International Journal of Pavement Engineering

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

Performance evaluation of geocell-reinforced

reclaimed asphalt pavement (RAP) bases in flexible
pavements

Anu Muthumala George, Aritra Banerjee, Anand J. Puppala & Manikanta

Saladhi

To cite this article: Anu Muthumala George, Aritra Banerjee, Anand J. Puppala & Manikanta
Saladhi (2019): Performance evaluation of geocell-reinforced reclaimed asphalt pavement
(RAP) bases in flexible pavements, International Journal of Pavement Engineering, DOI:
10.1080/10298436.2019.1587437

To link to this article: https://doi.org/10.1080/10298436.2019.1587437

Published online: 19 Mar 2019.

Submit your article to this journal

View Crossmark data

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

https://www.tandfonline.com/action/journalInformation?journalCode=gpav20
INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING
https://doi.org/10.1080/10298436.2019.1587437

Performance evaluation of geocell-reinforced reclaimed asphalt pavement (RAP)

bases in flexible pavements
Anu Muthumala Georgea, Aritra Banerjeeb, Anand J. Puppala a
and Manikanta Saladhia,c
a
Department of Civil Engineering, University of Texas, Arlington, TX, USA; bDivision of Automation and Intelligent System (AIS), University of Texas at
Arlington Research Institute, Fort Worth, TX, USA; cTolunay-Wong, Houston, TX, USA

ABSTRACT ARTICLE HISTORY

Reclaimed asphalt pavement (RAP) materials have been considered as one of the most sustainable options Received 22 July 2018
in the pavement industry. However, the low shear strength and high permanent deformation behaviour of Accepted 20 February 2019
RAP materials often limit their application in road bases. Stabilisation of RAP bases using high-density
KEYWORDS
polyethylene (HDPE) geocell was hence attempted in this study to assess the efficacy of three- HDPE geocell; repeated load
dimensional geosynthetic reinforcements in improving the resilient properties of RAP materials and tests; reclaimed asphalt
mitigating permanent deformation of them. The present paper discusses the results from a series of pavement (RAP); pavement
large-scale repeated load box tests conducted to study the effectiveness of HDPE geocell-reinforced design
RAP bases (GRRB). The performance of the geocell reinforcement was evaluated based on various
parameters including resilient deformation, cumulative permanent deformation, resilient modulus,
traffic benefit ratio (TBR), and rut depth reduction (RDR) factors. Test results show that HDPE geocell
layer increased the resilient modulus of the RAP base layer by 2.5–3.3 times and reduced the
permanent deformation of RAP base by 70–80% when compared to the unreinforced RAP bases. Test
results were then used to perform pavement design to quantify the base thickness reduction with
geocell reinforcement.

1. Introduction
The quest for sustainability in pavement infrastructure is focus- 1998, Taha et al. 1999, Rana 2004, Kazmee et al. 2016). This
ing primarily on recycling used materials such as quarried necessitates the adoption of a cost-effective chemical or mech-
aggregate material, reclaimed asphalt pavement (RAP) anical stabilisation method for the utilisation of RAP as pave-
materials, and crushed concrete (Puppala et al. 2019). Among ment base material.
these, RAP is the most widely used sustainable and cost-effec- Several researchers in the past have been studying the
tive option in the pavement industry (Papp et al. 1998). RAP is improvement in performance of pavement sections using
the reprocessed pavement material obtained through pavement RAP bases by mixing it with virgin aggregates or by treating
milling processes which generally contains asphalt and aggre- it with chemical stabilisers such as lime, cement, and fly ash
gates (Mulheron and O’Mahony 1990, Thakur et al. 2012). (Mohammad et al. 1995, Taha et al. 1999, Bennert et al.
Efforts have been made by various Department of Transpor- 2000, Taha et al. 2002, Gnanendran and Woodburn 2003, Pot-
tation (DOTs) and other federal agencies to increase the use turi 2006, Li et al. 2007, Potturi et al. 2007, Attia 2010, Deniz
of RAP materials in highway construction and rehabilitation et al. 2010, Hoyos et al. 2011, Puppala et al. 2011, Puppala
projects to mitigate issues regarding storage and to encourage et al. 2011, Mohammadinia et al. 2014, Avirneni et al. 2016,
the adoption of sustainable pavement materials (Das et al. LaHucik et al. 2016, Chakraborty et al. 2018, Puppala et al.
2018). According to Hansen and Copeland (2017), more than 2018). These studies recommend that the virgin aggregate-
76.9 million tons of RAP were used in new pavements in RAP mix treated with cementitious additive can be efficiently
United States during 2016 construction season. used in pavement bases. However, blending RAP with conven-
The use of higher amounts of RAP in pavements can signifi- tional aggregates still depletes non-renewable materials and the
cantly decrease the emission of greenhouse gases by minimising chemical treatment of RAP is not necessarily eco-friendly
the energy required to manufacture the raw materials for the (Sambodh 2017). This necessitates the adoption of a sustainable
virgin mix (NAPA 2009). Also, the recycling of RAP curtails and cost-effective mechanical stabiliser such as Geocell or other
the amount of debris to be discarded and reduces the cost of geosynthetic materials to enhance the utilisation and perform-
construction materials by replacing the virgin aggregates. How- ance of RAP base materials.
ever, unbounded RAP cannot be used as a total substitute for Geosynthetic reinforcements, primarily Geocell and others,
conventional virgin aggregates in pavement base layer due to have been widely used in pavement systems to provide struc-
its source-variability and inadequate strength and stiffness tural support and stability to the pavement layers. The use of
characteristics (Maher et al. 1997, Wilburn and Goonam geocell in reinforcing base materials has been proven to

CONTACT Anand J. Puppala anand@uta.edu

© 2019 Informa UK Limited, trading as Taylor & Francis Group
2 A. M. GEORGE ET AL.

