sintef report · sintef report title aircraft noise measurements at gardermoen airport, 2001. part...

SINTEF REPORT

TITLE

Aircraft Noise Measurements at Gardermoen Airport, 2001. Part 1: Summary of results.

AUTHOR(S)

S. Å. Storeheier, R. T. Randeberg, I. L. N. Granøien, H. Olsen, A. Ustad CLIENT(S)

SINTEF Telecom and Informatics Address: NO-7465 Trondheim NORWAY Location Trondheim: S.P. Andersens v 15 Location Oslo: Forskningsveien 1 Telephone: +47 73 59 30 00 Fax: +47 73 59 43 02 Enterprise No.: NO 948 007 029 MVA

Norwegian Air Traffic and Airport Management1, Oslo Airport AS2, Norwegian Defence Construction Service3.

REPORT NO. CLASSIFICATION CLIENTS REF.

STF40 A02032 Unrestricted K. H. Liasjø1, K. Holen2, N. I. Nilsen3 CLASS. THIS PAGE ISBN PROJECT NO. NO. OF PAGES/APPENDICES

Unrestricted 82-14-02519-2 403129 55/3 ELECTRONIC FILE CODE PROJECT MANAGER (NAME, SIGN.) CHECKED BY (NAME, SIGN.)

STF40-A02032.doc Svein Å. Storeheier Idar L. N. Granøien FILE CODE DATE APPROVED BY (NAME, POSITION, SIGN.)

2002-06-05 Odd Kr. Ø. Pettersen, Research Director ABSTRACT

In June of 2001, SINTEF Telecom and Informatics, with assistance from Norwegian Air Traffic and Airport Management, Oslo Airport AS and Norwegian Defence Construction Service, carried out an aircraft noise measurement study at Oslo Airport Gardermoen. The main objective of the study was to investigate the causes of observed differences between predicted and measured aircraft noise levels around the airport. The measurements were conducted on scheduled air traffic, at 5 measurement positions reasonably close to the airport area. Efforts were made to collect and synchronize relevant aircraft data with the acoustic measurements. An essential contributing factor to this was the access to flight recorder data from the airline company SAS for the monitored flights. The number of selected flights in this study totals 155, including both take-off and landing operations. Part 1 of this report will overview the measurement layout, including data collection and data processing routines. The main results relate to noise source directivity, differences between estimated and predicted lateral attenuation and aircraft noise emission (Noise-Power-Distance) data. The differences are discussed, and possible measures for correcting the aircraft noise prediction routines are recommended. Report Part 2 is released in additional volume and will include all technical memos produced during the study.

KEYWORDS ENGLISH NORWEGIAN

GROUP 1 Acoustics Akustikk GROUP 2 Noise Støy SELECTED BY AUTHOR Aircraft Flytrafikk Noise level prediction Støyberegning

2

PREFACE This report and the underlying investigation was managed by SINTEF Telecom and Informatics as contract research for a joint client group including Norwegian Air Traffic and Airport Management, Oslo Airport AS, and Norwegian Defence Construction Service. Flight recorder data for a number of flights were collected through arrangements made by Scandinavian Airline System (SAS). Oslo Airport AS and Norwegian Air Traffic and Airport Management made practical arrangements (transport, admission to airport area facilities, etc.). The client group participated during the field investigation by providing equipment and personnel for the acoustical measurements, and also in the collection of flight recorder data from SAS. SUMMARY

In June of 2001, SINTEF Telecom and Informatics, with assistance from Norwegian Air Traffic and Airport Management, Oslo Airport AS and Norwegian Defence Construction Service, carried out an aircraft noise measurement study at Oslo Airport Gardermoen (OSL). The main objective of the study was to investigate the causes of observed differences between predicted and measured aircraft noise levels around the airport. The measurements were conducted on scheduled air traffic, at 5 measurement positions reasonably close to the airport area. Efforts were made to collect and synchronize relevant aircraft data with the acoustic measurements. A contributing factor to this was the access to flight recorder data for a significant part of the monitored flights. The number of flights in this study totals 155, including both take-off and landing operations. An overview of the measurement program, data collection and data handling procedures are given in chapters 2 and 3 respectively. The main results are summarised and commented upon in chapter 4, major conclusions are given in chapter 5. Possible measures for correcting the aircraft noise prediction routines in short, intermediate and long terms are recommended in chapter 6. The main results are: 1. The basic assumption, used by all fixed wing aircraft noise models, that the aircraft noise

emission is cylindrical symmetric is wrong and need to be corrected for in next generation models.

2. The differences between observed and recommended lateral attenuation are significant, the order of magnitude have influence on prediction accuracy.

3. The differences between observed and in use Noise-Power-Distance data are statistically significant.

4. Flight profiles are different from the standard profiles used by the noise models. 5. The results indicated in this study are in broad accordance with recent findings in international

investigations. 6. The noise measured by the NTMS can be expected less than 1 dB higher relative to calculations,

because measurements take place at 6 meter height and calculation are done for 1.5 meter.

3

CONTENTS

PREFACE ..................................................................................................................................................................... 2 SUMMARY .................................................................................................................................................................. 2

1 INTRODUCTION .................................................................................................................................................. 5

1.1 BACKGROUND ...................................................................................................................................................... 5 1.2 OBJECTIVES.......................................................................................................................................................... 5

2 MEASUREMENT PROGRAM ............................................................................................................................ 6

2.1 MEASUREMENT SITE AND MEASUREMENT POSITIONS ........................................................................................... 6 2.2 ACOUSTICAL INSTRUMENTATION ......................................................................................................................... 6 2.3 MEASUREMENT SYNCHRONISATION ..................................................................................................................... 6 2.4 AIRCRAFT TRACKING ........................................................................................................................................... 6 2.5 METEOROLOGICAL AND GROUND TYPE DATA ..................................................................................................... 8 2.6 MEASUREMENT PROCEDURES .............................................................................................................................. 8

3 DATA REDUCTION AND PROCESSING, OVERVIEW................................................................................. 9

3.1 TIME-SPACE-POSITION ANALYSIS ........................................................................................................................ 9 3.2 NET THRUST DATA................................................................................................................................................ 9 3.3 NOISE EVENT EXTRACTION – EXCLUSION OF BACKGROUND NOISE INFLUENCE..................................................... 9 3.4 SEL CALCULATION............................................................................................................................................. 10 3.5 AIRCRAFT NOISE SPECTRA .................................................................................................................................. 11 3.6 METEOROLOGICAL AND GROUND TYPE DATA.................................................................................................... 11

4 DATA ANALYSIS................................................................................................................................................ 12

4.1 NOISE SOURCE DIRECTIVITY ............................................................................................................................... 12 4.1.1 Analysis tools ........................................................................................................................................... 12 4.1.2 Backward propagation parameters.......................................................................................................... 13 4.1.3 Data filtering............................................................................................................................................ 14 4.1.4 Directivity results ..................................................................................................................................... 14 4.1.5 Comments................................................................................................................................................. 21

4.2 LATERAL ATTENUATION .................................................................................................................................... 21 4.2.1 Lateral attenuation as measured.............................................................................................................. 21 4.2.2 Simulated lateral attenuation ................................................................................................................... 23

4.3 NOISE-POWER-DISTANCE DATA ......................................................................................................................... 25 4.3.1 Analysis method ....................................................................................................................................... 25 4.3.2 Results ...................................................................................................................................................... 25 4.3.3 Statistical analysis.................................................................................................................................... 28 4.3.4 Comments................................................................................................................................................. 29

4.4 COMPARISON OF ACTUAL FLIGHT PROFILES WITH DATABASE ............................................................................. 29 4.4.1 Examples of results................................................................................................................................... 29 4.4.2 Comments on database and flight recorder profile comparison .............................................................. 31

4.5 INFLUENCE OF METEOROLOGY ON AIR ABSORPTION ........................................................................................... 31 4.6 QUALITY CHECK ON THE NOISE AND FLIGHT TRACKING SYSTEM AT OSL........................................................... 32

4.6.1 Evaluation of microphone positions of the airport monitoring system..................................................... 33 4.6.2 Noise levels measured by the airport noise monitoring installation ........................................................ 35 4.6.3 Comparison of tracks and profiles from flight recorder with NTMS ....................................................... 36 4.6.4 Quality check on the monitoring system aircraft log................................................................................ 40

5 CONCLUSIONS................................................................................................................................................... 41

5.1 DIRECTIVITY ...................................................................................................................................................... 41 5.2 LATERAL ATTENUATION..................................................................................................................................... 41 5.3 NOISE POWER DISTANCE DATA .......................................................................................................................... 41 5.4 FLIGHT PROFILES ................................................................................................................................................ 41

4

6 RECOMMENDATIONS FOR FURTHER WORK.......................................................................................... 43

6.1 SHORT TERM MEASURES ..................................................................................................................................... 43 6.1.1 Correcting lateral attenuation.................................................................................................................. 43 6.1.2 Corrections to Noise-Power-Distance data ............................................................................................. 43 6.1.3 Adjustment to flight profiles ..................................................................................................................... 43

6.2 INTERMEDIATE TERM MEASURES........................................................................................................................ 43 6.2.1 Flight profiles from the monitoring system .............................................................................................. 43 6.2.2 Investigate effects of flight path segmentation ......................................................................................... 44

6.3 LONG TERM MEASURES ...................................................................................................................................... 44

7 REFERENCES ..................................................................................................................................................... 45

