air transportation is a complex adaptive system: not an air traffic control automation problem dr....
TRANSCRIPT
Air Transportation is a Complex Air Transportation is a Complex Adaptive System:Adaptive System:
Not an Air Traffic Control Not an Air Traffic Control Automation ProblemAutomation Problem
Dr. George L. DonohueGeorge Mason University
March 18, 2004 Harvard University© George Donohue 2004
OutlineOutline
• How did We Get Here?
• Why Should We Care?
• Capacity - Delay
• Capacity – Delay – Safety
• System Network Effects
• Observations and Recommendations
How did We Get Here?How did We Get Here?
• 1903 Wright Bros. produced a Heavier than Air Flying System: AGE OF INVENTION– Airfoils, L/W Structure, Controls, L/W Propulsion
• WW II ( +50 Yrs) System is Upgraded: AGE OF ALL WEATHER COMMERCIAL FLIGHT– Radar, Jet Aircraft, Radio Navigation and Communication
• 2003 ( +100 Yrs.) System Needs Upgrading Again: AGE OF RELIABLE INTERNATIONAL TRANSPORTATION NETWORK – Predictable under all Weather Conditions
– Maximum Airport Capacity Utilization
– Near Optimal Network Load Balancing
– Predicable Safety Operating Margins
Barriers to the Third Age VisionBarriers to the Third Age Vision
• The Technical Community has been Aware of the Transition Problem for Over 20 Years! – Increasing Delays, Flight Cancellations
– Increasing Runway Incursions, ATC Op Errors and TCAS RA’s
• The Technical Elements that will enable the Development of an Affordable, Reliable, and Predictable Mode of International Transportation Already Exist!– TCAS II Deployed Worldwide No System Credit
– FMS with ±30 Sec. RTA ? No System Credit
– GPS Navigation and Surveillance (ADS-B) No System Credit
– Digital Communication Data Links NOT DEPLOYED
– TMA, pFAST, aFAST, URET No System Credit
• The Barriers to Growth are Regulatory and Institutional!
Source: USDOT BTS NTS 2000; USDOT BTS update April, 2002; DOC BEA 2002 (*real GDP using 1996 chained dollars
Increasing Delay is a Frog in the Boiling Water Problem
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
1960 1965 1970 1975 1980 1985 1990 1995 1998
Year
Gro
wth
rat
e/G
DP
gro
wth
rat
e
Air carrierpassenger miles
Highway trip miles
Rail passenger miles
Deregulation
Air Transportation’s Air Transportation’s Contribution to GDPContribution to GDP
Total Output GDP ContributionAir Transportation $205 $80Aircraft Manufacturing $134 $94Tourism $94 $85Agents/Forwarders $3 N/CGovernment $2 N/CTotal Impact $438 $259
Economic Impact of Aviation Industry($billions 1999)
Civil Transport Share of Civil Transport Share of Aerospace IndustryAerospace Industry
$473
$100
$81
$65
$54$31 $6
Source: L. Anderson, op cit., from Forecast International
Large Civil
Transports
Fighter, Attack
and Trainer
Rotorcraft
(Military and
Civil)
Business /
Corporate
Regional /
Commuter
Military
Transport
General
Aviation
Total Projected Aircraft Market 1999 to 2008: $810 Billion
FUTURE MARKETS FOR AERONAUTICS PRODUCTS ARE LARGE
74% CIVIL TRANSPORT
Capacity and DelayCapacity and Delay
• System Capacity is Primarily Limited by Network Runway Availability
• ATC Workload is an important Secondary Limitation
• Runway Maximum Capacity is a function of Aircraft Landing Speed and Runway Occupancy Time (ROT)
• Delay is a Non-Linear function of Demand to Maximum Capacity Ratio– Stochastic FCFS System– Queuing Theory Applies
• Major Hub Airports are Over-Scheduled– Transportation Network Is NOT LOAD BALANCED– Market Mechanisms Could Achieve this Goal
Network Operational Capacity is Network Operational Capacity is a Limited Commoditya Limited Commodity
