FAA CENTER OF EXCELLENCE FOR ALTERNATIVE JET FUELS & ENVIRONMENT Takeoff/Climb Analysis to Support AEDT APM Development Project 45 Project manager: Bill He, FAA Lead investigator: Michelle Kirby, Georgia Institute of Technology Presenter: Matthew J. LeVine, Georgia Institute of Technology April 18 th & 19 th, 2016 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.
Introduction Accurate modeling of aircraft performance is a key factor in estimating aircraft noise, emissions and fuel burn Various assumptions are made for aircraft performance modeling (APM) within the AEDT with respect to: Aircraft load factor Takeoff weight Departure flight profiles, which model maximum engine thrust at takeoff The main objectives of this research are 1. Identify prior relevant research methods and benchmark the current APM assumptions 2. Conduct statistical analysis of real-world performance data 3. Develop a state estimator 4. Document recommendations for APM enhancements 2
Practical Outcomes Short term Assessment of current modeling assumptions within the APM Identification of modeling gaps to real world flight Identification of necessary flight data to represent real world flight Statistical analysis of real flight data Sensitivity investigation of modeling assumptions, including fuel burn, NOx, and noise Long term Recommendations for new algorithm to mimic real world takeoff performance Documentation of sensitivity analysis and implications of modifications to the procedures for the APM 3
Schedule and Status Task 1: Literature review - Completed ASCENT Project 35 Functional relationships between mission range and takeoff weights established for a subset of vehicles Observations of amount of reduced thrust, typically 15% ACRP 02-41: Estimating Takeoff Thrust Settings for Airport Emissions Inventories Reduced takeoff thrust analysis of airline data, typically 15% Takeoff Thrust-Setting Estimator Tool (TTREAT) AEDT APM module algorithms FAA AC 91-53A and ICAO PANS OPS Chapter 3 Volume II: Recommend that all carriers adopt no more than two procedures for each aircraft type; one for noise abatement of communities close to the airport and one for noise abatement of communities far from the airport AC 91-53A: Close-In and Distant Procedure ICAO Pans Ops: NADP1 and NADP 2 Task 2: Statistical Analysis of Flight Data (where available) and Modeling Data In progress and focus of today Task 3: Development of Aircraft State Estimators June 2017 Task 4: Develop APM Enhancement Recommendations Aug 2017 4
How Does It All Fit Together? Today ASCENT 35 Estimate Reduced Thrust Takeoff Refine Takeoff Weight Assumptions ASCENT 45 is closely linked to ASCENT 35 Regular correspondence with A35 team Impact of departure procedures in AEDT may have interactions with A35 findings on departure weight and thrust Also considering how to implement A35 findings directly into AEDT ASCENT 45 Model More Representative Departure Procedures ASCENT 43 Determine Impact of NPDC AEDT Projects shown are focused on AEDT noise predictions, not meant to exclude other ASCENT funded work supporting AEDT ASCENT 23 More accurate departure and arrival Noise estimates from adv. procedures *MIT 5
Quantifying the Impact of Takeoff Assumptions within AEDT Isolate the impact on terminal area performance due to changes in takeoff assumptions within the airplane performance module (APM) Weight, thrust, and departure procedure Establish the partial derivative impact due to these assumptions Weight: Utilize AEDT representative stage lengths within each bin and compare AEDT weight assumptions to Project 35 weights, which are heavier for takeoff Run existing departure procedures within AEDT; quantify environmental footprint Thrust: Utilize AEDT representative stage lengths within each bin for the AEDT weight assumptions with a 15% reduced thrust Run existing departure procedures within AEDT; quantify environmental footprint Procedure: Cannot utilize AEDT assumptions due to the fact that the PROFILE parameter definitions are not known as a function of TO weight and thrust Must use EDS/FLOPS or High Fidelity Validation data Finally, test the interactions of these assumptions 6
Preliminary Investigation of AEDT Thrust Assumption Looked at impact of reduced thrust takeoff Reduced COEFF_E in latter equation Real thrust reduction typically accomplished via Assumed Temperature method In reality, should adjust parameters in procedure definition as well NOTE: Highlighted contours = RTT 7
Preliminary Investigation of AEDT Weight Assumption Origin Destination Stage Length Range [nm] AEDT Assumed Weight [lb] Project 35 Calculated Range [nm] Weight [lb] SFO SAN 1 350 133300 388 141198 BOS ATL 2 850 139200 822 146539 ATL BOS 2 850 139200 822 147830 DEN JFK 3 1350 145500 1413 158522 BOS LAX 4 2200 156700 2269 168070 Takeoff Weight (lbs) 180000 170000 160000 150000 140000 130000 B737-800 Takeoff Weight Comparison Project 35 Outcomes B737-800 Actual B737-800 Actual B737-800 AEDT B737-800 AEDT AEDT assumes a take-off weight for the aircraft per stage length of the mission Actual weights are primarily a function of great circle distance Two missions with the same range can have different weights based on payload and departure atmospheric conditions AEDT discrete stage length assumption tends to underpredict takeoff weight 120000 0 500 1000 1500 2000 2500 3000 Great Circle Distance (nm) 8
