National Academies Press: OpenBook

Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis (2017)

Chapter: Chapter 3 - Findings and Applications

« Previous: Chapter 2 - Research Approach
Page 17
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 17
Page 18
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 18
Page 19
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 19
Page 20
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 20
Page 21
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 21
Page 22
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 22
Page 23
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 23
Page 24
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 24
Page 25
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 25
Page 26
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 26
Page 27
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 27
Page 28
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 28
Page 29
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 29
Page 30
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 30
Page 31
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 31
Page 32
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 32
Page 33
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 33
Page 34
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 34
Page 35
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 35
Page 36
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 36
Page 37
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 37
Page 38
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 38
Page 39
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 39
Page 40
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 40
Page 41
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 41
Page 42
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 42
Page 43
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 43
Page 44
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 44
Page 45
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 45
Page 46
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 46
Page 47
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 47
Page 48
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 48
Page 49
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 49
Page 50
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 50
Page 51
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 51
Page 52
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 52
Page 53
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 53
Page 54
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 54
Page 55
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 55
Page 56
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 56
Page 57
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 57
Page 58
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 58
Page 59
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 59
Page 60
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 60
Page 61
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 61
Page 62
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 62
Page 63
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 63
Page 64
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 64
Page 65
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 65
Page 66
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 66
Page 67
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 67
Page 68
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 68
Page 69
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 69
Page 70
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 70
Page 71
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 71
Page 72
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 72
Page 73
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 73
Page 74
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 74
Page 75
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 75
Page 76
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 76
Page 77
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2017. Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis. Washington, DC: The National Academies Press. doi: 10.17226/24704.
×
Page 77

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

17 Findings and Applications The measured data was reduced and analyzed during post-processing. (Table 2-1 provides nomenclature and summary information on the 20 test sites, which are used as designations in this chapter.) At each site, the researchers took OBSI,19 SIP,20 and acoustic beamforming measurements. The measurement results are reported in this chapter. Summary of OBSI Results Figure 3-1 shows the overall A-weighted OBSI levels for all 20 of the Northern California and North Carolina (NC) measurement locations. Given the speed limits, Sites NC8 and NC14 were measured at 35 mph instead of the standard speed of 60 mph. For higher speed locations, the levels ranged from 100.4 to 106.6 A-weighted decibels (dBA). For the asphalt pavements of the North Carolina measurements, the lowest levels were measured at the locations with smaller size aggregate (i.e., Sites NC4, NC5, and NC9). Several pavements around Elkin were measured during the North Carolina OBSI Rodeo in September 2010.22 The ranges in level were similar to those reported in Figure 3-1; however, measurements were not made on the same sections as the Rodeo. The 2010 asphalt concrete (AC) pavements included two types—S9.5 and NovaChip. The portland cement concrete (PCC) pavements had transverse tine texture. The S9.5 and NovaChip pavements ranged from 98.7 to 103.4 dBA compared with 100.7 to 104.1 dBA for the AC pave- ments in 2014. The PCC pavements ranged from 103.1 to 105.2 dBA in 2010 compared with 105.7 to 106.5 dBA for 2014. Compared with the Northern California results, the ground PCC for NC10 and NC11 at 103.7 and 104.5 dBA, respectively, were slightly higher than the 103.6 dBA measured along the ground PCC pavement at 505 NB1. As a point of reference, research con- ducted for NCHRP Project 10-76 indicated that the FHWA TNM Average Pavement appeared to correspond to an OBSI level of about 102.5 to 103 dBA. Appendix B (available in NCHRP Web- Only Document 225) discusses the one-third octave band spectra for each of the sites. Summary of Statistically Isolated Pass-By Results Under the SIP procedure, the maximum pass-by levels for a specific site are plotted against vehicle speed, fitted with a logarithmic regression, and compared to the SIP reference curve for a given vehicle type. The reference curve used in this project was based on the REMELs data- base, but it was not the REMELs curve defined in the FHWA report. The reference curve used the regression developed for the REMELs curves; however, the constant powertrain engine noise term was omitted, and the correction for converting between level-mean and energy-mean was not included. For comparison with individual vehicle pass-by noise levels, the level-mean is more appropriate. C H A P T E R 3

18 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis SIP Measurement Results Under the SIP procedure, the results are reported as a SIP Index (SIPI). This was done by determining the difference in level between the reference curve and the regression of the data points at the average speed of the SIP data for each site (as noted on each figure). The SIPI pro- vides a way to compare the performance of each pavement at different speeds normalized by the REMELs based on the reference curve. The pass-by results can also be compared by considering the regressions of the data for each site. In Table 3-1, the SIPI values are presented, along with the designated (site dataset average) vehicle speed, the pass-by level from the regression of the speed/SIP data at that speed, the SIP reference level at the designated speed, and level from the regression curves at 55, 65, or 35 mph, depending on the site. Excluding the slow speed site, NC8, the highest SIPI values were calculated at the Lakeville site (aged DGAC) in Northern California and at NC1 and NC2 (both with transverse tine PCC) in North Carolina. These were also the sites with the highest OBSI levels (see Figure 3-1). The lowest SIPI values were calculated at 505 SB1 and NC9, which were among the lowest OBSI lev- els of the 60 mph data. Figure 3-2 shows the SIPI values compared with the OBSI levels for all higher speed sites (50 mph and above). This data produces a low coefficient of determination (R2) value and a small slope in the linear regression line. However, this data contains pass-bys that are not only on different pavements but under different operating conditions for uphill and downhill cases. In Figure 3-3, the uphill cases are removed from the data set, and the resulting linear regression has a much higher R2 value and an increased slope. Appendix C (available in NCHRP Web-Only Document 225) provides the individual datasets for the heavy truck pass-by levels along the 20 pavements, along with further discussion and com- parison of different sites and the regression curves grouped by the sites in Northern California, Figure 3-1. Overall on-board sound intensity levels for pavements at all test sites.

Findings and Applications 19 Site Designated (Average) Speed, mph Level from Regression, dBA SIP Reference Level, dBA SIPI Level at 55mph, dBA Level at 65mph, dBA Level at 35mph, dBA Lakeville 51.4 90.3 87.8 2.5 90.7 505 SB1 57.3 86.0 89.6 -3.6 85.5 505 SB2 56.8 88.8 89.2 -0.4 88.1 505 NB1 58.4 90.5 89.6 0.9 88.6 NC1 65.9 93.2 91.6 1.6 92.9 NC2 66.8 94.0 91.8 2.2 93.5 NC3 63.4 91.9 91.0 0.9 91.9 NC4 67.7 90.9 92.0 -1.1 90.0 NC5 62.6 91.7 90.8 0.9 92.1 NC6 58.4 91.4 89.7 1.7 92.6 NC7 64.0 91.3 91.1 0.2 91.6 NC8 29.0 84.4 78.7 5.7 85.8 NC9 61.6 90.7 90.5 0.2 91.1 NC10 67.8 91.8 92.0 -0.2 90.9 NC11 63.4 91.9 91.0 0.9 92.3 NC12 63.8 90.2 91.1 -0.9 91.0 NC13 61.3 89.7 90.4 -0.7 90.0 NC14 38.0 82.5 83.0 -0.5 82.0 NC15 60.7 89.9 90.2 -0.3 90.4 NC16 53.2 88.9 88.2 0.7 92.5 Table 3-1. Results from SIP analysis and average pass-by levels at average speeds and 55, 65, or 35 mph. Figure 3-2. SIPI versus overall OBSI level for all higher speed measurement sites.

20 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis the uphill sites in North Carolina, the downhill sites in North Carolina, the flat sites in North Carolina, and the slower speed sites in North Carolina. Comparison With the REMELs Database One of the objectives of this research was to begin to populate a new heavy truck noise source database for potential use in future traffic noise models. One issue was whether or not the truck noise sources have changed in the 22 years since the original REMELs data collection. The SIP results from this research were, therefore, compared with the REMELs data. Thus, 6.3 dB was added to the REMELs data acquired 50ft (15m) from the centerline of vehicle travel lane to account for the difference in the 25ft (7.6m) measurement distance used in this research. This propagation difference is based on NCHRP Project 1-44,23 in which a median difference of 6.3 dB between SIP data at 25 and 50ft (7.6 to 15m) was documented for 11 measurement sites for microphone heights of 5ft (1.5m).23 Using the 6.3 dB offset, the SIP data from this research for all of the downhill and level test sites were plotted with the REMELs average cruise level for heavy trucks in Figure 3-4. The regression through the Project 25-45 data lies below that of the REMELs database by about 1 dB. The slope of the Project 25-45 trend line virtually parallels the REMELs cruise curve. From Project 1-44, the standard deviation of the 25 to 50ft (7.6 to 15m) offset is 0.7 dB, which is slightly less than the offset shown in Figure 3-4. This may indicate a slight decrease in heavy truck cruise noise in the past 22 years. The scatter about the Project 25-45 trend line of about 10 dB is similar to that of the REMELs database. Figure 3-5 presents a comparison of the uphill and the slower speed SIP data with the heavy truck interrupted/grade average. For this operating condition, the Project 25-45 results are below Figure 3-3. SIPI versus overall OBSI for all higher speed measurement sites with uphill sites removed.

Findings and Applications 21 the REMELs average, with differences of about 1.5 dB above 50 mph. At the slower speeds, the REMELs curve and the Project 25-45 trend line diverge considerably, with differences greater than 7 dB below 25 mph. This suggests that powertrain noise sources may have improved since the 1994 measurements. This would be consistent with the stricter, slow speed 80 dBA pass-by rules from the U.S. Environmental Protection Agency. Although these rules took effect in 1980, the reductions may not have made it into all truck fleets by the time of the REMELs data collection in 1994. Taken with the results shown in Figure 3-4, it appears that tire noise at higher speed cruise may not be significantly lower than the REMELs database; however, engine noise at slower speeds Figure 3-4. Comparison of SIP data from 11 downhill and level test sites with the REMELs heavy truck cruise average level. Figure 3-5. Comparison of SIP data from eight uphill and slow speed/accelerating sites with the REMELs heavy truck interrupted flow and grade average level.

22 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis has been reduced. In Figure 3-6, regression for the five higher speed, uphill sites is virtually identi- cal to the REMELs curve for heavy truck cruise noise. Compared with Figure 3-4, this suggests that the uphill grade adds about 1 dB to the truck noise levels at these higher speeds. Appendix C provides comparisons of results for individual sites with the REMELs data. Noise Source Mapping The beamforming data were acquired in 10-second intervals for each pass-by event. To deter- mine if the target vehicle selected during each run would be adequate for analysis, the 6 dB cri- terion for vehicle selection, defined in the AASHTO TP-98 SIP standard method,20 was applied to evaluate all vehicle data. According to the SIP procedure, the A-weighted sound pressure level (SPL) should be at least 6 dB below the target vehicle maximum sound level just prior to and just after the passage of the target vehicle. This criterion was applied to each run at the Northern California and North Carolina test sites. Table 3-2 summarizes the total vehicle counts of the accepted runs used for the noise source mapping analysis at each test site. Evaluation of Typical Noise Contours The contour figures for all Northern California and North Carolina test sites were constructed for each pass-by event. The grid size used for each run was -23 to +42.7ft (-7 to +13m) about the center of the array, with the 42.7ft (13m) being to the right of the grid. The grid spacing in the horizontal direction was 0.7ft (0.2m), while the vertical spacing was 0.3ft (0.1m). This grid size was found to be adequate in all cases for the 20 measurement sites. The contours captured at each site showed dominant noise concentrations at the ground, with a few special case contours having traces of noise at elevations of 13.1ft (4m) and above. Examples of the contours produced for a typical heavy truck pass-by (Run 38) on the Lake- ville site are shown in Figures 3-7 through 3-9 in individual one-third octave bands and overall A-weighted level. In these figures, the photograph of the truck is properly scaled and aligned with the grid. The yellow boxes included with the truck photograph are 8ft-by-8ft (2.44m-by-2.44m), Figure 3-6. Comparison of SIP data from five uphill highway sites with the REMELs heavy truck interrupted flow and grade average level.