increase the performance of the pavements by virtue of its performance of the base layer by reducing the resilient mod-
three-dimensional honey-comb structure which offers lateral ulus. This was prevented by placing non-woven geotextile at
confinement to the infill material (Al-Qadi and Hughes 2000, the interface of the base layer and subgrade as a separator to
Dash et al. 2007, Zhang et al. 2010, Han et al. 2011, George minimise the intrusion of clay particles into the RAP base
et al. 2019, George et al. 2019). Additionally, previous studies during loading. Geotextile was provided for both URRB and
on stabilisation of base materials using geocell indicate that GRRB sections. The effect of geotextile is not considered in
geocell reinforcement enhance the performance of the RAP this study as this is a comparative study of URRB and GRRB
material by offering an additional lateral confinement which with geotextiles acting as a separator for all test sections.
results in improved strength and stiffness behaviour (Pokharel Table 2 summarises the material properties of geotextile, as
et al. 2011, Bortz et al. 2012, Thakur et al. 2012). However, most provided by the manufacturer.
of these studies were limited to creep and permanent defor-
mation behaviour of geocell (Nano polymeric alloy) reinforced 2.1.2. Reclaimed asphalt pavement material
RAP and also the source-variability of the RAP material The Reclaimed Asphaltic Pavement (RAP) used in this study
necessitates further research in this area. was obtained from Texas Department of Transportation
In this study, the effectiveness of high-density polyethylene (TxDOT) stockpile in Arlington, Texas (West of Dallas). A
(HDPE) geocell reinforcement in RAP bases to reduce the series of laboratory box tests were initially carried out to
pavement thickness is being investigated. A series of large- characterise the RAP material. These tests included sieve analy-
scale repeated load box tests were performed on unreinforced sis (ASTM D1241-15 2015), compaction (TEX-2011-E 2011),
RAP base (URRB) sections and geocell-reinforced RAP base specific gravity (ASTM D854-00 2000), and repeated load triax-
(GRRB) sections to determine the structural support provided ial test (NCHRP 1-28A, Witczak 2004) and are listed in Table 3.
by geocell reinforcement. Test results were used to analyse the
elastic and plastic deformation response of GRRB under
repeated axial loading. The resilient modulus behaviour of 2.1.3. Clay subgrade
GRRB section was evaluated, and structural coefficients pro- The low plasticity clay, obtained from a site in Alvarado,
vided by the geocell-reinforced RAP materials were assessed Texas (South-west of Dallas), was used as the subgrade
for flexible pavement design. Additionally, pavement analysis material for this study. Material characterisation tests were
was also performed to evaluate the effect of geocell reinforce- carried out on the soil and are listed in Table 4. The subgrade
ment on rutting and fatigue behaviour of pavements. material was classified as lean clay (CL) based on Unified soil
classification system (USCS). The variation of moisture con-
tent and resilient modulus was also considered, as per rec-
ommendations of Banerjee (2017) and Banerjee et al. (2018).
2. Materials and methods
2.1. Test materials
Table 1. Properties of material used in cellular confinement system.
2.1.1. Geocell and geotextile Material properties Values
High-density polyethylene (HDPE) geocell was used as the Nominal expanded cell size (cm) 32 × 29
reinforcement to impart confinement to the RAP material. Nominal expanded cell area (cm2) 460
Cell depth (cm) 10.16
The geocell mattress used in this study is shown in Figure 1. Seam peel strength (N) 1423.43
Geocell with two different heights, 10, and 15 cm, were used Polymer density (kg/m3) 935.5–964.3
in this study. The properties of HDPE geocell including cell Carbon black content (% minimum by weight) 1.5
Nominal sheet thickness after texturing (mil) 60 (−5%, +10%)
size, cell depth, polymer density, and seam peel strength, pro-
vided by the manufacturer, are listed in Table 1. The contami-
nation of RAP base by clay particles can degrade the
Table 2. Properties of material used as separator.
Material properties Values
Grab tensile strength (N) 1110
Elongation (%) 50
Tear (N) 450
CBR puncture resistance 3110
Apparent opening size, AOS (microns) 150
Permittivity (s−1) 1.0
UV resistance (at 500 h) 70%
Flow rate (l/min/m2) 3056

Table 3. Properties of RAP material.

Material properties Values Standards
Specific gravity 2.49 ASTM D 854
Maximum dry density (kg/m3) 1960 Tex-113 E
Optimum moisture content (%) 7.9 Tex-113 E
Resilient modulus (MPa) 159 NCHRP I 28
Permeability (m/day) 1.6 ASTM D-2434
Figure 1. HDPE geocell used for testing.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 3

Table 4. Properties of clay subgrade.

Material properties Values Standards
Specific gravity 2.7 ASTM D 854
Liquid limit (%) 42.1 ASTM D 4318
Plasticity index (%) 17.1 ASTM D 4318
Maximum dry density (kg/m3) 1963 Tex-113 E
Optimum moisture content (%) 11.5 Tex-113 E
Resilient modulus (MPa) 76 AASHTO T 307–99

The resilient modulus of 76 MPa for the clay subgrade was

obtained by performing repeated load triaxial tests corre-
sponding to the last loading sequence of the AASHTO
T307-99 (2003) standard which is recommended for cohesive
soils. Since the overburden pressure was expected to be mod-
erate during the large-scale repeated load box testing, the resi-
lient modulus corresponding to the last loading sequence Figure 3. Large-scale laboratory testing facility.
(confining pressure = 13.8 kPa and maximum deviator stress
= 68.9 kPa) was adopted.
and a servo control unit. The repeated load cycles were applied
on the loading plate by pumping the hydraulic fluid into a ver-
2.2. Large scale laboratory tests tical actuator through a servo control unit.
A large-scale laboratory box setup was designed and fabricated The hydraulic fluid was recirculated through a flow control
to facilitate a series of repeated load tests to evaluate the per- system using an accumulator and hydraulic pump which was
formance of GRRB sections. The test set-up consists of a connected to the servo control unit. A constant high pressure
large steel tank of dimensions 1.83 m (length) × 1.83 m of 18 MPa (2600 psi) was maintained in the accumulator
(width) × 1.52 m (height), loading frame, accumulator, hydrau- throughout the test for carrying out the repeated load tests.
lic pump, cyclic load regulator, data acquisition system and lin- In addition, for controlling the repeated load application during
ear variable displacement transducers (LVDTs) (Saladhi 2017). the testing, a cyclic load regulator was connected to the servo
The schematic diagram for the entire test setup is shown in control unit. The cyclic load regulator can support simul-
Figure 2. The test box consists of detachable steel plates of taneous execution of multiple programmes and was connected
0.76 m (30 in.) height stacked and bolted in two layers on all with a Return Motion Control (RMC) software to control the
three sides of a 1.83 m × 1.83 m steel base plate, as shown in test using the motion controllers.
Figure 3. The algorithm offered by the RMC software allows the user
The front part of the box consists of three 0.51 m high to access position and control the speed of the actuator. It also
detachable steel plates for the easy placement of pavement allows the user to create their own test methods and protocols
layers inside the test box. A reference beam was placed at the for performing the repeated load tests at different frequencies.
top of the tank to attach the linear variable displacement trans- The axial load applied was measured using a load cell con-
ducers (LVDTs) with the help of adjustable magnetic holders. nected to the loading shaft and the corresponding displacement
Two adjustable iron beams were bolted on to the top of the at the surface of the loading plate was measured using two lin-
I-section which was fixed at the top of the tank to support ear variable displacement transducers (LVDTs). The data col-
the loading frame, which accommodates a hydraulic actuator lected from the large scale repeated load tests using the load
cell and LVDTs were transferred to the System 8000 StrainS-
mart data acquisition system.
Each test section included 0.3 m clay subgrade compacted at
95% maximum dry density (MDD) maintaining the water con-
tent at optimum moisture content (OMC). The subgrade was
placed in four equal lifts of 0.076 m (3 in.) height and each
lift was compacted to the required dry density using a vibratory
compactor (manual plate compactor). The compaction density
and subgrade strength were ensured by performing the sand-
cone test (ASTM D1556/D1556M-15e1 2015) and dynamic
cone penetrometer (ASTM D6951/D6951M-18 2018), respect-
ively. The URRB or GRRB layer was prepared over the clay sub-
grade by maintaining the density of RAP material at 95%
maximum dry density.
A non-woven geotextile was used between the base layer and
subgrade as a separator for preventing the penetration of RAP
into the subgrade during loading. The geocell mattress was
Figure 2. Schematic diagram of the large-scale laboratory setup. placed on the geosynthetic membrane and was filled with
4 A. M. GEORGE ET AL.

RAP material in three lifts by compacting each cell individually bed. The total surface deformation and the applied axial load
for each lift using a compactor. A 0.025 m (1 in.) RAP cover at the loading area were measured at an interval of 0.1 s
was placed over the GRRB section for the uniform application using two LVDTs and a load cell, respectively. The total surface
of the load on the surface of geocell reinforcement. This also deformation at the centre of the loading plate was measured by
helps the uniform compaction of the RAP inside the geocell averaging the deformation values obtained from the two
pockets. LVDTs placed on the surface of the loading plate.
A circular steel plate of 152.4 mm (6 in.) diameter and
12.7 mm (0.5 in.) thickness was used to simulate tire contact
area, which has a very high rigidity compared to the underlying 3. Analysis and test results
base layer, which is considered to simulate rigid loading plate. The stresses and strains developed at the loading surface
Repeated load tests were performed on the test bed by placing were calculated by analysing the data collected from the load
the circular steel plate at the centre of the actuator against the cell and LVDTs. A typical stress-strain plot from a repeated
reaction frame to prevent eccentric loading. Traffic load was load test with 100 load cycles is shown in Figure 5. The
simulated by applying a portion of equivalent single axle benefit of geocell reinforcement were also quantified by
wheel load (ESAL) repeatedly on the loading plate with the computing two indices such as traffic benefit ratio (TBR), and
help of a computer-controlled servo-hydraulic actuator. The rut depth reduction (RDR). Finally, the improvement in struc-
frequency of the load application was kept constant at 0.2 Hz tural coefficient was determined to quantify the reduction in
by applying a maximum load of 9.5 kN and a minimum load thickness of the pavement layers reinforced with geocell
of 0.95 kN in 5 s to replicate the traffic loading pattern. The reinforcement.
continuous load wave used for applying axial load is shown
in Figure 4. Tests were performed on both URRB and GRRB
(10-cm and 15-cm geocell height) sections for 1000 load cycles 3.1. Elastic deformation
as the resilient modulus and permanent deformation of the Elastic deformation at the loading surface of the test bed
GRRB sections were found to be constant after 700 cycles. under repeated loading was calculated from the rebound
Repeatability of tests was ensured by performing two trials curve, during unloading of each load cycle. The variation
for each case. of resilient deformation developed at the centre of the load-
ing plate with loading cycles for the URRB and GRRB sec-
tions are shown in Figure 6. The magnitude of resilient
2.3. Experimental program deformation for all the three cases decreased gradually to a
A series of six large-scale laboratory tests were conducted on constant value after 500 load cycles. The 15-cm GRRB sec-
different test sections: two-URRB sections, two 10-cm GRRB tion exhibit 2.5 times less resilient deformation compared
sections and two 15-cm GRRB sections. Thousand load cycles to the URRB section. This indicated that the lateral confine-
of 9.5 kN magnitude were applied on the loading plate for both ment offered by the geocell reinforcement played a major
URRB and GRRB sections and the performance of the geocell role in improving the resilient behaviour of the RAP base.
reinforcement on the RAP base layer was analysed based on This observation is in agreement with the results obtained
resilient modulus and permanent deformations of the test by Acharya (2011).

