Available online at www.sciencedirect.

com

ScienceDirect
International Journal of Pavement Research and Technology 11 (2018) 114–127
www.elsevier.com/locate/IJPRT

Experimental evaluation of geocell-reinforced bases under

repeated loading
Sanat K. Pokharel a,1, Jie Han b,⇑, Dov Leshchinsky c, Robert L. Parsons b
a
Stratum Logics Inc, St. Albert, Alberta T8N 7L5, Canada
b
Civil, Environmental, and Architectural Engineering Department, The University of Kansas, Lawrence, KS 66045, USA
c
Department of Civil and Environmental Engineering, The University of Delaware, Newark, DE 19719, USA

Received 9 September 2016; received in revised form 8 March 2017; accepted 12 March 2017
Available online 21 March 2017

Abstract

Geocells, one type of geosynthetics manufactured in a form of three-dimensional interconnected cells, have been reported to effectively
provide lateral confinement to infill material to increase the modulus and bearing capacity of base courses. Most studies so far have been
focused on the behavior of geocell-reinforced bases under static loading. Geocells used for pavement applications are subjected to
repeated loading. Limited studies have been conducted to investigate the performance of geocell-reinforced bases under repeated loading.
In this study, single and multiple geocell-reinforced granular bases with three types of infill materials (Kansas River sand, quarry waste,
and AB-3 aggregate) were tested and compared with the unreinforced bases under repeated loading. This study experimentally investi-
gated the effect of the geocell reinforcement on the permanent deformation and percentage elastic deformation of the granular bases. The
test results showed that the geocell reinforcement reduced the permanent deformation and increased the percentage elastic deformation
of the granular bases. Multiple geocell-reinforced sections demonstrated even better performance as compared with single
geocell-reinforced sections.
Ó 2017 Chinese Society of Pavement Engineering. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Geosynthetic reinforcement; Geocells; Permanent deformation; Elastic deformation; Repeated loading; Base course

1. Introduction tenance requirements [1,2]. Therefore, it is considered as a

sustainable option to overcome premature pavement fail-
Geosynthetic reinforcement, one of the established tech- ure. Majority of the research on geosynthetics for pave-
niques of ground improvement for over 40 years, has been ment applications so far, has focused on planar
developed extensively to improve the performance of both reinforcement, such as geogrid and geotextile, and has
paved and unpaved roads. Geosynthetic reinforcement has resulted in several design methods [1–5]. However, the
been used to increase bearing capacity and modulus, three-dimensional forms of interconnected honeycomb
reduce required base thickness, extend the service life of geosynthetic cells, known as geocells, are not that com-
the pavement, reduce operational cost, and minimize main- monly used as compared with geogrids and woven geotex-
tiles. One major reason for this situation is that theories
⇑ Corresponding author. Fax: +1 7858645631.
and design methods for geocells have been lagging far
E-mail addresses: [email protected] (S.K. Pokharel),
behind the applications in the field [6].
[email protected] (J. Han), [email protected] (R.L. Parsons). The idea of cellular confinement was first developed by
1
Fax: +1 7804082259. the United States Army Corps of Engineers in 1970s [7].
Peer review under responsibility of Chinese Society of Pavement The geocells then were made of paper soaked in phenolic
Engineering.

http://dx.doi.org/10.1016/j.ijprt.2017.03.007
1996-6814/Ó 2017 Chinese Society of Pavement Engineering. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127 115

