G Model

SCS-147; No. of Pages 9 ARTICLE IN PRESS

Sustainable Cities and Society xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Sustainable Cities and Society

journal homepage: www.elsevier.com/locate/scs

Sustainable roadway construction using recycled aggregates with

geosynthetics
Jie Han a,∗ , Jitendra K. Thakur b,1
a
Civil, Environmental, and Architectural Engineering Department, The University of Kansas, 1530W, 15th Street, Lawrence, KS 66045-7609, United States
b
Terracon Consultants, Inc., 1211 W. Florida Avenue, Midland, Texas 79701, United States

a r t i c l e i n f o a b s t r a c t

Keywords: Concrete, asphalt pavements, and ballast are removed during the re-construction of existing roads
Geosynthetic and have been increasingly recycled as aggregates for the construction of roadways. Due to existence
Recycled aggregate of asphalt, cement, and fines, mechanical properties of recycled aggregates may not be sufficient for
Stabilization
load support. They may also have long-term durability problems. Geosynthetics have been used to
improve mechanical properties and long-term durability of recycled aggregates. This paper reviews
recent research work on the use of geosynthetics to stabilize recycled aggregates in roadway construction
and summarizes the main findings on permanent deformation, creep deformation, degradation, stress
distribution, and/or crack propagation.
© 2014 Published by Elsevier B.V.

1. Introduction been in practice since 1930s. The U.S. FHWA estimated that 100.1
million tons of asphalt pavement materials are milled off each
According to the Federal Highway Administration (FHWA, year during resurfacing and widening of road projects, of which
2004), two billion tons of aggregates are quarried annually in 80.3 million tons are reclaimed and reused for roadbeds, shoul-
the United States and the quantity of quarried aggregates will ders, and embankments (Missouri Asphalt Pavement Association,
reach 2.5 billion tons by 2020, which force construction industries 2010). RCA is a removed and reprocessed construction material
to consider new sources of aggregates. Roadways (highways and from demolished concrete structures, such as high-rise buildings,
railways) that have reached the end of their service lives are fre- bridges, highways, railways, etc. containing cement and natural
quently rehabilitated by removing the existing roadway surfaces aggregates. The natural aggregates contain 60–75% of the total
and replacing the removed portion with new construction materi- volume of RCA (RMRC, 2008). Ballast is a free-draining granular
als. A large amount of recycled aggregates are created every year material composed of medium to coarse gravel-sized aggregates
during the rehabilitation and reconstruction of existing roadways. (10–60 mm in diameter) with a small percentage of cobble-sized
Currently, great emphasis is placed on sustainable construction and particles, commonly used as a load-bearing material in railway
infrastructure with green technologies because the demand for sus- tracks (Indraratna, Khabbaz, Salim, & Christie, 2006). The good qual-
tainable and environmental-friendly roads is increasing daily. More ity of ballast consists of angular particles with rough surface and
technologies for sustainable roadway construction are needed. One minimum hairline cracks and should have high specific gravity,
way to construct sustainable roads is through the use of recycled shear strength, toughness and hardness, and enough resistance to
aggregates. Recycled Asphalt Pavement (RAP), Recycled Concrete weathering (Indraratna et al., 2006). Railway ballast degrades and
Aggregate (RCA), and Recycled Ballast (RB) are the three types of deteriorates progressively under repeated cyclic loading. Degraded
recycled aggregates as shown in Fig. 1 and are discussed in this ballast is usually replaced by fresh ballast during routine track
paper. maintenance. The railway track constructed using recycled ballast
According to the Recycled Material Resource Center (RMRC, (RB) shows excessive settlement and lateral deformation, which
2008), RAP is a removed and reprocessed pavement material affect the performance of railroads.
from deteriorated asphalt pavements containing asphalt binder The use of recycled aggregates can reduce the cost of construc-
(3–7%) and aggregates (97–93%) by weight. The use of RAP has tion materials, reduce the amount of waste to be land-filled, reduce
the transportation and energy costs to import virgin aggregates,
and conserve natural resources by requiring less virgin aggre-
∗ Corresponding author. Tel.: +1 7858643714; fax: +1 7858645631. gates in road construction projects. Several agencies are seriously
E-mail addresses: [email protected] (J. Han), [email protected] (J.K. Thakur). considering the economic and environmental benefits of using
1
Tel.: +432 684 9600. recycled aggregates in roadways and facing challenges to maintain

2210-6707/$ – see front matter © 2014 Published by Elsevier B.V.

http://dx.doi.org/10.1016/j.scs.2013.11.011

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
2 J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx

Fig. 1. Types of recycled aggregates.