Figure 4. Haversine shaped load pulse used for repeated load testing. Figure 5. Typical stress-strain plot from repeated load test with 100 load cycles.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 5

Geocell reinforcement substantially improved the perma-

nent deformation behaviour of GRRB section by 70–80%
owing to the lateral confinement offered by the geocell pockets
on the infill RAP material. The improved performance may also
be due to the tension membrane effect (Thakur et al. 2012) in
which tensile forces are developed in geocell reinforcement
under repeated loading. This results in the formation of a
zone of tension at the lower section of geocell-reinforced
layer, which in turn increases the structural support offered
by the geocell thereby reducing the vertical deformation of
the subgrade. A similar trend in the variation of permanent
deformation was also observed by Thakur et al. (2017) in
which the permanent deformation of RAP base was improved
by a factor of 1.6 by the use of 15-cm geocell reinforcement.

3.3. Resilient modulus

Resilient modulus is the ratio of cyclic stress to the recoverable
strain. Resilient modulus was computed for the last five cycles
Figure 6. Variation of resilient deformation with number of load cycles. of 100, 250, 500, 750, and 1000th tests independently and the
average value was taken as the actual resilient modulus of the
test bed. The resilient modulus (Mr) of the entire test bed is
3.2. Cumulative permanent deformation given by
Cumulative permanent deformation is the plastic settlement sd
Mr = (1)
accumulated at the centre of the loading plate during each 1a
load cycle (Saride et al. 2015). The variation of cumulative per-
where σd is the deviatoric stress applied to the sample and εa is
manent deformation with the number of load cycles for the
the axial elastic strain developed due to the applied σd. Typical
URRB and GRRB sections are shown in Figure 7. The perma-
stress–strain plot for a load cycle and the elastic and plastic
nent deformation of both URRB and GRRB section increased
strains are shown in Figure 8.
with increase in load cycle, however, the rate of increase in per-
To evaluate resilient modulus, stresses developed at the mid-
manent deformation of GRRB section decreased with increase
height of each layer was calculated. For conservative analysis, a
in load cycles. The increase in permanent deformation is sig-
stress dispersion angle of 26° was used based on conventional 2
nificant until 500 load cycles and after that it remains relatively
Vertical to 1 Horizontal method for the unreinforced case and
constant.
30° (Thakur et al. 2011) for GRRB as shown in Figure 9.
Total elastic strain (εtot) developed in the test bed is the sum
of elastic strains developed in the individual layers as expressed

Figure 7. Variation of cumulative permanent deformation with number of load

cycles. Figure 8. Resilient Modulus from the stress-strain plot.
6 A. M. GEORGE ET AL.

respectively. The sharp increase in the initial phase might be

due to the rearrangement of particles under initial loading
and the confinement offered by the cellular structure of geocell
reinforcement, which resulted in a compact arrangement,
thereby increasing the (particle) interlocking between RAP par-
ticles and stiffness of the GRRB section. The initial increase in
stiffness is equivalent to the pre-conditioning cycles applied to a
traditional repeated-load triaxial test, where 500–1000 cycles
are applied prior to initiating the actual loading sequences.
Figure 9. Improvement in stress distribution angle due to geocell reinforcement. Similar observations were made by Thakur et al. (2010) for
RAP reinforced with NPA geocell, where the resilient modulus
in the following equation: of reinforced RAP was 1.6 times higher as compared to the
unreinforced RAP.
1tot = 1GR + 1s (2)

where εGR and εs are the elastic strains developed in GRRB layer 3.4. Traffic benefit ratio (TBR)
and subgrade, respectively. Equation (2) can also be expressed
as The extension of pavement life is typically indicated in terms of
a dimensionless index, Traffic Benefit Ratio (TBR). It is the
s1, GR s1, s ratio of the number of load cycles required to reach a given
1t = + (3)
Mr, GR Mr, s amount of rutting in GRRB section to the number of cycles
required in URRB section to reach the same amount of rutting.
where MGR and Ms are the resilient moduli of the geocell-
TBR shows the additional traffic loads that can be applied to the
reinforced RAP base layer and the subgrade, respectively; and
pavement when a geocell reinforcement is added, with all the
σ1,GR and σ1,s are the vertical stress at mid-height of GRRB
other pavement materials and geometry being equal. The
layer and the subgrade, respectively.
TBR is also denoted to as the traffic improvement factor and
Therefore, the resilient modulus of the GRRB layer is given
is given by
by
Nr
s1, GR TBR = (5)
Mr,GR = (4) Nu
1t − (s1,s /Mr,s )
where Nr is the number of load cycles on GRRB section and Nu
The variation of resilient modulus of URRB and GRRB sec- is the number of load cycles on the URRB section. Figure 11
tion with number of load cycles is shown in Figure 10. Resilient show the variation of TBR with number of load cycles. At the
modulus of GRRB section is 2.5–3.3 times higher than that/ end of 1000 load cycles, TBR value of 15-cm GRRB is 60%
those of the URRB section at the end of 1000 load cycles. more than that of the 10-cm GRRB showing thicker geocell
The rate of increase in resilient modulus of GRRB section is sig- layer provides stronger and resilient support to pavement
nificant till 750 cycles and then it reaches a constant value of system.
325 and 450 MPa for 10-cm and 15-cm height geocell,

Figure 10. Variation of resilient modulus with number of load cycles. Figure 11. Variation of traffic benefit ratio with number of load cycles.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 7