water resistant resin. Later metallic geocells were chosen to provide confinement and increase the modulus and
meet the strength requirements but they proved unfeasible strength of infill material, geocell-reinforced bases under
because of handling difficulty and high cost. Cellular struc- repeated loading have not been well investigated. In this
tures resembling geocells were also made from geogrids study, repeated load tests on single and multiple geocell-
forming the sides and diaphragms [8,9]. Geocells have also reinforced bases were carried out using three different infill
been made using geogrid sheets jointed by bodkin bars (for materials. In addition, NPA geocells manufactured using a
example, Carter and Dixon [10]). Commercially available new manufacture technology than HDPE were used in this
geocells are now made of high-density polyethylene study.
(HDPE) and novel polymeric alloy (NPA). NPA geocells This paper presents the results of an experimental study
are explained in the sections to follow. Geocells come in conducted with NPA geocells. Repeated plate load tests
different shapes and sizes; however, the most common were carried out on unreinforced, single geocell-
shape of geocell is nearly circular. reinforced, and multiple geocell-reinforced bases in-filled
Geocells provide enhanced confinement effect and with poorly-graded Kansas River sand (KR sand), QW,
impart apparent cohesion [11]; increase strength [12] and and well-graded AB-3. The influences of both single and
resilient modulus [13,14]; and significantly improve the multiple geocell reinforcements with different infill materi-
load-deformation and stress distribution characteristics of als are compared and evaluated in terms of permanent
poorly-graded materials [15]. The extent of bearing capac- deformation, traffic benefit ratio (TBR), and percentage
ity increase is correlated with the horizontal stiffness of the elastic deformation.
cell material [16] and the hoop stresses in the geocell wall
are the most significant contributing factor toward resisting 2. Material and test equipment
loads [17]. NPA geocell reinforcement reduces the plastic
deformation and increase the percent of elastic deforma- 2.1. Geocell type and characteristics
tion under repeated loading [18,19]. A series of static plate
load tests conducted by Pokharel et al. [20] showed that the Geocell made of novel polymeric alloy (NPA) was used
shape of geocell layout, the stiffness and type of geocell for the tests in this study. NPA is a nano-composite alloy
material, and the property of infill material all played vital of polyester/polyamide nano-fibers dispersed in polyethy-
roles in the behavior of geocell-reinforced bases under sta- lene matrix and characterized by flexibility at low temper-
tic loading. Pokharel et al. [20] recommended a near circu- atures similar to HDPE with elastic behavior similar to
lar shape of geocell layout as the most efficient one. engineering thermoplastic. NPA geocells have a lower ther-
Pavement failure is often caused by insufficient stiffness mal expansion coefficient and higher tensile stiffness and
and strength of the pavement structure including subgrade, strength than HDPE geocells. The coefficient of thermal
base, and asphalt or concrete surface, under heavy and expansion (CTE) of the Neoloy element used to make the
repeated traffic loading. Al-Qadi and Hughes [21] reported geocell, measured using ASTME831 was less than
an increase of the resilient moduli of aggregate layers by 80 ppm/°C in the measurement range from 30 °C to
about two times due to the installation of geocells within +30 °C. The geocell used in the experiments had the tensile
an asphalt paved road construction. Field tests with indus- strength of 19.1 MPa and the elastic modulus of 355 MPa
trial by-products as the infill material in the geocell over a at 2% strain. 2% strain was chosen to characterize the
period of 12-month was also found to achieve all the essen- stress–strain of geocell because the field studies have shown
tial performance requirements [22]. Han et al. [23] and Tha- that the measured strains in geosynthetics are typically
kur et al. [24] studied recycled asphalt pavement (RAP) within 2%. Almost identical test results on wide-width ten-
materials used as infill materials while in the Pokharel sile property of the geocell were obtained from three sam-
et al. [25] study three different infill materials were used ples of 100 mm wide NPA geocell tested by following the
including Aggregate Base Type 3 (AB-3), quarry waste ISO-527 test method. Fig. 1 shows the stress–strain curve
(QW), and RAP. Both studies showed the benefits of geo- of the NPA geocell, which was measured at a strain rate
cell reinforcement in reducing ruts if unreinforced and rein- of 10%/min at 23 °C in this study. The height of the geo-
forced sections are equally compacted. Under static cells used in this study was 100 mm and the wall thickness
loading, Pokharel et al. [20] found the modulus of the sin- was 1.1 mm. There were two perforations of 1 cm2 each on
gle geocell-reinforced bases improved by up to two times each pallet of the geocell. The spacing between the weld at
that of the unreinforced bases while the bearing capacities two ends of a single NPA cell and when expanded the cell
of the single geocell-reinforced bases were improved by up had internal dimension of 205 mm  235 mm. Fig. 2 shows
to 2.5 times those of the unreinforced bases. Thakur et al. the picture of the geocell used in this research.
[24] investigated the effect of geocell confinement on the
creep deformation on RAP base material and they found 2.2. Base course material
that the geocell confinement significantly reduced the creep
deformation of the RAP base material. In the present study, three different base materials, Kan-
Although a summary of the above-mentioned past stud- sas River (KR) sand, quarry waste (QW), and AB-3 aggre-
ies on geocell reinforcement confirms that the geocell can gate, were used as infill material. KR sand is locally
116 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127

20

16
Tensile stress (MPa)

12

0
0 4 8 12 16 20
Strain (%)

Fig. 1. Tensile stress–strain curve of the NPA geocell.