Findings and Applications 23 Test Site Testing Date, Time Heavy-Duty Vehicles Medium-Duty Vehicles Light-Duty Vehicles Buses Lakeville 12/17/2013, 12:20-15:30 44 9 0 0 505 SB1 12/20/2013, 12:20-14:10 57 0 2 0 505 SB2 12/21/2013, 9:30-13:20 49 8 2 1 505 NB1 12/20/2013, 15:50-17:1012/30/2013, 13:10-15:00 24 49 2 1 0 9 0 0 NC1 9/18/2014, 14:00-16:00 69 0 9 0 NC2 9/19/2014, 10:00-12:30 61 1 0 0 NC3 9/19/2014, 15:30-17:00 66 0 7 0 NC4 9/22/2014, 10:30-12:00 73 4 0 0 NC5 9/22/2014, 14:30-16:00 76 0 0 0 NC6 9/23/2014, 10:00-12:00 71 5 13 3 NC7 9/23/2014, 14:00-16:00 72 3 9 1 NC8 9/24/2014, 10:00-13:00 58 8 12 0 NC9 9/24/2014, 14:30-15:009/25/2014, 14:30-16:00 80 0 13 0 NC10 9/26/2014, 10:00-11:30 66 5 7 0 NC11 9/26/2014, 13:00-15:00 67 4 8 0 NC12 9/29/2014, 10:00-12:00 65 4 10 0 NC13 9/29/2014, 13:30-14:309/30/2014, 10:00-11:00 66 1 0 0 NC14 9/30/2014, 12:30-17:30 46 15 0 4 NC15 10/1/2014, 11:00-13:00 62 4 0 0 NC16 10/1/2014, 14:00-17:0010/2/2014, 9:30-12:00 68 9 0 2 Table 3-2. Total vehicle counts per test site used for the noise source mapping analyses. as determined from a photograph of a pole placed at the right wheel path of the lane of travel for the vehicle under test. The scale of the contours is indicated in Figure 3-7, where the upper pale yellow color represents the maximum level for any one plot. The colors correspond to 1 dB increments. The level of the maximum color is given on each contour plot. The approximate location of the pavement ground plane is shown with each contour, and sound reflected by the pavement is indicated below the ground plane. Given that the array software operates in metric units only, the tick marks on the left side of the contour plot indicate height above the ground in meters; however, these would correspond to approximately 3.3ft (1m) increments, as well. The red dot visible in all contour plots has no significance to the data presented or to its interpretation. Most of the tractor/trailer pass-by events produced contours similar to those of Figures 3-7 through 3-9. The contours of the overall level (Figure 3-7) indicate primary noise regions that are low and split by the ground plane. The regions are concentrated at the drive axles of the tractor and at the axles of the trailer, indicating that tire noise is the dominant noise source. The contours for the tractor are also stretched toward the very front of the truck, indicating tire noise for the front axle and/or powertrain noise coming beneath the truck. Some distortion of the contours to the right in the plots results from the large angle between the axis of the array and the measurement plane at +42.7ft (+13m). At the lower frequency bands shown in Figure 3-7, the noise sources are extended in the vertical direction to heights up to 13ft (4m); however, the cause for these low-frequency distortions cannot be determined from the plots. This low-frequency content does not appear to be affecting the measured overall contour plot. At the higher frequencies shown in Figures 3-8 and 3-9, powertrain noise is further indicated, given that the source region extends the entire length of the tractor. A muffler is apparent directly below the cab and in front of the gas tank. There are no elevated source regions. For this truck, the maximum overall A-weighted level was 81 dBA, with the 800 Hz one-third octave band producing the highest level of 76 dBA. Typical of all the data, the higher frequency bands (i.e., 2,000 Hz and above) produced the lowest sound levels. In interpreting the contours, (text continues on page 26)

24 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Figure 3-7. Sound level contours for a typical truck at the Lakeville site–overall A-weighted and one-third octave bands from 315 to 630 Hz.

Findings and Applications 25 Figure 3-8. Sound level contours for a typical truck at the Lakeville site–one-third octave bands from 800 to 2,000 Hz.

26 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis at lower frequencies it is difficult to localize the source given the size of the acoustic wavelengths involved. Figure 3-10 shows the 315 Hz contours for the loudspeaker along with the contours for the truck. The loudspeaker is about 4in-by-4in (0.1m-by-0.1m) square; however, the contour for the highest sound level is 3.9ft (1.2m) in diameter, which is only slightly more than an acoustic wavelength at this frequency. This further indicates the distortion that occurs at the lower frequencies. For higher frequencies, beamforming provides better spatial resolu- tion, as shown in Figures 3-11 (for 800 Hz) and 3-12 (for 1,600 Hz). The acoustic contours for all of the vehicle pass-bys were reviewed to identify trends and special cases, particularly for indications of exhaust outlet noise. Figure 3-13 shows an example of a truck with exhaust noise for another truck (Run 55) at the Lakeville measurement site. The case for the truck of Run 55 is quite different than that for Run 38. As shown in the contours for Run 55, exhaust system noise has a more profound influence on the contours. For overall A-weighted contours, the level at the top of the stack at a height of 13.8ft (4.2m) above ground is only 6 dB lower than the maximum level of 79 dBA at the pavement surface. Traces of exhaust noise can be found in nearly every frequency band of the spectra, including the 500- and 1,000-Hz bands, which are also shown in Figure 3-13. In the 2,500- to 4,000-Hz band contours, the exhaust outlet produces levels equal to the maximum levels produced by sources at pavement level, and at the higher frequency bands where other exhaust noise dominated the tire noise, the exhaust noise and tire noise measured in Run 55 were equally dominant noise sources. Out of a total of 44 heavy-duty vehicles measured at the Figure 3-9. Sound level contours for a typical truck at the Lakeville site–one-third octave bands from 2,500 to 4,000 Hz.

Figure 3-10. Comparison of truck and loudspeaker contours for the 315 Hz one-third octave band. Figure 3-11. Comparison of truck and loudspeaker contours for the 800 Hz one-third octave band.

28 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Lakeville test site, only 9 (20.5% of the total) were identified as having some exhaust system noise; however, only in Run 55 were the levels high enough to be within 10 dB of the maximum A-weighted level. Analysis of Overall A-Weighted Profiles The primary purpose of the beamforming measurements was to determine the height of the truck noise sources. To determine the vertical distribution of vehicle noise source levels, contours were exported to Microsoft Excel in a two-dimensional array of level versus x-y location. The levels were converted to energy and summed in the horizontal rows for the entire length of the contour figure, yielding vertical distributions of the sound pressure levels for each vehicle tested. Vertical distributions were calculated as the overall levels and for each one-third octave band frequency ranging from 315 to 4,000 Hz. The profiles of overall A-weighted level are presented and discussed in this section, and, for each of the vertical profiles shown in this section, each heavy-duty run was normalized to the maximum profile level at each respective site. Northern California Sites The four test sites in Northern California had relatively flat grades and posted maximum heavy truck speed limits of 55 mph. From Table 3-2, the number of acceptable heavy vehicle runs collected at the Lakeville, 505 SB1, 505 SB2, and 505 NB1 totaled 44, 57, 40, and 73, respectively. Figure 3-12. Comparison of truck and loudspeaker contours for the 1,600 Hz one-third octave band.

Findings and Applications 29 Lakeville Site. Figure 3-14 shows the vertical distributions for the acceptable heavy truck runs measured at the Lakeville site. The highest overall levels for each run were at ground level. This indicates that the dominant noise source is tire noise, whether from the front or rear axles or a combination of both, as found for Runs 38 and 55 (Figures 3-7 through 3-9 and 3-13). Other noise sources that would be found below 9.8ft (3m) include engine noise (which could radiate from under the front of the vehicle), from the front wheel well, and possibly fan noise radiating from the front grille of the truck. The profiles for the two trucks shown in Figures 3-7 through 3-9 (Run 38) and 3-13 (Run 55) are identified in Figure 3-14, in order to compare the contours. The typical profile in Figure 3-14 for the Lakeville site reduces from the maximum profile level by 10 dBA at heights ranging from 5.3 to 8ft (1.6 to 2.4m). In some cases, the profiles start to reduce like the typical truck runs and show a jump in height at a particular noise level. Run 2, identified in Figure 3-14, “bumps-up” at 88 dBA. This may indicate a small contribution of exhaust outlet noise (see Figure D1 in Appendix D for the contours for this case). Appendix D (available in NCHRP Web-Only Document 225) provides analysis of additional truck contours. Two profiles in Figure 3-14 produced levels at 12ft (3.7m) within 10 dBA of the ground-level maximum level. For Run 55, this was due to exhaust noise. For Run 17, the cause was high levels of low-frequency noise. Figure 3-13. Overall, 500, and 1,000 Hz contours of truck pass-by Run 55 at Lakeville.

30 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis 505 SB1 Site. Figure 3-15 shows the vertical distributions measured at 505 SB1. From Table 3-2, a total of 57 heavy-duty vehicles were measured at this site. Typical heavy-duty vehicles measured along 505 SB1 reduced from the maximum noise level by 10 dBA at heights ranging from 7.5 to 11.2ft (2.3 to 3.4m). One truck (Run 54) produced levels at 12ft (3.7m) within 10 dBA of the ground-level maximum. This resulted from exhaust stack noise. Although the maximum overall levels at the pavement surface were lower at 505 SB1 than at the Lakeville test site, the relative noise profile levels at various heights above the pavement sur- face were mostly higher at 505 SB1. Part of the reason for this could be the lower tire/pavement noise at 505 SB1. With the quieter pavement, the tire noise would be reduced, allowing other noise sources (e.g., engine noise and, exhaust noise) to have greater influence on the overall noise profile. Also, at the Lakeville site, the average speed for heavy-duty vehicles was approximately 51 mph, while at the 505 SB1 site the average speed was 58 mph. Even though tire/pavement noise would be higher at this higher speed, engine RPM might also be higher. 505 SB2 Site. Figure 3-16 shows profiles for the overall noise levels measured at the second site along I-505 southbound, 505 SB2. At heights ranging from 5.6 to 9.8ft (1.7 to 3m) above the pavement, overall noise levels for typical heavy-duty vehicles reduced by 10 dBA from the maxi- mum level at the pavement surface. The lower to the ground the noise level was concentrated, the greater the indication that the source of the measured noise levels was dominated by tire noise. The performance of the pavement at 505 SB2 should have been similar to that of TNM Average pavement, based on the SIPI value of Table 3-1 and the measured OBSI levels that were within 1 dB of that average pavement.24 Although the profiles of Figure 3-16 were slightly elevated from those of the Lakeville site, the contours were generally similar. As with the Lakeville results, contour plots for 505 SB2 (see Appendix D) indicated tire noise from trailer tires as being a significant source in almost all plots Figure 3-14. Overall A-weighted levels for all heavy-duty vehicles measured at Lakeville normalized to 95.9 dBA.

Findings and Applications 31 Figure 3-15. Overall A-weighted levels for all heavy-duty vehicles at Site 505 SB1 normalized to 93.1 dBA. Figure 3-16. Overall A-weighted levels for all heavy-duty vehicles at 505 SB2 normalized to 95.8 dBA.

