3

9th International Conference on Geosynthetics, Brazil, 2010

Geosynthetics in pavements: North American contributions

Zornberg, J.G. & Gupta, R.
The University of Texas at Austin, USA

Keywords: geosynthetics, mechanisms, geogrids, geotextiles, reinforcement, road, pavement, design

ABSTRACT: Geosynthetics have been used to reinforce the base layer of flexible pavement systems for the
past thirty years. However, and in spite of the good field evidence that geosynthetic reinforcements can improve pavement performance, the specific conditions or mechanisms that enable and govern the reinforcement
function are, at best, unclear as they have remained largely unmeasured. Overall, the selection of design parameters for geosynthetics has been complicated by the difficulty in associating their relevant properties to the
improved pavement performance. Nonetheless, significant research has been conducted with the objectives
of: (i) determining the governing mechanisms and relevant properties of geosynthetics that contribute to the
enhanced performance of pavement systems, (ii) developing appropriate analytical, laboratory and field methods capable of quantifying the above properties for geosynthetics, and (iii) enabling the prediction of pavement performance depending on the various types of geosynthetics used. This paper summarizes research
conducted to address these objectives with specific focus on contributions made in North America.
1 INTRODUCTION

are smaller). Adherence to this principle makes

possible the use of local materials and usually results
in an economical design (Huang, 1993). A typical
flexible pavement system includes four distinct layers: asphalt concrete, base course, subbase, and subgrade (Figure 1). The surface layer is typically asphalt concrete, which is a bituminous hot-mix
aggregate (HMA) obtained from distillation of crude
petroleum. The asphalt concrete is underlain by a
layer of base course, typically consisting of 0.2 m to
0.3 m of unbound coarse aggregate. An optional
subbase layer, which generally involves lower quality crushed aggregate, can be placed under the base
course in order to reduce costs or to minimize capillary action under the pavement. The constructed layers are placed directly onto a prepared subgrade,
which is generally graded and compacted natural insitu soil.

The use of geosynthetic reinforcements in regions of

an earth structure under tensile stresses helps inhibiting extension strains within the soil, thereby increasing the overall strength of the composite material
(Jewell 1981, Palmeira 1987). Common uses of
soil-geosynthetic reinforcement include the construction of roads, retaining walls, foundations and
embankments. This paper focuses on the North
America contributions towards the use of geosynthetic reinforcement in the base-course layer of flexible pavements (paved roads).
2 BACKGROUND
To highlight the significance regarding the use of
geosynthetics in flexible pavements, this section
provides an overview of flexible pavement design
and of the functions of geosynthetics in pavement
applications.

2.1.1 Critical stresses

Flexible pavements allow redistribution of traffic
loads from the contact surface to the underlying layers. As the pavement flexes under the load, stresses
are redistributed over a greater area than that of the
tire-footprint. Figure 2 illustrates the stress redistribution under the wheel load. Design of flexible
pavement pays particular attention to two critical locations within the pavement structure: (1) the hori-

2.1 Flexible pavements

Flexible pavements can be defined as layered systems that include materials on top (where the contact
stresses are high) that have improved qualities than
those towards the bottom (where the contact stresses
379

weaknesses. Cycles of wetting and drying (or freezing and thawing) cause base course material to
breakdown, generating fines in the subgrade and
leading to crack development. Construction practices also affect pavement distress conditions. For example, the use of aggregates with excessive fines
and inadequate inspection may lead to rapid pavement deterioration. Finally, pavement distress is also
a function of maintenance or, more correctly, lack of
maintenance (Yoder and Witczak 1975). For example, sealing cracks and joints at proper intervals and
maintaining the shoulders help improve pavement
performance. Ultimately, a pavement's intended longevity represents a calculated decision on the part of
the engineer who has to balance increased initial
construction costs against increased maintenance
costs during the design process.

zontal tensile strain at the bottom of the asphalt

layer, which should be minimized in order to prevent
fatigue cracking, and (2) the vertical stress on the
top of subgrade, which should be minimized in order
to reduce permanent deformations. The allowable
vertical stress on a given subgrade depends on the
shear strength of the subgrade. The granular base in
flexible pavements should be thick enough so that
the vertical compressive subgrade stress is reduced
to some limit value below the allowable distress level (Yoder and Witczak 1975).

2.2 Reinforcement function

Typical functions of geosynthetics used in the construction of roadways include reinforcement, separation, filtration, lateral drainage and sealing (Koerner
2005). Geosynthetics used for separation minimize
intrusion of subgrade soil into the aggregate base or
sub-base. The potential for the mixing of soil layers
occurs when the base course is compacted over the
subgrade during construction and also during operation of traffic. Additionally, a geosynthetic can perform a filtration function by restricting the movement of soil particles while allowing water to move
from the subgrade soil to the coarser adjacent base.
In addition, the in-plane drainage function of a geosynthetic can provide lateral movement of water
within the plane of the geosynthetic. Finally, geosynthetics can be used to mitigate the propagation of
cracks by sealing the asphalt layer when used in the
overlay of the pavement.

Figure 1: Cross-section of flexible pavement system

(Muench 2006)

Figure 2: Stress distributions with depth in a flexible

pavement (a) high stress area directly under wheel load;
(b) reduced load at subgrade level

This paper focuses on the reinforcement function of

geosynthetics in flexible pavements. Reinforcement
is the synergistic improvement of the pavement
created by the introduction of a geosynthetic into a
pavement layer. While the function of reinforcement
has often been accomplished using geogrids, geotextiles have also been used as reinforcement inclusions
in transportation applications (Bueno et al. 2005,
Benjamin et al. 2007). The stresses over the subgrade are higher in unreinforced flexible pavements
than in geosynthetic-reinforced pavement, as shown
schematically in Figure 3. The geosynthetic reinforcement has generally been placed at the interface
between the base and sub-base layers or the interface
between the sub-base and subgrade layers or within
the base course layer of the flexible pavement.

