 Altitude (ft)

Altitude (ft)

25000

20000

20000

15000

15000

10000

10000

5000

5000

0 1800

1600

1400

1200

SFO

1000

800

600

400

200

350

300

250

200

150

Along track distance to ORD (nm)

Figure 6. Corresponding vertical profiles for five flights from SFO to ORD. JANESVILLE

0 400

0 ORD

Along track distance to ORD (nm)

BRADFORD

100

50

0 ORD

Figure 7. Expanded scale showing level-offs for flights from SFO to ORD.

KNOX

ROYKO

SAYERS

DEPARTURES

380 360 340 320 300 280 Flight Level (100 ft)

260 240 220 200 170 150 130 110 90 70 50 30 10 20

30

40

10

20

30

40

10

20

30 40

10

20 30 40

10 20 30

40

10

20 30

40

Average Level-off Distance (nm)

Figure 8. Level-off penalties by altitude for all sampled flights to and from ORD. E. Aircraft Type Equivalence An analysis of the Official Airline Guide (OAG) schedule of commercial airlines operating at Chicago for May 2008 revealed that nineteen different aircraft types were represented in the schedule. Each aircraft type has a different fuel burn profile based on thrust, speed, weight, and atmospheric conditions. A detailed fuel burn analysis would involve simulated flight conditions, and would have been data and resource intensive, beyond what was 4 American Institute of Aeronautics and Astronautics

50.00 747

45.00

777

40.00

Long Range

35.00 Fuel per NMi

possible in this study. In order to simplify the analysis it was decided to divide the nineteen different aircraft types into three categories. The fuel/seat-nm was divided by the fuel/nm for each aircraft type. Break-points were drawn for the group of aircraft for long-range (such as the Boeing 777), mid-range (such as the Boeing 767), and single-aisle (such as the Boeing 737) aircraft and are presented in Figure 9.

25.00

D10

M11

30.00

A30

A34 A33

Mid Range

767

20.00

727 757

15.00

73G

10.00

M80 M90 A3273S

Single Aisle

D9S CRJ

ERJ

5.00 0.00 0.0600

0.0800

0.1000 0.1200 Fuel per Seat-NM i

0.1400

0.1600

Figure 9. Categorization of aircraft types into long-range, mid-range, and single-aisle airplanes.

F. Fuel Penalty Measurement for Arrivals and Departures Once the aircraft selection and categorization was completed, the fuel burn per nautical mile in level flight was calculated for each aircraft type. Simplifying assumptions of a standard operating weight and speed/drag profile were used. A plot of the relative fuel burn per nautical mile for the B737, B767, and B777 is presented in Figure 10. It was assumed that the optimal fuel burn for each aircraft would occur at cruise altitude. The fuel penalty measurement for each altitude for each aircraft type was then calculated as the difference between the fuel burn at the lower altitude and that at cruise for the level-off distance recorded. To evaluate the fuel efficiency for shortrange flights where optimum cruise flight levels are not achievable, a reduced efficiency altitude was assumed for comparison with the various level-off altitudes. A graphical presentation of the fuel burn penalty measurement is presented in Figure 11. 400

350 737 767

300

777

Fuel burn increases for 737 cruising at FL 290

Flight Level

250

Altitude

Fuel burn increases for 737 cruising at FL 210

200

150

100

50

0 400

Fuel Burn (lb/nm)

Figure 10. Relative fuel burn per nautical mile for B737, B767, and B777.

350

300

250

200

150

100

50

0

Distance from ORD (n.mi.)

Figure 11. Sample of fuel burn penalty measurement for level-off segments for 737.