APPENDIX 1 MEASUREMENT SITE AND MEASUREMENT POSITIONS........................................................................... 46 APPENDIX 2 METEOROLOGICAL DATA ..................................................................................................................... 51 APPENDIX 3 ESTIMATION OF GROUND IMPEDANCE BY ACOUSTICAL MEASUREMENT. .............................................. 55

5

1 Introduction

1.1 Background Since 1994, aircraft noise exposure near Norwegian airports has been calculated by the computer program NORTIM. The program is in accordance with international methods for calculating aircraft noise. Adjustments have been made to incorporate the topography in the calculations. The accuracy of NORTIM has previously been justified from comparison with other international calculation methods, and with measurements at Fornebu Airport in 1989 [15,16]. The conclusion was then that measured aircraft noise level is on average 0.5 dB lower than calculations with NORTIM, with a standard deviation of 1.9 dB. A special version of NORTIM has been developed for use at the Norwegian main airport OSL. The program, called GMTIM, calculates the aircraft noise exposure automatically a posteriori, based on the registered traffic and the flight tracks measured by radar. Both NORTIM and GMTIM use the INM database with respect to noise1 and climb profiles. As a part of the validation of GMTIM, SINTEF Telecom and Informatics has investigated possible causes of deviation between calculated aircraft noise near OSL, and measurement results obtained by the noise and track monitoring system (NTSM) at the airport [1,2]. The conclusion of the investigation, based on measurements in 6 positions, using nearly 70000 noise events, was that the measured aircraft noise level is on average 3.1 dB higher than calculated. The airport noise monitoring system has microphones at different heights above ground (5-8 meters) as opposed to GMTIM that calculates for 1.5 meter above ground. When this is corrected for, the deviation between measurements and calculations has mean values 2.0 dB for MFN (maximum noise level) and 2.7 dB for EFN (equivalent noise level). The deviations are statistically significant. This experience is broadly in line with results from recent investigations [10,11] involving the use of SAE recommendations [5,6] and the INM database.

1.2 Objectives The project had two main goals: i) To determine the main factors that cause the deviations between measured and calculated

noise levels, and ii) To prescribe possible actions to reduce or eliminate the deviations. A major part of the investigation was to be based on specially designed noise measurements at OSL. The objective of the present report (Part 1) is to give an overview of the noise measurement program including data collecting and data processing routines. In addition this report presents main results from the investigation. A number of technical memos with detailed results are collected in Part 2 of the report in a separate volume.

1 Noise data for MD80 class aircraft are adjusted by +1.5 dB for SEL and LAMAX for take-off thrust levels.

6

2 Measurement Program

2.1 Measurement site and measurement positions The site for the noise measurements was located to the north and east of the western runway system, as shown in Appendix 1. The measurement positions are numbered 1 - 5. Position 1 is situated inside the airport area, directly below the flight path and 400 m from the runway end. Position 2 is also in line with the runway, just outside the airport area at a distance of 1150 m from the runway end. Positions 2 - 5 are situated on a straight line slightly tilted to the perpendicular to the runway extension. The distances from position 2 are 400 m, 800 m and 1250 m for positions 3, 4 and 5 respectively. The terrain along the line of measurement positions is relatively flat, although smooth terrain undulation occurs causing terrain height variations of a few meters. The ground type in the area is mainly sandy soil, with varying extent of grass coverage. Position 2 is situated at the border of a spruce forest area. Position 5 is placed in an area sparsely covered by tall spruce or deciduous trees. Else, the area of the measurement positions is open land. An agreement with the Airline Company SAS was vital to the program to attain flight-recorder data from selected flights. The SAS operations at OSL mainly consist of narrow-body aircraft of the MD 80 family, “Next Generation” B737s and a few MD90s.

2.2 Acoustical Instrumentation The measurement positions 2 - 5 all had microphones at two heights. The lower microphone was put at 1.5 m above ground. The higher microphone height varied between 6.5 - 10.5 m above ground. At measurement position 1 only one microphone was used. This microphone was attached to the runway approach light bar arrangement, approximately 6.5 m above ground. The acoustic measurement equipment consisted of condenser microphones (1/2" with standard 9 cm diameter windscreen) connected to Nor-121 Environmental Analysers manufactured by Norsonic AS. Instrumentation is described in detail in Part 2 of the report. A schematic layout is shown in Appendix 1 of this report. The instruments were pre-set to measure LEQ and LMAX(FAST) in 0.25-second intervals in 1/3-octave frequency bands 25 - 10,000 Hz and both A-weighted and C-weighted. The data were stored electronically at each position and at the end of every session the data were transferred to a computer and recorded to CD-ROM.

2.3 Measurement synchronisation For synchronisation purposes, the measurement instruments were connected to a PC with specially designed software for generating trigger signals and reading/storing of trigger signal time from an external GPS clock. Three GPS clocks were used; one for position no. 1, one for position number 2 - 5, and one for synchronisation of aircraft movements on the runway.

2.4 Aircraft Tracking The task was to determine the position of an aircraft at a certain moment of time during the noise measurement. Additionally, other flight parameters such as velocity and thrust level had to be synchronised with the noise measurements. Data from the airport radar tracking system was available, as well as flight-recorder data from selected flights. Below is an outline of the procedures that were investigated.

7

1. Synchronising the airport radar tracking system with the noise measurement system.

In measurement position 1 the microphone was placed close to one of the microphones of the airport noise and track monitoring system. The latter microphone is connected to the radar tracking system. By comparing acoustic signatures for coincident flights in the two systems, any time-shift between them could be found. The time indication of the airport radar tracking system was then adjusted to match the time of the noise measurement system. Each measurement session (see section 2.6) was adjusted separately.

2. Synchronising aircraft flight recorder time with the noise measurement system (I). The times for touchdown (or lift-off) of nose wheel and main wheels were manually clocked with a GPS, and correlated with details in the aircraft flight recorder. Out of several possibilities, the Air/Ground indicator in the flight recorder was chosen and correlated with the time of nose wheel touch down or lift-off. The time resolution for this parameter in the flight records is ± 1 second for the 737s and MD90s and ± 2 seconds for the MD80s. In addition there will be some uncertainty due to the manual operation of the GPS.

3. Position calculated from speed and direction. The GPS clock used in (2) was also used to determine the time when the aircraft passed a known reference position on the runway. The aircraft time was synchronised with the measurements as described in (2). With basis in the reference position, the aircraft position could be calculated from the aircraft speed and direction given in the flight recorder. This method was however abandoned due to large deviations in the position compared to the other methods described below.

4. Offset of flight recorder GPS position. The 737s and the MD90s are equipped with a GPS that records the aircraft's position. It was found that the recorded position from the aircraft while passing (presumably) straight over measurement position 1, was 30 m offset due west. The same offset was seen relative to the runway. Accordingly, all subsequent aircraft GPS positions were adjusted 30 m east.

5. Offset of airport radar tracking system. The airport radar tracking system gives the aircraft position vs. time, with a time resolution of 1 second. Through an iterative process (ref. point 6), the radar data was time synchronised with flight recorder time. By then comparing the positions from radar with flight recorder positions, an average west offset was found of 140 m (median for positions with offset less than 500m). The north offset was not statistically significant, and were assumed to be zero.

6. Synchronising aircraft flight recorder time with the noise measurement system (II). The lateral distance from measurement position 2 to the aircraft was calculated from positions given by the aircraft GPS and the airport radar tracking system. The distances were plotted as function of time. By manually adjusting the aircraft time, a time synchronisation better than described in (2) could be obtained.

7. Adjusting pressure height. The pressure height taken from the flight recorder was used to determine the height of the aircraft. This value had to be adjusted, by defining the pressure height at the runway to be zero.

8. Correcting for magnetic deviance. The heading indicated in the flight recorder data, was adjusted 0.5 degrees east.

For the MD80s, the position vs. time were taken from the airport radar tracking system, after time synchronising as described in (1) and adjusting the position with the average offset found in (5). All

8

other flight parameters were taken from the flight recorder after time synchronising as described in (2), and correcting as described in (7) and (8). For the 737s and MD90s, the position vs. time were taken from the GPS position data in the flight recorder, after time synchronising and position adjustment as described in (2), (4) and (6). All other flight parameters were taken from the flight recorder data after time synchronising as described in (2), and correcting as described in (7) and (8).

2.5 Meteorological and Ground Type Data A 10 meters high mast with meteorological equipment was erected at position no. 4 as shown in Appendix 2. The following parameters were recorded every minute:

- Temperatures (°C) at two heights, 0.5 m and 9.25 m above ground, - Wind velocity and wind gust (m/s), wind direction, all at 9.25 m height, - Relative humidity (%) at 9.25 m height, - Precipitation (rainfall) (mm) - Net radiation (W/m2)

The net radiation was measured in an attempt to estimate the relative cloud cover in the sky. All data are average values, except for wind gusts that represent maximum values. Graphical presentations are presented in Appendix 2. The acoustic impedance of the ground surface was measured at the measurement positions except for position no. 1. A typical measurement layout is shown in Appendix 3.

