A A A Study on the ground attenuation of engine run-up and APU noise for developing the airport noise model in Japan Takatoshi Yokota 1 , Koichi Makino 2 Kobayasi Institute of Physical Research 3-20-41 Higashi-Motomachi, Kokubunji, Tokyo 185-0022, Japan Toshiyasu Nakazawa 3 , Masayuki Sugawara 4 Naoaki Shinohara 5 Aviation Environment Research Center, Organization of Airport Facilitation 1-3-1 Shiba-koen, Minato-ku, Tokyo 105-0011, Japan Kazuyuki Hanaka 6 Narita International Airport Promotion Foundation 76 Amadutsumi Kayamashinden, Shibayama, Sanbu, Chiba 289-1601 Japan ABSTRACT The correction for ground effect around airports is modeled based on the results of numerical anal- ysis to estimate the propagation of noise caused by ground operations in airports. The excess atten- uation due to ground effects on noise of engine run-ups and APU is calculated using a PE method for typical jet aircraft and propeller aircraft. On the assumption that the ground in and around an airport consists of a mixed surface of hard and soft surfaces, the excess attenuation for the mixed impedance ground is calculated with the Fresnel-zone method using the results of the PE calculation. The mixed ratio of ground surface is assumed based on the ratio of asphalt-paved surface and grass- covered surface along the propagation path from the engine run-ups spot to the airport site boundary of six major airports in Japan. In this paper, we introduce the excess attenuation calculated with the PE method. Based on the calculation results, the correction model for ground effects on noise of engine run-ups and APU is discussed to implement it in the airport noise prediction model used in Japan. 1. INTRODUCTION Ground-to-ground (GtoG) aircraft noise propagation is one of the major interests in the field of air- craft noise prediction around airports. Especially in Japan, the prediction of GtoG propagation has become more important because such ground activities as engine run-ups, taxiing and APU operation have been included in the items of the evaluation of noise impact around airports. As a method of predicting GtoG aircraft noise propagation, AIR5662 [1] has been adopted in ECAC-CEAC Doc 29 [2] and ICAO Doc 9911 [3] which are the common methods of computing noise contours around airports. In AIR5662, the attenuation due to ground effects for GtoG propagation is 1 t-yokota@kobayasi-riken.or.jp 2 makino@kobayasi-riken.or.jp 3 t-nakazawa@aeif.or.jp 4 m-sugawara@aeif.or.jp 5 n-shinohara@aeif.or.jp 6 hanaka@napf.or.jp 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW described as over-ground attenuation. By using the method, the over-ground attenuation for A- weighted sound pressure level can be calculated even at an elevation angle of 0 degree, however, the model was developed based on the frequency characteristics of aircraft noise under straight, steady, level flight conditions. In addition, it has been confined to the case of sound propagation over soft level ground, whereas the ground surface in airport sites consists of not only soft absorptive surfaces but also hard reflective surfaces. The assumption for calculating ground attenuation of ground activ- ities such as engine run-ups and APU is deviated from the assumptions in AIR5662, regarding fre- quency characteristics of sound source and ground surface conditions. In this study, we have assumed the ground surface in and around an airport consists of a mixed surface of hard and soft surfaces. The ratio of asphalt-paved surface and grass-covered surface along sound propagation path of six major airports in Japan has been researched. The excess attenuation for the mixed impedance ground has been calculated with the Fresnel-zone method using the calculation results of a Parabolic Equation (PE) method. In this paper, the correction model for ground effects on noise of engine run-ups and APU is discussed to implement it in the airport noise prediction model used in Japan. 