3.5. Rut depth reduction (RDR) practice (AASHTO 2008). For unreinforced sections, the tra-
ditional AASHTO 1993 method was adopted.
The rutting behaviour of GRRB is quantified in terms of a par-
According to AASHTO 1993, the number of cumulative 18-
ameter called rut depth reduction (RDR) factor. It is the ratio of
kip ESALs over the design life of the pavement is given by
the difference between cumulative permanent deformations of
the URRB section (Du) and the GRRB section (Dr) to that of the log W18 = ZR × S0 + 9.36 × log (SN + 1) − 0.2
URRB section at a specific number of load cycle (Saride et al.
log (DPSI/2.7)
2015). Hence, RDR for an nth load cycle can be expressed as + + 2.32

 0.4 + [1094/(SN + 1)5.19 ]
Dr
RDR = 1 − × 100 % (6) × log (Mr ) − 8.07 (7)
Du
where ZR is the standard normal deviate for reliability level, S0
Figure 12 shows the comparison of RDR of 10-cm and 15-
is the combined standard error of the traffic prediction and per-
cm GRRB sections. RDR initially increases and then attains an
formance prediction, SN is the structural number of the pave-
approximately constant value of 70% for 10-cm and 76% for
ment, ΔPSI is the allowable loss in serviceability, and Mr is the
15-cm GRRB section, respectively. This indicates that the geo-
resilient modulus of the subgrade.
cell height has a substantial effect on the rutting behaviour of
The structural number of the pavement is calculated using
GRRB sections. When the height of geocell is less than the
the nomograph of Equation (7) corresponding to the desired
width of loading plate, a membrane effect is developed under
design life and allowable loss in serviceability (W18 =
the application of wheel load and as the geocell height increases
1,000,000 ESALs, reliability = 95%, standard deviation = 0.45,
to more than the width of loading plate, a beam bending effect
ΔPSI = 1.7, and subgrade modulus = 76 MPa are adopted for
is developed in the geocell reinforced RAP layer (Thakur et al.
this study). Once the required overall SN has been estimated,
2012) which reduce the rutting of the pavement layers.
the base layer thickness for the unreinforced pavement (h2,u)
can be designed using the following equation:
3.6. Design of geocell-reinforced flexible pavement
SN − (a1 × h1 )
system h2, u = (8)
(a2 × m2 )
A modified American Association of State Highway and Trans-
portation Officials (AASHTO) 1993 method based on TBR by where a1 and a2 are the layer coefficient of surface and base
Holtz et al. (2008) was used for designing the geocell-reinforced courses, respectively; h1 is the thickness of the surface course;
pavement sections. Mechanistic-Empirical Pavement Design and m2 is the drainage coefficient (m2 = 1.2) for the base
(MEPD) method was not used in this study as it requires infor- layer. The surface and base layer coefficients are usually com-
mation regarding a number of complex and intricate variables puted from the resilient or elastic moduli of the respective
which are generally not available to practitioners. Additionally, layers. The structural coefficient of the surface layer a1 (corre-
the option for accommodating geocell reinforcement in the sponding to Mr,s = 2900 MPa) and the unreinforced RAP base
flexible pavement design is not currently available in AASHTO a2 are determined using AASHTO equation:
Mechanistic-Empirical Pavement Design Guide (MEPDG) of a1 = 0.398 × log Mr,s − 0.953 (9)

a2 = 0.249 × log Mr,b − 0.977 (10)

where Mr,s and Mr,b are the resilient moduli of surface (asphalt
concrete) and base courses, respectively.
For the geocell-reinforced case, a modified form of
AASHTO procedure based on TBR proposed by Holtz et al.
(2008) was used to estimate the reinforced base thickness.
The extended pavement life, W18r with geocell reinforcement,
expressed in terms of TBR and the design life without reinforce-
ment, W18u is given by
W18r = TBR × W18u (11)
The modified structural number, SNr corresponding to the
extended pavement life can be calculated from Equation (8).
The increase of structural number, ΔSN, which quantifies the
benefit of geocell reinforcement is given by the Equation (12).
DSN = SNr − SNu (12)
The increase in structural number with geocell reinforce-
ment is shown in Figure 13. The structural number increased
from 3.4 to 4.2 and 4.5 for 10-cm and 15-cm GRRB section,
Figure 12. Variation of rut depth reduction with number of load cycles. respectively.
8 A. M. GEORGE ET AL.

3.7. Fatigue and rutting analyses

To compare the influence of geocell reinforcement on the fati-
gue and rutting behaviour of pavement, a layered elastic analy-
sis was performed on the an URRB and 15-cm GRRB sections
with same base thickness (180 mm = 7 in.). A typical steer axle
load of 32 kN (7200 lbs) was applied at a radius of 11 cm
(4.37 in.) to simulate a tire pressure of 827 kPa (120 psi) as
shown in Figure 15. Strains developed at the two critical
location of the pavement were considered for the analysis:
The horizontal tensile strain (εt) developed at the bottom of
the HMA layer and the vertical compressive strain (εv) at the
top of the subgrade. The fatigue life (Nf) was computed from
εt and the rutting life (Nr) from εv using fatigue (Timm et al.
1999) and rutting (Asphalt Institute 1982) models, respectively
and are tabulated in Table 5.

 k2
1 −6
Nf = k1 × 10 (14)
1t
Figure 13. Improvement in SN with geocell reinforcement.
k4
10−6
Nr = k 3 × (15)
1v
The reduced thickness of the GRRB can be calculated from,
where k1 = 2.830, k2 = 3.148, k3 = 1.077 × 1018 and k4 = 4.484.
The analysis result shows that the geocell reinforcement
(SNu − DSN) − a1 h1 increased the fatigue and rutting life by approximately 8 and
h2,r = (13)
a2 × m2 30 times, respectively. The rutting life enhancement is attribu-
ted to an increase in resilient modulus of the base layer, which
in turn resulted from the presence of additional lateral confine-
Figure 14 shows the reduction in thickness of the base layer
ment offered by the geocell reinforcement.
with the geocell reinforcement. The reduction in base thickness
Overall, the present experimental programme and sub-
for 10-cm and 15-cm GRRB sections were computed to be
sequent pavement design analyses showed that the Geocell
around 30% and 50% of the URRB section, respectively. This
analsysis indicates that both the improved strength and stiff-
ness characteristic of the RAP material due to the lateral
confinement offered by geocell walls will result in an improve-
ment pavement design with lesser distress and improved pave-
ment performance.