2.1.2 Type of failure modes

During its lifetime, a flexible pavement can experience two different types of failure modes: structural and functional. Structural failure leads to the collapse of the pavement, thereby making it incapable
of sustaining the surface loads. Functional failure, on
the other hand, renders the pavement incapable of
carrying out its intended function, causing discomfort to passengers. Structural failure requires a complete rebuilding of the pavement whereas functional
failure can be remediated by maintenance.
Pavement distress may occur due to either traffic or
environmental loads. Traffic loads result from the
repetition of wheel loads, which can cause either
structural or functional failure. Environmental loads
are induced by climatic conditions, such as variations in temperature or moisture in the subgrade,
which can cause surface irregularities and structural

The improved performance of the pavement due to

geosynthetic reinforcement has been attributed to
three main mechanisms, as follows: (1) lateral re380

straint, (2) increased bearing capacity, and (3) the

tensioned membrane effect (Giroud and Noiray
1981, Giroud et al. 1984, Perkins and Ismeik 1997,
Holtz et al. 1998). These three mechanisms are illustrated in Figure 4. The rationale behind these three
mechanisms was originally based on observations
and analyses done for unpaved roads. The relevance
of these mechanisms for flexible pavements is discussed next.

restraint develops through interfacial friction between the geosynthetic and the aggregate, thus the
mechanism is one of a shear-resisting interface (Perkins 1999). When an aggregate layer is subjected to
traffic loading, the aggregate tends to move laterally
unless it is restrained by the subgrade or geosynthetic reinforcement. Interaction between the base aggregate and the geosynthetic allows transfer of the
shearing load from the base layer to a tensile load in
the geosynthetic. The tensile stiffness of the geosynthetic limits the lateral strains in the base layer. Furthermore, a geosynthetic layer confines the base
course layer thereby increasing its mean stress and
leading to an increase in shear strength. Both frictional and interlocking characteristics at the interface
between the soil and the geosynthetic contribute to
this mechanism. For a geogrid, this implies that the
geogrid apertures and base soil particles must be
properly sized. A geotextile with good frictional capabilities can also provide tensile resistance to lateral aggregate movement.

Figure 3: Relative load magnitudes at subgrade layer level for (a) unreinforced flexible pavement and (b) geosynthetic-reinforced flexible pavement

2.2.2 Increased bearing capacity

As illustrated in Figure 4b, the increased bearing capacity mechanism leads to soil reinforcement when
the presence of a geosynthetic imposes the development of an alternate failure surface. This new alternate plane provides a higher bearing capacity. The
geosynthetic reinforcement can decrease the shear
stresses transferred to the subgrade and provide vertical confinement outside the loaded area. The bearing failure mode of the subgrade is expected to
change from punching failure without reinforcement
to general failure with reinforcement.
2.2.3 Tensioned membrane effect
The geosynthetic can also be assumed to act as a
tensioned membrane, which supports the wheel
loads (Figure 4c). In this case, the reinforcement
provides a vertical reaction component to the applied
wheel load. This tensioned membrane effect is induced by vertical deformations, leading to a concave
shape in the geosynthetic. The tension developed in
the geosynthetic contributes to support the wheel
load and reduces the vertical stress on the subgrade.
However, significant rutting depths are necessary to
realize this effect. Higher deformations are required
to mobilize the tension of the membrane for decreasing stiffness of the geosynthetic. In order for this
type of reinforcement mechanism to be significant,
there is a consensus that the subgrade CBR should
be below 3 (Barksdale et al. 1989).

Figure 4: Reinforcement mechanisms induced by geosynthetics (Holtz et al. 1998): (a) Lateral restraint; (b) Increased bearing capacity; and (c) Membrane support

2.2.4 Relevance of the Various Mechanisms

The aforementioned mechanisms require different
magnitudes of deformation in the pavement system
to be mobilized. Since the early studies on geosynthetic reinforcement of base course layers focused
on unpaved roads, significant rutting depths (in

2.2.1 Lateral Restraint

The primary mechanism associated with the reinforcement function for flexible pavements (Figure
4a) is lateral restraint or confinement (Bender and
Barenberg 1978). The name is misleading as lateral
381

tations of these approaches for designing geosynthetic-reinforced flexible pavements are discussed
next.

excess of 25 mm) may have been tolerable. The increased bearing capacity and tensioned membrane
support mechanisms have been considered for paved
roads. However, the deformation needed to mobilize
these mechanisms generally exceeds the serviceability requirements of flexible pavements. Thus, for
the case of flexible pavements, lateral restraint is
considered to contribute the most for the improved
performance of geosynthetic-reinforced pavements.

3.1 AASHTO Method

The American Association of State Highway and
Transportation Officials (AASHTO) guide for design of pavement structures is one of the most widely used methods for flexible pavement design in
North America (AASHTO 1993). The AASHTO
method uses empirical equations developed from the
AASHO road tests, which were conducted in the late
1950s. The method considers the pavement as a multi-layer elastic system with an overall structural
number (SN) that reflects the total pavement thickness and its resiliency to repeated traffic loading.
The required SN for a project is selected such that
the pavement will support anticipated traffic loads
and experience a loss in serviceability no greater
than established by project requirements. The SN is
determined using a nomograph that solves the following equation:

3 NORTH AMERICAN DESIGN

METHODOLOGIES FOR GEOSYNTHETICREINFORCED FLEXIBLE PAVEMENTS
The design philosophy of flexible pavement systems
was initiated by the Romans, evolving into the current design approaches. As mentioned, the design
approach involves providing a protective layer over
the subgrade that improves the serviceability under
traffic and environmental loads. Figure 5 shows the
evolution of road design methods in the US since the
1930s.

log
log W 18 = Z R S O + 9.36 log( SN + 1) 0.2 +

Level of
sophistication

Experience
Subgrade CBR
Experience

0.4 +

SN
PSI
ESAL
ESG

EAC,BC,SG
Pratio
ESAL
Transfer function
(X,Y)critical

EAC,BC,SG
Pratio
Transfer function
(X,Y)critical
Non linear properties

AASHTO Method

Use what works

1960

1970

2000

2.7

+ 2.32 log M R 8.07

1094
( SN + 1)

5.19

(1)

where W18 is the anticipated cumulative 18-kip

Equivalent Single-Axle Loads (ESALs) over the design life of the pavements, ZR is the standard normal
deviate for reliability level, SO is the overall standard
deviation, PSI is the allowable loss in serviceability,
and MR is the resilient modulus (stiffness) of the underlying subgrade. Once the required overall SN has
been determined, the individual layers can be designed according through a series of iterations using
the following equation:

MEPDG

Linear M-E Method

CBR Cover Design Method

1930

PSI

2010

Figure 5: Evolution of pavement design methods (Reck

2009)

The Cover Based Design Method was developed after the great depression in the 1930s. It required a
single input in terms of the California Bearing Ratio
(CBR), but it also required use of significant engineering judgment. Subsequently, and after completion of the American Association of State Highway
Officials (AASHO) Road Test in the 1960s, a series
of design methods were proposed. They were more
sophisticated than the Cover Based Method, requiring a greater number of design parameters as input.
For example, in the 1970s, the linear mechanisticempirical (M-E) design method was proposed by researchers from South Africa. Since the early 1990s,
the focus in the US has shifted to M-E design methods that incorporate features from purely empirical
methods to sophisticated non-linear finite element
methods. Attempts have been made to incorporate
the use of geosynthetic reinforcements into AASHO
and M-E design methods. The advantages and limi-

SN = (a d ) hma + (a d m) base + (a d m) subbase (2)

where a is the coefficient of relative strength, d is
the thickness in inches of each layer, and m is the
modifier accounting for moisture characteristics of
the pavement.
The purposes of using geosynthetics as reinforcement in flexible pavements have been: (1) to extend
a pavements life-span, or (2) to enable the construction of a pavement with a reduced quantity of base
course material without sacrificing pavement performance. Early design approaches for reinforced
flexible pavements focused at modifying Equations
1 and 2 to reflect the benefit achieved by the addition of geosynthetics. These improvements to the
pavement system provided by geosynthetic reinforcement have been measured in terms of the Traf382

fic Benefit Ratio (TBR) and the Base Course Reduction (BCR).

ber corresponding to the equivalent W18 for the unreinforced pavement.

The TBR is defined as the ratio between the number

of load cycles on a reinforced section (NR) to reach a
defined failure state (a given rutting depth) and the
number of load cycles on an unreinforced section
(NU) with the same geometry and material constituents that reaches the same defined failure state (Berg
et al. 2000). Specifically, the TBR can be defined as:
TBR =

NR
NU

(3)

Use of the TBR in pavement design leads to an extended pavement life defined by:
W18 (reinforced) = TBR * W18 (unreinforced)

Figure 6: Typical TBR values for an unreinforced and

reinforced pavement to reach a given rutting depth
(Shukla 2002)

(4)

The BCR has been determined from laboratory and

field tests. Anderson and Killeavy (1989) constructed test sections with different base course
thicknesses. The study showed that geotextilereinforced section with a 350 mm thick base layer
performed similarly to an unreinforced section with
a 450 mm thick base layer. Miura et al. (1990) reported the construction of field reinforced sections
that contained a base course that was 50 mm thinner
than that of unreinforced sections. The reinforced
sections were observed to perform better than the
control sections for all rutting depths. Also, at a site
with a subgrade of CBR 8, Webster (1993) showed
that a section containing a geogrid with a 150 mmthick base showed a performance equivalent to that
of an unreinforced section with a 250 mm-thick
base. Thus, BCRs ranging from 20% to 40% have
been reported in the literature, with greater percentage reduction for stronger subgrade materials.

The TBR is sometimes referred to as the traffic improvement factor (TIF), which is commonly used to
relate the long-term performance of reinforced and
unreinforced pavements. As shown in Figure 6, the
TBR can also be used to calculate the number of
traffic passes that a reinforced pavement can withstand as compared to an unreinforced pavement for a
given rutting depth. For most geotextiles, the TBR
value ranges from 1.5 and 10, and for geogrids between 1.5 to 70 (Shukla 2002).
The BCR is defined as the percent reduction in the
base-course thickness due to an addition of geosynthetic reinforcement (TR) in relation to the thickness
of the flexible pavement with the same materials but
without reinforcement (TU), to reach the defined
failure state. The BCR is defined as follows:
BCR =

TR
TU

The AASHTO design method is empirical in nature

and does not directly consider the mechanics of the
pavement structure, climatic effects, or changes in
traffic loads and material properties over the designlife of the pavement. Extension of this design methodology to geosynthetic-reinforced pavements has
been limited to the case of specific products, materials, geometries, failure criteria and loads used in
test sections to quantify their values. Thus, this approach lacks desirable generality as experience cannot be easily transferred from one site to another.

(5)

The BCR is sometimes referred to as the layer coefficient ratio (LCR). A modifier has been applied to
the SN of the pavement, as follows:
SN = (a d )

hma

+ BCR.(a d m)

base

+ ( a d m)

subbase

(6)

When designing a pavement using the BCR, the reduced depth of the base course can be estimated as
follows:

base, ( R)

SN (a d )
(a d m)
u
hma
subbase
BCR.(a m)
base

This method has been unable to provide a consistent

groundwork for performance comparison among
various geosynthetics. In addition, it has been difficult to incorporate the BCR and TBR ratios into
the design where the objective of the reinforcement
is to provide both an increased pavement life and a
reduced base course thickness. Although research
conducted to date has supported some of the proce-

(7)

where dbase,(R) is the reduced base course thickness

due to reinforcement and SNu is the structural num-

383

where d is the cyclic deviator stress (or cyclic principal stress difference) and r is the recoverable
(elastic) strain. Thus, both MR and the Youngs
Modulus (E) represent the strain response of the material to applied stresses. However, they are not considered the same due to differences in the rate of
load application, as shown in Figure 8. The value of
E refers to the initial deformation (with some permanent component) of the material, whereas MR refers to the elastic deformation of the material after
cyclic loading.

dures, long-term performance information of

projects designed using this method is not available
in order to establish confidence limits.
3.2 NCHRP Mechanistic-Empirical Method (2004)
The National Cooperative Highway Research Program (NCHRP) has recently developed a guide for
M-E design of new and rehabilitated pavement
structures (NCHRP 2004). The method uses mechanistic principles and detailed input data to minimize
design reliance on empirical observations and correlations that may be applicable for a specific project.
The M-E method attempts to improve design reliability, reduce life-cycle costs, characterize better the
effects of drainage and seasonal moisture variations,
and prevent premature failures (Olidis and Hein
2004).
While the M-E design method involves two key
components (mechanistic and empirical), they are
both considered interdependent on each other. The
calculation models require input parameters regarding pavement layers, traffic conditions, climatic
conditions and materials. The generated output is
then compared against parameters used as hypothesis for the original design. If the comparison fails,
the design is then modified using an iterative process
and re-evaluated. The flowchart of the various components in the M-E design method is shown in Figure 7.

Figure 8: Comparison of Resilient Modulus, MR, and

Modulus of Elasticity, E (Abusaid 2006)

The M-E method uses a hierarchical approach to design, based on the project importance and available
information. Level 1 is the highest confidence level,
typically reserved for research or very high-volume
roads. Level 2 corresponds to moderate confidence
level, intended for routine pavement design. Level 3
is the lowest confidence level, typically reserved for
low-volume roads. Based on the selected design level, material properties are determined using the specific materials to be used in actual construction
(Level 1), or estimated from the correlations using
routine tests (Level 2), or are defined using default
values from the database (Level 3).
The mechanistic properties of pavement materials
are used to estimate stresses and displacements under loading. These estimates are in turn converted
into pavement surface distresses using regression
models of the Long Term Pavement Performance
(LTPP) program database, which contains comprehensive data from field-scale road test sections. Surface distresses are broadly classified into three
groups: fracture, deformation, and degradation.
These surface distresses can be used to evaluate performance, estimate life cycle and anticipate failure
modes of the pavement.

Figure 7: Flowchart for M-E Design (NCHRP 2004)

The main parameters used in M-E method are the

mechanistic properties of each pavement layer, including their Poissons ratio () and resilient modulus (MR). The Poissons ratio (ratio of lateral to
axial strains exhibited in response to axial loading)
typically ranges from 0.15 to 0.5 for pavement materials. The MR is a measure of the material stiffness
after cyclic loading, represented by:

MR =

d
r

Design of pavements using the M-E approach involves measuring the traffic load cycles that correspond to a limited level of surface distress. This approach could be applied to geosynthetic-reinforced
pavements. The M-E design approach is better

(8)

384

However, flexible pavements are designed for

smaller rutting depths, and are subjected to higher
volume traffic loading than unpaved roads. Giroud
and Han (2004) developed a procedure for the design of geosynthetic reinforced unpaved roads,
which considers stress distribution at depth, base
course resilient modulus, and degradation of material stiffness with repeated loading. Unfortunately, the
theoretical basis that exists for the design of geosynthetic-reinforced unpaved roads is still lacking in the
case of geosynthetic-reinforced pavements. Yet,
numerous research studies have been performed to
better understand the behavior of reinforced pavements using field-scale testing, laboratory testing
and numerical simulations.

suited than the AASHTO approach to incorporate

geosynthetic benefits. This is because the M-E approach requires input from the user to define the local materials, thus providing a more consistent basis
for evaluation of geosynthetic properties.
In the mechanistic model, the contribution of a thin
layer such as a geosynthetic has been incorporated
as an equivalent resilient modulus and Poissons ratio. Yet, in the empirical design, calibration of the
equivalent damage model in terms of subgrade rutting has not provided similar results for thin and
thick asphalt geosynthetic-reinforced flexible pavements. Specifically, in thin asphalt pavements the
geosynthetic contribution has been incorporated into
the properties of the base course layer, whereas in
thick asphalt pavements it has been simulated as an
equivalent delay in the onset of fatigue cracking
(when compared to the onset in an unreinforced
pavement section). Consequently, the benefits of
geosynthetics have not been consistently defined using the M-E design.

4 ASSESSMENT OF THE PERFORMANCE OF

GEOSYNTHETIC-REINFORCED FLEXIBLE
PAVEMENTS
Assessment of the performance of pavements has
been conducted using field scale tests, laboratory
tests, and numerical simulations. These three methods not only differ widely, but have also provided
different perspectives on performance data, as illustrated in Figure 9. Ultimately, the quality of pavement performance data depends on the cost and the
method used for its collection (Reck 2009).

The M-E design approach has been deemed more

appropriate method for estimating field behavior of
flexible pavements than a multi-layered elastic analysis because it is more rigorous and adaptable (AlQadi, 2006). However, the practicality of the method
is compromised since a significant amount of information and test data are required to characterize the
pavement and its anticipated performance. Only few
test agencies have been are capable of performing
the complex tests required to determine properties
such as MR, and even when they are, the associated
costs could be unjustifiably high. Finally, as in the
AASHTO method, the M-E approach also relies
heavily on correlations to material properties.
In summary, prediction of the behavior of flexible
pavements is complex, as the overall performance is
controlled by numerous factors, including load magnitude, subgrade strength, layer thickness, interlayer
mixing, material degradation, cracking and rutting,
and seasonal and climactic fluctuations (WDOT
2007, Dougan 2007, Al-Qadi 2006). While beneficial, the use of geosynthetic reinforcement adds
complexity to the system understanding by introducing a new set of variables. These include the reinforcement mechanism, geosynthetic types and stiffness, tensile strength, aperture size and placement
location. Therefore, due to uncertainty in quantifying the mechanisms of geosynthetic-reinforcement,
neither the AASHTO (1993) nor the NCHRP (2004)
approaches incorporate specific geosynthetic properties fully in design of pavements.

Figure 9: Interrelationship between different facets of

pavement design (Hugo et al. 1991)

Full-scale tests include field studies and accelerated

pavement tests that simulate actual pavement behavior. In both cases, the cost of data gathering is high
and, consequently, only the number of tests typically
conducted has been limited. Thus, the testing scope
in reported research programs has been often expanded by small-scale laboratory studies or numerical simulations. Laboratory tests are cheaper than
field tests and can be performed under controlled
conditions. However, it has been difficult to replicate the actual behavior of the pavement system under laboratory tests. Finally, numerical studies have
been conducted to simulate both field and laboratory

In addition, the design of geosynthetic-reinforced

pavements has relied significantly on empirical data
used to calibrate the response of unpaved roads.
385

tests. Numerical simulations are particularly valuable for parametric evaluations.

Deflection measurements have also been made using

non-destructive testing (NDT) devices in order to
evaluate the pavement structural capacity and to calculate the moduli of various pavement components.
The device most widely used to measure pavement
deflections is the Falling Weight Deflectometer
(FWD) (Figure 11a). This approach involves applying a series of impulses on the pavement using a
trailer-mounted device that is driven to the desired
test locations. A loading plate is hydraulically lowered to the pavement surface, after which an impulse is applied to the pavement by dropping a
weight from a known height onto the loading plate.
The magnitude of the load is measured using a load
cell while deflections are measured using seven velocity transducers. An equipment known as a Rolling Dynamic Deflectometer (RDD) (Figure 11b),
has been recently developed for assessing the conditions of pavements and determining pavement deflection profiles continuously (Bay and Stokoe 1998).
Unlike the FWD, the RDD performs continuous rather than discrete deflection measurements. The ability to perform continuous measurements makes
RDD testing an effective approach for expeditious
characterization of large pavement sections. The
equipment applies sinusoidal forces to the pavement
through specially designed rollers. The resulting deflections are measured by rolling sensors designed to
minimize the noise caused by rough pavement surfaces.

4.1 Field tests

Full-scale field tests have been performed on both
public roadways and in-service roads. As previously
discussed, M-E design processes have been recently
developed that require data for calibration and validation purposes (Watts and Blackman 2009). The
monitoring of in-service roads is a time consuming
process. Consequently, useful data has also been
generated using accelerated pavement testing (APT).
APT facilities consist of test tracks located either indoor or outdoor (Figure 10). They involve the use of
automated, one or two axle, single wheel loads that
repeatedly runs over the test track surface. APT may
provide a good simulation of the performance of inservice pavements and can be particularly useful to
provide rapid indication of pavement performance
under severe conditions.

(a)
(b)
Figure 10: APT test facilities (a) ATLAS at the Illinois
Center for Transportation, USA; (b) pavement fatigue carousel at LCPC, France

Several approaches have been implemented to evaluate and compare pavement performance in fieldscale test sections. In flexible pavements, the two
most commonly quantified variables are surface deflection and cracking (including longitudinal, transverse and fatigue). Surface deflection is the most
common performance criterion for both reinforced
and unreinforced pavements. Distress has been evaluated using: (1) measurement of existing surface
deflections in terms of rutting depth, and (2) measurement of surface deflections in response to an applied load to determine its structural capacity.
Rutting occurs because of the development of permanent deformations in any of the pavement layers
or in the subgrade. Rutting is generally measured in
square meters of surface area for a given severity
level, as defined from data collected with a dipstick
profiler every 15 m intervals. Measurements of rutting depth are comparatively easy to obtain, as they
are taken at the pavement surface, and provide a
simple method of comparing pavement performance
among multiple test sections.

Figure 11: Non-destructive testing methods used in

pavement evaluation (a) Falling Weight Deflectometer
and; (b) Rolling Dynamic Deflectometer

Field tests on full-scale road sections have been

conducted to evaluate the effect of geosynthetic reinforcement in flexible pavement systems. Perkins and
Ismeik (1997) compared the results from nine sections, among which four were constructed on indoor
test tracks, three on outdoor test tracks, one on a
public roadway and one in a field truck-staging area.
386

tions (without geogrid reinforcement) had a 0.20 mthick base course layer. FWD testing showed a comparatively higher pavement modulus for the geogridreinforced section with a 0.20 m-thick base while
lower modulus value were obtained for the geogridreinforced section with a 0.127 m-thick base. Yet,
field visual assessment showed cracking in the control section while the two geogrid-reinforced sections performed well. While the geogrid-reinforced
sections outperform the unreinforced section, the results of FWD testing showed a different trend. This
study illustrated the inadequacy of the currently
available evaluation techniques involving nondestructive testing for the purpose of quantifying the
benefits of geosynthetic reinforcements.

The indoor test tracks used a single moving wheel to

load the test sections (Brown et al. 1982, Barksdale
et al. 1989, Collin et al. 1996, Moghaddas-Nejad
and Small 1996). The outdoor test tracks involved a
single moving wheel (Barker 1987, Webster 1993),
and a two-axle, dual wheel truck to load the pavement (Halliday and Potter 1984).
Additional studies have been recently reported on
geosynthetic-reinforced test sections using APT
equipment (Cancelli and Montanelli 1999, Perkins
2002, Perkins and Cortez 2005, Al-Qadi et al. 2008,
Reck et al. 2009). Assessment of these test sections
indicated that rutting depth continued to be the most
common method to evaluate pavement distress. A
total of nine field test sections and four APT sections
were reported involving measurements from profilometer readings at the end of design loading cycles.
However, FWD tests were conducted only at four
field sections and at one APT section.

The lessons learned from these field case studies,

provided the basis for a field monitoring program to
evaluate the performance of geosynthetic-reinforced
pavements constructed over expansive clays. This
involved the rehabilitation of a low-volume road in
Texas by use of geosynthetic reinforcements. A
comparative evaluation with 32 test sections was
conducted. This included 8 different reinforcement
schemes (3 reinforcement products and an unreinforced control section, as well as lime stabilized sections). Also, and in order to account for variability
due to environmental, construction and subgradetype, a total of 4 repeats were constructed for each
one of the 8 schemes. Therefore, a total of 32 test
sections (4 reinforcement types x 2 stabilization approaches x 4 repeats) were constructed (Figure 12).

Zornberg and Gupta (2009) reported three case studies conducted in Texas, USA, for geosyntheticreinforced pavements on which FWD testing was
conducted on in-service roads. One of the cases involved a forensic investigation conducted in a newly
constructed pavement. Longitudinal cracks were observed in a geogrid-reinforced pavement before it
was open to traffic. However, the investigation revealed that the contractor had laid rolls of geogrid
leaving a portion of the pavement unreinforced.
Cracks only appeared in unreinforced locations
within the pavement. Accordingly, the difference in
response within and beyond reinforced portions of
the pavement illustrated that use of geogrid can prevent pavement cracking.
The second case study reported the field performance of geogrid-reinforced pavements built over
highly plastic subgrade soils. The pavement sections
had been reinforced using two different types of
geogrids that met project specifications. Although a
section reinforced with one type of geogrid was
found to be performing well, the other section reinforced with second type of geogrid showed longitudinal cracking. The reviews of the material properties lead to the preliminary conclusion that poor
performance in the second section was due to inadequate junction efficiency. Further inspection indicated a higher tensile modulus of the geogrid used in
the better performing section. This study highlighted
the need for better material characterization and the
possible inadequacy of commonly used specifications for geosynthetic-reinforced pavements.

Figure 12: Schematic layout of test sections at FM 2 site

(Gupta et.al 2008)

Due to unique characteristics of this field study, the

reinforced pavement was considered experimental
and an extensive post-construction performance
monitoring program was implemented. This included the installation of moisture sensors to characterize the patterns of moisture migration under the

The third case involved three pavement sections.

