Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 72
72
noise. Data from the Illinois Tollway showed similar noise, however, may be influenced by the conditions at the
results, with standard-depth (0.125 in. [3.2 mm]) tines time of testing (e.g., wind speed and direction, barometric
producing greater noise than shallow-depth (0.075 in. pressure, and other site characteristics).
[2 mm]) tines.
· Longitudinal Drag
Statistical Analyses
Drag type--Despite similar texture depths, a broom
drag in California produced greater near-field and inte- Texture Depth Measurement Procedure
rior noise than burlap drag.
During field testing, a difference was observed between
texture depth measured by the high-speed profiler and that
Relationship of Near-Field Noise measured by the CT Meter. This difference could be attrib-
with Interior Noise and Pass-By Noise uted to the difference in sampling and collecting the data,
Figure 5-32 is a plot of near-field SI noise and interior Leq transforming the raw data into the MPD statistic, and con-
noise for the 70 test sections (57 existing and 13 newly con- verting MPD into EMTD (high-speed profiler) or MTD
structed sections). The data shown are the average values of (CT Meter).
repeated runs made in the wheelpath and lane center posi- The high-speed profiler uses a laser to measure the eleva-
tions of each test section. A linear trend-line through the data tion profile (sampling/recording interval = 1 point every
shows a general relationship (R2 = 0.51) between the two 0.016 in. [0.4 mm]) along the length of a section at a distinct
noise sources and one that somewhat reflects the qualitative position within a lane (e.g., wheelpath or lane center). The CT
noise levels for the five ranges of SI. The interior Leq values in Meter uses a laser to measure the circular elevation profiles
Figure 5-32 are all within 3% of the values established previ- (radius = 11.2 in. [284 mm]) at individual spots selected along
ously in Table 5-13. the length of a section. Thus, the high-speed elevation profile
Figure 5-33 is a plot of near-field SI noise versus far-field CPB represents a continuous set of measurements taken in a vir-
noise in which most of the data points are for measurements tual straight line (longitudinal), whereas the CT Meter eleva-
taken on the newly constructed test sections. A linear trend-line tion profile represents one discrete set of measurements taken
through these data shows a general relationship (R2 = 0.51) across the horizontal plane (longitudinal and transverse).
between the two non-equal noise types. Multiple test locations are required by the CT Meter to give a
The interior noise is influenced by the physical properties more accurate indication of the texture depth along the length
of the test vehicle and tires, which determine the type and of the section.
degree of dampening or attenuation that takes place on the Texture depth data for both the existing and the newly con-
various noise frequencies produced at the source. Pass-by structed sections were compiled and statistically analyzed to
76.0
y = 0.6419x + 3.2032
75.0
R2 = 0.5066
74.0
73.0
Interior Leq, dBA
72.0
71.0
70.0
69.0
68.0
67.0
66.0
98.0 100.0 102.0 104.0 106.0 108.0 110.0 112.0
Near-Field SI, dBA
Figure 5-32. Near-field noise versus interior noise.
OCR for page 73
73
85.0
1 to 1 Line
84.0
83.0
y = 0.6331x + 13.942
2
R = 0.5072
82.0
Far-Field CPB, dBA
81.0
80.0
79.0
78.0
77.0
76.0
100.0 102.0 104.0 106.0 108.0 110.0 112.0
Near-Field SI, dBA
Figure 5-33. Near-field noise versus far-field CPB noise.
determine the extent of agreement between the two methods in measuring texture depth--the moving laser tends to stay
of measurement. Figure 5-34 compares high-speed MPD and within a longitudinal groove or atop a longitudinal ridge for
CT Meter MPD and high-speed EMTD and CT Meter MTD. extended distances. As Figure 5-34 shows, the data points for
The high-speed data shown are the mean values of three runs the longitudinally grooved or ground sections are far to the
at a particular position (i.e., wheelpath, lane center) for each right of the equality line, and the CT Meter texture depth read-
texture test section. The CT Meter data are the mean values of ings for these sections are nearly twice those obtained by the
15 test locations at a particular position for each test section. high-speed profiler.
