A methodological proposal to measure rolling noise under real road use conditions Pedro Atanasio-Moraga 1 Laboratorio de Acústica (Lambda), Departamento de Física Aplicada, INTERRA, Universidad de Extremadura. Avda. de la Universidad, s/n, 10003 Cáceres, Spain. Manuel Sánchez Fernández 2 Universidad de Extremadura, INTERRA, NEXUS, Avda. de la Universidad s/n, 10003 Cáceres, Spain. Guillermo Rey-Gozalo 3 , David Montes González 4 , Rosendo Vílchez-Gómez 5 , Alicia Bachiller León 6 , Juan Miguel Barrigón Morillas 7 Laboratorio de Acústica (Lambda), Departamento de Física Aplicada, INTERRA, Universidad de Extremadura. Avda. de la Universidad, s/n, 10003 Cáceres, Spain. ABSTRACT

inter noise 21-24 AUGUST SCOTTISH EVENT CAMPUS GLASGOW

The consideration of temperature as a factor influencing the sound power generated by tyre contact with the road has been known for several decades. However, the introduction of this effect in the methods recommended for the elaboration of strategic maps is very recent. So far, there is an extensive scientific literature that studies this issue using different standardised methods. They generally involve controlled conditions for different variables, for example, the type of tyre. On the other hand, studies presenting methodologies for measuring the effect of temperature on rolling noise, under normal road use conditions and using the basic indicators for the preparation of noise maps, are very recent. This study presents a methodological proposal to measure the effect of air and pavement temperature. The spectral analysis differs in low band frequency from the results reported in the literature for controlled conditions or with the recommendations of CNOSSSOS-EU, being similar in mid and high bands frequencies.

1 pedroam@unex.es 2 msf@unex.es 3 guille@unex.es 4 davidmg@unex.es 5 vilchez@unex.es 6 aliciabachiller@unex.es 7 barrigon@unex . es

1. INTRODUCTION

Road traffic is the main source of noise in cities even in quiet areas or in vulnerable buildings [1, 2]. In fact, most noise maps only model road traffic [3]. Noise generated by road traffic is conditioned by multiple factors. Thus, the characteristics of the infrastructure itself, the characteristics of the road traffic and environmental factors influence the sound levels recorded on roads. The own characteristics of infrastructure are parameters that can be controlled in the design, modification, or maintenance process. Some of these characteristics are the type of pavement, the state of preservation, the design of the geometry, etc. In contrast to the factors associated with the infrastructure itself, those associated with traffic characteristics, such as vehicle flow, speed, type of vehicles, their inherent characteristics or the tires used, are more complex to control generally due to the degree of accessibility of roads. Regards meteorological factors, these affects both the propagation and the generation of sound energy. The influence of air and pavement temperature on sound pressure level generation is studied in scientific literature [4-9]. These works are developed on the basis of reference standards for the acoustic characterization of a pavement. In general, the relationship between the sound pressure level and the temperature of the air or roadway is carried out through linear relationships. The variability found are between -0.03 and -0.11 dBA/°C. Some authors [4-6] analyse the effect described in 1/3 octave bands, showing different behaviours depending on the frequency. Generally, a higher ratio is observed at high and low frequencies, meaning low frequency as 31.5 Hz to 630 Hz, medium frequency as 800 Hz to 1250 Hz and high frequency as 1.6 kHz to 5 kHz, and negative coefficients of temperature-sound level ratio, based on a linear relationship. Anfosso-Lédée et al. [4] observe that temperature affects the sound pressure level with greater magnitude in the low frequency range from 125 Hz to 500 Hz, and in high frequency from 1.6 kHz to 5 kHz. In general, these results show a trend of decreasing sound level with increasing temperature. However, in the analysis in 1/3 octave bands between 125 Hz and 315 Hz, it is evident that an increase from 21 °C to 26 °C produces an increase in the sound level registered at these frequencies. The 1/3 octave frequency analysis performed by Bühlmann et al. [6] shows that the sound level is inversely proportional to temperature. The quality criterion established in this work, studied by linear regression, is standard error ≤ 0.2dBA or R 2 ≥ 0.9. The values found vary according to the type of tire, the material, and the road porosity. The frequencies that showed the most significative relationship respect the sound level recorded were found at medium frequency (800 Hz to 1250 Hz). The sound level-temperature relationship in 1/3 octave bands is also analysed by Bueno [5] at frequencies from 200 Hz to 16 kHz. Overall, it shows an inversely proportional relationship between temperature and sound level. The 200 Hz and 250 Hz frequencies show a coefficient of regression close to zero (˂-0.02). The value of the largest magnitude (≈ -0.10) is recorded at the 2 kHz frequency. In the work of Bühlmann the coefficient of linear variation of sound level with respect to temperature is -0.06 in broadband. This paper proposes a method to measure the effect of air and pavement temperature on sound levels on a two-way road with high traffic flow. Eighteen measurements were carried out in each of the two consecutive days, with similar atmospheric and traffic conditions to reduce the variability related to these factors, in a flat and absorbing area from the acoustic point of view. Relationship between L N and the air and pavement temperature for each day was obtained. 2. METHODOLOGY