2. BASIC EQUATION FOR PREDICTING GROUND ACTIVITY NOISE Equation (1) is a basic equation to predict airport ground activity noise at a prediction point r [m] from a sound source. 𝐿 ሺ𝑟ሻൌ𝐿 ሺ1ሻെ10log ଵ 𝑟 ଶ ∆𝐿 ,ୟ୧୰ ሺ𝑟ሻ∆𝐿 ,୰୬ୢ,୫ୣ୲ ሺ𝑟ሻ (1) where 𝐿 ሺ𝑟ሻ [dB] is the A-weighted sound pressure level at a distance of r [m] from a sound source, 𝐿 ሺ1ሻ [dB] is the A-weighted sound pressure level converted to the value at the reference distance of 1 m, ∆𝐿 ,ୟ୧୰ ሺ𝑟ሻ [dB] is the correction for atmospheric absorption, ∆𝐿 ,୰୬ୢ,୫ୣ୲ ሺ𝑟ሻ [dB] is the correc- tion for ground and meteorological effects. In the airport noise prediction model used in Japan, Equa- tion (1) is implemented as 𝐿 ሺ𝑟ሻൌ𝐿 ,ୈ ሺ𝑟ሻ∆𝐿 ,୰୬ୢ,୫ୣ୲ ሺ𝑟ሻ (2) where 𝐿 ,ୈ ሺ𝑟ሻ is the A-weighted sound pressure level at a distance of r [m] from a sound source obtained from the NPD data which has been prepared by measurements. The ∆𝐿 ,୰୬ୢ,୫ୣ୲ ሺ𝑟ሻ for each aircraft is calculated using the following equations. 𝐿 ,୵୧୲୦୰୭୳୬ୢ ሺ𝑟ሻൌ𝐿 ሺ1ሻെ10log ଵ 𝑟 ଶ ∆𝐿 ,ୟ୧୰ ሺ𝑟ሻ∆𝐿 ,୰୬ୢ,୫ୣ୲ ሺ𝑟ሻ (3) 𝐿 ,୰ୣୣ ሺ𝑟ሻൌ𝐿 ሺ1ሻെ10log ଵ 𝑟 ଶ ∆𝐿 ,ୟ୧୰ ሺ𝑟ሻ (4) 𝐿 ,୵୧୲୦୰୭୳୬ୢ ሺ𝑟ሻൌ10 ,౭౪ౝ౨౫ౚ ሺሻା∆ ଵ (5) 𝐿 ,୰ୣୣ ሺ𝑟ሻൌ10 ,౨ ሺሻା∆ ଵ (6) ∆𝐿 ,୰୬ୢ,୫ୣ୲ ሺ𝑟ሻൌ𝐿 ,୵୧୲୦୰୭୳୬ୢ ሺ𝑟ሻെ𝐿 ,୰ୣୣ ሺ𝑟ሻ (7) 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW where f [Hz] is the center frequency of 1/3 octave band, the subscript f denotes the frequency band , 𝐿 ,୵୧୲୦୰୭୳୬ୢ ሺ𝑟ሻ [dB] is the sound pressure level at a distance of r [m] from a sound source in a hemi- free field, 𝐿 ,୰ୣୣ ሺ𝑟ሻ [dB] is the sound pressure level at a distance of r [m] from a sound source in a free field, 𝐿 ሺ1ሻ [dB] is the sound pressure level converted to the value at the reference distance of 1 m, ∆𝐿 ,ୟ୧୰ ሺ𝑟ሻ is the correction for atmospheric absorption, ∆𝐿 ,୰୬ୢ,୫ୣ୲ ሺ𝑟ሻ [dB] is the correction for ground and meteorological effects, ∆A [dB] is the A-frequency-weighting for the frequency band of f [Hz]. 3. CALCULATION AND MODELING OF GROUND ATTENUATION FOR JAPANESE AIPORT NOISE PREDICTION MODEL 3.1. Modeling of the ground surface conditions in and around airports To model the ground surface condition assumed in the noise prediction model, the current ground conditions in and around six major airports in Japan, where the engine run-ups are conducted, were researched by using aerial photographic maps. In the airport sites, we checked the sound propagation path from a target source position (engine run-ups position) to an airport boundary adjacent to the nearest residential area. It has been confirmed that the types of ground surfaces in the airport sites are mainly composed of asphalt-paved surfaces and grass-covered surfaces. Table 1 shows the length of the sound propagation path and the mixed ratio of asphalt-paved surface and grass-covered surface along the path. In contrast, in the surrounding area outside of the airport sites, it has been confirmed that the types of ground surfaces and the density of buildings vary according to the location of the airports. Therefore, we have decided to set a soft level ground which is assumed in the general aircraft noise prediction model AIR5662 as the initial model of ground condition around airports. From the results, we have decided to model the ground surface as follows: the surface in the range of 500 m from a sound source is a mixed impedance surface of the