Content uploaded by Cherif Amer-Yahia
Author content
All content in this area was uploaded by Cherif Amer-Yahia on Mar 13, 2018
Content may be subject to copyright.
Content uploaded by Cherif Amer-Yahia
Author content
All content in this area was uploaded by Cherif Amer-Yahia on Mar 13, 2018
Content may be subject to copyright.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
7
INVESTIGATIONS OF PREMATURE FAILURE
OF AN ASPHALT OVERLAY
Cherif Amer-Yahia, University of Phoenix, 4635 E Elwood St, Phoenix, AZ 85040, USA. Email:
chay@email.phoenix.edu
Julia Miller, Todd Majidzadeh, Chhote Saraf & Kamran Majidzadeh, Resource International, Inc.,
6350 Presidential Gateway, Columbus, OH 43231, USA
doi: 10.1515/ijpeat-2016-0005
ABSTRACT
Premature pavement failures and distresses continue to happen despite significant
advancements in pavement technology. This paper reports an integrated approach
used to evaluate subsurface conditions of 72 centerline kilometers of a highway
pavement, as part of an investigation prompted by premature pavement failures of a
recent mill and asphalt overlay project. The integrated approach involved ground
penetrating radar (GPR), coring, and laboratory testing, in conjunction with a review
of existing falling weight deflectometer (FWD) data, design plans, construction and
post-construction documents, and other available data. A visit to the site was arranged
to assess the conditions of the overlay and identify the type(s) of distresses which
were visible on the pavement surface. GPR data was used to identify contributing
factors to the premature surface failures, such as large areas of stripped pavement
beneath the new asphalt overlay, and localized areas of wet or moist old asphalt base.
Pavement cores provided data to verify the results of GPR Tests. FWD test data was
used to determine the resilient moduli of the existing pavement layers. The premature
pavement failures in this study were believed to be caused by infiltration of water into
the milled surface prior to overlaying, questionable structural integrity of the milled
surfaces of old asphalt base material, and de-bonding due to tack coat placement
issues. Recommendations based on the findings from this study are included in the
paper.
INTRODUCTION
Premature pavement failures and distresses continue to happen despite significant
advancements in pavement technology. The root causes of these failures need to be
identified to prevent them from occurring and to find corrective actions. This paper
reports an integrated approach used to evaluate subsurface conditions of 72 centerline
kilometers of a highway pavement, as part of an investigation for the owner/agency
prompted by the premature pavement failures within two months of a mill and asphalt
overlay project. The integrated approach involved ground penetrating radar (GPR)
(David, 2004), coring, and laboratory testing, in conjunction with a thorough review
of existing falling weight deflectometer (FWD) (Bush and Baladi, 1989) data, design
plans, construction and post-construction documents, and other available data.
A pavement rehabilitation was designed and constructed (by others) for a section of
the tested 72 centerline kilometers (6.61 km) toll road with an AADT of 257,000. The
roadway generally consists of 4 to 6 lanes in each direction. The project is located in
a tropical climate zone with an average year round temperature of 27ᴼ C and an
average of 1,550 mm of rainfall per year.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
8
The pavement rehabilitation was performed between August 2012 and March 2013.
Premature pavement failures appeared within two to three months following the final
overlay placement. Premature pavement failure for this project is defined as the
occurrence of defects and distresses in the completed asphalt courses appearing
within two months to one year of placement creating a loss of expected performance
and requiring ongoing maintenance. Distresses were indicated by potholes and the
appearance of fines on the surface; usually fines appeared on the pavement surface
first, followed by cracking, and finally potholes.
The pavement investigation occurred in two stages, first at the project level in May
2013 and second by an outside independent investigation in October 2013. The first
investigation initiated by the owner/agency in May 2013 asked the original
construction inspection/firm for the project to investigate and perform testing on a
pavement section that exhibited premature failure, and on two pavement sections that
did not. Pavement cores were taken and tested for moisture susceptibility,
permeability, and pavement saturation. The construction inspection firm in
conjunction with the contractor performed additional core testing and postulated that
water was moving upward from the subgrade and through compromised existing
asphalt material to the new overlay causing stripping. Additionally, because the new
asphalt overlay was essentially impermeable, the presence of water under the overlay,
coupled with temperature changes and traffic loading, caused pore water pressure to
build up and fractured the asphalt-aggregate bond.
In October 2013 the owner/agency asked for an independent review and investigation
to evaluate possible causes of the premature pavement failures. This paper addresses
the methodology and findings of October - December 2013 investigation.
