REPORT. Revision of Nordtest Methods NT ACOU 039 and ACOU 056 for Measuring Noise from Road Traffic Client: Nordtest. Revised 15 March 2001

Transcription

1 Page 1 of 25 REPORT DELTA Danish Electronics, Light & Acoustics Building 356 Akademivej DK-2800 Kgs. Lyngby Denmark Revision of Nordtest Methods NT ACOU 039 and ACOU 056 for Measuring Noise from Road Traffic Client: Nordtest Revised 15 March 2001 Tel. (+45) Fax (+45) The report must not be reproduced, except in full, without the written approval of DELTA. 80rap-uk-a

2 Page 2 of 25 Title Revision of Nordtest Methods NT ACOU 039 and ACOU 056 for Measuring Noise from Road Traffic Journal no. Project no. Our ref. AV 1570/00 P 8803 JK/lm Client Nordtest PB 116 FIN Esbo Finland Client ref. Project No , Laila Törnroos Summary Since the present Nordtest methods for measuring road traffic noise were issued in the 1980s, significantly improved knowledge has become available, and a need for up-dating was identified by the Nordtest expert group on acoustics. The present report contains a proposal for revised Nordtest methods. The main changes are the introduction of a new meteo-window, a completely new section on measurement uncertainty and definition of maximum noise levels from single noise events. The text has been restructured and thoroughly elaborated, and an informative annex has been added on the influence of road surfaces on traffic noise levels. The Nordtest project group recommends for road and environmental authorities to require that measurement results used as a basis for making decisions on noise reduction or compensation shall be provided by laboratories accredited or persons certified for road traffic noise measurement according to the Nordtest methods. DELTA, 15 March 2001 Jørgen Kragh Acoustics & Vibration

3 Page 3 of 25 Contents 1. Background and Purpose Delimitation Method of Work Development and Updates since Measurement Instruments New Standards Noise Barriers Facade Sound Insulation Road Surface Influence on Traffic Noise Vehicle Noise New Nordic Prediction Method New Noise Indicators Weather Influence on Sound Propagation SP Workshop DELTA Project DELTA Project Present Nordtest Method Uncertainty General L Aeq L AFmax Contributions to σ Road Surface Influence Maximum Sound Pressure Level Definition and Determination Correction for Background Noise and Reverberation Time Background Noise Reverberation Time Requirements on Reporting Needs Identified for Future Development References... 25

4 Page 4 of 25 Appendix 1 Proposal for Nordtest method - Acoustics: Road Traffic Measurement of Noise Immission. Engineering Method. Appendix 2 Proposal for Nordtest method - Acoustics: Road Traffic. Measurement of Noise Immission. Survey Method.

5 Page 5 of Background and Purpose Nordtest approved in March 1982 the method NT ACOU 039 Road Traffic: Noise [1], and three years later the simplified method NT ACOU 056 [2]. Since then the knowledge on road traffic noise emission and on sound propagation has increased significantly, and measurement instrumentation has changed. Road traffic noise is regularly measured in order to quantify the acoustical environment adjacent to trafficked roads, even though road traffic noise levels nowadays are more often characterised by results of calculation made according to the Nordic prediction method for road traffic noise. Thus, a need for up-dating Nordtest method NT ACOU 039 as well as the simplified version NT ACOU 056 was identified, and a project group was formed, consisting of Jørgen Kragh (project leader), DELTA, Acoustics & Vibration, Denmark Juhani Parmanen, VTT Building Technology, Acoustics, Finland Steindor Gudmundsson, Icelandic Building Research Institute, Iceland Svein Å. Storeheier, SINTEF, Telecom & Informatics, Norway Hans G. Jonasson, SP Swedish National Testing and Research Institute, Sweden The group has met twice, in Lyngby on 29 th September and in Borås on 15 th November Delimitation The task of the project group has been to incorporate available knowledge in up-dated versions of NT ACOU 039 and NT ACOU 056. The project should standardise methods already available, while no new development was foreseen within the project. During its work the project group identified some topics where future development is needed. These are given at the end of the present report. 3. Method of Work DELTA wrote a method proposal which was discussed at the first project group meeting held in Lyngby on 29 th September 2000.

