International Journal of Pavement Engineering

ISSN: 1029-8436 (Print) 1477-268X (Online) Journal homepage: http://www.tandfonline.com/loi/gpav20

Effects of using screening materials in the graded

aggregate base layer of flexible pavements
Sung-Hee Kim, Jidong Yang & Jayhyun Kwon
To cite this article: Sung-Hee Kim, Jidong Yang & Jayhyun Kwon (2015): Effects of using
screening materials in the graded aggregate base layer of flexible pavements, International
Journal of Pavement Engineering, DOI: 10.1080/10298436.2015.1036868
To link to this article: http://dx.doi.org/10.1080/10298436.2015.1036868

Published online: 29 Apr 2015.

Submit your article to this journal

Article views: 24

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=gpav20
Download by: [Deakin University Library]

Date: 07 January 2016, At: 19:53

International Journal of Pavement Engineering, 2015

http://dx.doi.org/10.1080/10298436.2015.1036868

Effects of using screening materials in the graded aggregate base layer of flexible pavements
Sung-Hee Kima, Jidong Yanga* and Jayhyun Kwonb
a

Civil and Construction Engineering, Kennesaw State University, 1100 South Marietta Parkway, Marietta, GA 30060, USA; bApplication
Technology Manager, Tensar International Corporation, Alpharetta, GA 30009, USA

Downloaded by [Deakin University Library] at 19:53 07 January 2016

(Received 2 December 2014; accepted 29 March 2015)

In this paper, the potential use of screening materials in Graded Aggregate Base (GAB) layer of pavements was investigated
using aggregate sources in Georgia. Three content levels of screening materials in GAB, i.e. 0%, 25% and 50%, were
studied. Morphological analysis of aggregates and Proctor test were conducted to reveal the characteristics of the modified
GAB materials, followed by the California Bearing Ratio (CBR) test and Mechanistic-Empirical Pavement Design Guide
software simulation to evaluate how the GAB strength and overall pavement performance change with varying amount of
screening materials in the GAB layer. The results indicated that the effect of using screening materials in GAB varies and
depends on the amount of screening materials used, which changes gradation, and the aggregate types and sources, which
exhibit significantly different morphological properties.
Keywords: graded aggregate base; screening materials; proctor test; CBR test; morphological indices

Introduction
The ever-increasing cost of oil and gas by-products has
motivated road management agencies and the aggregate
industry to pursue alternative materials and/or construction technologies to reduce the cost for pavement
rehabilitation and construction. Over years, the Georgia
aggregate industry has stockpiled several million tons of
screening materials (No. 810 and M10), which are
underutilised finer aggregates for pavement construction.
This staggering amount of screening materials provides a
potential and less expensive source for pavement
construction in Georgia. The main objective of this study
was to investigate the potential impacts of using No. 810
screening materials in a Graded Aggregate Base (GAB)
layer, and in particular, how GAB strength and pavement
performance change with varying amount of No. 810
screening materials in the GAB layer. As such, a practical
amount of screening materials can be determined and
adopted for pavement construction and rehabilitation in
Georgia.
In consultation with the Georgia Department of
Transportation (GDOT) and Georgia Construction Aggregate Association, and considering the States current
practices, two aggregate groups, limestone (Group I) and
granite (Group II), were selected for this study. One source
(013C) was chosen for Group I and three sources (028C,
048C, and 158C) were chosen for Group II. The source
locations are indicated in Figure 1.
To ascertain the morphological difference of aggregates
across sources selected, University of Illinois Aggregate
Image Analyzer (UIAIA) was utilised to obtain three

*Corresponding author. Email: jyang12@kennesaw.edu

q 2015 Taylor & Francis

morphological indices, i.e. angularity index (AI), flat and

elongated (FE) ratio and surface texture (ST). Statistical
analyses were performed to determine if the morphological
properties are significantly different between the aggregate
types and among the sources. Besides the morphological
properties, particle size distribution was examined against
the gradation requirement established by the GDOT. Then,
Proctor tests were followed to assess the optimum moisture
content and maximum dry density. Finally, California
Bearing Ratio (CBR) tests were conducted to evaluate the
strength of GAB. The results of morphological and
gradation analyses were correlated with the Proctor and
CBR test results to discern the effect of aggregate types and
sources, and the amount of screening materials on the
strength of resulting GAB.
Additionally, since the GAB layer is a constitutive
component of pavement structure, the effect of varying
content of screening materials was further evaluated by
simulating overall pavement performance in terms of
alligator cracking, permanent deformation and international roughness index (IRI), using MechanisticEmpirical Pavement Design Guide (MEPDG) software.