32 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis for trucks with trailers. In general, the contours corresponding to the typical truck profiles of Figure 3-16 were similar to the contours from typical heavy-duty vehicles at the Lakeville site, (shown in Figures 3-7 through 3-9). For the 505 SB2 site, most of the runs also included engine noise. More specifically, out of the 49 heavy-duty vehicles, approximately 45 indicated the pres- ence of engine noise, which is approximately 91.8%. At 12ft (3.7m), three trucks displayed levels within 10 dBA of the ground-level maximum. Two of these (Runs 96 and 128) were found to have exhaust stack noise, and the third (Run 91) was due to low frequency content. 505 NB1 Site. The final site in Northern California was along I-505 northbound. The researchers conducted testing on this site on December 21 and December 30, 2013. As sum- marized in Table 3-2, a total of 73 heavy trucks was measured on these days. Compared to the average maximum profile levels for the heavy-duty vehicles at the other Northern Cali- fornia sites, the average overall maximum profile level at 505 NB1 was approximately 1 dB higher than the Lakeville data, 3.8 dB higher than 505 SB1, and 1.1 dB higher than 505 SB2. Figure 3-17 shows the heavy-duty vertical distributions measured at 505 NB1. The overall noise levels reduce by 10 dBA from the maximum at heights ranging from 5.2 to 8.9ft (1.6 to 2.7m). As with 505 SB2, all trucks with trailers displayed tire noise from both the drive axles and trailer. In addition to tire noise, runs from the 505 NB1 site contained some engine noise (two only had traces). As shown in Figure 3-17, two trucks had levels at 12ft (3.7m) within 10 dBA of the ground- level maximum. One truck (Run 65) had indicated exhaust stack noise, while the other (Run 11) had high levels of low frequency content. The one heavy truck with exhaust noise had levels with 5 dBA of the ground-level maximum at elevated heights. This was the only truck measured in Northern California that resulted in elevated noise sources with levels as high as tire noise. One truck (Run 53) had levels between 1 and 2ft (0.3 and 0.6m) that were greater than the ground- level maximum, which is indicative of engine noise. Figure 3-17. Overall A-weighted levels for all heavy-duty vehicles at Site 505 NB1 normalized to 96.9 dBA.

Findings and Applications 33 North Carolina Sites The profiles of the North Carolina sites are categorized as follows: (1) uphill, (2) downhill, (3) level, and (4) slower speed. For the sites in Categories 1, 2, and 3, the maximum level for the heavy truck runs was found at the ground level (except for 4 cases out of 894). This indi- cates that the dominant noise source is tire noise, whether from the front or rear axles or a combination of both. Uphill Sites. For most of the trucks tested at the uphill sites, noise sources concentrated low to the ground were the only contributors to the profiles. A few runs at each of the uphill sites had high noise levels at heights above 11.5ft (3.5m). In some cases, the noise contribution remained relatively high from the ground to 14.5ft (4.4m) or above. At NC1 (see Figure 3-18), the profile levels for three runs (69, 74, and 58) remained within 10 dBA of the maximum profile level for the entire array window. The contours for these three runs were similar to the contours of Run 55 at the Lakeville site (see Figure 3-13) and to the contours of Run 128 at 505 SB2 (see Figure D13 in Appendix D). Two trucks (Runs 20 and 68) reduced below 80 dBA with increasing height, and, given a secondary noise source at an elevated height, the profile increased to levels above 80 dBA at heights above 11ft (3.4m). All five of the trucks with profile levels greater than 87 dBA above 12ft (3.7m) at NC1 indicated the presence of exhaust noise. For an additional truck (Run 12) at NC1, the levels reduced by 10 dBA from the ground-level maximum at a height of 10.2ft (3.1m). After a further examination of the one-third octave band contours and profiles for this truck, it was determined that, in addition to being influenced by some low frequency content (discussed above for Run 38 at Lakeville), traces of exhaust noise, especially in the higher frequency bands, was present. It was possible at these higher frequen- cies that the noise was aerodynamic and produced by “jetting” of the exhaust gases through the restriction of the outlet. Out of 69 trucks at NC1, 4 (68, 58, 74, and 69) had exhaust noise Figure 3-18. Overall A-weighted levels for all heavy-duty vehicles measured at NC1 (uphill) normalized to 101.1 dBA.

34 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis ranging from 3 to 9 dB lower than the maximum at ground level (5.8% of the total). Two other trucks had some indication of the exhaust. At NC5 (Figure 3-19), five trucks indicated some exhaust noise. Three trucks (Runs 8, 19, and 23) had levels within 10 dB of the ground-level maximum at elevated heights. Out of 76 total acceptable pass-bys, this equated to 3.9%. In one case (Run 8), the level above 13ft (4m) was actually 3 dB greater than the ground-level maximum. Additionally, one (Run 15) of the five trucks with exhaust had some low frequency content. Another truck with exhaust noise (Run 52) had a pronounced 87.4 dBA exhaust noise “bubble” at a height of 12ft (3.7m) with levels below 81 dBA between 6.5 and 11ft (2 and 3.4m). Three additional heavy trucks, which reduced by 10 dB from the ground-level maximum at heights above 9ft (2.7m), also had low frequency content apparent in the overall profiles. In Figure 3-20, which shows the heavy truck runs measured at NC6, eight trucks had levels greater than 83 dBA above 13ft (4m). Two of these profiles (Runs 15 and 23) were dominated by low frequency content and did not show any trace of noise from the exhaust outlet. The remaining six (Runs 74, 42, 6, 61, 7, and 12) resulted from exhaust noise. All six of these runs produced levels ranging from 5 to 10 dBA below the ground-level maximum, corresponding to 8.5% out of 71 acceptable heavy truck runs. At Site NC9 (see Figure 3-21), three heavy trucks resulted in levels within 12 dBA at heights above 12ft (3.7m), and one truck had a secondary noise concentration between 11 and 12.5ft (3.4 and 3.8m). After examining the one-third octave band profiles and contours, it was deter- mined that only three of these trucks indicated exhaust noise, while the other was influenced by noise in the lower frequency bands. Out of 80 trucks measured at Site NC9, only one (1.3%) had exhaust noise within 10 dB of the ground-level maximum. Five trucks at Site NC11 (Figure 3-22) had high levels above 13ft (4m), and in each case, these high levels were due to exhaust noise (Runs 8, 34, 17, 42, and 72). These range from being 2 dB greater than the ground-level maxi- mum to 7 dB below. This corresponds to 7.5% of the 67 trucks tested at Site NC11. Three Figure 3-19. Overall A-weighted levels for all heavy-duty vehicles measured at NC5 (uphill) normalized to 98.6 dBA.

Findings and Applications 35 Figure 3-20. Overall A-weighted levels for all heavy-duty vehicles measured at NC6 (uphill) normalized to 99.3 dBA. Figure 3-21. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC9 (uphill) normalized to 98.3 dBA.

36 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis additional trucks showed excessive low frequency content. Of all the uphill test sites, Site NC13 (Figure 3-23) had the most number of heavy trucks with high levels at heights above 11.5ft (3.5m). Three trucks (Runs 66, 40, and 58) resulted in elevated noise levels within 10 dBA of the ground-level maximum, and three more trucks (Runs 22, 63, and 57) resulted in elevated noise levels within 13 dBA of the maximum. All six of these resulted from exhaust noise. Five additional trucks indicated some noise within 15 dBA of the ground-level maxi- mum at heights above 11ft (3.4m), and for three of these trucks, exhaust noise was present. Of the 66 total heavy trucks measured at Site NC13, 4.5% had levels within 10 dB of the ground-level maximum. In addition to trucks with exhaust noise, two trucks indicated some low frequency content. Downhill Sites. Figures 3-24 through 3-28 show each heavy truck run measured at the downhill sites (NC2, NC4, NC7, NC10, and NC12, respectively). Most of the heavy trucks mea- sured at the downhill sites had reduced by 10 dBA from the ground-level maximum at heights of 10ft (3m) or below, with fewer trucks having higher noise levels at elevated heights than those measured at the uphill sites. At Site NC2 (Figure 3-24), all trucks reduced by 10 dBA at heights at or below 10ft (3m). Two trucks had low frequency content, while no exhaust noise was observed at any trucks at Site NC2. One truck at Site NC4 (Figure 3-25) had traces of exhaust noise, and another truck indicated low frequency content. Out of 73 trucks measured at Site NC4, none had exhaust noise within 10 dB of the ground-level maximum. Of all the downhill test sites, Site NC7 (Figure 3-26) had the most trucks with noise at heights above 11.5ft (3.5m). Three heavy trucks (Runs 20, 13, and 68) at this site had noise levels within 10 dBA of the ground-level maximum at heights above 12ft (3.7m). Out of a total heavy truck count of 72, this corresponds to 4.2%. Four trucks (Runs 21, 75, 4, and 44) had noise levels within 10 dBA of the ground-level maximum at heights above 11.5ft (3.5m). Figure 3-22. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC11 (uphill) normalized to 99.2 dBA. (text continues on page 40)

Findings and Applications 37 Figure 3-23. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC13 (uphill) normalized to 97.0 dBA. Figure 3-24. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC2 (downhill) normalized to 101.9 dBA.

38 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Figure 3-25. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC4 (downhill) normalized to 97.9 dBA. Figure 3-26. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC7 (downhill) normalized to 99.0 dBA.

Findings and Applications 39 Figure 3-27. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC10 (downhill) normalized to 99.2 dBA. Figure 3-28. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC12 (downhill) normalized to 97.9 dBA.

40 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Of these seven trucks, six had noise radiating from the exhaust outlets. The seventh truck (Run 21) was identified with elevated noise levels, as well as one other truck at NC7, that con- tained low frequency content. One truck at Site NC10 (Figure 3-27) resulted in noise levels within 10 dBA of the ground-level maximum (Run 40). In fact, the exhaust noise for this truck was about 1 dB greater than the ground-level maximum, similar to the results of the two uphill sites, Sites NC5 and NC11. A second truck at Site NC10 (Run 4) had low frequency content. Sixty-six heavy trucks were measured at Site NC10, and only one (1.5%) had exhaust noise. At Site NC12 (Figure 3-28), three trucks (Runs 8, 66, and 37) out of the 65 measured had noise levels within 11 dBA of the ground-level maximum at heights above 11.5ft (3.5m), and from the contours, it was determined that all three trucks had exhaust noise. Of these, only one (Run 37) had exhaust noise levels within 10 dB of the ground maximum, which is 1.5% of the heavy trucks measured at Site NC12. Four additional trucks (Runs 60, 69, 56, and 47) at this site had low frequency content. Not one heavy truck at Sites NC2, NC4, or NC7 had noise levels above 90 dBA at heights above 13ft (4m), while Sites NC10 and NC12 had one truck each with levels at a height of 13ft (4m) or above. In contrast, five out of the six uphill sites had multiple heavy trucks with noise levels above 90 dBA at heights above 13ft (4m). The Northern California sites were medium-speed sites with level grades, compared to the higher speeds measured on uphill and downhill grades in North Carolina; however, one heavy truck had levels exceeding 90 dBA at 13ft (4m) or above at the Northern California sites (505 NB1). Exhaust noise was within 10 dB of the ground-level maximum for 1.3% to 8.5% of the heavy trucks measured at the uphill sites, while 0% to 4.2% was found at the downhill sites. At the Northern California sites, exhaust noise was within 10 dB of the maximum for 1.8% to 6.1% of the heavy trucks. Level Sites. The heavy truck runs shown in Figure 3-29 for Site NC3 had consistent verti- cal noise profiles from truck to truck, with all trucks reducing by 10 dBA from the ground-level Figure 3-29. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC3 (level) normalized to 99.6 dBA.

