Noise of individual vehicles in a high-rise city

Journal of Sound and Vibration (1977) 55(l),

NOISE OF INDIVIDUAL

39-48

VEHICLES IN A HIGH-RISE

CITY

N. W. M. Ko Department of Mechanical Engineering, University of Hong Kong, Hong Kong

(Received 1 February 1977, and in revisedform 21 June 1977)

Extensive roadside noise measurements of 20 000 vehicles in 100 measurement sites in the high-rise city, Hong Kong, are reported. The vehicles are classified into petrol-powered saloon, diesel-powered saloon, mini-bus and small lorry, and bus and big lorry. The survey was mainly concentrated in the urban areas. However, rural areas were also included in the investigation such that comparison with the urban areas could be made. The results obtained illustrate the effect of enclosed environment on the noise emitted by the vehicles and support the simple classification of the sites into closed, semi-closed and open environments. Distinct differences in the sound pressure levels observed in these environments have been found.

1. INTRODUCTION In an urban environment the measured noise level of an individual vehicle would be different from the level obtained by adopting the standard test procedure [I, 21. This difference in the sound pressure level may be due to the local geographical condition, the presence of buildings, the presence of traffic obstacles, the nature of the road surface, the composition of the local traffic, the type of vehicle, the speed of the vehicle, the age of the vehicle, the mechanical condition of the vehicle, the driving behaviour of the motorists and other factors. The main objection to adopting the standard test procedure in an urban environment is the difficulty in having similar test sites which satisfy the specific conditions laid down by the procedure [l, 21. The test procedure requires the absence of any reflecting surface near the site and a comparatively large distance between the microphone and the road. This absence of any reflecting surface means that reverberation is excluded from the consideration. Studies of the noise of individual vehicles have been reported by different workers, notably by Hammarfors and Kajland, Stephenson and Vulkan, Olson, Safeer and Knerr, Lewis, Waters, and Martins Da Silva [3-91. However, the work reported was mainly concerned with noise levels of the individual vehicles in rural or country areas where the specific conditions of absence of buildings and sufficient distance for locating the microphone were satisfied. Furthermore, the few urban measurements of Stephenson and Vulkan [4] and of Lewis [7] also satisfied the conditions of the standard test procedure. This means that the overall effect of enclosed conditions, which are usually found in urban areas, has not really been looked into. The present investigation was aimed at understanding more fully the overall effect of an enclosed environment on the noise generated by individual vehicles in urban areas where buildings and roads are closely interwoven. The present study was carried out in the high-rise city of Hong Kong. Because of the limited space available, very close proximity of tall buildings and busy roads is found [lo]. As has been shown by the findings of aircraft noise in 39

40

N. W. M. KO

the same city [II], this enclosed higher noise level observed.

environment

2. INSTRUMENTS

in a high-rise

AND

city was responsible

for the

METHODS

The basic set of instruments used for the present investigation consisted of a Briiel and Kjaer sound level meter Type 2204 with a 1 inch microphone and a Nagra IV magnetic tape recorder. During the survey the random incidence corrector and wind shield were used such that the directional and wind effects were eliminated. The microphone was about 1.2 m above ground and wherever possible, it was located 1 m away from the kerb. During the present investigation, it has been occasionally found that the microphone has to be located closer than 1 m from the kerb of the road. All the recordings were carried out in dry weather so that noise due to wet roads did not affect the results. Although the recordings were made in calm conditions, the shielding and wind-tunnelling effect gave rise to complications in the estimation of speed and direction of the wind. The output of the sound level meter was recorded on to one channel of the magnetic tape. The other channel was used by the observer to record a commentary on the types of vehicles passing the site. Vehicles which were too close together were neglected and ignored in the analysis, The instruments used for the analysis were a Briiel and Kjaer frequency spectrometer Type 2113 and a Briiel and Kjaer level recorder Type 2305. It was during the analysis that, from the commentary, the peaks on the sound level traces which corresponded to the pass-bys of single vehicles were identified. However, peaks which were too closely spaced were usually neglected. The vehicles were classified into five main groups, which were petrol-powered saloon (private), diesel-powered saloon (taxi), diesel-powered mini-bus and small lorry, bus and big lorry, and motorcycle. The capacity of the engine of the private cars could vary from about 500 C.C. to 6000 C.C. However, the engine capacity of the taxis was mostly from 2000 to 2200 C.C. and of the mini-buses and small lorries about 2200 C.C. The range of the engine capacity of the last two categories was much narrower than that of the private cars. The range of the engine capacity of the bus and big lorry group, as for the private car, could be very large indeed. The lane in which the single vehicle was travelling was also specified in the commentary. The near lane was the one nearest to the microphone. The distance of an individual vehicle in another lane was expressed in terms of the distance between the kerb (on the microphone side of the road) and the centreline of that particular lane. The measurement sites chosen were mainly inside the urban areas, where the traffic was subjected to a 48 km/h (30 mile/h) limit. At each site, the section of the road chosen for recording was level. Although it was initially planned to have roughly the same proportions of sites under the category of closed, semi-closed and open environment, as suggested by Ko [I I], difficulty in finding open sites in the urban areas was soon realized. Thus, the open sites chosen were mainly located in the rural areas. Unfortunately, the vehicles in the open roads were usually travelling at speeds higher than 48 km/h. Similar to the simple classification suggested by Ko [I I], a site which was classified as a closed environment was one which had tall buildings on two, or more than two, sides of the road. An open environment was one without any building on the sides of the road. A semiclosed environment was the one in between. Since the recording of the noise at each site was over a period of one and half hours, some of the sites chosen did not have sufficient samples of individual vehicles in each category. This was particularly true for roads having very high and low traffic volume and density. For the former the vehicles tend to be too closely spaced.

