Upload
others
View
8
Download
0
Embed Size (px)
Citation preview
FAA CENTER OF EXCELLENCE FOR ALTERNATIVE JET FUELS & ENVIRONMENT
Project manager: Hua He and Aniel Jardines, FAA Lead investigators: Hamsa Balakrishnan (MIT) and Tom Reynolds (MIT LL)
Surface Analysis to Support AEDT Aircraft Performance Module Development
Project 46
April 18-19, 2017 Alexandria, VA
Opinions, findings, conclusions and recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of ASCENT sponsor organizations.
2
Motivation
• Taxi phase in the Aviation Environmental Design Tool (AEDT) is currently modeled using default or user-specified taxi times, coupled with engine idle fuel and emissions assumptions from ICAO Aircraft Engine Emissions Databank
• These assumptions reduce the accuracy of the taxi performance modeling
• For some applications, higher fidelity surface modeling approaches may be needed
3
Objectives
• Identify needs and evaluate methods for improving taxi performance modeling in AEDT in order to better reflect actual operations
• Develop and validate enhanced taxi models by combining ASDE-X surface track data with statistical models of engine performance (using Flight Data Recorder information)
• Make recommendations for enhancements to AEDT Aircraft Performance Module (APM) based on knowledge gained
4
Outcomes and Practical Applications
AEDT APM
Total fuel burn
per time-period, per airport
Fuel flow rate profiles
ASDE-X data
Current scope
5
Outcomes and Practical Applications
AEDT APM
Total fuel burn, noise and emissions
per time-period, per airport
Fuel flow rate profiles
Thrust profiles
ASDE-X data
Potential extensions
6
Approach
AEDT APM
surface model needs
assessment
Aircraft surface
performance modeling
enhancements
Aircraft taxi performance
model validation
AEDT APM
enhancement recommendations
Stakeholder input,
supporting documents & prior research
ASDE-X data
FDR data
Model development data Validation data
7
Schedule and Status
• Assess AEDT aircraft surface performance modeling needs
• Develop enhanced aircraft surface performance models
• Validate enhanced aircraft surface performance models
• Recommend AEDT APM enhancements
[Sept.-Nov. 2016]
[Oct. 2016-present]
[Jan. 2017-present]
[~Aug. 2017]
8
Recent Accomplishments [1]
• Assessment of AEDT surface APM modeling needs – Synthesis of findings from AEDT documentation, ACRP studies
02-45 and 02-27, and stakeholder input • Airport average taxi-time is either default (7 min taxi-in/ 19 min taxi-out)
or airport-specific for 75 airports – Default taxi database is outdated – Analysis in ACRP 02-45 suggests 16 min for taxi-out/ 7 min for taxi-in – Airport specific taxi-time estimates off by 4-50% [recent ASPM data]
• Taxiway network model of airport; taxi-path for each flight – Taxi-time based on taxi-path length and speed (AEDT default: 15 knots)
» ACRP report suggests ~13 knots – Taxi-in time is considered to be unimpeded; taxi-out time includes a queuing
delay at the runway. Queuing delay is sum of runway occupancy times of aircraft already in the queue
» Queues are assumed only at the runway; effects of acceleration not included
» Deterministic runway occupancy times. Variation by aircraft type/weather condition not known
• No surface-specific regression model for fuel flow rates
9
Recent Accomplishments [2]
• Assessment of AEDT surface APM modeling needs
• Identified needs: – Models that are representative of a wide range of taxi
conditions, aircraft types, airports, airlines, and weather conditions
– Identify key locations and events at major airports (e.g., non-movement area, taxiway intersections, spot and departure runway queues, runway crossings, etc.)