Literature review
Previous research indicates that the resilient behaviour of
granular materials is associated with the degree of
saturation, dry density, fines content, aggregate gradation,
aggregate type and shape. For mix design, it is extremely
important to consider how the strength changes when these
factors vary in content. This section provides a review of

S.-H. Kim et al.

Downloaded by [Deakin University Library] at 19:53 07 January 2016

aggregate base. On the other hand, Thom and Brown

(1987) showed that dry density has relatively insignificant
effect for a crushed-limestone road base. The effect of dry
density varies by aggregate types, fine contents and stress
state (Haynes and Yoder 1963). Hicks and Monismith
(1971) found that dry density plays a more important role
for the partially crushed aggregates than for the fully
crushed aggregates and the effect of dry density decreases
as the fine content increases.

Figure 1.

GAB source locations.

representative previous studies on the effect of those

factors.

Degree of saturation and dry density

Several researchers demonstrated that the effect of the
degree of saturation on resilient behaviour of granular
materials varies by aggregate type, gradation, stress state
and fine content (Hicks and Monismith 1971, Barksdale
and Itani 1989, Dawson 2000). In particular, Dawson
(2000) studied the behaviour of granular materials with
high degree of saturation and found that the resilient
modulus of granular materials decreases when approaching the complete saturation. When exceeding the
optimum moisture content, the stiffness decreases rapidly
due to the development of excess pore pressure. Haynes
and Yoder (1963) observed a 50% reduction in resilient
modulus of gravels when the degree of saturation varied
from 70 to 97%. Raad et al. (1992) demonstrated that the
moisture content has significant effect on well-graded
materials with high proportion of fines because the water
has better chance to be held in the pores in such gradation
while the water can drain or infiltrate freely in open
gradation.
The increase in dry density or the degree of
compaction of aggregate materials makes the aggregate
medium stronger and stiffer. Kolisoja (1997) showed that
the effect of dry density or degree of compaction is a
significant influencing factor for the stiffness of unbound

Gradation and fine content

A change in aggregate gradation effects a change in
moisture content and dry density in GAB, and thus affects
the stiffness. Kim (2004) investigated the effect of
gradation on resilient modulus and found that the opengraded limestone had a higher resilient modulus while no
significant changes were observed for gravel. Thom and
Brown (1987) reported that the stiffness generally
decreases as the fine content increases. Hicks and
Monismith (1971) found that the resilient modulus
decreases as fines content increases for partially crushed
aggregates, but for fully crushed aggregates, the trend
reverses. They also found that stiffness initially increased
and then decreased as fines were added to the crushed
aggregates. They argued that the initial increase in stiffness
was due to the increase of the contacts as voids were filled
with fines and the decrease in stiffness afterward was due to
the displacement of coarse particles as excess fines were
added, which results in the loss of aggregate particle
interlocks and thus the load-carrying capability lies only on
the fines. As such, it could be inferred that the aggregate
gradation and amount of fines impose an indirect effect on
the stiffness of GAB by changing the moisture and density
of the system. A more direct effect of gradation on the
stiffness occurs when the fine particles fill the voids and
influence the interaction among the coarser, angular
particles. This can be visualised in the extreme situations
when one compares a floating matrix where the coarse
aggregate floats in the fines, thus preventing interaction
among coarse aggregates with a lack of fines where only
coarse aggregate interaction provides resistance to movement. The optimal case lies somewhere in between, where
the coarse aggregates and fine aggregates are blended in
such a way to provide optimum density and maximum
particle interaction.
Aggregate type and shape
Aggregate type and shape are significant factors influencing the resilient behaviour of granular materials. The
angular and rough-textured aggregates provide stronger
and stiffer mass by way of better locking, while the
rounded and smooth-textured aggregates tend to slide
against each other. Studies have indicated that the crushed

International Journal of Pavement Engineering

Table 1.
QPL ID

Aggregate source and physical properties.

