KSCE Journal of Civil Engineering (0000) 00(0):1-5 Copyright ⓒ2018 Korean Society of Civil Engineers DOI 10.1007/s12205-018-9501-3

Highway Engineering

pISSN 1226-7988, eISSN 1976-3808 www.springer.com/12205

TECHNICAL NOTE

Tire-Pavement Noise Prediction using Asphalt Pavement Texture Seong Jae Hong*, Sung-Wook Park**, and Seung Woo Lee*** Received June 21, 2017/Revised October 20, 2017/Accepted November 19, 2017/Published Online January 8, 2018

··································································································································································································································

Abstract Implementation of noise as an additional measure in pavement service evaluation has come forth since noise seriously influences on driver comfort. Due to a considerable amount of time and cost for traditional road-side noise measurement such as pass-by method, evaluation of tire-pavement noise based on MPD (Mean Profile Depth) has been attempted by numerous researchers. MPD is a part of texture characteristics that influences tire-pavement noise and it can be easily measured by Portable Laser Profiler (PLP). In general, macrotexture is a pavement surface characteristic having the significant impact on tire-pavement noise. Generally, it can confirm that tire-pavement noise increases by the increase of texture depth and wavelength. However, the theoretical study showed that the influences of pavement texture characteristics including texture depth and wavelength on tire-pavement noise are insufficient. In this study, the statistical analysis on relationship between tire-pavement noise, MPD, and wavelength for asphalt concrete pavement was performed. The collected data of measured tire-pavement noise, MPD, and wavelength on asphalt pavement sections from Donghae Expressway, and National Highway 35 Line of South Korea, were conducted and analyzed. Based on the multiple regression analysis results, it was found that tire-pavement noise is linearly increased due to the increases of MPD and average wavelength. In addition, multiple regression analysis of tire-pavement noise level, MPD, and average wavelength gave a similar trends compared to the single regression analysis between tire pavement noise with MPD and tire pavement noise with average wavelength. This study also showed that aggregate gradation used in asphalt mixture significantly affect the MPD and average wavelength. Therefore, tire-pavement noise would be affected by aggregation gradation used in asphalt mixture. Keywords: Texture Characteristic, Mean Profile Depth (MPD), Tire-Pavement Noise, Average Wavelength ··································································································································································································································

1. Introduction The serviceability level of pavement is generally evaluated on the basis of the severity and amount of pavement distress and surface roughness. However, it may be necessary to include pavement noise in serviceability evaluation since noise seriously influences driver comfort. However, a considerable amount of time and cost for the measurement of traditional road-side noise, such as the pass-by method, may be a major obstacle to including noise as one of monitoring parameters in pavement management systems. Therefore, developing a costeffective and speedy method of noise evaluation to determine a monitoring parameter in pavement management system is a crucial challenge. Recently, there has been an evaluation of tire-pavement noise based on the Mean Profile Depth (MPD) (Hong et al., 2013). The MPD can be obtained from a Portable Laser Profiler (PLP), which is used to evaluate road roughness as shown in Fig. 1. The PLP emits and receives a returning laser beam. The time consumed by the beam’s travel is translated to the texture depth of the road surface. Simultaneously, measurement of speed and

distance enable us to obtain information on the texture depth at a given location. Thus, the PLP can measure the texture depth with a fixed interval regardless of the vehicle’s speed. Indices representing the texture depth of road surfaces are widely accepted. It is calculated from the data of 100 measured profile depths (level) at every 1 mm interval of every 100 mm baseline with the data of 100 depths. The final MPD is defined as the difference between the average level of the baseline and the average peak level as depicted in Fig. 2.

