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 17

17
Accelerated Pavement Test Results 0
Accelerated pavement tests were conducted at the APT 5
facility at the Indiana DOT Research Division in West
Average Total Rut Depth, mm
Lafayette, Indiana. Up to four test lanes can be constructed in 10
this facility at a time using conventional paving equipment.
Details of the facility, test section construction, and data col- 15
lection can be found in Appendix D, which is available in
NCHRP Web-Only Document 82. 20
25
Rutting
Rutting was monitored by recording transverse surface 30
profiles at increments of traffic applied with dual tires and
without wander. These profiles were captured with software 35
that automatically reduces and stores the data in a spread- 1 10 100 1000 10000 100000
Number of Wheel Passes, N
sheet (7). An initial profile was recorded before traffic appli-
CA-1 CA-2 CA-3 CA-4 CA-5b
cation and used as the baseline reference for determining
rutting from subsequent profiles. Figure 6. Coarse-graded mixture rut
Each test lane was trafficked until a total rut depth of 20 depth development.
mm was achieved or 20,000 wheel passes were applied,
whichever occurred first. Profiles were recorded at nine
locations over the length of a given test section; however, to 5 (worst). The mixtures were ranked by four parameters;
three consecutive sections nearest the center of the test three of the mixtures were ranked the same by the four
section (Sections 4, 5, and 6, see Figure D.8) were averaged parameters; however, the early rut rate inverts 1 and 2, and
and used as a single result in the subsequent analyses. 3 and 4. This would seem to indicate that after approxi-
These three locations were used because the APT wheel mately 200 wheel passes are applied with the APT, one can
carriage travels at a constant speed over this portion of the obtain a good indication of the relative rutting rank of an
test lane. HMA mixture.
Fine-Graded Mixtures. APT rutting of the fine-graded
Performance mixtures is shown in Figure 7. Testing on Mixture FA-1 (Nat-
Coarse-Graded Mixtures. Rutting in the APT as a ural Sand A) was terminated at 1,000 wheel passes because of
function of wheel passes for the coarse-graded mixtures is excessive total rut depth. Testing on Mixture FA-2 was termi-
shown in Figure 6. In addition to total rut depth, the rut rate nated inadvertently at 12,500 wheel passes; however, the trend
was also computed. Rut rate is simply the slope of the regres- for this mixture shows that it would have most likely accom-
sion line and has units of mm/log(N) where N is the num- modated additional passes with a minimal increase in rutting.
ber of wheel passes. Rutting data for mixtures CA-1, CA-2, Rut development curves for fine-graded mixtures also
CA-3, and CA-5b exhibited a bilinear trend; these data were appear to be bilinear. As above, the first regression line repre-
fitted by two logarithmic functions. The rut rate during the sents early traffic stage rutting and the second later traffic
early traffic stage is the slope of the first regression line while
the rut rate during later traffic stage is the slope of the sec-
ond regression line. The regression equations for all mix- Table 11. Coarse-graded mixture rutting
tures are given in Table 11. regression equations.
Rutting performance parameters for coarse-graded mix- Mixture ID Regression Equation Condition
tures are shown in Table 12. These parameters include total CA-1
Total rut depth = 1.56 log(N) + 0.28 for N < 400
rut depth at 5,000 and 20,000 wheel passes as well as the rut Total rut depth = 2.25 log(N) -1.50 for N > 500
Total rut depth = 1.89 log(N) + 0.17 for N < 200
rate during early and later traffic stages (early and late CA-2
Total rut depth = 3.56 log(N) -3.74 for N > 200
stages are defined by the N break point shown in the table). CA-3
Total rut depth = 4.88 log(N) 1.27 for N < 200
Total rut depth = 14.72 log(N) 26.02 for N > 200
As an example, for mixture CA-2, early traffic is where CA-4 Total rut depth = 1.75 log(N) 0.13 for all N
N200, while late traffic is N200. For each of the rutting CA-5b
Total rut depth = 2.09 log(N) 0.01 for N < 200
Total rut depth = 3.27 log(N) 2.66 for N > 200
performance parameters, mixtures are ranked from 1 (best)