Findings and Applications 41 maximum at heights at or below 9ft (2.7m). No trucks at this site had exhaust noise or low frequency content. All noise sources were relatively close to the ground. Compared to the uphill (NC1) and downhill (NC2) sites with transverse tine PCC, only Site NC1 had heavy trucks with exhaust noise within 10 dB of the ground-level maximum (approximately 5.8%), while Sites NC2 and NC3 did not have any trucks with exhaust noise. Furthermore, Site NC3 did not have any truck runs with low frequency content, while Sites NC1 and NC2 had one and two trucks, respectively, with low frequency content. Compared to NC3 where all trucks reduced by 10 dBA from the ground-level maximum at heights of 9ft (2.7m) or below, NC1 and NC2 reduced by this amount at heights at or below 8.9ft (2.7m) and at or below 9.1ft (2.8m), respectively. All but three heavy trucks at Site NC15 (Figure 3-30) reduced by 12 dBA from the ground- level maximum at heights of 10.2ft (3.1m) or below. Of the three trucks in Figure 3-30 with elevated noise levels, two had exhaust noise (Runs 27 and 63), and the other (Run 21) had low frequency content. Out of 62 trucks tested at Site NC15, only one (1.6%) had exhaust noise within 10 dB of the ground-level maximum (Run 27). Site NC15 had a DGAC pavement similar to Sites NC13 and NC12, which were uphill and downhill, respectively. While 1.6% of the total number of heavy trucks measured at Site NC15 had exhaust noise within 10 dB of the maxi- mum profile level, 4.5% at Site NC13 and 1.5% at Site NC12 had exhaust noise within 10 dB of the maximum. The uphill DGAC Site NC13 had approximately 2.9% to 3% more exhaust noise than its corresponding downhill (NC12) and flat (NC15) sites. Additionally, one truck at Site NC15 had low frequency content, while Sites NC13 and NC12 had two and four trucks, respectively. Slower Speed Sites. Figures 3-31 through 3-33 show the heavy truck profiles for the slower speed sites in North Carolina: Sites NC8, NC14, and NC16, respectively. There were more heavy trucks with traces of elevated noise at these sites than with the uphill, downhill, or flat test sites Figure 3-30. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC15 (level) normalized to 97.1 dBA.

42 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Figure 3-31. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC8 (slower speed) normalized to 91.8 dBA. Figure 3-32. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC14 (slower speed) normalized to 89.1 dBA.

Findings and Applications 43 in North Carolina. More engine and exhaust noise would be expected for heavy trucks when accelerating. At the I-77 on-ramp (Site NC8), most of the heavy trucks reduced by 10 dBA from the ground-level maximum at heights below 10ft (3m). Approximately 17 trucks indicated elevated noise levels above 10ft (3m). With further examination of the one-third octave band profiles and contours, it was determined that 16 out of the total 58 trucks measured at Site NC8 had exhaust noise, while two additional trucks had low frequency content. Out of the trucks with exhaust noise, 11 of the 58 total (19%) produced levels within 10 dB of the ground-level maximum, and two (Runs 56 and 50) had levels greater than ground-level at heights of 10ft (3m) and greater. This site was compared to Site NC1 because both sites had transverse tine PCC on an upgrade. With a 37-mph difference in average measured speeds and a 9.3 dB difference in maximum pro- file level, only one or two trucks at both sites had low frequency content. However, the number of trucks with elevated noise levels was greater at Site NC8. At Site NC1, 4 out of 69 trucks—5.8%— had exhaust noise. At Site NC8, the percentage increased to 19%. This directly pertained to the driving scenario of the heavy truck operator. The trucks measured at Site NC8 were actively accelerating to full speed, but with the long on-ramp at Site NC8, the speed for individual trucks (see Figure C24 in Appendix C) varied by a factor of two, indicating different acceleration rates and potentially different engine loading. All vehicles at this site showed varying degrees of engine noise. This is apparent in the two profiles (Runs 24 and 9) where the levels around 1.6ft (0.5m) bulge to be greater than the ground-level maximum. At Site NC1, the roadway was headed uphill, and many of the vehicles along I-77 were close to the speed limit, which would suggest that some trucks were not fully into the throttle. This might explain why only 5.8% were observed to have exhaust noise. The average speeds measured at Site NC14 were approximately 9.6 mph faster than at Site NC8. Unlike Site NC8, most of the 46 trucks measured at Site NC14 were at a cruising speed when Figure 3-33. Overall A-weighted levels for all heavy-duty vehicles measured at Site NC16 (slower speed) normalized to 93.7 dBA.

44 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis passing the measurement location, with a few accelerating from stop at the traffic light. For most of the trucks measured at this site, the noise levels reduced by 10 dBA from the ground-level maximum at heights at or below 10.5ft (3.2m), but after examination of the one-third octave band profiles and contours for all heavy trucks, it was determined that 11 had at least some trace of exhaust noise. Of these, six (Runs 46, 60, 36, 41, 27, and 71) had exhaust noise levels within 10 dB of the ground-level maximum. This is 13% of the 46 total. While consisting of DGAC pavement on a slight upgrade similar to Site NC13, Site NC14 had average speeds approximately 23.2 mph slower and slightly more truck-to-truck variation than at Site NC13. At Site NC13, there were 4.5% of the heavy trucks with exhaust noise within 10 dB of the maximum profile level, which was 8.5% fewer than at Site NC14. Similar to Site NC8, two of the trucks (Runs 25 and 40) at Site NC14 indicate some engine noise as the levels at 1.6ft (0.5m) are again higher than those of the ground-level maximum. Of the three sites with slower posted speed limits, Site NC16 had the highest average speed measurements of 53.4 mph. At this site, the pavement was level, but just prior to the location of the test site, the grade of the roadway was slightly uphill. Therefore, it could be expected that some trucks would be in some degree of throttle. As shown in Figure 3-33, most of the trucks at Site NC16 reduced by 10 dBA from the ground-level maximum at heights of 9.5ft (2.9m) or below. Sixteen trucks indicated elevated noise levels above 11ft (3.4m). The one-third octave band profiles and contours for these 16 trucks were reviewed, and 11 heavy trucks were found to have exhaust noise. Nine trucks out of 68 (13.2%) were found to have noise levels within 10 dB of the ground-level maximum. Of these, eight (11.8%) were due to exhaust noise (Runs 18, 31, 63, 79, 6, 68, 66, and 13), and one (Run 18) was due to high levels of low frequency noise. In comparison with Site NC15, which was the same type of pavement on a level grade but with a speed limit 10 mph faster than Site NC16, the results varied. As discussed above, the aver- age maximum profile level at Site NC16 was only 3.4 dBA lower, and the average vehicle speeds were only 7.3 mph slower. However, the presence of exhaust noise in the measured trucks was significantly higher at Site NC16. With 11.8% of the 68 trucks at Site NC16 with exhaust noise, only 1.6% had exhaust noise at Site NC15—a difference of approximately 10.2%. Further, only one truck at Site NC15 had low frequency content, while five showed it at Site NC16. The differ- ence in exhaust noise may be due to the throttle position of the trucks as they pass through each site. Given that an upgrade existed just prior to Site NC16, it is probable that more trucks were still into the throttle at the time of passing the array than at Site NC15. The Lakeville and NC16 sites are both relatively flat two-lane highways with vehicles operating under cruise conditions. The average speed measured at Lakeville was 51.4 mph, and the average speed at NC16 was only 2 mph greater. In comparing the vertical distributions, 9.1% of the total heavy truck count at Lakeville resulted in exhaust noise, with only 4.5% having exhaust noise within 10 dB of the maximum profile level. Although 7.1% more overall heavy trucks showed some indication of exhaust noise (16.2%) at Site NC16, there were 11.8% with levels within 10 dB of the maximum level at the ground, which is greater than the Lakeville site by 7.3%. This may be due to the grade just prior to the NC16 site. Summary of A-Weighted Profile Analysis On average, the highest overall levels for each run at each site were at the ground level, as shown by the above overall profile levels. These would be associated with tire noise, and for this report, these maximum overall levels are referred to as maximum profile levels. Due to the summation process, these levels are not equal to pass-by levels. Table 3-3 summarizes the range of maximum profile levels, average maximum profile levels, and the average vehicle speed mea- sured at each test site and grouped by pavement grade. Comparing the individual downhill sites

Findings and Applications 45 with uphill sites of the same/similar pavement, it was determined that the difference in the aver- age maximum profile levels was less than 1 dBA in each case, and the average speeds measured at the downhill sites were faster than the corresponding uphill sites, which may be expected due to the grade. Similar to the OBSI results, the sites with the highest average maximum profile levels were Sites NC1 and NC2, which were the uphill and downhill sites, respectively, with transverse tine PCC. The average maximum profile level at Site NC2 was approximately 0.8 dBA higher than at Site NC1. Although NC3 had transverse tine PCC and the third highest average maxi- mum profile level, it was approximately 1.5 to 2.3 dBA less than Sites NC1 and NC2. Given that the grade of NC3 is relatively level, less engine noise or brake noise may have been applied at the site, resulting in less noise near the pavement surface. The ground PCC pavement sites, Sites NC10 (downhill) and NC11 (uphill), resulted in the same average maximum profile level. Two uphill test sites (Sites NC5 and NC9) and one down- hill test site (Site NC4) consisted of fine aggregate DGAC. At these sites, the uphill sites were 0.4 to 0.7 dBA higher than the downhill sites. Similar with the larger aggregate DGAC, the average maximum profile level at the uphill site (Site NC6) was approximately 0.3 dBA higher than the downhill site (Site NC7). The largest difference was found between Sites NC12 and NC13, which both had DGAC. In this case, the average maximum profile level at the uphill Site NC13 was lower by approximately 0.9 dBA than the downhill Site NC12. In addition to varying vehicle speeds affecting the beamforming measurements, operational differences would also affect the average maximum profile levels. Sites 505 SB1 and NC9 were among the higher speed sites and resulted in the lowest OBSI levels. Although 505 SB1 resulted in the third lowest average maximum profile level of 93.1 dBA, the profile level at NC9 was 98.3 dBA. At 505 SB1, the grade of the pavement was relatively level, and the average speed was 57.5 mph, but at NC9, the grade was a slight upgrade of +1.1% with an average speed of 61.5 mph. This may indicate that other factors, such as engine noise, could be increasing the maximum profile levels at the ground level. Test Site Range of Maximum Profile Levels Average Maximum Profile Levels Average Vehicle Speeds N or th er n Ca lif or ni a sit es Lakeville 92.1-98.3 dBA 95.9 dBA 51.4 mph 505 SB1 89.5-96.4 dBA 93.1 dBA 57.5 mph 505 SB2 92.4-98.9 dBA 95.8 dBA 56.8 mph 505 NB1 90.2-103.7 dBA 96.9 dBA 58.4 mph U ph ill si te s NC1 96.1-104.6 dBA 101.1 dBA 65.7 mph NC5 94.4-105.2 dBA 98.6 dBA 62.6 mph NC6 94.0-105.4 dBA 99.3 dBA 58.5 mph NC9 92.5-105.5 dBA 98.3 dBA 61.5 mph NC11 93.3-105.1 dBA 99.2 dBA 63.7 mph NC13 93.1-102.3 dBA 97.0 dBA 61.5 mph D ow nh ill sit es NC2 97.0-106.2 dBA 101.9 dBA 66.0 mph NC4 93.7-103.6 dBA 97.9 dBA 67.5 mph NC7 95.4-102.8 dBA 99.0 dBA 63.9 mph NC10 94.7-104.8 dBA 99.2 dBA 68.0 mph NC12 93.5-103.8 dBA 97.9 dBA 64.1 mph Le ve l sit es NC3 95.9-102.7 dBA 99.6 dBA 63.9 mph NC15 93.0-102.8 dBA 97.1 dBA 60.7 mph Sl ow er sit es NC8 85.6-101.1 dBA 91.8 dBA 28.7 mph NC14 75.6-94.2 dBA 89.1 dBA 38.3 mph NC16 87.9-98.0 dBA 93.7 dBA 53.4 mph Table 3-3. Summary of maximum profile levels and vehicle speeds for heavy trucks at all measurement sites.