Aggregate group

Source location

GAB character

LA abrasion (%)

Bulk specify gravity

I
II
II
II

Dalton
Hitchcock
Norcross
Walton County

Limestone
Mylonitic Gneiss
Granite Gneiss
Biotite Gneiss

25
18
45
41

2.702
2.697
2.684
2.640

013C
028C
048C
158C

aggregate, which typically has high angularity and rough

texture, provides higher load-carrying capacity and
stiffness than the rounded gravel (Lekarp et al. 1996).

Downloaded by [Deakin University Library] at 19:53 07 January 2016

Laboratory tests and results

The sources and physical properties of selected aggregates
are presented in Table 1.
Morphological properties
Morphological properties of mineral aggregates are known
to affect pavement performance in terms of strength,
modulus and permanent deformation. Current practices
consider the effects of physical properties of aggregates,
such as shape, texture and angularity, on the strength,
stability and performance of the pavement base layer.
To quantify those properties of selected aggregate
materials, the image analysis device, UIAIA, was utilised
to analyse images and compute three morphological
indices, i.e. AI, FE ratio and ST. The UIAIA uses three
orthogonally positioned cameras to capture three-dimenTable 2.

sional shape properties of aggregates. These indices are

determined based on the particle image outlines obtained
from each of the top, side and front views. First, aggregate
samples were sieved through three sieve sizes: 19, 12.5
and 4.75 mm, and washed and air dried. Then, they were
scanned and particle morphological indices were computed. The low bound of AI is 0, which represents a perfect
sphere or no angularity, while the upper end could reach
700 800 degrees, indicating very high angularity. The ST
index typically exhibits values up to 1 for smooth gravel
with higher values for increasing angularity, crushed faces,
corners and jagged edges in the case of 100% crushed
stone (Al-Rousan et al. 2007). The morphological indices
of aggregate samples are presented in Table 2. As seen, the
computed indices vary considerably across the selected
sources. To distinguish the levels of difference in terms of
these indices, hypothesis tests were conducted and the
results are presented in Tables 3 and 4.
First, the morphological indices were compared
between Groups I and II. In this case, 013C was compared
with each of the three sources (028C, 048C and 158C)
selected in Group II. One-tail t-test was performed to

Morphological index summary.

Retained on 19 mm
Angularity

Retained on 12.5 mm

Retained on 4.75 mm

FE ratio

Surf texture

Angularity

FE ratio

Surf texture

Angularity

FE ratio

Surf texture

Sample ID: 013C

Maximum
598.41
Minimum
319.46
Average
425.38
COV
16.77

4.68
1.52
2.70
32.81

2.66
1.40
1.74
19.27

523.95
279.25
420.11
15.16

6.01
1.88
2.94
32.71

3.52
1.26
2.06
27.97

681.48
306.54
490.71
18.42

3.91
1.39
2.65
25.43

5.38
1.46
3.16
30.58

Sample ID: 028C

Max
587.11
Min
299.19
Avg
458.34
COV
17.79

4.11
1.86
2.91
25.60

3.18
0.97
2.18
27.30

614.64
305.22
466.04
14.50

5.75
1.36
2.88
36.81

3.58
1.26
2.33
25.05

652.01
426.25
524.89
11.76

4.24
1.85
2.62
25.84

5.25
1.76
3.30
30.00

Sample ID: 048C

Max
667.81
Min
267.79
Avg
424.29
COV
22.02

5.50
1.73
3.13
31.74

3.42
1.18
2.10
28.04

716.23
309.36
482.37
20.09

5.20
1.63
2.84
24.02

3.98
1.28
2.47
29.43

635.05
301.52
468.01
18.50

4.73
1.35
2.58
25.82

4.97
1.74
2.89
28.90

Sample ID: 158C

Max
563.84
Min
322.84
Avg
421.08
COV
15.41

4.34
1.58
2.79
25.32

3.08
1.15
1.77
24.01

562.97
262.50
418.40
16.32

4.45
1.66
2.64
27.79

3.63
1.05
1.91
28.42

690.39
310.15
465.27
19.82

5.27
1.56
2.82
29.30

5.63
1.45
2.81
34.13

4
Table 3.

S.-H. Kim et al.