coefficient of uniformity = 2.73. The KR sand is classified

as SP according to the Unified Soil Classification System
(USCS). The peak angle of internal friction of the KR sand
at 70% relative density was 41°, which was obtained from
three triaxial tests. The details of these triaxial tests on
the KR sand can be found in the paper by Bhandari and
Han [26]. The QW used for the tests had a mean particle
size (d50) = 1.2 mm, liquid limit = 20, plastic limit = 12,
specific gravity = 2.76 at 20 °C, coefficient of curva-
ture = 0.77, and coefficient of uniformity = 12. The QW
is classified as SP-SC according to the USCS. Standard
Proctor tests and California Bearing Ratio (CBR) tests
were conducted for the QW. As shown in Fig. 4, the opti-
mum moisture content of the QW was 9% and its maxi-
mum dry density was 2.1 g/cm3. Fig. 5 shows the CBR
value of 57% at 7% moisture content and 38% at the opti-
Fig. 2. Picture of the NPA geocell used in the research. mum moisture content. The AB-3 aggregate used for the
tests is a well-graded base material widely used in pavement
applications in Kansas (USA). It had a mean particle size
available in the Lawrence area in Kansas, USA. KR sand
(d50) = 7.0 mm, liquid limit = 20, plastic limit = 13, specific
was chosen in this research to examine the possibility of
gravity = 2.69 at 20°C, optimum moisture content = 10%,
poorly graded and comparatively cheaper material in the
California Bearing Ratio (CBR) = 75% at 7.1% moisture
future with geocells. AB-3 is crushed limestone aggregate
content and 49% at the optimum moisture content, coeffi-
available in local quarry sites in Kansas and is the most
cient of curvature = 1.55, and coefficient of unifor-
common aggregate used in the road bases and unpaved
mity = 21. The AB-3 is classified as GW-GC according to
road construction in Kansas. QW is the waste material pro-
the USCS. Standard Proctor tests and CBR tests were con-
duced during aggregate production in quarries. The alter-
ducted for the AB-3 also. As shown in Fig. 4, the optimum
native QW was chosen to check if the waste material can
moisture content of the AB-3 was 10% and its maximum
be used as an infill in the geocell thus enhancing the sus-
dry density was 2.06 g/cm3. The compaction and CBR
tainable use of the waste at the quarry sites. The QW
curves for the AB-3 aggregate are also shown in Figs. 4
and AB-3 aggregate used in the tests were brought from
and 5.
a local limestone quarry site in Kansas. The grain size dis-
tribution curves of these infill materials are shown in Fig. 3.
The KR sand is a poorly-graded sub-rounded river sand 2.3. Model test setup
having a mean particle size (d50) = 2.6 mm, minimum void
ratio = 0.354, maximum void ratio = 0.583, specific grav- Model tests were conducted in a medium-scale loading
ity = 2.65 at 20°C, coefficient of curvature = 0.98, and apparatus designed and fabricated at Department of Civil,
 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127 117

100

80

% passing
60

40

Kansas River sand

20 Quarry waste
AB-3 aggregate

0
0.01 0.1 1 10 100
particle size in mm

Fig. 3. Grain size distribution curves of KR sand, QW, and AB-3 aggregate.

2.10
Dry density (Mg/m 3)

2.05

2.00

1.95
Quarry waste
AB-3 aggregate
1.90
6 7 8 9 10 11 12
Moisture content (%)

Fig. 4. Standard Proctor compaction curves of QW and AB-3 aggregate.

75

60
CBR (%)

45

30

15 Quarry waste

AB-3 aggregate
0
6 7 8 9 10 11
Moisture content (%)

Fig. 5. California Bearing Ratio curves of QW and AB-3 aggregate.

Environmental, and Architectural Engineering at the 900 kPa. The loading plate was 15 cm in diameter. Fig. 6
University of Kansas. The loading system had a 15-cm shows the details of the wooden test box, which was square
diameter air cylinder with a maximum air pressure of and had a plan area of 6400 cm2 with an adjustable height.
118 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127

Fig. 6. Test box with geocell layout.

Fig. 7 shows the pictures of the loading machine and Fig. 8 mum dry density on the drier side (approximately 7% mois-
shows the single and multiple geocell reinforcement placed ture content) of the compaction curve. The AB-3 was also
inside the box. For single geocell-reinforced sections, the compacted to a dry density approximately equal to 95% of
geocell was placed at the center of the box and laid out the maximum dry density on the drier side (approximately
in a near circular shape as suggested by Pokharel et al. 9% moisture content) of the compaction curve. For com-
[20]. The geocells were placed in a near circular shape as parison purposes, unreinforced bases were prepared in a
well for multiple geocell-reinforced bases. All geocells in similar way. No subgrade existed for all the tests because
this study were 10 cm high with a 2-cm thick fill cover the primary purpose of this research was to evaluate the
(same material as the base). For all the tests, the geocells influence of the geocell on the behavior of the base course.
were filled and embedded in the infill material. The KR A loading plate was placed at the center of the geocell
sand was placed and compacted to 70% relative density for the reinforced case or at the center of the box for the
inside and outside the cell in three layers: two 5-cm thick unreinforced case. Loads were applied in increments by
layers and one 2-cm cover layer. The QW was compacted adjusting air pressure in the air cylinder. The repeated load
to a dry density approximately equal to 95% of the maxi- was applied at 1 min/cycle in a trapezoidal form. The load
 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127 119

Fig. 7. Loading machine.