high-quality road infrastructure. The mechanical properties geogrids available in the market. Uniaxial geogrid provides tensile
(mainly strength and stiffness) and long term durability (break- resistance in only one direction, biaxial geogrid provides tensile
age and abrasion) of recycled aggregates may not be sufficient for resistance in two directions, and triaxial geogrid can provide near-
load support due to the existence of asphalt and cement, or loss uniform tensile resistance (Dong, Han, & Bai, 2010; Qian, 2009)
of angularity of ballast. In the past, most of the research studies on when it is subjected to tension in different directions. Qian (2009)
the improvement of RAP and RCA quality focused on blending them reported that the confinement of granular base aggregates was
with virgin aggregates or stabilizing them using chemical additives. obtained through the interlocking between geogrid apertures and
However, the blending of RAP and RCA with virgin aggregate still aggregate particles as shown in Fig. 3b. Webster (1992) reported
consumes natural resources and the chemical stabilization is not that the degree of interlocking depended on geogrid aperture size
always environmental friendly. and aggregate particle size and the effectiveness of interlocking
The use of 100% recycled aggregate with geosynthetic is a depended on the in-plane stiffness, rib strength, and junction
sustainable solution. Liu, Scarpas, Blaauwendraad, and Genske strength of the geogrid. Thakur et al. (2012) reported that the
(1998) were few early researchers to explore such a possibil- geocell-reinforced bases had improved bending resistance. The
ity. Recent research work done by Indraratna, Salim, and Christie beam effect of geocell-reinforced bases is demonstrated in Fig. 3c.
(2002), Indraratna et al. (2006), Han et al. (2011), Thakur, Han,
Pokharel, and Parsons (2012), Thakur, Han, and Parsons (2013),
and others has further evaluated the behavior and performance 3. Geosynthetic-reinforced recycled asphalt pavement
of geosynthetic-reinforced recycled aggregates. This paper reviews (RAP)
these studies on the use of geosynthetics to stabilize recycled aggre-
gates (RAP, RCA, and RB) in roadway construction and summarizes Geotextile, geogrid, and geocell have been used to stabilize RAP
the main research findings. bases. This section discusses the effects of geosynthetic reinforce-
ment on the permanent deformation, resilient deformation, creep
deformation, and stress distribution of RAP bases.
2. Reinforcement mechanism of geosynthetic
Foye (2011) presented the work of a design-build contractor
who used a geosynthetic stabilization technique for reconstruction
Geosynthetics manufactured from polymeric materials have
of 19,500 m2 asphalt parking lot on a site with very weak subgrade
been widely used as construction materials to solve many civil engi-
(CBR ranging from 1 to 3%). The remedial design parking lot sec-
neering problems since 1970s. Geosynthetics are used to improve
tion consisted of very weak subgrade soil overlaid by 200 mm thick
the performance of unpaved and paved roads for over 40 years
geocomposite (a 271 g/m2 needle-punched nonwoven geotextile
(Giroud & Han, 2004). The use of geotextile, geogrid, and geocell
– geogrid)-stabilized blended RAP aggregate base, 64 mm thick
with recycled aggregates are discussed in this paper. Geotextile
dense-graded asphalt course, and 25 mm thick asphalt wearing
and geogrid are planar geosynthetics whereas geocell is a three-
course. The geocomposite was placed at the interface of subgrade
dimensional honeycomb type of geosynthetic. Geogrid and geocell
and granular base course to provide separation and reinforcement.
improve the performance of aggregate layers by providing lat-
It was found that the geocomposite-stabilized parking lot section
eral confinement whereas geotextile improves the performance of
showed little rutting or deflection under proof rolling and the use
aggregate layers by providing a tensioned membrane effect. Differ-
of the geocomposite reduced the cost of construction from about
ent types of geosynthetics used in roadway construction are shown
$890,000 (estimated for the original cut and replacement specifica-
in Fig. 2.
tion) to about $200,000. In addition, the geocomposite stabilization
The most efficient and convenient location of geosynthetic in
technique saved time, resources, and energy as compared with the
roadway construction is at the interface of subgrade and granular
traditional cut and replacement technique.
base course (Das & Shin, 1998). Geosynthetic installed at this
Han et al. (2011) conducted moving wheel tests on five geocell-
location provides full or partial separation, lateral confinement
reinforced and two unreinforced RAP bases over weak subgrade
of granular base materials, a tensioned membrane or beam effect
(target CBR = 3%) to evaluate the effect of geocell reinforcement on
when a road deforms extensively. The tensioned membrane
rut depth and stress distribution angle at a certain number of passes
or beam effect is referred to as the tension developed in the
of the wheel load. Two types of recycled asphalt materials, named
curved geosynthetic-reinforced base to resist the vertical load
RAP and FRAP (fractioned RAP or RAP with finer gradation) were
(Rajagopal, Krishnaswamy, & Madhavi Latha, 1999). The tensioned
used in this study. The following base sections were prepared and
membrane effect mechanism is shown in Fig. 3a. Nonwoven
tested:
geotextile provides separation, filtration, and drainage whereas
woven geotextile provides separation and reinforcement. Geogrid
and geocell provide reinforcement to aggregate base and subgrade (1) 300 mm thick unreinforced RAP.
by providing lateral confinement due to their tensile strength and (2) 150 mm thick geocell-reinforoced RAP with a 20 mm thick RAP
stiffness. Uniaxial, biaxial, and triaxial geogrids are three types of cover.

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx 3

Fig. 2. Types of geosynthetics. (a) Non-woven geotextile and geocell. (b) Triaxial geogrid. (c) Biaxial geogrid. (d) Uniaxial geogrid. (e) Woven geotextile.