Morphological index differences between aggregate Groups (I and II).
19 mm

12.5 mm

4.75 mm

Morphological index

Between sources

t statistic

p value

t statistic

p value

t statistic

p value

Angularity

013C
013C
013C
013C
013C
013C
013C
013C
013C

1.237
2 0.041
2 0.199
0.733
1.465
0.356
2.628
2.299
0.296

0.225
0.967
0.843
0.469
0.150
0.724
0.013*
0.026*
0.769

3.133
3.116
20.111
20.252
20.501
21.497
2.080
2.560
21.157

0.002**
0.003**
0.912
0.802
0.618
0.139
0.041*
0.013*
0.251

1.392
2 1.158
2 1.182
2 0.134
2 0.458
0.953
0.470
2 1.346
2 1.570

0.170
0.250
0.241
0.894
0.649
0.344
0.640
0.182
0.121

FE ratio
ST

028C
048C
158C
028C
048C
158C
028C
048C
158C

Downloaded by [Deakin University Library] at 19:53 07 January 2016

Notes: *5% significance; **1% significance.

evaluate if the angularity, FE ratio and ST are different

between the two groups. As shown in Table 3, a positive t
statistic value indicates a lower morphological index value
of 013C compared to other sources in Group II. As shown,
for aggregates retained on 12.5 mm sieve, 013C has much
lower angularity (at the 0.01 significance level) and ST (at
the 0.05 significance level) than 028C and 048C. For
aggregates retained on the 19 mm sieve, 013C also has
lower ST (at the 0.05 significance level).
Morphological indices were also compared across the
sources (028C, 048C and 158C) in Group II. For this
comparison as shown in Table 4, F tests were performed.
The smaller p-values with asterisk indicate a rejection of
the null hypothesis that the morphological indices for all
three sources are same. As seen, for aggregates retained on
the 12.5 mm sieve, significant differences were found in
terms of angularity and ST across the sources in Group II.

Particle size distribution

Figure 2 presents the gradation of each aggregate source
used in this study. Overall, the GAB gradation is much
closer to the 0.45 maximum density curve representing a
gradation where the aggregate particles fit together in their
densest possible arrangement. It should be noted that 013C
aggregate shows more passing No. 4 and less passing No. 8
sieve materials. This implies that adding finer screening
materials may create a denser mix similar to Group II

Table 4.

Proctor test
The optimum moisture content and maximum dry density
were measured in conformance with ASTM D1557 using
the four different aggregate sources prior to CBR tests. For
each source of aggregates, the proctor test was conducted
with varying levels of screening materials in the mix, as
follows:
100% regular GAB
75% GAB with 25% 810 screenings
50% GAB with 50% 810 screenings

An automatic compaction apparatus was employed to

perform the 56 blows per layer as prescribed by ASTM
D1557. Figure 4 shows proctor test results.

Morphological index differences across aggregate sources in Group II.

19 mm

Morphological index
Angularity
FE ratio
ST

aggregates. To verify this, proctor tests were conducted for

each source with varying screening contents as described
in the next section.
Figure 3 shows the particle size distribution of No.810
screening materials from each source. It should be noted
that the 028C screening materials exhibited a gap
gradation with most particle sizes lying between 2.36
and 4.75 mm. When this type of screening material is
added in GAB, it could be expected to provide aggregate
interlocking to a certain degree.

12.5 mm

4.75 mm

F statistic

p value

F statistic

p value

F statistic

p value

1.174
1.050
3.298

0.316
0.356
0.043*

7.168
0.893
8.643

0.001**
0.412
0.000**

3.197
1.081
1.712

0.045*
0.343
0.186

Notes: *5% significance; **1% significance.

Downloaded by [Deakin University Library] at 19:53 07 January 2016

International Journal of Pavement Engineering

Figure 2.

Particle size distributions of virgin GAB materials.

Figure 3.

Particle size distributions of screening materials.

S.-H. Kim et al.

144

142

013C
Poly. (100%)

142

Poly. (75%/25%)
Poly. (50%/50%)

138

138
Dry Density (pcf)

Dry Density (pcf)

140

136
134
132

028 C
Poly. (100%)
Poly. (75%/25%)
Poly. (50%/50%)

140

136
134
132

130
130

128
126

128
0

10

12

14

137

135

6
8
10
Moisture Content (%)

12

158 C
Poly. (100%)

134

Poly. (75%/25%)
Poly. (50%/50%)

132
Dry Density (pcf)

134
133
132
131

14

136

048C
Poly. (100%)
Poly. (75%/25%)
Poly. (50%/50%)

136

Dry Density (pcf)

Downloaded by [Deakin University Library] at 19:53 07 January 2016

Moisture Content (%)

130

128

126
130
124

129
128
0

10

12

Moisture Content (%)

Figure 4.

122
0

10

12

14

Moisture Content (%)

Proctor test results.

It can be observed in Figure 4 that for Group I (013C)

proctor curves tend to shift upwards and leftwards.
It indicates that the maximum dry density increases as the
proportion of screening materials increases if the optimum
moisture content remains approximately 6% or less. This
reveals that the 013C material is not completely well
graded and the finer screening particles occupy the
available voids, thereby increasing the maximum dry
density. The increasing trend of the maximum dry density
for 013C material is an interesting discovery because the
performance of Group I aggregate could be improved by
adding No.810 screening materials. However, it should be
noted that finer mixes generally retain more water, and
thus their stiffness and strength are likely to decrease.
Therefore, the best material at constant water content may
not be the best material when the actual water content
likely to occur in equilibrium in the road construction is
considered. Under an asphalt layer, a somewhat more open
graded material, albeit still well-graded, is required to
ensure necessary drainage capability.

In contrast, Group II materials (028C, 048C and 158C)

all display a similar but reversing trend. The compaction
curves tend to move towards the lower right with
decreasing maximum dry density and increasing optimum
moisture content as the proportion of screening materials
increases. This trend indicates that Group II GAB sources
are well graded and adding screening materials reduces the
maximum dry density. This could result in a decrease in
stiffness because of the loss of coarse aggregate particle
interlocking due to excess fines being added.

California bearing ratio test

In this study, the CBR test was used to assess the quality of
compacted granular materials with varying contents of
screening materials. Specimens were compacted at
optimum moisture with maximum dry density and
subjected to CBR test in accordance with ASTM
Standards D1883-05. After setting, the piston was
penetrated into the specimen with a rate of 1.27 mm/min.

Downloaded by [Deakin University Library] at 19:53 07 January 2016

International Journal of Pavement Engineering

Vertical penetration depth and load of piston were
recorded. CBR value was calculated from the ratio of the
vertical stress at a penetration depth of 2.5 and 5 mm
penetration depths. Figure 5 summarises the CBR test
results.
Group I (013C) shows a CBR value of 22 when 100%
GAB was used, which is the lowest compared to those
sources in Group II. As discussed in the previous section,
angularity and ST of 013C were significantly lower than
those of other sources. Therefore, it could be inferred that
higher angularity and ST contribute to a stiffer GAB and a
higher CBR value. An interesting finding is that CBR
values of the 013C source increased as more No.810
screening materials were added. This is likely due to the
interplay between the primary structure (coarse aggregates) and the secondary structure (finer screening
materials), where an appropriate amount of screening
particles occupying the spaces between coarse aggregates
results in improved stability of overall aggregate skeleton
(Yideti et al. 2013).
Even though an increase of CBR was observed by
increasing No.810 screening contents, the resulting CBR
values are much lower than those of other sources in
Group II. Figure 6 illustrates gradation changes for Group
I (013C) after replacing GAB with 25% and 50%
screening materials. Considering GDOTs current GAB
gradation requirement, it might be possible to replace 25%
of GAB with No.810 screening materials. However, this
only resulted in a marginal CBR gain of 4. As such, a

Figure 5.

CBR test results.

benefit cost analysis and pavement performance evaluation may be needed to identify if adding 25% screening
materials is beneficial to lower the construction cost while
maintaining a structurally sound pavement.
In comparison, CBR values for Group II (028C, 048C
and 158C) sources are higher and in the range of 46 53
when 100% GAB was used. A decreasing trend in CBR
was observed for all Group II sources as more GAB was
replaced by the screening materials. This may imply that
the gradations of Group II GAB sources were already well
graded and adding more fines could result in a loss of
interlocking among aggregates. The level of decrease in
CBR varied depending on aggregate sources and grading.
In particular, 048C source experiences the largest drop in
CBR as the screening materials were added. Figure 7
shows the gradation changes for Group II sources when the
original GAB is replaced by 25% screening materials.
As shown in Figure 7, replacing 25% GAB with screening
materials results in gradation curves exceeding the upper
limit of GDOTs GAB gradation requirement.
Pavement performance evaluation
MEPDG simulation conditions
The design of the aggregate base layer, which is a
constitutive component of pavement structure, might be
better evaluated by way of the overall pavement
performance. For this evaluation, SR 17/US78 project
was utilised as a test section. MEPDG program Version

Downloaded by [Deakin University Library] at 19:53 07 January 2016

Figure 6.