GROUND PENETRATING RADAR TESTING
In October 2013 GPR testing was performed in the right wheel path of each lane of
the toll road. The GPR utilized a non-contact horn antenna suspended 475 mm over
the pavement surface executed at speeds of up to 110 km/h. Large areas of low
density asphalt zones (stripped asphalt areas) were determined in the existing asphalt
material indicating the material had prior defects (Resource International, Inc., 2013).
These defects were subsequently overlaid during the rehabilitation of the roadway.
Asphalt stripping is defining as moisture related mechanism that deteriorates the bond
between asphalt and aggregate, with a resulting low-density zone. A stripped area
shows up on a radar survey as an out-of-phase reflection peak because the radar wave
is propagating from a higher dielectric material to a lower one. The low dielectric is a
function of the low density of the dry stripped material. The GPR scan of Figure 1
shows 12 meters stripping zone and the corresponding out-of-phase signal reflection
peak.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
9
Figure 1: Detection of stripping zones
Additionally, based on the estimated dielectric constants, the new and existing asphalt
material had low moisture content and existing aggregate base and subgrade were
generally dry with some pockets of moisture in the existing asphalt.
GPR uses radio waves as an energy source which are transmitted into the pavement
and reflected at layer interfaces. When the radar waves pass through a medium other
than a vacuum, the velocity of propagation becomes a function of the dielectric
constant of the medium. The moisture content has a large influence on the dielectric
constant, and therefore affects the two-way travel time, so that the greater the amount
of water saturation, the lower the radar wave velocity, and hence the higher the
dielectric constant of the material. The dielectric constant of air is 1, while the
dielectric constant of water is 81. Asphalt is considered free of moisture if its
calculated dielectric constant varies between 3 and 6 (David, 2004). Dielectric
constants of 7 or above indicate a moist asphalt material.
INDEPENDENT REVIEW AND INVESTIGATION
In December 2013, as part of the independent review and investigation, a site visit
and project documentation review of the rehabilitation design and construction
inspection was completed.
Rehabilitation design review
A review of the design procedure for the pavement rehabilitation indicated that the
design was based on the evaluation of the FWD data, pavement cores, and visual
examination of the project. The FWD data provided the resilient moduli of the
existing pavement layers and was used to determine the pavement strength and
required rehabilitation. The design was intended to address areas with poor moduli,
areas of known pavement damage, and restore the International Ride Index (IRI) to
acceptable limits.
The design team reported that, based on the FWD data, the existing pavement
structure was generally in good condition, and the pavement surface exhibited typical
12 meters stripping zone
Out of phase reflection
peak
Bottom of asphalt
Layer depth in mm
Distance in m
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
10
signs of fatigue including alligator cracking and some areas of bleeding. The team
further indicated that in general there was nothing unusual about existing pavement
condition or pavement cores. They recommended drainage improvements in areas
that were showing premature failures. The design process was standard and
applicable based on the data obtained during the pre-design phase.
Construction inspection documents review
Based on the review of the project documents the pavement rehabilitation project
consisted of three specific construction activities;
Type A (50.8 mm of cold milling and 50.8 mm surface asphalt course).
Type B (76.2 mm cold milling, 25.4 mm leveling asphalt course, interlayer
reinforcement system (25.4 mm x 25.4 mm geogrid), 50.8 mm surface asphalt
course).
Type C (full depth removal of existing asphalt pavement, 203.2 mm of lean
cement concrete pavement, 101.6 mm of base asphalt course, 50.8 mm of
asphalt surface course).
The Type A rehabilitation was used widely throughout this project. The Type B
rehabilitation, utilizing the geogrid, was also used throughout the project; however,
the Type C rehabilitation was used in very limited locations.
The surface course was an owner/agency approved mix design for Warm Mix Asphalt
(WMA). Based on the mix design information provided, the WMA was produced
with the addition of Rediset WMX. Wetfix 312 was added as anti-strip chemical.
Both of these chemicals are proprietary and were added to the PG 64-22 binder at the
terminal. The mixed binder was then shipped to the asphalt plant by tanker trucks.
Contractor-produced limestone and “quarry sand” was used for the coarse and fine
aggregates with the addition of coarse sand from another source. Because of possible
future litigation, there were no data made available to the authors regarding the
physical or chemical properties of the fine or coarse aggregates.
It was reported that, prior to January 2012, cold milling was typically performed three
to five days ahead of the paving operation but there were many sections that were
open substantially longer, up to 25 days. It was also reported that the cold milled
surface appeared to be sound even after being exposed to traffic. There were some
instances where additional milling was required as well as the placement of a
levelling course to correct unsound existing asphalt surface. However, these areas
were reported as limited. It was noted that no moisture was observed on the milled
surface immediately after milling or prior to paving.