OCR for page 17

18
Table 12. Coarse-graded mixture rutting performance.
Total Rut
Mixture Total Rut Depth
Depth at Total Rut Rate (mm/log(N)) Overall
ID at 20,000 Passes
5,000 Passes
Early Later Rank
mm Rank mm Rank Rank Rank
traffic Traffic
CA-1 7.1 2 7.6 2 1.6 1 2.3 2 2
CA-2 9.5 4 11.3 4 1.9 3 3.6 4 4
CA-3 29.5 5 --1 --1 4.9 5 14.7 5 5
CA-4 6.3 1 7.2 1 1.8 2 1.8 1 1
CA-5b 9.2 3 11.1 3 2.1 4 3.3 3 3
1 Tested to only 5,000 wheel passes.
stage rutting. Regression equations for the data are given in Moisture Susceptibility
Table 13.
Table 14 lists total rut depths at 1,000; 5,000; and 20,000 Before APT testing, six cores were taken from each of the
wheel passes and rutting rates for both early and later traffic moisture susceptibility test lanes. In-place densities and air
stages. These rutting parameters were used as the basis for voids of the test lanes were determined from the cores, which
ranking mixture performance from 1 (best) to 6 (worst). were then used to test the plant-produced mixtures for mois-
Overall rank is the rank appearing the most number of times ture susceptibility in accordance with AASHTO T 283. In
for each mixture. Three of the parameters (rut depth at 5,000 addition to moisture conditioning, the conditioned speci-
and 20,000, and rut rate at later traffic) rank the mixtures the mens were also subjected to one freeze/thaw cycle before
same. The remaining two parameters rank the mixtures the being tested in indirect tension.
same, but have 3 and 4 inverted from the previous three The AASHTO T 283 test results are shown in Table 15. Aver-
parameters. Again, it appears that a good indication of the rut ages of VTM ranged from 7.1 to 9.9 percent. Degree of satu-
resistance can be gained after approximately 200 wheel passes ration for the conditioned specimens varied from 67.4 to 78.2
of the APT. percent. The tensile strength ratio (TSR) [the ratio of the indi-
rect tensile strength of conditioned specimens to that of dry
(unconditioned) specimens] varied from 1.10 to 0.79. Some
specimens were loaded until they cracked. The interior sur-
0
faces were then inspected for stripping and photographs were
taken. Visual observation indicated that stripping occurred.
5 Typically, conditioned specimens lost their glossy appearance
and the interior surface exhibited a brownish tint (see Figures
E.1 to E.5 in Appendix E provided in NCHRP Web-Only Doc-
10
Average Total Rut Depth, mm
ument 82). Stripping in terms of lost binder film was also
observed in specimens with dolomite coarse aggregate.
15
Performance
20
Rutting accumulation for all test lanes is shown in Figure 8.
A 20-mm total rut depth criteria was adopted for traffic ter-
25 mination. Mixture FAM1 (Natural Sand A) was terminated at
1,000 wheel passes, while mixture FAM5 (Natural Sand B)
30 reached the 20-mm total rut depth criteria and was terminated
at 7,500 wheel passes. The FAM5 mixture exhibited a signifi-
cant increase in rutting between 5,000 and 7,500 wheel passes.
35
1 10 100 1000 10000 100000
There was also an increase in rutting after 3,000 wheel passes
Number of Wheel Passes, N
on mixture FAM3 (Granite Sand). In general, this type of
FA-1 FA-2 FA-3 FA-4 FA-5b FA-6b
rutting does not occur in dry rutting tests and may be an indi-
cation of stripping. The rutting data for mixtures FAM1,
Figure 7. Fine-graded mixture rut depth FAM2, and FAM4 do not show a change in rate of rutting
development. accumulation.

