Validation of calculation method of road traffic noise behind building complex in ASJ RTN-Model 2018 by field measurements Shinichi Sakamoto 1 Institute of Industrial Science, The University of Tokyo 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan Wenrui Xu 2 Graduate school of Engineering, The University of Tokyo 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8654, Japan Taiki Fukuda 3 West Nippon Expressway Company Limited 1-6-20 Dojima, Kita-ku, Osaka 530-0003, Japan Miki Yonemura 4 Institute of Industrial Science, The University of Tokyo 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan ABSTRACT Road traffic noise calculation method behind building complex included in the ASJ RTN-Model 2018 is an empirical model formula based on experimental results of a scale model. Therefore, the scope of the calculation method is formally limited in terms of building height, density, and distance from the road, dependent on the range of the experimental conditions. However, the roadside conditions in Japanese urban and suburban areas are diverse, and it is necessary to quantitatively investigate the validity of the calculation in such various conditions in order to widely expand the applicability of the calculation method. In this study, we examine the validity of the road traffic noise calculation method in building complex by comparing the actual measurement results in some build-up areas. 1. INTRODUCTION

The authors have been studying an efficient estimation method of road traffic conditions in a view- point of noise emission from the roads[1]. In order to create accurate noise maps, continuing from the study, it is necessary to investigate an accurate estimation method of noise propagation in various build-up areas. In Japan, noise mapping for entire city is not obliged, but rates for meeting environ- mental quality standard are collected nationwide. For road traffic noise, which is the most familiar

1 sakamo@iis.u-tokyo.ac.jp 2 xu-wr7@iis.u-tokyo.ac.jp 3 tikfkd@gmail.com 4 m-yone@iis.u-tokyo.ac.jp

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

noise source to citizens, the ASJ RTN-Model[2] has been periodically updated and has been referred in environmental noise assessment and monitoring as representative noise prediction method. The ASJ RTN-Model includes calculation method of noise in building complex, which was proposed by Fujimoto et. al [3.4]. (they have named the prediction method of insertion loss of detached houses “F2012, and its simplified version, F2012*”). The calculation methods, F2012 and F2012*, are basi- cally an empirical method based on scale model experiments, and therefore it is necessary to further investigate the applicability of the method to various siting conditions of detached houses in order to utilize the calculation method to create noise maps in Japan. In the presenting paper, the authors applied the calculation method to two areas which have different features regarding the siting condi- tions and examined the applicability of the calculation method by comparing the calculation results with measurement results of the road traffic noise in the areas. 2. CALCULATION METHOD

According to the ASJ RTN-Model 2018[2], A-weighted sound pressure level at a prediction point P generated by a point source, i , (discrete sound source located on a driving lane), L A, i , is calculated as,

A, A, 10 B, 8 20log i W i i L L r L     (1)

where, L W A, i is A-weighted sound power level of the source i , r is a distance from the source to the prediction point P,  L B, i is a correction value due to the buildings along the propagation path from the point source i to the prediction point P. The ASJ RTN-Model 2018[2] gives the correction value  L B, i as follows.

    

i i L p L q

B, BB,

 

   

(2)

p H h

0.017 8.8 1

p

 

  

q H h

0.063 8.8

p

where, H is the height of the buildings [m] and h p is the height of the prediction point [m].  L BB, i is the correction value [dB] when H and h p are 10 m and 1.2 m, respectively, and is given by

            (3)

  

SP, 0.0904 BB, 0 1 2 10lg 10 i i d i i i L b b b

where, b 0 = 0.046, b 1 = 1.01, b 2 = 0.554 and these values were determined based on scale model experiments by Fujimoto et. al.[2,3].  i ,  i ,  i and d SP, i are defined as shown in Figs. 1 (a) and 1 (b).

7.5 m 7.5 m

5 m 5 m



Area

P 

P 

Area +

(a) Perspective angle (b) Building density Figure 1: Definition of the parameters in Eq. (3).

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

Here, all of the parameter values of  i /  i ,  i and d SP, i which are needed to obtain the insertion loss due to the buildings can be explicitly obtained by the plane coordinates of the buildings, the sound source i and the prediction point P, and the computational load is comparably light. On the other hand, basic equation regarding sound propagation includes correction term(s) related to various attenuation factors in sound propagation as follows.