The owner\agency’s project records showed that the cold milled surface was power-
broomed prior to the application of a tack coat. The tack coat application was
reported as “good” with complete coverage obtained. The tack coat was allowed to
break prior to paving, but tack pick up by the paver tires and asphalt trucks was noted.
The project records did not include tack coat application rate or cure time. The
project specifications required application at a rate of 0.30 to 0.45 liter per square
meter. The project records did not provide detailed information regarding actual
application rates.
The mix production temperature was reported at 127°C to 138°C with the mix
placement temperature reported at 113°C to 135°C. The coarse aggregate was
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
11
reported to have been kept under cover at the asphalt plant, but the fine aggregate was
not. The asphalt plant was a standard counter flow drum mix plant with a cyclone-
type dust control unit and baghouse to capture fines. Fines were introduced into the
mixture from the baghouse and the asphalt was stored in silos during production.
Records for the operation of the asphalt plant were not available for review due to
possible litigation, nor was it possible to visit the production facility. Release agents
in truck beds were reported to be a mineral oil or paraffin type coating and there were
no reports of its overuse.
The contractor used a Material Transfer Vehicle (MTV) between the end dump
asphalt delivery trucks and the paver. A standard wheeled paver was used for
placement. The mat during placement was reported to be uniform. However it was
reported that the mixture was tender, internally unstable mixture that displaces
laterally and shoves rather than compacts under roller loads, as evidenced by a bow
wave in front of the rollers. There were times that the rollers lagged far behind the
paver; which can indicate that the roller operators felt the material was tender. Rain
was reported during the paving of 10 of the 41 lots. The contractor power-broomed
the surface to remove any moisture and paving was resumed when possible.
During production quality control cores were extracted from the completed pavement
lots and tested to determine in-place compaction, gradation, and asphalt binder
content for payment. Based on the provided data, the average lot in-place compaction
met the requirement of 92 to 97% in all but two lots. However, a review of the sub-
lots showed there were twenty-two locations where the in-place compaction was not
met.
Site visit observations
During the December 2013 site visit, all visual observations were performed from a
vehicle or from the pavement shoulder because of the very high average daily traffic.
Potholes were the primary distress in the new overlay and were not located along the
longitudinal construction joint. There seemed to be a generalized pattern of potholes
forming approximately 0.3 to 0.6 m offset from longitudinal construction joints.
Areas of surface staining were observed in locations with and without potholes and
appeared to be consistent with the deposition of fines from the asphalt mix to the
surface through a “pumping” action. On some ramps, shoving of the new overlay was
observed. The depth of potholes appeared to be the same as the thickness of the new
overlay. The new overlay material appeared to be delaminated from the milled
surface and there was loose material in the associated pothole. The distresses
appeared consistent with stripping; there was either a cohesion failure between the
aggregate surface and asphalt binder or an adhesion failure within the asphalt binder
itself. It was also noted that there were areas where asphalt binder material was
visible on the surface of the new overlay as fat spots (isolated areas where the asphalt
cement came to the surface of the mix versus bleeding which is characterized by
flushed longitudinal streaks in the wheel path.)
Potholes were observed in lower elevation areas and at higher elevations such as a
ramp on embankments. The project area is part of a coastal plain and a portion is
adjacent to a canal widening project that is to be used for flood control. During a
rainfall event, the lower elevation areas of the project showed evidence of ponding
water in ditches. In some areas, water remained ponded in ditches, while other areas
appeared to drain.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
12
Discussion of testing and data review
Based on the design, construction, and post-construction documents and the available
data, the distressed areas in the new asphalt surface overlay appeared consistent with
a pavement that has stripped; where the bond between the aggregate and the binder,
or the bond within the binder itself, has failed in the presence of water.
The contractor and the construction inspection firm claimed that the infiltration of
surface water through the new asphalt overlay was not a cause of the stripping and
pointed to the testing performed on two control sections, one distressed section, and
the laboratory produced samples (“pills”) from the mix design. This testing included
permeability using Florida Department of Transportation Test Procedure FM 5-565;
Tensile Strength Ratio (TSR) Test Procedure AASHTO T-283 to determine potential
for stripping reported as tensile strength ratio (TSR); and AASHTO T-324 (Hamburg
Wheel-Track Testing of Compacted Hot Mix, HMA) to determine susceptibility for
premature damage reported as Stripping Inflection Point (SIP). The first control
section was labeled as “initial” and the second as “final”. The “final” control section
was selected as representative due to the presence of cracks in the existing asphalt
pavement layer in the “initial” control section. Table 1 summarizes the test results.
Table 1: Section 1 Summary of Test Results by Contractor and Owner/Agency
Test
Section.