6 Page 6 of 25 At a meeting with Nordic road authorities two weeks after the first project group meeting, the project leader presented the proposed method to the road authority representatives, who after the meeting gave their written comments [3], [4]. The project group members collected points of view from other national experts [5]-[8] on the revised proposal. Based on these comments the project leader produced a third revised method proposal together with a draft project report. The third revised method proposal and project report was discussed at the final project meeting held in Borås on 15 th November Comments from that meeting have been introduced in the present final project report with the method proposals, which have been approved via mail correspondence by the project group. 4. Development and Updates since Measurement Instruments Since 1982 the general trend has been for calculation to replace measurement of road traffic noise. However, measurements are carried out regularly, and the equipment used for data recording and analysis has changed a lot, among other things because computer availability and applicability have exploded. New standards for sound level meters, calibrators, etc. have been specified in the Nordtest method proposals in the appendices, and the text concerning requirements for instruments has been reworked. 4.2 New Standards Various international standards and Nordtest methods have been published since None of these standards have direct implications for the proposed Nordtest method Noise Barriers ISO 10847:97, Acoustics - Determination of insertion loss of outdoor noise barriers of all types EN , Road Traffic Noise Reducing Devices Test method for determining the acoustic performance Part 1: Intrinsic characteristics of sound absorption EN , Road Traffic Noise Reducing Devices Test method for determining the acoustic performance Part 2: Intrinsic characteristics of airborne sound insulation

7 Page 7 of 25 EN , Road Traffic Noise Reducing Devices Test method for determining the acoustic performance Part 3: Normalized traffic noise spectrum Facade Sound Insulation ISO 140-5:98, Acoustics - Measurement of sound insulation in buildings and of building elements - Field measurements of airborne sound insulation of facade elements and facades Road Surface Influence on Traffic Noise ISO , Acoustics Method for Measuring the Influence of Road Surfaces on Traffic Noise Part 1: The statistical pass-by method; Part 2: The close proximity method Vehicle Noise ISO 9645:1990, Acoustics Measurement of noise emitted by two-wheeled mopeds in motion - Engineering method ISO 7188:1994, Acoustics - Measurement of noise emitted by passenger cars under conditions representative of urban driving ISO 5130:1984, Acoustics - Measurement of noise emitted by stationary road vehicles - Survey method ISO 10844:1994, Acoustics - Specification of test tracks for the purpose of measuring noise emitted by road vehicles New Nordic Prediction Method The latest version of the Nordic prediction method for road traffic noise is [9]. Among the changes in this latest version relatively to the 1978 version available in 1982 when the Nordtest method for measuring road traffic noise was published are a) revised noise source emission b) new definition of maximum noise levels, and c) information on the influence of the road surface on the traffic noise level. 4.3 New Noise Indicators In its proposal for a directive on a European noise policy the European Commission decided to use as indicators the Day, Evening, Night equivalent noise level, L DEN and the equivalent nighttime noise level, L N. The member state authorities are to decide individually on their definition of day, evening, and night. Besides, the Commission left it open for member states to decide to use supplementary noise indicators for certain spe-