46 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis The lowest average maximum profile levels shown in Table 3-3 were measured at Sites NC14 (located along SR 211, downstream from a stoplight in a 45 mph zone) and NC8 (located at the southbound I-77 entrance ramp). While all heavy trucks were accelerating at Site NC8, the trucks at Site NC14 consisted of a mixture of accelerating and cruising. At both sites, however, the average speeds were measured to be less than 40 mph. Figure 3-34 shows the correlation between aver- age maximum profile levels and average speeds. The coefficient of determination is 0.7, which is significantly better than with OBSI levels. Test Site Summary All the runs shown above for the vertical distributions of noise sources were averaged for each site. These averages included the runs containing engine and exhaust noise, as well as overall levels influenced by low frequency content. Figure 3-35 shows the average overall noise levels for the heavy-duty vehicles measured at each site. The average profiles were normalized to the average maximum profile level calculated at each specific site and color-coded by category— Northern California sites in red, uphill sites in North Carolina in blue, downhill sites in North Carolina in green, flat sites in North Carolina in purple, and slower speed sites in North Carolina in orange. As mentioned, the slower speed sites in North Carolina and the medium speed sites in Northern California resulted in the lowest average maximum profile levels. To compare the vertical characteristics of the profiles, the averages shown in Figure 3-35 were normalized to the same maximum profile level of 97.1 dBA, which is the average of all the sites, and the normalized profiles are shown in Figure 3-36. The curvatures for each of the profiles were similar. However, the North Carolina site curva- ture profiles seemed to decrease from the maximum profile level slightly faster than those for the Northern California sites. At 5ft (1.5m), the profile levels at the North Carolina sites had decreased by approximately 4.6 to 6.8 dBA from the maximum at 0ft (0m). At the Northern California sites, the profile levels had decreased by 4 to 5 dBA at this same height. The profiles from North Carolina and Northern California really diverged by 6.5ft (2m). In North Carolina, the pro- files decreased by 9.8 to 25.8 dBA from the ground-level maximum by 6.5ft (2m), while levels decreased in Northern California by 6.6 to 12.6 dBA at this height. At heights above 6.5ft (2m), Figure 3-34. Average maximum profile level versus average measured vehicle speed for all measurement sites.

Findings and Applications 47 Figure 3-35. Average overall A-weighted levels for all heavy-duty vehicles. Figure 3-36. Average overall A-weighted levels for all heavy-duty vehicles normalized to 97.1 dB.

48 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis the levels for all the profiles dropped quickly. At heights of 10ft (3m) and above, the average profile levels for each site were effectively zero. Of the 20 sites tested in Phases I and II, 2 were slow-speed sites (posted speed limits were 45 mph), 5 were medium-speed sites (posted speed limits were 55 mph), and 13 were high-speed sites (posted speed limits were 65 mph and up). While the average profiles shown in Figures 3-35 and 3-36 do not indicate elevated sources above 9ft (2.7m), a few heavy trucks were found to have some noise in the vicinity of the exhaust stack outlet at heights around 11ft (3.4m) and above. The occurrence and source strength of elevated noise sources was a primary concern of this research. Table 3-4, which summarizes the number of heavy trucks at each site that showed some indica- tion of elevated noise sources, demonstrates that 8.8% of the 1,289 trucks had some measurable noise at heights commonly taken to be the top of the exhaust stack. As mentioned, the current assumption that TNMv2.5 uses for vertical noise distribution is that 50% of the total noise is generated at the ground level and 50% at an elevated exhaust source, which implies an equal magnitude of noise from both heights. Although Table 3-4 shows that only 8.8% of all heavy trucks tested resulted in traces of exhaust noise, even fewer are comparable in magnitude to the noise source at ground level. Figure 3-37 shows the number of trucks at each test site that resulted in elevated noise levels at 12ft (3.7m) within 10 dBA of the profile level at the ground (shown in black), noise levels at 12ft (3.7m) within 5 dBA of the profile level at the ground (shown in red), and noise levels at 12ft (3.7m) equal to or greater than the profile level at the ground (shown in blue). At Sites NC2 and NC3, there were no trucks with traces of a noise source at 12ft (3.7m), and at Site NC4, the only truck with an elevated source resulted in levels at 12ft (3.7m) that were more than 10 dBA below the profile level at 0ft (0m). Trucks with noise levels at 12ft (3.7m) within 10 dBA of the profile level at 0ft were measured at Sites NC6, NC7, NC12, Lakeville, 505 SB1, and 505 SB2, but at each of these sites, the noise levels at 12ft (3.7m) were not within 5 dBA of the ground-level source. Trucks with 12ft (3.7m) noise levels that were within 5 dBA of the ground-level source but not equal to the ground source were measured at Sites NC1, NC13, NC15, 505 NB1, and NC14. Only six trucks out of a total count of 1,289, which were measured at NC5, NC10, NC11, NC16, and NC8, resulted in noise levels at 12ft (3.7m) that were equal to or greater than noise levels at 0ft. This comes to 0.5% of all heavy trucks, not 50%. Test site Heavy truck count No. with exhaust noise % with exhaust noise Lakeville 44 4 9.1% 505 SB 1 57 7 12.3% 505 SB 2 49 9 18.4% 505 NB 1 73 9 12.3% NC1 69 6 8.7% NC2 61 0 0% NC3 66 0 0% NC4 73 1 1.4% NC5 76 5 6.6% NC6 71 6 8.5% NC7 72 6 8.3% NC8 58 16 27.6% NC9 80 3 3.8% NC10 66 1 1.5% NC11 67 5 7.5% NC12 65 3 4.6% NC13 66 9 13.6% NC14 46 11 23.9% NC15 62 2 3.2% NC16 68 11 16.2% Total No. 1,289 114 8.8% Table 3-4. Summary of heavy trucks with elevated exhaust noise source regions.

Findings and Applications 49 For the trucks with noise levels at 12ft that were within 10 dB of the ground-level maximum, not all were due to exhaust noise. Of the 63 cases in Figure 3-37, only 56 (4.3%) have identifi- able exhaust noise. The others are due to high levels of frequency noise that extend up to 12ft. Comparison to Previous Research Compared with the Caltrans truck source mapping research,2 where tire noise was found to be the dominate source, approximately 1% to 2% of the 125 heavy trucks measured had noise sources from the exhaust stack outlet. Engine noise was also identified at the higher frequen- cies to be forward on the tractor in many cases. Figure 3-38 shows the Caltrans and average profiles for all 20 sites normalized to the maximum profile level of the Caltrans research, which was 95 dBA. The curvature of the Caltrans profile was similar to those of the uphill pavements, which were typically on the lower range of profile levels for all 20 test sites, for the first 2ft (0.6m). Starting at 2ft (0.6m), the Caltrans profile started to decrease at a slower rate. At 5ft (1.5m), the Caltrans profile decreased by approximately 4.2 dBA, while each Northern California site had decreased by 4 to 5 dBA and each North Carolina site had decreased by 4.6 to 6.8 dBA. At a height of 6.5ft (2m), the profiles measured at the Northern California and North Carolina sites had decreased by 6.6 to 12.6 dBA and by 9.8 to 25.8 dBA, respectively, while the Caltrans profile had decreased by 6.2 dBA. Above this height, the profile for the Caltrans average tended to flatten with increasing height, while the profiles from this research dropped quickly in level to essentially zero. This was due in part to better performance of the newer array and better data processing to exclude background noise in the profile calculation as discussed in previous sec- tions of this report. Comparing the current results with those of NCHRP Report 635,1 similar trends were found as those of the Caltrans comparison. The conclusions of NCHRP Project 08-56 were that tire noise was the dominant noise source, followed by powertrain noise and some exhaust stack Figure 3-37. Heavy truck counts at each site with noise source strength at 12ft (3.7m) within 10 dBA of the ground-level source.

50 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis noise in limited cases. In the Project 08-56 research, trailer tires were often identified as noise sources (similar to the current research). The average of the profiles for heavy trucks from the Project 08-56 research is shown along with the current results and those from the Caltrans work in Figure 3-39. All profiles shown in the figure were normalized to the profile level of the NCHRP Project 08-56 research at 0ft. Like the Caltrans profile, the Project 08-56 profile flattens in level for heights above about 6.5ft (2m). It is suspected that this reflects the same reasons as in the Caltrans study: array performance and data processing methods. Between 2 and 5.2ft (0.6 and 1.6m), the Project 08-56 profile is roughly parallel to the uphill sites in North Carolina. Below 2ft (0.6m), the Project 08-56 profile decreases with decreasing height down to the level of the pavement. Analysis of One-Third Octave Band Profiles As discussed above for the overall noise profiles, the vertical distribution of vehicle noise source levels for each one-third octave band frequency ranging from 315 to 4,000 Hz was deter- mined by summing the horizontal energy at each height for the entire length of the contour figure. This resulted in vertical distributions of the sound pressure levels for each frequency band. These one-third octave band profiles were developed for each vehicle tested in Northern California and North Carolina. Need for Frequency Weighting At lower frequencies, the “hot spot” size is determined by the acoustic wavelength and the ability of the array to resolve the noise source. Although the array used in the current research is superior in resolution to those of the earlier studies1,2 as discussed previously, the size of the hot Figure 3-38. Comparison of average overall A-weighted profiles for heavy trucks at all test sites normalized to the average profile from the Caltrans research.

Findings and Applications 51 spot at lower frequencies is enlarged, getting larger with decreasing frequency. This occurs even if the actual source is small and acoustically compact, as illustrated by the contours for the 400 and 1,000 Hz frequency bands shown in Figure 3-40, where the actual noise source is the 4in-by- 4in (0.1m-by-0.1m) loudspeaker that was used for site calibration. Given that summing across the horizontal rows adds more apparent source strength due to the (lack) of resolution of the array, the summation that defines the vertical profile level in the lower frequencies will be exag- gerated, relative to the higher frequencies. As a result, the profiles for different frequency bands cannot be directly compared in terms of their contribution to the overall A-weighted profile. This effect is illustrated in Figure 3-41 for the loudspeaker located 3.3ft (1m) above the pave- ment, as measured at Site 505 SB2. In this figure, the profile for the overall level is shown, along with the energy summation of each of the one-third octave band levels. For overall levels less than 75 dBA, the summation of the bands is greater than the profile determined from the directly measured overall level. The purpose of this section is to determine a method for weighting the individual one-third octave band levels to better represent their contribution to the overall level profile. This is a necessary step in defining source level as a function of frequency for traffic noise modeling and barrier prediction. Method for Frequency Weighting The data in Figure 3-42 compares the level at the center of each one-third octave band “hot spot” determined from the loudspeaker contours to the maximum profile level measured at each frequency band profile shown in Figure 3-41. The overall A-weighted levels for each spec- trum are also shown. The results indicate that the levels determined by the profiles are con- sistently higher and that the difference decreases with increasing frequency. The maximum profile level exceeds the spot level by 13 dBA at 315 Hz and by 2.6 dBA at 4,000 Hz. This is due to the broadening of the beamformer at increasingly lower frequency, with the difference in Figure 3-39. Comparison of average overall A-weighted profiles for heavy trucks at all test sites normalized to the average profile from the NCHRP Project 08-56 research.

Figure 3-40. Contours for a small loudspeaker noise source on the pavement, as measured by the beamforming array for 400 and 1,000 Hz. Figure 3-41. Vertical profiles measured for a small loudspeaker 3.3ft (1m) above the pavement.

Findings and Applications 53 the overall being approximately 4.8 dBA. The calculated differences between the spot levels and the maximum profile level at each frequency band were used as adjustment factors at each corresponding frequency band profile shown in Figure 3-41. To match the overall A-weighted levels for the spot and profile, the difference between them, which was determined to be 4.8 dBA, was also subtracted from each one-third octave band profile. The following formula illustrates the adjustment factor calculations: (3-1)A SPL Spot Max Spot Maxi i i i overall overall[ ]( ) ( )= − − − − where i represents the frequency bands ranging from 315 to 4,000 Hz; SPLi represents the verti- cal profile for a given frequency band (i); Spoti represents the level measured at the center of the “hot spot” for a given frequency band (i); Maxi represents the maximum profile level at a given frequency band (i); Spotoverall represents the level measured at the center of the “hot spot” for an overall contour; and Maxoverall represents the maximum profile level of the overall profile. For the loudspeaker example shown in Figure 3-41, these adjustment factors range from 8.2 dBA at 315 Hz to -2.3 dBA at 4,000 Hz. When applied to the profiles in Figure 3-41 for each frequency band, the resultant profiles are those shown in Figure 3-43. Compared to Figure 3-41, the lower frequency band profiles in Figure 3-43 are shifted down in level. This results in the summation more closely following the measured overall A-weighted profile. While the summation profile in Figure 3-41 is practically equal to the overall measured profile level down to 75 dBA, the sum- mation in Figure 3-43 is practically equal down to 70 dBA. The same method was used to produce adjustment factors for the loudspeaker results mea- sured at the ground level of 10 additional sites in Northern California and North Carolina. In Figure 3-42. One-third octave band levels for a small loudspeaker 3.3ft (1m) above the pavement, determined by the maximum level of the contour (spot) and the maximum profile maximum level.

