Influence of traffic noise emission spectra on the design of barriers

Transcription

1 Influence of traffic noise emission spectra on the design of barriers Vitor C. T. Rosão a) and Maria F. F. Neto b) (Received 2006 March 13; revised 2007 February 15; accepted 2007 March 03) Acoustical barriers are widely used and the methods to estimate the noise attenuation they produce are well established. Although data for noise spectra for different conditions of road traffic are available, the interim method recommended in the European Union considers only one kind of spectrum. This has important implications in the dimensions of the barriers used for noise attenuation, generally leading to over sizing with the corresponding economical costs. The intent of this paper is to show the practical importance of the spectra used when determining dimensions of acoustical barriers and the level of uncertainty that may occur when only one kind of spectrum is considered Institute of Noise Control Engineering. Primary subject classification: 31.1; Secondary subject classification: INTRODUCTION An assessment of the environmental impact caused by sources of noise should be performed before the implementation of any construction or activity that may cause any form of damage to the environment. Concern with the acoustic comfort and the quality of life is common to any citizen, but the State must be active and promote, guide and inform the citizens about the necessity to control the noise generation with the subsequent goals of reducing noise pollution which require accurate assessments of activities that may produce noise and acoustic discomfort. In relation to the noise in cities, road traffic is one of the most significant sources of acoustic discomfort, when compared to other sources such as industry, airports and even that produced by people in their daily activities. 1 In the study presented by Ali and Tamura 2 the authors report a close relation between traffic noise and the sensation of discomfort and irritability in the population. Road traffic is of paramount economical importance but measures to reduce the noise produced by road traffic are urgently needed. The automobile is a complex source of noise. The main sources are the engine, exhaust, gears, tires and car structure. 3 Reduction of the noise produced by road traffic can be achieved, amongst other options, by introducing mechanical improvements or by improving the quality of the pavement. However, the most usual a) b) SCHIU Protugal, Engenharia de Vibração e Ruído, Lda., Rua de Faro, Bloco B, 2 F, Estoi, Faro, PORTU- GAL; vitorrosao@schiu.com SCHIU Brazil, Engenharia de Vibração e Ruído, Lda., Av. Mascote, 621, ap 11, , São Paulo BRAZIL; fatimaneto@schiu.com process to reduce the impact of traffic noise is to introduce acoustical barriers. Acoustical barriers have been extensively studied. Neto 4 has presented a report on the performance of acoustical barriers composed of different materials, regarding efficiency and sound quality. Several other studies have addressed this subject, 5,6 while others focused on the barrier geometry. 7 There are several software programs developed to predict road traffic noise and the effectiveness of acoustical barriers. 8 These software programs are often based in national models, for example: 9,10 Austria, RVS 3.02 Lärmschutz (RVS 3.02), 1997 (only overall levels); France, NMPB/XP S31-133, 1996/ 2001 (1/1-octave Hz; only one kind of spectrum); Germany, Richtlinien für Lärmschutz an Straßen (RLS-90), 1990 (only overall levels); Joint Nordic, New Nordic Prediction Method (Nord 2000), 2000 (1/3-octave Hz; different spectra); U. Kingdom Calculation of Road Traffic Noise (CRTN), 1988 (only overall levels); USA, Traffic Noise Model (TNM), 1998 (1/ 3-octave Hz; different spectra). Directive 2002/49/EC, of the European Parliament and the Council, relating to the assessment and management of environmental noise, 11 was published in 2002, and recommends the use of French method NMPB/XP S31-133, 12 for road traffic noise modelling, until there is agreement on a common European method. 13,14 Although the French method considers only one kind of spectrum, there are six countries that have adopted it as their national method: 15 Belgium, 322 Noise Control Eng. J. 55 (3), 2007 May-June

