Method for Estimating Aircraft Noise Corresponding to Changes in Thrust and Speed during Start of Takeoff Roll Toshiyasu NAKAZAWA 1 Naoaki SHINOHARA Organization of Airport Facilitation 1-3-1-5F, Shiba-koen, Minato-ku, Tokyo, Japan Kazuyuki HANAKA Narita International Airport Promotion Foundation Chiba, Japan

ABSTRACT In aircraft noise calculation by international guidelines such as ICAO Doc.9911, NPD (Noise Power Distance) data that relates aircraft thrust and noise are used. In these guidelines net-thrust use as power in NPD data. However, net-thrust of an aircraft engine cannot be measured directly. Therefore, net-thrust has to be estimated using the conversion equation with engine parameters such as N1% RPM and EPR, as well as speed. Meanwhile, for recent aircraft types, the coefficients for the N1% and speed terms of the conversion equation are not available. For this reason, Japanese aircraft noise models use N1% and EPR directly in place of net-thrust to create NPD. These parameters change the most at the start of the takeoff roll, and it is necessary to confirm the validity of the noise calculation in this area. In this study, noise measurements at the side of the runway and aircraft onboard data were analyzed to investigate an alternative method of estimating noise changes to net-thrust. The result shows that it is reasonable to correct the noise according to the speed change in addition to the noise calculation in the section of lower speed state of takeoff roll.

1. INTRODUCTION

Aircraft require the strongest thrust for takeoff, and the noise level is very high especially during the acceleration from a standing start. In other words, the noise impact around the runway during takeoff is particularly large. Therefore, it is necessary to evaluate the noise impact around the runway with validity. As part of this process, noise calculation has been conducted at various airports.

However, for recent generation aircraft, the sound quality may have changed due to technological improvements, such as an increase in the engine bypass ratio and the introduction of geared turbofan engines. For this reason, the validity of the calculation needs to be confirmed. As part of the validation activity, noise directivity was measured in November 2020 behind the start of the takeoff roll. We reported last year that the behind directivity of noise far from the aircraft was found to be slightly different from that of the conventional correction method and may need to be

1 t-nakazawa@aeif.or.jp

Jai. inter noise 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O ? . GLASGOW

improved [1]. This paper reports on the results of simultaneous measurements made in November 2020 for the purpose of validating the noise at the side of the aircraft during the takeoff roll.

1.1. Takeoff Roll Procedure

In the normal takeoff roll procedure, after entering the runway, the aircraft accelerates to lift off. In some cases, the aircraft comes to a complete stop on the runway before beginning the takeoff roll (standing takeoff), while in other cases, the aircraft may begin its takeoff roll by taxiing off the taxiway and not stopping even on the runway (rolling takeoff). Rolling takeoff has a slightly higher initial thrust and speed, but the aircraft is immediately in the same state as standing takeoff.

Typically, in the case of passenger jet aircraft, the control of the aircraft thrust is delegated to a system called auto-throttle or auto-thrust, which is operated by the flight control computer a short time after the aircraft starts its takeoff roll. In this case, the system fixes engine parameter such as engine rotation ratio (N1%) and engine pressure ratio (EPR) at specified values before the aircraft reaches 80 kts. Therefore, the engine parameters remain constant until liftoff, although the thrust increases for about 150 meters after the start of the takeoff roll. Regarding the change in speed, the aircraft continues to accelerate during the takeoff roll. Immediately after the start of the takeoff roll, the aircraft accelerates largely, and then, although the acceleration gradually decreases, the aircraft continues to accelerate and lifts off at about 140 ~ 160 kts.

1.2. Calculation Method for Takeoff Roll

According to existing noise calculation guidelines, such as ICAO Doc. 9911 [2] and ECAC Doc. 29 [3], the noise at the aircraft during takeoff roll is determined primarily by the aircraft status: horizontal path, altitude, speed, and thrust. These are the same as in flyover. In a takeoff roll, the horizontal path and altitude can be assumed to go through the centerline on the runway. The speed affects the L AE noise with a duration that depends on the length of the finite length segment. These segments during the takeoff roll are to be subdivided in order to reduce the effect of large speed changes in the calculation guidelines.