8 Page 8 of 25 cial cases. Table 1 summarises the present noise indicators used by Nordic authorities in their traffic noise regulation. In Norway the 5 th percentile of the distribution of maximum noise levels is enforced, while in Sweden the 5 th highest noise level occurring during a vehicle pass-by during the night is the basis for regulation. Indicator Time interval Day Night Day Night EU L DEN L N local local DK L Aeq,24h FIN L Aeq,Day L Aeq,Night ISL L Aeq,24h N L Aeq,24h L AFmax S L Aeq,24h L AFmax Weather Influence on Sound Propagation SP Workshop 1988 In 1988 SP Swedish Testing and Research Institute arranged a workshop on how the weather conditions influence sound propagation. The main outcome was a proposal for widening the present Meteo-window and to base its definition on the sound path curvature [10] DELTA Project 1991 In 1991 DELTA carried out a project [11] for the Danish Environmental Protection Agency. The project resulted in a proposed new Meteo-window according to the principles reached at during the workshop at SP. The proposed meteo-window is applicable for point sources above flat terrain. Concerning extended sources, see Section The basis of the project was a new series of sound propagation measurements carried out by DELTA, together with results of measurements found in the literature. Octaveband attenuations were collected from a total of more than 1,200 combinations of source-receiver positions and weather conditions. Overall A-weighted noise levels of pink noise were calculated for each measurement together with the value of the normalised sound path curvature k defined in (1).

9 Page 9 of 25 k 0.6 T + u = [km -1 ] (1) 3.2 T = difference between air temperature at 10 metres and at 0.5 metres above the ground, respectively [K]. u = difference between wind speed component in the direction of propagation, at 10 metres and at 0.5 metres above the ground, respectively [m/s]. Analyses proved that a meteo-window as specified in Figure 1 could be defined, based on the normalised curvature k, and that the standard deviation of A-weighted noise levels from sources emitting pink noise was as specified in the figure. Reference is made to Appendix 1 or [11] for more detailed explanation, including the definitions of High and Low situations. For the purpose of such assessment, the frequency spectrum of road vehicle noise emission can be assumed to be approximately pink, cf. Figure 2. High: No Ingen restriction krav σφ m = 1.5 db k > -0.1 σ m = 2 db Low: k > 0.1 σ m = 2 db Effective source-receiver distance d [m] Figure 1 Requirements on sound path curvature k and associated measurement uncertainty - expressed as standard deviation σ m - due to weather influence, for various combinations of source-receiver distances and source/receiver heights. At distances d of more than 400 m: k > 0.1; σ m = 1 + d/400 [db].

10 Page 10 of 25 A-w 1/3 oct. rel. tot [db] Pink 50 km/h 90 km/h k 2k 4k Frequency [Hz] Figure 2 Normalised frequency spectra of road traffic noise (10% heavy vehicles) [12] compared with pink noise. The new meteo-window is wider than the old meteo-window for measuring environmental noise from industry. This is illustrated in Figure 3. The figure shows the old meteo-window for measuring environmental noise from industry as an inverted L -shaped surface. This surface is delimited by a) a wind speed component between 1.4 m/s 1) and 5 m/s for temperature gradients between minus 0.05 C/m and zero, or b) wind speed component between 0 and 5 m/s for temperature gradients between zero and plus 0.05 C/m. The new meteo-window is the whole area above the slant lines corresponding to k = -0.1 or k = 0.1 [km -1 ], respectively 2). For example, the new meteo-window allows measurement under upwind conditions as long as the total effect of wind speed and temperature gradient ensures that the normalised curvature is larger than the value required for that situation (defined by distance and heights). This latter means that with the new meteo-window, measurements of environmental noise from industrial plants will generally be cheaper because fewer situations outside the meteo-window will be encountered during attempts to measure the noise. 1) 2) The minimum wind speed 2 m/s at an angle of 45. The slant lines in Figure 3 have been calculated for ground surface roughness length z o = 0.02 m.