OCR for page 17

19
Table 13. Fine-graded mixture rutting regression
equations.
Mix ID Regression Equation Condition
Total rut depth = 7.26 log(N) - 2.78 for N < 100
FA-1
Total rut depth = 21.10 log(N) - 33.33 for N > 200
Total rut depth = 3.81 log(N) 0.74 for N < 200
FA-2
Total rut depth = 5.00 log(N) 3.29 for N > 300
Total rut depth = 2.78 log(N) 0.52 for N < 200
FA-3
Total rut depth = 4.02 log(N) 3.56 for N > 300
FA-4 Total rut depth = 1.06 log(N) 0.61 for N < 200
Total rut depth = 1.63 log(N) 1.84 for N > 300
Total rut depth = 2.75 log(N) 0.004 for N < 400
FA-5b
Total rut depth = 4.45 log(N) 4.70 for N > 500
Total rut depth = 2.10 log(N) 0.57 for N < 200
FA-6b
Total rut depth = 3.31 log(N) 3.54 for N > 300
Table 14. Fine-graded mixture rutting performance.
Rut Rut Rut Total Rut Rate (mm/log (N))
Depth Depth Depth Early Later
Rank
Rank
Rank
Mix at at at Traffic Traffic
ID 1,000 5,000 20,000
Rank
Rank
Passes Passes Passes Overall
(mm) (mm) (mm) Rank
FA-1 30.6 6 NA1 NA1 7.3 6 21.1 6 6
FA-2 11.1 5 15.4 5 18.12 5 3.8 5 5.0 5 5
FA-3 8.6 4 11.3 3 15.5 3 2.8 4 4.0 3 3
FA-4 2.8 1 4.3 1 5.4 1 1.1 1 1.6 1 1
FA-5b 8.4 3 11.7 4 16.8 4 2.8 3 4.5 4 4
FA-6b 6.4 2 8.7 2 10.4 2 2.1 2 3.3 2 2
1Testing was terminated at 1,000 wheel passes because total rut depth was more than 20mm
2Testing was inadvertently terminated at 12,500 wheel passes; total rut depth at 20,000 wheel
passes was predicted using the regression equation.
The rutting data were also plotted as a function of the log stages. These rutting parameters were used in ranking per-
of the number of wheel passes (see Figure 9). This plot formance. Among these rutting parameters, only total rut
shows that the rutting of mixtures FAM1, FAM2, FAM3, and depth at 1,000 wheel passes and total rutting rate are available
FAM5 appears to be bilinear. As a result, data for these mix- for all mixtures. Based on these rutting parameters, the mix-
tures were fitted with two logarithmic regression lines. The tures were ranked from 1 (best) to 5 (worst). Overall per-
regression equations for all mixtures are shown in Table 16. formance rank from best to worst are FAM2 (Crushed Gravel
Table 17 shows total rut depths for 1,000; 5,000; and 20,000 Sand), FAM3 (Granite Sand), FAM4 (Traprock Sand), FAM5
wheel passes and total rutting rates for early and later traffic (Natural Sand B), and FAM1 (Natural Sand A).
Table 15. AASHTO T 283 and MBV test results.
Mixture ID FAM1 FAM2 FAM3 FAM4 FAM5
Crushed
Natural Natural
Aggregate Type Gravel Granite Traprock
Sand A Sand Sand B
Dry Specimens
Tensile Strength, kPa 452.3 689.6 837.7 652.8 808.7
Average Air Voids, % 8.4 9.1 8.3 9.9 7.1
Conditioned Specimens
Tensile Strength, kPa 499.4 621.2 708.1 537.0 640.2
Average Air Voids, % 9.1 9.2 8.2 9.4 7.5
Degree of Saturation, % 70.9 78.2 67.4 69.2 71.4
TSR 1.10 0.90 0.85 0.82 0.79
Methylene Blue Value 3.3 1.3 8.0 5.1 5.0

OCR for page 17

20
30 Table 16. Regression equations.
Mix ID Regression Equation Condition
Total rut depth = 6.82 log(N) - 2.40 for N < 200
25 FAM1
Total rut depth = 26.88 log(N) 52.53 for N > 300
Total rut depth = 1.62 log(N) + 0.77 for N < 200
FAM2
Average Total Rut Depth, mm
Total rut depth = 2.31 log(N) 0.89 for N > 300
20 Total rut depth = 2.18 log(N) 1.32 for N < 1000
FAM3
Total rut depth = 4.88 log(N) 10.16 for N > 2000
FAM4 Total rut depth = 2.53 log(N) 0.69 for all N
15
FAM5 Total rut depth = 3.31 log(N) 1.14 for N < 400
Total rut depth = 8.03 log(N) 13.94 for N > 500
10
When APT traffic application was complete, cores were
5 collected and split open to determine visually if stripping had
occurred. Photographs of the split surfaces are shown in Fig-
0
ures E.6 through E.10 (Appendix E). Visual inspection
0 5000 10000 15000 20000 revealed no stripping of FAM1, FAM2, and FAM4 cores.
Number of Wheel Passes, N There was a loss of glossiness on the split surfaces of FAM3
FAM1 Natural Sand A, IN FAM2 Crushed Gravel Sand, IN (Granite Sand) and FAM5 (Natural Sand B). These two mix-
FAM3 Granite Sand, NC FAM4 Traprock Sand, VA tures also exhibited signs of stripping in their rutting data.
FAM5 Natural Sand B, OH Signs of stripping were observed on the bottom of cores taken
from all of the test lanes.
Figure 8. Rut depth development.
Fatigue
Relationships between fatigue cracking and coarse and fine
aggregate properties were evaluated through construction
and testing of six mixtures in the APT as indicated in Table
18. Fatigue performance was characterized by percentage of
fatigue cracking in the wheel path. The experiment is similar
0 to the rutting experiment, with the exception that a conven-
tional flexible pavement was installed consisting of 100 mm
5
of HMA and 200 mm of a crushed stone on a subgrade. An
attempt was made to control the pavement test temperature
Average Total Rut Depth, mm
at approximately 10 to 20°C. However, because of the unavail-
10 ability of a cooling system, this temperature range was
exceeded for the tests conducted in June and July.
15
The six experimental fatigue mixtures listed in Table 18
were selected based on the earlier APT rutting performance
and aggregate quality. Mixtures were selected to have as
20 wide a range in both rutting performance and aggregate
quality as possible. Of the six mixtures, three were coarse-
25 graded and three were fine-graded. Mixture FA-1 (Natural
Sand A) had the poorest aggregate qualities and exhibited
the worst rutting performance. The mixture was included,
30
even though such a mixture probably would be replaced
1 10 100 1000 10000 100000
Number of Wheel Passes, N
in the field before failing in fatigue. Testing mixtures with
the greatest range of aggregate quality, as determined by
FAM1 Natural Sand A, IN FAM2 Crushed Gravel Sand, IN
the aggregate test methods, were expected to provide the
FAM3 Granite Sand, NC FAM4 Traprock Sand, VA
most useful information for determining the strength of
FAM5 Natural Sand B, OH
the relationships between the aggregate properties and
Figure 9. Rut depth development. fatigue performance.