For thrust, there are problems with obtaining it. First, in order to obtain thrust, onboard data is necessary, but it is difficult to obtain without a relationship with airlines. In comparison, other values such as altitude and speed can be easily obtained and confirmed externally with ADSB data. Second, corrected net thrust cannot be measured directly. In the noise calculation guidelines, it is described that this corrected net thrust of the aircraft engine is related to the noise. The corrected net thrust is estimated by calculation using the engine parameters from the onboard data and the measured environmental parameters. Equations 1 and 2 show the calculation method in the existing calculation guideline ECAC Doc. 29 appendix B.

𝐹 ௡ 𝛿 ⁄ = 𝐸+ 𝐹∙𝑉 ஼ + 𝐺 ஺ ∙ℎ+ 𝐺 ஻ ∙ℎ ଶ + 𝐻∙𝑇+ 𝐾 ଵ ∙𝐸𝑃𝑅+ 𝐾 ଶ ∙𝐸𝑃𝑅 ଶ (1)

ଶ

𝐹 ௡ 𝛿 ⁄ = 𝐸+ 𝐹∙𝑉 ஼ + 𝐺 ஺ ∙ℎ+ 𝐺 ஻ ∙ℎ ଶ + 𝐻∙𝑇+ 𝐾 ଷ ∙൬ 𝑁 ଵ

൰+ 𝐾 ସ ∙൬ 𝑁 ଵ

(2)

൰

√𝜃

√𝜃

where:

F n / δ is the corrected net thrust per engine, lbf V C is the calibrated airspeed, kt T is the ambient air temperature in which the aeroplane is operating, °C h flight altitude, ft E, F, G A ,

are engine thrust constants or coefficients for temperatures below the engine flat rating temperature at the thrust rating in use (on the current segment of the takeoff/climbout or approach flight path), lb.s/ft, lb/ft, lb/ft 2 , lb/°C. Obtainable from the ANP database. EPR is defined by some manufacturers as engine pressure ratio. K 1, K 2 are coefficients from the ANP database that relate corrected net thrust and engine pressure ratio in the vicinity of the engine pressure ratio of interest for the specified aeroplane Mach number.

G B , H

N 1 is the rotational speed of the engine’s low-pressure compressor (or fan) and turbine stages, % θ = (T + 273)/288.15, the ratio of the absolute total temperature at the engine inlet to the absolute standard air temperature at mean sea level K 3, K 4 are constants derived from installed engine data encompassing the N 1 speeds of interest. It can be assumed that there are no changes in altitude or air pressure, which are variables in this equation, as long as the aircraft is moving within the runway. Therefore, what should be considered in this equation are the engine parameters such as N1 and EPR and the change in speed. For speed, the corrected net thrust will decrease as the speed increases because the coefficient F in the equation is usually negative. N1 and EPR, on the other hand, do not vary except during the first section of the takeoff roll, since they increase and stabilize soon after the start of the run.

se Scorn Gm

ECAC Doc. 29 states that the coefficients used in this equation can be obtained from the ANP database [4], but for takeoffs of recent aircraft types, only the coefficients for maximum thrust takeoffs are published. In reality, many airlines operate using thrust reduction, which means that noise calculations are overestimated. In addition, the latest aircraft types that are operated more frequently, such as the A320neo, are not published. For this reason, Japanese aircraft noise models use N1%, EPR, etc. as they are in place of corrected net thrust when creating NPD for new aircraft types through measurement. However, the relationship with the noise value has not been sufficiently verified, and this method needs to be validated. Therefore, noise was measured during the takeoff roll and its variation with aircraft condition was investigated.

2. MEASUREMENT

Aircraft noise during takeoff was measured around Narita Airport in Japan. The noise to be measured was takeoff in the 34L direction (toward north) of runway-A, and the measurement was carried out one week during from October 28th to November 3rd, 2020. 576 aircraft took off in direction 34L, and these noises were measured. These were all passenger jet aircraft. Unfortunately, during these periods, the number of passenger aircraft was reduced due to the pandemic. So, there was a possibility of a slightly lighter takeoff weight, but other than that, the takeoff procedure was the same as usual. During the measurement, it rained for about one day, but the rest of the time was cloudy or sunny, the average temperature was 16.2 C, and the wind was mostly calm with less than 6 meters / sec.