In this study, in situ measurements were carried out in the N-521 national road (Figure 1), which connects Cáceres with Malpartida de Cáceres, in Extremadura (Spain). This two-way road has two lanes and a 1 m wide roadside. The pavement is considered NL01 class [10], porous asphalt [11], composed of 8 cm of bituminous concrete AC22s in the lower layer, and 3 cm of discontinuous

bituminous mix BBTM 11B in the upper part [12]. According to the General Department of Roads of the Ministry of Public Works of the Government of Spain [13], this road supported an Average Daily Intensity (IMD), that is, the number of vehicles passing through a stretch of road in 24 hours, of 8542 vehicles, with a percentage of 4.07 % of heavy vehicles and 95.93 % of light vehicles in 2018. Most of these vehicles transit during daylight hours, so initially we can predict a traffic flow of 600 vehicles per hour during this day time period.

15m

Figure 1.- Measurement point. The main purpose of this study is to analyze the relationship between traffic noise and air and asphalt temperature in the N-521 road. Taking into account the flow previously indicated, 2 campaigns of 18 noise measurements of 10 minutes each were carried out, ensuring at least 100 vehicles for each measurement. Since each vehicle passing through the measurement point with a different speed, it is of a specific range and it has different maintenance and tire conditions, it is required to establish a compensation so that the deviations in the sound level associated with the differences in speed and with the characteristics and conditions of the vehicles, are counterbalanced by using an average sound pressure level recorded during a given period of time, thus obtaining an appropriate average of its effects on the noise level. In addition, the same number of vehicles do not always pass through each measurement, so that the equivalent continuous sound level recorded is influenced by the number of vehicles which pass in front of the equipment in each measurement. It is therefore necessary to establish a normalization of the results with respect to a reference flow, set at 780 vehicles/hour after vehicle counting in situ . 𝐿 ே ൌ𝐿 ଴ ∗10𝑙𝑜𝑔 ଵ଴ ቀ ௏ ೘ ଻଼଴ ቁ, (1) where 𝐿 ை is the recorded sound pressure level; 𝑉 ௠ is the total number of vehicles recorded in 10 min. Measurements were performed with a class 1 sound level meter located 15 m from the central axis of the road [14], without obstacles or reflective surfaces between them that could influence the registered sound levels [15,16], and at a height of 1.5 m [17]. The sound level meter was calibrated before and after each measurement season, recording the equivalent A-weighted sound level in spectrum in 1/3 octave bands. These distances were chosen because in this study we aim to have an overview of the effect of temperature in the in situ noise that occurs in the environment of a road with high traffic flow without reflection of the vehicles. The measurement area is considered to be flat and absorbing from the acoustic point of view. Before and after each measurement, road and air temperature, as well as humidity and wind were recorded. Road temperature was taken with a thermographic camera, pointing at the pavement of the nearest lane, while air temperature and humidity were taken with a thermohygrometer placed in the shade at a height of 1.5 m. Wind speed during both campaigns was zero or close to zero. Vehicles were categorized for each line of traffic according to the CNOSSOS-EU method indicated in the European Directive on Environmental Noise for establishing common noise assessment methods according to Directive 2002/49/EC [11]. In the selected section, the road is limited to a maximum speed of 100 km/h. The speed of passing vehicles was controlled with a radar

placed in the proximity of the measurement point, obtaining speed differences in a range of ±5 km/h, which would mean a maximum variation of 1 dBA according to the Vehicle Technology Institute [18]. 3. RESULTS AND DISCUSSION