S.-H. Kim et al.

Gradation changes for Group I source.

1.10 was used. For more accurate and rigorous

comparison, design parameters for SR17/US78 Pavement
Reconstruction Project in McDuffie/Wilkes County were
utilised as summarised in Table 5.
A three-layer asphalt pavement structure was considered for this study. The asphalt structures for the

Figure 7.

Gradation changes for Group II sources.

MEPDG simulations were composed of a 7.5 inch HMA, a

12 inch Aggregate Base, and subgrade, detailed in Table 5.
Specifically, HMA E* values were obtained from GDOTs
material input library (Kim 2013) and GAB CBR values
were obtained from the current study. Three design
alternatives for the aggregate base layer, (1) 100% GAB,

International Journal of Pavement Engineering

Table 5.

MEPDG inputs.

Traffic data

Structure

Materials inputs

Downloaded by [Deakin University Library] at 19:53 07 January 2016

2018 Average Daily Traffic, one-way or directional: 3500

2038 Average Daily Traffic, one-way or directional: 5300
Lane Distribution factor: 0.90
2018 Average Annual Daily Truck Traffic: 781 (this value represents 22% trucks based on an AADT of 3500
Average Annual Growth Rate of Trucks (%): 2.6%; assumed to be the same as AADT
Design 18 KESALs; Cumulative over Design: 5,701,300
1.500 SMA
200 HMA with PG 64-22
400 HMA with PG 64-22
1200 Non-stabilised Base
HMA E* GDOT RP 12-07 Report
GAB CBR was measured in this study.
Soil: 2 Silty Clay/2 Clayey Silt/4Silty Clay
Soil support value of 3.5

(2) 75% GAB/25% screening materials, and (3) (2) 50%

GAB/50% screening materials, were evaluated.

Pavement performance
In MEPDG simulations, alligator cracking, permanent
deformation and IRI were estimated based on the inputs in
Table 5. Figure 8 shows the evolution of alligator
cracking. Except for the GAB layer, other layers inputs
remained the same. Level 2 CBR values were entered in
Pavement ME depending on the screening contents as
shown in Figure 5.
The upper limit of vertical axis (y-axis) represents the
design limit value of each performance index. The design
limit for alligator cracking was set as 15%, which is
commonly used in Georgia. As shown in Figure 8, the
amount of alligator cracking was predicted to decrease as
more screening materials were added for limestone source
(013C). This decreasing trend for 013C is likely due to the
resulting higher density and stiffer GAB layer as more
screening materials are added. For Group II sources,
alligator cracking was predicted to increase over time, but
not exceed the failure criterion. However, alligator
cracking increased dramatically as 25% of screening
materials was added. As shown previously, Group II
sources with 25% of screening materials exceeded the
upper limit of GDOTs GAB gradation requirement, and
further investigations are needed prior to the adoption of
screening materials in the GAB layer.
Figure 9 shows the total rut depths of the pavement
structures simulated. There appears to be little
performance difference (about 0.01 inches) in rut depth
between the two design alternatives because the trends of
evolution are quite similar and overlapping each other.
For a 20-year simulation, the rut depths were predicted
to be lower than the design limit of 0.5 inch. Therefore,
one could conclude that use of screening materials up to
25% of GAB would likely not result in significant
performance difference in terms of permanent
deformation.

Figure 8. Evolution of alligator cracking.

S.-H. Kim et al.

Downloaded by [Deakin University Library] at 19:53 07 January 2016

10

Figure 9.

Evolution of permanent deformation.

IRI, has been used to measure road roughness for

evaluating and managing road systems. IRI is calculated
based on the longitudinal profiles of wheel paths and is a
function of pavement distresses, including fatigue and
thermal cracking. Figure 10 illustrates IRI progression
over time. As shown, IRI trends associated with different
source materials converge and little difference was found
with different percentages of screening materials. Therefore, it could be inferred that the aggregate types and
amount of screening materials would not likely effect a
difference in the performance of asphalt pavements in
terms of IRI.