2.6 Measurement Procedures The noise measurements were carried out during June 20 – 27, 2001. The measurements were arranged in two separate sessions each day. The sessions lasted for 2.5 – 4 hours. After acoustic calibration the equipment were brought in position and checked for proper operation. The equipment were all battery operated. The measurements were started independently at site I (position 1) and site II (positions 2 – 5) by the site PC giving a trigger signal, and simultaneously recording the GPS time. The measurement of sound levels according to section 2.2 then proceeded continuously through the session. Positions 2 – 5 were manned in order to register events of background noise that might affect the aircraft noise. Personnel situated in the airport control tower registered the aircraft movements as described in section 2.4. The registrations included the GPS time for the special movements along with the flight identification and the arrival/departure time. All the registrations were saved on data files. At the end of the sessions the measurements were stopped manually by the attending personnel. After the last session each day the equipment were collected for acoustic calibration, transferring of data files to a main PC, making backup CDs, etc.

9

3 Data Reduction and Processing, Overview

3.1 Time-Space-Position Analysis The position of the aircraft as function of time was extracted from the airport radar tracking system, the aircraft flight recorder, and manual GPS clock operations, as described in section 2.4. For all aircraft, the position data and the other flight parameters were interpolated to have a regular, 0.5-second step length. There were a total of 155 valid flights, 70 with position taken from the aircraft GPS, and 85 with position taken from the airport radar tracking system. The aircraft position data was given as east/north co-ordinates, and height above runway. Using the co-ordinates of the measurement positions and the height of the microphones above ground, the triplet (east, north, height) was replaced with the triplet (lateral distance, elevation angle, azimuth angle relative to north). In this way the position of the aircraft was expressed in units needed for the following data processing.

3.2 Net thrust data Flight recorder data provide engine thrust settings that has to be transformed into net thrust. Equations for this purpose are found in the SAE AIR 1845 [6]. When the aircraft engines are being operated at thrusts other than rated thrust, the thrust developed is a function of the thrust-setting parameter. When engine pressure ratio (EPR) is available the corrected net thrust per engine is calculated by: )()/( 1

2 EPRKHThGhGFVEF amBACamn +++++=δ (3.1)

Fn is the net thrust per engine, δam is the ratio of the ambient air pressure at the aeroplane to the standard air pressure at mean sea level, Vc is the calibrated airspeed (kts), h is the pressure altitude above sea level, Tam is the ambient air temperature, E, F, GA, GB, K1 and H are constants or coefficients, which are taken from the INM 6.0c database for the relevant operating conditions, EPR, Vc and h are available from the flight-recorder data.

When the thrust is defined by N1, the corrected net thrust is calculated by: 2

13122 )/()/()/( TTamBACamn NKNKHThGhGFVEF θθδ ++++++= (3.2)

N1 is the engine's low pressure rotor speed, θT is the ratio of absolute total air pressure at the engine inlet to the absolute standard air temperature at mean sea level, K2 and K3 are coefficients, which are taken from the INM 6.0c database for the relevant operating conditions.

3.3 Noise event extraction – exclusion of background noise influence The raw noise level measurement results were filtered to determine a background noise level. Background noise level was defined separately for each 1/3-octave band as the level that was exceeded 90% of the time (L90). The background noise level was determined for a 30-minute wide time window, which was moving in 5-minute steps.

10

The personnel at each measurement site registered random noise disturbances like vehicle noise, bird singing, voices, etc. Not all of these possible disturbances had significant sound levels compared to the aircraft noise. Therefore, the time intervals of these disturbances had to be manually defined afterwards, based on the shape of the time history. Both the A-weighted and the 1/3-octave band data were used to find and define the time intervals of the significant disturbances. To locate noise data corresponding to a specific flight, and to eliminate the effects of background noise and noise from preceding/succeeding flights, the following criteria were applied in the order listed (only landing criteria are listed; takeoff criteria were similar): 1. Noise data time window must be inside time window of aircraft position data (the delay of the

sound propagation is accounted for). 2. Noise data must not include time intervals with disturbance. 3. A peak must be found within 60 seconds before touchdown, and minima must be found

between touchdown and 90 seconds before touchdown. 4. Noise level must be higher than 10 dB above the levels at the minima. 5. Noise level must also be higher than 10 dB above the background noise level. 6. The remaining noise data must span at least 10 dB down at either side of peak. The table below shows the number of valid flights after the criteria above were applied to the A-weighted noise data. For the 1/3-octave band data, the number of valid noise events were dependent on frequency.

Table 3.1 Number of valid noise events after applying filtering criteria

Aircraft type Pos. 1 Pos. 2 Pos. 3 Pos. 4 Pos. 5 Total B736 31 20 24 27 30 132 B737 35 23 31 31 27 147 B738 2 1 2 1 1 7 MD81 24 18 17 19 19 97 MD82 48 35 36 38 40 197 MD83 1 1 1 1 1 5 MD87 11 8 9 7 7 42 MD90 3 3 3 3 2 14 Total 155 109 123 127 127 641

The result of the extraction process was data files containing one line per aircraft position and measurement position, at a time resolution of 0.5 second. Each line contained all relevant flight parameters and the corresponding measured sound level, i.e. the 0.5-second LEQ was the energy average of two consecutive 0.25-second LEQ.

3.4 SEL calculation For each flight, the closest point of approach (CPA) was assumed to equal the position where the lateral distance to the measurement position was smallest. At this point, the lateral distance, elevation angle and slant distance was extracted, together with all the other flight parameters. From the noise data the sound exposure level (SEL) was determined according to:

( )∑ ⋅= 10/Leq105.0log10SEL (3.3)

11

The sum is taken over all 0.5-second time steps for the current noise event, which fulfil the criteria listed in section 3.3. In addition the maximum level (MAX) was found, based on 1-second equivalent sound pressure levels derived from the 0.5-second equivalent sound pressure levels.

3.5 Aircraft noise spectra For all aircraft types, average noise spectra were found for takeoff and landing. The spectra were based on SEL values for each 1/3-octave band. Only data from measurement positions 1 and 2 were used. The sound attenuation in air was accounted for by using the ISO 9613-1 standard [4]. A temperature of 20 degrees Celsius and a relative humidity of 60% were used, in accordance with measured average weather conditions. The distance used for the sound attenuation was the height of the aircraft above the microphone at CPA.

3.6 Meteorological and Ground Type data The meteorological data are already suitably processed by recording values for subsequent 1 minute intervals. This is assumed to be sufficient to link meteorological parameters to single flights. As can be seen from the accumulated weather description in Appendix 2, wind speeds were mostly low (i.e. wind speeds well below 5 m/s). The wind direction was mostly due North or South, which means minor components of down- or up-wind at the measurement positions. The temperature gradient revealed prevailing negative values, corresponding to summer conditions with varying sunlight brightness. The acoustic impedance of the ground surface was assumed to follow the Delaney/Bazley model. The important parameter is the flow resistivity (Pa s/m2) which was determined from acoustic measurements according to Nordtest standard [3], for a number of test areas at the measurement positions. The main result is the average flow resistivity relevant to positions 2 - 5, as given in Table 3.2.

Table 3.2 Measured Flow Resistivity (kPa s/m2)

Pos. no. 2 Pos. no. 3 Pos. no. 4 Pos. no. 5 400 250 250 - 400 100

12

4 Data Analysis The data analysed in this chapter are all based on the 155 valid flights shown in Table 3.1.

4.1 Noise source directivity This chapter presents noise source directivity2 of the measured aircraft types separated for each 1/3-octave frequency band from 25 to 10,000 Hz. The directivity is calculated from the measured data, using a propagation model that eliminates the influence of distance, ground surface and weather, thus giving the actual sound emission from the aircraft. Part 2 of the report includes a memo with all results. This chapter only highlights a few of the major findings.

4.1.1 Analysis tools The sound pressure level of every measurement sample (2 per second) is propagated back to the source, using the model Nord 2000 [12]. This is a modern ray-tracing model, which accounts for the influence of topography, ground impedance, wind- and temperature gradients, and turbulence effects in the atmosphere. The calculations are done in a special developed MATLAB program. To describe the 3D directivity as precise as possible, the program accounts for all aircraft angles (heading, pitch and roll), as well as the full source-to-receiver geometry. A separate computer program has been developed to visualise the source directivity. The result is displayed as a colour map of the hemisphere below the aircraft, projected to a plane. The hemisphere above the aircraft is disregarded in these analyses. The projection is a hybrid polar projection with side angle in the wing-plane (Theta) along the projection circle, and azimuth angle (Phi) displayed as radius. Theta is set to 0 in front, 90 degrees to the right, 180 degrees behind etc. Phi is 0 straight above the aircraft, 90 degrees in the wing-plane, and 180 degrees straight below (see Figure 4.1). This projection is chosen because it preserves the real area of any shape projected from the hemisphere to the plane, which means that grouping, smoothing, and other statistical analyses can be based on the x, y co-ordinates of the projection plane. The projection however, does not preserve the shape of any object. To help visualise any axis-symmetry in the directivity, lateral lines are added to the diagram (see red lines in Figure 4.1). These connect points in the projection with constant angle to the flight-heading vector. Any axis-symmetry (i.e. cylindrical symmetry) in the data set will display as coloured areas parallel to these lines. In the source directivity program all source data for one aircraft type can be displayed. Different filters can be added to the data to investigate how the directivity changes due to operation (arrival and departure), thrust interval, frequency, measurement position and microphone height. In addition data can be excluded based on user defined parameters for background noise, source – receiver distance, aircraft elevation angle, and quality indicators in the propagation model.

2 The term “directivity” used herein is the distribution of backwards propagated sound pressure levels over the hemisphere below the aircraft. This should not be mixed with the normalised directivity that usually characterises single sound sources.