11 Page 11 of 25 Wind Speed u(10) [m/s] Old Window New Window k>0.1 k> ,1 Temperature gradient [oc/m] Figure 3 Meteo-windows. Old meteo-window for measuring environmental noise from industry: inside the frame shown. New: above the slant lines. The new meteo-window distinguishes between three classes of situations, depending on source and receiver height and on source-receiver separation. This means that steps will occur in the required curvature and in the associated standard deviation between classes as a result of minor changes in heights or distance. With the new Nordic sound propagation models available, future simulation would allow a more continuous approach than is presently possible, based on available measurement results DELTA Project 1999 In a project for the Danish Road Directorate and the Danish Environmental Protection Agency, the meteo-window has been generalised to be valid for extended sources such as a road. For roads the meteo-window for point sources shown in Figure 1 can be used, provided the normalised curvature k is determined in a vertical plane through the microphone position and perpendicularly on the road centre line 3). The mean wind direction shall be in the interval ±60 degrees around the normal from the road through the microphone position. The effective source-receiver distance d shall be determined along the dividing line of the angle between the mean wind speed vector and the normal from the road to the microphone position, cf. Figure 4. This distance shall be used to determine the standard deviation σ m by means of Figure 1 in order to determine the measurement uncertainty, cf. Section ) I.e. the curvature shall correspond to the wind speed component perpendicularly to the road.

12 Page 12 of 25 Mean wind direction Centre line d d Microphone position Figure 4 Illustration of the allowed wind direction interval and the effective source-receiver distance d, measured along the dividing line of the angle between the average wind speed vector and the normal from the road to the microphone position. The generalisation was based on simulations of the sound exposure level at various distances from a straight road. A point source was successively positioned at 20 locations on the road centre line, and the contributions at the receiver were calculated and added up. In the absence of a general model for downwind ground attenuation, an approximate model was used. First, PE-calculations were made for a point source at 0.5 m height and receiver at 2 m height, at distances m, for various wind speeds between 0 m/s and 10 m/s. Then the ground effect L M at a receiver 2 m above porous ground was calculated, for no-wind condition, according to Equations (2.36)-(2.37) in [9]. As an approximate model, the ground effect L M was expressed as L M = L M (1 p cos α) (2) p is a factor characterising the wind speed and scaling the Nordic no-wind ground attenuation to simulate downwind ground effect. α is the wind direction relatively to the normal from the receiver to the road centre line.

13 Page 13 of 25 The values of p as a function of wind speed were determined so that the deviation between PE-results and the scaled no-wind results were minimum at distances up to 200 m 4). Calculations were made of L AE for a variety of wind directions (0 α 90 ). In these simulations, the ground effect was calculated according to Equation (2), and the effect of sound absorption was included with an attenuation of db/m. Based on the calculation results, the allowed wind direction interval of plus and minus 60 was determined. For 6 m/s wind speed, the deviation from the result found for wind perpendicularly to the road was less than 1 db at 400 m distance up to a wind direction of 50, and less than 1.5 db up to a wind direction of 60. A wind speed of 6 m/s is quite high, and smaller wind speeds yield smaller deviation Present Nordtest Method The present Nordtest method [1] specifies the following conditions for measurements to be carried out with microphone positions at least 4 m above the ground, i.e. only the situations denoted High in Figure 1 are dealt with in [1]. For distances less than 30 m: no restrictions. At m distance and less than 10 db screening: wind speed more than 2 m/s measured at 10 m above the ground with a component from road to microphone position. At more than 100 m distance: wind speed more than 2 m/s measured at 10 m above the ground and less than 5 m/s measured at 2 m above the ground 5) ; component from road to microphone position; full cloud cover is required. These meteo-conditions have been summarised in Table 2. Table 2 also shows the minimum value of the normalised curvature that would occur on a midsummer day in Denmark with a wind direction just below 90. Normalised curvatures as small as -0.7 or -0.2 would be accepted in the present Nordtest method. Thus, the new meteo-window in Figure 1 seems to be more restrictive than the present meteowindow for measuring road traffic noise according to [1]. On the other hand, the present Nordtest method prescribes the measurement to be carried out twice and the highest measured noise level to be used as the final result. To always have to measure twice is very restrictive. The proposal in Appendix 1 for a new Nordtest method does not require measurements performed twice, although it does provide for better accuracy when measurements are repeated and the average result is used. 4) 5) p was in the order of the wind speed component [m/s] divided by 100. Corresponding to less than 7-9 m/s at 10 m above ground, depending on surface roughness.