OCR for page 17

21
Table 17. Moisture susceptibility rutting performance.
Total Total Total
Overall Rank
Rut Rut Rut Total Rut Rate (mm/log(N))
Rank
Rank
Rank
Mix Depth Depth at Depth at
ID at 1000 5000 20000
Rank
Rank
Passes Passes Passes Early Late
Traffic Traffic
(mm) (mm) (mm)
FAM1 28.5 5 --1 --1 6.8 5 26.9 5 5
FAM2 6.1 2 7.6 1 9.2 1 1.6 1 2.3 1 1
FAM3 5.4 1 7.7 2 11.0 2 2.2 2 4.9 3 2
FAM4 7.1 3 8.8 3 9.9 3 2.5 3 2.5 2 3
FAM5 10.1 4 16.1 4 --2 3.3 4 8.0 4 4
1Testing terminated at 1,000 wheel passes; 20 mm rut depth reached.
2Testing terminated at 7,500 wheel passes; 20 mm rut depth reached.
Based on rutting performance from best to worst, the fine- grade was so "springy" that construction was made difficult.
graded mixtures selected for the fatigue study were FA-4 It was later discovered that the drain for the APT pit was
(Granite), FA-3 (Natural Sand B), and FA-1 (Natural Sand A). clogged and that excess moisture could not be drained from
Likewise, coarse-graded mixtures were CA-4 (Granite), CA-2 the soil. Once the drain was repaired, the excess moisture
(Limestone), and CA-3 (Uncrushed Gravel). drained and the soil became stiffer.
After subgrade compaction, a geotextile fabric was placed
on the subgrade and an unbound, crushed stone base course
Test Section Construction
was placed and compacted such that the finished base course
The first step in constructing the fatigue test sections was 200 mm in depth. The 100-mm deep HMA test section
involved removing the previous rutting test sections. The mixtures were then constructed on the base course using con-
underlying Portland concrete cement (PCC) slabs were then ventional HMA construction techniques as described in
removed, along with the underlying pea gravel fill. Subse- Appendix D (available in NCHRP Web-Only Document 82).
quently, a subgrade soil was installed and compacted to a
depth of 1.5 m. Moisture was added to the soil, and the two
Fatigue Testing
were mixed using two motorized tillers. Compaction was
accomplished using vibrating plate compactors. The Proctor Performance data were collected throughout the loading
curve for the soil is shown in Figure 10. An attempt was made process, including transverse profiles when needed. Longitu-
to compact the soil on the wet side of optimum in order to dinal and fatigue cracking were measured by counting the
render the soil more plastic and thereby aid the pavement number of cracks that developed during loading. The fre-
fatigue process. The optimum moisture content is approxi- quency of measurement varied with test section and
mately 15 percent and the soil was compacted at 18- to 20- depended on how quickly the cracking occurred. From a
percent moisture. As can be seen from Figure 11, the fatigue standpoint, the criterion was established that a test
California bearing ratio (CBR) value for moisture content in
the desired range was approximately 2. The result was a
"springy" subgrade that was indeed plastic. In fact, the sub- 1850
1800
Density (kg/m3)
Table 18. Fatigue experiment design.
1750
Aggregate Performance Aggregate Type
Tests Category Coarse Aggregate Fine Aggregate
Coarse Aggregate Test CA-2 (Limestone) 1700
Methods Evaluation
CA-3 (Uncrushed Gravel) Natural Sand A
(Coarse-Graded
Mixtures) CA-4 (Granite)
1650
FA-1 (Natural Sand A) 5 10 15 20 25
Fine Aggregate Test
Methods Evaluation Uncrushed Gravel FA-3 (Natural Sand B) Moisture Content (%)
(Fine-Graded Mixtures) FA-4 (Granite)
Figure 10. Subgrade Proctor curve.