The two geogrid-reinforced sections (Sections 1 and
2) had base course thicknesses of 0.20 m and 0.127
m, respectively. On the other hand, a control sec387

pavement. A total of eight horizontal moisture and

vertical moisture sensor profiles, each containing an
array of four sensors was installed below the pavement. Field monitoring involving visual inspection,
surveying and FWD was conducted before reconstruction and immediately after reconstruction of the
road. The final construction of the reinforced pavement was completed in January 2006 and performance evaluation of the newly reconstructed road
has been conducted on a regular basis since then.
The results obtained from the field study are providing good understanding of the underlying mechanisms governing the performance of the geosyntheticreinforced pavements. Also, the collected data is
useful to quantify the mechanisms of longitudinal
cracking and effectiveness of the geosynthetic reinforcements in mitigating such distresses.

formance of the geosynthetic-reinforced pavements.

Tests in this category include the wide-width tensile
test, biaxial loading test, junction efficiency test, and
torsional rigidity test. While the wide-width tensile
test can be conducted using any type of geosynthetics (geogrid, geotextile), the other three tests are
specific for the characterization of geogrids.

Overall, the results from field studies reported in the

literature have indicated that the geosyntheticreinforced test sections led to less rutting depth than
the unreinforced sections. The improved performance has been attributed to the ability of the geosynthetics to control lateral spreading of the base
layer.

Figure 13: Mechanisms due to soil-geosynthetic interaction in geosynthetic- reinforced pavement that have been
tried to be represented in laboratory tests (Perkins 1999)

The tensile strength of geosynthetic materials has often been deemed as the most important property for
projects involving reinforcement applications. While
tensile strength may not be particularly relevant for
the case of pavement design, tensile strength has often been incorporated into pavement design and specifications. The current state of practice for measuring the tensile properties of a geosynthetic involves
placing the material within a set of clamps, positioning this assembly in a load frame, and tensioning the
geosynthetic until failure occurs. The test is generally performed at a constant strain rate. Currently, two
ASTM standards are available for tensile tests. The
grab tensile test (D4632) is used for manufacturing
quality control, as it involves a narrow geosynthetic
specimen. Instead, the wide-width tensile test
(D4595) has been used in design applications. The
load frame for a wide-width tensile test conducted
using roller grips is shown in Figure 14. The tensile
test provides the tensile stiffness at different strain
values (1%, 2%, and 5%), as well as the ultimate
tensile strength. Methods used for unpaved road design have included the tensile stiffness at 5% in
product specifications. Based on full scale model
studies for the paved roads, Berg et al. (2000) reported accumulated in-service tensile strain of 2% in
geosynthetics and thus recommended the tensile
stiffness at this strain level for design. However, the
actual strain level representative of field conditions
is certainly smaller for the case of pavement applications.

4.2 Laboratory Tests

A number of laboratory tests have been proposed to
quantify the mechanisms governing the performance
of geosynthetic-reinforced flexible pavements. The
primary objective of laboratory tests has been to
quantify the soil-geosynthetic interaction mechanisms in flexible pavement systems either by measuring the geosynthetic index properties or by replicating the field conditions. An important field condition
to be replicated is the effect of interface shear provided by geotextiles and interlocking provided by
geogrids when used under or within the base course
layer of pavements (Figure 13). Depending on the
adopted approach, the tests reported in the literature
can be grouped into two main categories: unconfined
and confined tests. In unconfined tests, geosynthetic
properties are measured in-air, while in confined
tests they are measured within confinement of soil.
The advantages and limitations of the various tests
developed in North America in each of these two
categories are discussed next.
4.2.1 Unconfined Tests
As mentioned, unconfined tests are conducted using
geosynthetic specimens in isolation. Advantages of
these tests include expedience, simplicity, and cost
effectiveness. They can be run in short periods of
time using conventional devices, which facilitates
the assessment of repeatability of test results. However, correlations are required between the index
property obtained from these tests and the field per-

388

of the reinforcement during the pavement construction and subsequent traffic load. However, the geogrid ability to transfer stress under low strains is a
consideration probably of more relevance for the
case of flexible pavements. However, junction stiffness requirements for pavement projects have not
been properly defined. Also, since this test was originally developed for geogrids with integral junctions,
it does not incorporate newer geogrids with entangled fibers or those with heat bonded or laser welded
junctions.

Figure 14: Wide-width tensile test conducted with roller

grips at the University of Texas at Austin

Bray and Merry (1999) investigated the stress and

strain conditions in wide-width tensile tests. They
concluded that strains vary across the specimen from
a plane-strain, biaxial condition near the grips, to a
uniaxial condition near the center of the specimen.
Thus, there may be a misconception that the test
measures geosynthetic behavior under the 1-D condition that is representative of field applications. It
should be noted that most geogrids tested using uniaxial methods suffer distortions, non-uniform
stresses (particularly at the junctions), premature
specimen rupture and problems with clamping
(McGown et al. 2005). Kupec and McGown (2004)
suggested a biaxial test method, which focused primarily on geogrids and allowed characterization of
the combined strength of tensile ribs and junctions in
a single test. The test specimen involved 5 ribs in
each direction (i.e. 25 junctions within the central
section of the specimen), as shown in Figure 15.
Loading was applied to the specimens under isotropic deformation conditions. Test results indicated that
the biaxial load-strain time behavior differs significantly from the uniaxial behavior. An increased
stiffness was obtained, which was attributed to the
behavior of junctions under tensile stresses in the
principal direction, to the Poissons ratio effect and
to re-orientation of the fibers in the junction areas.

Figure 15: Geogrid specimens for biaxial testing

(McGown et al. 2005)

A torsional rigidity test was developed by Kinney

and Yuan (1995) to measure the in-plane rotational
stiffness of the geogrids (Figure 16). The test aimed
at quantifying the performance of geogrid-reinforced
paved road tests constructed by the US Army Corps
of Engineers at the Waterways Experiment Station.
While the test focuses on the interlocking capacity
of the geogrid, a relationship between geogrid torsional rigidity and the performance of geogrid reinforced road sections could not be established. The
test provides a higher torsional rigidity for stiff geogrids than for flexible geogrids. However, a study
conducted by the Texas Research Institute (TRI
2001) reports a lack of correlation between torsional
rigidity and the confinement performance of the
geogrids.