Fig. 8. Geocell layout.

was increased from the minimum (0 kPa) to the maximum after 150 cycles except the unreinforced sand that could
(345 kPa or 552 kPa) in 15 s. The maximum load was not hold the applied load (i.e., failure occurred). As
applied for 20 s and the load was released slowly in a per- Pokharel et al. [18–20] and Han et al. [27] showed, the ulti-
iod of 15 s to the minimum, which was maintained for 20 s mate bearing capacities of the unreinforced and geocell-
before starting the next loading cycle. The test was termi- reinforced AB-3 and QW bases are higher than the tire
nated when the percentage of elastic deformation reached pressure of a typical highway truck (i.e., 552 kPa). There-
constant at many loading cycles. To establish consistency fore, a repeated load of 552 kPa was applied for all the tests
in the comparison, all the tests therefore were terminated with AB-3 and QW bases. Pokharel et al. [20] showed that
120 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127

the ultimate bearing capacities of multiple geocell- confined test, the geocell was filled and embedded in the
reinforced, single geocell-reinforced, and unreinforced granular material. The infill material was placed into the
KR sands were 715, 500, and 230 kPa, respectively. Since box including the geocell and compacted to the desired
the ultimate bearing capacity of the multiple geocell- density. The layout of single and multiple geocell in the test
reinforced KR sand is higher than 552 kPa, a repeated load box are shown in Fig. 8a and b. A single geocell-reinforced
of 552 kPa was applied on the multiple geocell-reinforced section after the test is shown in Fig. 9.
KR sand. Since the ultimate bearing capacity of the single
geocell-reinforced KR sand was lower than 552 kPa, the 3.1. Pressure–displacement cycles
repeated load test on the single geocell-reinforced KR sand
was done at an applied pressure of 345 kPa (corresponding The pressure–displacement cycles of the geocell-
to approximately 70% of its ultimate bearing capacity, reinforced KR sand under repeated loading of 345 kPa were
which is also close to the tire pressure of typical construc- presented in Pokharel et al. [18] as a preliminary study. The
tion equipment). Due to the low ultimate bearing capacity pressure–displacement cycles of the unreinforced and multi-
of the unreinforced KR sand (230 kPa), a repeated load ple geocell-reinforced AB-3 under repeated loading of
test at either 552 kPa or 345 kPa pressure was impossible. 552 kPa are presented in Fig. 10. The pressure–displacement
For a comparison purpose, a repeated load test was also curves with the unloading and re-loading process were
performed on the multiple geocell-reinforced KR sand at obtained for all the unreinforced, single geocell-reinforced,
an applied pressure of 345 kPa. and multiple geocell-reinforced sections. However, due to
the space limit, only two curves from the tests on AB-3 aggre-
3. Test results and discussion gate sections are shown in Fig. 10(a) and (b). The permanent
deformations of the KR sand, QW, and AB-3 bases under
Before the repeated tests, the possible boundary effect of repeated loading are shown in Figs. 11–13, respectively.
the test box and the repeatability of the test results were The permanent deformation kept on accumulating with the
examined. As discussed by Pokharel et al. [20], the box size increasing number of the loading cycles during the test. It
of 80 cm  80 cm is large enough to eliminate the bound- is shown that in each cycle, there are elastic and plastic (also
ary effect. Pokharel et al. [20] also showed that this test called permanent deformation) deformations.
device had good repeatability of the test results. Test results The modulus values of the unreinforced and reinforced
of unreinforced, single geocell-reinforced, and multiple bases at the first loading cycle can be determined based
geocell-reinforced bases with three different infill materials on the slopes of the initial linear portions of the pres-
(KR sand, QW, and AB-3) under repeated loading are dis- sure–displacement curves. In a similar test with the KR
cussed below. For single geocell tests, unconfined and con- sand, the modulus of the single geocell-reinforced section
fined tests were conducted. For an unconfined test, the was approximately 1.5 times that of the unreinforced sec-
geocell was filled with granular material inside the cell only tion [18]. In the present study, the modulus improvement
without any surrounding soil outside the geocell. For a factors of the reinforced base to the unreinforced base

Fig. 9. Single geocell-reinforced section after the test.

 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127 121

Applied pressure (kPa)

0 100 200 300 400 500 600
0

1
Displacement (mm)

3 1 cycle

4 5 cycle
10 cycle

5 25 cycle
50 cycle
6 100 cycle
150 cycle

(a) Unreinforced AB-3 aggregate

Applied pressure (kPa)
0 100 200 300 400 500 600
0

1
Displacement (mm)

2 1 cycle

3 5 cycle
10 cycle
25 cycle
4 50 cycle
100 cycle
150 cycle
5

(b) Multiple geocell-reinforced AB-3 aggregate

Fig. 10. Pressure–displacement cycles under repeated loading.