A, A, 10 cor, 8 20log i W i i L L r L     . (4)

Δ L cor, i includes correction terms for sound diffraction Δ L dif, i , ground effect Δ L grnd, i and air ab- sorption Δ L air, i as follows.

cor, dif , grnd, air, i i i i L L L L        (5)

In this study, correction terms by air absorption and ground effect were also taken into con- sideration other than Δ L B, i in E.(1), because the distances from source points to prediction points tended to be very large in order to create noise map regarding road traffic noise. 3. SITES UNDER INVESTIGATION

Two sites which are located in and around Tokyo were chosen to validate the calculation method. In the two sites, houses and buildings are spreading along arterial roads which have heavy traffic vol- ume.

3.1 Site A

Figure 2 (A) and 2 (B) show a situation of buildings and measurement points of the Site A, which is located along an arterial road in Tokyo. Along the road (Tokyo Metropolitan Road No. 318, Kannana- Dori Route 7), apartment houses, office buildings and educational facilities line up as buffer buildings with a height of three stories or more and many detached houses are spreading behind the buffer buildings. Figure 3 (A) and 3 (B) shows statistical data on the area of the buildings. This area is a high-density residential area, and the histogram of the statistical data of the area of individual build- ings has a peak at around 40 to 60 m 2 , and building coverage ratio to the total area is comparably high as 0.41.

3.2 Site B

Figure 4 (a) and 4 (B)shows a situation of buildings and measurement points of the Site B, which is located along an arterial road near Tokyo. Along the road (National highway No. 122) , farm fields are spreading and detached houses are scattered. Figure 5 (A) and 5 (B) shows statistical data on the area of the buildings. The histogram of the statistical data of the area of individual buildings has two peaks at less than 20 m 2 and 80 to 100 m 2 . Building with area of 20 m 2 or less, number of which is largest in the histogram, seem to be a shed or a garage. From the histogram, many of the detached houses have their area of 60 m 2 or more, and the area is larger than that of the Site A. The total building coverage ratio is low as 0.09.

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

(A)Source road and measurement points (B)Building situation around the road Figure 2: Summary of the site A

0.2

Total Area ratio

0.15

0.1

0.05

0

0-25 25-100 100-200 200-400 400-

(A)Distribution of respective building area (B)Total area of buildings Figure 3: Statistics of area of buildings

Area/building [m 2 ]

Ratio Bdgs : 0.41 1200 gggeg° 88328 ‘s6pig Jo “WAN,

(A)Source road and measurement points (B)Building situation around the road Figure 4: Summary of the site B

0.2

Total Area ratio

0.15

0.1

0.05

0

0-25 25-100 100-200 200-400 400-

(A)Distribution of respective building area (B)Total area of buildings Figure 5: Statistics of area of buildings

Area/building [m 2 ]

Ratio Bdgs

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

4. MEASUREMENT AND ANALYSIS OF ROAD TRAFFIC NOISE

Two kinds of measurement points, the first ones were reference points (M00 in each area) located at roadside and the second ones were propagation measurement points (M01, M02, etc.) behind build- ings, were set. In measurements, sound pressure at the reference point, M00, were continuously rec- orded, and simultaneously, recording sound pressure at the propagation measurement points were repeated by moving the microphone. Temporal duration of recording the road traffic noise at each propagation measurement points were 5 minutes or more. In the measurement, class 1 sound level meters (RION NL-52) was used, the microphones were set at the height of 1.2 m using tripods and the sound was recorded by using the recording function of NL-52. Recorded signals were passed through the A-weighting filter and then they were squared. The squared signals were time averaged by the time weighting F (125 ms time constant) to obtain wave forms of A-weighted sound pressure levels (SPLs). Obtained A-weighted SPL variation characteris- tics for the Sites A and B are shown in Figs. 6 (A) and 6 (B), respectively. In each graph, A-weighted SPL variation in time at the reference point (M00) and the sound propagation points (M01, M02 and etc., recorded serially) were synchronously overlapped. In order to obtain attenuated road traffic noise level.  L A , due to buildings, recorded signals on each sound propagation point (M01, M02, and etc.) and corresponding part of recorded signals on M00 were analyzed to obtain their time averaged A- weighted sound pressure levels. Gray parts in the graphs include obstructive noise other than road traffic noise from the target road and the noises in the parts were excluded from the analyses.

100

A-weighted SPL [dB]

M01 M10 M12 M13 M14 M11 M00 M00 M00 M00 M00 M00

90

80

70

60

50

40

30 0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440 1500 1560 1620 1680 1740 1800 Time [s]

(A) Site A

100

A-weighted SPL [dB]

M00 M00 M00 M00 M00

90 80

M01 M05 M04 M03 M02

70

60 50

40

30 0 60 120 180 240 300 360 420 480 540 600 660 720 780 840 900 960 1020 1080 1140 1200 1260 1320 1380 1440 1500 1560 1620 1680 1740 1800 Time [s]

(B) Site B Figure 6: Time variation of measured A-weighted sound pressure levels and analysis periods 5. COMPARISON BETWEEN CALCULATION AND MEASUREMENT

Comparisons between calculation and measurement for the Sites A and B are shown in Fig. 7 (A) and 7 (B).