41

INDIVIDUALVEHICLENOISE 3. RESULTS

AND DISCUSSION

3.1. SITES Sound pressure levels of about 20 000 individual investigation. Out of these 20 000 vehicles, 7000 diesel-powered saloons or taxis, 3500 mini-buses lorries. During the survey, motorcycles were also extremely difficult to obtain sufficient samples in are excluded from the following discussion.

vehicles have been obtained in the present were petrol-powered or private cars, 5000 and small lorries, 4500 buses and big included. However, for motorcycles it was nearly all the sites considered. Thus, they

TABLE 1

The distribution of the individual vehicles in the near lane

__Number of sites Average number of individual vehicles at site Standard deviation Maximum number of individual vehicles at all sites

Petrol-powered saloon

Diesel-powered saloon

Mini-bus and small lorry

Bus and big lorry

66 79

44 71

38 68

44 69

52 263

36 166

36 161

35 177

Although about 100 measurement sites have been chosen and the sound pressure levels have been recorded and analyzed, the local traffic composition and traffic density render the results from some of the sites unusable. This is because of insufficient samples of individual vehicles in the different categories from roads having very high and very low traffic density. Furthermore, the dominance of vehicles in one or two categories in the traffic yield insufficient samples in other categories. Because of this difficulty, results having a sample as low as 25 vehicles have to be included in the consideration, so that a reasonable number of sites is available for consideration. Thus, out of the 100 sites considered in the present investigation, only the results from the vehicles in the near lanes, say, of 66 sites for private cars, 44 sites for taxis, 38 sites for mini-buses and small lorries, and 44 sites for buses and large lorries were usable. The average number of individual vehicles measured in these sites, the standard deviation and the maximum are shown in Table 1. It is found that roughly the same value has been found for the average number of vehicles under the four categories. In addition, except for the private vehicles, the standard deviations of the other categories are the same.

3.2. HISTOGRAMS A representive histogram of the measured peak sound pressure levels, in dB(A), of individual vehicles passing along a four-lane road of two-way traffic is shown in Figure 1. The distribution which has the larger sample is not significantly different from a normal distribution. Also from the figure the mean sound pressure levels of the histograms are lower in lanes further away from the microphone. The standard deviations of the measured sound pressure levels under the four categories of vehicles have been obtained. Since the results of the four categories are basically the same, only the ones for the private vehicles are shown in Figure 2. The few available open site results of Stephenson and Vulkan [4] and of Lewis [7] are also included. In the figure there is not any distinct difference in the standard deviation obtained in the closed, semi-closed and open sites. Furthermore, no distinct difference has been found for

42

N.

W. M. KO I

Lone 2

1

Lane 4

1

u I

b ;_

65k--b--

0

0

IO

Number

Figure 1. Distribution

of vehicles

20

IO

of sound pressure level of private cars.

the results from different lanes. Linear regression analysis of the results presented in Figure 2 gives a value of 3.3 dB(A) and a slope of zero. This is slightly higher than the open site results of Stephenson and Vulkan [4], which are 2.6 to 2.9 dB(A). The open site results of the light vehicles of Lewis [7] obtained in the rural and urban dual carriageways are also slightly lower. Linear regression analysis of the results of the taxis, mini-buses, buses and big lorries also yields zero slope. The respective values of the standard deviation are 2.9 dB(A), 3.3 dB(A) and 3.7 dB(A). The value for the mini-buses and small lorries is exactly the same as that for the private cars. The lower standard deviation of the taxis may be due to the adoption of basically two to three models of diesel-powered saloons for taxis, and the compulsory regular inspection of the taxis by the authority. Correspondingly, the higher standard deviation of the buses and big lorries may be due to the large variation in the types and capacities of these vehicles and the large age distribution. Furthermore, the standard of maintenance of these heavy vehicles is usually much lower than that for the vehicles of other categories. By using these histograms of the measured peak sound pressure levels of the vehicles in the near lanes, the mean accumulative histograms of all the usable sites have been found and are shown in Figure 3. The normal distribution curve and the corresponding standard deviations