ID
Locatio
n
Layer
Acceptanc
e Criteria
Results
Permeabilit
y per FM5-
565
Final
Control
KM 4.4
Lane 4
WB
Surface
- New
125 cm/s
x10-5
0 cm/s
x105
Base
AC -
New
0 cm/s
x105
In-situ
Moisture*
Distresse
d
KM
3.23
Lane 3
EB
Surface
- New
NA
58%
Base
AC
existin
g
75%
Initial
Control
KM
2.68
Lane 3
EB
Surface
- New
NA
27%
Base
AC
existin
g
Core
crumble
d – no
test
*% Saturation = {( % Moisture by Wt. x Bulk SG)/ % Air Voids } x 100
Test
Area
Locatio
n
Layer
Dry
Strength
Criteria,
psi
Dry
Strengt
h, psi
Wet
Strengt
h, psi
TSR
Criteria
, %
TSR
, %
AASHTO
T-283
Final
Control.
KM 4.5
Lane 4
WB
Surface
- New
80
124
119
75
96
137.8
143.4
104
Initial
Control
KM
2.68
Lane 3
EB
99
105
105
Distresse
d
KM
3.23
Lane 3
EB
79
40
51
93
30.8
33
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
13
Final
Control
KM 4.5
Lane 4
WB
Base
AC -
New
70
126
107
65
85
Initial
Control
KM
2.68
Lane 3
EB
Base
AC -
Existin
g
NA
179
135
NA
75
Distresse
d
KM
3.23
Lane 3
EB
117
88
75
Design
Lab
Surface
- New
80
135
131
75
97
Test
Area
Locatio
n
Layer
Acceptanc
e Criteria,
TX
Rut
Passes
SIP
Passes
AASHTO
T-324
Design
(AC
Mix)
Lab.
12.5
mm
10,000
9,412
6,083
Final
Control
KM 4.4
Lane 4
WB
Surface
- New
10,973
6,667
Design
(AC mix)
Lab
19 mm
16,333
10,250
Final
Control
KM 4.4
Lane 4
WB
Base
AC -
New
20,000
No SIP
Evaluating these test data, the contractor and the construction inspection firm
concluded that moisture could not be entering the pavement from the top and
infiltrating into the base materials, and thus had to be a result of saturation of the
existing asphalt material, the existing aggregate base, and the subgrade from the
bottom or sides. This theory included water infiltration from the nearby canal,
flooding of ditches and wetting of base material, and a substandard underdrain system
that does not drain water away from the pavement. In all of these cases, it would be
possible for water to enter the subgrade and potentially travel upward, particularly
during flooding situations and due to the proximity to the canal dredging project. The
data presented by the contractor and the construction inspection firm were not
conclusive for the following reasons:
1. Permeability of the surface course in the distressed section was not tested.
Permeability testing was performed on a control section with a corresponding
higher average in-place compaction (Lot Production Number, LPN 33, IPC =
96%) than other lots that also have not shown signs of distress. For example,
LPN 15 had an average in-place compaction of 91.1% and testing could reveal
the pavement to be more permeable than reported from the testing of LPN 33.
The contractor did not test the existing asphalt mixture layer for permeability.
2. The contractor felt that the TSR data from the distressed section was not
relevant due to the presence of water and cracks in the pavement layers that
would skew the test results. However, since the location of the cores is near an
area of potholes and stripping, it would follow that the low TSR results could
indicate susceptibility to water damage or stripping.
3. The in-situ moisture (% Saturation) is inconsistent with the results of the GPR
testing performed in October 2013. Overall, the GPR test data indicated that
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
14
the asphalt material (new and existing) typically had low moisture content,
and the aggregate base material was generally dry. Dielectric constants in two
areas showing potholes were evaluated to assess the potential for trapped
moisture in the asphalt layers; the distressed section (LPN 14) in lanes 3 and 4
between km 3.389 and km 3.445 that was used in the contractor’s testing and
in another section (LPN 6). As can be seen in Table 2, the dielectric constants
indicate some moisture in both the new surface course and existing asphalt
base material. It is interesting that the higher dielectric constant in LPN 6
coincided with the greater number of days the cold milled surface was open to
traffic (19 days) as compared to LPN 14 (2 to 5 days).
Table 2: Dielectric Constants from GPR Testing
Location
Lane
LPN
Average Dielectric
Constant - Surface
Average Dielectric
Constant – Existing
Asphalt Base
Km 3.389 –
3.438 EB
3
14
6.02
6.53
4
14
6.22
6.63
Km 4.478 –
4.438 EB
5
6
6.83
7.31
Ramp
Lane
6
6.70
7.10
A review of the pavement core data from 2012 (by others) shows that the aggregate
base is mostly an AASHTO A-2-4 to A-2-6 material that could become saturated
under flooded conditions but it is granular and should quickly drain as water recedes.
A confining layer of clay, such as an AASHTO A-7-6, could trap water in the
aggregate base material. However, GPR test results did not confirm that the
aggregate base layer was saturated at these locations.
Additionally, the Hamburg Wheel testing was performed on cores extracted from the
final control section (see Table 1); the new asphalt base material did not indicate
stripping, based on the reported Stripping Inflection Point (SIP), the new surface
asphalt did indicate stripping with an SIP of 6,667 passes (3,333 cycles). The
industry standard and specification used by most state DOTs is 10,000 passes (5,000
cycles).
The asphalt production process for the WMA surface mixture was not available and
there remain questions regarding the potential for dusty/dirty aggregates or residual
aggregate moisture that could be a cause of the stripping. The mineralogical
properties of the aggregate are not known and there are no data available. It was
stated that the coarse aggregate was kept under cover but the fine aggregate was not.
Fine aggregate moisture can range from 0 to 16% due to rain and dry cycles. The
owner/agency could not provide access to the asphalt production plant or plant
records due to possible litigation regarding this project.
2013 core evaluation
Additional cores were obtained by the owner/agency from LPN 15 and represent the
following distressed and non-distressed areas. (Rehabilitation types were described on
page 3):
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
15
1. Area with visible distress at a relatively low pavement elevation area. This
area was rehabilitated using Type B.
2. Area with visible distress at a relatively higher pavement elevation area. This
area was rehabilitated using Type B.
3. Area without visible distress. This area was rehabilitated using Type A.
All cores were taken from the eastbound left lane between approximate km 2.4 and
2.8. A total of seven cores were taken in each area. Six 100 mm diameter cores and
one 150 mm diameter core were obtained. The 150 mm diameter core was a full-
depth sample of the pavement from surface to base or to subgrade material. The cores
were examined and the following tests were performed.
Moisture susceptibility testing, Tensile Strength Ratio (TSR) (AASHTO T
283, 2014).
In-place air voids content (ASTM D2726, 2014).
Gradation, Method A (ASTM D2172, 2011).
Percent binder (asphalt) content (ASTM D6307, 2010).
Percent passing sieve #200 (75 μm), Method A (ASTM D2172, 2011).
Visual examination of the cores
The cores were visually examined. Cores in Areas 1 and 2 had at least three visible
distinct layers of asphalt mixture. These layers were: (1) top layer of new surface
course mixture, (2) second layer of new leveling course mixture, and (3) third layer of
old asphalt mixture, which typically included more than one layer of asphalt mixture.
Cores in Area 3 had a visible new surface course mixture and an old asphalt mixture
that included surface, intermediate, and base courses. All asphalt below a new surface
course or leveling course was designated as “old asphalt base” mixture. Table 3
provides an overview of the visual observations of the cores.
Table 3: Summary of 2013 Core Defects
DEFECT
Delamination between new
asphalt courses
Delamination between new
and exisitng asphalt
courses
Stripping evident
CORE AREA
17 of 7 cores 2 of 7 cores 3 of 7 cores
27 of 7 cores 3 of 7 cores 1 of 7 cores
30 of 7 cores 3 of 7 cores 0 of 7 cores
2013 laboratory test results
The cores were tested in the laboratory to determine the in-place characteristics of the
new asphalt surface mixture and old asphalt base mixture. The results of these tests
are summarized below.
Percent air voids content
Percent air voids content and bulk specific gravity of new surface mixture samples
and old asphalt base course mixture samples were measured in the laboratory. The
results are summarized in Table 4 for cores obtained from Areas 1, 2, and 3
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
16
Table 4: Summary of Laboratory Test Results, % Air Voids Content
Area
Location
No. of Samples
Tested
Avg. Bulk Sp.
Gr.
Avg. Air Voids
Content, %
1
New Surface Mixture
6
2.42
4.82
Old Base Course
Mixture
6
2.28
10.13
2
New Surface Mixture
6
2.41
5.43
Old Base Course
Mixture
6
2.35
7.19
3
New Surface Mixture
6
2.38
6.42
Old Base Course
Mixture
6
2.37
6.28
Gradation, Percent Passing Sieve #200, and Percent Binder (Asphalt) Content
The gradations of asphalt mixtures used in the new asphalt surface mixture were
determined from the cores of Areas 1, 2, and 3. The results of these tests are
summarized in Figure 2. The values in the last column “JMF” (Job Mix Formula)
refer to the values that were approved by the owner/agency for the Marshall design
Warm Mix Asphalt (WMA).
Figure 2: Surface Mix Gradations of Areas 1, 2, 3 and JMF.
The percent binder (asphalt) content and percent Passing Sieve # 200 (dust
determination) were also determined. The results of these tests are listed in Table 5.
The following results are from the quantitative extraction method. The ignition
method was used and provides slightly different results as shown in parentheses. The
extraction method is considered more accurate.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
17
Table 5: Results of Percent Binder (Asphalt) Content and F/A Ratio, New
Surface Course Mixture
Test Area
% Asphalt Content
% Passing
Sieve #200 (Fine)
F/A (Fines/Asphalt) Ratio
1
6.7 (6.49)
2.7 (3.8)
0.40 (0.58)
2
5.9 (6.18)
2.4 (3.9)
0.41 (0.63)
3
6.0 (6.13)
2.4 (4.2)
0.40 (0.56)
JMF
5.52*
6.5
1.31**
*Does not include WMA Additive and Antistripping agent.
**Dust to Effective Asphalt Ratio.
Tensile Strength (Unconditioned and Conditioned)
Indirect Tensile Strength tests were performed on core samples obtained from all
three areas (1, 2 and 3). For this purpose, three cores of each area were tested
unconditioned (dry) and three cores were tested after conditioning the cores according
to the procedure described in AASHTO T 283. The results of these tests are
summarized in Table 6.
Table 6: Results of Indirect Tensile Strength Tests New Surface Course (Top)
and Old Asphalt Base Course (Bottom) Mixtures
Test Area
Average Strength,
Unconditioned (Dry) Samples,
MPa
Average Strength,
Conditioned Samples,
MPa
Strength Ratio
= (Col. C/Col. B)
Column B
Column C
Column D
1-Top
0.83
0.72
0.87
2-Top
0.72
0.53
0.74
3-Top
0.72
0.60
0.84
JMF
0.90
0.90
0.97
1-Bottom
0.87
0.49
0.56
2- Bottom
0.84
0.53
0.64
3- Bottom
106
0.71
0.67
DISCUSSION OF 2013 LABORATORY TEST RESULTS
The results of tests performed on the cores obtained from three areas (1, 2 and 3) are
as follows:
1. The gradation tests performed on the new surface course mixture (see Table 4)
indicated that the aggregates of the mixtures in all three areas had similar
gradation characteristics with little or no variability. The gradations of the
tested cores met the JMF requirements with the exception of percent passing
the #200 sieve. A low percent passing the #200 sieve can result in a tender
mixture that is highly permeable.
2. Average bulk specific gravity (BSG) of the surface mix ranged from 2.42 to
2.38 and average percent air voids content ranged from 4.8 to 6.4 (see Table
3). These results indicated that Area 1 with visible failures and relatively low
elevation had the maximum average BSG and lowest average air voids
content (highest average in-place compaction of 95.2%), whereas, Area 3 with
no visible failure had the lowest average BSG and highest average air voids
content (lowest average in-place compaction of 93.6%). In all cases, the
average in-place compaction of the surface mixture in all areas ranged from
93.5% to 95.1%; all are within specification and not statistically different.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
18
3. The lowest single in-place compaction result for the new surface course was
91.7% in Area 3 and was only slightly out of specification.
4. Average bulk specific gravity (BSG) of the old asphalt base mixture ranged
from 2.28 to 2.37 and average percent air voids content ranged from 10.1 to
6.3 (89.9 % to 93.7% average in-place compaction) (see Table 4). These
results indicate that Area 1 with visible failures and relatively low elevation
had the lowest average BSG and highest average air voids content with an
average in-place compaction of 89.9%, whereas, Area 3 with no visible failure
had highest average BSG and lowest average air voids content with an
average in-place compaction of 93.7%. On the basis of these observations, the
development of premature failure may be related to lower BSG and higher air
voids content (lower in-place compaction) in the old asphalt base mixture.
5. The discussion of Percent Binder (Asphalt) Content and the Percent Passing
Sieve #200 is based on the quantitative extraction method. The Percent Binder
(Asphalt) Content of the new surface mixture ranged from 5.9 to 6.7 as
compared to the JMF of 5.52. The Percent Binder (Asphalt) Content of 6.7
from Area 1 did not meet the specification tolerance of ± 0.52 (5.52 + 0.52 =
6.04). The Percent Passing Sieve #200 of the surface mixture ranged from 2.4
to 2.7 as compared to the JMF of 6.5 and did not meet the specification
tolerance of ± 3 (6.5 – 3 = 3.5). Note that these data are based on unwashed
samples. It is noted that LPN 15 which represented the placement quality
control testing for Areas 1, 2, and 3, did not indicate that the F/A ratio was out
of tolerance.