• CMAX = 2 C AR MAX S i (XG)i Ri {Airports}
–K AK(t) {Airspace Management Intervention}
• S = f ( Safety, ATC , Wake Vortex, etc.) ~ 0.6 to 0.8
• AK(t) = (A/CREQUEST – A/CACCEPT) ~ [ 0 to >1,000]
– AK(t) = f ( GDP:Weather, Sector Workload Constraints )
• C AR MAX ~ 40 Arrivals/Hour (set by Runway Occupancy Time)
• Ri = Number of Runways at ith Airport
• XGi = Airport Configuration Factor at ith Airport
• i = 1 to N, where N is approximately 60 Airports
• K = 1 to M, where M is typically much less than 100 Sectors
ATS Delays Grow Exponentially ATS Delays Grow Exponentially with Increasing Capacity Fractionwith Increasing Capacity Fraction
PREDICTED DELAY vs. CAPACITY FRACTION
0.05.0
10.015.020.025.030.035.040.045.050.0
0% 20% 40% 60% 80% 100%
AIRPORT CAPACITY FRACTION
NU
MB
ER
DE
LA
YS
( >
15 m
in /
1000
ope
rati
ons)
Aircraft Arrival Rate:Aircraft Arrival Rate:Distance-Time RelationshipDistance-Time Relationship
0
10
20
30
40
50
60
70
80
0 1 2 3 4 5 6 7
DISTANCE ( NMi)
AR
RIV
AL
S /
RW
/ H
R 120 Knots
130 Knots
140 Knots
Spacing
(sec)
60
90
180
120
72ROT?
WV?
NY LaGuardia: A non-Hub NY LaGuardia: A non-Hub Maximum Capacity AirportMaximum Capacity Airport
• 1 Arrival Runway
• 1 Departure Runway
• 45 Arrivals/Hr (Max)
• 80 Seconds Between Arrivals
• 11.3 minute Average Delay
• 77 Delays/1000 Operations
• 40 min./DelayTOTAL SCHEDULED OPERATIONS AND CURRENT OPTIMUM RATE BOUNDARIES
0
4
8
12
16
20
24
28
32
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Schedule Facility Est. Model Est.
New York LaGuardia Airport New York LaGuardia Airport Arrival- Departure Spacing VMCArrival- Departure Spacing VMC
0
10
20
30
40
50
60
0 10 20 30 40 50 60
Departures per Hour
Arr
ival
s pe
r H
our
ASPM - Apr 2000 - Visual Approaches
ASPM - Oct 2000 - Visual Approaches
Calculated VMC Capacity
Optimum Rate (LGA)
40,40
Each dot represents one hour of actual traffic
during April or October 2000
45 Arr./Hr/RW@ 80 sec separation
DoT/FAA
LGA Arrival - Departure IMCLGA Arrival - Departure IMC
0
10
20
30
40
50
60
0 10 20 30 40 50 60
Departures per Hour
Arr
ival
s p
er H
ou
r
ASPM - April 2000 - Instrument Approaches
ASPM - October 2000 - Instrument Approaches
Calculated IMC Capacity
Reduced Rate (LGA)
32,32
Capacity-Delay-SafetyCapacity-Delay-Safety
• ATM System Safety and Capacity are Non-Linearly Related
• Wake Vortex Separation sets the Current System Capacity Limit in Instrument Meteorological Conditions– Safety Limitation
• ICAO System Safety Goal is 10-9 / Operation
• Small number Statistics leads us to use Accident Precursors as Safety Indicators
• Safety Analysis must be Analytical
Time Separation Is an Important Time Separation Is an Important Determinant of the SAFETY LimitationDeterminant of the SAFETY Limitation
Time (seconds)
Probability
ROT
A/C Inter-arrival Time
35+ Arrivals/RW/Hr
Accident Pre-Cursor Incidents Accident Pre-Cursor Incidents Safety – Capacity RelationshipSafety – Capacity Relationship
Haynie, GMU 2002
Hazard Reports 1988-2001
0
20
40
60
80
100
120
35 40 45 50 55 60 65 70
Percentage Capacity Used
Nu
mb
er R
epo
rts
Fil
ed
NMAC
RWY Inc
Legal Sep
ATL, BWI, LGA, DCA
ATL Estimated ATL Estimated Collision ProbabilityCollision Probability
Collision probability per SRO for each combination
Small-Heavy
Small-Large
Small-B757
Leader - Trailer
Richard Y. Xie, GMU research in progressRichard Y. Xie, GMU research in progress
Wake Vortex Accident Rate in Wake Vortex Accident Rate in Safety-Capacity CoordinatesSafety-Capacity Coordinates
Single Runway Estimated Wake Vortex Accident Rate50% Mix B747 & B737: S-Wake Calculation
y = -25004Ln(x) + 593477