14 Page 14 of 25 Distance [m] Wind speed u(10) [m/s] Wind direction α [ ] Cloud cover [-/8] Minimum k [km -1 ] < 30 Any > 2 ±90 Any -0.7 > ± Table 2 Meteo-window in the present Nordtest method [1] and corresponding minimum values of the normalised curvature. 4.5 Uncertainty General The uncertainty δ of a measurement result is defined in the Nordtest method as half the width of the 90% confidence interval. This has become practice in environmental noise matters. In that case, when testing whether a measurement result is below a noise limit or above a noise limit, 95% confidence is obtained in one-sided tests L Aeq After having carried out one measurement, the uncertainty δ of the measured L Aeq -value can be determined by Equation (3) where σ is the standard deviation of the distribution of measurement results. Only one measurement has been carried out, and the value of σ must be taken from experience obtained in similar measurements. The numbers given for contributions to σ in Sections and are based on such experience. δ = 1.65 σ (3) When more than one measurement has been carried out, the mean value shall be used as the measurement result, and the uncertainty can then be estimated by Equation (4) with n = the number of independent measurements. To be independent, measurements should in general be separated by at least 24 hours δ = σ n (4) The total standard deviation is considered consisting of a contribution σ i from instruments, a contribution σ k from variation in vehicle noise emission, a contribution σ r from

15 Page 15 of 25 the effect of reflections, and a contribution σ m from weather induced variation in sound transmission path attenuation. The total standard σ deviation is calculated by Equation (5). Typical values of σ i, σ k, and σ r are given in Section 4.5.4, while the values of σ m are mentioned in Section σ = σ + σ + σ + σ 2 i 2 k m r (5) In case three or more independent measurements have been carried out, the value of δ can be estimated directly from the N measurement results by Equation (6) t δ = N 1 N s (6) t N-1 (Student s t) for 95% confidence in one-sided tests is listed in Annex D of the method proposal in Appendix 1. s is the sample standard deviation of the actual measurement results L AFmax An estimate of the 5 th percentile of the distribution of maximum sound pressure levels can be determined by Equation (7) based on the (arithmetic) average L AFmax,avg and the sample standard deviation s of the recorded maximum levels during vehicle pass-bys L AFmax,5% = L AFmax,avg s (7) The uncertainty of L AFmax,5% can be determined in a way similar to Equations (3)-(6). For a normal distribution of data, and sample units that have been selected at random and are independent, guidelines are given in ISO 3207 [13]. The standard defines a statistical tolerance interval as an interval such that there is a fixed probability (confidence level) that the interval will contain at least a proportion p of the population from which the sample is taken. The statistical tolerance interval may be one-sided or twosided. The limits of the interval are called statistical tolerance limits. Figure 5 illustrates a normal distribution with a mean value of 85 db and a standard deviation of 1 db. For a one-sided interval to the right, there is a probability 1-α that at least a proportion p of the population is below the value determined by Equation (8) L i = x + k2 u(n,p, 1-α) s (8)