54 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis each case, the calculated adjustment factors were similar, with variation at any given frequency of slightly more than 1 dBA. Figure 3-44 shows the correction factors calculated for the 3.3ft (1m) high loudspeaker compared to the average correction factors calculated for all 10 ground- level loudspeaker results. Both the 3.3ft (1m) and ground-level data display essentially the same slope with frequency and have about a 1-dB offset. The average of the 3.3ft (1m) and ground- level results is also shown in Figure 3-44. Given that the noise sources for heavy trucks would not all occur at the ground level, the 3.3ft (1m) and ground-level average was the optimal choice for a correction factor. The uncorrected profiles for the typical truck (Run 38 at Lakeville) of Figures 3-7 through 3-9 are shown in Figure 3-45. The summation of the profiles is almost the same as the measure- ment for the overall A-weighted levels, down to a sound pressure level of about 88.5 dB. Just below that, the measured levels drop to about 4 dBA lower than the summation, indicating that the 315, 400, and 500 Hz profiles do not contribute to the measured overall level for vertical distances above about 6ft (1.8m). When the above adjustment factors developed based on the point source are applied to the profiles in Figure 3-45, there is an offset between the summation profile and the measured overall profile of approximately 2.7 dBA. To account for this offset, an additional, uniform adjustment of -2.7 dBA would be required at each frequency band. The resultant profiles and summation are shown in Figure 3-46. The corrections reduce the con- tribution of the 315, 400, and 500 Hz bands to the summation profile, reducing the levels for vertical distances above 6ft (1.8m). This more closely replicates the behavior of the measured overall, which drops to 0 dBA above the vertical distance of 6ft (1.8m). This maximum profile level method was performed on other heavy trucks measured at other test sites (see Appendix E for examples). These included a mix of Northern California and North Figure 3-43. Corrected vertical profiles measured for loudspeaker 3.3ft (1m) above the pavement.

Findings and Applications 55 Figure 3-44. One-third octave band frequency corrections from ground level & 3.3ft (1m) loudspeaker data, with average of both. Figure 3-45. Uncorrected vertical profiles measured for a typical truck (Run 38) at the Lakeville site.

56 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Carolina sites and a mix of trucks with relatively normal profiles; that is, where either the overall A-weighted profile did not extend above 8ft (2.4m) or where there was some profile content above 8ft (2.4m). Typically, the weighted summation profiles were similar to Figure 3-46, with discrepancies between the summation and the measured overall profiles increasing with source height. Above the drop off, the adjusted summations were reduced, compared to the unad- justed profile, better replicating the measured overall level profile. For heavy trucks with higher source content, the profiles generally follow the same trend as the lower source trucks; however, depending on the particular truck run, specific frequency band profiles may contribute more to the summation profile than others. However, there was no apparent consistency about which frequency bands had the most effect on the summation. It varied from truck-to-truck. As a result, arbitrarily raising specific frequency bands based on one or even a few examples would not necessarily produce closer agreement between the summation and overall profiles. In gen- eral, the adjustment factors should be considered to apply more in an average sense with some variance from truck-to-truck. An alternative approach comparing the spot level to profile energy sum was also considered, but due to a low frequency bias in the energy summations, it was determined that using the maximum profile levels was more advantageous. The energy summation method is discussed in Appendix E (available in NCHRP Web-Only Document 225). Average Adjusted Profiles for Each Frequency Band Each individual truck profile measured at all 20 test sites was averaged per one-third octave band. These uncorrected one-third octave band profile averages are shown in Figure 3-47. For Figure 3-46. Corrected vertical profiles measured for a typical truck (Run 38) at the Lakeville site, based on the point source with a -2.7 dB adjustment.

Findings and Applications 57 frequencies ranging from 2,500 to 4,000 Hz, the average profiles were adjusted based on ambi- ent levels determined at each site and for each frequency band. As shown in the high frequency contours of Run 38 at Lakeville (see Figure 3-9), there were typically “ghost” noise levels at the higher frequency bands that were not zeroed out when the horizontal energy was summed to develop the vertical profiles. With the inclusion of the ghost noise levels, the average profiles for 2,500; 3,150; and 4,000 Hz bands would reduce from the maximum level at the ground, similar to the average profiles in Figure 3-47, but instead of reducing to zero like the lower frequency profiles, these high frequency profiles would reduce to a constant noise level between 53 and 61 dBA (depending on the site). Once this noise floor was reached, the profiles would become vertical for the full height of the profile. These noise levels are not representative of truck noise sources and, therefore, were zeroed out, resulting in the average profiles shown in Figure 3-47 for frequency bands ranging from 2,500 to 4,000 Hz. Similar to the Run 38 example at Lakeville (see Figure 3-45 for the unadjusted profiles), the curvature of the summation is similar to the measured overall profile until a height of about 4ft (1.2m). The measured overall profile reduces to zero at a fairly constant rate starting at a height of 6ft (1.8m), while the summation profile shows excessive influence from the 315 and 400 Hz frequency bands. Additionally, there is a 0.8 dBA offset between the measured overall and the summation at the ground. When the above maximum profile adjustment factor is applied to the one-third octave band profiles of Figure 3-47, the resultant profiles are shown in Figure 3-48. The summation profile does more accurately replicate the behavior of the measured overall value in Figure 3-48. When the energy from the corrected summation profile and the overall measured profile are compared, the calculated energy lost from the profiles shown in Figure 3-48 was approximately 0.5 dBA. Figure 3-47. Uncorrected average vertical profiles measured at each test site.

58 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Source Height Distribution Analysis The first objective of this research was to develop heavy truck vertical source height distribu- tions that could be used in traffic noise modeling. For this purpose, the average profiles from Figure 3-48 were developed as complete vertical distributions of source heights for each one-third octave band. However, for ease of implementation into a noise model, equivalent distributions with several different discrete source heights were considered. Profiles were to be developed for two- and three-point source models and compared to the full distributions from the beamforming array, as given in Figure 3-48. To accomplish this, it was first necessary to establish the profile for a single point source at each frequency. Point Source Height Distributions As discussed in regard to Figure 3-40, the sound pressure contours varied with frequency even for a compact point source, as shown for the 400 and 1,000 Hz one-third octave band contours in that figure. The size of the contour enclosing the highest level (or “hot spot”) was found to decrease with increasing frequency. This effect is also mirrored in the profiles (see Figure 3-41 for the loudspeaker at a height 3.3ft [1m] above the pavement and Figure 3-49 for the loudspeaker at the ground level). To approximate a profile for the truck pass-bys, the point source model used at each height and frequency needed to reflect this behavior properly. To develop the point source model profile at each frequency band, the results from the loud- speaker site calibration measurements were used. These site calibration measurements were typically collected by placing the 4in-by-4in (0.1m-by-0.1m) loudspeaker on the pavement surface. Ground-level loudspeaker profiles averaged for 10 sites were found to be essentially the same, as shown in Figure 3-50 for the 630 Hz one-third octave band and in Figure 3-51 Figure 3-48. Corrected average vertical profiles measured at each test site.

Findings and Applications 59 Figure 3-49. One-third octave band and overall level profiles for a loudspeaker at ground level. Figure 3-50. 630 Hz profiles for ground-level loudspeaker at sites in North Carolina and Northern California.

60 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis for the 2,000 Hz band. Further, the profile for the loudspeaker measured 3.3ft (1m) above the pavement was found to be almost identical to the ground-level profile for the portion of the profile from 3.3ft (1m) and higher, as shown in Figure 3-49. This same behavior was found for the other frequency bands. For each frequency band, a point source profile was defined based on these data and the assumption that the distribution is symmetric about the maximum level of the profile. For Figure 3-49, this is not the case, as reflection from the pavement modifies the profile; however, for modeling such as that in TNM, the ground reflection is accounted for in the model and not in the source distribution. As a result, point source profiles, such as those of Figure 3-52 for 630 Hz, were developed and used for each band, and the profile is taken to be symmetric about the height where the point source is positioned. The point source profiles do not extend below the pavement surface when the source is located at some height above the pavement. The profiles for each frequency band are shown in Figure 3-53. Using these frequency-dependent point source distributions, combinations of point sources could then be produced by summing the individual profiles to produce a complete multi-source distribution to compare to the beam- forming array results. Models Using Two Point Sources In consideration of replicating the source distribution in TNM, vertical distributions for each one-third octave band would need to correspond to two point sources. These would preferably be at two of the three sources currently used by different vehicle types in TNM, which include ground level, 5ft (1.5m), and 12ft (3.7m). Based on the beamforming results, heights of 0 and 5ft (0 and 1.5m) would be the starting point of the two-point source distribution because cases where the source level at 12ft (3.7m) equaled that of ground level were 0.3% or less for all test sites measured in the 55 to 70 mph speed range. Considering Figures 3-50 and 3-51, it also appears that the upper source height would have to vary with frequency. The two-source height distribution Figure 3-51. 2,000 Hz profiles for ground-level loudspeaker at sites in North Carolina and Northern California.

Findings and Applications 61 Figure 3-52. Comparison of 630 Hz profiles for ground level and 3.3ft (1m) high loudspeaker at Site 505 SB2 in Northern California. Figure 3-53. Point source profiles developed for each one-third octave band.

62 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis in TNM for heavy trucks consists of the same two heights (0ft and 12ft [0m and 3.7m]) for all one-third octave bands, although the relative strength at the two heights is not necessarily the same for each frequency band. To develop the source height distributions for heavy trucks, the averages of the profiles from Figure 3-48 were used for each one-third octave band. As was observed for the profiles of the overall levels at each site, the average profiles at 630 Hz were similar for each site, even though the sites represented various pavements and operating conditions. This is illustrated in Figure 3-54, which presents the average profile for each site for the 630 Hz band along with the average of all sites at this frequency. For each profile shown in Figure 3-48, a two-point source height distribution was fitted to the curve by varying the height and amplitude of the sources. The sound pressures of the two sources were summed on an energy basis and compared to the measured average profile. Adjustments were made until the closest fit was obtained. One source was always assumed to be at ground level, representing tire-pavement noise, as in the current TNM source distributions. Initially, the goal was to use a 5ft (1.5m) source height, consistent with the current TNM distribution. However, as illustrated by Figure 3-55 for 630 Hz, this was not feasible for the point source lev- els and percentages noted in the plot legend. In this case, in order to conform to the measured profile up to a height of about 4.6ft (1.4m), the ground-level source had to be 3.6 dB greater than the 5ft (1.5m) source. Above 4.6ft (1.4m), the calculated profile diverged significantly from the measured values. After some trial, the best fit was produced with an upper source of 2.3ft (0.7m), as shown in Figure 3-56. The strength of the 2.3ft (0.7m) source was 1.5 dB less than the ground-level source. For simplicity, only one upper height would be used for all frequency bands if the current upper source height of 5ft (1.5m) was not viable. However, this was not possible either, as illus- trated in Figure 3-57 for 2,000 Hz and a 2.3ft (0.7m) upper source height. Ultimately, a second Figure 3-54. Average 630 Hz profiles from all measurement sites and the average of all sites.

Findings and Applications 63 Figure 3-55. Initial application of a ground level and 5ft (1.5m) high point sources to match the measured truck profile at 630 Hz. Figure 3-56. Final matching of measured 630 Hz profile to a ground level and 2.3ft (0.7m) high point sources.