No specific relationship exists between the two methods. Figure 5-35 shows the texture depth data for the newly
While the difference in sampling rates is likely a factor, the tex- constructed test sections. Clear relationships exist, but they
ture type and direction are more profound factors. Because are affected by type and direction of texture. The diamond-
of their same basic direction as the path of high-speed profiler ground section and one of the two longitudinal-groove sec-
texture measurement, longitudinal textures (particularly tions provided substantially lower high-speed profiler texture
those that are cut instead of formed) create greater difficulties depth readings.
2.000
1.800 1 to 1 Line
1 to 1 Line
1.800
1.600
1.600
1.400
1.400
High-Speed EMTD, mm
High-Speed MPD, mm
1.200 CO 1007 WP
(Long Groove) 1.200
AZ 1003 LC
1.000 AZ 1003 WP AZ 1003 WP
(Long DG)
CA 1004 WP (Long DG) 1.000 (Long DG)
y = 0.1077x + 0.6714 (Long Groove) y = 0.1305x + 0.6998
0.800 R2 = 0.0599 AZ 1003 LC
R2 = 0.0288 CA 1007 WP 0.800 CA 1004 WP
CA 1004 LC (Long Groove) (Long DG)
(Long Groove)
(Long Groove)
0.600 CA 1007 LC CA 1007 WP
0.600 (Long Groove) (Long Groove)
CA 1005 LC
(Long DG)
0.400 0.400
0.200 0.200
0.000 0.000
0.000 0.200 0.400 0.600 0.800 1.000 1.200 1.400 1.600 1.800 0.000 0.200 0.400 0.600 0.800 1.000 1.200 1.400 1.600 1.800 2.000
CT Meter MPD, mm CT Meter MTD, mm
(a) Mean Profile Depth (b) Mean Texture Depth
Figure 5-34. Texture depths for existing test sections.
OCR for page 74
74
1.400 1.400
(5B) Long Tine (5B) Long Tine
(heavy turf) (heavy turf)
1.200 1.200
(7) Long Groove y = 0.9263x + 0.0414 (7) Long Groove y = 0.8048x + 0.1483
R2 = 0.7765 2
R = 0.775
1.000 1.000
High-Speed EMTD, mm
High-Speed MPD, mm
(8) Long Groove
(8) Long Groove
0.800 0.800
0.600 0.600
(3) Long DG
0.400 0.400
(3) Long DG
0.200 0.200
1 to 1 Line 1 to 1 Line
0.000 0.000
0.000 0.200 0.400 0.600 0.800 1.000 1.200 1.400 0.000 0.200 0.400 0.600 0.800 1.000 1.200 1.400
CT Meter MPD, mm CT Meter MTD, mm
(a) Mean Profile Depth (b) Mean Texture Depth
Figure 5-35. Texture depths for newly constructed test sections.
Statistical analysis (SAS Proc CORR procedure) of the tex- Newly Constructed Texture Test Sections
ture depth data was performed to further examine the correla-
Table 5-16 shows the Tukey rankings for each of the per-
tion between the two methods. Analysis of the immediate
measure (MPD) and the extrapolated measure (MTD) showed formance variables for the newly constructed sections. For
weak correlation among the existing test sections and fairly each performance variable, statistically significant differ-
strong correlation among the newly constructed test sections ences existed among the various texture types. Also, the effect
(correlation coefficients of 0.24 and 0.86, respectively). In both of test position (i.e., lane center versus wheelpath) on each
cases, significantly higher correlations would result if longitu- performance variable was not statistically significant. The
dinal diamond-ground and grooved sections were excluded interactive effect of texture type and test position was statis-
from the analysis. tically significant in most cases, largely due to one or two
cases where statistical differences in test position were found
to exist for a given texture.
Test Site/Location Performance Analysis
The SAS Mixed Procedure was used to evaluate the perfor- Analysis of Texture, Friction, and Noise
mance characteristics of the test sections at individual sites/
locations, including texture depth (CT Meter MPD and MTD), In this analysis, the texture, friction, and noise measure-
texture orientation parameters (CT Meter RMS and TR), near- ments collected for the 57 existing test sections and 13 newly
field noise (SI), micro-texture friction (DFT(20)), and locked- constructed test sections were combined with other pertinent
wheel friction (newly constructed sections only). In each available test section data. The texture, friction, and noise data
case, the null hypotheses (H0) of (a) all textures being equal and included results from replicate tests using the high-speed tex-
(b) the test positions (lane center versus wheelpath) are equal, ture profiler (MPD and EMTD), the CT Meter (MPD, MTD,
were tested. Statistical rankings using the Tukey Least Signifi- RMS, and TR), the DF Tester (DFT(20) and extrapolated
cant Differences (LSD) method then were developed. F(60)), locked-wheel friction tester (FN40 and extrapolated
F(60)), and noise-testing equipment (near-field SI and interior
Leq). The other pertinent data included
Existing Texture Test Sections
Table 5-15 shows the Tukey rankings for each performance · Age/Traffic
variable for the existing test sections. In most cases, statistically Pavement age at time of testing.
significant differences between texture types within a test Estimated cumulative overall traffic at time of testing.
site/location were identified for each performance variable. Estimated cumulative truck traffic at time of testing.
With the exception of the Colorado US 287 site where only · Climate
wheelpath measurements were taken, statistically significant LTPP climatic zone (WF, DF, WNF, DNF).
differences existed between the wheelpath and lane center Average annual precipitation (AvgPrecip).
measurements, indicating the effects of traffic wear. Average annual snowfall (AvgSnow).
OCR for page 75
Table 5-15. Tukey rankings for existing test sections.
TEXTURE DEPTH TEXTURE ORIENTATION FRICTION NOISE
Avg Avg
CTM CTM Avg Avg Avg
SECT MPD, MTD, Tukey CTM Tukey CTM Tukey Avg Tukey SI, Tukey
ID TEXTURE DESCRIPTION mm mm Rank RMS Rank TR Rank DFT(20) Rank dB(A) Rank
AZ-1001 Long DG (no jacks), 0.235-in. spacing (0.11-in. spacers) 0.99 1.06 2 0.43 2 2.30 1 80.5 1 104.5 1
AZ-1002 Long DG (jacks), 0.235-in. spacing (0.11-in. spacers) 1.01 1.09 2 0.43 2 2.34 1 77.0 1 2 105.7 1 2
AZ-1003 Long DG (no jacks), 0.245-in. spacing (0.12-in. spacers) 1.58 1.65 1 0.64 1 2.45 1 81.0 1 106.35 2
AZ-1004 Long DG (jacks), 0.245-in. spacing (0.12-in. spacers) 0.70 0.78 3 0.41 2 1.73 2 67.5 3 104.8 1
CA-1002 Long DG (no jacks) 0.245-in. spacing (0.12-in. spacers) 0.74 0.81 4 0.39 3 1.87 1 72.0 1 105.65 2
CA-1003 Long Groove (0.75-in. spacing, 0.125-in. depth), burlap drag 1.04 1.10 3 1.48 2 0.72 3 71.5 1 2 3 104.8 2
CA-1004 Long Groove (0.75-in. spacing, 0.25-in. depth), burlap drag 1.23 1.30 2 2.16 1 0.54 4 73.0 1 2 105.25 2
CA-1045 Long Burlap Drag 0.27 0.35 5 0.16 3 1.70 2 71.5 1 2 3 104.45 1
CA-1005 Long DG (no jacks), 0.23-in. spacing (0.105-in. spacers) 0.73 0.81 4 0.40 3 1.83 1 2 68.5 1 2 3 104.45 1
CA-1007 Long Groove (0.375-in. spacing, 0.25-in. depth), broom drag 1.53 1.58 1 1.76 2 0.87 3 67.5 2 3 105.6 2
CA-1075 Long Broom Drag 0.25 0.34 5 0.14 3 1.77 1 2 66.0 3 105.25 2
CO-1007 Long Groove (0.75-in. spacing, 0.125-in. depth), turf drag 1.28 1.35 1 1.38 1 0.94 2 74.5 1 105.9 2
CO-1008 Long Turf Drag 0.27 0.36 3 0.16 3 1.70 1 75.0 1 104.4 1
CO-1009 Long Tine (0.75-in. spacing, 0.125-in. depth), turf drag 0.82 0.89 2 0.77 2 1.07 2 76.5 1 106.1 2
CO-3001 Long Heavy Turf Drag1 0.81 0.88 1 2 0.43 3 1.87 2 92.0 1 103.0 1
CO-3002 Long Tine (0.75-in. spacing, 0.1875-in. depth), no pretexture1 0.95 1.03 1 2 0.76 1 1.26 3 95.0 1 104.3 2