The aircraft speed may also vary by altitude. However, the relationship of Mach and calibrated-airspeed versus true airspeed and ground speed is complex. In general, aircraft descending with constant Mach show accelerating true airspeed and those descending along a constant calibrated airspeed have accelerating true airspeeds. In a busy terminal airspace region within 200 nautical miles of Chicago O’Hare, aircraft speed is often limited by ATC using miles-in-trail (MIT) restrictions. Therefore, a conservative assumption was made that any time penalties for level-off operations would not be considered in this analysis. G. Fuel Penalty Measurement for En Route Cruise A simple analysis of en route fuel burn penalties was performed for the flights in the sampled set. For all of the flights to/from Chicago and Denver, those flights’ distances and altitudes where en route operations were recorded at off-optimal altitudes (excluding off-optimal altitudes already measured in climb and descent) were identified. To evaluate the fuel efficiency for short-range flights where optimum cruise flight levels are not achievable, a reduced efficiency altitude was assumed for comparison. 5 American Institute of Aeronautics and Astronautics

H. Extrapolating Results to the NAS With a methodology proposed for evaluating the vertical flight path inefficiencies for Chicago (representing a high complexity airport), Denver (representing a medium complexity airport), and the en route segments for those flights, the question remained as to how to extrapolate the results to the rest of the NAS. The OAG for May 2008 was used to determine the weekly frequency of commercial flights, by aircraft type, at all of the remaining OEP and non-OEP airports. A map of the airports considered in the analysis is presented in Figure 12. The methodology used to extrapolate the results from the sampled flights to the FOR EACH AIRPORT Calculate % 737 767 777 Calculate Annual number Of flights Determine Airport Category

HIGH

LOW

MEDIUM

Apply Chicago Level-off Penalty by Altitude

Apply Denver Level-off Penalty by Altitude

Apply 50% of Denver Level-off Penalty by Altitude

Sum level-offs For all Flights

Sum level-offs For all Flights

Sum level-offs For all Flights

Figure 13. Methodology used to extrapolate sampled results to the rest of the NAS.

Figure 12. Collection of airports having commercial service that were included in the NAS-wide analysis.

rest of the NAS is presented in Figure 13. For each airport, the aircraft types operating at that airport were categorized into the three aircraft performance classes – B737, B767, and B777 to evaluate the percentage of B737, B767, and B777 aircraft at that airport. The weekly frequency of flights was multiplied by 52 to calculate the annual number of commercial flights, by aircraft type, for each airport. If the airport was categorized as a high complexity airport, the level-off penalty by altitude for Chicago was applied to that airport. If the airport was categorized as a medium complexity airport, the level-off penalty by altitude for Denver was applied to that airport. Since no flights were collected for the low complexity airports, for simplicity, it was assumed that the level-off penalty for low complexity airports was 50% of that calculated for Denver. Airports in the vicinity of a major airport inherited the complexity of the nearby major airport. The FACT2 Study, published in 2007, was used to assign airports in a metropolitan area.7 Airports in the vicinity of San Francisco and Los Angeles International Airports that were included in the analysis are presented in Figure 14. Typical arrival profiles for high, medium, and low complexity airports are presented in Figure 15. 400 350

Flight Level

300

LOW

MEDIUM HIGH

250 200 150 100 50

Figure 14. Airports in the vicinity of San Francisco and Los Angeles included in the analysis.

0 400

350

300

250

200

150

100

50

Range from destination (n.mi.)

Figure 15. Typical arrival profiles for high, medium, and low complexity airports.