5.1 Site A

The Site A is a densely built residential area which spreads along an arterial road. Its ground surface is reflective except for some ground parks. Therefore, a correction term for the ground effect, Δ L grnd, i , in Eq. (5) was set to zero. From the figure, the calculation results are in good agreement with the measurement results.

5.2 Site B

In the Site B, most of ground surface is farmland, and residential buildings are scattering in the field as shown in. Fig. 4 (A). On a part of the ground along the arterial road, there exists a reflective concrete asphalt area (hatching part in Fig. 4 (B)). Therefore, influence of the ground effect on the calculation results behind residential buildings was examined by comparing the calculation result on several ground effect conditions assumed in the calculation with the measurement results. As the grounds, four cases of a rigid surface (concrete or asphalt pavement), a loose soil, a grassland and a compact ground, which were treated in the ASJ RTN-Model 2018, were adopted, and their calculation results were investigated. In the ASJ RTN-Model 2018, effective flow resistivities for the loose soil, the grassland and the compact ground are assumed as 75 kPa s/m 2 , 300 kPa s/m 2 and 1,250 kPa s/m 2 , respectively. On the calculation of the ground effect prescribed in the ASJ RTN-Model 2018, aver- aged propagation height over the ground was fixed as 0.6 m (source and receiver heights were 0 m and 1.2 m, respectively) in this study. Actually, the sound propagation height strongly affects the calculation results. Therefore, the averaged propagation height in a case with buildings should be investigated as another issue. From the comparison result shown in Fig. 7 (B), following findings were obtained. (1) Calculation as sound propagation over reflective ground gave around 6 dB greater result than the measured ones ( y = x + 6.58). It is necessary to take the ground effect into consideration. (2) When the ground effect was included in the calculation, the case of compact ground best fitted to the measurement result ( y = x + 0.1). The cases of grassland or loose soil gave around 4 dB smaller than the measurement results.

10

10

y = x +0.43 SE=1.95

y = x +6.58,SE=1.51 y = x - 4.48,SE=2.82 y = x - 3.52,SE=2.37 y = x +0.10,SE=1.55

Rigid Loose soil Grassland Compacted ground

5

5

0

0

Calculated  L A [dB]

Calculated  L A [dB]

-5

-5

-10

-10

-15

-15

-20

-20

-25

-25

-30

-30

-35

-35

-40

-40

-40 -35 -30 -25 -20 -15 -10 -5 0 5 10 Measured  L A [dB]

-40 -35 -30 -25 -20 -15 -10 -5 0 5 10

Measured  L A [dB]

(A) Site A (B) Site B Figure 7: Comparisons of ∆ L A between measurement and calculation

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

5. CONCLUSIONS

In this paper, the road traffic noise calculation method in build-up areas included in the ASJ RTN- Model 2018 was validated by comparing the calculation results with field measurement acquired in two measurement sites which had different features in siting conditions. From the investigations, the following findings have been obtained. (1) To dense build-up area of residential houses on rigid ground, the calculation method can be ac- curately applied. (2) In areas in which buildings exist on soft ground, consideration of ground effects is needed to ensure a certain level of calculation accuracy. From our experimental examination, a ground con- dition of a compact ground described in the ASJ RTN-model 2018 fitted to measurement results the best. 6. REFERENCES

1. Shinichi Sakamoto, Taiki Fukuda, Miki Yonemura, Hyojin Lee, Road traffic noise mapping based on aerial photographs - sound power level determination of road vehicles. Proceeding of INTER- NOISE 2021, pp. 5523–5527. Washington D.C., USA, August 2021. 2. Shinichi Sakamoto, Road traffic noise prediction model “ASJ RTN-Model 2018”: Report of the Research Committee on Road Traffic Noise, Acoust. Sci. & Tech. , 41 (3), 529–589 (2020). 3. Kazutoshi Fujimoto, Kyosuke Tsuji, Toru Tominaga and Kengo Morita, Prediction of insertion loss of detached houses against road traffic noise using a point sound source model. Acoust. Sci. & Tech. , 36 , 109–119 (2015). 4. Kazutoshi Fujimoto and Ken Anai, Prediction of insertion loss of detached houses against road traffic noise using a point sound source model: Simplification of prediction formula F2012. Acoust. Sci. & Tech. , 38 , 287–294 (2017).

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW