1. Noise Prediction Modeling I
John-Paul Clarke Nhut Tan Ho Liling Ren
DLR-DLH-MIT Workshop on Noise Abatement Procedures August 17-19, Seeheim, Germany

2. Key Question How accurate do our noise model need to be?
All aircraft versus only aircraft that drive noise?
What is the right modeling approach?
Single event versus extended duration noise levels?
Detailed spectra-directivity versus noise-power-distance curves?
Specific weather effects versus standard conditions?
How could we use use these models in the design of noise abatement procedures?

3. How accurate?
Developing models for all aircraft would require more money than we have
Too many different aircraft types
Prudent approach would be to:
Analyze traffic and noise monitor data from all airports to determine aircraft that are main noise drivers
Conduct flyover measurements to determine noise levels for these aircraft
Data would of course also be used to support development of detailed component models
Develop correlations with other aircraft types and use them to estimate noise levels of other aircraft

4. Modeling approach?
Extended duration noise levels are used throughout the world for deciding on compensation for aircraft noise impact, but …
Knowing the single-event spectral-based weather-affected noise level is critical for developing noise abatement procedures.
Tailor operations to specific conditions (when weather conditions are included) and to specific population distribution
Capture effects of variability in weather and operations on noise impact
Can also be aggregated in post processing to give extended duration measures

5. Source Directivity-Spectra

6. How to use? Evaluate and compare noise impact of candidate procedure
Straight-forward application of modeling
Select best procedure for given conditions from list of pre-evaluated candidates procedures
Initial study for NASA Ames suggests that this could be a useful tool for airport configuration management
No current work being done in this area
Optimize procedure parameters (e.g. targets) given variability in operating environment
Very useful for developing pilot procedures

7. Effects of Weather
Models generally ignore the effects of atmospheric gradients
Airports assume ‘no weather’ for annual average noise levels
The validity of this assumption was tested through a numerical study using the aircraft noise simulation model NMSIM [WYLE]
Five years of meteorological data at 7 major US airports
Model airport
With the compliments of

8. Wyle Airport Noise Study
Five years of weather at seven airport locations
Upper air soundings – twice daily
Surface weather - hourly
Surface flux model
Generate full profiles for each hour
Model airport
Composite of major air carrier airports
Three runways, various directions
Four nominal aircraft types
Hourly operations based on annual operations at component airports
Tracks straight in and out
With the compliments of

9. No-Weather Leq (65 dB and higher)
Leq contours for no gradients – the usual assumption.
With the compliments of

10. U.S. Standard Atmosphere
See how the shape differs – narrower alongside runways, where upward refraction causes shadow zones.
With the compliments of

11. Sample Hour
With the compliments of

12. Annual average Leq contours
Areas in square miles
With the compliments of

13. MIT Study
MIT study (main campus and Lincoln Laboratory) of two airports agreed with Wyle study
Significant impact due to weather
Study went further to include effects of weather on aircraft performance

14. Logan Case Study Low Noise Departure Existing Departure North wind
Scenarios/ dBA SEL
60-70
70-80
80-90
90-100
Above 100
Existing departure
405,885
235,083
43,825
5,764
192
Weather-specific dep.
293,745
94,574
18,882
5,320
143

15. Fast-time Simulators (1)
Functionality
Evaluation aircraft deceleration and maneuverability
Simulation of noise abatement approach procedures (example: decelerate at idle thrust, 3º slope angle)
Monte-carlo simulation studies for capacity and noise impact
General features
Actual aerodynamic models
Aircraft : B , B , B , B
Control system functions: auto-pilot and navigation control logic
Modes: level flight, level turn, climb/airspeed, 3º slope angle hold
Models of pilot response, wind, disturbances
Coded in Matlab

16. Fast-time Simulators (2)
Point-mass modeling assumptions
Aircraft behaves as a point mass (no rotation, inertia)
No slide-slip {b = 0, v = 0}, coordinated turn
Thrust is along the body axis
Controls pilot has goes directly to applied forces
Other body rates (q & r) are determined by curvature of trajectory i.e. cannot pitch without climbing

17. Fast-time Simulators (3)
Point-mass simulation loop

18. Integrated Tool (1) Integration of simulators with Noise Model
Automated to perform repeated computation
A/C
States
Simulator input:
A/C Initial conditions
Pilot Response models
Wind models
Noise
Model
Fast-time
Simulator

19. Integrated Tool (2)
Tradeoff study: Noise Reduction and Target Achievement
Pilot response models
Approaches
PF Cues PNF to extend flap
PF cues PNF to
extend flap in turbulence

20. Integrated Tool (3)
Highest noise saving and highest target achievement for
Non-turbulent pilot response model: Flaps schedule is designed for 1050 ft and HI equal 5000 ft
Turbulent pilot response model: Flaps schedule is designed for 1200 ft and HI equal 7000 ft
To design for worst case (i.e. turbulence): high value of HI provides best noise saving and target achievement probability
Non-turbulent pilot response model
Turbulent pilot response model