0

I

I 20

I

I

40

60

I 80

I

I 100

I20 Number

I 140

I

160

I

180

I 200

I

220

I 240

of vehicles

Figure 2. Standard deviation of sound pressure level of private cars. A, Closed; x, semi-closed; 0, Lewis (open, urban); o’, Lewis (open, all sites); ‘0, Stephenson and Vulkan (open).

c:, open.

43

INDIVIDUAL VEHICLE NOISE

-3 0

I IO

I 20

I 30

I 40

I 50

I 60

I 70

1 80

I 90

1 IO0

Percentage of vehdes

Figure 3. Accumulative histograms of sound pressure level (mean). 0, Private; small lorry; D, bus and big lorry. -, Normal distribution.

A,

taxi; 0, mini-bus and

from the means are also shown in Figure 3. The ordinate scale, Y,, is based on the ratio of the difference of the sound pressure level from the mean to the standard deviation. From the distribution curves good agreement of the distributions with the normal curve is found. However, slight deviation from the normal distribution is found for Y0 2 1, which suggests that the measured peak sound pressure levels of less than 20% of the vehicles do not follow the normal distribution. These vehicles have slightly higher peak sound pressure levels than the ones showing a normal distribution. Also from Figure 3 similarity of the accumulative histograms of the vehicles in all the different categories indicates that the measured sound pressure levels of the other 80% of the vehicles, which follow the normal distribution, are found within a range of about 11 dB(A). This bigger contribution to the increase in the measured peak sound pressure level by the noisier 20 % of the vehicles indicates the importance of and the necessity for controlling these vehicles. 3.3. SOUND PRESSURE LEVELS IN NEAR LANE From the accumulative histograms of the measured peak sound pressure levels of the vehicles in the near lane obtained at each site, the sound pressure levels which are exceeded by 10 %, 50 % and 90 % of the sampled vehicles can be determined. In the following paragraphs the sound pressure level (in dB(A)) LNsowill first be discussed, and then L,,, and LN9,,. The effect of the environment of the site will then be isolated and discussed. The above definition is slightly different from the usual definition of Llo,L,, and L,,, based on the observation period, used by other workers. The LNsOvalues for the vehicles in the four categories thus obtained are correlated with the width of the road. However, only the detailed results of the private vehicles are shown in Figure 4. In the correlation the width of the road, instead of the distance between buildings, was adopted so that comparison between the findings in the closed, semi-closed and open sites could be done. Because the speed of the vehicles at the open sites was much higher than 48 km/h, the sound pressure levels observed in these sites were much higher. Thus, they were excluded from the following consideration. The figure indicates distinct differences in the L N50levels obtained in the closed and semi-closed environment. The peak sound pressure levels observed in the closed environment are usually higher than the corresponding ones in

44

N. W. 100

z z

I

70

50

I

I

A

I

&A ‘ rt I $&_x;__,*__ x ,___

’

-4

----7

:: 3

I

SC

z

I

M. KO

b

I

I

I

I

I

A

, _ _’ _z

A

___--

----

ix,

0

I%!

?

P

I

I

I IO

I

I Width of

I

I

I 20

I

I

I

road (ml

Figure 4. Variation of LNso of private cars in near lane. A, Closed; x, semi-closed. Vulkan (open); 0, Lewis; 6, Hammarfors and Kajland; ,o, Olson.

‘0, Stephenson

and

the semi-closed site, by 3 to 5 dB(A). Besides the higher sound pressure level in the closed environment, Figure 4 also shows that the LN,,levels observed in the closed sites are nearly independent of the width of the road. This independence of the width of the road is true over the whole range of roads found in the urban areas, 6 m to 24 m. The open site results of Hammarfors and Kajland [3], Stephenson and Vulkan [4], Olson [5], and Lewis [7] are also shown in Figure 4. Generally, they are lower than the corresponding semi-closed site resuIts of the present investigation. The linear regression analysis of the detailed results of the vehicles in the four categories is summarized in Figure 5. The closed site LNsOvalues are found to be nearly independent of the four vehicle categories. Due to the limited results in the semi-closed sites, it is not certain whether the differences between the sound pressure levels in the closed and semi-closed site would diminish for wider roads. Further comparison between the closed site results of the vehicles in the different categories indicates that the LN,,of the private cars is about 2 dB(A) lower than that of the taxis. The LNS,, of the mini-buses and small lorries, however, is only 1 dB(A) higher than that of the taxis. Similarly, the L,,,of the buses and big lorries is about 2 dB(A) higher than that of the mini-buses and small lorries. The near lane results for the LN,,and L,,,levels of the vehicles in the four categories have also been obtained. The detailed results for the private cars are shown in Figure 6 and the