13

Figure 4.1 Directivity projection. Theta angles in blue (outside circle). Phi angles in grey

(inside circle). Lateral lines (axis-symmetric) display equal angle to flight vector. Finally the program is able to calculate statistics for a resulting averaged source directivity, and produce report files containing source spectra, directivity plots and tables. Directivity parameters are averaged according to a grid. We assume symmetry between right and left side of the aircraft. Thus all data are transferred to the right-hand side of the aircraft before statistical analyses. To cover all angles, the grid-cells with less than 10 samples are automatic set equal to the closest grid-cell with more than 10 samples. The resulting averaged source directivities with corresponding statistics are tabulated in ASCII text files. They are not included in this report due to space limitations.

4.1.2 Backward propagation parameters The backward propagation from receiver to source is based on:

• Nord 2000 algorithms. • Actual geometry including flight position and orientation sampled every 0.5-second. • Actual ground impedance measured close to the receiver. • Air absorption based on actual measured temperature and humidity, smoothed over 30

minutes. • Actual wind speed and direction smoothed over 7 minutes.

14

• Temperature gradient is set to –1 degrees per 100 meter, in accordance with measured

meteorology and expert judgement [13] of the total weather situation. • No topography corrections (horizontal ground assumed). • After backward propagation, the data is normalised to a fixed thrust setting of 10,000

Lbs. This is done by subtracting 1.1 dB per 1000 Lbs. thrust exceeding 10,000 Lbs.

4.1.3 Data filtering A series of test calculations of source directivity were done to verify the quality, and find optimum settings for filtering parameters. The goal was to exclude suspicious data, and gain highest possible statistic strength in the results. Preliminary results indicated that the directivity does not change much with varying thrust, so the figures display the directivity as an average for all measured thrust settings, after separation between arrival and departure. The optimised filter settings where:

• No specific adjustment for background noise other than what was done with the raw measurement data (se Section 3.3).

• All data propagated more than 2 km where excluded. • On departures, all data propagated at an elevation angle less than 15 degrees were

excluded.3 On arrivals no data where excluded due to elevation angle. • All data with air and ground attenuation more than 30 dB were excluded.

4.1.4 Directivity results In the technical memo in part 2 of the report the directivity plots are shown for each aircraft type. The data is organised as one spectrum table and 27 directivity plots, one for every 1/3-octave band. The table covers all measured directions from the aircraft, and has the following columns:

• Freq.: 1/3-octave centre frequency from 25 to 10,000 Hz. • Num.: Number of 0.5-second measurement samples. • Avg.: Arithmetic average sound level 1 meter from a point representing the aircraft • Std: Standard deviation of the 0.5-second sound levels. • 95%Avg: The +/- deviation for the average sound level, within 95% confidence interval. • P1, …P6: Average level for each engine power setting, grouped according to the NPD

table. (No data is denoted –1). The directivity plots show all 0.5-second samples as noise levels 1 meter from a point representing the aircraft. They are plotted in the projection of the hemisphere below the aircraft, described above, as one small circle per sample. The colour of the circle indicates the deviation from the average of all samples, according to a colour scale ranging +/- 15 dB. Black circles indicate that the sample is outside the colour scale. At the top of both tables and directivity plots, the aircraft type is indicated, along with the one letter indicator “A” for arrival and “D” for departure. In the plots, the 1/3-octave centre frequency, and average sound level is indicated. Table 4.1 and 4.2 show examples of directivity data for the aircraft B737-600 and MD82 on departure.

3 This filter only removes data in front of the aircraft and does not influence data to the side or behind.

15

Table 4.1 B737 600 mean spectrum on departure

SPECTRUM Source Emission

B736 D

Freq. Num. Avg. Std 95%Avg P1 P2 P3 P4 P5 P6 25 7426 120.2 6.7 0.2 -1.0 -1.0 119.4 120.3 -1.0 -1.0 31.5 8568 121.9 6.0 0.1 -1.0 -1.0 121.1 121.9 -1.0 -1.0 40 8738 122.5 5.6 0.1 -1.0 -1.0 121.9 122.6 -1.0 -1.0 50 9146 123.1 5.2 0.1 -1.0 -1.0 122.1 123.2 -1.0 -1.0 63 9345 124.2 5.0 0.1 -1.0 -1.0 122.5 124.4 -1.0 -1.0 80 9799 125.4 4.7 0.1 -1.0 -1.0 122.9 125.6 -1.0 -1.0 100 9982 125.3 4.3 0.1 -1.0 -1.0 122.5 125.6 -1.0 -1.0 125 10195 124.6 4.3 0.1 -1.0 -1.0 122.0 124.8 -1.0 -1.0 160 10203 123.6 4.6 0.1 -1.0 -1.0 120.7 123.9 -1.0 -1.0 200 10298 123.0 5.0 0.1 -1.0 -1.0 119.5 123.3 -1.0 -1.0 250 10331 122.5 4.9 0.1 -1.0 -1.0 120.0 122.7 -1.0 -1.0 315 10365 121.4 5.3 0.1 -1.0 -1.0 118.0 121.7 -1.0 -1.0 400 10344 120.5 5.6 0.1 -1.0 -1.0 117.1 120.8 -1.0 -1.0 500 10330 119.9 6.0 0.1 -1.0 -1.0 116.0 120.3 -1.0 -1.0 630 10286 119.0 5.7 0.1 -1.0 -1.0 115.1 119.3 -1.0 -1.0 800 10254 118.1 5.3 0.1 -1.0 -1.0 113.9 118.5 -1.0 -1.0 1000 10210 117.5 5.4 0.1 -1.0 -1.0 113.2 117.9 -1.0 -1.0 1250 10219 117.2 5.5 0.1 -1.0 -1.0 112.8 117.5 -1.0 -1.0 1600 10188 117.3 5.6 0.1 -1.0 -1.0 112.8 117.7 -1.0 -1.0 2000 9856 118.2 6.1 0.1 -1.0 -1.0 114.1 118.6 -1.0 -1.0 2500 9198 119.0 5.2 0.1 -1.0 -1.0 116.7 119.2 -1.0 -1.0 3150 7390 120.7 5.2 0.1 -1.0 -1.0 120.1 120.8 -1.0 -1.0 4000 4334 121.2 5.4 0.2 -1.0 -1.0 121.2 121.2 -1.0 -1.0 5000 1772 121.1 5.2 0.2 -1.0 -1.0 120.8 121.1 -1.0 -1.0 6300 463 120.7 5.8 0.5 -1.0 -1.0 121.6 120.6 -1.0 -1.0 8000 81 119.5 7.2 1.6 -1.0 -1.0 122.0 119.1 -1.0 -1.0 10000 7 111.6 7.4 5.5 -1.0 -1.0 -1.0 111.6 -1.0 -1.0

16

Figure 4.2 Noise source directivity pattern

B737-600 Departure for some 1/3 octave bands

17


B737-600 Arrival for some 1/3 octave bands

18

Table 4.2 MD 82 mean spectrum on departure

SPECTRUM Source Emission

MD82 D

Freq Num Avg Std 95%Avg P1 P2 P3 P4 P5 P6 25 13546 124.2 8.3 0.1 -1.0 -1.0 -1.0 -1.0 125.2 -1.0 31.5 14365 126.1 8.1 0.1 -1.0 -1.0 -1.0 -1.0 127.1 -1.0 40 15589 127.0 8.0 0.1 -1.0 -1.0 -1.0 -1.0 128.0 -1.0 50 15535 128.3 8.2 0.1 -1.0 -1.0 -1.0 -1.0 129.3 -1.0 63 15698 130.0 8.5 0.1 -1.0 -1.0 -1.0 -1.0 131.0 -1.0 80 16139 131.4 8.1 0.1 -1.0 -1.0 -1.0 -1.0 132.4 -1.0 100 17310 131.5 7.4 0.1 -1.0 -1.0 -1.0 -1.0 132.3 -1.0 125 18411 131.0 6.3 0.1 -1.0 -1.0 -1.0 -1.0 131.6 -1.0 160 18694 130.6 5.8 0.1 -1.0 -1.0 -1.0 -1.0 131.0 -1.0 200 18727 130.4 5.9 0.1 -1.0 -1.0 -1.0 -1.0 130.8 -1.0 250 18716 130.4 5.6 0.1 -1.0 -1.0 -1.0 -1.0 130.7 -1.0 315 18689 130.2 5.7 0.1 -1.0 -1.0 -1.0 -1.0 130.3 -1.0 400 18745 129.8 5.8 0.1 -1.0 -1.0 -1.0 -1.0 129.8 -1.0 500 18690 129.3 5.6 0.1 -1.0 -1.0 -1.0 -1.0 129.2 -1.0 630 18749 129.1 5.5 0.1 -1.0 -1.0 -1.0 -1.0 129.0 -1.0 800 18610 128.8 5.4 0.1 -1.0 -1.0 -1.0 -1.0 128.7 -1.0 1000 18695 128.3 5.5 0.1 -1.0 -1.0 -1.0 -1.0 128.3 -1.0 1250 18687 127.6 5.6 0.1 -1.0 -1.0 -1.0 -1.0 127.6 -1.0 1600 18675 126.9 5.7 0.1 -1.0 -1.0 -1.0 -1.0 126.9 -1.0 2000 18504 126.6 5.7 0.1 -1.0 -1.0 -1.0 -1.0 126.6 -1.0 2500 17610 126.8 5.6 0.1 -1.0 -1.0 -1.0 -1.0 126.9 -1.0 3150 13494 126.5 5.6 0.1 -1.0 -1.0 -1.0 -1.0 127.0 -1.0 4000 7188 126.0 5.7 0.1 -1.0 -1.0 -1.0 -1.0 126.7 -1.0 5000 2948 124.4 6.5 0.2 -1.0 -1.0 -1.0 -1.0 124.7 -1.0 6300 803 121.0 7.6 0.5 -1.0 -1.0 -1.0 -1.0 121.5 -1.0 8000 150 117.1 8.5 1.4 -1.0 -1.0 -1.0 -1.0 117.1 -1.0 10000 19 110.9 10.4 4.7 -1.0 -1.0 -1.0 -1.0 117.2 -1.0

19


MD82 Departure for some 1/3 octave bands

20


MD82 Arrival for some 1/3 octave bands

21

4.1.5 Comments The directivity analyses can be summarised in the following main results:

• The directivity is strongly dependent on frequency. • The directivity is fundamentally different between the aft-fuselage mounted-engine

aircraft (MD8x, MD90) and wing-mounted-engine aircraft (B73x). This is especially the case below 1000 Hz, which is the dominating frequency range in noise zoning situations. Aft-fuselage-mounted-engine aircraft has a very distinct directivity lobe behind and below the aircraft at mid and low frequencies, while wing-mounted-engine aircraft are more omni directional.