Fig. 1. Road Roughness Profiling Device (Budras, 2001)

*Member, Full-time Researcher, Dept. of Civil Engineering, Gangnung-Wonju National University, Gangwondo 25457, Korea (E-mail: [email protected]) **Associate Professor, Dept. of Electronic Engineering, Gangnung-Wonju National University, Gangwondo 25457, Korea (E-mail: [email protected]) ***Member, Professor, Dept. of Civil Engineering, Gangnung-Wonju National University, Gangwondo 25457, Korea (Corresponding Author, E-mail: [email protected]) −1−

Seong Jae Hong, Sung-Wook Park, and Seung Woo Lee

Fig. 2. Definition of MPD

Fig. 3. Pavement Surface Characteristic Classifications and Their Impact on Pavement Performance Measures (Rasmussen and Bernhard, 2007)

( Peak Level 1st ) + ( Peak Level 2nd ) MPD = --------------------------------------------------------------------------------------- – Average Level 2

(1)

The MPD can be obtained easily without traffic control (Hong et al., 2013), and used it to estimate tire-pavement noise. Hanson et al. (2004) investigated the relationship between the MPD and tire-pavement noise using 46 measured pairs of MPD and tirepavement noise in asphalt concrete pavement sections in Alabama, Nevada, Texas and Colorado. However, the absolute relationship was not observed. Using only a small amount of data in the analysis was probably the cause of a scattered relationship. CTRE (2006) showed the variation of the On-board Sound Intensity (OBSI) and the variation of maximum texture height for various texture patterns of concrete pavement and summarized the range of the OBSI according to pavement texture types of concrete pavement. Pouliot and Langlois (2006) collected MPD and Tire-Pavement noise data in pavements in Quebec. ARAN and NCAT-CPX trailers were used to measure the MPD and tirepavement noise, respectively, and the linear relationship between the MPD and tire-pavement noise was observed. Hong et al. (2013) discussed the need for comprehensive consideration of the pavement texture characteristics such as wavelength and shape for reliable prediction of tire-pavement noise based on a review from the CTRE report (2006) and the study of Rasmussen and Bernhard (2007). Figure 3 describes the impact of texture wavelength on various pavement surface characteristics. This figure addresses the macrotexture area of asphalt pavement where the wavelength spans from 0.5 to 51 mm and the texture depth from 0.1 to 20 mm influences tire-pavement noise. More specifically, Rasmussen and Bernhard (2007) reported that as the wavelength of macrotexture got longer the tire-pavement noise level may get bigger. Index of Average Wavelength was suggested to represent wavelength of texture for evaluating tire-pavement noise. An equation for estimating tire-pavement noise by using two parameters, MPD and Average Wavelength, was proposed in this study. Also, the improvement of tire-pavement noise estimation using two parameters compare to the one using only MPD was explored for test road in Korean highway.

2. Database construction of MPD, Average Wavelength and Tire-Pavement Noise for Test Sections in Korean Highway According to Rasmussen and Bernhard (2007), tire-pavement noise level increases as the wavelength of pavement texture gets longer. The wavelength of pavement is difficult to analyze since it has various wavelengths of one tire-pavement noise. Therefore, a simple index, Average Wavelength, is suggested to represent the wavelength of pavement texture in this study. Average Wavelength is calculated from a number of peaks as defined in Eq. (2). Fig. 4 illustrates the relationship between Average Wavelength and texture wavelength. As Average Wavelength is large, the texture wavelength has a tendency to be small. Reversely, if Average Wavelength is small, the texture wavelength has a tendency to be large and it is expected to induce high noise level. Average Wavelength = 100 mm/number of inflection points above average level (2) Unit length of each dataset was designed to be 22 meters which is the passing distance during 1 second for 80 km/hour vehicle speed. Each dataset consists of average tire-pavement noise, MPD and Average wavelength of unit length. After collecting profile depth measured at every 1 mm interval at the test roads, it is necessary to convert those data to MPD and

Fig. 4. Definition of Average Wavelength −2−

KSCE Journal of Civil Engineering

Tire-Pavement Noise Prediction using Asphalt Pavement Texture

Table 1. Test Section for Collection of Profile and Tire-Pavement Noise Road Routes