<= runway-A 34L

205 m

eeceo'

O :measurement points

150 m

Figure 1: Location of noise measurement points (Narita airport, Japan)

Red point: measurement points, Blue: planning lines at 150 m intervals, Green: planning lines 205 m from runway

To measure the noise during the takeoff roll, multiple sound level meters were placed along the side of the runway. Assuming that the aircraft runs in the center of the runway, the distance between the measurement points and the aircraft can be kept constant by placing them parallel to the runway centerline. The measurement points needed to be as close to the runway as possible to remove the effects of the ground surface and weather. In this measurement point layout, the distance was

planned to be 205 meters from the runway centerline to obtain safety permission from the airport administrator and to ensure that the measurement points could be aligned on the same line. In addition, in order to measure the noise corresponding to changes in the condition of the aircraft, 11 measurement points were planned to be located at intervals of 150 meters from the taxiway to the end of the runway and from where the aircraft was expected to line up in the direction of the runway, covering a distance of 1500 meters.

However, in the actual measurement, the measurement points were placed slightly off from the planned locations because of the difficulty of placement due to the indentations in the ground and the lack of precise surveying of the placement points. The most deviated point was point 4, which had to be placed 5 meters away from the runway and 25 meters in the direction of the runway from the planned position in order to keep it away from the old taxiway. The next deviated point was at point 10, which was placed about 25 meters off in the direction of the runway. The others were able to be placed within 10 meters of the planned position.

A sound level meters (Rion NA-28 or NL-62/52) were placed at these locations at a microphone height of 1.5 m. The L AS and 1/3-octave bands every 0.1 second were continuously recorded on a memory card in the sound level meter. In addition, the ADSB emitted by the aircraft was received to obtain the aircraft's hourly position and speed. Operational logs were also obtained to identify departure time, flight number, and aircraft type. Meteorological data were obtained to determine temperature, pressure, wind direction, wind speed, and precipitation.

3. ANALYSIS

The measured noise results were separated into individual flights and matched the flight operation logs based on the time of day and other factors. These were classified into four aircraft types (A320, B787, B777, and B767) for which a sufficient number of operations could be measured. L ASmax and L AE were calculated for each flight and measurement point. Next, we attempted to determine the noise value at the approach position to each measurement point. Unfortunately, the ADSB times and the times of the 11 sound level meters were not exactly synchronized with enough precision to accommodate the speed of the aircraft. Therefore, the maximum noise value at each measurement point for each individual flight was assumed to be the noise value at the closest point of the aircraft. In addition, the speed at the closest point of the measurement point for each flight was also picked out.

First, simply looking at the data by location for each flight operation, noise varied but tended to drop slightly, and for speed, there was a continuous acceleration. The degree of these changes was not constant, and the acceleration was greater at the point immediately after the start of the run. No aircraft lifted off the ground at any of the locations.

The reason for the large variation in noise values is that different aircraft types are included, as well as variations in reference takeoff thrust due to actual takeoff weight and weather conditions even within the same type of aircraft. Since the purpose of this study is to check the variation within individual operations, it is necessary to remove these effects. Therefore, it is better to use some single point as a reference point and look at the changes in noise when passing through each point for each individual operation. Assuming that the positional relationship between the aircraft and the measurement point is constant, the L ASmax , which does not take the duration into account, should give a clearer picture of the change in noise due to changes in the condition of the aircraft than the L AE , which takes the duration into account.

Therefore, by checking the difference with respect to point 10, which is estimated to have already reached the specified takeoff thrust, the tendency of the noise value to change was confirmed. Since the influence of speed variation was considered to be mainly significant, the difference noise for each point and the speed were plotted in Figure 2. In addition, a regression line is also drawn.

Figure 2 shows that the noise is decreasing as the aircraft speed increases. This may be due to the effect of the corrected net thrust decreasing in accordance with the change in speed, and the noise may be decreasing in accordance with the takeoff roll distance. On the other hand, the effect of

10

10

5

5

LAmax relative to pt.10(dB)

LAE relative to pt.10(dB)

0

0

-5

-5

-10

-10

20 60 100 140 180

20 60 100 140 180 B767 B777 B787 A320

GS(kt)

GS(kt)

B767 B777 B787 A320

Relative L ASmax Relative L AE

Figure 2: Relationship between noise level and speed during takeoff roll Horizontal: Ground Speed (kt), Vertical: Relative noise from point 10 for each operation (dB) and EPR changes on noise was considered to be observable at points 0, 1, and 2 immediately after the start of roll. However, these locations were strongly affected by the jet engine noise directivity behind the diagonal, in addition to the large effect of speed change, and there were many cases where the maximum noise value appeared after the passage. Therefore, a clear relationship between changes in N1% and EPR and noise values could not be confirmed.