16 Page 16 of 25 x = average of n observations k2 u (n,p, 1-α) = coefficient tabulated in ISO 3207, cf. Table 3 6) s = sample standard deviation The sample standard deviation s in Equation (8) is an estimation of the population standard deviation σ. Normally contributions to σ from instruments, weather influence and reflections are less important than the variation between individual vehicle noise levels. If in special cases the sample deviation s is smaller than 1.5 or 2 db, an estimate of σ should be used in Equation (8) instead of s. σ can be calculated by means of Equation (5) with s inserted instead of σ k. L i in Equation (8) represents the upper limit of the 95% confidence interval of the 5 th percentile of the maximum sound pressure level during pass-bys of the vehicle category considered, and probably this will be the most interesting descriptor of measured maximum noise levels from road vehicles. If this upper limit does not exceed a given noise limit, there is 95% probability that the 5 th percentile is below the noise limit. The lower limit of the confidence interval of the 5 th percentile can be determined by Equation (9). If this lower limit exceeds a given noise limit, there is 95% probability that the 5 th percentile is above the noise limit. L i = x + k2 l(n,p, 1-α) s (9) x = average of n observations k2 l (n, p, 1-α) = coefficient tabulated in Table 3 7) s = sample standard deviation (estimation of the population standard deviation σ) The coefficients k 2l in Table 3 have been calculated by P. Thyregod, Institute of Mathematical Modelling, Technical University of Denmark [14]. t0. 05( n 1, 1.64 n) k2 l ( n, p,1 α) = (10) n n = sample size 6) 7) With an accuracy better than 10% the coefficients can be determined from k 2u = 2.7 (lg n) With an accuracy better than 10% the coefficient can be determined from k 2l = 1.0 (lg n)

17 Page 17 of 25 t 0.05 (f,δ) = 5 th percentile of a non-central t-distribution with f degrees of freedom and with the non-central parameter δ = u p n u p is the p th percentile of the standardised normal distribution, p = 0.05 => u p = Percentage [%] i LAFmax [db] Figure 5 Normal distribution with illustration of the 5 th percentile (full vertical line) and confidence limits calculated according to Equations (8) (10) with n = 30.

18 Page 18 of 25 Sample size n k 2 5 th percentile Sample size n k 2 5 th percentile Lower Upper Lower Upper infinity Table 3 Coefficients k 2 (n, p, 1-α) for one-sided statistical tolerance interval depending on the number n of observations, with confidence level 1-α = 0.95 and proportion of population p = 95%, after [13] and [14]. Figure 6 illustrates the development of the sampling process random numbers of a Gaussian distribution with mean value 85 db and a standard deviation of 2.5 db (as for heavy vehicles at 80 km/h) have been drawn, and the upper part of the figure shows as a

19 Page 19 of 25 function of the sample size how the value of L AFmax,5% might develop during 30 vehicle pass-bys when estimated using Equation (11). L AFmax,5% = L AFmax,avg σ (11) The figure also shows the confidence limits estimated by Equations (12) (13). Upper = L + k (n, p, -α) σ AF max, avg 2u 1 (12) Lower = L + k (n, p, -α) σ AF max, avg 2l 1 (13) The lower part of the figure shows the corresponding values determined according to Equations (7) (10), using the sample standard deviation s instead of the population standard deviation σ.

20 Page 20 of Upper Avg σ Lower LAFmax,5% [db] Sample Size [-] 95 Upper Avg s Lower LAFmax,5% [db] Sample Size [-] Figure 6 Illustration of the estimated values and the upper and lower statistical tolerance limits of the 5 th percentile of a distribution with an 85 db mean value and 2.5 db standard deviation. In the upper part of the figure, the true standard deviation σ = 2.5 db has been applied for estimating L AFmax,5%, while in the lower part the sample standard deviation of the actually sampled data has been used.

21 Page 21 of Contributions to σ This section gives guidelines for the contributions to the standard deviation in Equation (5) for overall A-weighted traffic noise levels. The measurement instrument contribution σ i depends on the equipment. When using precision instruments and handling them as specified by their manufacturer, the contribution σ i < 1 db, provided the instruments have been properly maintained, controlled, and calibrated. The contribution σ k from variation in individual vehicle noise emission due to vehicle, speed, and driving pattern variation depends mainly on the number of vehicle pass-bys during measurement. An empirical relation has been established for equivalent noise levels in [15], see Equation (14). Graphs have been included as Figure 4 in Appendix σ k [db] (14) n The specified microphone location, Section 7 of the method proposal in Appendix 1, ensures the variation in reflection contributions to be σ r < 1 db in outdoor microphone positions. The uncertainty caused by using a limited number of indoor microphone positions depends on the room volume, shape, and damping. The contribution σ r to the standard deviation of overall A-weighted levels of traffic noise is typically between 1 db and 3 db. The contribution σ m from weather induced variation has been described in Section Road Surface Influence The road surface influences the traffic noise level. As guidance a table is given in an informative annex of the Nordtest method proposal. This information can be used as a basis for judgement or comparison of measurement results. 4.7 Maximum Sound Pressure Level Definition and Determination In the Nordic prediction method for road traffic noise [9] a recommendation is given for characterising the maximum noise level by the 5 th percentile of the distribution of maximum noise levels from individual vehicle pass-bys, i.e. the maximum noise level exceeded by 5% of the vehicles of a given category. The maximum noise level should normally be determined on the basis of measured noise levels from at least 30 individual pass-bys of vehicles in the category considered. An estimate of the 5 th percentile of the distribution of maximum sound pressure levels