Impact of AEDT Weight Assumption on STANDARD procedures Sample results show impact of AEDT weight assumption for STANDARD procedures Project 35 projected a higher weight for this flight distance Weight impacts ground roll length, initial climb velocity, etc. Combination of weight assumption and reduced thrust will lead to further differences in departure contours and terminal area fuel burn Note that STANDARD procedure for this aircraft DOES NOT match current Noise Abatement Departure Procedures in place at Boston Logan airport 9
NADP Procedures from CAEP/7-WP/25 NADP1 ICAO-A ICAO-B NADP2 Note that NADP procedures resemble but don t necessarily match ICAO procedures ICAO-A procedure matches Procedure 2 in the table Procedure 1 is almost identical but with earlier cutback NADP1 procedures are identical to ICAO-A, but different cutback altitudes are possible ICAO-B procedure matches Procedure 3 in the table Procedure 4 represents an NADP2 procedure (although other cutback altitudes are possible) Primary differentiation between NADP2 and ICAO-B is cutback occurring respectively before or after flap retraction steps Many aircraft in AEDT database use STANDARD procedure that is identical to ICAO-B Informal discussion with an airline pilot suggests NADP2 is actually most common procedure used at most airports, typically with thrust cutback at 1000-ft 10
Process for Validation of NADP Modeling Phase 1: FLOPS Phase 2: Calibrate aerodynamics Match takeoff weight and thrust Validate procedures o Trajectory o Velocity Profile o Flight Path Angle HFVD Assume Standard Day Sea-Level Translate Trajectories o Accelerated Climb Rates o Transition Velocities AEDT PROCEDURE PARAMETERIZATION Phase 3: FLOPS Compare SEL Contours Validate procedures o Trajectory o Velocity Profile o Thrust Profile AEDT TESTER HFVD 11
FLOPS Validation Sample: NADP-1 DoE NADP-1 Sample Design of Experiments (1000 runs) Validation velocity and trajectory data was captured in the DoE ranges 12
FLOPS Validation Sample: Best Cases Best overall DoE case for each NADP Solution that comprised of the lowest trajectory, speed, and flight path angle RMS errors RMS Errors for the Best Overall DoE Case Ground Distance [nm] Speed [kts] Flight Path Angle [deg] NADP-1 0.217 1.249 0.847 NADP-2 0.194 1.570 0.801 13
Impact of Modeling Different Procedures in AEDT Contours, trajectories, and terminal area fuel burn reflect impact of modeling different procedures given the AEDT assumed weight and thrust Notable differences in contours due to the fact that STANDARD procedure features later thrust cutback Primary differences between NADP1 and NADP2 observable in trajectory plots below 3000-ft NADP1 and NADP2 exchange terminal area fuel burn (and emissions) above or below 3000-ft 14
Summary Summary statement Current procedures in AEDT do not match real world conditions for departure procedures Combination of better weight estimates, reduced thrust, and modeling of current Noise Abatement Departure Procedures will yield more realistic noise and emissions results Results of this research will provide better understanding of the combined impacts of these factors Next steps? Explore interactions of three partial derivatives Comparison of AEDT profiles versus high fidelity data for actual airport altitude and atmospheric conditions Repeat process for other aircraft (Small Twin Aisle, Large Twin Aisle, Regional Jet, Large Quad) Key challenges/barriers Access to real flight data and other validation data Iteration/automation of validation process 15
References Global and Regional Environmental Aviation Tradeoff (GREAT) CO2 Emission Metrics for Commercial Aircraft Certification: A National Airspace System Perspective, A PARTNER Project 30 Findings Report, NO. PARTNER-COE-2012-002 Airport Noise Grid Interpolation Method (ANGIM) Bernardo, Kirby, & Mavris, Development of a Rapid Fleet-Level Noise Computation Model, AIAA Journal of Aircraft, Nov. 2014 ASCENT Project 35: Airline Flight Data Examination to Improve flight Performance Modeling ACRP 02-37: Integrated Noise Model Accuracy for General Aviation Aircraft ACRP 02-41: Estimating Takeoff Thrust Settings for Airport Emissions Inventories ACRP 02-55: Enhanced AEDT Modeling of Aircraft Arrival and Departure Profiles Contributors Staff: Prof. Dimitri Mavris (PI), Dr. Michelle R. Kirby (Co-PI), Dr. Matthew LeVine, Dr. Yongchang Li, Dr. Dongwook Lim, Dr. Holger Pfaender, Mr. Chris Perullo, Prof. JP Clarke, Mr. Jim Brooks FAA-AEE and Volpe: Bill He (PM), Joseph DiPardo, Dr. Mohammed Majeed, AEDT Development Team, Dave Senzig (Volpe) Graduate Students: Ameya Behere, Dylan Monteiro, Vu Ngo 16