R2 = 0.8731
y = -37168Ln(x) + 832913
R2 = 0.9698
-
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
450,000
- 1,000,000 2,000,000 3,000,000 4,000,000 5,000,000 6,000,000 7,000,000Safety - Arrivals / WV Accident
Cap
aci
ty
- A
rriv
als
/ Y
ear Log. (Hazardous Accident)
Log. (Catastrophic Accident)
NLR Stochastic AnalysisNLR Stochastic Analysis
30 Years3 Years
System Network EffectsSystem Network Effects
• Aprox. 10 Major Hub Airports are Operating at D/C max > 0.65
• Delays at these Airports spread Non-Linearly throughout the Network
• Runway Additions at one Airport May have Little Network Effect
• System-wide improvements have a Larger Effect than Individual Airport Improvements– Except at Major Airports like ORD, LGA and ATL!
Major US Airport CongestionMajor US Airport Congestion
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
DEN
BWI
JFK
PIT
CLT
DFW
DTW
IAD
EWR
PHL
SFO
LGA
MSP
SEA
ORD
STL
ATL
LAXA
IRP
OR
T
DEMAND / CAPACITY RATIO
J. D. Welch and R.T. Lloyd, ATM 2001
Queuing Delays Grow Rapidly
Market Does Not Act to Minimize Delay or Market Does Not Act to Minimize Delay or Maintain SAFETY: LGA Air 21 ImpactMaintain SAFETY: LGA Air 21 Impact
Source: William DeCota, Port Authority of New York
LaGuardia Airport
0
2040
60
80
100120
140
160180
200
06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Time of Day
Historic Movements AIR-21 Induced Svc.
Maximum Hourly Operations Based on Current Airspace & ATC Design
Atlanta: A Maximum Capacity Atlanta: A Maximum Capacity Fortress Hub AirportFortress Hub Airport
• 2 Runways – Arrivals
• 2 Runways – Departures
• 50 Arrivals/Hr/RW – Max
• 72 Seconds Between Arrivals
• 8.5 minutes Average Delay
• 36 Delays/1000 Operations
• 38 min./delay
TOTAL SCHEDULED OPERATIONS AND CURRENT OPTIMUM RATE BOUNDARIES
0
10
20
30
40
50
60
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Schedule Facility Est. Model Est.
Simulated AuctionSimulated Auction Delay Benefit at ATL Delay Benefit at ATL
45 min maximum schedule deviation allowed, no flights are rerouted
0
10
20
30
40
50
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
0
5
10
15
20
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Time (15-min bins)
Original Schedule Auctioned Schedule
Scheduled arrivals (#operations/quarter hour)
Estimated Average Runway Queuing Delay (min)
ATL reported optimum rate
Loan Le Research Loan Le Research in Progressin Progress
Observations – NAS SafetyObservations – NAS Safety
• We are approaching the Point that the existing system may be demonstrably less safe (at current and future capacity fractions) than a new, more synchronous, aircraft FMS/ADS-B separation based system
• System is Safe BUT Safety Margins are Diminishing!
• This case has not been Analyzed nor even Suggested as a RATIONAL for CHANGE to date!
Central Research QuestionsCentral Research Questions
• Both Safety and Efficiency Concerns lead us to the conclusion that the network should be operated as a Synchronous System– with economic incentives to use the largest aircraft
affordable and economically viable
• Time Window Auctions at Airport Metering Fix may provide the Economic Incentives necessary to maximize/optimize Network Capacity
• Central Research Questions:
– How Synchronous Can We Make this System – All Weather?
– What Should the Design Target Level of Safety Be?
• Backup Slides
Wake Normalized Aircraft Time Wake Normalized Aircraft Time Separation: LGA in VMC & IMCSeparation: LGA in VMC & IMC
02468
10
-60
-20 20 60 100
140
Seconds Deviation per Aircraft From Perfect WVSS Adherence Value
Air
craf
t / R
W /
Hr (
20 S
ec.