22 Page 22 of 25 can be determined by Equation (15) based on the (arithmetic) average L AFmax,avg and the sample standard deviation s of the recorded maximum levels during vehicle pass-bys. L AFmax,5% = L AFmax,avg s (15) In cases when it is not possible to record maximum noise levels from at least 30 vehicle pass-bys, it is recommended in Equation (15) to use the standard deviation as a function of speed given in Equations (16) (17) from the Nordic prediction method for road traffic noise. Such calculation takes advantage of the fact that numerous measurements made earlier have given information on the typical spread of data on noise from individual vehicles. Heavy s = 4.1 db 30 v 50 km/h v s = 10 e db v 50 km/h (16) v Light s = 5.5 e db v 30 km/h (17) If other than the 5 th percentile of the distribution is wanted, the factor 1.65 in Equation (15) shall be replaced by another coverage factor as described in Section 5.2 of the method proposal in Appendix 1. The above is valid for determining the maximum noise level generated by one individual vehicle passing the microphone position, presuming that is what is of interest for characterising the potential disturbance of residents sleep at sites with low traffic intensity at night. In case more vehicles pass the microphone at the same time, higher maximum noise levels occur than estimated by the above procedure, depending on the number of vehicles simultaneously contributing to the noise levels. This latter depends on the number of lanes, the traffic intensity, and the distance from the road to the microphone. On motorways, trucks often travel in groups, and in such cases higher maximum noise levels are likely to occur than determined from the above procedure for individual vehicles. One question is whether along such roads the maximum noise levels exceed the noise limit without the L Aeq,24h exceeding the limit for that noise level. This depends on the noise limits. No well established practice exists for assessing the annoyance, sleep disturbance or other measures of noise effects in such situations, and no well developed practice for measuring or predicting the relevant noise metric exists for such conditions. The project group decided to list this issue as topic for future investigation. Until results of such investigations are available, it is recommended to apply the same procedure for

23 Page 23 of 25 groups of vehicles as for individual vehicles. This implies among other things that the maximum noise levels from at least 30 vehicle group pass-bys shall be recorded. 5. Correction for Background Noise and Reverberation Time 5.1 Background Noise The proposed Nordtest methods in Appendix 1 and 2 require the background noise level to be at least 10 db below the road traffic noise level to be measured, be it an overall noise level or a frequency band noise level. Provided such a low background noise level is ensured, no correction is needed for the influence of background noise. The limit for traffic noise indoors is often L Aeq,24h = 30 db, and thus a direct measurement would require a background equivalent noise level of no more than 20 db, which is not very common in normal buildings. By measuring simultaneously outdoor and indoor during periods of high traffic intensity, with as little indoor activity as possible in the building by people and installations, an improved signal-to-noise ratio can be obtained than is available on a 24-hour basis. In order to be able to accurately correct the measurement result for the effect of background noise, the background noise level must be determined with a high degree of accuracy. This is normally not possible, and often the background noise level will have to be estimated based on recordings made during periods between vehicle pass-bys. If correction for background noise were made, there would be a high risk of overcorrection, i.e. for subtracting too much from the measured total noise level. 5.2 Reverberation Time The noise level indoors depends on the damping in the room, and indoor noise limits are sometimes defined based on a specific room reverberation time. In such cases it may be necessary to correct the measured indoor noise level for the influence of the reverberation in the room. This was not mentioned in [1] or [2]. In the method proposal in Appendix 1 reference is made to ISO [16] in which the correction as a function of the room volume is given in tables for different types of room. ISO assumes a reference reverberation time T 0 = 0,5 s. Alternatively, the room reverberation time can be measured according to European standard EN ISO [17]. The measurement result can then be corrected to any reference reverberation time.