OCR for page 17

22
15
10
CBR Value
5
0
5 10 15 20 25
Moisture Content (%)
Figure 11. Soil CBR values.
section was considered to have failed when the center one- Figure 12. CA-3 test section after 1,000 wheel passes.
third of the test section had exhibited fatigue cracking exceed-
ing 10 percent of the area. This center area of the test lane was in the section during application of the last 60,000 wheel
chosen because it was expected to have the most uniform passes as shown in Figure 13. Subgrade failure was repaired
mixture properties. Random wheel wander was also incorpo- and traffic continued; however, damage at the edge continued
rated during testing to help avoid rutting; it was done by a to accumulate and it was necessary to discontinue trafficking
random number generator within the APT control program. of the section to avoid damage to the APT equipment.
The fatigue results are shown in Table 19. The third mixture to undergo fatigue testing was the FA-1
The first test section to be trafficked in the fatigue experi- mixture produced using Natural Sand A. This test section
ment was the CA-3 (Uncrushed Gravel) mixture. Loading proved difficult to compact during construction because of
consisted of a 40 kN load on dual wheels with a tire pressure mixture tenderness and the resiliency of the section against
of 690 kPa. The section deformed quickly and failed after which it was being compacted. As a result, numerous cracks
only 1,000 passes. The failure is shown in Figure 12. Three to occurred during compaction. These "roller"cracks were painted
four cracks appear in the center one-third of the test lane; with a lime-water solution so as to appear white to distinguish
however, inspection of the failure revealed that the subgrade them from cracks caused by APT loading. The existence of these
failed before the fatigue properties of the mixture could be cracks from the outset of the testing most certainly influenced
fully tested. the fatigue results of the section. The section before traffic is
Testing began on the next test section, CA-2 (Limestone),
with a reduced load of 26.7 kN applied to the dual wheels and
the tire pressure reduced to 620 kPa. After 8,000 passes, no
signs of cracking were observed, so the load was increased to
33.3 kN and an additional 12,000 passes were applied. The
load was then increased to 40 kN and the tire pressure
increased to 690 kPa. Testing was stopped at 80,000 wheel
passes because the section showed signs of subgrade failure
near the edge of the test pit. Minor fatigue cracking developed
Table 19. Fatigue results.
Test Percent Average Total Number of
Mixture Cracking Rut Depth (mm) Wheel Passes
CA-2 1 None 80,000
CA-3 30 None 1,000
CA-4 None 40.0 20,000
FA-1 25 None 2,000
FA-3 None 33.6 20,000 Figure 13. CA-2 test section after 80,000
FA-4 None 43.8 20,000
wheel passes.

OCR for page 17

23
Figure 14. Initial FA-1 test section. Figure 15. FA-1 test section after 2,000 wheel passes.
shown in Figure 14. Trafficking of this test section began with fatigue results of the last three APT test sections. Finally, in
an applied load of 31.1 kN and tire pressure of 690 kPa. As August 2003, trafficking was discontinued because 20,000
shown in Figure 15, the section exhibited substantial fatigue passes were applied on each section with no evidence of sig-
cracking when the testing was stopped after 2,000 passes. nificant fatigue cracking; however, each section did show rut-
Fatigue testing of mixtures CA-2, CA-3, and FA-1 was ting. The FA-3 mixture (Natural Sand B) had an average rut
finally completed in June 2003, and subsequently, the test sec- depth of 33.6 mm, while the FA-4 (Granite) and CA-4 (Gran-
tions were removed. In July 2003, test sections of mixtures ite) mixtures had average rut depths of 43.8 and 40.0 mm,
FA-3 (Natural Sand B), FA-4 (Granite), and CA-4 (Granite) respectively. Recently, cooling capabilities have been installed
were placed in the APT facility. Testing immediately began on in the APT building and fatigue tests have been conducted.
the FA-3 mixture; rutting began to develop with no sign of These tests indicate that, had the desired temperature range
fatigue cracking. Without the ability to control temperature in been maintained during the testing of the FA-3, FA-4, and CA-
the APT facility, the test pavement temperatures were well over 4 mixtures, these mixtures probably would have sustained
the desired temperature range of 10 to 20°C. This affected the 80,000 to 100,000 APT wheel passes before failing in fatigue.