asphalt-paved (hard) surface and the grass-covered (soft) surface and its ratio is 80: 20, the surface in the range over 500 m from the source is a soft surface. Table 1: Ground surface conditions in the six major airports in Japan. Airports Distance from the source position to the airport site boundary [m] Mixed ratio Asphalt-paved Grass-covered A 570 100% 0% B 128 92% 8% C 902 64% 36% D 614 100% 0% E 214 100% 0% F 601 50% 50% Average 505 79% 21% model 500 80.0% 20.0% 3.2. Frequency characteristic and height of target sound sources Table 2 shows the target operations and aircraft types, and assumed source heights. The source height was assumed at the height of the center point of main engine or at the height of the APU exhaust. Figure 1 shows the 1/3 octave band sound pressure level under consideration, which is converted to 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW the value at the reference distance of 1 m and the overall sound pressure level of each aircraft is set to be 0 dB, respectively. Table 2: Sound sources under consideration. Target sound source Source Height [m] Operation Aircraft type Engine run-ups (E) A321neo 1.98 Engine run-ups (E) B767-300 2.13 Engine run-ups (E) B787-8GE 2.74 Engine run-ups (E) B777-300PW 3.02 Engine run-ups (E) DHC8-400 3.96 APU (A) B767-300 4.88 APU (A) B787-8GE 6.10 A321neo_E B767-300_E B787-8GE_E B777-300PW_E DHC8-400_E B767-300_A B787-8GE_A 0 Relative SPL [dB] -10 -20 -30 -40 63 125 250 500 1k 2k 4k 1/3 octave band center frequency [Hz] Figure 1: Frequency characteristics of target sound sources. 3.3. Calculation of ground attenuation The excess attenuation under the condition that sound propagates above an impedance ground was calculated using Green’s Function Parabolic Equation (GF-PE) method [4]. In the cases that sound propagates above a mixed impedance ground, the excess attenuation was calculated with the Fresnel- zone method [5] using the results of GF-PE calculation. In the GF-PE calculation, meteorological conditions were assumed as in Table 3. Although the effects of atmospheric turbulence were not considered in the GF-PE calculation, a lower limit for the attenuation was set to -20 dB for each frequency band. Figure 2 shows the calculation condition. The source height was set at the height shown in Table 2 and receiving points were set at a height of 1.2 m and 50 m intervals. The surface of the ground within the range of 500 m from the sound source was assumed as the mixed impedance surface modeled in Section 3.1 (Hard: 80%, Soft: 20%), and the ground was assumed as the soft surface over the range of 500 m. The impedance of the ground surface was modeled by the Miki’s impedance model [6] and the effective flow resistivity of the hard surface was assumed to be 20,000 kPa s/m 2 and 300 kPa s/m 2 for the soft surface. 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW Table 3: Meteorological conditions under consideration. Vector wind speed No wind : 0 m/s Downwind: +2 m/s, +5 m/s Upwind : -2 m/s, -5 m/s Wind speed profile Logarithmic profile Roughness length: 0.3 m (Soft), 0.01 m (Hard) Temperature, Relative humidity, Atmospheric pressure 25 °C, 70 %, 1013.25 hPa Temperature gradient No gradient (0.0 °C / 100 m) Upwind Area Downwind Area Wind speed: 0, 2, 5m/s with logarithmic profile Sound source Height: position of engine center or APU Receiving points: 50 m intervals, height = 1.2 m 0 500 m 500 m soft ground surface soft ground surface Mixed ground with hard (80 %) and soft (20 %) surfaces Figure 2: Calculation condition. Figure 3 shows the calculated ∆𝐿 ,୰୬ୢ,୫ୣ୲ for each aircraft in case