6. The average unconditioned (dry) Indirect Tensile Strength (ITS) of the new
surface course mixture was 0.83 MPa, 0.72 MPa, and 0.72 MPa in Areas 1, 2,
and 3 respectively. These results are lower than the JMF value of 0.90 MPa
but met the specification requirement of 0.55 MPa (see Table 6). Also, in Area
1, where the pavement elevation was relatively low and premature failures
were observed, the average unconditioned ITS was higher than the average
unconditioned ITS of Area 2 (higher elevation and visible failures) and Area 3
(no visible failures). The average conditioned (wet) ITS of the new surface
course mixture was 0.72 MPa, 0.53 MPa, and 0.60 MPa in Areas 1, 2, and 3,
respectively. These results are lower than the JMF value of 0.90 MPa. All
but Area 2 met the specification requirement of 0.55 MPa (see Table 6). Also,
in Area 1, where the pavement elevation is relatively low and premature
failures were observed, the average conditioned ITS was higher than Area 2
and Area 3. The ITS Ratio of new surface mixture (see Table 6) ranges from
74% to 87% as compared to the JMF value of 96.87%. All but Area 1 met the
specification requirement of 75%. These ITS test values indicate that the new
surface course is not a mix that is particularly susceptible to moisture induced
damage. However some evidence of stripping is visible in the cores.
The ITS test data for old asphalt base mixture indicated that the average
unconditioned ITS values for all three areas were higher than the corresponding JMF
values for the new surface mixture (see Table 6). The average conditioned ITS
values were lower and therefore resulted in a lower ITS Ratio of 56% to 67%
indicating that the old asphalt base mixture is susceptible to moisture induced
damage. Stripping in the old asphalt layer was confirmed using GPR imaging and is
visible in some of the extracted cores.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
19
CONCLUSIONS
Based on the GPR test results; review of the project records and test data; visual
observation of pavement cores; and testing performed on the cores of the new surface
course mixture and the old asphalt base course mixture, our findings are as follows:
1. The existing asphalt base material had significant areas of stripping as
determined using GPR. After milling, existing asphalt base material may have
absorbed significant amounts of moisture when it was exposed to surface
water runoff and traffic for several days before overlaying:
A review of the available construction records for Areas 1, 2, and 3
(LPN 15) indicated that the areas were open to weather and traffic
for approximately six to eleven days.
GPR and core testing showed that the existing asphalt base mixture
had areas of relatively high air void content and other damage
related to stripping (as visible in some cores) that would allow
surface water runoff to enter the pavement matrix prior to the
overlay application.
The highest air voids content in the existing asphalt base layer
were measured in Area 1, which exhibited the greatest amount of
surface failures; lesser air voids contents were measured in the
existing asphalt base layer in Area 2 and corresponded to lesser
amounts of surface failures; and the least air voids content in the
existing asphalt base layer measured in Area 3 corresponded to no
visible surface failures.
Additionally infiltrated surface water that was trapped in the old
asphalt base layer would condense with temperature variations and
create pressure within the matrix that could cause distress in the
new surface or leveling courses, eventually allowing potholes to
form and thus allowing additional water migration into the existing
pavement layer(s). Also, paving over an existing damaged asphalt
surface can manifest in future pavement failures even if moisture is
not an issue.
2. Considering the inconsistency of premature surface failures in the different
coring areas, it is possible that Area 1 absorbed more moisture than Area 2
due to the elevation difference. Area 2 is at a relatively higher elevation;
therefore, rainfall on this area would run off faster than rainfall on Area 1. In
both Areas 1 and 2, there were old asphalt base failures that would allow
surface water infiltration which could result in damage to the new pavement
courses. The old asphalt base of Area 3 was comparatively denser than the
other two areas, having the lowest air voids content. Additionally, the old
asphalt base cores from this Area 3 were intact and did not have significant
damage. Therefore, the infiltration of surface water runoff into the old asphalt
base is less and the area had no visible failures.
3. Delamination without stripping was observed in cores from Areas 1, 2, and 3.
The delamination appeared in Areas 1 and 2 at the surface/leveling course
interface (at the location of the geotextile), at the leveling course/old asphalt
base course; and at the surface/old asphalt base course in Area 3.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
20
Delamination without other apparent distress mechanisms generally occurs
due to the lack of an adequate bond between layers. Without a secure bonding
face, the pavement structure does not act as a unit and the individual layers
must resist shear stressed induced by traffic. Tack coat between the layers at
delaminated areas was difficult to see. It is possible that tracking and pick up
of the tack coat under construction traffic was a factor in these failures.