• There is no fundamental difference in directivity between MD8x and MD90, even though the engine type is different.

• There is a fundamental difference between the directivity of B73x and MD90, even though the engine type is similar. Connected to the above two points, this indicates that a major difference in directivity depends more on engine placement, than engine type.

• The sound radiation downwards is generally stronger than to the side. This is clearly most evident behind aft-fuselage- mounted-engine aircraft, at mid and lower frequencies.

• The directivity does not change significantly within the thrust intervals covered by the measurements.

4.2 Lateral Attenuation Two methods have been used to determine lateral attenuation: • Through measurements where the A-weighted data were used to estimate the lateral attenuation

by comparison of noise levels measured directly below the flight and at positions to the side. • Through simulations based on the source directivity spectra found in 4.1

4.2.1 Lateral attenuation as measured It was assumed that the measurement positions 2 – 5 were situated on a line orthogonal to the flight path. The lateral attenuation (LA) was calculated according to equation (4.1):

( ) 53log10,2

2 K=Φ−

−−=θ k

hd

SELSELlLA kk

kkk (4.1)

Here, k is the measurement position, (lk,θk,dk) is the lateral distance, elevation angle and slant distance from the measurement position to the aircraft respectively, and h2 is the height of the aircraft above the microphone at position 2. Φk is the A-weighted air absorption from the aircraft to position k, which was calculated according to equation (4.2) :

[ ] ( )[ ]kk dRHTSS ,,α−−=Φ AA (4.2) Here A[] represents A-weighting, S is the noise source spectrum of the aircraft, T is the temperature, RH is the relative humidity, and α( ) is the frequency dependent air absorption for the given temperature, humidity and distance. For each flight, the temperature and relative humidity values at the time of closest point of approach were used.

22

4.2.1.1 Comparison of measured lateral attenuation and SAE 1751 The estimated lateral attenuation was developed per aircraft type and split on landing (LA) and takeoff (TO) operations. In Figure 4.6 the estimated lateral attenuation is compared with the standard lateral attenuation given in SAE AIR 1751 [5]. The straight line in the figure indicates conformity of values.

Figure 4.6 Lateral attenuation, measured versus SAE prediction.

23

In Figure 4.7 the estimated lateral attenuation is presented as a function of elevation angle. Also shown is the lateral attenuation according to SAE AIR 1751, corresponding to values at 300-m distance, and to maximum attenuation (distance 914 m or more) which covers the actual span in lateral distance in this study.

Figure 4.7 Measured lateral attenuation versus elevation angle.

4.2.1.2 Comments on measured lateral attenuation The results indicated above may be influenced by thrust cutback during take off. Cutback is assumed to occur at an aircraft height of approximately 450-500 m above runway elevation. This means that a small part of the takeoffs (TO) may be affected. However, the engine power cutback represents a certain noise level reduction in both SEL2 and SELk in equation (4.1). The net effect in Figures 4.6 and 4.7 is therefore expected to be rather small. The measured lateral attenuation shows clearly that SAE AIR1751 overestimates the effect for these aircraft. Since lateral attenuation in the recommendation combines effects from directivity, installation of engines and ground attenuation, it is not clear what contributes most to the differences measured. There is a tendency that B737 models have less lateral attenuation than MD80 models. This may be due to any of the effects, since both noise source spectra, directivity and installation are different. Further investigations have therefore been undertaken to distinguish between the possible causes.

4.2.2 Simulated lateral attenuation The noise resulting from horizontal straight line flights have been simulated. The sound emission of the source is modelled by the spectra and directivities described in chapter 4.1. Flight speed is fixed to 160 knots, while flight height is variable. Sound pressure levels are calculated at different receiver points 1.5 meters above the ground below the flight line and to the side. Nord 2000 algorithms have been used to calculate the noise

24

propagation from points along the flight line to the receiver point. Distance from receiver point to closest point of approach is kept constant at two distances. The propagation conditions are set to:

• Horizontal ground with a flow resistivity = 250,000 Pa s/m2. • Temperature = 15° Celsius. • Relative humidity = 70%. • Wind speed = 0. • Temperature gradient = -1°C/100m. • Wind turbulence parameter = 0.5.

The turbulence parameter expresses the strength of turbulence due to wind. The value of this parameter is adjusted by comparison between simulated and measured vertically separated sound pressure levels. The sound pressure levels at the receiver points are accumulated to A-weighted SEL values for the whole flight. The following figure shows the difference between SEL below the flight line, and SEL to the side at the same distance from the point. The results are plotted as function of elevation angle for MD82 and B737-600. To indicate distance dependency, the curves are given at both 300 and 1200 meters.

Simulated propagation

-2

-10

12

34

5

67

89

10

0 10 20 30 40 50 60 70 80 90

Elevation Angle (degrees)

Late

ral A

ttenu

atio

n in

cl. G

roun

d Ef

fect

MD82 at 300mB736 at 300mMD82 at 1200mB736 at 1200m

Figure 4.8 Simulated lateral attenuation (dBA)

The spectral source directivity and the ground attenuation both influence the curves. The difference between the two aircraft is caused by the spectral source directivity. Since directivity seems to be more dependent on engine position than engine class, it can be concluded that the main difference between the two aircraft lateral attenuation is due to engine position.

25

4.3 Noise-Power-Distance data The measured A-weighted data were used to estimate the noise-power-distance data. Only data from measurement positions 1 and 2 were used, i.e. the positions directly below the flight path.

4.3.1 Analysis method The measured levels were normalised from the actual atmosphere represented by ISO air absorption coefficients for the ambient temperature and relative humidity to the atmosphere represented by the air absorption coefficients tabulated in SAE AIR 1845 [6]. The A-weighted SEL levels for each flight were corrected for speed by adding the correction Vcorr = 10log(V/160) according to SAE AIR 1845, where V is the velocity (kts) of the aircraft at CPA. The A-weighted SEL- and MAX levels were compared to the corresponding noise curves from the NORTIM database noise-power-distance tables. For comparison with the NORTIM curves, the thrust levels of the measured data were grouped in intervals corresponding to the thrust levels of the NORTIM curves. The measured level at a specific thrust value was normalised to the curve value by linear interpolation between the database noise levels. For the B737-600, the measured noise data are compared with the NPD curves for B737-500 from the database, since the latter lacks data for this specific aircraft type. Also, at this point, no data for the B737-700 were included in the NORTIM database, and comparison for this aircraft type was with the data for B737-400.

4.3.2 Results A technical memo with all results is incorporated in Part 2 of the report. The memo contains figures with comparisons of measurement results and database values for all the different aircraft types that were in the study for both SEL and LAMAX values. The following figures show the results for the two dominant aircraft types in the study, the MD 82 and the B737-600 (B736). In the following figures, the thrust levels of the NORTIM curves are annotated along the curves.

26

Figure 4.9 Comparison of database (NPD) and measured SEL for B737-600.

Figure 4.10 Comparison of database (NPD) and measured LAMAX for B737-600.

27

Figure 4.11 Comparison of database (NPD) and measured SEL for MD 82.

Figure 4.12 Comparison of database (NPD) and measured LAMAX for MD82.

28

Note that the NORTIM database levels for MD82 has been adjusted by + 1.5 dB for both SEL and LAMAX for the two upper thrust settings as compared to the INM database.

4.3.3 Statistical analysis The deviations (in dBA) between the NPD curves from the database and the normalised data have been analysed and the statistics on the differences are given in the following tables: for differences in SEL values and LAMAX values respectively. The number of observations, mean value, standard deviation and limits for the 95% confidence interval are shown for each aircraft and operation. A positive mean value suggests that the database NPD curves give too high noise value. Mean differences are not significant if the number 0 lies between the limits of the 95% confidence interval.