(3) 100 mm thick geocell-reinforced RAP with a 70 mm thick RAP on the road surface for each section. The vertical stress increased
cover. or remained constant with the number of passes for unreinforced
(4) double layered geocell-reinforced RAP with a 30 mm thick RAP sections but decreased with the number of passes for reinforced
cover above a 100 mm thick bottom geocell layer and a 70 mm sections. They attributed this phenomenon to the beam or slab
thick RAP cover above a 100 mm thick top geocell layer. effect of the geocell-reinforced bases.
(5) 250 mm thick unreinforced FRAP. Thakur et al. (2012) conducted large-scale cyclic plate load-
(6) 100 mm thick geocell-reinforced FRAP over a 100 mm thick ing tests on one unreinforced RAP base (300 mm thick) and three
unreinforced FRAP base course with a 50 mm thick FRAP cover. geocell-reinforced RAP bases (150, 230, and 300 mm thick) over
(7) 75 mm thick geocell-reinforced FRAP over a 100 mm thick weak subgrade (target CBR = 2%) to evaluate the performance of
unreinforced FRAP base course with a 75 mm thick FRAP unreinforced and geocell-reinforced RAP bases over weak sub-
cover. grade. The permanent deformation, the resilient deformation, the
vertical stress at the interface of subgrade and base, and the strains
They found that the geocell improved the life of unpaved sections by in the geocell wall were measured during the cyclic plate loading
a factor of 1.3 using one layer of 75 mm high geocell and 1.8 using tests. Fig. 5 shows the permanent deformations of the unrein-
one layer of 100 mm high geocell at a rut depth of 75 mm. They forced and reinforced bases over weak subgrade at the center of
concluded that the geocell reinforcement reduced the rut depth the loading plate versus the number of loading cycles. The per-
and vertical stresses transferred to the subgrade by distributing the manent deformation increased with the number of loading cycles.
load over a wider area at the same number of passes. For a demon- The rate of increase in the permanent deformation decreased with
stration purpose, the vertical stresses at the interface of subgrade the number of loading cycles. On the weak subgrade, the geocell-
and FRAP base versus the number of passes are shown in Fig. 4. reinforced RAP bases (150, 230, and 300 mm thick) improved the
The measured vertical stresses at the interface of subgrade and performance (i.e., the number of cycles) by a factor of 6.4, 3.6, and
base were much lower than the tire pressure of 552 kPa applied 19.4 as compared with the 300 mm thick unreinforced RAP base,

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
4 J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx

Fig. 3. Different reinforcement mechanism of geosynthetics. (a) Tensioned membrane effect mechanism (from Maxwell, Kim, Edil, & Benson, 2005). (b) Aggregate interlocking
between geogrid aperture and aggregate particle (from Tensar International). (c) Beam effect of geocell.

respectively, at 75 mm permanent deformation. The 230 mm thick Thakur et al. (2013) conducted plate loading tests to investi-
geocell-reinforced base had a lower improvement factor than the gate the vertical stress-displacement responses of the following
150 mm thick geocell-reinforced base because of the lower CBR RAP specimens:
values of the base and subgrade in the 230 mm thick base as com-
pared with those in the 150 mm thick base. They concluded that
geocell-reinforced RAP bases provided a sustainable solution for (1) unreinforced sample (unreinforced RAP sample extruded from
roadway construction technology by improving the performance a Proctor compaction mold)
of RAP bases. In addition to reducing the permanent deformation (2) unreinforced base (a RAP base prepared in a test box without
of bases, geocell reinforcement reduced the vertical stresses trans- geocell)
ferred to the subgrade and increased the percentage of resilient (3) single geocell-reinforced base (a RAP base prepared by placing
deformation of RAP bases. RAP into the single geocell pocket and the test box)

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx 5

200 Vertical stress (kPa)

Unreinforced FRAP
FRAP reinforced with 10 cm high geocell
FRAP reinforced with 7.5 cm high geocell
0 100 200 300 400 500 600
0
160
Vertical stress (kPa)

Displacement (mm)
5
120
10

80 15

Unreinforced base
20
40
Single geocell-confined base
25
Multi geocell-confined base
0
500 2000 4000 8000 12000 30
Number of passes
Fig. 6. Vertical stress-displacement curves for unreinforced and geocell-reinforced
Fig. 4. Vertical stress at the interface of base and subgrade versus number of passes RAP bases (data from Thakur et al., 2013, ©ASCE).
(data from Han et al., 2011, ©ASCE).