3.1. General conditions

Two campaigns of 18 measurements were carried out in consecutive days, at similar time of day and with similar atmospheric conditions in order to reduce the variability related to these meteorological factors. The air temperature is very similar in the central time of measurement in both campaigns, finding small differences in the initial and final sections (Figure 2a). These variations are also observed in the relative humidity of the air in the initial and final sections of the measurement (Figure 2b).

Air Temperature C) PBRRBEBS 9:00 10:00 11:00 12:00 Time of day —e-first season second season 13:00 14:00 15:00

Figure 2: a) Evolution of air temperature (°C) over time in the two measurement campaigns; b) Evolution of relative air humidity (%) over time in the two measurement campaigns Regarding to the pavement, there is a small difference in temperature in the initial measurements between the two seasons, being very similar in the rest of the day (Figure 3).

sa Humidity( %) 0 8:00 first season, —e—second season 9:00 10:00 11:00 12:00 Time of the day 13:00 14:00 15:00

Figure 3: Evolution of pavement temperature (°C) over time in the two measurement campaigns Concerning road traffic, there is an increase in the number of vehicles in the initial and final hours of the measurement time period (Figure 4a). It should be recalled that this road is used to access to the western area of the city of Cáceres, where different commercial areas and companies are located, in addition to providing access to the city centre, justifying the increase in the number of vehicles in those hours of the measurement timetable. Considering the two categories of heavy vehicles indicated in CNOSSOS-EU [11], the percentage of this type of vehicles is similar in the two campaigns, between 2 % and 8 % (Figure 4b, Figure 4c).

Pavement temperature (°C) BBRERS 8:00 9:00 10:00 11:00 12:00 Time of the day efirst season —esecond season 13:00 14:00

a)

Py 3 first season . ‘@second season S sor 2 2 E 4.0% 3.0% 3 . 20% oa z . . z *|. E06 4 . S oe? eS ge ° 0.0% aces oe ee t. 00 9:00——«10:00» «11:00» 12:00 1350040015100 ‘Time of the day

b)

Normalysed total vehicles 210 200 190 180 170 160 150 140 130 120 110 8:00 9:00 10:00 11:00 12:00 Time of day @ first season second season 13:00 14:00 15:00

c) Figure 4: a) Total number of vehicles in both seasons; b) Percentage of medium heavy vehicles according to CNOSSOS-EU; c) Percentage of large heavy vehicles according to CNOSSOS-EU Around 95 % of the vehicles is light traffic, while only 5 % is heavy traffic, including both types of heavy vehicles contemplated in CNOSSOS-EU [11]. Although it could have a minor effect on the variability of the equivalent level of each measurement, a second normalization was performed considering the differences in the sound power emitted between CNOSSOS-EU category 1, light

3 es 2 3 3 2,0% 1,0% Percentage of medium heavy vehicles % 0,0% 8:00 9:00 3 10:00 11:00 12:00 Time of the day 13:00 first season © second season 14:00 15:00

vehicles, and categories 2, 3 and 4 [11], considering the equivalence between the noise level emitted by category 1 vehicles and the rest of the vehicles of the other categories [19]. The coefficients proposed by Sandberg [17] were applied for categories 2, 3 and 4. Finally, the equivalent vehicle value was normalised based on the average total equivalent category 1 vehicle flow in the measurements (a total equivalent of 930 category 1 vehicles per hour). 𝐿 ே ൌ𝐿 ଴ ∗10𝑙𝑜𝑔 ଵ଴ ቀ ௏ ೘ ଽଷ଴ ቁ, (2) where 𝐿 ை is the recorded sound pressure level in 1/3 octave band; 𝑉 ௠ is the total number of vehicles recorded in 10 min 3.2. Analysis in 1/3 octave bands

A detailed analysis has been conducted in 1/3 octave bands in the 50 Hz to 10 kHz range for air and pavement temperature for the two seasons of measurement. Figure 5 shows the results for the relationship between 𝐿 ே and the air and pavement temperatures for each day. Considering that the number of data used in the regression is the same for all frequencies, the coefficient of determination for which it was significant, with a probability of 95 %, 99 % and 99.9 %, for the air and pavement temperatures in the different days was obtained (Figure 5a, Figure 5c).