sound propagates above the mixed ground assumed in the section 3.1 under “No wind” condition. In the figure, the results are shown in three categories, engine run-up of jet aircraft, engine run-up of propeller aircraft and APU. In each category, the difference in ∆𝐿 ,୰୬ୢ,୫ୣ୲ due to aircraft types is small. Therefore, the arithmetic average of ∆𝐿 ,୰୬ୢ,୫ୣ୲ in each distance for each category is used in the following discussion. Figure 4 shows the ∆𝐿 ,୰୬ୢ,୫ୣ୲ for each category under each wind speed condition. Under no wind and downwind conditions, the tendency of the variation in ∆𝐿 ,୰୬ୢ,୫ୣ୲ against distance is changed at 500 m from the sound source. On the other hand, the ∆𝐿 ,୰୬ୢ,୫ୣ୲ is smoothly changed with distance and it reaches the minimum value in the relatively short range in cases under upwind conditions. It indicates that the effects of hard ground near the sound source is more apparent in cases under no wind and downwind conditions than upwind conditions. Under downwind conditions, it is also seen that the difference in ∆𝐿 ,୰୬ୢ,୫ୣ୲ due to wind speed and the variation in ∆𝐿 ,୰୬ୢ,୫ୣ୲ due to distance are small and ∆𝐿 ,୰୬ୢ,୫ୣ୲ is about –5 dB at the distance of 3 km. On the other hand, under downwind conditions, the difference in ∆𝐿 ,୰୬ୢ,୫ୣ୲ due to wind speed is large in the short range until ∆𝐿 ,୰୬ୢ,୫ୣ୲ reaches the minimum value. Focusing on the difference in sound source category, the attenuation for engine run-ups of propel- ler aircraft is smaller than that of jet aircraft, because the dominant frequency components of engine run-ups noise generated from propeller aircraft are lower than those of jet aircraft. It is also seen that the attenuation for APU is smaller than that for engine run-ups of jet aircraft, because the source height of APU is higher than that of the engine run-ups of jet aircraft. 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW 10 10 5 5 0 0 Δ L grnd,met [dB] Δ L grnd,met [dB] -5 -5 -10 -10 -15 -15 A321neo B767 B787 B777 Average -20 -20 -25 -25 DHC8 100 200 300 400 500 1000 2000 3000 -30 -30 100 200 300 400 500 1000 2000 3000 horizontal range [m] horizontal range [m] (a) Jet aircraft engine run-ups (b) Propeller aircraft engine run-ups 10 5 0 Δ L grnd,met [dB] -5 -10 -15 B767APU B787APU Average -20 -25 -30 100 200 300 400 500 1000 2000 3000 horizontal range [m] (c) APU Figure 3: Attenuation under “No wind” condition for each source condition. 10 10 5 5 0 0 Δ L A,grnd,met [dB] Δ L A,grnd,met [dB] -5 -5 -10 -10 -15 -15 -20 -20 -5 m/s -2 m/s No wind +2 m/s +5 m/s 100 200 300 400 500 1000 2000 3000 -30 -5 m/s -2 m/s No wind +2 m/s +5 m/s 100 200 300 400 500 1000 2000 3000 -25 -25 -30 horizontal range [m] horizontal range [m] (a) Jet aircraft engine run-ups (b) Propeller aircraft engine run-ups 10 5 0 Δ L A,grnd,met [dB] -5 -10 -15 -20 -5 m/s -2 m/s No wind +2 m/s +5 m/s 100 200 300 400 500 1000 2000 3000 -25 -30 horizontal range [m] (c) APU Figure 4: Average attenuation for each source category under each wind speed condition. 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW 3.4. Modeling the ground attenuation based on the calculation results In order to predict the effects of ground and meteorology by the airport noise prediction model used in Japan, the relationship between ∆𝐿 ,୰୬ୢ,୫ୣ୲ and propagation distance obtained by the calculations has been modeled by curve fitting. We adopted a curve expressed in Equation (8) and parameters a , b , and K were obtained by curve fitting. ∆𝐿 ,୬ୢ,୫ୣ୲ ൌ 𝐾 ሺ1 𝑏∙exp൫െ𝑎 √𝑥 ൯ሻ 𝑐 (8) where x is the horizontal distance from a sound source, a, b and K are regression coefficients, and c is the