are provided in Table 1. The modulus improvement factor cycles of loading. The same load in the case of multiple
is defined as the modulus ratio of a reinforced base to an geocell-reinforced KR sand section produced only 84% of
unreinforced base with the same infill material. that deformation after 150 loading cycles. This comparison
demonstrates that multiple geocell reinforcement further
3.2. Permanent deformation vs. number of cycles improved the performance of reinforced bases. Under a
higher applied pressure of 552 kPa, the multiple geocell-
Fig. 11 presents the cumulative deformation vs. number reinforced section deformed more than that under a lower
of cycles of the KR sand under three different conditions. pressure of 345 kPa. It is interesting to note that the defor-
For each test, there are two lines, which correspond to mation vs. number of cycle curves for the multiple geocell-
the cumulative deformations under a load and zero load. reinforced sections under these two different pressures are
The difference between these two curves at the same num- nearly parallel after 50 cycles.
ber of cycles is the elastic deformation. Since the unrein- Figs. 12 and 13 present the cumulative deformations of
forced KR sand section failed at 230 kPa [19,27], a unreinforced, single geocell-reinforced, and multiple
repeated load test under a pressure of 345 kPa was not pos- geocell-reinforced QW and AB-3 bases, respectively under
sible. However, the single geocell-reinforced KR sand sec- repeated loading of 552 kPa. It is clearly shown that both
tion under the repeated loading of 345 kPa survived 150 single and multiple geocell reinforcements reduced the
122 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127

20

Cumulative deformation (mm)

16

12

8 Single cell-reinforced - 0 kPa (unloaded from 345 kPa)

Single cell-reinforced - 345 kPa
Multicell-reinforced - 0 kPa (unloaded from 345 kPa)
4 Multicell-reinforced - 345 kPa
Multicell-reinforced - 0 kPa (unloaded from 552 kPa)
Multicell-reinforced - 552 kPa
0
0 25 50 75 100 125 150
Number of loading cycle

Fig. 11. Deformations of the KR sand bases under repeated loading.

Fig. 12. Cumulative deformations of the QW bases under repeated loading.

Fig. 13. Cumulative deformations of the AB-3 bases under repeated loading.
 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127 123

Table 1 552 kPa. Since the unreinforced KR sand could not sustain
Modulus improvement factors of the bases. the applied pressure of 552 kPa, no test data is shown in
Reinforced base Modulus improvement factor this figure. However, the test data of the multiple geocell-
Multiple geocell-reinforced KR 2.04 reinforced KR sand in Fig. 15 clearly show the significant
Single geocell-reinforced QW 1.26 benefit of geocell reinforcement in stabilizing the KR sand.
Multiple geocell-reinforced QW 1.46 Fig. 15 also shows that the AB-3 and QW bases had the
Single geocell-reinforced AB-3 1.32
Multiple geocell-reinforced AB-3 1.73
similar performance under the repeated loading, which
was much better than the reinforced KR sand. It is worth
pointing out that since the QW is more sensitive to mois-
cumulative deformations as compared with the unrein- ture than the AB-3, it may behave differently from the
forced base. The reduction in the cumulative deformation AB-3 when they are saturated. Further research is needed
started from the first loading cycle. The cumulative defor- to evaluate their behavior under a saturated condition.
mations for the QW and AB-3 bases with a single geocell
reinforcement measured at 150 cycles were reduced by a 3.3. Traffic benefit ratio
factor of 1.50 and 1.33 as compared with the unreinforced
QW and AB-3 bases, respectively, at the same number of The benefit of geocell reinforcement for extending pave-
cycles. The inclusion of multiple geocells reduced the ment life can be evaluated using a parameter of traffic ben-
cumulative deformations of the QW and AB-3 bases by efit ratio (TBR). TBR is defined as the ratio of the number
factors of 1.55 and 1.40, respectively. These comparisons of cycles necessary to reach a given rut depth (i.e., the per-
demonstrate that multiple geocell reinforcement further manent deformation) for a geocell-reinforced test section to
improved the performance of reinforced bases slightly. that for an unreinforced section at the same rut depth with
Pokharel et al. [20] showed that the geocell reinforcement the same section thickness and base and subgrade proper-
did not show any benefit in the performance of QW bases ties. The base thickness in all the test sections was 12 cm
under static loading because QW had apparent cohesion. and the subgrade in the present study was a hard wood sur-
This study clearly shows that geocell reinforcement face. Past research on other geosynthetic-reinforced bases
improved the performance of QW and AB-3 bases under showed that the TBR values depended on the level of per-
dynamic loading even though these materials have appar- manent deformation. Since the permanent deformations of
ent cohesion. most base sections at the end of 150 loading cycles in this
Fig. 14 presents a special case where the cumulative study were less than 4 mm, the TRB values were calculated
deformation of the confined, single geocell-reinforced at the permanent deformation of 3 mm. Since the unrein-
QW under repeated loading is compared with an uncon- forced KR sand failed under static loading before the max-
fined case. It is clearly shown that the confinement of the imum pressure was applied, the number of cycles required
geocell by the surrounding soil reduced the cumulative to reach a permanent deformation of 3 mm would be less
deformation. It is understandable that soil confinement than 1. Therefore, although the benefit of geocell reinforce-
and geocell confinement have the same effect. ment in the weak and poorly-graded KR sand was very
Fig. 15 shows a comparison of the permanent deforma- high compared to that of the stronger materials QW and
tions of unreinforced and geocell-reinforced bases with AB-3, the TBR values for KR sand sections were not cal-
three different infill materials under repeated loading of culated. Table 2 presents the TBR values for single and