Allowing the milled surface to remain open to traffic and precipitation over a
period of days can also create problems with cleanliness and dampness that
will inhibit a good tack bond. It is additionally possible that the “tack coat” for
the interlayer reinforcement system (geogrid) was not as specified. The
project specifications required the use of the “same grade asphalt cement as
used in the hot mix asphalt”. Records for the material used in this application
could not be found. The delamination issue could exacerbate potential
problems with surface water runoff (as postulated in Items 1 and 2 above).
Delamination or de-bonding would allow water to infiltrate into the pavement
structure as the weak pavement layer is peeled away by the action of traffic.
Water would move along the layer interface causing further delamination and
the potential for stripping. Based on the available records, it is unlikely that
the mixture design itself, as produced, is a cause of premature pavement
failure. Field compaction of the mixture appears consistent in Areas 1, 2, and
3. It is improbable that the source of water is from the subgrade, as was
proposed by the construction contractor and the construction inspection
company. The geotechnical firm hired to extract cores reported that they did
not observe free water in any of the core holes. This is consistent with GPR
testing that found no excess moisture in the subgrade or base materials but
some moisture in the asphalt pavement layers.
In summary, and based on the mechanisms of premature failure outlined above, the
causes of premature failure are: (1) surface water infiltration into the milled surface
prior to overlaying; (2) questionable structural integrity of the milled surfaces of old
asphalt base material (stripping); (3) delamination/de-bonding due to tack coat
placement issues.
RECOMMENDATIONS
Based on our findings from this study as well as review of previous studies, it was
recommended to the owner/agency that for the constructed project:
1. The correction of areas of pavement distress requires full depth repairs. The
majority of failures from Areas 1, 2, and 3 were located in the wheel path. In
addition to surface water infiltration during construction, there are areas of
fatigue cracking and other existing asphalt base deficiencies (including
stripping) that need repair. These areas will need to be located based on
current observations and any other pavement records and past pavement
surveys. GPR and FWD can be used to assess the areas of significant failures
under the overlay.
2. The existence of underdrains should be determined and if they do not exist or
are not operational at this time, they should be installed. Functional
underdrains can help outlet water that is trapped in pavement layers.
The International Journal of Pavement Engineering and Asphalt Technology (PEAT) ISSN 1464-8164.
Volume:17, Issue:2, December 2016
21
Based on our findings from this study as well as review of previous studies, it was
recommended to the owner/agency that for the future projects:
1. Because of the local weather conditions (frequent rainfall) and the known
potential for stripping in the existing asphalt the use of GPR and targeted
FWD testing can help locate low density areas for additional evaluation in the
design phase.
2. Because of the local weather conditions (frequent rainfall) projects which
require milling and overlay should be milled only in lengths that can be paved
almost immediately so that the milled surface is not left open to rain and
traffic for any significant length of time. Also, the damaged areas which may
become visible after milling should be appropriately repaired (partial or full
depth) before overlaying.
3. The use of trackless tack should be considered to avoid tracking and pickup of
the tack coat to prevent potential delamination problems.
REFERENCES
AASHTO T 283-14 (2014). Standard Method of Test for Resistance of Compacted
Asphalt Mixtures to Moisture-Induced Damage.
Anon (2003). “Moisture Sensitivity of Asphalt Pavements”. A National Seminar,
Transportation Reasearch Board, San Diego, CA.
ASTM D2172 / D2172M – 11 (2011). Standard Test Methods for Quantitative
Extraction of Bitumen From Bituminous Paving Mixtures
ASTM D2726/2726M-14 (2014). Standard Test Method for Bulk Specific Gravity
and Density of Non-Absorptive Compacted Bituminous Mixtures.
ASTM D6307 - 10 (2010). Standard Test Method for Asphalt Content of Hot-Mix
Asphalt by Ignition Method.
Bush, A. J. and Baladi, G.Y. (1989). “Nondestructive Testing of Pavements and
Backcalculation of Moduli”. American Society for Testing and Materials.
Philadelphia, PA.
David, J. D. (2004). “Ground Penetrating Radar”. The Institution of Electrical
Engineers, UK, 2nd Edition.
Kandhal, P. S. and Richards, I. J. (2001). “Premature Failure of Asphalt Overlays
from Stripping: Case Histories”. Annual meeting of the AAPT. Clear Water, FL.
Parker, F. (1989). “A Field Study of Stripping Potential of Asphalt Concrete
Mixtures”. Highway Research Center, Auburn University, Auburn, AL.
Resource International, Inc. (2013). “Premature pavement failure”. Report No. W-13-
110. Columbus, OH.
Scullion T. (2004). ICHPD Workshop on GPR Implementation. Columbus, OH.