To address perceived deficiencies of uniaxial tensile

test, a complementary uniaxial test, known as the
junction strength test, was developed. It is conducted as per the procedure recommended in GRIGG2 specifications and involves gripping the cross
member of a geogrid rib on both sides of the junction with a clamping device. Load is then applied
until the junction breaks. The force required to fail
the junction is defined as the junction strength of the
geogrid. Junction strength provides quantification of
the contribution to stability that may lead to rupture

Figure 16: Torsional rigidity tests (Kinney and Yuan

1995)
389

The geosynthetic behavior observed in the laboratory from unconfined tests has to be correlated with
the performance in field applications, which have
different loading and boundary conditions. In general, it has been difficult to replicate field conditions
using the aforementioned unconfined tests. Consequently, unconfined tests should only be considered
as index parameters rather than actual design properties for geosynthetic-reinforced flexible pavements.

able, thus restricting the application of this test to research studies. In addition, the cyclic plate loading
test was considered to have important drawbacks associated with the testing procedures, time demands,
and appropriate simulation of rolling wheel loads
(Han et al. 2008).

4.2.2 Confined tests

Geosynthetics used for base reinforcement are under
the confinement of soil and subjected to dynamic
loading (traffic). These conditions cannot be simulated by monotonic unconfined tests. Geosyntheticsoil confinement depends not only on the macrostructure and properties of geosynthetics but also on
the properties of soil and, most importantly, on the
interaction between geosynthetics and soil particles
(Han et al. 2008). The interaction between soil and
geosynthetics under confinement, specifically the
confined stress-strain properties of the geosynthetics, has been focus of previous research. A Federal
Highway Administration (FHWA) sponsored study
focusing on existing confined tensile tests for geosynthetics concluded that the unconfined response of
geosynthetics is overly conservative and that confinement significantly improves their mechanical response (Elias et al. 1998). Recently, a number of
confined tests have been proposed, out of which six
tests have focused on characterizing the behavior of
geosynthetics used to reinforce flexible pavements.
These tests include the cyclic plate load test, cyclic
triaxial test, cyclic pullout test, bending stiffness
test, modified pavement analyzer test, and the pullout stiffness test.

Figure 17: Cyclic plate load test (Perkins 1999)

The cyclic triaxial test (Figure 18) has been used to

measure the ability of soils to develop shear stresses
induced by cyclic loading (ASTM D5311 2004).
The resilient modulus, Mr, of the soil aggregates
computed using this test has been specifically used
as input in the M-E design (NCHRP 2000). This test
was modified by Perkins et al. (2004) to quantify the
change in resilient modulus and permanent deformation behavior due to the addition of geosynthetics to
the aggregate layer of pavements. The results from
cyclic triaxial tests indicate that the use of reinforcements does not affect the resilient modulus of
the aggregates, although it reduces significantly the
pavement permanent deformations. Also, reinforcements were observed to increase the stiffness of the
aggregate in zones above and below the geosynthetic
(the thickness of these zones was observed to be
similar to the specimen radius, or 150 mm). However, relatively poor repeatability of the test results was
achieved, making it difficult to quantify differences
in the performance expected from the different geogrid products. Also, an appreciable permanent deformation was not observed until reaching mobilized
friction values of approximately 30.

The cyclic plate load test has generally involved

large scale laboratory experiments on reinforced and
unreinforced pavement sections (Al-Qadi et al.
1994, Cancelli et al. 1996, Haas et al. 1988, Miura et
al. 1990, Perkins 1999). The test setup designed by
Perkins (1999) consisted of a 2 m wide and 1.5 m
high reinforced concrete tank (Figure 17). The model pavement section was constructed with a geosynthetic at the interface of the base course and subgrade layers. The load was applied by a pneumatic
actuator in the form of a trapezoidal wave pulse,
which generated a maximum surface pressure of 550
kPa on the pavement. The force and displacement
responses were measured using a load cell and eight
surface LVDTs. TBRs ranging from 1 to 70 and
BCRs ranging from 20% to 50% were obtained using cyclic plate load tests in sections involving geotextile and geogrid reinforcements (Hsieh and Mao
2005). These tests were reported to have successfully demonstrated the effect of soil confinement and
dynamic loading. However, facilities in which cyclic
plate loading can be conducted are not readily avail-

Cyclic pullout tests were conducted by Cuelho and

Perkins (2005) by modifying the standard pullout
test (ASTM D6706) to resemble the loading protocol
used in a cyclic triaxial test (Figure 19). Cyclic shear
load cycles (ranging from 100 to 300) were applied
at different confinement level beginning with a seating load of 51 kPa until pullout failure was reached.
Based on the test results, a parameter known as geosynthetic-soil resilient interface shear stiffness (Gi)
was defined to describe the reinforcement-aggregate
390

interaction under cyclic loads. This parameter is defined as:

(9)
Gi = i
i

(a)

where i is the relative displacement between the

aggregate and reinforcement and i is the shear stress
applied to the interface. The units of Gi are kN/m3.
The parameter, Gi was assumed to closely resemble
Mr as it depends on both the shear load and confinement. Therefore, the three parameter log-log equations for Mr reported in NCHRP (2001) was modified and used to calibrate Gi for a given soilgeosynthetic interface, as follows:

Gi = k1 .Pa . i
pa

k2

. i + 1
pa

(b)

k3

(10)
(c)

where i is the normal stress on the interface, pa is

the normalized atmospheric pressure, Pa is the atmospheric pressure per unit length and k1, k2 and k3
are dimensionless material. The purpose of this test
was to provide a property useful to characterize the
interface shear moduli in finite element simulations
conducted to calibrate the M-E approach. However,
pullout test results conducted on six geosynthetics
indicated that correlations between the predicted and
measured values were erratic. The results were sensitive to small changes in displacement magnitudes.
Also, the shear load was cycled while normal load
was kept constant, which may not be representative
of field conditions. Accordingly, additional research
was deemed necessary to improve the testing
equipment and to establish testing protocols.