confinement significantly reduced the creep deformation of the

(4) multi geocell-reinforced base (a RAP base prepared by placing RAP bases and multiple geocell-reinforced base crept least followed
RAP into the multiple geocell pockets and the test box). by the single geocell-reinforced base and the unreinforced base.
Han, Acharya, Thankur, and Parsons (2012) conducted large-
The applied vertical stress versus displacement curves are shown scale cyclic plate loading tests on asphalt pavements with
in Fig. 6. They found that the unreinforced RAP sample failed at geocell-reinforced RAP bases over moderate subgrade (target
172 kPa while other sections did not fail up to a vertical stress of CBR = 5%) and concluded that geocell reduced the permanent defor-
586 kPa and showed a linear vertical stress-displacement response. mations of HMA layer, RAP base, and subgrade, reduced the vertical
The stress-displacement responses were analyzed in terms of a stress at the interface of base and subgrade, and increased the elas-
modulus improvement factor. The modulus improvement factor ticity of the RAP base.
is the ratio of the slope of the initial portion of the vertical stress Bortz, Hossain, Halami, and Gisi (2012) conducted moving wheel
versus displacement curve for the geocell-reinforced base to that of tests on eight asphalt pavement sections, in which two unre-
the unreinforced base. The test results showed that the moduli of inforced sections had well-graded crushed limestone aggregate
the single geocell-reinforced and the multi geocell-reinforced bases (AB-3) bases and six geocell-reinforced sections had AB-3, quarry
were increased by 1.2 and 1.6 times compared to the unreinforced waste (QW), and RAP bases. The details of these bases are provided
base, respectively. below:
Thakur et al. (2013) also conducted static plate loading tests
in a test box and a compaction mold at a room temperature (1) 300 mm thick unreinforced AB-3
of about 25 ◦ C to investigate the effects of confinement, stress, (2) 75 mm thick geocell-reinforced QW with 25 mm thick cover
and cover material on creep deformations of unreinforced and (3) 75 mm thick geocell-reinforced RAP with 25 mm thick cover
geocell-reinforced RAP bases. The plate loading tests were con- (4) 75 mm thick geocell-reinforced AB-3 with 25 mm thick cover
ducted at vertical stresses of 276 and 552 kPa under five confining (5) 200 mm thick unreinforced AB-3
conditions including single and multiple geocell-reinforced RAP (6) 150 mm thick geocell-reinforced QW with 50 mm thick cover
bases. They concluded that geocell confinement reduced the ini- (7) 150 mm thick geocell-reinforced RAP with 50 mm thick cover
tial deformation and the rate of creep of the RAP bases, RAP bases (8) 150 mm thick geocell-reinforced AB-3 with 50 mm thick cover.
crept more at a higher vertical stress and at a lower degree of
confinement, and the well-graded aggregate cover significantly All the cover materials were the same as the infill materials. The
reduced the creep of geocell-reinforced RAP bases as compared subgrade was AASHTO A-7-6 clay and was prepared to obtain tar-
with the RAP cover. Fig. 7 shows the creep deformations of the get CBR values of 6% and 12%. The bases (1)–(4) were prepared over
geocell-reinforced RAP bases at the vertical stress of 276 kPa as the subgrade with a CBR of 6% and were paved with a 50 mm thick
compared with the unreinforced RAP bases. It is shown that geocell
30
80 Unreinforced base
Permanent deformation (mm)

25 Single geocell-confined base

70
Axial creep strain (%)

Multi geocell-confined base

60
20
50
150 mm reinforced
40 15
230 mm reinforced

30 300 mm reinforced
10
20 300 mm unreinforced
5
10

0 0
0 20 40 60 80 100 120 1 500 2000 4000 6000
Number of loading cycles Time (minute)

Fig. 5. Permanent deformations at the center versus the number of loading cycles Fig. 7. Creep behavior at vertical stress of 276 kPa (data from Thakur et al., 2013,
for RAP bases over weak subgrade (Thakur et al., 2012). ©ASCE).

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
6 J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx

30 90

Stress intensity factor, K (N/mm3/2)

80
25
RAP (moderate) AB-3 (moderate) 70
Rut depth (mm)

QW (moderate) RAP (firm)

20 60
AB-3 (firm) QW (firm)
50
15 Unreinforced CNA
40
Unreinforced RCA
30
10 Reinforced RCA
20 Unreinforced RMA
5 Reinforced RMA
10

0
0 0 50 100 150 200
12500 50000 100000 500000
Crack length (mm)
Number of passes
Fig. 9. K factor distribution versus crack length (redrawn and modified from Liu
Fig. 8. Rut depth at wheel path versus number of passes of geocell-reinforced sec- et al., 1998).
tions (redrawn and modified from Bortz et al., 2012).

unreinforced CNA base followed by the unreinforced RCA base, the

reinforced RMA, and the unreinforced RMA.
HMA layer, whereas the bases (5) to (8) were prepared over the
Donovan (2011) conducted dynaflect tests on five pavement
subgrade with a CBR of 12% and were paved with a 100 mm thick
test sections (150 m long each) constructed at city of Edmonton,
HMA layer. All eight sections were tested to evaluate the effects of
Canada to investigate the possibility of using RCA in roadway con-
base, cover and HMA thicknesses and geocell reinforcement on rut
struction. The pavement structure of each section consisted of
depth and stress distribution at the interface of subgrade and base
cement stabilized subgrade (CSS) overlaid by granular aggregate
at a certain number of passes of the wheel load. They concluded that
base (GAB), asphalt concrete base (ACB), and asphalt concrete over-
a minimum cover thickness of 50 mm over geocell and a minimum
lay (ACO). The city had added the second lift of ACO after two
HMA thickness of 100 mm over a base were necessary for better
years of original construction in 2004. The total thickness of asphalt
performance of pavements. The geocell-reinforced waste mate-
layer in each section was same (50 mm thick ACB + 60 mm thick
rials (RAP and QW) performed as well as the geocell-reinforced
first ACO + 50 mm thick second ACO = 160 mm) with a variation in
AB-3. The pavement sections constructed over the firm subgrade
type and thickness of the aggregate base course. Sections 1 and 2
(CBR = 12%) performed better than those over the moderate sub-
used natural aggregate and crushed natural aggregate, respectively,
grade (CBR = 6%) as shown in Fig. 8. In this figure, RAP (moderate)
whereas Sections 3–5 used recycled crushed aggregate as the gran-
and RAP (firm) stand for asphalt pavements with geocell-reinforced
ular base course. Recycled crushed concrete consisted of 60% RCA,
RAP bases over moderate and firm subgrades, respectively. The
25% RAP, and 15% other materials such as cement treated granular
same representation holds for AB-3 and QW. Fig. 8 does not show
base aggregate, brick and other recyclable materials. The thick-
the rut behavior at 100,000 and 500,000 passes for the pavement
nesses of GAB for Sections 1–4 were 325 mm while that for Section
sections constructed over the moderate subgrade because these
5 was 150 mm. Each section had 150 mm thick CSS except Section 3
sections reached the failure criterion of 12.5 mm rut depth before
which had 325 mm thick CSS. Geotextile and geogrid were installed
100,000 passes.
on top of subgrade in Sections 3 and 5, respectively. All other
remaining sections were unreinforced. The deflection test results
4. Gesosynthetic-reinforced recycled concrete aggregate are shown in Fig. 10. The recycled crushed aggregate sections
(RCA) performed better than the natural aggregate and crushed natural
aggregate sections. The deformation of each section increased with
Geotextile and geogrid have been used to stabilize RCA bases. time. The deformations measured in 2005 were less than those in
This section discusses the effects of geosynthetic reinforcement on 2004 because of the placement of the second lift of ACO in 2004.
the permanent deformation and crack propagation of RCA bases.
Liu et al. (1998) performed finite element analysis on asphalt 0.05
Section 1 Section 2
concrete pavements with geogrid reinforced recycled concrete Section 3 Section 4
aggregate and natural aggregate bases to evaluate the performance 0.045 Section 5
of the geogrid-reinforced RCA base as compared with the recycled
Deformation (mm)