(dB diff.)

10

5

0

-5

-10

20 60 100 140 180 B767 B777 B787 A320 (kt)

Figure 3: Relative L AE removed duration correction of noise calculation guideline

Horizontal axis Ground Speed, Vertical axis relative L AE in Figure 2 minus log ଵ଴ (160𝑘𝑡/𝐺𝑟𝑜𝑢𝑛𝑑𝑆𝑝𝑒𝑒𝑑)

In an attempt to remove the effect of duration on L AE , log ଵ଴ (160𝑘𝑡/𝐺𝑟𝑜𝑢𝑛𝑑𝑆𝑝𝑒𝑒𝑑) was subtracted from the relative L AE , following the method of correcting for duration in the noise calculation guidelines (Figure 3). However, unlike the L ASmax trend, the slope of the regression line has disappeared. Since the change in speed is large during the takeoff run of an aircraft, it is

presumably difficult to represent the duration before and after the closest approach to the measurement point in terms of ground speed. Note that this is not a problem in actual noise calculation, since segments are subdivided at each constant speed change to reduce the effect of the speed change.

Finally, a regression equation was obtained from the L ASmax data in Figure 2. Since it is more convenient to use the reference speed V ref =160 kt in the noise calculation as a reference, the regression equation in Equation 3 was calculated by converting it to a 160 kt reference rather than using the speed at point 10 as the reference.

∆𝐿 ஺ = −0.02268𝑉+ 2.53050 (3) where ∆𝐿 ஺ is the noise value correction for the modified net thrust in response to changes in speed, V is the speed kt

Using this correction equation, 2.3 dB is added at a speed of 10 kt, and it becomes zero at a speed of 160 kt.

4. CONCLUDING REMARKS

Noise measurements were carried out at the side of the runway at Narita Airport in Japan on 2020/11 to validate the takeoff roll calculation method. Eleven sound level meters were placed on a line 205 m apart on runway 34L, at intervals of 150 m, over a distance of 1500 m. The purpose was to validate the validity of noise calculations when coefficients for corrected net thrust calculations were not available. From the results, L ASmax and L AE for each individual flight were analyzed to see the relationship with the status of each aircraft.

The results show that the noise varies with takeoff roll distance. Considering the aircraft status, it can be assumed that speed change is the cause, since the noise is increasing even in the section where the engine parameters such as N1% and EPR are assumed to be constant. Finally, a regression equation between speed and noise was obtained for possible use in noise calculation.

Correcting for noise due to speed may result in more valid noise calculation results. However, the amount of correction by the regression equation is 2.3 dB additional at 10 kt just after the start of roll, which is not so large considering that the duration correction by the calculation guideline is as much as 12 dB (= log ଵ଴ (160𝑘𝑡/10𝑘𝑡) ). Also, this formula is only an estimate based on measurements during the takeoff roll, and should not be applied after liftoff. Since this is only an alternative method, it would actually be better to get the coefficients for calculating corrected net thrust from the manufacturer. 5. ACKNOWLEDGEMENTS

Authors are grateful to Narita International Airport Corporation (NAA), for their kindly supporting to carry out the aircraft noise measurements around Narita airport. 6. REFERENCES

1. Nakazawa T. et al., Study on aircraft noise directivity of behind the start of takeoff roll, Proc. INTERNOISE 2021, (2021) 2. International Civil Aviation Organization, Recommended Method for Computing Noise Contours Around Airports , ICAO Doc 9911 2nd edition, 2018. 3. European Civil Aviation Conference, Report on Standard Method of Computing Noise Contours Around Civil Airports , ECAC Doc 29 4th Edition, 2016. 4. EUROCONTROL, The Aircraft Noise and Performance (ANP) Database, Retrieved April 21, 2022, from https://www.aircraftnoisemodel.org