Fig. 14. Cumulative deformations of unconfined and confined, single geocell-reinforced QW.
124 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127

Fig. 15. Comparison of permanent deformations under 552 kPa repeated loading.

Table 2 3.4. Elastic deformation

Calculated traffic benefit ratios (TBRs).
Infill Reinforcement Maximum applied pressure TBR The elastic deformations of unreinforced and single
material type (kPa)
geocell-reinforced sections with the number of loading
QW Single geocell 552 8.0 cycles are plotted in Fig. 16 for the QW and AB-3 sections.
QW Multiple geocell 552 12.0
The elastic deformations of the multiple geocell-reinforced
AB-3 Single geocell 552 8.5
AB-3 Multiple geocell 552 12.5 KR sand and unreinforced and multiple geocell-reinforced
QW and AB-3 sections are plotted in Fig. 17. Fig. 16 shows
multiple geocell-reinforced sections having different infill that the single NPA geocell-reinforced sections had higher
material. The geocell-reinforced QW and AB-3 bases had elastic deformation than the unreinforced sections of QW
comparable TRB values. The multiple geocell reinforce- and AB-3. This comparison demonstrates that the
ment had higher TBR values than the single geocell rein- geocell-reinforced bases had more elastic responses than
forcement. It should be pointed out that these TRB the unreinforced bases due to the inclusion of the geocell.
values were calculated to demonstrate the benefit of geocell The amount of elastic deformation was slightly higher at
reinforcement in pavement life, but they should not be used the beginning but later stabilized to a constant value.
for design because actual pavement sections may be much Fig. 17 shows that the multiple geocell-reinforced AB-3
different from the base sections in this study, for example, base had more elastic response than the multiple geocell-
the bases in this study were on a firm subgrade. reinforced sand, followed by the quarry waste.

0.9
Elastic deformation (mm)

0.6

Unreinforced QW
0.3 Single geocell-reinforced QW
Unreinforced AB-3
Single geocell-reinforced AB-3
0.0
0 25 50 75 100 125 150
Number of loading cycle

Fig. 16. Elastic deformations of unreinforced and single geocell-reinforced QW and AB-3 bases under 552 kPa repeated loading.
 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127 125

0.9

Elastic deformation (mm)

0.6

Multiple geocell-reinforced sand

0.3
Multiple geocell-reinforced QW

Multiple geocell-reinforced AB-3

0.0
0 25 50 75 100 125 150
Number of loading cycle

Fig. 17. Elastic deformations of multiple geocell-reinforced bases under 552 kPa repeated loading.

100

80
% elastic deformation

60

40 Single cell-reinforced loaded at 345 kPa

Multicell-reinforced loaded at 345 kPa

20 Multicell-reinforced loaded at 552 kPa

0
0 25 50 75 100 125 150
Number of loading cycle

Fig. 18. Percentage of elastic deformation for geocell-reinforced KR sand with loading cycle.

3.5. Percentage of elastic deformation Figs. 18 and 19 show that the representative percentage
of elastic deformation for all the test sections increased
Fig. 18 presents the elastic deformation as a percentage with the number of the loading cycles. At the initial loading
of the total deformation for geocell-reinforced KR sand cycles, the plastic deformation was more pronounced,
with the number of loading cycles under the repeated load- however, at around 10 cycles, the percentage of elastic
ing of 345 kPa and 553 kPa. Fig. 19 presents the percent- deformation increased rapidly with the loading cycles and
ages of elastic deformation of multiple geocell-reinforced became relatively stable. After 10 cycles, the percentage
sections with different base materials under 552 kPa of elastic deformation was more than 80% for the single
repeated loading. Similar results were obtained for all three geocell-reinforced KR sand and more than 95% for the
base materials: reinforced KR sand and both unreinforced geocell-reinforced QW and AB-3 bases. At 150 load
and reinforced QW and AB-3 bases. Both Figs. 18 and 19 cycles, the percentage of elastic deformation was 95.2%
represent the trend of those curves as well; therefore, those for the geocell-reinforced sand and more than 99% for
curves are not included here. The percentage of elastic the geocell-reinforced QW and AB-3 bases. The higher
deformation was calculated by dividing the elastic defor- percentage of elastic deformation is desirable for a longer
mation induced by each load cycle to the total deformation service life of a pavement section. Figs. 18 and 19 show that
(i.e., the sum of elastic and plastic deformations) at that geocell reinforcement increased the percentage of elastic
cycle. deformation in the reinforced section as compared with
126 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127