CO-3003 Long Meander Tine (0.75-in. spacing, 0.125-in. depth), no pretexture1 1.01 1.08 1 0.74 1 1.36 3 92.0 1 104.4 2
CO-3004 Long Groove (0.75-in. spacing, 0.125-in. depth), turf drag1 0.96 1.03 1 2 0.72 1 2 1.47 3 81.0 1 104.3 2
CO-3005 Long DG (no jacks), 0.22-in. spacing (0.095-in. spacers)1 0.83 0.91 1 2 0.34 3 2.44 1 89.0 1 102.8 1
CO-3006 Long Tine (0.75-in. spacing, 0.125-in. depth), turf drag1 0.73 0.81 2 0.49 2 3 1.51 3 89.0 1 103.8 1 2
IA-1002 Tran Tine (0.5-in. spacing, 0.075-in. depth), turf drag 106.35 3
IA-1003 Long Tine (0.5-in. spacing, 0.075-in. depth), turf drag 105.4 1
IA-1004 Long Tine (0.75-in. spacing, 0.15-in. depth), turf drag 106.1 2 3
IA-1061 Tran Groove (1-in. spacing, 0.18- to 0.25-in. depth), turf drag 109.0 4
IA-1007 Long Turf Drag 105.7 1 2
IA-2001 Long Tine (0.75-in. spacing, 0.125-in. depth), turf drag 0.63 0.69 1 0.50 2 1.31 1 43.0 1 103.8 1
IA-2002 Long Tine (0.75-in. spacing, 0.125-in. depth), burlap drag 0.70 0.75 1 0.64 1 1.13 2 38.5 2 105.3 2
KS-1002 Long DG (no jacks), 0.235-in. spacing (0.11-in. spacers) & standard-sawed joints 0.77 0.85 2 3 0.36 4 2.15 2 60.5 4 104.65 1
KS-1004 Long DG (no jacks), 0.245-in. spacing (0.12-in. spacers) & single-sawed joints 0.79 0.86 2 3 0.35 3 4 2.29 1 60 .5 3 4 105.5 2
KS-1005 Long DG (jacks), 0.255-in. spacing (0.13-in. spacers) & standard-sawed joints 0.95 1.01 1 0.45 2 3 2.13 2 64. 0 2 3 105.3 3
KS-1006 Long DG (jacks), 0.255-in. spacing (0.13-in. spacers) & single-sawed joints 0.97 1.03 1 0.45 2 2.13 2 62.5 2 3 4 105.4 3
KS-1007 Long DG (no jacks), 0.255-in. spacing (0.13-in. spacers) & standard-sawed joints 0.98 1.04 1 0.44 2 2.21 1 62 .5 2 3 4 105.1 3
KS-1008 Long DG (no jacks), 0.255-in. spacing (0.13-in. spacers) & single-sawed joints 0.91 0.98 1 2 0.42 2 2.18 1 64. 5 2 106.4 3
KS-1010 Long Tine (0.75-in. spacing, 0.15-in. depth), turf drag 0.65 0.72 3 0.53 1 1.22 3 70.0 1 105.35 3
MN-7001 Long Turf Drag 0.40 0.48 2 0.27 2 1.47 1 71.5 2 106.8 1
MN-8001 Long Tine (0.75-in. spacing, 0.125-in. depth), turf drag 0.91 0.98 1 0.82 1 1.12 2 75.0 1 108.15 2
ND-2001 Long Heavy Turf Drag 0.57 0.65 1 0.47 1 1.22 2 78.5 1 110.3 2
ND-2002 Tran Tine (variable spacing, 0.1-in. depth), turf drag 0.39 0.47 2 0.22 2 1.77 1 81.0 1 105.5 1
1
Mean values based on right wheelpath measurements only.
Shaded items indicate no data (or rankings) were available.
1 in. = 25.4 mm
OCR for page 76
Table 5-16. Tukey rankings for the newly constructed test sections.