Coefficient of determination 1,00 090 080 0.70 0,60 p-value = 0.001 oo 7 bo wf ooo $e¢s 3 § $s Fess Fs FP RFS FF FF FZ Z “pavement temperature day 1 -*-pavement temperature day 2

a)

b)

ssa (oavap g) 1uaxo43909 woyssaui 333 3 3 a p01 Dg arI¢9 ans 7p apse is ae AD anise ary *H008 *H0¢9 M005 *HO0p sig Hose “Hove 40091 sz “Hoo 7H03 7H¢9 Hos 3 Frequencies lay 2 -+pavement temperature d re day 1 pavement temperature day

c)

d) Figure 5: a) Coefficient of determination for pavement temperature for day 1 and for day 2; b) Regression coefficient for pavement temperature for day 1 and for day 2; c) Coefficient of determination for air temperature for day 1 and for day 2; d) Regression coefficient for air temperature for day 1 and for day 2.

p—value = 0.001 p-value = 0.01 p-value = 0.05 7 g253922533 3 “1001 aang “N16 as aay ase aS aayz 7119 arse ar “008 "069 “00s =00y "sie “sz "002 "091 “usc "001 7H0g “189 Hos Frequencies ir temperature day 2 air temperature day 1

Even with similar conditions in both measurement days, differences were found in the results achieved. Significant relationship is observed in the 63 Hz band in pavement temperature (p-value < 0.05) and air temperature (p-value < 0.01) on the second day of measurement. In both cases, the negative slope (Figure 5b, Figure 5d) indicates that an increase in temperature implies a decrease in sound level. There is no significant relationship in the 63 Hz band on the first day of measurement. In the pavement, the next frequency bands in which we found a significant relationship (p-value < 0.05) between 𝐿 ே and temperature are those in the range of 160 Hz to 400 Hz for the first day of measurement. In this case and for this range of frequency bands, the slope indicates that an increase in temperature implies an increase in sound level. In this frequency range, no significant relationships are established for the second day of measurement. In the case of air, significant relationships (p-value < 0.05) were found in the 160 Hz, 200 Hz and 400 Hz bands for the first day of measurement. Again, the slope indicates that an increase in temperature implies an increase in sound level for the indicated frequency bands. No significant relationships are established for the second day of measurement in these 3 frequency bands. The next significant relationship has been found for pavement in the 800Hz and 3.15 kHz range of frequency bands. The relationship is highly significant (p-value < 0.001) for all bands for the first day of measurement except for 800 Hz, which is highly significant (p-value < 0.01), and for 3.15 kHz which is significant (p-value < 0.05). For the second day of measurement, the relationship is highly significant (p-value < 0.001) for the 800 Hz to 2 kHz bands, being highly significant (p-value <0.01) for the 2.5kHz band and significant (p-value < 0.05) in the 3.15 kHz band. In both cases the slopes are negative, indicating that if the pavement temperature increases, the sound level decreases.

oyssou ava g)aus}945909 woy Z1DV01 aD DIe9 ans aby 4Dsre anise itr D9" anise ay 7H008 7H0¢9 oad Asie Hose “002 7091 sz *H001 708 *He9 “Hos Frequencies air temperature day 2 air temperature day 1