64 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis source at 1.6ft (0.5m) was required to get the fit shown in Figure 3-58. This trial-and-error process was implemented at each frequency band, and it was determined that multiple second source heights would be required to fit the profiles for all of the frequencies. The minimum number of different heights was determined to be a total of four heights, in addition to the ground-level source height: 3.3ft (1m), 2.3ft (0.7m), 1.6ft (0.5m), and 1ft (0.3m). The source heights, levels, and split percentages are given in Table 3-5. Appendix F (available in NCHRP Web-Only Document 225) provides plots showing the final matching of the two-point source models to the measured profiles for all of the remaining one-third octave bands. Models Using More Than Two Point Sources Although the measured truck profiles could be fairly well replicated with two-point source models at each frequency, attempts were made to improve this by using three point sources. For the lower frequencies, these attempts placed an additional point source at 12ft (3.7m) to mir- ror what is done in TNM. An example of this is presented in Figure 3-59 for the 315 Hz band. Compared to the two-point source model (see Figure F1 in Appendix F), the three-point source model provides a better match of the measured profile up to a vertical height of almost 10ft (3m), while the two-point source model diverges from the measured profile starting at about 7.5ft (2.3m). However, above 10ft (3m), the three-point source model significantly overpredicts the measured profile. Due to the curvature of the point source profile, it is not possible to improve the two-point source match to the measured profile with added profiles greater than 3.3ft (1m). This was found in all of the other lower frequency bands. For the higher bands, such as 2,000 Hz (see Figure 3-58), the two sources are already so low to the ground that a third source cannot be added to improve the match to the measured profile. As a result of this analysis, it was concluded that the two-point source models defined by Table 3-5 are the optimal choice, aside from using the actual measured profiles. Figure 3-57. Initial application of a ground level and 2.3ft (0.7m) high point sources to match the measured truck profile at 2,000 Hz.

Findings and Applications 65 REMELs and TNM Source Height Distributions In the REMELs report,21 the issue of source height splits between ground level and 12ft (3.7m) for heavy trucks was addressed using the research performed at Florida Atlantic Uni- versity (FLAU) in the 1990s.25,26 The best fit to these results gave a sub-source height ratio (rht) given by 1 1 (3-2)log[ ]( ) ( )( )= + − − + ( )+r f L L M eht N f P Q Figure 3-58. Final matching of measured 2,000 Hz profile to a ground level and 1.6ft (0.5m) high point sources. 1/3 Octave Band Lower Source Upper Source Height Level Source Strength Height Level Source Strength 315 Hz 0ft (0m) 80.5 dBA 44% 3.3ft (1m) 81.5 dBA 56% 400 Hz 0ft (0m) 82.5 dBA 56% 3.3ft (1m) 81.6 dBA 44% 500 Hz 0ft (0m) 86.6 dBA 65% 3.3ft (1m) 84.0 dBA 35% 630 Hz 0ft (0m) 87.0 dBA 59% 2.3ft (0.7m) 85.5 dBA 41% 800 Hz 0ft (0m) 88.5 dBA 67% 2.3ft (0.7m) 85.5 dBA 33% 1,000 Hz 0ft (0m) 88.0 dBA 67% 2.3ft (0.7m) 85.0 dBA 33% 1,250 Hz 0ft (0m) 85.5 dBA 74% 2.3ft (0.7m) 81.0 dBA 26% 1,600 Hz 0ft (0m) 82.5 dBA 69% 1.6ft (0.5m) 79.0 dBA 31% 2,000 Hz 0ft (0m) 80.0 dBA 78% 1.6ft (0.5m) 74.5 dBA 22% 2,500 Hz 0ft (0m) 77.3 dBA 77% 1.6ft (0.5m) 72.0 dBA 23% 3,150 Hz 0ft (0m) 73.8 dBA 62% 1ft (0.3m) 71.7 dBA 38% 4,000 Hz 0ft (0m) 71.2 dBA 65% 1ft (0.3m) 68.5 dBA 35% Frequency Center Table 3-5. Heights for two-point source distribution with source strength and percentage of source split.

66 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis where L is the sub-source height ratio at low frequencies, M is an empirical constant such that 1-M is the sub-source height ratio at high frequency, and N, P, and Q control the transition between the low and high frequencies For heavy trucks at cruise, it was determined that L = 0.054276, M = 0.973749, N = -36.503587, P = 102.627995, and Q = -132.679357. It was also noted that this value of L put only 5.4% source energy at 12ft (3.7m) for low frequencies and 2.6% at high frequencies. From the database, a heavy truck cruising at 55 mph would have a source level at the ground of 83.9 dBA and at 12ft (3.7m) of 69.9 dBA, corresponding to the rht value of 0.0385. To account for the differences in propagation for the REMELs data made at 50ft (15m) over soft ground and the FLAU measure- ments made over hard ground, the rht(f) values were multiplied by Frequency Correction Factors (FCFs) for the frequency bands from 50 to 10,000 Hz. These propagation factors were specific to either heavy trucks or all other vehicles and were applied uniformly to the rht(f ) values. The resultant spectrum for a heavy truck at 55 mph is shown in Figure 3-60, both uncorrected and corrected. The FCFs have some influence on the source split, because the 12ft (3.7m) source is decreased to 68.7 dBA while the lower source remains at 83.9 dBA. When implemented in TNM 1.0 and 2.0,27,28 modifications were made to the source height distribution. The values for L, M, N, P, and Q for heavy trucks were all changed. The new value for L was 0.594848, which put 59% of the energy at the 12ft (3.7m) source height for lower frequencies. The value for M was set to 0.643317, putting 37% at 12ft (3.7m) for the higher frequencies. Without FCFs, the levels for a heavy truck at 55 mph became 79.1 dBA for the 12ft (3.7m) source and 82.4 dBA for the ground-level source. The FCFs were split into 12ft (3.7m) and 0ft (0m) values, again as a function of frequency. These values ended up with a level of 79.2 dBA at 12ft (3.7m) and 79.2 dBA at 0ft (0m). The resulting spectra for the upper, lower, and overall levels are shown in Figure 3-61. Figure 3-59. Three-point source model for the 315 Hz one-third octave band.

Findings and Applications 67 Figure 3-60. Heavy truck pass-by spectrum for 55 mph cruise with and without the Frequency Correction Factors from the REMELs Report. Figure 3-61. Heavy truck pass-by spectrum for 55 mph cruise for the upper and lower source levels and the overall pass-by levels, using coefficients from TNM 2.0.

68 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis For TNM 2.5,29 the values of for L, M, N, P, and Q were all changed again. The value of L was taken to be 0.8500, putting 85% of the energy at 12ft (3.7m) for lower frequencies. Additionally, M was changed to -0.33, increasing the energy at 12ft (3.7m) to 133% for the low frequencies and 85% for the higher frequencies. The FCFs for the 12ft (3.7m) and 0ft (0m) source heights were also changed for heavy trucks. With these changes the level for the 12ft (3.7m) source height was 76.8 dBA for a 55 mph heavy truck at cruise and 74.8 dBA for 0ft (0m). The resulting spectra for the upper, lower, and overall levels are shown in Figure 3-62. Source Distributions and Barrier Performance Although source distribution can be a factor for highway noise prediction, with respect to ground reflection in unobstructed sound propagation, it is critical for predicting highway barrier performance. To assess the effect of source height profiles, a standard model of sound propaga- tion over barriers was used to calculate insertion loss.30 In these calculations, ground reflection on either side of the barrier was not included and constant distances of 35ft (10.7m) from the truck to the barrier and 50ft (15m) from the receiver to the barrier were used. The receiver height was taken as 5ft (1.5m). Calculations were done for barrier heights of 10, 12, 14, and 16ft (3, 3.7, 4.3, and 4.9m). These calculations were done only to examine how different profiles and source distributions influenced the performance of the barrier. They were not intended to be a rigorous representation of actual highway cases, such as would be done in a fully developed traffic noise model. The spreadsheet calculation was done for a source at each discrete height produced by the beamforming array. The benchmark for comparison is that calculated for the average pro- files developed in this research for each frequency band, as shown in Figure 3-48. These profiles were attenuated using the barrier calculations for a 12ft (3.7m) barrier, producing the results shown in Figure 3-63. Comparing Figures 3-48 and 3-63, the shapes of the profiles are similar Figure 3-62. Heavy truck pass-by spectrum for 55 mph cruise for the upper and lower source levels and the overall pass-by levels, using coefficients from TNM 2.5.

Findings and Applications 69 with and without the barrier; however, the contributions of some the frequency bands change significantly. As expected, the relative levels of the bands from 1,000 to 4,000 Hz are reduced significantly. As a result, the lower frequency bands from 315 to 800 Hz become more significant, although the relative change between them is small. The 315 and 400 Hz bands contribute about the same as they do in the no barrier case; however, the 500 Hz band is the largest contributor to the overall at the lower heights with the barrier. To calculate the total reduction by a barrier, the profile for each one-third octave band was summed on an energy basis for each height increment from 0 to 15.4ft (0 to 4.7m), with and without the barrier. The summed levels for each band were then summed to produce overall levels, with and without the barrier, and noise reduction taken as the difference of these. The results of these calculations for the site average profiles of Figures 3-48 and 3-63 are shown in Figure 3-64. Similar analysis of the beamforming results has been performed for some of the individual sites and is presented in Appendix G (available in NCHRP Web-Only Document 225). This same type of barrier analysis was performed for various other source height distributions. These included distributions for the original REMELs, TNM 2.0, TNM 2.5, the optimized two- point source distributions produced in this research, the three-point source distributions, and a constant distribution, in which the sound level at each profile height was the same. The inser- tion loss (IL) for these cases is shown in Figure 3-65 for a 12ft (3.7m) high barrier. These results display two groupings. The two TNM and constant profile results are consistently lower than the IL calculated for the average profiles of this research. Below 630 Hz, the differences range from 2 to 6 dB, with the TNM 2.5 IL producing the lowest IL. At 630 Hz and above, the range is from 6 to 9 dB, with the largest differences occurring at the higher frequencies. The TNM 2.5 results hover around the constant profile, typically within 1 dB. For the higher IL grouping, the two- point source models developed from the NCHRP Project 25-45 average tend to give the highest IL values, particularly below 1,000 Hz where the difference between the NCHRP Project 25-45 Figure 3-63. Site averaged profiles, as modified by a 12ft (3.7m) barrier.

70 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Figure 3-64. Levels with and without a 12ft (3.7m) barrier, calculated for the site average profiles and resultant noise reduction. Figure 3-65. Insertion losses for a 12ft (3.7m) high barrier, calculated from different heavy truck noise source distributions.

Findings and Applications 71 average and the two-point source model ranges from about 1 to 2.5 dB. By adding the third source to the two-point source model at frequencies below 630 Hz, the three-point source model better replicates the NCHRP Project 25-45 average. In general, the original REMELs distribution compares well to the NCHRP Project 25-45 average, except in the 315 and 400 Hz bands, where the REMELs IL is 1 to 2 dB greater. The insertion loss for the source distributions of Figure 3-65 are shown in Figure 3-66 in terms of overall A-weighted level losses for barrier heights of 10, 12, 14, and 16ft (3, 3.7, 4.3, and 4.9m). Similar to the trends of Figure 3-65, the overall IL values are in two groupings, with the TNM and constant distributions producing lower values. In addition to lower values, this grouping also produces a different, higher rate of increase in IL with barrier height. For the two TNM distributions, the increase in IL from 10 to 16ft (3 to 4.9m) is about 9 dB, and for the constant distribution, it is about 8 dB. For the grouping of the higher values, REMELs, the NCHRP Project 25-45 average, and the two- and three-point source distributions, the increase from 10 to 16ft (3 to 4.9m) is about 5.5 dB. Figure 3-66 indicates that the TNM distributions are particularly biased to producing lower IL values for lower barrier heights than those based on the NCHRP Project 25-45 data and these barrier geometry scenarios. The barrier calculation was applied to profiles for individual trucks. Typically, they displayed IL behavior similar to the 25-45 average. The results from the Lakeville site shown in Figures 3-7 through 3-9 for Run 38 (LK 38) are an example. However, for those few cases where significant source strength was indicated at the upper profile heights, the calculated IL was more similar to the TNM IL values. Run 19 at the North Carolina Site NC5 (NC5 19) illustrates such a case. The profiles for NC5 19 are shown in Appendix G. The results of the barrier calculations for these two trucks are shown in Figure 3-67, along with the results from the two TNM distributions and the NCHRP Project 25-45 average distribution. This figure indicates that some trucks replicate Figure 3-66. Overall insertion losses for 10, 12, 14, and 16ft (3, 3.7, 4.3, 4.9m) high barriers calculated from different heavy truck noise source distributions.