correction term for the ground condition which 𝐿 ሺ1ሻ in Equation (1) is obtained. Table 4 shows the values of the coefficients a , b , and K . The value of c is set to be 3 dB in the case that 𝐿 ሺ1ሻ is obtained in free field and -3 dB is adopted when 𝐿 ሺ1ሻ is obtained on the hard surface in hemi-free field, respectively. In other cases, the appropriate value for c should be set according to the condition that 𝐿 ሺ1ሻ is obtained. Figure 5 shows ∆𝐿 ,୰୬ୢ,୫ୣ୲ as a function of horizontal distance from sound source in case that the value of c set to be 3 dB. 4. COCLUSIONS In this study, we have modeled the attenuation of aircraft ground activity noise due to the effects of ground and meteorology in order to implement it into the airport noise prediction model used in Japan. We have set a default ground surface condition and the attenuation for each aircraft has been obtained based on the calculation results with the GF-PE method. We have classified the sound sources of ground activities into three categories as “engine run-ups of jet aircraft”, “engine run-ups of propeller aircraft” and “APU”, and the attenuation model has been proposed as a function of horizontal distance from a sound source. Since the model has been proposed based on the numerical simulation results, it is important to examine the prediction accuracy of the proposed model by conducting measurements around airports in future work. 5. REFERENCES 1. Society of Automotive Engineers: AIR5662, Method for predicting lateral attenuation of airplane noise (2006). 2. Report on standard method of computing noise contours around civil airports. 4th ed. European Civil Aviation Conference; 2016. (ECAC-CEAC Doc 29; vol. 2, Technical Guide). 3. Recommended method for computing noise contours around airports. International Civil Aviation Organization; 2018. (Doc 9911). 4. E. M. Salomons, “Computational atmospheric acoustics,” Kluwer Academic Publishers, 2001. 5. D. C. Hothersall & J. B. N. Harriott, Approximate models for sound propagation above multi- impedance plane boundaries, J. Acoust. Soc. Am ., 7 , 918-926 (1995). 6. Y. Miki, “Acoustical properties of porous materials -Modifications of Delany-Bazley models -,” Acoust . Sci . & Tech ., 11, 19–24 (1990). 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW T able 4(1): Coefficients a , b and K for engine run-ups of j e t a i rc raf t. Wind condition -5 m/s -2 m/s No wind +2 m/s + 5 m / s a 0.48 0.23 0.16 0.21 0.25 b 257.72 55.07 73.90 925.54 2935.72 K -22.97 -23.03 -18.33 -9.96 -8.95 Tabl e 4(2): Coefficients a , b and K for engine run-ups of pr opeller air c raft. Wind condition -5 m/s -2 m/s No wind +2 m/s + 5 m / s a 0.30 0.17 0.10 0.06 0.16 b 50.88 26.20 21.38 13.15 156.61 K -23.00 -23.22 -18.19 -11.92 -7.52 Table 4(3): Coefficients a , b and K for APU . Wind condition -5 m/s -2 m/s No wind +2 m/s + 5 m / s a 0.39 0.23 0.17 0.04 0.04 b 223.51 97.28 118.37 14.66 12.19 K -22.99 -23.02 -18.14 -25.00 -17.53 10 10 5 5 0 0 Δ L A,grnd,met [dB] Δ L A,grnd,met [dB] -5 -5 -10 -10 -15 -15 -20 -20 -5 m/s -2 m/s No wind +2 m/s +5 m/s 100 200 300 400 500 1000 2000 3000 -30 -5 m/s -2 m/s No wind +2 m/s +5 m/s 100 200 300 400 500 1000 2000 3000 -25 -25 -30 horizontal range [m] horizontal range [m] (a) Jet aircraft engine run-ups (b) Propeller aircraft engine run-ups 10 5 0 Δ L A,grnd,met [dB] -5 -10 -15 -20 -5 m/s -2 m/s No wind +2 m/s +5 m/s 100 200 300 400 500 1000 2000 3000 -25 -30 horizontal range [m] (c) APU Figure 5: ∆𝐿 ,୰୬ୢ,୫ୣ୲ as a function of horizontal distance from a sound source. 21-24 AUGUST SCOTTISH EVENT CAMPUS ? 02 GLASGOW Previous Paper 425 of 769 Next