‘ooT--r

501

I

I

I

I

I

I

10

I

I

I

I

I

20

Width

of road

(ml

Figure 5. Comparison of LNsOinnear lane. Semi-closed: -, private; ---, ---, taxi; -.-, mini-bus and small lorry; -. .-, bus and big lorry.

taxi. Closed: -,

private;

INDIVIDUAL

50

I

I

I

4.5

VEHICLE NOISE

I

I

I

I

I

I

I

1

20

IC Width

of road

im)

Figure 6. Variation of LNIOand LNgOof private cars in near lane. LNIO: A, closed; x, semi-closed. L$,,,: ?? , closed; +, semi-closed.

__>

P_-_-_-_---_-.-._ -----

--------____

vUdth

Figure 7. Comparison

of road (ml

of Lyle and LNgOin near lane. Legends same as Figure 5.

results from the linear regression analysis are shown in Figure 7. As before, distinct differences between the measured sound pressure levels in the closed and semi-closed sites have again been observed. The difference tends to be slightly larger than the one obtained from LNSO. Also similar to the results for the&,,, the&,, and LNgOare nearly independent of the width of the road. In addition, the differences in the LNgo values over the four categories of vehicles are smaller than those for the LN1,. 3.4. SOUND PRESSURE LEVELS IN OTHER LANES It has been shown in the last section that there is a distinct difference in the sound pressure level obtained in the near lane of the closed and semi-closed sites. The present section will present and discuss the results in other lanes of the road. The measured peak sound pressure levels of the vehicles in other lanes, besides the near lane, are presented in the form of ALNbovalues. AL,,, is defined as the difference in the LNSOobtained at the near lane to that of other lanes. The distance used is the distance between the centre-line of that particular lane to the kerb of the road where the microphone was located. The detailed results thus obtained for the private cars and for all the vehicles in the four categories are shown in Figures 8 and 9, respectively. Similar to the results presented in the former sections, there are distinct differences between the results obtained in the closed sites from those for the semi-closed and open sites: the AL,_, for the closed environment has a smaller value than those for the other two types of

N. W. M. KO

46

I

-2

I

I

I IO

I

I

I

I

I

20

Distance

(m)

Figure 8. Variation of A&,, of private cars with distance. -A--, open.

Closed;---x---,

semi-closed;-.-:.\-.-,

environments. This suggests that reverberation of the noise between tall buildings may be the dominant factor responsible for the higher levels obtained in other lanes. For the ALNsOvalues of the semi-closed and open environments, respectively, it seems that there is a difference between them. Generally, the ALNso obtained in a semi-closed site is lower than that in an open site. Again, reverberation may be responsible for this. The summarized results of the linear regression analysis of the vehicles in the different categories, shown in Figure 9, indicates that, besides having lower AL,v50 values, the slopes of the straight lines for the closed site measurements are lower than those for the other two types. For the three types of environment, the largest slope is found for the open sites. The values for the semi-closed sites are somewhere between the other two. Except for the mini-buses and small lorries, the reduction in the sound pressure level in the open sites is less than the inverse square law of propagation: namely, -6 dB per doubling

-2

I

I

I

IO D~stonce

Figure 9. Comparison and big lorry.

of dLNS,,. -,

I

I

I

Private; ---,

I

I

I

I

J

20 (m)