masonry aggregate (RMA) base and the crushed natural aggregate 0.04
(CNA) base. Influence of the material characteristics of the recycled
aggregate and the geogrid reinforcement on the development and
0.035
rate of propagation of reflective cracking in the top layer of the
pavement was selected as the criterion for the comparison pur-
0.03
pose. They used stress intensity factor (K) distributions for both
unreinforced and reinforced pavements to evaluate the perfor-
mance in terms of the energy available at a crack tip for additional 0.025
crack propagation. Fig. 9 shows that geogrid reinforcement reduced
the K factor for different base materials thus improved the pave- 0.02
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
ment performance. They concluded that geogrid reduced the rate
Year
of crack propagation into the top layer of the pavement, improved
the load spreading in the base layer, and enhanced the pavement Fig. 10. Deformation behavior of pavement sections (redrawn and modified from
life. The geogrid-reinforced RCA base performed better than the Donovan, 2011).

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx 7

30 Number of loading cycles

Permanent deformation (mm)

0 100000 200000 300000 400000 500000 600000

25 0
Unreinforced fresh ballast (dry)
Unreinforced recycled ballast (dry)

Settlement (mm)
Geogrid-reinforced recycled ballast (dry)
20 5 Geogrid-reinforced recycled ballast (wet)
Geogrid-geotextile-reinforced recycled ballast (dry)
Geogrid-geotextile-reinforced recycled ballast (wet)
15 Unreinforced FA 10
Unreinforced RCA
10 Reinforced FA 15
Reinforced RCA
5 20

0 25
0 10000 20000 30000 40000 50000 60000
(a)
Number of cycles
Number of loading cycles
Fig. 11. Permanent deformation versus number of loading cycles (redrawn and
0 100000 200000 300000 400000 500000 600000
modified from Gongora & Palmira, 2012). 0
Unreinforced fresh ballast (dry)
Unreinforced recycled ballast (dry)