6 American Institute of Aeronautics and Astronautics

0

III. NAS-Wide Vertical Flight Path Inefficiency Analysis Results The methodology described above was applied to the 142 sampled flights to evaluate the excess fuel burn due to vertical flight path inefficiency for commercial flights into and out of Chicago O’Hare and Denver International Airports. The OAG for May 2008 was used to evaluate the impact of a full schedule at a specific airport. For each airport the weekly frequency of commercial flights, by aircraft type, was determined. The excess fuel burn for each aircraft performance class, for each phase of flight, was multiplied by the weekly frequency of each aircraft performance class to evaluate the weekly fuel penalty. The excess fuel burn/flight, for each phase of flight, was calculated by adding the weekly fuel penalties for the three aircraft performance classes and dividing by the total number of weekly flights. Vertical flight path inefficiency analysis results for Chicago O’Hare and Denver are presented, along with the inefficiency results for en route as well as NAS-wide results. A. Results for Chicago O’Hare and Denver International Airports Sixty-three arrivals and eighteen departures were sampled to evaluate the vertical flight path fuel penalty for Chicago O’Hare International Airport. Each Chicago arrival from the May 2008 OAG was assigned a corner-post based on the JANESVILLE SAYRS inbound direction from the departure airport, as presented in Figure 16. Departures were all KNOX grouped together since the level-off penalty for departures was minimal. The excess fuel burn by aircraft type by corner-post was calculated by BRADFORD ROYKO multiplying the level-off distance (nm) for each altitude in Figure 8 by the fuel burn penalty (lbs/nm) described in Section IIF. The excess fuel burn for the weekly number of flights were summed and divided by the total weekly number of flights to calculate the average fuel penalty/flight for arrivals to and departures from Chicago. The Figure 16. Categorization of arrivals into Chicago by results, presented in Figure 17, show that the corner-post. average fuel penalty for Chicago is 345 lbs per arrival and 41 lbs per departure. The results vary considerably by aircraft type; however, results are JANESVILLE dominated by the large number of single-aisle airplanes, represented in this analysis by the B737 BRADFORD aircraft. Surprisingly, the results did not differ by corner-post. The fuel burn penalty for B767 and 737 KNOX 767 B777 arrivals ranged from 500 to 900 lbs/flight. 777 This is in line with actual fuel savings for B777 and Avg B747 partial tailored arrivals conducted into San ROYKO Francisco in December 2007.8 Results from those flights show that the fuel savings for full tailored SAYRS arrivals was approximately three times these values. These results indicate that Departure our analysis results may be conservative or that the potential fuel savings for oceanic flights may be 0 100 200 300 400 500 600 700 800 900 1000 greater due to their increased fuel and payload. Average fuel penalty/flight (lbs) The analysis for Denver was conducted Figure 17. Average fuel penalty/flight by corner-post for similarly to that for Chicago. Results for Denver ORD. show that the average fuel penalty for Denver is 216 lbs per arrival and 19 lbs per departure. The results from the methodology described for Denver are within 8% of the estimated average fuel savings of 200 lbs/flight for flights into Denver International Airport during October 2007.9 Figure 18 presents the typical excess fuel burn for arrival and departure level-offs at the high, medium, and low complexity airports. The inefficiency results for Chicago O’Hare (representing a high complexity airport) are 7 American Institute of Aeronautics and Astronautics

50% higher than those for Denver (representing a medium complexity airport). The analysis for these two airports shows that the excess fuel burn for departures is approximately 10% of that for arrivals. This 90%/10% split was also found to hold true for an analysis of 45,000 flight tracks for October 5, 2006.10 Figure 19 presents the annual excess fuel burn for arrivals and departures at the 35 OEP airports. 400

200

Annual Arrival & Departure Fuel Inefficiency (million lb)

180

300 High Medium Low 200

150

100

50

0

Low

Arrival 120

100

80

60

40

20

Departures

Figure 18. Typical excess fuel burn for level-offs for arrivals and departures at high, medium, and low complexity airports.

0 PH L SF O D EN LG A IA D BO S D C A C LT PH X LA S M SP BW I M IA C LE M D W C VG SE A SL C ST L FL L PI T M C O SA N H N L M EM TP A PD X

Arrivals

Medium

140

IA H EW R D TW D FW

250

High Departure

160

AT L O R D LA X JF K

Average Fuel Inefficiency per Trip (lb)

350

Figure 19. Total annual arrival inefficiency at the 35 OEP airports.

and

departure

Annual Fuel Inefficiency (million lb)

B. En Route Vertical Flight Path Inefficiency Results The 142 sampled flights that 3000 comprise the network of flights into and out of Chicago O’Hare and Denver International Airports were used for the 2500 en route inefficiency analysis. The network of flights was previously 2000 presented in Figure 3. Similarly to the methodology applied to arrivals and 1500 departures, the excess fuel burn for en route for each aircraft performance class was calculated by multiplying the 1000 difference between the fuel burn at the off-optimal altitude and that at the 500 optimal cruise altitude by the distance recorded at the off-optimal altitude. The excess fuel burn for each aircraft 0 Total Departure Enroute Arrival TRACON performance class was multiplied by the Transition weekly frequency of each aircraft Figure 20. NAS-wide vertical flight path inefficiency results. performance class to calculate the weekly fuel penalty. The en route fuel penalty/flight was calculated by adding the weekly fuel penalties for the three aircraft performance classes and dividing by the total number of weekly flights. The en route penalty was calculated to be 27 lbs/flight. C. NAS-Wide Results NAS-wide vertical flight path inefficiency results for each phase of flight and the total are presented in Figure 20. The arrival phase of flight was divided into two parts – Arrival Transition and TRACON. The level-off penalties by altitude in Figure 8 showed that all flights exhibited level-offs when they were flying between 7,000 and 13,000 ft, irrespective of corner-post. This is probably due to air traffic control having to space aircraft from different routes to sequence them for arrival. Therefore, level-off penalties between en route and 13,000 ft were considered to be in the arrival transition phase of flight and those from 13,000 ft and below in the TRACON phase of flight. The total NAS-wide annual excess fuel burn due to vertical inefficiency is 2.6 billion lbs. This is significant, representing 3% of the total fuel consumed by US domestic commercial aircraft in 2006.11 In comparison, the report on flight delays published by the United States Congress Joint Economic Committee in May 2008 claimed that 8 American Institute of Aeronautics and Astronautics

Departure 9%

delays accounted for 5% of the total fuel consumed by US domestic commercial aircraft in 2006.12 The vertical flight path inefficiency, by phase of flight, is presented in Figure 21. It shows that departure combined with en route account for approximately 20% of the vertical flight path inefficiency while arrival transition accounts for another 20% and TRACON accounts for 60% of the vertical flight path inefficiency.

IV. Implications for Future Work

Enroute 11%

TRACON 62%

Arrival Transition 18%

The development of a baseline of NAS operations is critical to the development of technology business cases. It is important to understand where the inefficiencies are, by phase of flight, so that Figure 21. Vertical flight path the right technologies can be applied. The work in this paper inefficiency by phase of flight. demonstrates an approach to the development of a vertical flight path efficiency metric. In this analysis, the sampling was limited to 142 flights. Arrivals and departures for Chicago O’Hare and Denver International Airports were analyzed to evaluate the excess fuel burn due to level-offs during arrival, departure, and en route phases of flight. A more comprehensive study should be conducted that samples a much greater number of flights and takes into consideration the number of airports, airport conditions such as weather and runway configurations, and traffic levels. In this study three aircraft performance classes were defined. However, a future study should consider additional aircraft types. This study used simple graphical evaluation of flight profiles; however, a more comprehensive study might require a simple aircraft performance model such as BADA13 as well as speed difference effects. The performance baseline should be validated and/or verified; actual airline estimated fuel burn variations, by flight, could be used to calibrate such a model. If the model is developed correctly, the number of airports and airspace state variables should be used to “explain” the variability measured in the NAS. The methodology used for this analysis is presented so that others in the industry that have access to more extensive data and automated techniques will be motivated to continue the analysis and add to and refine the results. Future Boeing studies will combine the results of this vertical inefficiency analysis with delay and lateral inefficiency results from other industry studies to evaluate the total excess fuel burn by phase of flight. Potential avionics and ground infrastructure solutions for each phase of flight will be proposed and benefits, in the form of fuel and time savings, will be evaluated.

V. Conclusions The aviation industry is focusing its efforts on reducing costs, fuel usage, and emissions. In order to define and prioritize the steps that need to be taken to reduce the environmental impact, it is necessary to understand where the inefficiencies are, by phase of flight, because different technologies can be applied to mitigate those impacts. A methodology was developed to evaluate the current vertical flight path inefficiency in the US by measuring the vertical flight path inefficiency of sampled flights in the National Airspace System, extrapolating the inefficiency of the sampled flights to an annual level using the schedule published in the Official Airline Guide, and estimating the potential benefits of avionics and ground infrastructure that support closer conformance to optimum vertical profiles. Results show that the vertical flight path inefficiency represents 3% of the total fuel consumed by domestic commercial aircraft in 2006, compared with delays accounting for 5% of the total fuel consumed by those same aircraft in 2006.12 These two factors, though caused by different mechanisms and having different solutions, combine to represent 8% of the potential improvements in efficiency that could be enabled by Air Traffic Management claimed by the Intergovernmental Panel on Climate Change.1 Analysis results show that vertical flight path inefficiency in the TRACON accounts for 60%, arrival transition for 20%, and departures and en route each account for approximately 10% implying that the aviation industry needs to focus its efforts on arrival management processes and procedures, where the majority of the vertical inefficiency occurs.

Acknowledgments The authors appreciate the contributions to this work by the following: Dan Boyle and Larry Parrent who provided advice on the categorization of airports and Michael Coates who graphically represented the raw data. Our gratitude also goes out to John Olsen, Joe McPherson, Tim Tuttle, Mike Murphy, and Scott Pelton, who continue to encourage our team and support our work with program management and budgets. 9 American Institute of Aeronautics and Astronautics

References 1

Intergovernmental Panel on Climate Change, Aviation and the Global Atmosphere, Cambridge University Press, 1999. 2 Reynolds, T. G., “Analysis of Lateral Flight Inefficiency in Global Air Traffic Management,” 26th International Congress of the Aeronautical Sciences/8th AIAA Aviation Technology, Integration & Operations Conference, Alaska, 2008. 3 Reynolds, T. G., “Development of Flight Inefficiency Metrics for Environmental Performance of ATM,” 8th USA/Europe Air Traffic Management Research and Development Seminar, Napa, 2009. 4 Kettunen, T., Hustache, J.-C., Fuller, I., Howell, D., Bonn, J., & Knorr, D., “Flight Efficiency Studies in Europe and the United States,” 6th USA/Europe Air Traffic Management Seminar, Baltimore, 2005. 5 DeArmon, J., Dorfmann, G., Solomos, G., Bluncher, M., Knorr, D., and Yi, H., “Excess Flying Time in the National Airspace System,” Air Traffic Control Quarterly, Vol. 14(4), 2006, pp 283 – 309. 6 United States Federal Aviation Administration, “FAA Airport Capacity Benchmark Report,” Washington, DC, September 2004, URL: http://www.faa.gov/about/office_org/headquarters_offices/ato/publications/bench/DOWNLOAD/pdf/2004_Benchmark_Report.p df. 7 United States Federal Aviation Administration, “Capacity Needs in the National Airspace System, 2007 – 2025,” Washington, DC, May 2007, URL: http://www.faa.gov/airports/resources/publications/reports/media/fact_2_main.pdf. 8 Elmer, K., Mead, R., Bailey, L., Cornell, B., Follett, J., Lanier, R., and Aircraft Noise Abatement Office, “Operational Implementation of Tailored Arrivals at SFO and Expected Environmental Benefits,” 26th International Congress of the Aeronautical Sciences/8th AIAA Aviation Technology, Integration & Operations Conference, Alaska, 2008. 9 Shresta, S., Neskovic, D., and Williams, S., “Analysis of Continuous Descent Benefits and Impacts During Daytime Operations,” 8th USA/Europe Air Traffic Management Research and Development Seminar, Napa, 2009. 10 Melby, P., Mayer, R., “Benefit Potential of Continuous Climb and Descent Operations,” 26th International Congress of the Aeronautical Sciences/8th AIAA Aviation Technology, Integration & Operations Conference, Alaska, 2008. 11 United States Environmental Protection Agency, “Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990 – 2007,” Washington, DC, April 2009, URL: http://epa.gov/climatechange/emissions/usinventoryreport.html. 12 United States Congress Joint Economic Committee Majority Staff, “Your Flight Has Been Delayed Again,” Washington, DC, May 2008, URL: http://jec.senate.gov/index.cfm?FuseAction=Reports.Reports&ContentRecord_id=11116dd7-973c61e2-4874-a6a18790a81b&Region_id=&Issue_id= 13 Eurocontrol, “Base of Aircraft Data (BADA 3.6),” The EUROCONTROL Experimental Centre, Bretigny, France, 2004, URL: http://www.eurocontrol.int/eec/public/standard_page/proj_BADA.html.

10 American Institute of Aeronautics and Astronautics