taxi; -.-.-,

mini-bus and small lorry; -. .-. .-, bus

INDIVIDUAL

VEHICLE NOISE

47

distance. This smaller reduction shown in Figure 8 may partly be due to the reflection of the sound by the ground surface. It also may be due partly to the fact that vehicles in the far lanes tend to travel faster. This higher velocity would result in the higher measured sound pressure levels [5, 71, and correspondingly lower reduction. The higher reduction of the mini-buses and small lorries is mainly due to the inaccuracy from having insufficient measurement sites. The insufficient number of measurement sites for the mini-buses and small lorries has also been observed for the semi-closed environment. The reason for the difficulty in finding enough sites was the concentration of these mini-buses on a few busier roads. This means that the available suitable sites for measurement were limited. Further from Figure 9, it is interesting to find that the closed site curves of all the four categories of vehicles fall very nearly on to one another. This interesting phenomenon suggests that in the closed environment the reduction in the measured peak sound pressure levels, which are exceeded by 50 % of the vehicles, is independent of the category of the vehicle. When compared with the near lane, a reduction of about 4 dB(A) at a distance of 20 m is observed. The curves of the semi-closed sites. however, do not collapse on to the same curve, as is the case for the closed sites. Even though they do not collapse on to one curve, they mainly lie within a band of about 1 to 1.5 dB(A). As has been pointed out, the results of the mini-buses and small lorries are not very reliable. The curves of the open sites have wider variations than those of the open sites (see Figure 9). The results for the private cars and taxis agree very well with each other. The results fat buses and big lorries tend to be 1.5 to 2 dB(A) higher.

4. COMMENTS From the present investigation into the measured sound pressure levels of 20 000 individual vehicles in a high-rise city, better understanding of the overall effect of an enclosed environment on the individual vehicle’s noise has been obtained. Even though the urban environment is very complicated in nature and the number of factors which may affect the noise is large, in the present work a fairly simple approach to the understanding of these effects has been found. The simple approach involves the classification of the measurement sites into closed, semi-closed and open environments. Distinct differences have been found between the results for the closed, as opposed to the semi-closed and open, sites. The sound pressure levels obtained in the closed sites are generally higher than the corresponding ones in the semi-closed and open sites. This is true not only for the results obtained from vehicles in the near lane but also from vehicles in other lanes of the road. The simple approach includes implicitly the individual and combined effects of the numerous parameters or factors. Even with the number of vehicles and sites of the present investigation, isolation of the individual effect of each parameter (such as the height of buildings, distance between buildings, the shape and exterior decoration of buildings, and the difference in speed of vehicles in different lanes) can be extremely difficult. One of the obvious difficulties is the unique geographical conditions at each site. The other difficulties are the local variation of the vehicle composition (in respect to engine capacity, type and mechanical condition) in each category at different lanes and sites, and the local variation of the driving behaviour of motorists in different lanes, sites and areas. In this respect, even the bestcontrolled experimental isolation of individual vehicle results would still lead to uncertainty in the findings. Thus, it is through this simple approach to the complicated problem that better understanding of the overall effect of enclosed environment on the vehicle noise can be found.

N. W. M. KO

48

With this understanding the permissible sound pressure levels for different types of vehicles in urban areas can be realistically set and the difficulty in enforcement can be easily overcome.

ACKNOWLEDGMENTS This work was part of the author’s survey of noise pollution in the high-rise city of Hong Kong. Help has been given by past and present students of the Department of Mechanical Engineering and by people outside the University. Without their help the survey would never have been started, not to mention completed. During the experimental work, help from Messrs W. M. Tsang, A. S. H. Kwan, W. T. Chan and S. Y. Chan was greatly appreciated.

REFERENCES 1. British Standards Institution, noise emitted by vehicles.

London 1966 BS 3425: 1966. Method for the measurement

of the

2. International Organization for Standardization,

3. 4. 5. 6. 7.

Geneva 1964 IS0 Recommendation R362. Measurement of noise emitted by vehicles. P. HAMMARFORS and A. KAJLAND1963 Acustica 13, 258-269. Sound-pressure analyses of noise from motor vehicles. R. J. STEPHENSON and G. H. VULKAN1968 Journal of Sound and Vibration 7, 247-262. Traffic noise. N. OLSON1970 National Research Council of Canada, N.R.C. 11270. Statistical study of traffic noise. N. B. SAFEERand B. F. KNERR 1971 Department of Transportation, Washington, Report No. OST-ONA-71-8. Community sound levels. A comparison of measured and estimated data. P. T. LEWIS 1973 Journal of Sound and Vibration 30, 191-206. The noise generated by single

vehicles in freely flowing traffic. 8. P. E. WATERS1974 Journal of Sound and Vibration 34, 560-561. Vehicle noise levels in freely

flowing traffic. 9. P. MARTINSDA SILVA 1975 MOP Laboratdrio National de Engenharia Civil, Lisboa, Znformacrlo tecnica edificios 7. Ruido de trafego rodoviario. 10. N. W. M. Koand K. C. Lo 1972 Hong Kong University Engineering Journal 34,5-15. Preliminary investigation of traffic noise in Hong Kong. 11. N. W. M. Ko 1975 Journal of Sound and Vibration 38, 512-516. Aircraft noise in a high-rise city.