Table 4.3 Statistical analysis of SEL value differences

Type Operation No of Obs Mean St.dev CI_low CI_high B736 LA 20 4.1 1.6 3.4 4.9B737 LA 22 3.8 1.4 3.1 4.4B738 LA 3 1.4* 2.1 -3.8 6.6MD81 LA 13 -2.6 0.7 -3.0 -2.2MD82 LA 25 -1.6 1.6 -2.3 -1.0MD87 LA 8 -2.4 1.4 -3.5 -1.2MD90 LA 2 8.9* 6.5 -49.5 67.2B736 TO 31 -0.9 1.2 -1.4 -0.5B737 TO 36 -0.8 1.8 -1.4 -0.2MD81 TO 29 -0.8 1.1 -1.3 -0.4MD82 TO 58 -1.4 1.5 -1.8 -1.0MD83 TO 2 0.1* 1.6 -14.5 14.8MD87 TO 11 -1.1 1.2 -1.9 -0.3MD90 TO 4 -2.0 1.0 -3.5 -0.5

* not significant

Table 4.4 Statistical analysis of LAMAX value differences

Type Operation No of Obs Mean St.dev CI_low CI_high B736 LA 20 6.0 2.5 4.9 7.2B737 LA 22 6.3 2.6 5.1 7.4B738 LA 3 4.2* 2.7 -2.4 10.8MD81 LA 13 -1.8 1.1 -2.4 -1.1MD82 LA 25 -0.4* 2.5 -1.5 0.6MD87 LA 8 -1.6 1.8 -3.1 -0.1MD90 LA 2 12.3* 10.9 -85.9 110.5B736 TO 31 0.2* 1.6 -0.4 0.8B737 TO 36 0.4* 2.0 -0.3 1.1MD81 TO 29 -0.5* 1.9 -1.2 0.2MD82 TO 58 -1.2 2.4 -1.8 -0.6MD83 TO 2 0.7* 2.0 -17.1 18.6MD87 TO 11 -1.0* 1.6 -2.1 0.1MD90 TO 4 -3.8 1.0 -5.4 -2.2

* not significant

29

4.3.4 Comments It is clear from the results that there are statistically significant differences between measured noise data and the database values for similar thrust setting and distance for some aircraft. The results for SEL data is generally more stable with higher confidence than the LAMAX data. The deviations for the B737 family is not clear since substitutions have been used from the database. However, the finding of a little, but statistically significant, higher SEL noise levels for take-off from both –600 and –700 as compared to –500 and –400 respectively, is surprising, since the newer variants are supposedly quieter. Norwegian authorities decided that 1.5 dB should be added for MD80 family aircraft at take-off thrust values based on the Yuma tests in the early 1990’s [14]. Still the database underestimates the SEL values with up to 1.4 dB.

4.4 Comparison of actual flight profiles with database A flight profile is defined as altitude, velocity and engine thrust settings as a function of distance from a reference point along the flight track. For a take-off profile, the reference point is the brake release point. For approach profiles, the reference point is the runway threshold at the touch down end of the runway. Flight recorder data give the possibility of constructing actual flight profiles. All 155 profiles have been extracted for the valid flights. Part 2 includes a technical memo with these profiles compared to the corresponding database profiles.

4.4.1 Examples of results Below are examples for the MD82 aircraft, both for approaches and departures. For the arrivals, a standard glide path of 3 degrees has been assumed for all flights. For the departures, the stage length has been calculated based on the ICAO code of the destination by the same procedures as is used in NORTIM. The figures show height, speed and net thrust as a function of distance from the reference point.

30

Figure 4.13 Comparison of database

approach profile and 15 measured profiles for MD 82.

NOTE: The database thrust level for reverse thrust is not a physical value. It refers to a database noise level in the NPD curve.

Figure 4.14 Comparison of database take-off profile and 29 measured profiles for MD

82.

31

4.4.2 Comments on database and flight recorder profile comparison For approach and landing there is little deviation from the 3 degrees glide slope with a few exceptions where the aircraft is coming in steeper. Differences in speed and net thrust are evident. Outside 5000 meter (3 NM), the measured speed is higher than the database profile. Outside 9000 meter (5 NM) the thrust is also generally lower, which results in less noise produced than the database profile would suggest. Between 9000 and 5000 meter from threshold there is an attitude shift in thrust enforcement. From a stable low thrust, one enters a region with higher mean value and larger variations. Compared with the database profile, this higher mean value would result in higher equivalent noise. The noise level will fluctuate and prediction of maximum levels will be uncertain. For the shown variations between net thrust 4300 and 8000 Lbs., the database NPD gives a difference in maximum noise level of 8-10 dB for this aircraft type. The measured take off profiles show that the INM database profile is based on different climb out procedures. Lower thrust is applied both during take off run and throughout the climb. Power cutback can only be seen on around half of the events, most likely due to frequently use of derated thrust procedures. The effects on noise level are mixed and will vary along the flight path and to the sides. A comparison test will be necessary to illustrate the effects of the different procedures.

4.5 Influence of meteorology on air absorption Appendix 2 shows recorded values of meteorology parameters during the exercise. It has previously been argued that the difference in air absorption between the table given in SAE AIR 1845 and normal Nordic weather conditions in it self would create an under-estimation of predicted aircraft noise in the order of several dBs. The figure shows the combination of temperature and relative humidity during the test, sampled at the time of closest approach at the meteorology system.

Figure 4.15 Temperature and relative humidity relation during each flight.

32

Just a few of the flights were measured under conditions close to international standardised atmosphere (ISA) of 15°C and 70% RH. Most of the time the temperature was higher and humidity was less. The following figure shows the resulting air absorption, expressed in dBA, for the measured mean noise source spectra for all operations. Three conditions are calculated based on the ISO standard, reference [4] together with the absorption based on SAE AIR 1845.

Figure 4.16 Air absorption in dBA for a mean spectrum of all flights, as a function of source to

receiver distance. The figure shows that air absorption was generally higher than AIR 1845 coefficients during the test. On colder days the air absorption may be less at some distance away from the source. However, the effect could be at a magnitude of tenths of a decibel and can therefore not alone explain the differences that are experienced between long-term measurements and calculations.

4.6 Quality check on the noise and flight tracking system at OSL As part of the investigation, a quality check on the noise and track monitoring system at OSL was performed. The quality check consisted of a number of tasks: • Influence on background noise at the measurement positions. • Influence on local microphone placement with respect to microphone height above ground and

influence of local surroundings. • Validation of calibration and data treatment. • Validation of flight tracks and profiles registered by the flight tracking system. • Validation of information on aircraft type, operation, runway use. In Part 2 a summary of a survey of the monitoring positions is presented. The evaluation of the first two bullet points above is based on this survey. A comparison of the measurement results at position 1 with the monitoring systems position 4 is given in the following. Further, the flight tracks and profiles reported by the radar tracking system are compared to the flight recorder data.

33

The last part of this chapter concentrates on the NTMS aircraft log data as compared to the measurement crew observations during the program.

4.6.1 Evaluation of microphone positions of the airport monitoring system A technical memo of this survey is posted in the report part 2. Pictures of the microphone locations are included in the memo.

4.6.1.1 The investigation The microphone positions belonging to the permanent noise and track monitoring system at OSL are numbered NMT 4 - 11. NMT 4 and NMT 5 are located at the north end of runway west and south end of runway east respectively. The rest of the microphones are located in areas some kilometres from the runways, mainly north and south of the airport in the runway directions. When judging the influence of local surroundings, a number of parameters were investigated during an inspection trip to all the positions mentioned above, exclusive pos. 5. The parameters recorded were: • type of area (rural, residential, urban), • general background noise situation, • noisy equipment/installations/industrial plants close to the microphone location, • distance to the nearest building, • roads near to the location, or audible, • ground cover.

4.6.1.2 The findings The findings of this survey investigation are listed in Table 4.5 below.

Table 4.5 Summary of registrations according to the listed parameters Type of area Background

noise Equipment, installations

Distance to nearest building, m

Type of roads

Type of ground cover

NMT 4 Airfield - - 20 - Grass NMT 5 Airfield - - - - Grass NMT 6 Rural Low / agric. no 275 - Agriculture NMT 7 Rural Low / agric. no 75 local/Rv120

100m Agriculture

NMT 8 Rural * Low / agric. no 15 E6, 500 m Agriculture NMT 9 Rural * Low / agric. no 100 Fv462,

100m Agriculture

NMT 10 Rural Low / agric. no 140 local Agriculture NMT 11 Rural Low / agric. no 30 local Agriculture

* Residential area in the vicinity.

All locations NMT 6-11 have generally low background noise levels, but the noise level will increase during periods of agriculture activities. Location NMT 7 is situated close to a local road. Vehicle traffic on this road may occasionally give rise to high maximum noise levels. NMT 8 is located relatively close (approximately 15 m) from the nearest building. This may represent a cause to sound reflections from the nearest vertical building façade.

34

As far as one could experience during the inspection survey, none of the locations NMT 6 - 11 were exposed to noise from permanent technical installations or local industrial plants. At location NMT 8 noise from garden activities (for instance grass cutting) may represent a certain background problem during summer time. At this location traffic noise from the distant E6 main road was clearly audible. The nearby ground covers are in most cases agricultural fields. This means that their acoustical properties will vary during the summer season, typically between 100 and 400 kPa s/m2. In principle, this may influence on the recorded aircraft noise levels in the magnitude of a few decibels, especially in the lower end of the noise spectra.