100

80
% elastic deformation

60

Sand
40
QW
AB-3
20

0
0 25 50 75 100 125 150
Number of loading cycle

Fig. 19. Percentages of elastic deformation of multiple geocell-reinforced sections with different base materials under 552 kPa repeated loading.

the unreinforced section, especially for the sections with the #DT0S59-06-G-00047, provided by the US Department
KR sand base. of Transportation – Research and Innovative Technology
Administration and PRS Mediterranean, Inc. in Israel.
4. Conclusions Their support is greatly appreciated. Mr. Milad Jowkar,
a former graduate student in the Department of Civil,
This paper presents the results of experimental work Environmental, and Architectural Engineering at the
conducted to investigate the behavior of novel polymer University of Kansas provided assistance in the testing.
alloy (NPA) geocell-reinforced bases under repeated load-
ing. The unreinforced and geocell-reinforced base courses References
with three different infill materials, Kansas River sand
(KR sand), quarry waste (QW), and AB-3 aggregates were [1] J.P. Giroud, J. Han, Design method for geogrid-reinforced unpaved
roads. I. Development of design method. ASCE, J. Geotech.
tested under repeated loading. The experimental investiga- Geoenviron. Eng. 130 (8) (2004) 775–786.
tions included the effect of infill material on the perfor- [2] J.P. Giroud, J. Han, Design method for geogrid-reinforced unpaved
mance of geocell-reinforced granular bases. The following roads. II. Calibration of applications. ASCE, J. Geotech. Geoenvi-
conclusions can be drawn from this experimental study: ron. Eng. 130 (8) (2004) 787–797.
[3] J. Steward, R. Williamson, J. Mohney, Guidelines for Use of Fabrics
in Construction and Maintenance of Low-Volume Roads. Report
1. The geocell-reinforced bases had higher initial modulus FHWA-TS-78-205, United States Department of Transportation,
than the unreinforced bases. Their modulus improve- Federal Highway Administration, Washington D.C., 1977.
ment factors ranged from 1.26 to 2.04. [4] J.P. Giroud, L. Noiray, Geotextile- Reinforced Unpaved Road
2. Geocell reinforcement significantly reduced the perma- Design, ASCE J. Geotech. Eng. Div. 107 (GT9) (1981) 1233–1254.
nent deformation as compared with the unreinforced [5] J.S. Tingle, S.R. Jersey, Empirical design methods for geosynthetic-
reinforced low-volume roads, J. Transport. Res. Board, No. 1989 2
bases of all three infill materials under repeated loading. (2007) 91–101.
Multiple geocells further reduced the permanent defor- [6] J. Yuu, J. Han, A. Rosen, R.L. Parsons, D. Leshchinsky, Technical
mation as compared with single geocell. Review of Geocell-Reinforced Base Courses over Weak Subgrade.
3. The traffic benefit ratio (TBR) values from this study Proceedings of the First Pan American Geosynthetics Conference &
were equal to or greater than 8.0 for the single geocell- Exhibition, 2–5 March 2008, Cancún, Mexico, 2008, pp. 1022–1030.
[7] S.L. Webster, Investigation of Beach Sand Trafficability Enhance-
reinforced bases and 12.0 for multiple geocell- ment Using Sand-Grid Confinement and Membrane Reinforcement
reinforced bases, respectively. Concepts. U.S. Army Engineer Waterways Experiment Station,
4. The geocell-reinforced QW and AB-3 bases had a higher Vicksburg, MS, Report GL-79-20 (1), 1979.
percentage of elastic deformation than the unreinforced [8] S. Edgar, The use of a high tensile polymer grid mattress on the
bases due to the contribution of the geocell. Except for Musselburgh and Protobello Bypass, in: Proceedings of the Polymer
Grid Reinforcement Conference, Thomas Telford, London, 1984,
the sand bases, geocell-reinforced AB-3 and QW bases 103–111.
reached 90% elastic deformation after the initial few [9] J. Paul, Economics and construction of blast embankments using
cycles (mostly 10 cycles). Tensar geogrids, in: Proceedings of the Polymer Grid Reinforcement
Conference, Thomas Telford, London, 1985, 191–197.
[10] G.R. Carter, J.H. Dixon, Oriented polymer grid reinforcement,
Acknowledgements Constr. Build. Mater. 9 (6) (1995) 389–401.
[11] R.J. Bathurst, R. Karpurapu, Large-scale triaxial compression testing
This research was funded jointly by the University of of geocell-reinforced granular soils, Geotech. Test. J., GTJODJ 16
Kansas, Transportation Research Institute from Grant (32) (1993) 296–303.
 S.K. Pokharel et al. / International Journal of Pavement Research and Technology 11 (2018) 114–127 127