Settlement (mm)
Section 4 performed best followed by Sections 5, 3, 1 and 2 in long 5 Unreinforced recycled ballast (wet)
Geotextile-reinforced recycled ballast (dry)
term as shown in Fig. 10. They concluded from this study: (1) the Geotextile-reinforced recycled ballast (wet)
geotextile had a positive effect when the subgrade soil was very 10
weak; (2) the geogrid improved the life of pavement section and
reduced the thickness of granular base by 50% as compared with a 15
similar unreinforced base to provide the same performance; (3) the
recycled crushed aggregate was successfully used as the granular
20
base course.
Gongora and Palmira (2012) conducted laboratory cyclic plate
25
loading tests on unreinforced and biaxial geogrid-reinforced
(b)
unpaved road test sections prepared inside a steel tank (750 mm in
diameter and 530 mm high) to investigate the performance of unre- Fig. 12. Ballast settlement versus number of loading cycles. (a) Redrawn and mod-
inforced and geogrid-reinforced unpaved roads on weak subgrade ified from Indraratna et al. (2002). (b) Redrawn and modified from Indraratna and
(CBR ≈ 4%). The cyclic load was applied with a peak force of 17.6 kN Salim (2003).
at a wave frequency of 1 Hz. Each test section consisted of 300 mm subgrade, a 100 mm thick subbase layer (i.e. a capping layer of
thick subgrade overlaid by a 230 mm thick granular aggregate base. gravel and sand), 300 mm thick load bearing ballast, and a 150 mm
Fresh aggregate (FA) and recycled concrete aggregate (RCA) were thick crib ballast layer of fresh or recycled ballast. They found
used as granular aggregate base materials. RCA was coarser in size that the geosynthetics (woven geotextile, geogrid, and geogrid-
than FA. The geogrid was installed at the interface of subgrade and geotextile composite) at the interface of ballast and the capping
base course in case of the reinforced sections. The deformations at layer reduced the amount and rate of vertical and lateral deforma-
different numbers of cycles were measured using linear variable tions. The wet RB had more deformations than the dry RB. They
differential transducers (LVDT). The deformation behavior of unre- developed a semi-logarithmic equation (S = A + B logN) to predict
inforced and geogrid reinforced sections is shown in Fig. 11. The the settlement of RB, where A and B are empirical constants depend-
geogrid-reinforced sections had much smaller permanent defor- ing on the initial compaction, type of ballast, type of reinforcement,
mations than the unreinforced sections. The geogrid-reinforced magnitude of cyclic loading, and degree of saturation; N is the num-
RCA performed best, followed by the reinforced FA, RCA, and FA ber of load cycles; and S is the tie (also called sleeper) settlement.
as shown in Fig. 11. The test results indicated that the geogrid Fig. 12 shows the effects of geosynthetic reinforcement and satu-
improved the life of reinforced RCA and FA sections by factors of ration on the deformation behavior of fresh and recycled ballasts
33.5 and 15 at a permanent deformation of 25 mm as compared in two different studies.
with the corresponding unreinforced sections, respectively. Indraratna et al. (2005) reported that the optimum location
of the geosynthetic was 200 mm beneath the tie to improve the
5. Gesosynthetic-reinforced recycled ballast (RB) railway track performance; however, it was easy to place the
geosynthetic at the ballast-capping interface. Indraratna et al.
Railroad ballast provides a support for railroad tracks and dis- (2005) also investigated the effects of geosynthetic reinforcement
tributes loads to weak subgrade. Geotextile and geogrid have been and saturation on degradation behavior of fresh and recycled bal-
used to stabilize fresh and recycled ballasts. This section discusses lasts under cyclic loading by sieving each ballast sample before and
the effects of geosynthetic reinforcement on the deformation and after the test, and recording the change in percentage retained on
degradation of RB. each sieve size (Wk ). They calculated a breakage index (Bg ) for
Indraratna et al. (2002) conducted large-scale laboratory cyclic each specimen using the method proposed by Marsal (1967). Break-
load tests in a cubical triaxial chamber, which simulated the field age index (Bg ) is the sum of the positive values of Wk . Higher Bg
load and boundary conditions on unstabilized, geogrid-stabilized, indicates higher potential to degradation. The degradation behav-
and geogrid-geotextile composite stabilized-fresh and recycled ior of unreinforced and geosynthetic-reinforced ballasts is shown
ballasts. Indraratna and Salim (2003), Indraratna, Shahin, and in Fig. 13. The unreinforced RB had 97 and 95% more breakage
Salim (2005), and Indraratna et al. (2006) conducted cyclic labo- as compared with the unreinforced FB in dry and wet conditions,
ratory triaxial tests in a prismoidal chamber, which simulated field respectively. Each specimen had slightly higher degradation in the
loading conditions on unstabilized and geotextile-stabilized fresh wet condition than that in the dry condition. Geogrid-reinforced
and recycled ballasts. The test sections consisted of 50 mm thick RB, geotextile-reinforced RB, and geocomposite-reinforced RB

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
8 J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx

3.5 Number of loading cycles

Unreinforced FB
Unreinforced RB 0 100000 200000 300000 400000 500000 600000
3 Geogrid-reinforced RB 0
Geotextile-reinforced RB

Vertical deformation (mm)

2 Unreinforced FB
2.5
Breakage index

Geocomposite-reinforced RB
4 Unreinforced RB
Geocomposite reinforced FB
2 6 Geocomposite reinforced RB
8
1.5
10
1 12
14
0.5
16
0 18
Dry Wet (a)

Fig. 13. Breakage indices of unreinforced and geosynthetic-reinforced ballast

Number of loading cycles
(redrawn and modified from Indraratna et al., 2005). 0 100000 200000 300000 400000 500000 600000
0

Lateral deformation (mm)

Unreinforced FB
2 Unreinforced RB
improved the performance by 41, 48, and 50% as compared with Geocomposite reinforced FB
unreinforced RB, respectively. The geosynthetic-reinforced RB per- Geocomposite reinforced RB
4
formed as good as or even better than the unreinforced FB in terms
of breakage consideration. 6
Indraratna, Nimbalkar, Christie, Rujikiatkamojorn, and Jayan 8
(2010) conducted full-scale moving wheel load tests on test
track sections. Four test sections (two unreinforced and two 10
geocomposite-reinforced, 15 m each in length) were prepared
12
and tested. The geocomposite consisted of biaxial geogrid and
nonwoven geotextile. The track had an overall bed thickness of 14
450 mm with a 300 mm thick ballast layer and a 150 mm thick (b)
capping layer. The particles smaller than 9.5 mm were removed
Fig. 14. Deformation behavior of unreinforced and geocomposite-reinforced ballast
from the recycled ballast. The capping materials consisted of a (redrawn and modified from Indraratna et al., 2010). (a) Vertical. (b) Lateral.
mixture of sand and gravel. The particle shapes of the fresh bal-
last (FB), the recycled ballast (RB), and the capping materials were
highly angular, semi-angular, and angular to rounded, respectively. Parsons, Jowkar, and Han (2012) conducted full-scale labo-
The geocomposite-reinforced FB performed best, followed by the ratory cyclic loading tests on one unreinforced and one triaxial
geocomposite-reinforced RB, RB, and FB. The RB performed bet- geogrid-reinforced recycled ballast railroad sections to investi-
ter than the FB because of the difference in gradation of ballast. gate the deformation and degradation behavior of unreinforced
The RB was moderately graded whereas the FB was relatively uni- and geogrid-reinforced RB. A full-scale trapezoidal railroad section
formly graded. The geocomposite improved the performance of FB consisted of a 600 mm thick recycled ballast layer over a 600 mm
significantly as compared with the RB. The geocomposite installed thick fat clay subgrade. The top and bottom widths of the railroad
at the interface of ballast and capping layer improved the perfor- cross-section were 2700 mm and 7500 mm, respectively. The sub-
mance of the reinforced track sections by reducing the vertical and grade and ballast layers were prepared at 2:1 slope with 2700 mm
lateral strains in the ballast layer as compared with that of the top width. The ties were embedded in the ballast to a depth of
unreinforced sections as shown in Fig. 14. 175 mm. The railroad track was placed over ties. The geogrid was
Fatahi and Khabbaz (2011) conducted finite element modeling installed 175 mm below the tie in case of the reinforced section.
using PLAXIS to investigate the effect of shoulder ballast width, bal- They concluded that geogrid reduced the vertical settlement of
last type, and geosynthetic on the deformation of a track due to train
load. The track section consisted of 3000 mm thick subgrade over-
62
laid by 150 mm thick sub-ballast, 300 mm thick ballast, 15 mm thick Unreinforced
Maximum settlement (mm)