Location

Construction

DongHae Expressway National Highway 35 Line

S. Gangnenung IC ~ N. Gangnenung IC SungSan ~ WangSan

2008 2010

average wavelength. Since MPD and average wavelength were calculated at every 0.1 m, 220 of MPD and average wavelength values were collected at every second. The average value of MPD, average wavelength during 1 second was calculated and saved in each dataset. Tire-pavement noise, on the other hand, was measured 10 times in every second. The average of tire pavement noise within 1 second was calculated and saved in each data set as well. Five hundred ninety seven datasets were constructed for test sites of DongHae Expressway (S. Gangnenung IC ~ N. Gangnenung IC) and National Highway 35 Line (SungSan ~ WangSan) as shown in Table 1. Figure 5 shows the procedure of MPD, average wavelength calculation. MPD and average wavelength calculation are explained in Eq. (1) and (2). The driving speed of test vehicle

Collected data set (MPD, Aver age Wavelength, Tire-pavement noise) 244 353

was kept as 80 km/hour and PLP collected the profile depths at every interval of 1 mm. MPD and average wavelength were calculated at every interval of 100 mm based on the collected profile depths and stored in database. Since the vehicle pass the unit length (22 m) during 1 second, MPD and average wavelength are calculated 220 times per unit length. The averages of MPD per unit length was calculated and saved in database. Tire-pavement noise was measured according to the instructions of Leeuwen and Reubsaet (2007) who used the microphones at eight inches from the center of the tire and four inches above the surface pavement following Close-Proximity Method (CPX) ISO Standard (ISO11819-2, 1997). Fig. 6 shows the procedure of collecting tire-pavement noise levels. The experiment vehicle was driven at 80 km/h constantly while the tire-pavement noise

Fig. 5. Procedure for MPD and Average Wavelength Calculation

Fig. 6. Construction Procedure of Tire-Pavement Noise Database Vol. 00, No. 0 / 000 0000

−3−

Seong Jae Hong, Sung-Wook Park, and Seung Woo Lee

measured at every 0.1 second was delivered to computer and the average of tire-pavement noises during 1 second was calculated and saved in the database.

3. Relationship between Tire-Pavement Noise and Texture Measured MPD in DGA (Dense Graded Asphalt) and SMA (Stone Mastic Asphalt) are ranged between 0.69 to 1.5 mm and from 1.26 to 2.17 mm, respectively. The correlation between on MPD Data and Tire-pavement noise Data is performed and shown in Eq. (3). Fig. 7 represents the correlation results of the MPD and tire-pavement noise, with 0.59 of determination coefficient and 0.0 of p-value. As shown, tire-pavement noise linearly increases when the MPD increased. This might be caused by increase of air pumping when the texture depth in pavement surface increases and also increasing the entrapped air in pavement surface texture. Estimated Noise, dB(A) = 3.13 MPD + 95.57, R2 = 0.59, P-value = 0

(3)

where Estimated Noise, dB(A) = Estimated Tire-Pavement Noise, MPD = Mean Profile Depth

Fig. 9. Comparison of Predicted and Measured Tire-Pavement Noise

The measured average wavelength data were ranged from 3.64 to 4.09 mm with DGA and from 3.80 to 4.24 mm with SMA. The correlation analysis result on average wavelength data and tire-pavement noise data showed that average wavelength and tire-pavement noise linearly increase as specified in Fig. 8, with 0.58 determination coefficient, and 0.0 of p-value. The results from this study were consistent with that of Rasmussen and Bernhard (2007), who reported that when the wavelength of macrotexture got longer, the tire-pavement noise level got bigger. Estimated Noise, dB(A) = 7.65Average Wavelength + 70.00, (4) R2 = 0.58, P-value = 0

Fig. 7. Correlation of MPD and Tire-Pavement Noise

Depth and wavelength of surface texture significantly affect the tire-pavement noise as explained in Fig. 7 and 8. Based on these relationships, a multiple regression analysis on tire-pavement noise level, MPD, and average wavelength is developed and presented in Eq. (5). Fig. 9 presents the comparative results from measured and estimated tire-pavement noise based on Eq. (5). As can be seen, it gave a similar result to single regression. A Linear relationship between MPD and average wavelength was observed for asphalt concrete pavement. This result indicated that tire-pavement noise is linearly proportional to MPD and average wavelength. Estimated Noise, db(A) = 2.67MPD + 2.88Average Wavelength, (5) R2 = 0.61, P-value = 0