Figure 15. Geometry of typical steer axle wheel load configuration.

Table 5. Fatigue and rutting life of pavements with same base thickness.
Thickness εt εv Nf Nr
Pavement type (mm) (×10−6) (×10−6) (cycles) (cycles)
Unreinforced 130 802 885 15,756 65,650
10-cm geocell- 130 542 552 54,095 545,178
reinforced
Unreinforced 180 782 868 16,788 71,616
15-cm geocell- 180 397 396 139,665 2,417,526
reinforced
Notes: εt is the horizontal tensile strain, εv is the vertical compressive strain, and Nf
Figure 14. Reduction in base thickness with geocell reinforcement. and Nr are the fatigue and rutting life of the pavement, respectively.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 9

system can be utilised with reclaimed asphalt pavement requires field implementation of GRRB sections under real-
aggregates that could lead to enhanced pavement perform- time traffic loading to study the long-term behaviour.
ance with minimal distress. Significant cost savings can be
realised over a longer design life period as the savings
obtained from aggregate layer reductions can be offset by Acknowledgement
the costs of geocell layer itself. Other intangible sustainable Authors would like to thank their support. Authors would like to express
and green pavement system benefits can be realised as the thanks to Geo Products, LLC for providing Envirogrid geocells for the
use of RAP materials will promote use of recycled materials research, and also thank other researchers including Dr. Aravind Pedarla,
which in turn can result in lesser landfilling costs and also Leopoldo C. Martizez, and others for helping with large-scale laboratory
testing.
can offer benefits of reducing native aggregate production
costs associated with air polluting emission costs from the
crushers and other equipment used in the aggregate pro-
Disclosure statement
duction plants.
No potential conflict of interest was reported by the authors.

4. Summary and conclusions

Large-scale experimental box tests were performed to quantify Funding
the benefits of using HDPE geocell-reinforced RAP material The research was funded by Texas Department of Transportation
as the pavement base layer in terms of resilient deformation, (TxDOT) – Fort Worth District (Mr. Richard WIlliammee, PE and Mr.
cumulative permanent deformation, resilient modulus, Paul Spraggins, PE; Project 02-7XXIA001) and NSF Industry-University
Cooperative Research Center (IUCRC) program funded ‘Center for the
traffic benefit ratio (TBR), and rut depth reduction (RDR). Integration of Composites into Infrastructure (CICI)’ site at UTA (NSF
The HDPE geocell increased the resilient modulus of the PD: Andre Marshall; Award #1464489).
base layer by 2.5–3.3 times and reduced the permanent defor-
mation of RAP base by 70–80% for 10-cm and 15-cm GRRB,
respectively. The improvement in performance is primarily ORCID
due to the increase in stress distribution angle caused by the
Anand J. Puppala http://orcid.org/0000-0003-0435-6285
lateral distribution of stresses through the interconnected geo-
cell pockets.
The repeated loading on GRRB sections resulted in a confi- References
ning mechanism within the geocell pockets, that resulted in lat-
eral distribution of stresses developed under loading. This AASHTO, 2008. Mechanistic-empirical pavement design guide, a manual
of practice. Interim edition Washington, DC: American Association
lowered the vertical stresses and permanent deformations on of State Highways and Transportation Officials.
the pavement subgrade. The magnitude of elastic deformation Acharya, B., 2011. Experimental study on geocell-reinforced flexible pave-
for all the three cases decreased gradually to a constant value ments with recycled asphalt pavement (RAP) bases under cyclic loading.
after 500 load cycles. Additionally, the resilient modulus and Thesis (MS). The University of Kansas.
permanent deformation values were observed to gradually Al-Qadi, I. and Hughes, J., 2000. Field evaluation of geocell use in flexible
pavements. Transportation Research Record: Journal of the
reach a constant after 500 load cycles. Hence, a minimum num- Transportation Research Board, 1709, 26–35.
ber of 500 load cycles is recommended as pre-conditioning American Association of State Highway and Transportation Officials
cycles for large scale repeated load box tests. (AASHTO), 1993. Guide for design of pavement structures.
The test results were used to perform pavement design to Washington, DC: American Association of State Highway and
quantify the reduction in base thickness with geocell Transportation Officials (AASHTO).
American Association of State Highway and Transportation Officials
reinforcement. It was observed that the geocell reinforcement (AASHTO) T307-99, 2003. Design guidelines standard method of test
significantly reduced the thickness of the RAP base layer. This for determining the resilient modulus of soils and aggregate materials.
is due to the increase in stiffness of the RAP material caused Washington, DC: American Association of State Highway and
due to the lateral confinement and tension membrane effect Transportation Officials (AASHTO).
offered by the geocell reinforcement. The increase in base stiff- Asphalt Institute, 1982. Research and development of the Asphalt Institute’s
thickness design manual. College Park, MD: Asphalt Institute.
ness reduced the vertical stress transferred to the subgrade ASTM D1241-15, 2015. Standard specification for materials for soil-aggre-
thereby improving the performance of the entire pavement gate subbase, base, and surface courses. West Conshohocken, PA: ASTM
section. A linear elastic analysis was also performed on the International.
designed pavement section to analyse the influence of geocell ASTM D1556/D1556M-15e1, 2015. Standard test method for density and
reinforcement in the rutting and fatigue life of pavement. The unit weight of soil in place by sand-cone method. West
Conshohocken, PA: ASTM International.
analysis result showed that the geocell reinforcement ASTM D6951/D6951M-18, 2018. Standard test method for use of the
increased the fatigue and rutting life by approximately 8 and dynamic cone penetrometer in shallow pavement applications. West
30 times, respectively. Cost savings of these layers will be Conshohocken, PA: ASTM International.
high as this system promotes the use of RAP types of recycled ASTM D854-00, 2000. Standard test methods for specific gravity of soil
base aggregates while reducing landfilling and air pollution solids by water pycnometer. West Conshohocken, PA: ASTM
International.
emissions. Attia, M. I. E., 2010. Characterization of the structural behavior of
This study is limited to large scale laboratory studies under reclaimed asphalt pavement as pavement base layer. Dissertation
repeated loading on one type of RAP material and further (PhD). The North Dakota State University.
10 A. M. GEORGE ET AL.