Geogrid at Subgrade-sub ballast interface

tie, and 100 mm thick rail. The ballasts used were fresh ballast (FB), 60
Geogrid at Sub ballast-ballast interface
recycled ballast (RB), and blended ballast (BB). The blended ballast Geogrid at fresh ballast-recycled ballast interface
consisted of 50% FB and 50% RB (FB and RB, each of 150 mm thick). 58
The widths of shoulders were varied from 0 to 700 mm to inves-
56
tigate the shoulder width effect. The geogrid was installed at the
interface of subgrade and sub-ballast for the reinforced RB section 54
whereas for the reinforced BB sections, the geogrid was installed at
different locations to investigate the location effect. They concluded 52
that track settlement was reduced by increasing the shoulder bal-
50
last width; the reinforced BB section performed best, followed by
the reinforced RB, unreinforced FB, and unreinforced BB sections; 48
and the effective location of geogrid was at the interface of sub-
grade and sub-ballast. Fig. 15 shows the maximum settlements of 46
Fresh ballast Recycled ballast Blended ballast
the track in different sections. The test results for unreinforced FB
and unreinforced RB are not shown in Fig. 15 as the analysis was Fig. 15. Settlement behavior of unreinforced and geogrid reinforced ballasts
not conducted for these sections. (redrawn and modified from Fatahi & Khabbaz, 2011).

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011
G Model
SCS-147; No. of Pages 9 ARTICLE IN PRESS
J. Han, J.K. Thakur / Sustainable Cities and Society xxx (2014) xxx–xxx 9