TEXTURE DEPTH TEXTURE ORIENTATION FRICTION
Avg Avg
CTM CTM Avg Avg
SECT MPD, MTD, Tukey CTM Tukey CTM Tukey Avg Tukey
ID
TEXTURE DESCRIPTION
mm mm Rank RMS Rank TR Rank DFT(20) Rank
IT-1a Long Heavy Turf Drag 0.18 8 9 45.6 1 2 3
IT-1b Long Heavy Turf Drag (modified) 0.28 0.36 4 0.15 9 1.96 7 32.8 5
IT-2 Long Tine (0.75-in. spacing, 0.125-in. depth), no pretexture 0.60 0.66 2 3 4 0.55 4 1.10 2 36.5 3 4 5
IT-3 Long DG (no jacks), 0.235-in. spacing (0.11-in. spacers) 0.65 0.71 2 0.27 7 8 2.41 7 48.3 1 2
IT-5a Long Tine (0.75-in. spacing, 0.125-in. depth), turf drag 0.51 0.58 2 3 4 0.43 5 1.19 2 3 4 33.3 5
IT-5b Long Tine (0.75-in. spacing, 0.125-in. depth), heavy turf drag 1.02 1.06 1 0.81 3 1.25 3 4 5 42.9 2 3 4
IT-6 Long Tine (0.75-in. spacing, 0.075-in. depth), turf drag 0.49 0.56 2 3 4 0.40 5 6 1.23 3 4 5 34.5 4 5
IT-7 Long Groove (0.75-in. spacing, 0.25-in. depth), burlap drag 1.02 1.06 1 1.37 2 0.74 1 44.9 2 3
IT-8 Long Groove (0.75-in. spacing, 0.25-in. depth), turf drag 1.22 1.25 1 1.57 1 0.78 1 53.6 1
IT-9 Tran Tine (0.5-in. spacing, 0.125-in. depth), burlap drag 0.48 0.55 2 3 4 0.34 6 7 1.41 6 42.5 2 3 4
IT-10 Tran Tine (variable spacing, 0.125-in. depth), burlap drag 0.62 0.68 2 3 0.45 5 1.36 5 6 36.8 3 4 5
IT-11 Tran Tine (1.0-in. spacing, 0.125-in. depth), burlap drag 0.44 0.51 3 4 0.40 5 6 1.10 2 3 34.3 4 5
IT-12 Tran Skew Tine (variable spacing, 0.125-in. depth), turf drag 0.58 0.65 2 3 4 0.43 5 1.28 4 5 6 36.9 3 4 5
Shaded items indicate no data (or rankings) were available.
1 in. = 25.4 mm
FRICTION NOISE
Avg Avg
Avg Int CPB
SECT Avg Tukey Avg Tukey SI, Tukey Leq, Tukey Lmax, Tukey
ID
TEXTURE DESCRIPTION
FN40R Rank FN40S Rank dB(A) Rank dB(A) Rank dB(A) Rank
IT-1a Long Heavy Turf Drag 47.4 2 3 101.5 2 69.5 3
IT-1b Long Heavy Turf Drag (modified) 46.4 2 3 24.4 2 101.7 2 3 69.2 3 79.2 2 3 4
IT-2 Long Tine (0.75-in. spacing, 0.125-in. depth), no pretexture 45.9 3 32.5 1 2 103.2 6 69.5 3 79.4 3 4
IT-3 Long DG (no jacks), 0.235-in. spacing (0.11-in. spacers) 50.3 2 3 100.5 1 67.6 1 77.5 1
IT-5a Long Tine (0.75-in. spacing, 0.125-in. depth), turf drag 49.7 2 3 37.8 1 102.3 4 71.1 4 77.6 1
IT-5b Long Tine (0.75-in. spacing, 0.125-in. depth), heavy turf drag 50.2 2 3 43.4 1 105.3 8 72.1 5 82.3 7
IT-6 Long Tine (0.75-in. spacing, 0.075-in. depth), turf drag 48.2 2 3 34.0 1 2 102.2 3 4 69.7 3 78.5 1 2 3
IT-7 Long Groove (0.75-in. spacing, 0.25-in. depth), burlap drag 47.5 2 3 101.7 2 68.1 1 2 79.1 1 2 3 4
IT-8 Long Groove (0.75-in. spacing, 0.25-in. depth), turf drag 55.7 1 2 102.4 4 5 67.6 1 77.9 2
IT-9 Tran Tine (0.5-in. spacing, 0.125-in. depth), burlap drag 44.7 3 102.6 4 5 67.7 1 80.7 5
IT-10 Tran Tine (variable spacing, 0.125-in. depth), burlap drag 62.1 1 37.1 1 102.8 5 6 68.8 2 3 81.3 6 7
IT-11 Tran Tine (1.0-in. spacing, 0.125-in. depth), burlap drag 48.6 2 3 31.5 1 2 104.0 7 69.3 3 80.4 4 5
IT-12 Tran Skew Tine (variable spacing, 0.125-in. depth), turf drag 49.2 2 3 29.0 2 102.6 4 5 69.0 3 80.0 4 5
Shaded items indicate no data (or rankings) were available.