2 Fig. 1 Schematic illustration showing the difference of path of the sound from the source (left) to the receptor (right) without a and with a sound barrier b+c. France, Greece, Italy, Portugal and Spain. The objective in this paper is to show the importance of knowing the effect of the spectra produced by the road traffic on the effectiveness of acoustical barriers, since the changes in spectrum will determine the dimensions of the acoustical barriers (height and length) that result in similar noise attenuation. 2 NOISE ATTENUATION PRODUCED BY AN ACOUSTICAL BARRIER Any common noise source can be decomposed into a set of more or less complex single point sources, and it is relevant to ascertain the reduction produced by an acoustical barrier relatively to these point sources. According to ISO for a given acoustical barrier fixed in the ground, three different diffracted paths can be considered: one at the top of the barrier, another at its left side and one at its right side. For each diffracted path, the attenuation of acoustical barrier, A, is given by: 2 A = 10 log f/c if 40f/c 2 A = 0 if 40f/c 1 where f is the frequency in study, c is the velocity of sound, and is the difference in length between the diffracted path under study and the direct path in the absence of the barrier. Figure 1 shows the difference in paths,, from the source to the receiver, with and without a barrier. Considering the center frequencies of the octave bands between 63 Hz and 4000 Hz, Fig. 2 presents the attenuation of an acoustical barrier for these frequency bands for values of between 0 and 4 metres. The analysis of Fig. 2 shows that for a reduction of 15 db for the 63 Hz octave band, it is necessary to create a value of of approximately 4 metres, for 125 Hz a value of of approximately 2.1 metres is necessary, for 250 Hz a value of of approximately 1 metre is necessary, for 500 Hz a value of of approximately 50 centimetres is necessay, for 1000 Hz Fig. 2 Attenuation of an acoustical barrier in function of the sound frequency and the difference of path. a value of of approximately 25 centimetres is necessary, for 2000 Hz a value of of approximately 12.5 centimetres is necessary and for 4000 Hz a value of of approximately 6.75 centimetres is necessary. These results are the consequence of the known principle that acoustical barriers are more efficient in reducing high frequency noise. 3 INFLUENCE OF THE SPECTRUM The single normalized spectrum established for the road traffic by the French standard XP S [values in octave band normalized in relation to the global broad band value (A-weighted)] is shown in Fig. 3. Figure 4 presents the spectra obtained in situ for normal road traffic on a smooth and a rough pavement. Using actual data from several road traffic noise databases from Harmonoise, 13 Imagine, 14 SILVIA 17 and the Technical Manual of USA Traffic Noise Model (TMN), 18 more realistic figures can be built. Figure 5 presents the spectra for automobiles on a typical pavement surface at 10 mph 16 km/h and at 80 mph 129 km/h considered by USA TNM. 19 Figure 6 presents the normalized spectra for cars and trucks, considered by the European working group SILVIA for the reference pavement. 20 The overall profiles are similar to the normalized Fig. 3 Normalized spectrum of road traffic noise. Noise Control Eng. J. 55 (3), 2007 May-June 323

3 Table 1 Attenuation in db of acoustical barriers in function of the spectrum and the difference of path (point source). Fig. 4 Spectra of two different pavement types. The noise produced over a smooth pavement has a higher component of lowfrequencies when compared with that produced over a rough pavement, which produces a higher amount of high frequencies. standard spectrum shown in Fig. 3, but important differences can be pointed out. The data in Table 1 represents the attenuations resulting from the use of the seven types of spectra previously described and different values of. It is observed that for the same overall noise emission, maximal variations in attenuation of up to 3 db can be Fig. 5 Spectra for automobiles for average pavement at 10 mph 16 km/h and at 80 mph 129 km/h (USA TNM). Fig. 6 Spectra for cars and trucks over reference pavement (EU SILVIA). Path difference [m] Spectra Normalized Smooth Pavement Only Trucks mph mph Rough Pavement Only Cars obtained by using different model spectra. In this case, the effective noise attenuation is less when the normalized spectrum is considered. Equation (1) is applicable to point sources, so the validity of its application on the case of linear sources is questionable. Thus, we modelled a road as a linear source at 0.5 metres above the ground, in software that includes the ISO method, and noise levels for receivers at different distances from the linear source and at different heights above the ground. The noise levels were calculated for the seven types of spectra described above, with and without a long acoustical barrier with a height of 4 metres, located at 2 metres from source. The receiver s positions were selected to obtain the same values of used in Table 1. The different attenuations given by the acoustical barrier is presented on Table 2. As shown in these tables, the attenuation produced by the barrier is lower (by 2 to 6 db) for linear sources than for point sources, however, for the same overall noise emission, identical variations in attenuation occur. Maximal variations in attenuation of up to 2 db can be obtained. Table 2 Attenuation in db of acoustical barriers in function of the spectrum and the difference of path (linear source). Path difference [m] Spectra Normalized Smooth Pavement Only Trucks mph mph Rough Pavement Only Cars Noise Control Eng. J. 55 (3), 2007 May-June