Figure 10.

IRI prediction.

Conclusions
This paper documents a study on the effect of No. 810
screening content in GAB layer on the GAB strength and
overall pavement performance. One limestone source (013C)
and three granite sources (028C, 048C and 158C) in Georgia
were selected. Three content levels of screening materials in
GAB materials, i.e. 0%, 25% and 50%, were studied. The
results indicated that the effect varies and depends on the
source and the amount of screening materials used.
Morphological analysis showed that Group I limestone
source has significantly lower angularity and ST compared
to Group II granitic gneiss sources. This partially explains

Downloaded by [Deakin University Library] at 19:53 07 January 2016

International Journal of Pavement Engineering

the lower CBR values for the limestone source than those
of the granitic gneiss sources. For limestone source, higher
content of No.810 screening materials results in higher
density of the assemblies and less percentage of alligator
cracking. The increasing trend of the maximum dry density
for the limestone source indicated that the performance of
the aggregate could be improved by adding No.810
screening materials. However, the percentage of alligator
cracking still exceeded the failure criterion by the end of
the design period based on the MEPDG simulation.
It should be noted that finer mixes generally retain more
water, and thus their stiffness and strength are likely to
decrease. Therefore, the best material at constant water
content may not be the best material when the actual water
content likely to occur in equilibrium in the road
construction is considered. Under an asphalt layer, a
somewhat more open-graded material, albeit still well
graded, is required to ensure necessary drainage capability.
In contrast, for granite sources, higher content of
No.810 screening materials results in lower density of the
assemblies and thus higher percentage of alligator
cracking. The gradation of those sources with screening
materials did not meet GDOTs requirement. To satisfy
the gradation requirement for Group II sources, a lower
amount of screening materials may be considered and
verified through further refined studies focusing on the
permeability, plasticity index, resilient modulus, permanent deformation and life cycle cost prior to the adoption
of any screening materials in the GAB layer.
Disclosure statement
No potential conflict of interest was reported by the authors.

References
Al-Rousan, T., et al., 2007. Evaluation of image analysis
techniques for quantifying aggregate shape characteristics.

11

Journal of Construction and Building Materials, 21 (5),

978 990. doi:10.1016/j.conbuildmat.2006.03.005.
Barksdale, R.D. and Itani, S.Y., 1989. Influence of aggregate
shape on base behaviour. In: Transportation research record
1227. Washington, DC: TRB, National Research Council,
173 182.
Dawson, A.R., 2000. Unbound aggregates in road. Rotterdam,
London: Balkema.
Haynes, J.G. and Yoder, E.J., 1963. Effects of repeated loading
on gravel and crushed stone base course materials used in
the AASHO road test. In: Highway Research Record 39.
Washington, DC: National Research Council, HRB,
82 86.
Hicks, R.G. and Monismith, C.L., 1971. Factors influencing
the resilient properties of granular materials. In: Transportation research record 345. Washington, DC: Transportation Research Board, National Research Council,
15 31.
Kim, S., 2004. Determination of aggregate physical properties
and its effect on cross-anisotropic behavior of unbound
aggregate materials. Ph.D., Dissertation, Texas A&M
University, College Station, TX.
Kim, S., Measurements of dynamic and resilient moduli for
roadway test sections (Final Report) GDOT research project
no. 12 07 2013.
Kolisoja, P., 1997. Resilient deformation characteristics of
granular materials. Ph.D., Dissertation, Tampere University
of Technology, Tampere, Finland.
Lekarp, F., Richardson, I.R., and Dawson, A., 1996. Influence on
permanent deformation behavior of unbound granular
materials. In: Transportation research record 1547. Washington, DC: TRB, National Research Council, 68 75.
Raad, L., Minassian, G., and Gartin, S., 1992. Characterization of
saturated granular bases under repeated loads. In: Transportation research record 1369. Washington, DC: TRB,
National Research Council, 73 82.
Thom, N.H. and Brown, S.F., 1987. Effect of moisture on the
structural performance of a crushed-limestone road base. In:
Transportation research record 1121. Washington, DC:
TRB, National Research Council, 50 56.
Yideti, TF., et al., 2013. Packing theory-based framework to
evaluate permanent deformation of unbound granular
materials. International Journal of Pavement Engineering,
14 (3), 309 320. doi:10.1080/10298436.2012.736620.