Avirneni, D., Peddinti, P. R. T., and Saride, S., 2016. Durability and long- Maher, M. H., Gucunski, N., and Papp Jr, W. J., 1997. Recycled asphalt
term performance of geopolymer stabilized reclaimed asphalt pavement pavement as a base and sub-base material. ASTM Special Technical
base courses. Construction and Building Materials, 121, 198–209. Publication 1275. West Conshohocken, PA: American Society of
Banerjee, A., 2017. Response of unsaturated soils under monotonic and Testing and Materials, 42–53.
dynamic loading over moderate suction states. Doctoral Dissertation. Mohammad, L.N, Puppala, A.J., and Alavilli, P., 1995. Resilient Properties
University of Texas at Arlington, Arlington, TX. of Laboratory Compacted Subgrade Soils. Transportation Research
Banerjee, A., Patil, U.D., Puppala, A.J., et al., 2018. Suction-controlled Record, 1196, 87–102.
repeated load triaxial test of subgrade soil at high suction states. In: Mohammadinia, A., et al., 2014. Laboratory evaluation of the use of
C.W.W. Ng, A.K. Leung, A.C.F. Chiu and C. Zhou, eds. Unsaturated cement treated construction and demolition materials in pavement
Soils, Proceeding of Seventh International Conference on Unsaturated base and subbase applications. Journal of Materials in Civil
Soils. Hong Kong: HKUST, 667–672. Engineering, 28 (7), 899–1561.
Bennert, T., et al., 2000. Utilization of construction and demolition debris Mulheron, M. and O’Mahony, M.M., 1990. Properties and performance of
under traffic-type loading in base and subbase applications. recycled aggregates. Journal of Highway and Transportation Research
Transportation Research Record: Journal of the Transportation and Development, 37 (2), 35–37.
Research Board, 1714, 33–39. NAPA, 2009. Black and green sustainable asphalt, now and tomorrow.
Bortz, B.S., et al., 2012. Low-volume paved road improvement with geocell Special report 200. National Ashpalt Pavement Association. Available
reinforcement. No. 12-3662. Washington, DC: Transportation Research from: https://www.asphaltpavement.org/images/stories/sustainability_
Board. report_2009.pdf [Accessed 21 January 2018].
Chakraborty, S and Nair, S, 2018. Impact of curing time on moisture- Papp Jr, W. J., et al., 1998. Behaviour of construction and demolition debris
induced damage in lime-treated soils. International Journal of in base and subbase applications. Proceedings of geo-congress: recycled
Pavement Engineering, 1–13. http://dx.doi.org/10.1080/10298436.2018. materials in geotechnical applications. Reston, VA: American Society of
1453068. Civil Engineers, 122–136.
Das, J.T., Banerjee, A., Chakraborty, S., et al., 2018. A Framework for Pokharel, S., et al., 2011. Accelerated pavement testing of geocell-
Assessment of Sustainability and Resilience in Subgrade Stabilization reinforced unpaved roads over weak subgrade. Transportation
for a High-Volume Road. Transportation Research Board 97th Research Record: Journal of the Transportation Research Board, 2204
Annual Meeting, January 7-10, 2018, Washington, DC. (1), 67–75.
Dash, S.K., Rajagopal, K., and Krishnaswamy, N.R., 2007. Behaviour of Potturi, A.K, 2006. Evaluation of resilient modulus of cement and cement-
geocell-reinforced sand beds under strip loading. Canadian fiber treated reclaimed asphalt pavement (RAP) aggregates using
Geotechnical Journal, 44 (7), 905–916. repeated load triaxial test. Thesis (MS). The University of Texas at
Deniz, D., Tutumluer, E., and Popovics, J. S., 2010. Evaluation of expensive Arlington.
characteristics of reclaimed asphalt pavement and virgin aggregate used Potturi, A., Puppala, A. J., and Hoyos, L. R., 2007. Resilient characteristics
as base materials. Transportation Research Record: Journal of of cement treated reclaimed asphalt pavement aggregates. No. 07-1838.
Transportation Research Board, 2167, 10–17. Washington, DC: Transportation Research Board.
George, A.M., Banerjee, A, Puppala, A.J., et al., 2019. An integrated LCA- Puppala, A.J., Congress, S.S.C., and Banerjee, A., 2019. Research
LCCA framework for selection of sustainable pavement design. Advancements in Expansive Soil Characterization, Stabilization and
Transportation Research Board 98th Annual Meeting, January 13–17, Geoinfrastructure Monitoring. Frontiers in Geotechnical Engineering,
2019, Washington, DC. 15–29. http://dx.doi.org/10.1007/978-981-13-5871-5. Book section
George, A.M., Banerjee, A., Taylor, T., et al., 2019. Large-Scale (Springer Publishers, Singapore).
Experimental Studies to Evaluate the Resilient Modulus of Geocell- Puppala, A.J., Congress, S.S.C., Bheemasetti, T.V., et al., 2018.
Reinforced Reclaimed Asphalt Pavement Bases. Geosynthetics Visualization of Civil Infrastructure Emphasizing Geomaterial
Conference 2019, February 10-13, Houston, Texas.. Characterization and Performance. Journal of Materials in Civil
Gnanendran, C. T. and Woodburn, L. J., 2003. Recycled aggregate for pave- Engineering, 30 (10), 10.1061/(ASCE)MT.1943-5533.0002434.
ment construction and the influence of stabilization. Proceedings on Puppala, A.J., Hoyos, L.R., and Potturi, A.K., 2011. Resilient moduli
conference of the Australian road research board. Leederville, response of moderately cement-treated reclaimed asphalt pavement
Australia: 21, 1755–1768. aggregates. Journal of Materials in Civil Engineering, 23 (7), 990–998.
Han, J., et al., 2011. Performance of geocell-reinforced RAP bases over Puppala, Anand J., Manosuthkij, Thammanoon, Nazarian, Soheil, et al.,
weak subgrade under full-scale moving wheel loads. Journal of 2011. Threshold moisture content and matric suction potentials in
Materials in Civil Engineering, 23 (11), 1525–1534. expansive clays prior to initiation of cracking in pavements. Canadian
Hansen, K.R. and Copeland, A., 2017. Annual asphalt pavement industry Geotechnical Journal, 48 (4), 519–531. http://dx.doi.org/10.1139/t10-087.
survey on recycled materials and warm-mix asphalt usage: 2016, 7th Rana, A. S. M. A., 2004. Evaluation of recycled material performance in
annual survey (IS 138). Lanham, MD: National Asphalt Pavement highway applications and optimization of their use. Dissertation
Association. (PhD). Texas Tech University.
Holtz, R.D., Christopher, B.R., and Berg, R.R., 2008. Geosynthetic design Saladhi, M., 2017. Experimental study on resilient behaviour of geocell-
and construction guidelines, reference manual (NHI course No. reinforced recycled asphalt pavement base layer: model development.
132013). FHWA Publication No. FHWA-HI-95-038. Washington, Thesis (Masters). The University of Texas at Arlington.
DC: Federal Highway Administration. Sambodh, A., 2017. Mechanical properties of soil-RAP-geopolymer for the
Hoyos, L. R., Puppala, A. J., and Ordonez, C. A., 2011. Characterization of stabilization of road base/subbase. Thesis (MS). University of
cement fibre-treated reclaimed asphalt pavement aggregates: Louisiana at Lafayette.
Preliminary investigation. Journal of Materials in Civil Engineering, Saride, S., Rayabharapu, V. K., and Vedpathak, S., 2015. Evaluation of rut-
23 (7), 977–989. ting behaviour of geocell reinforced sand subgrades under repeated
Kazmee, H., Tutumluer, E., and Beshears, S., 2016. Using accelerated pave- loading. Indian Geotechnical Journal, 45 (4), 378–388.
ment testing to evaluate reclaimed asphalt pavement materials for pave- Taha, R., et al., 1999. Evaluation of reclaimed asphalt pavement aggregate
ment unbound granular layers. Journal of Materials in Civil in road bases and subbases. Transportation Research Record: Journal of
Engineering, 29 (2), 1–13. the Transportation Research Board, 1652, 264–269.
LaHucik, J., et al., 2016. Cement-treated bases containing reclaimed asphalt Taha, R., et al., 2002. Cement stabilization of reclaimed asphalt pavement
pavement, quarry by-products, and fibers. Transportation Research aggregate for road bases and subbases. Journal of Materials in Civil
Record: Journal of the Transportation Research Board, 2580, 10–17. Engineering, 14 (3), 239–245.
Li, L., et al., 2007. Evaluation of recycled asphalt pavement material stabil- TEX-113-E, 2011. Laboratory compaction characteristics and moisture-
ized with fly ash. Soil and Material Inputs for Mechanistic-Empirical density relationship of base materials. Austin, TX: Texas Department
Pavement Design, 169, 1–10. of Transportation.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 11