72 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis the TNM distribution; however, the results of this research indicate that, with the current fleet of trucks in service, there are relatively few of these examples. Medium Truck Results The primary focus of this research was on the heavy vehicle noise source heights and, as a result, priority was placed on these measurements over other vehicle types. However, depend- ing on circumstance and practicality, some pass-bys of medium trucks, light vehicles, and buses were captured with the beamforming array. The numbers of these types of vehicles are shown in Table 3-2. In total, 83 medium trucks, 100 light vehicles, and 11 buses were measured with the array and the SIP acquisition system. The concentration of data for these vehicle types varied from site to site. For medium trucks, 52 of the total 83 pass-bys were captured on the sites with speed limits at or below 55 mph—medium trucks were more prevalent and/or easier to mea- sure on these sites. For light vehicles, the measured pass-bys were better distributed; however, for eight sites, no data was acquired. Given the lack of data on buses, these vehicles were given some priority in the field measurements; however, due to their scarcity, only 11 were captured. For light vehicles, some sites, particularly those with noisier pavement, had general background noise from the interstate that limited the usable data even further. Sufficient data for conducting statistical analysis on the results was not obtained at any site. This section presents the overall findings for medium trucks at the 20 test sites. For contour and profile analysis of the medium trucks, see Appendix H (available in NCHRP Web-Only Document 225). Data measured for light vehicles and buses are included in Appendix I. Discussion of Overall A-Weighted Profiles Medium trucks were measured at three of the Northern California sites, and in North Carolina at two of the uphill sites, all five of the downhill sites, one of the flat sites, and all three slower Figure 3-67. Comparison of TNM and 25-45 average source distributions to a typical truck (LK 38) and a truck with unusual upper source strength (NC5 19).

Findings and Applications 73 speed sites. Maximum profile level, average speed, and maximum height information is sum- marized in Table 3-6 for all medium trucks. Similar for the heavy trucks, the three slow speed sites in North Carolina resulted in the lowest maximum profile levels, ranging from 87.2 dBA at Site NC14 to 91 dBA at Site NC8. The highest maximum profile level was measured at the uphill ground PCC Site NC11, which was approximately 98.1 dBA. On average, the highest maximum profile levels were found at the uphill site, which were calculated to be approximately 96.9 dBA. In comparison, the aver- age maximum profile levels for the downhill sites were approximately 0.8 dBA lower, while the maximum profile levels at the Northern California sites, the level grade sites in North Carolina, and the slower speed sites in North Carolina were approximately 3.4, 3.0, and 7.8 dBA lower, respectively. For the heavy trucks, the maximum profile levels averaged higher at the downhill sites, with the uphill sites being only 0.3 dBA lower. The difference between the average down- hill maximum profile levels and those calculated at the Northern California sites, the flat sites in North Carolina, and the slower speed sites in North Carolina was approximately 3.8, 0.8, and 7.6 dBA, respectively. There is a greater difference between the flat sites and the maximum sites for the medium trucks than for the heavy trucks; however, the difference with the North- ern California sites and the slower speed sites in North Carolina was approximately the same. Test Site Summary As shown in Table 3-6, all the medium truck runs for the vertical distributions of noise sources were averaged for each site. Figure 3-68 shows the average overall noise levels for the medium-duty vehicles, normalized to the average maximum profile level of 94.2 dBA. The average profiles are color-coded by category as follows: Northern California sites (red); uphill sites in North Carolina (blue); downhill sites in North Carolina (green); flat sites in North Carolina (purple); and slower speed sites in North Carolina (orange). Unlike the heavy truck averages shown in Figure 3-36, the curvatures for medium truck averages shown in Figure 3-68 show more variation. The heavy trucks were practically identical within 2ft (0.6m) of the ground, but the medium trucks varied by 1.6 dBA at 2ft (0.6m). By 4ft (1.2m) and above, the medium truck profiles for each Northern California site and for each uphill site in North Carolina seemed to decrease from the maximum profile level slower than the other sites. At 5ft (1.5m), one slower speed site in North Carolina had reduced from the maximum by over 20 dBA, while the other North Carolina sites had decreased by approximately 5.7 to 8 dBA at this height. The Northern California sites showed a reduction of 5 to 6.4 dBA at 5ft (1.5m). Starting approximately 14 dBA below the maximum, the medium truck averages at a few sites show a jump in profile. One of the sites in Northern California, for Test Site Number of Vehicles Range of Maximum Profile Levels Average Maximum Profile Levels Average Vehicle Speed Lakeville 9 90.9-95.8 dBA 93.3 dBA 53.5 mph 505 SB 2 8 91.1-96.5 dBA 93.3 dBA 59.8 mph 505 NB 1 3 91.2-96.5 dBA 93.8 dBA 57.3 mph NC2 1 -- 94.7 dBA 65.0 mph NC4 4 93.5-99.8 dBA 96.4 dBA 67.0 mph NC6 5 94.8-98.4 dBA 96.9 dBA 58.2 mph NC7 3 93.4-97.3 dBA 95.9 dBA 59.0 mph NC8 8 87.7-95.4 dBA 91.0 dBA 37.1 mph NC10 5 94.5-102.3 dBA 97.6 dBA 67.6 mph NC11 4 93.2-102.3 dBA 98.1 dBA 66.3 mph NC12 4 94.9-97.2 dBA 95.9 dBA 66.0 mph NC13 1 -- 95.7 dBA 66.0 mph NC14 15 81.2-93.8 dBA 87.2 dBA 37.0 mph NC15 4 90.2-96.6 dBA 93.9 dBA 64.5 mph NC16 9 86.9-91.2 dBA 89.0 dBA 52.4 mph Table 3-6. Summary of maximum profile levels and vehicle speeds for medium trucks at all test sites.

74 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis instance, jumps from 79.4 dBA at 6.6ft (2m) to 75.2 dBA at 9.2ft (2.8m). This indicates the pres- ence of elevated noise sources in the 6.5 to 9.5ft (2 to 2.9m) range. Due to the variation in medium truck profiles and the limited sample size measured at each site, conclusions cannot be made based on these results. More data needs to be collected for medium trucks. Various vehicles with different configurations are classified as medium trucks. Studying the medium trucks in subcategories would help to better understand the noise dis- tributions of each configuration and then the types could be compared. Additional research is required for medium trucks. Discussion of Average One-Third Octave Band Profiles Figure 3-69 shows the average unadjusted one-third octave band profile for all medium trucks tested. As discussed for the heavy trucks, the profiles for frequency bands ranging from 2,500 to 4,000 Hz were modified at heights above 6.2ft (1.9m), which is where each of these profiles stopped reducing from the maximum profile level and reached the noise floor. The noise floor was zeroed out so heights above 6.2ft (1.9m) were not contaminated by noise unrelated to the trucks. As was observed for the heavy trucks, the summation profile has a maximum level at the ground slightly lower than the measured overall profile. The curvature of the summation is simi- lar to the measured overall profile until a height of about 4ft (1.2m), which was also true for the heavy trucks. The measured overall profile reduces to zero at a fairly constant rate, starting at a height of 5.1ft (1.6m), while the summation profile shows excessive influence from the 315 and 400 Hz frequency bands. These lower frequency bands do not extend in the vertical direction as high as the lower frequency band profiles did for the heavy trucks. At the higher frequency bands (3,150 and 4,000 Hz), the profiles reduce from the maximum level at the ground at a slower rate than the heavy truck profiles at these frequency bands. In fact, there is only a noise Figure 3-68. Site average overall A-weighted levels for all medium-duty vehicles, normalized to 94.2 dBA.

Findings and Applications 75 level reduction of approximately 5 dBA from the medium truck maximum profile level to 6ft (1.8m) in the 4,000 Hz band. The profile maximum correction factors, developed above for point sources, were also applied to the unadjusted medium truck average profiles. When the point-source adjustment factors were applied, the summation profile was approximately 2.6 dBA lower than the measured over- all profile at the ground level. This additional adjustment was applied and the resultant profiles are shown in Figure 3-70. The adjusted summation profile shows more of an agreement with the measured overall profile. The energy lost by the corrected summation profile, when compared to the overall measured profile, was calculated to be approximately 0.6 dBA. For the heavy truck average profiles, two-point source distributions were developed. Given the limited amount of medium truck data, development of two-point source distributions was not attempted for these vehicles. However, comparing the profiles for the medium trucks (Fig- ure 3-70) to those for the heavy trucks (Figure 3-48), it is apparent that the source regions are closer to the ground for medium trucks. For the heavy trucks, it was found that with a two-point source model where the upper source was positioned at 5ft (1.5m), the contours for even the lowest frequency of 315 Hz could not be replicated and a lower height was required. Based on this comparison, it can be concluded that the upper source height for the lower frequency would be 5ft (1.5m) or less and equal to or lower than those of the heavy truck upper source heights for higher frequencies. Comparison to REMELs Database Similar to the heavy truck SIP results, the SIP values for medium trucks were compared to the REMELs database. These results are shown in Figure 3-71 for the 83 medium truck pass-by Figure 3-69. Uncorrected average one-third octave band vertical profiles for medium trucks.

76 Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Figure 3-70. Corrected average one-third octave band vertical profiles for medium trucks. Figure 3-71. Medium truck SIP data points compared to REMELs medium truck curves.

Findings and Applications 77 events, for which these data are available, compared to the cruise and interrupted flow REMELs curves. As discussed above, given the very low number of medium trucks, comparison to the REMELs curve is not as significant as was found in the similar comparison for heavy trucks. For Figure 3-71, the data for speeds at and below 45 mph are from the slower speed sites and include some trucks under acceleration. Given the small size of the medium truck database, the data cannot be meaningfully segregated into uphill, downhill, and level sites, as was done for the heavy trucks. Generally, the medium truck results are consistent with the REMELs curves; although, there is more scatter about the curves than was found for the heavy trucks. Eliminat- ing the uphill and slower speed sites, as shown in Figure 3-72, the comparison to the REMELs cruise curve contains as much scatter as it does in Figure 3-71. A logarithmic trend through these points defines a trend line that parallels the REMEL’s medium truck cruise curve, about 1.5 dB lower at the higher speeds; however, the coefficient of determination is only R2 = 0.16. Although it appears that the medium truck results from this research fall below the REMELs curves, there is too little data to be conclusive. Figure 3-72. Medium truck SIP downhill and level site data points only, compared to REMELs medium truck cruise curve.

Next: Chapter 4 - Conclusions and Suggested Research »
Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB's National Cooperative Highway Research Program (NCHRP) Research Report 842: Mapping Heavy Vehicle Noise Source Heights for Highway Noise Analysis provides an analysis to determine height distributions and spectral content for heavy vehicle noise sources. The report also explores establishing and beginning the development of an extended heavy vehicle (truck and bus) noise source database for incorporation into traffic noise models, including future versions of the U.S. Federal Highway Administration (FHWA) Transportation Noise Model (TNM) acoustical code.

Accompanying the report is Web-Only Document 225: Appendices to NCHRP Research Report 842.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!