– Need to model operational variability • Assess uncertainty associated with fuel burn estimates • Evaluate sensitivity to various factors (e.g., takeoff mass,
ambient conditions) – Need data-driven validation of models
10
Recent Accomplishments [3]
• Surface APM model enhancements – In order to determine the “full” fuel burn
profile, need to consider both the pre-ASDE-X (frequently the non-movement area) portion and the ASDE-X portion of surface operations
– The pre-ASDE-X portion may include engine start and ramp area movements
– Can vary by airport; needs characterization by airport
– For the ASDE-X portion, prior work suggests that the taxi time and number of acceleration events are significant
Extract taxi time & # of
accelerations Fuel burn estimate
Regression model
ASDE-X
Gate
Enginestart
Taxi
Push-back
FDR dataASDE-X data
Pre-ASDE-X contribution
11
Recent Accomplishments [4]
• Initial identification of Non-Movement Area (NMA) & “pre-ASDE-X” area (PAXA) for US airports in FDR data sets of European carrier (BOS, JFK, EWR, MIA, ORD, LAX) – Determine NMA (based on airport markings) & PAXA (based on
ASDE-X data analysis) polygons for each of the US airports
BOS JFK
MIA ORD
NMAPAXA
12
Recent Accomplishments [5]
• Determining PAXA regions
ASDE-X Archives
Select airport, analysis days & plot
data
Generate coverage grid
Identify PAXA polygon based on interior edges of largest cluster
Latit
ude
Longitude
Latit
ude
Longitude
e.g. MIA
e.g., MIA
13 750 800 850 900 950Time (s)
-2
0
2
4
6
8
10
Spee
d (m
/s)
DataFilterSmoother
Recent Accomplishments [6]
• Smoothing/filtering algorithms for FDR and ASDE-X data – Need to “match” events (e.g., turns and accelerations) – Resolution and noise characteristics vary between datasets – Infer ground truth data (equivalent tracks) for algorithm validation
100 200 300 400 500 600 700Time
0
2
4
6
8
10
12
Spee
d (m
/s)
DataFilterSmootherASDE-XFDR
(Not the same flight)
14
Summary
• Development and validation of enhancements to AEDT surface APM
• Next steps – Model range of taxi conditions, aircraft types, airports, airlines, and
weather conditions – Analyze ASDE-X data from major airports to identify key locations
and events (e.g., non-movement area, taxiway intersections, spot and departure runway queues, runway crossings, etc.)
– Model operational variability and evaluate sensitivity – Develop and validate models by synthesizing data from
• 2006 FDR (European carrier) • 2016 FDR (European carrier, including 123 arrivals and 368 departures
at six US airports) • Aggregate (i.e., total over taxi phase) fuel use data from A4A • 2016 ASDE-X data
15
References
• ACRP 02-45, “Methodology to Improve EDMS/AEDT Quantification of Aircraft Taxi/Idle Emissions”, Transportation Research Board, 2016.
• ACRP 02-27, “Aircraft Taxi Noise Database for Airport Noise Modeling”, Transp. Research Board, 2013.
• Chati, Y.S., and Balakrishnan, H., “Analysis of Aircraft Fuel Burn and Emissions in the Landing and Takeoff Cycle using Operational Data," Intl. Conf. on Research in Air Transportation, Istanbul, 2014.
• DOT-VNTSC-FAA-16-11, “Aviation Environmental Design Tool – Technical manual, Version 2b”, U.S Department of Transportation (FAA), 2016.
• H. Khadilkar and H. Balakrishnan, “Estimation of Aircraft Taxi Fuel Consumption using Flight Data Recorder Archives”, Transportation Research Part D: Transport and the Environment, 17, (7), 2012.
• Särkkä, Simo. "Unscented Rauch-Tung-Striebel Smoother." IEEE Transactions on Automatic Control 53.3 (2008): 845-849.
• Lopez, Remy, and Patrick Danes. "Low-Complexity IMM Smoothing for Jump Markov Nonlinear Systems." IEEE Transactions on Aerospace and Electronic Systems (2017).
Contributors • PIs: Hamsa Balakrishnan and Tom Reynolds • MIT Lincoln Laboratory: Emily Clemons • MIT Students: Sandeep Badrinath, Yashovardhan Chati • FAA: Chris Dorbian
16