Thakur, S.C., et al., 2010. Laboratory evaluation of physical and mechan- Timm, D.H., Newcomb, D.E., Birgisson, B., et al., 1999. Incorporation of
ical properties of recycled asphalt pavement as a base course material. reliability into the Minnesota mechanistic-empirical pavement design
Geotechnical Special Publication, 203, 255–263. method. Minnesota: Minnesota Department of Transportation, (Report
Thakur, J.K., 2011. Geocell-reinforced unpaved and paved roads with No. MN/RC-1999-35)..
recycled asphalt pavement (RAP) bases: experimental study and damage Wilburn, D. R. and Goonam, T. G., 1998. Aggregates from natural and
model development. (Doctoral dissertation). Department of Civil, recycled sources: economic assessments for construction applications –
Environmental, and Architectural Engineering, University of Kansas, a materials flow analysis. USGS Circular, 1176. Reston, VA: US
Lawrence, KS.. Department of Interior US Geological Survey.
Thakur, J.K., et al., 2012. Performance of geocell-reinforced recycled Witczak, M. W., 2004. Harmonized test methods for laboratory determi-
asphalt pavement (RAP) bases over weak subgrade under cyclic plate nation of resilient modulus for flexible pavement design – NCHRP 1-
loading. Geotextiles and Geomembranes, 35, 14–24. 28A. Washington, DC: National Cooperative Highway Research
Thakur, J. K., Han, J., and Parsons, R. L., 2017. Factors influencing defor- Program, Transportation Research Board.
mations of geocell-reinforced recycled asphalt pavement bases under Zhang, L., et al., 2010. Bearing capacity of geocell reinforcement in
cyclic loading. Journal of Materials in Civil Engineering, 29 (3), 899– embankment engineering. Geotextiles and Geomembranes, 28 (5),
1561. 475–482.