the reinforced recycled ballast between the ties and geogrid by Donovan, H. (2011). Recycled aggregate and geosynthetic study – City of Edmon-
37–65% as compared with that of the unreinforced test section and ton. In Annual conference of the transportation association of Canada Edmonton,
Alberta, 19p.
increased the resistance against breaking under cyclic loading. Fatahi, B., & Khabbaz, H. (2011). Enhancement of ballasted rail track performance
using geosynthetics. In GeoHunan international conference (pp. 222–230).
6. Conclusions Federal Highway Administration (FHWA). (2004). Transportation applications of
recycled concrete aggregates. FHWA State of the Practice National Review.
Foye, K. C. (2011). Use of reclaimed asphalt pavement in conjunction with ground
This paper reviews the research work done in the past on the use improvement: A case history. Advances in Civil Engineering, 2011. Hindawi Pub-
of geosynthetics to stabilize recycled aggregates including recycled lishing Corporation, 7p.
Giroud, J. P., & Han, J. (2004). Design method for geogrid-reinforced unpaved roads.
asphalt pavement (RAP), recycled concrete aggregate (RCA), and I. development of design method. Journal of Geotechnical and Geoenvironmental
recycled ballast (RB). The following conclusions can be made: Engineering, 130(8), 775–786.
Gongora, I. A. G., & Palmira, E. M. (2012). Influence of fill and geogrid character-
istics on the performance of unpaved roads on weak subgrades. Geosynthetics
(1) 100% RAP and 100% RCA have been used with geosynthetic rein-
International, 19(2), 191–199.
forcement as base course materials for sustainable roadway Han, J., Acharya, B., Thankur, J. K., & Parsons, R. L. (2012). Onsite use of recycled
construction. asphalt pavement materials with geocells to reconstruct pavements damaged by
heavy trucks. Report no. 25-1121-0001-462. Mid-America Transportation Center.,
(2) 100% RB has been used with geosynthetic reinforcement
121p.
(woven geotextile, geogrid, and geogrid and non-woven geo- Han, J., Pokharel, S. K., Yang, X., Manandhar, C., Leshchinsky, D., Halahmi, I., et al.
textile composite) as a load bearing layer in railway track (2011). Performance of geocell-reinforced RAP bases over weak subgrade under
construction. full-scale moving wheel loads. ASCE Journal of Materials in Civil Engineering,
23(11), 1525–1534.
(3) Geocell improved the performance of RAP bases by reducing the Indraratna, B., Khabbaz, H., Salim, W., & Christie, D. (2006). Geotechnical properties
permanent and creep deformations and vertical stresses trans- of ballast and the role of geosynthetics in rail track stabilization. Journal of Ground
ferred to the subgrade, increasing the percentage of resilient Improvement, 10(3), 91–102.
Indraratna, B., Nimbalkar, S., Christie, D., Rujikiatkamojorn, C., & Jayan, V. (2010).
deformation and the modulus of the RAP bases. Field assessment of the performance of a ballasted rail track with and without
(4) Geogrid improved the performance of RCA bases in flexible geosynthetics. Journal of Geotechnical and Geoenvironmental Engineering, 136(7),
pavements by reducing the rate of crack propagation into the 907–917.
Indraratna, B., & Salim, W. (2003). Deformtion and degradation mechanism of
top layer of the pavement and spreading the load in the base recycled ballast stabilized with geosynthetics. Soils and Foundations, 43(4),
layer in a wider area. 35–46.
(5) Geogrid improved the life of reinforced RCA and FA base sec- Indraratna, B., Salim, W., & Christie, D. H. (2002). Improvement of recycled ballast
using geosynthetics. Rail International, 33(4), 20–29.
tions by factors of 33.5 and 15 at a permanent deformation
Indraratna, B., Shahin, M. A., & Salim, W. (2005). Use of geosynthetics for stabilizing
of 25 mm as compared with the corresponding unreinforced recycled ballast in railway track substructures. In Proceedings of NAGS2005/GRI
sections, respectively. 19 cooperative conference (pp. 1–15).
Liu, X., Scarpas, A., Blaauwendraad, J., & Genske, D. D. (1998). Geogrid reinforcing of
(6) Geosynthetics (geotextile, geogrid, and geotextile–geogrid
recycled aggregate materials for road construction: Finite element investigation.
composite) improved the degradation resistance of the rein- Transportation Research Record, 1611, 78–85.
forced RB by 41 to 50% as compared with the unreinforced RB. Marsal, R. J. (1967). Large scale testing of rockfill materials. Journal of Soil Mechanics
The geosynthetic-reinforced RB performed as good as or even and Foundation Engineering, 93(SM2), 27–43.
Maxwell, S., Kim, W., Edil, T. B., & Benson, C. H. (2005). Effectiveness of geosynthetics in
better than the unreinforced FB in terms of breakage consider- stabilizing soft subgrades. Report to the Wisconsin Department of Transportation.
ation. Missouri Asphalt Pavement Association. (2010). Recycling of asphalt pave-
(7) The performance of ballast depended on its gradation. The geo- ment. http://www.moasphalt.org/facts/environmental/recycling.htm Retrieved
10.12.10
composite (geotextile–geogrid) reduced the vertical and lateral Parsons, R. L., Jowkar, M., & Han, J. (2012). Performance of geogrid reinforced ballast
deformations of the uniform graded ballast more significantly under dynamic loading. Report no. 25-1121-0001-363. Mid-America Transporta-
as compared with the well-graded ballast. tion Center., 119 pp.
Qian, Y. (2009). Experimental study on triangular aperture geogrid-reinforced bases
(8) Geosynthetics (woven geotextile, geogrid, and nonwoven over weak subgrade under cyclic loading (MS thesis). CEAE Department, The Uni-
geotextile–geogrid composite) reduced the vertical and lat- versity of Kansas.
eral deformations and enhanced the long-term durability of Rajagopal, K., Krishnaswamy, N. R., & Madhavi Latha, G. (1999). Behaviour of sand
confined with single and multiple geocells. Geotextiles and Geomembranes, 17(3),
recycled ballast.
171–184.
Recycled Material Research Center (RMRC). (2008). User guideline for byproducts and
References secondary use materials in pavement construction. http://www.recycledmaterials.
org/tools/uguidelines/rcc4.asp Retrieved 24.05.12
Thakur, J. K., Han, J., & Parsons, R. L. (2013). Creep behavior of geocell-reinforced
Bortz, B. S., Hossain, M., Halami, I., & Gisi, A. (2012). Low-volume paved road
recycled asphalt pavement (RAP) bases. ASCE Journal of Materials in Civil Engi-
improvement with geocell reinforcement. In Transportation Research Board,
neering, 25(10), 1533–1543.
annual meeting (online publication).
Thakur, J. K., Han, J., Pokharel, S. K., & Parsons, R. L. (2012). Performance of geocell-
Das, B. M., & Shin, E. C. (1998). Strip foundation on geogrid-reinforced clay: Behavior
reinforced recycled asphalt pavement (RAP) bases over weak subgrade under
under cyclic loading. Geotextiles and Geomembranes, 13(10), 657–666.
cyclic plate loading. Geotextiles and Geomembranes, 35, 14–24.
Dong, Y. L., Han, J., & Bai, X. H. (2010, 20–24 February). A numerical study on
Webster, S. L. (1992). Geogrid reinforced base courses for flexible pavements for light
stress–strain responses of biaxial geogrids under tension at different directions.
aircraft: test section construction, behavior under traffic, laboratory tests, and design
In D. O. Fratta, A. J. Puppala, & B. Muhunthan (Eds.), Advances in analysis, mod-
criteria, final report, DOT/FAA/RD-92/25. U.S. Department of Transportation and
eling and design, Geotechnical special publication no. 199 (pp. 2551–2560). West
Federal Aviation Administration, 91 p.
Palm Beach, FL: GeoFlorida.

Please cite this article in press as: Han, J., & Thakur, J.K. Sustainable roadway construction using recycled aggregates with geosynthetics.
Sustainable Cities and Society (2014), http://dx.doi.org/10.1016/j.scs.2013.11.011