Similar results were observed in the case of air. In the first day of measurement highly significant relationships (p-value < 0.001) have been found in the 800 Hz to 2.5 kHz 1/3 octave bands range, being significant (p-value < 0.05) in the 3.15 kHz band. The trend is similar in the second day of measurement. Highly significant relationships (p-value < 0.001) have been found in the 800 Hz to 2 kHz 1/3 octave bands range, being highly significant (p-value < 0.01) for the case of the 2.5 kHz band, and significant (p-value < 0.05) for the 3.15 kHz band. Again, the negative slope indicates that an increase in air temperature implies a decrease in sound levels. Regarding the results found in the low-frequency bands, specially for the first season of measurement in wich we found a positive linear relationship, except for 63 Hz in wich we found a negative linear relationship, between sound pressure level and the pavement and air temperature, it is remarkable the study of Bühlmann and Ziegler [6] in which positive values were obtained for the coefficients of variation in the sound pressure level with temperature, although this result was discarded in the subsequent analysis made for Bühlmann and Ziegler [6]. Furthermore, the finding results in this work are similar trend to those reported for several authors [4-6] for the range 800 Hz to 3.15 kHz, in wich a negative linear relationship between sound pressure and temperature were found. In another research about the dependence of pavement texture on road traffic noise using the CPX method, A. del Pizzo et al. [20] found a positive linear relationship between sound pressure level and megatexture at low frequency associated with tire vibration, and negative at high frequencies and macrotexture associated with aerodynamic mechanisms. Our work shows similars results in 1/3 octave frequency bands for the pavement. 4. CONCLUSIONS

An in-situ study of two seasons of 18 measurements was carried out on a national road in order to analyse the relationship between road traffic noise and air and asphalt temperature, with real traffic conditions. Normalizations have been carried out both for the levels of the sound sources and for the type of vehicles circulating on the studied road. In this work, we can conclude: Significant relationships between sound levels and air and pavement temperature were found in the 63 Hz band and 800 Hz and 3.15 kHz 1/3 octave bands range for both days of measurement. The negative slope indicates that, in these bands when the temperature increases, the sound level decreases. Significant relationships have been found in the 160Hz to 400Hz 1/3 octave bands range for the first day of measurement for pavement temperature, with positive regression coefficient. The relationships are similar on the first day of measurement for air except for the 315 Hz band where no significant relationship was found. Even with similar conditions, no significant relationships were found in this range of 1/3 octave frequency bands for the second day of measurement in any case. A more detailed study in the low frequency emission range can be developed in this line of work to determine the implications of this aspect. 5. ACKNOWLEDGEMENTS

This project was co-financed by European Regional Development Fund (ERDF) and Junta de Extremadura, Consejería de Economía, Ciencia y Agenda Digital (IB18050 and GR21061). This work was co-financed by Consejería de Educacion y Empleo of Junta de Extremadura , the Extremadura Public Employment Service (SEXPE) and European Union (FSE) through grants to promote the hiring of research support personnel in the Autonomous Community of Extremadura (PAI20/40). This work was also supported by Consejería de Economía, Ciencia y Agenda Digital of Junta de Extremadura and European Regional Development Fund (ERDF) through grants for the financing of industrial research and experimental development projects (IDA-19-0022-3).

This work was also supported by Consejería de Economía, Ciencia y Agenda Digital of Junta de Extremadura through grants for attracting and returning research talent to R&D&I centres belonging to the Extremadura Science, Technology, and Innovation System (TA18019), where University of Extremadura was the beneficiary entity. REFERENCES