4.6.1.3 The influence of microphone height All the locations have the same microphone height above ground, i.e. approximately 6 m. Aircraft noise level calculations are made for a receiver height of 1.5 m above ground. The effect of this difference was investigated theoretically in an earlier investigation [1]. Two examples of the influence of microphone height were prepared. At measurement position 2 the two microphone heights were 1.5 and 10.5 m respectively. This position is directly below the aircraft flight path. At measurement position 5 the two heights were 1.5 and 8 m. This position is approximately 1250 m to the side of the flight track. Measured sound exposure levels averaged over all departure flights with B736 indicate some significant deviations in the low frequency range below approximately 200-250 Hz. Although the deviation amount to 7 dB in some 1/3-octave bands, the influence was measured to 0.7 and 0.3 dBA for position 2 and 5 respectively. The high microphone measured higher noise levels than the 1.5-meter microphone.

4.6.1.4 Discrimination algorithms in NTMS According to [17] the algorithms that discriminate aircraft noise from other noise sources in the measurements reported by NTMS, are as based on the following criteria: • The noise event must exceed a sound pressure level threshold set individually at each NMT,

typically 55 dBA to 57 dBA for positions in the rural area, 75 dBA for the two close to the runways. It is possible to have a different threshold for start and end of the noise event.

• Maximum sound pressure level during the noise event must be at least 3 dBA higher than threshold.

• The noise event must have duration of 5 to 60 seconds for the rural positions, 5 to 140 seconds for the positions close to the runway.

• Measured noise events is correlated to aircraft movements by use of the radar tracking system and the following criteria apply: - Distance between aircraft and microphone must be less than a given value, set individually for each position, ranging from 1.800 to 3.000 meter. - Time interval between point of closest approach to the position and measured LAMAX must not exceed 20 second.

If a noise event fails to comply with any of these criteria, it is automatically neglected.

4.6.1.5 Concluding remarks All the microphone locations are situated mainly in rural areas, with generally low background noise. No specific permanent background noise sources were found. But agriculture activities can give rise to high maximum noise levels during specific parts of the summer season. If the noise

35

monitoring system is activated due to aircraft traffic (i.e. all criteria in 4.6.1.4 are fulfilled), there is a chance that a false aircraft noise level may be recorded. The microphone height of 6 m above ground could give slightly higher noise level relative to 1.5 m, as far as A-weighted SEL values are concerned.

4.6.2 Noise levels measured by the airport noise monitoring installation The equipment at each position of the monitoring system is calibrated acoustically every 6 months. This is considered sufficient in addition to the daily actuator calibration checks. Measurement position 1 in this investigation coincides with one of the measurement positions (NMT-4) of the airport noise monitoring installation. Thus, the same noise levels should be measured since the microphones of the two systems were less then 0.3 m apart. It should be noted that the two microphones were not equally oriented. The monitoring system’s microphone has its diaphragm positioned horizontally, while the investigation microphone’s diaphragm was vertically oriented. Taking general microphone response curves and associated tolerances into account, no significant difference between the microphones should be anticipated below approximately 3 - 4 kHz. The differences in microphone sensitivity at higher frequencies will not be crucial. A comparison of LAMAX and A-weighted SEL levels measured by the two microphones was done. The results are shown in Table 4.6 – 4.8 below, given as mean level differences and standard deviations per aircraft type. Positive level differences indicate that higher noise levels are measured by the airport noise monitoring installation. The level difference observed seem to be negative for LAMAX and positive for SEL. The magnitude of the differences is however what can be expected for two independent measuring systems. In addition the maximum level detection in the two systems may not be quite coincident.

36

Table 4.6 Differences in measured noise levels at position 1 / NMT-4, landings.

Aircraft type Operation MAX St.dev SEL St.dev Count B736 LA -1.3 1.0 0.4 1.2 13 B737 LA -1.1 0.7 0.4 0.5 14 B738 LA -1.2 0.3 0.3 0.2 2 MD81 LA -0.9 0.4 0.4 0.2 7 MD82 LA -0.8 0.6 0.3 0.5 16 MD83 LA 0 MD87 LA -1.0 0.4 0.5 0.4 5 MD90 LA -1.3 0.0 -0.5 0.0 1

ALL LA -1.0 0.7 0.4 0.7 58

Table 4.7 Differences in measured noise levels at position 1 / NMT-4, take-off.

Aircraft type Operation MAX St.dev SEL St.dev Count B736 TO -0.2 0.3 -0.8 0.4 15 B737 TO -0.2 0.3 -0.6 1.0 19 B738 TO 0 MD81 TO -0.5 0.5 0.4 0.2 17 MD82 TO -0.5 0.4 0.5 0.4 32 MD83 TO 0.2 0.0 1.1 0.0 1 MD87 TO -0.4 0.4 0.5 0.3 6 MD90 TO -0.3 0.4 0.1 0.1 2

ALL TO -0.4 0.4 0.1 0.8 92

Table 4.8 Differences in measured noise levels at position 1 / NMT-4, all operations.

Aircraft type Operation MAX St.dev SEL St.dev Count B736 TO+LA -0.7 0.9 -0.2 1.0 28 B737 TO+LA -0.6 0.7 -0.1 0.9 33 B738 TO+LA -1.2 0.3 0.3 0.2 2 MD81 TO+LA -0.7 0.5 0.4 0.2 24 MD82 TO+LA -0.6 0.5 0.4 0.5 48 MD83 TO+LA 0.2 0.0 1.1 0.0 1 MD87 TO+LA -0.7 0.5 0.5 0.3 11 MD90 TO+LA -0.6 0.7 -0.1 0.3 3

ALL TO+LA -0.6 0.6 0.2 0.8 150

4.6.3 Comparison of tracks and profiles from flight recorder with NTMS A technical memo on this subject is presented in Part 2. Of the 155 flights investigated, 70 flight records included GPS position data. The pressure altitude is also reported in the flight recorder. A comparison has been made between the flight recorder data and the data reported from the radar tracking part of the noise and flight monitoring system at OSL.

37

The following two figures compare the position data given in the flight recorder and reported by the radar tracking system. The blue lines represent the GPS position (smoothed). The red dots represent the position given by the radar. The magenta crosses mark the position of the five measurement sites. The western runway is shown as a thick, black line. The co-ordinates are given relative to measurement position 1. The figures show that near the ground (i.e. close to the runway), the radar data is subject to significant random errors. At greater altitudes, there are no large random errors, although the deviation is in the order of ± 50–100 m in the north/south direction, and ± 20–60 m in the east/west direction.

38

Figure 4.17 Comparison of position data from NTMS (dots) and flight recorder GPS based flight

tracks (lines).

Figure 4.18 Comparison of position data from NTMS (dots) and flight recorder GPS based flight

tracks (lines) near the runway.

39

The figures below compare the altitude given in the flight records and reported by the radar tracking system. Only three randomly selected flights are shown for each operation. The altitudes of both sources have been adjusted by setting the altitude at the runway to zero.

Figure 4.19 Altitudes from flight recorder (thick lines) and NTMS (thin lines) as a function of

time for 3 randomly selected arrivals.

Figure 4.20 Altitudes from flight recorder (thick lines) and NTMS (thin lines) as a function of

time for 3 randomly selected departures.

40

The figure below shows the difference between the reported altitude for all 70 flights. Red dots represent arrivals, while blue dots represent departures. The data show that the average deviation is approximately –50 m for arrivals, and approximately 50–100 m for departures.

Figure 4.21 Comparison of altitudes from flight recorder and NTMS as a function of time for

arrivals (red dots) and departures (blue). The average deviation may be due to time delay in the altitude reported by the radar tracking system. A time adjustment of approximately 1.5 – 3 seconds applied to the radar data results in better agreement.

4.6.4 Quality check on the monitoring system aircraft log Project personnel at the air traffic control tower kept record of all flight movements in the north west quadrant of the airport during the exercise. Their log has been compared to the aircraft movement log produced by the noise and track monitoring system (NTMS) at OSL. The latter is a vital input for the GMTIM noise model. Major findings are: • The time difference between observation and NTMS log is typically 1 minute equal to the

resolution, where the NTMS systematically is late. NTMS time matches with the time the aircraft is clear from the runway.

• There is a one-to-one match between observation and the aircraft registration number in NTMS. • There is a one-to-one match on runway use and on operation type. • Translation from registration number to aircraft type in NTMS has been checked with the

Norwegian Civil Aircraft Register or “JP airline – fleets international” (97/98 edition) and shows good match with a few exceptions.

41

5 Conclusions The investigations show that several effects influencing the accuracy of aircraft noise modelling have been isolated.

5.1 Directivity The most important result in this study is the directivity patterns that have been extracted. One basic assumption for state of the art models, that never has been certified, is that sound radiation from the aircraft is cylindrical symmetric with the fuselage centre line as the symmetry axis. The results from this investigation clearly show that this is not the case. All of today’s noise models fall short of taking this result into account. 3D directivity can only be modelled in full extent with a new method of describing the noise source. The task is therefore put in the category “long term measures”.

5.2 Lateral attenuation Both measurements and subsequent simulations show that lateral attenuation is less than predicted by SAE AIR 1751. This trend is true for all aircraft types in this study, at both departure and landing operations. This trend is also recognised in other investigations [7,8]. The results also indicate that the aircraft family B737 exhibits less lateral attenuation than the aircraft family MD80. The reason for this is related to effects of engine position (wing mounted and rear fuselage mounted engines respectively). This is also seen in other recent investigations [8,9]. The lateral attenuation according to SAE AIR 1751 enters into the noise level prediction performed by the NORTIM/GMTIM methods. Over-estimation of the lateral attenuation leads to under-estimation of predicted noise levels. The findings in this study show that typical effect could be in the order of 2 - 4 dB. The over all effect on the noise zones is dependent on where the calculation point is located relative to the flight path. The findings concerning lateral attenuation can be taken into account in “short term measures”.