4 Table 3 Length (in metres) of the necessary and adequate acoustical barriers. Attenuation of acoustical barrier [db] Distance of the receptor to the road [m] DIMENSIONS AND COSTS OF THE ACOUSTICAL BARRIERS As a first approach, we may consider that a given acoustical barrier fixed in the ground is sufficiently long (the resulting diffractions from the ends are negligible in relation to the diffractions from the top of the barrier), if the angle ( in degrees) for different paths via the ends of the barrier, in relation to a given receptor, satisfies the following equation: A/ A/10 2 In Eqn. (2), A is the broad band attenuation of the acoustical barrier, and should be higher than 0 db. In this case, considering a straight road of infinite length, an acoustical barrier parallel to the road, and a receptor located at a distance d, along a perpendicular line to the road that crosses the barrier in its middle point, the barrier is sufficiently long if: 22 l 2 d Tan 2 3 Table 3 shows the indicative values of l in relation to possible A and d. In agreement with the reasoning in the previous section, we verify that variations in noise levels of 1 db can occur due to slight variations of the spectrum. From the data in Table 3, to obtain a change of 1 db in the attenuation provided by the barrier (due to a slight variation in the spectrum) it would be necessary to significantly change the length of the barrier. In the case of a change from an effective attenuation of 14 or 12 db to 13 db, to compensate for the variation in predicted values between the model and the actual spectrum, one would need to either shorten the barrier by approximately 650 m or enlarged by almost 525 m respectively, for a receptor located at a distance of 100 m along the perpendicular to the middle point of the barrier. The validity of Eqn. (3) for the determination of minimum effective barrier lengths (when diffractions from the ends are negligible in relation to diffraction over the top) was tested using the software and model described previously. If we choose, for example, the spectrum for automobiles at 80 mph, and decrease the related length of acoustical barrier to reduce attenuation by 1 db, i.e., to give the same attenuation as the acoustical barrier related to only trucks, we obtain a new length of 1200 metres. This new length is 1000 metres shorter then the related length of only trucks, which confirms the conclusion obtained by using Eqn. (3). In Portugal, the most common height of the acoustical barriers is 4 metres, and assuming an overall cost (material and labor) of 150 /m 2, a modification in length of 1000 metres will alter the costs by , a significant value, especially if one considers that we are altering the attenuation of the barrier by just 1 db due to a slight change in the spectrum. 5 CONCLUSIONS Considering the data and reasoning described, it is very important that the available software and software to be developed take into account the spectral characteristics of the noise sources, which does not happen currently in road traffic noise modelling when the NMPB/XP S method is used. Nonetheless it would be of great interest that the modelling software could include, in addition to several default spectra, the possibility for the user to define one or more custom spectra especially when the characteristics of the road traffic are known and the spectrum can be obtained in situ. Such possibility would produce improved predictions and design of barriers for specific cases, and ultimately result in economically and environmental benefits. According to this study, reductions in costs of , and differences in attenuation of up to 2 db could be achieved. 6 REFERENCES 1. Ö. Gündoğdu, M. Gökdağ and F. Yüksel, A Traffic Noise Pre- Noise Control Eng. J. 55 (3), 2007 May-June 325

5 diction Method Based on Vehicle Composition Using Genetic Algorithm, Appl. Acoust., 66(7), , (2005). 2. S. A. Ali and A. Tamura, Road Traffic Noise Levels, Restrictions and Annoyance in Greater Cairo, Egypt, Appl. Acoust., 64, , (2003). 3. H. G. Jonasson and S. Å. Storeheier, Nord New Nordic Prediction Method for Road Traffic Noise, SP Swedish National Testing and Research Institute, SP Report:10, Borås, (2001). 4. M. F. F. Neto, Estudo de Barreiras Acústicas ao ar Livre sob a Perspectiva da Eficiência e Qualidade Sonora, Master Dissertation, Campinas, Brasil, D. Aylor, Perception of Noise Transmitted Through Barriers, J. Quant. Spectrosc. Radiat. Transf., 59(2), , (1976). 6. G. Watts., L. Chinn and N. Godfrey, The Effects of Vegetation on the Perception of Traffic Noise, Appl. Acoust. 56, 39 56, (1999). 7. T. Ishizuka and K. Fujiwara, Performance of Noise Barriers With Various Edge Shapes and Acoustical Conditions, Appl. Acoust., 65, , (2004). 8. C. Stelle, A Critical Review of Some Traffic Noise Prediction Models, Appl. Acoust. 62, , (2001). 9. State of the Art, Improved Methods for the Assessment of the Generic Impact of Noise in the Environment (IMAGINE), (2004). 10. Source modelling of road vehicles, HARMONOISE, (2003). 11. Assessment and management of environmental noise, EURO- PEAN DIRECTIVE 2002/49/EC. 12. Acoustique-Bruit des infrastructures de transports terrestres- Calcul de l atténuation du son lors de sa propagation en milieu extérieur, incluant les effets météorologiques, French Standard XP S31-133, (2001) Acoustics-Attenuation of sound during propagation outdoors- Part 2: General method of calculation, ISO :1996, (1996) Traffic Noise Model, Technical Manual, FHWA, (1998). 20. Guidance Manual for the Implementation of Low-Noise Road Surfaces, Sustainable road surfaces for traffic noise control (SILVIA), (2006). 21. Ministère de l Environnement et du Cadre de Vie; Ministère des Transports; CETUR, Guide du Bruit des Transports Terrestres, Prévision des Niveaux Sonores. [s.l.]: ed. A., (1980). 22. Vitor C. T. Rosão, Desenvolvimento de Modelo de Avaliação do Impacte Ambiental Devido ao Ruído de Tráfego Rodoviário, Master Dissertation, Lisboa, Portugal, (2002). 326 Noise Control Eng. J. 55 (3), 2007 May-June