Bin
s)
VFR 33.8 Arr/hrIFR 34 Arr/HrVFR 30.9 Arr/HrVFR 27 Arr/Hr
Observed WV Separation Observed WV Separation Violations vs. Capacity RatioViolations vs. Capacity Ratio
Figure 6-5 Ratio of Incidents to Capacity Used
0
2
4
6
8
0 50 100 150 200
Percent of Capacity Used in 15 Minutes
Nu
mb
er
of
< W
VS
S
Inc
ide
nts
Ex
pe
cte
d i
n
15
Min
ute
s
BWI LGA Quadratic Model
Haynie, GMU 2002
ATL Arrival - Departure IMCATL Arrival - Departure IMC
0
20
40
60
80
100
120
0 20 40 60 80 100 120
Departures per Hour
Arr
iva
ls p
er
Ho
ur
ASPM - April 2000 - Instrument Approaches
Calculated IMC Capacity
Reduced Rate (ATL)
84,90
ATL and LGA Inter-Arrival Time in IMC ATL and LGA Inter-Arrival Time in IMC and VMC:32 - 39 Ar/Rw/Hrand VMC:32 - 39 Ar/Rw/Hr
LGA & ATL Arrival Histograms
-2
0
2
4
6
8
10
12
14
0 50 100 150 200 250
Inter-Arrival Time (Seconds)
Air
craf
t / R
W /
Hr
(20
Sec
. Bin
s)
LGA in VMC N=168
LGA in IMC N=124
ATL IN VMC N=114
ATL in VMC N=323
Observations - RecommendationsObservations - Recommendations
• FAA Culture – Barriers to Change
• NASA Culture – Barriers to Change
• State of NAS Safety
• Proposed Grand Experiment/OPEVAL
FAA Barriers to ChangeFAA Barriers to Change
• FAA has an Operational and Regulatory Culture– Inclination to follow training that has seemed
to be Safe in the Past
• FAR has NOT Changed to Provide Operational Benefits from Introduction of New Technology
• Assumption that Aircraft Equipage would be Benefits Driven did not account for Lack of an ECONOMIC and/or SAFETY Bootstrapping Requirement
FAA Investment Analysis Primarily FAA Investment Analysis Primarily focus on Capacity and Delayfocus on Capacity and Delay
• OMB requirement to have a B/C ratio > 1 leads to a modernization emphasis on Decreasing Delay
• In an Asynchronous Transportation Network operating near it’s capacity margin, Delay is Inevitable
• Delay Costs Airlines Money and is an Annoyance to Passengers BUT– is Usually Politically and Socially Acceptable
NASA Barriers to ChangeNASA Barriers to Change
• NASA has become more Process Oriented than Product Oriented
• Frequently Stated Objective of 25 year Implementation Goal Avoids Accountability and renders NASA TRL 4/6 Product unusable by either Government Agencies or Industry
• NASA needs a Cadre of Engineers/System Analysts with a long range goal of becoming the USG Technical Experts in Aviation System Safety Analysis
Proposed Grand Experiment:Proposed Grand Experiment:OPEVAL to FOCUS EffortsOPEVAL to FOCUS Efforts
• FY 2008 One Year of Night Operations – 12pm to 8 am
• DAG-TM + aFAST+CDM + WV
• Entire US Air Cargo Fleet
• Inter-Agency IPT– DoT, NASA, FAA, DoD, NTSB, Boeing, CAA,
Airlines
Hypothesis: Most Major Changes to the Hypothesis: Most Major Changes to the NAS have been due to NAS have been due to Safety ConcernsSafety Concerns
• 1960’s Mandated Introduction of Radar Separation
• 1970’s Decrease in Oceanic Separation Standards Required a Landmark Safety Analysis
• 1970’s Required A/C Transponder Equipage
• 1970’s Required A/C Ground Proximity Equipage
• 1990’s Required A/C TCAS Equipage
• 1990’s Required A/C Enhanced Ground Prox. Equipage
• 1990’s TDWR & ITWS Introduction
• 1990’s Mandated Development of GPS/WAAS