5.3 Noise Power Distance data The analyses show that there are statistically significant differences between measured noise and the database NPD curves. Further analysis need to be performed for the B737-600/700 aircraft where new database NPD curves are compared with the measurements. The statistically significant differences can be used as corrections in the NORTIM/GMTIM database as a “short term measure”.

5.4 Flight profiles It is obvious from this investigation that the standard take off profiles generated by INM and used in GMTIM/NORTIM are based on different procedures than those used by SAS at Oslo Airport. The main reason is probably that SAS uses derated thrust frequently and INM procedures are not constructed as such. It should be noted that, according to SAS, the procedures they use are in agreement with criteria set by OSL. The resulting noise level difference will vary along the flight track and to the sides.

42

It is also clear that the thrust adjustments typically seen from 5 NM to touch down are not taken into account in the standard approach profiles. The effect on maximum levels can be in the order of up to 10 dB. Both these findings should be corrected for in the noise model. Corrections may be implemented on a short to intermediate term. The noise and track monitoring system at the airport shows results in good agreement with the aircraft recorder data. Actual flight profiles could therefore be taken directly from the NTMS into GMTIM. The implementation calls for the development of algorithms to estimate applied thrust for each flight. This effort is therefore seen as an “intermediate term measure”.

43

6 Recommendations for further Work This chapter discusses a number of actions to be taken in order to minimise the difference between measured and predicted aircraft noise levels. In order to sort different actions, the presentation is divided into measures on a short-, intermediate- and long term time scale.

6.1 Short term measures The time scope for these implementations are 3 months.

6.1.1 Correcting lateral attenuation The correction should be based on the results obtained in the present investigation combined with corresponding results from recent international investigations. The new routines will have an intermediate status until international recommendations are updated. Present proposal will divide aircraft in three categories: propeller driven aircraft, wing mounted turbofan engine aircraft and ditto fuselage mounted. Implementation of the new algorithms into NORTIM/GMTIM will also affect the routine that handles terrain topography, which must be properly adjusted. When the implementation is accomplished, the final step is to test the corrected noise level predictions against measurement results. This test will be performed on the base of measured noise levels used in earlier investigations [1,2].

6.1.2 Corrections to Noise-Power-Distance data The deviations that are statistically significant should be implemented in the database. For the B737-600/700 the comparison should be regenerated with new sets of data (INM 6.0c database for B737-700). New statistical analysis is also recommended for the MD80 family, where the aircraft should be grouped together.

6.1.3 Adjustment to flight profiles Flight profiles as registered by the aircraft flight recorder combined with the airport radar system can be implemented in the NORTIM/GMTIM database. Empirically based takeoff profiles can be developed for the limited numbers of aircraft in this study. These aircraft are the dominant types at Oslo Airport and such an implementation would improve the modelling accuracy. Standard profiles can also be improved with information from the airline companies on what set of procedures are in use. This input can be used to create new database profiles.

6.2 Intermediate term measures These actions can be accomplished within 12-15 months.

6.2.1 Flight profiles from the monitoring system Further investigations need to be initiated to determine the difference between height reported by the radar transponder and the flight recorder. As pointed out in chapter 4.6.3 time synchronisation may be the cause. When this is sorted out, take off profiles from the track and monitoring system may be imported into GMTIM and used instead of the standard profiles. An implementation must be supplemented with algorithms to calculate what thrust have been applied to fly the profile. Take off weight for each flight will probably be one needed input parameter.

44

For landing profiles it is recommended that a small study be initiated to firstly determine the nature of the varying thrust applied, secondly to develop routines to model this in standard profiles. These routines will have a more generic character and be applicable to other noise models.

6.2.2 Investigate effects of flight path segmentation The noise level data collected in the present investigation represent a potential for further analysis of differences between measured and predicted aircraft noise levels. An evident task is to check the flight segmentation procedures used in NORTIM/GMTIM in order to investigate the applicability of a point source model in aircraft noise level prediction.

6.3 Long term measures Relevant tasks are not described in detail, but obvious activities are to prepare for a reconsidering of the aircraft noise level prediction basis, including: • New aircraft noise source description, possible point source. • New frame of noise emission data, including 3D directivity. • New ground attenuation/topography model/weather model. • Harmonising with international development on noise level prediction regarding community

noise.

45

7 References [1] R. T. Randeberg: “Validating GMTIM", SINTEF Memo 40-2000-5930, 2000-10-18. (In

Norwegian.) [2] R. T. Randeberg: “Analysis of differences between measured and calculated aircraft noise

levels at Gardermoen - Statistical analysis of existing data", SINTEF Memo 40-2001-2698, 2001-05-15.

[3] Nordtest Method NT ACOU 104: "Ground Surfaces: Determination of the Acoustic Impedance". Approved 1999-11.

[4] ISO 9613-1:1993, " Acoustics -- Attenuation of sound during propagation outdoors -- Part 1: Calculation of the absorption of sound by the atmosphere”.

[5] SAE AIR 1751:"Prediction Method for Lateral Attenuation of Airplane Noise During Takeoff and Landing", 1981.

[6] SAE AIR 1845:"Procedure for the Calculation of Airplane Noise in the vicinity of Airports", March 1986.

[7] K. J. Plotkin et al.: ”Examination of the Lateral attenuation of Aircraft Noise”. NASA/CR-2000-210111, April 2000.

[8] D. A. Senzig et al.: ”Lateral Attenuation of Aircraft Sound Levels Over an Acoustically Hard Water Surface: Logan Airport Study”. NASA/DR-2000-210127, DOT-VNTSC-NASA-00-01, May 2000.

[9] D. A. Senzig et al.: ”Measured Engine Installation Effects of Four Civil Transport Airplanes”. Proc. NOISE-CON 2001, Portland, Maine, October 29-31.

[10] J. A. Page et al.: ”Validation of Aircraft Noise Prediction Models at Low Levels of Exposure”. NASA/CR-2000-210112, April 2000.

[11] N. P. Miller et al.: ”Examining INM Accuracy Using Empirical sound Monitoring and Radar Data”. Harris, Miller, Miller & Hanson Inc., Report No. 294520.03, October 1999.

[12] B. Plovsing, J. Kragh: “Nord2000. Comprehensive Outdoor Sound Propagation Model”, 31. Dec. 2001. Part 1: ‘Propagation in an Atmosphere without Significant Refraction’, Lyngby 2000 Part 2: ‘Propagation in an Atmosphere with Refraction’, Lyngby 2000.

[13] Personal communication with Dr. Hans Olav Hygen, Det Norske Meteorologiske Institutt. [14] T. Conner et al.: ”Accuracy of the Integrated Noise Model (INM): MD-80 Operational

Noise Levels”. Joint report of Swedish CAA and US DoT, Stockholm/Washington May 1995.

[15] K. H. Liasjø, I. L. N. Granøien: “Comparison of Aircraft Noise Programs INM-2/6, INM-3/9, INM-3/10, DANSIM and NOISEMAP”, SINTEF Report STF40 A93043. (In Norwegian.) Trondheim, April 1993.

[16] H. Olsen et al.: “Topography influence on aircraft noise propagation, as implemented in the Norwegian prediction model, NORTIM”, SINTEF Report STF40 A95038. Trondheim, May 1995.

[17] “Noise and track monitoring system”, Technical specification in the Contract: OHAS-N-SP-0005 E03 - C-16726, Oslo September 1997. Personal communication with Knut Holen, Manager HSE, Oslo Airport Gardermoen.

46

Appendix 1 Measurement site and measurement positions.

The measurement positions 1 - 5 are indicated. In the following a picture of the equipment and site for all the 5 positions.

1

2 3 4 5

1 1

2 3 4 5

47

Position 1

Position 2

48

Position 3

Position 4

49

Position 5

50

Schematic layout of the measurement system.

POS. 2 B C

POS. 3 D E

POS. 1 NMT4 A

POS. 4 F G

POS. 5 H I

POS. 0 (REF) 220VAC

HEADQUARTER 2 PC w/ CD-recorder Batteries, Chargers Spare sound analyser Spare microphones Calibrators Ground Impedance Equipment Spare parts Tools Cables, Tape etc

LIST OF SYMBOLS

Microphone Sound Analyser Battery PC Trigger System GPS antenna Voltage adapter Meteorology sensor Meteorology logger Mast

R U N W A Y

51

Appendix 2 Meteorological data

Picture shows the 10 meters mast (type Clark QTM) with meteorological sensors. The following sensors were coupled to the meteorology logger Aanderaa Type 3660:

• Wind direction (°) at 9.25 m. Sensor adjusted with compass. • Mean and max wind (m/s) speed at 9.25 m. • Air temperature (°C) at 9.25 m. • Temperature difference (°C) between 9.25 and 0.5 m. • Net Radiation (W/m2). • Relative air humidity (%RH) at 9.25 m. • Precipitation (mm) at 1.5 m.

52

Graphical presentation of wind direction and wind speed (values representing mean speed and gust)

53

Graphical presentation of air temperature (at 9.25 m above ground), relative humidity and rainfall.

54

Graphical presentation of air temperature difference between 9.25 and 0.5 m height and net radiation.

55

Appendix 3 Estimation of ground impedance by acoustical measurement. Typical measurement layout.

15° 500 500 200 35 1750 mm

sintef report · sintef report title aircraft noise measurements at gardermoen airport, 2001. part...

Documents