1. Rey Gozalo, G., & Barrigón Morillas, J. M. Perceptions and effects of the acoustic environment in quiet residential areas. Journal of the Acoustical Society of America , 141 (4) , 2418–2429. (2017). https://doi.org/10.1121/1.4979335 2. Montes González D., Barrigón Morillas J.M., Gómez Escobar, V., Vílchez-Gómez, R., Rey Gozalo, G., Atanasio Moraga P., & Méndez-Sierra, J.A. Environmental noise around hospital areas: A case study. Environments 6, 41 . (2019). https://doi.org/10.3390/environments6040041 3. Rey Gozalo, G., Gómez Escobar, V., Barrigón Morillas J.M., Montes González D., & Atanasio Moraga P. Statistical attribution of errors in urban noise modeling. Applied Acoustics 153 , 20- 29. (2019). https://doi.org/10.1016/j.apacoust.2019.04.001 4. Anfosso-Lédée, F., & Pichaud, Y. Temperature effect on tyre-road noise. Applied Acoustics 68 , 1–16. (2007). https://doi.org/10.1016/j.apacoust.2006.06.001 5. Bueno, M., Luong, J., Viñuela, U., Terán, F., & Paje, S.E. Pavement temperature influence on close proximity tire/road noise. Applied Acoustics 72 , 829–835. (2011). https://doi.org/10.1016/j.apacoust.2011.05.005 6. Bühlmann, E., & Ziegler, T. Temperature effects on tyre/road noise measurements. 40th International Congress and Exposition on Noise Control Engineering 2011, INTER-NOISE 2011 , pp. 557–564. Osaka Japan, september 4-7, 2011 7. Kneib, G., Belcher, D., Beckenbauer, T. & Beyeler, H.P. Continuous road traffic noise monitoring and aging of asphalt surfaces. Proceedings of the INTER-NOISE 2016 - 45th International Congress and Exposition on Noise Control Engineering: Towards a Quieter Future . pp. 7122–7133. Hamburg, Germany, August 21-24 2016. 8. Liao, G., Heitzman, M., West, R., Wang, S. & Ai, C. Temperature effects on the Correlations between Tire-Pavement Noises and Pavement Surface Characteristics . New Frontiers in Road and Airport Engineering - Selected Papers from the 2015 International Symposium on Frontiers of Road and Airport Engineering, IFRAE 2015 . pp. 219–232. Shangai, China, October 26-28, 2015. https://doi.org/10.1061/9780784414255.021 9. Yuan, M., Ni, D., Zhang, D., Wei, X., Wang, Z. & Wang, C. Effect of temperature on road pass- by noise of light vehicle. IOP Conference Series: Earth and Environmental Science. IOP Publishing , p. 22003. (2019) 10. ISO/TS 13471-1. Acoustics. Temperature influence on tyre/road noise measurement-Part 1: Correction for temperature when testing with the CPX method . (2017)

11. European Directive. Commission Directive (EU) 2015/996 of 19 May 2015 establishing common noise assessment methods according to Directive 2002/49/EC. European Union . (2015)

12. Ministry of Transport of Spain. Parte 5a Firmes. Pliego de Prescripciones Técnicas Generales para Obras de Carreteras y Puentes (PG-3). (2015) 13. https://www.mitma.es/carreteras/trafico-velocidades-y-accidentes-mapa-estimacion-y-

evolucion/datos-mensuales-de-trafico/2018 . With access 22/04/2022. 14. Montes González, D., Barrigón Morillas, J.M. & Rey Gozalo, G. The influence of microphone location on the results of urban noise measurements. Applied Acoustics 90 , 64–73. (2015). https://doi.org/10.1016/j.apacoust.2014.11.001.

15. Montes González D., Barrigón Morillas J.M., & Rey Gozalo, G. Acoustic behaviour of plates made of different materials for measurements with the microphone flush mounted. Applied Acoustics , 132 , 135-141. (2018). https://doi.org/10.1016/j.apacoust.2017.11.011

16. Montes González D., Barrigón Morillas J.M., & Rey Gozalo, G. (2020). Microphone position and noise exposure assessment of building façades. Applied Acoustics 160 , 107157. https://doi.org/10.1016/j.apacoust.2019.107157 17. Sandberg, U. Vehicle categories for description of noise source. HARMONOISE , 2003.

18. Investigations on Noise Emission of Motor Vehicles in Road Traffic. Institute for Vehicle Technology . Wuerselen, Germany, 2005.

19. Sánchez-Fernández, M., Barrigón Morillas, J.M., Montes González, D. & Rey Gozalo, G. Relationship between temperature and road traffic noise under actual conditions of continuous vehicle flow. Transportation Research Part D: Transport and Environment 100 , 103056. (2021). https://doi.org/10.1016/j.trd.2021.103056

20. Del Pizzo, A., Teti, L., Moro, A., Bianco, F., Fredianelli, L. & Licitra, G. Influence of texture on tyre road noise spectra in rubberized pavements. Applied Acoustics . 159 , 107080. (2020). https://doi.org/10.1016/j.apacoust.2019.107080