[12] R. Gourves, P. Reiffsteck, J.F. Vignon, Study of confinement effect in [20] S.K. Pokharel, J. Han, D. Leshchinsky, R.L. Parsons, I. Halahmi,
geocells, Geosynth. Appl. Design Construct. 1 (1996) (1996) 455–458. Investigation of factors influencing behavior of single geocell-
[13] M.J. Mengelt, T.B. Edil, C.H. Benson, Reinforcement of Flexible reinforced bases under static loading, J. Geotextile Geomemb. 28
Pavements using Geocells. Geo Engineering Report No. 00–04, (2010) 570–578.
Geotechnical Engineering Program, Department of Civil & Environ- [21] I. Al-Qadi, J. Hughes, Field evaluation of geocell use in flexible
mental Engineering, University of Wisconsin-Madison, Madison, pavements, Geotech. Aspects Pavements, Transport. Res. Record,
Wisconsin, USA, 2000. No. 1709 (2000) 26–35.
[14] M.J. Mengelt, T.B. Edil, C.H. Benson, Resilient modulus and plastic [22] T.B. Edil, C.H. Benson, M.S.J. Shafique, B.F. Tanyu, W.H. Kim, A.
deformation of soil confined in a geocell, Geosynth. Int. 13 (5) (2006) Senol, Field Evaluation of Construction Alternatives for Roadway
195–205. over Soft Subgrade. Transportation Research Board, 81st annual
[15] B. Kazerani, G.H. Jamnejad, Polymer Grid Cell Reinforcement in meeting, January 13–17, 2002, Washington, D.C.
Construction of Pavement Structures, Section 1A, Unpaved and [23] J. Han, S.K. Pokharel, X. Yang, C. Manandhar, D. Leshchinsky, I.
Paved Roads. Geosynthetic ’87 Conference, New Orleans, USA, Halahmi, R.L. Parsons, Performance of geocell-reinforced RAP bases
1987. over weak subgrade under full-scale moving wheel loads, ASCE J.
[16] M. Shimizu, T. Inui, Increase in the Bearing Capacity of Ground with Mater. Civ. Eng. 23 (11) (2011) 1525–1534.
Geotextile Wall Frame. In: Proceedings of the Fourth International [24] J.K. Thakur, J. Han, R.L. Parsons, Creep behavior of geocell-
Conference on Geotextiles, Geomembranes and Related Products, reinforced recycled asphalt pavement (RAP) bases, ASCE J. Mater.
Vol. L. Hague, the Netherlands, 1990, pp. 254. Civ. Eng. 25 (10) (2013) 1533–1543.
[17] S.Y. Mhaiskar, Subgrade stabilization using geocells, ASCE Geotech. [25] S.K. Pokharel, J. Han, C. Manandhar, X.M. Yang, D. Leshchinsky,
Special Publ. 2 (30) (1992) 1092–1103. I. Halahmi, R.L. Parsons, Accelerated pavement testing of geocell-
[18] S.K. Pokharel, J. Han, D. Leshchinsky, R.L. Parsons, I. Halahmi, reinforced unpaved roads over weak subgrade, J. Transport. Res.
Behavior of Geocell-Reinforced Granular Bases under Static and Board, No. 2204 2 (2011) 67–75, Low-Volume Roads.
Repeated Loads. Contemporary topics in Ground Modification, [26] A. Bhandari, J. Han, ‘‘DEM study of a shallow foundation under
Problem Soils, and Geo-Support, in: M. Iskander, D.F. Laefer, M.H. vertical loading”. Contemporary Topics in Ground Modification,
Hussein (Eds.), International Foundation Congress & Equipment Problem Soils, and Geo-Support, Geotechnical Special Publication
Expo, March 15–19, 2009, Orlando, Florida. ASCE Geotechnical No. 187, in: M. Iskander, D.F. Laefer, and M.H. Hussein (Eds.),
Special Publication 187, 2009, pp. 409–416. International Foundation Congress & Equipment Expo 2009 -
[19] S.K. Pokharel, J. Han, R.L. Parsons, Y. Qian, D. Leshchinsky, I. IFCEE ’09, March 15–19, Orlando, Florida, 2009, 465–472.
Halahmi, Experimental Study on Bearing Capacity of Geocell- [27] J. Han, S.K. Pokharel, D. Leshchinsky, R.L. Parsons, I. Halahmi,
Reinforced Bases, in: Proceedings of 8th International Conference Effect of Infill Material on the Performance of Geocell-reinforced
on Bearing Capacity of Roads, Railways and Airfields, June 29–July Bases, in: 9th International Conference on Geosynthetics, ICG 2010,
2, 2009, Champaign, Illinois. May 23–27, 2010, Brazil.