1 in. = 25.4 mm
OCR for page 77
77
Number days >90°F [>32°C] and number days <32°F DFT ( 20 ) = 76.39 - 18.99 × GTIEAC - 5.20 × GTILDG
[<0°C]. - 8.99 × GTILGr - 0.34 × Precip + 0.14 × Snow
Freezing index (FI). - 0.47 × ln ( Truck ) + 10.12 × MPDCTM Eq. 5-1
· Pavement
Number of joints per 1,000 ft (305 m) (#Jts). where
Joint width (JW). GTIEAC = GTI for EAC texture (=1 if EAC, 0 otherwise).
General pavement condition (PvtCond) (excellent, good, GTILDG = GTI for diamond ground texture (=1 if dia-
fair, etc.). mond ground, 0 otherwise).
GTILGr = GTI for longitudinal grooved texture (=1 if
The SAS Proc ANOVA and REG procedures were used to longitudinal grooved, 0 otherwise).
determine the effect of independent variables on each of two Precip = Average annual precipitation (in.).
dependent variables--DFT(20) and near-field SI. Multiple Snow = Average annual snowfall (in.).
regression techniques also were used to augment the under- Ln(Truck) = Log-normal cumulative truck applications.
standing of the influence of the independent variables on MPDCTM = MPD from CT Meter (mm).
friction/micro-texture and noise measurements.
The results of the ANOVA and regression analyses high-
Micro-Texture/Friction lighted the effect of traffic (e.g., truck traffic) on micro-texture,
the favorable micro-texture of transverse textures over uni-
Using DFT(20) data from 55 of the 70 test sections (data form and longitudinal textures, and the possible influence
were not available for some sections in Illinois, Iowa, Kansas, of macro-texture and mega-texture (roughness) on micro-
Michigan, and Minnesota), ANOVA testing was conducted to texture. The relationship of climatic variables to DFT(20) could
identify variables that significantly influence this micro-texture not be determined in this project.
parameter. Because initial results showed conflicting indica-
tions of the effect of traffic on DFT(20) (i.e., increased DFT(20)
corresponding to increased traffic), the 13 newly constructed SI Noise
test sections and 2 sections in Iowa that exhibited unusually
Based on SI data from all 70 test sections, ANOVA testing
low DFT(20) data were removed from analysis. Thus, data
showed a statistically significant relationship of traffic, texture
from 30 of the 70 total test sections were used. ANOVA testing
depth (CT Meter MTD), and general texture type/texture
indicated the following findings regarding the effects of the
direction to SI at the 95% confidence level. SI increases as tex-
dependent variable DFT(20):
ture depth and traffic increase. With respect to general texture
· Log-normal truck traffic (lnTruck)--Increased cumulative
type/texture direction, SI is primarily driven by the significant
range in noise differences exhibited by asphalt-surfaced pave-
truck traffic, reduced DFT(20) values.
· CT Meter MTD--Higher DFT(20) values were measured
ments at the low end and EAC at the high end. However, sig-
for pavements with greater texture depth. nificant differences between grooved and ground textures and
· Precipitation (AvgPrecip)--Lower DFT(20) values were transverse-tine textures were also noted.
measured for locations with higher annual average pre- Multiple regression using SAS Proc REG yielded various
cipitation. models linking independent variables with log(SI), the best of
· Roughness--Higher DFT(20) values corresponded to which had an R2 of 0.61 as follows:
increased IRI values.