24 Page 24 of 25 In the case of dwellings a simple procedure is proposed. If the room is furnished, no correction for reverberation is made. If the traffic noise level was measured in an empty room, a reasonably accurate estimate of the traffic noise level to be expected in the room with furniture is obtained by subtracting 3 db from the measurement result. 6. Requirements on Reporting It is the experience of authorities that reports of road traffic noise measurements are often rudimentary, meaning that important information is missing, even though the requirements in the Nordtest measurement methods are rather strict. In the proposed new Nordtest method these requirements are even stricter, and it is recommended for authorities to request measurements performed by recognised laboratories, preferable by laboratories with accreditation to measure according to the method or by persons certified to do so. Such laboratories and persons are likely to follow method specifications and reporting having done so. 7. Needs Identified for Future Development The project group identified the following topics where new research and development is needed. An updated meteo-window without the discontinuities between requirements in classes of cases. ( High / Low and distance classes, respectively). A new meteowindow could be based on series of simulations using the new Nordic models ( Nord2000 ) for sound propagation under various weather conditions. Better definition of the maximum noise levels to use in regulation. Maximum noise levels predicted according to the Nordic prediction method for road traffic noise are noise levels from the pass-by of individual vehicles. If more vehicles pass more or less simultaneously, higher noise levels will occur. The increase in maximum noise level depends on the number of vehicles and on their separation in time and space which in turn depends on the traffic intensity, the number of lanes, and the distance from road to measurement position. The project group recommends for road- and environmental authorities in the future to require traffic noise measurements to be carried out by laboratories accredited or persons certified for such measurements in order to control the quality of measurement results and documentation. Such requirements could be put forward in regional or national regulation.

25 Page 25 of References [1] Nordtest Method NT ACOU 039, Road Traffic: Noise. [2] Nordtest Method NT ACOU 056, Road Traffic: Noise Simplified Method. [3] H. Bendtsen, personal communication, November [4] A. Neergård, personal communication, November [5] S. Gudmundsson, personal communication, November [6] H. G. Jonasson, personal communication, November [7] J. Parmanen, personal communication, November [8] S. Å. Storeheier, personal communication, November [9] TemaNord 19996:525, Road Traffic Noise, Nordic Prediction Method, Nordic Council of Ministers [10] H. G. Jonasson, How to expand the meteorological window for noise immission measurements, SP Report 1988:37, Borås [11] J. Kragh, New Meteo-Window for measuring environmental noise from industrial plants (in Danish), DELTA report No. 148, A summary can be found in Proceedings Internoise 91, p , Sydney [12] Nordtest Method NT ACOU 061, Windows: Traffic Noise Reduction Indices. [13] ISO 3207, Statistical interpretation of data Determination of a statistical tolerance interval. [14] P. Thyregod, personal communication, July-August [15] S. Å. Storeheier, Measurement of noise immission from road traffic (in Norwegian), SINTEF Report No. STF44 A78025, Trondheim [16] ISO/DIS 10052:2000, Acoustics Field measurements of airborne and impact sound insulation and equipment sound Survey method 8). [17] ISO 140-5, Acoustics - Measurement of sound insulation in buildings and of building element - Field measurements of airborne sound insulation of facade elements and facades. 8) At present at the stage of draft.