4. Relationship between MPD and Average Wavelength

Fig. 8. Correlation of Average Wavelength and Tire-Pavement Noise

Relationship between MPD and average wavelength is illustrated in Fig. 10. MPD increased with the increase of average wavelength. Therefore, the relationship between MPD and average wavelength of asphalt pavement can be expressed as shown in Eq. (6). As shown in Fig. 10, high relationship between MPD and average wavelength is observed with 0.87 determination coefficient and 0.0 of p-value. −4−

KSCE Journal of Civil Engineering

Tire-Pavement Noise Prediction using Asphalt Pavement Texture

were from 3.64 to 4.09 mm with DGA and from 3.80 to 4.24 mm range with SMA. According to the analysis result of the relationship between MPD and tire-pavement noise of asphalt pavement, it was found that tire-pavement noise increases as the MPD and wavelength of asphalt pavement increase. In addition, multiple regression analysis on tire-pavement noise level, MPD, and average wavelength gave a similar result to single regression due to the high correlation between MPD and average wavelength. This study also showed that aggregate gradation used in asphalt mixture significantly affect the MPD and average wavelength. Therefore, tire-pavement noise would be affected by aggregation gradation used in asphalt mixture.

References Fig. 10. Correlation of MPD and Average Wavelength

Average Wavelength = 0.37MPD + 3.4, R2 = 0.87, P-value = 0

(6)

Figure 10 also illustrated the effects of aggregate gradation used in asphalt material on values of MPD and average wavelength. MPD and average wavelength are significantly affected by asphalt pavement aggregate gradation. As shown, SMA material gives a high both values of MPD and average wavelength while the DGA gives a low amounts for these parameters. Therefore, tirepavement noise can be affected by asphalt pavement aggregate gradation. However, it could not be confirmed that either MPD or average wavelength is the most influencing factor inducing tire-pavement noise. Also, it could not be indicated that whether or not surface texture shape is the influence factor causing the tire-pavement noise.

5. Conclusions In this study, the statistical analysis on the relationship between MPD, average wavelength and tire-pavement noise of asphalt concrete pavement was performed. The measured MPD data were from 0.69 to 1.50 mm for asphalt concrete pavement with DGA (dense graded asphalt) and from 1.26 to 2.17 mm with SMA (Stone Mastic Asphalt). The measured average wavelength

Vol. 00, No. 0 / 000 0000

Budras, J. (2001). “A synopsis on the current equipment used for measuring pavement smoothness.” Federal Highway Administration, U.S Department of Transportation Centre for Transportation Research and Education (2006). “Evaluation of U.S. and european concrete pavement noise reduction methods.” National Concrete Pavement Technology Center, Iowa State University, Ames, IA, pp. 46-50. Hanson, D. I., Jame, R., and NeSmith, C. (2004). “Tire/Pavement Noise Study.” NCAT Report 04-02, National Center for Asphalt Technology, Auburn University, Alabama, United State, pp. 12-20. Hyun, T. J., Hong, S. J., Kim, H. B., and Lee, S. W. (2013). “Estimation of Tire-Pavement noise for asphalt pavement by using mean profile depth.” Journal of the Korean Society of Civil Engineers, Vol. 33, No. 4, pp. 1631-1638, DOI: 10.12652/Ksce.2013.33.4.1631. International Organization for Standardization (1997). Measurement of the Influence of Road Surfaces on Traffic Noise-Part 2, CloseProximity Method ISO Standard 11819-2. Leeuwen, H., Kok, A., and Reubsaet, J. (2007). “The Uncertainty of Acoustical Measurements on Road Surfaces using the CPXMethod.” INTER-NOISE 2007, Istanbul, Turkey. Pouliot, N., Carter, A., and Langlois, P. (2006). “Close-proximity measurement of tire-pavement noise on the ministry of transportation of Quebec’s Road Network.” Annual Conference of the Transportation Association of Canada. Rasmussen, R. O. and Bernhard, R. J. (2007). “The little book of quieter pavements.” FHWA–IF–08-004, Federal Highway Administration, U.S Department of Transportation.

−5−