· Test position--Higher DFT(20) values were obtained for Log ( SI ) = -0.35 + 0.37 × TDTRAN + 0.62 × GTIEAC + 0.20
lane center compared with the wheelpath position. × GTILTi + 0.01 × # Jts + 0.06 × ln ( Traffic )
· General texture indicator (GTI)--Significant differences + 0.53 × MPDHS Eq. 5-2
were obtained for some general texture types, particularly
the substantially lower DFT(20) values for the shotblasted where
section in Texas. TDTRAN = Transverse texture direction (=1 if transverse,
· Texture direction (TD)--Higher DFT(20) values were 0 otherwise).
obtained for transverse textures compared with both GTIEAC = GTI for EAC texture (=1 if EAC, 0 otherwise).
longitudinal and uniform/isotropic textures. GTILTi = GTI for longitudinal-tine texture (=1 if
longitudinal-tine, 0 otherwise).
Multiple regression using SAS Proc REG yielded various #Jts = Number of joints per 1,000 ft (305 m).
models linking independent variables with DFT(20) the best Ln(Traffic) = Log-normal cumulative traffic applications.
of which had an R2 of 0.85 as follows: MPDHS = MPD from high-speed profiler (mm).
OCR for page 78
78
In this model, the key variables, aside from the EAC texture the full test segment length, and the second analysis used dis-
type, are texture depth, transverse texture direction, and crete location values of the texture PSD parameters. For both
longitudinal-tine texture type. Traffic and joints (in terms of analyses, discrete location values of SI were computed corre-
the number/frequency of joints) also are seen as factors in this sponding to each of the five short (6 ft [1.8 m]) segments
model. Thus, the ANOVA and regression analyses indicated where CT Meter tests were performed. The mid-lengths of
that near-field SI noise is influenced to a large extent by tex- these discrete segments were located 100, 200, 300, 400, and
ture depth (the deeper, the louder) and texture type and direc- 500 ft (30.5, 61, 91.5, 122, and 152.5 m) from the start of each
tion. Also, SI is increased as the number of traffic applications test section. All replicate test data were included in the analy-
increased and the number/frequency of joints increased. sis, which consisted of the SAS REG procedure. Results of the
first analysis yielded the following model with an R2 of 0.772:
Noise-Texture Relationship
SI = 106.63 - 14.28 × A1 A 2 + 2.79 × RMS - 1.25 × Dir Eq. 5-3
Near-field SI data and various texture parameter data from
the 13 newly constructed test sections were used to develop a where
statistical model relating pavement texture and pavement Dir = 0, for transverse or uniform/isotropic texture.
tire noise to better understand the specific texture parameters = 1 for longitudinal texture.
that significantly influence the generation of noise. The model
was intended to establish SI as a function of one or more of Figure 5-36 plots the actual versus predicted SI values using
the following variables: this partially discrete model. The data and corresponding
trend line on the left are based on actual averaged discrete
· Texture direction (longitudinal, transverse, or uniform/ location SI values. The model gives a near 1-to-1 relationship
isotropic). of actual and predicted SI.
· CT Meter MTD The data and corresponding trend line on the right are based
· CT Meter RMS on actual average SI values for the full-length (528 ft [161 m])
· CT Meter TR of a test segment. While other factors influence the shift
· Texture PSD L4/L63 (derived from texture profiles from high- between these two trend lines, the joint slap may be a major rea-
speed profiler) son for the nearly 2 dB(A) difference because the discrete loca-
· Texture PSD A1/A2 (derived from texture profiles from high- tions from which discrete SI values were derived did not include
speed profiler) the pavement joints (0.25-in. [12.7 mm] wide, 15-ft [4.6-m]
· Texture PSD Peak Wavelength (PW) (derived from texture spacing), but the full-length SI did include the joints.
profiles from high-speed profiler) As part of the second statistical analysis, discrete location
PSD parameter data were computed using extracted texture
Two sequential statistical analyses were performed. The profile data corresponding to each of the five short (6 ft [1.8 m])
first analysis used texture PSD parameter values representing segments where CT Meter tests were performed. Because of
Discrete SI Entire/Full-Length Segment SI
Linear (Discrete SI) Linear (Entire/Full-Length Segment SI)
106
1-to-1 line
105
2
104 R = 0.7716
Predicted SI, dBA
2
103 Adj. R = 0.686
102
101
100
99
Effect of Joints
98
98 99 100 101 102 103 104 105 106
Actual SI, dBA
Figure 5-36. Actual versus predicted SI, based on partially discrete
SI-texture model applied to full-length texture PSD parameter data.