Figure 19: Cyclic pullout test (a) Plan view (b) Side view
(c) Loading protocol (Cuelho and Perkins 2005)

The bending stiffness test was developed by Sprague

et al. (2004) as a small-scale index test procedure
aimed at predicting the behavior of geosynthetics
used for reinforcement of pavements. The test apparatus is a modified version of the multi-axial tension
test for geomembranes (ASTM D 5617) (Figures
20a and 20b). Details regarding the testing procedures are provided by Abusaid (2006) and Finnefrock (2008). Testing involves applying a uniform
vacuum pressure of 5 kPa on the soil-geosynthetic
interface to induce a confinement representative in a
pavement structures. Subsequently, cycles of air
pressure cycles are applied to the soil-geosynthetic
system and the center-point deflection is measured
by a dial gauge to quantify the response. A property
identified as the bending stiffness (BS) is obtained
from the test results, which is defined as the ratio between the deviator stress d and the recoverable deformation r (Figure 23c). Specifically, BS is defined as the slope of the curve representing the
applied pressure versus the center-point deflection,
as follows:
BS =

Figure 18: Cyclic triaxial test (a) Test equipment (Perkins

et al. 2004); (b) Schematic of test setup (Tutumluer 2004)

d
r

(11)

BS is reported in units consistent with the resilient

modulus, Mr, and thus is not a measure of the
391

strength but rather of the stiffness of the system.

Test results reported by Finnefrock (2008) show
reinforcement benefits of 20% to 25% in terms of BS
ratio, indicating a clear difference between geogridreinforced and unreinforced specimens. However,
the relative performance among different geogrid
products could not be well identified due to scatter
in the test results. Also, a theoretical analysis conducted by Yuan (2005) indicates significant influence of edge shear resistance on the test results. The
tests performed on geotextile-reinforced base course
sections did not indicate benefits over control sections.

reduction ratio (RRR), defined as the ratio between

the rutting in the reinforced base and that of in the
unreinforced base for a given service life (8,000
cycles). Geosynthetics leading to a lower RRR value
are expected to result in better field performance.
Tests were conducted using four geosynthetics
(three geogrids and one geotextile) and two base
course materials. TBR values ranging from 1 to 36
and RRR values ranging from 0.3 to 1.2 were obtained. When a surcharge was applied to the soilgeosynthetic system to simulate confinement in the
field, lower rutting depths were observed as compared to unconfined tests for a similar setup.

(a)

(b)

(c)

Figure 20: Bending stiffness test (a) Actual test apparatus

(b) Schematic of the test (c) Deviator stress and recoverable deformation plots for a typical test (Abusaid 2006)

Figure 21: Modified pavement analyzer test (a) schematic of the test (b) APA testing machine (c) modified box
(d) geosynthetic placed in the middle of the box (e) base
course layer over the geosynthetic (f) test with loaded
wheels (g) rut observed at the end of test (h) rut measurement using dial gauge ( Han et al. 2008)

Han et al. (2008) proposed a test method involving

the use of an asphalt pavement analyzer (APA) to
evaluate the benefits of geosynthetic-reinforcement
in the base course layer of the pavement (Figure
21a). The APA is a multifunctional wheel-loaded
test device used to quantify permanent deformation,
fatigue cracking, and moisture susceptibility of both
hot and cold asphalt mixes (Figure 21b). A conventional box was modified in order to conduct the test
on a geosynthetic-reinforced base course. The
loaded wheel is moved back and forth on the surface
of base course as shown in Figures 21c to 21h. A relationship was established between the measured
rutting depth and the number of passes. The results
could be used to compare the response of geosynthetic-reinforced layers with the unreinforced base
layer. Besides evaluating the TBR for pavement sections, the authors proposed a parameter known as rut

A Pullout Stiffness Test (PST) was recently developed by Gupta (2009) at the University of Texas,
Austin in order to quantify the soil-geosynthetic interaction in reinforced pavements. The equipment
involves a modified large-scale pullout test modified
to capture the stiffness of the soil-geosynthetic interface under small displacements. Research conducted
using the PST has shown that monotonic pullout
tests (Figure 22) aimed at characterizing the soilgeosynthetic interaction under low displacements are
promising. Although these pullout tests did not replicate the cyclic nature of traffic load conditions, it
simulated the interface transfer mechanisms between
392

soil and geosynthetic reinforcements that are expected in the field.

unconfined tests. Accordingly, and based on the current body of literature, unconfined tests are considered inadequate for assessment of the performance
of geosynthetic-reinforced pavements.

Reaction frame

Wooden boards

Air cylinders

Load cell

Roller grips

Connecting wire
Lvdt support
frame

Geosynthetic

Lvdt

Support system
for grips

(a)
Roller grips

Hydraulic Pistons
Reaction frame

LVDT
support frame
Lvdt's

Support system
for grips

Load cell

(b)

Figure 22: Pullout stiffness test to quantify soilgeosynthetic interaction (a) Side view; (b) Plan view
(Gupta 2009)

An analytical model was proposed to predict the

confined load-strain characteristics of soilgeosynthetic systems under small displacements using the results obtained from the PST. This approach
takes into account both the confined stiffness (Jc)
and ability of geosynthetic to mobilize shear or interlock (y), which are two important parameters governing the performance of geosynthetic interfaces.
The two parameters can be combined to define a
unique coefficient of soil-geosynthetic interaction
(KSGI) that characterizes the soil-reinforcement interface. This coefficient is computed as:
K SGI = 4. y .J C

A summary of the confined test methods developed

for the evaluation of geosynthetic-reinforced pavements is presented in Table 1. The tests provide
quantification of the soil-geosynthetic interaction
behavior, although they are comparatively more expensive and time consuming than unconfined tests.
The tests quantify the performance of the soilreinforcement system in the terms of reduced deflections (e.g. TBR, BS, RRR) or increased confinement
modulus (e.g. Mr , Gi ,KSGI). Results from confined
tests are deemed more appropriate as input in design
methods such as the AASHTO and M-E design approaches. The various studies indicated that reinforced systems provided improvement over control
sections without geosynthetics. However, drawbacks
were also identified in several of the proposed confined test approaches. Specifically, these tests require specialized equipment and, at least in several
of the proposed methods, the variability of test results was significant. Overall, confined testing approaches were considered more representative and
appropriate to assess the improvement of geosynthetic reinforcements in pavements than unconfined
testing methods. The main characteristics and relative merits of the various confined tests are summarized in Table 1.
Based on this evaluation, it may be concluded that a
reasonable test method should include the following
features: (a) ability to capture the mechanism of lateral restraint; (b) provide parameter(s) suitable for
M-E design; (c) provide good repeatability of test
results; (d) utilize parameter(s) that distinguish between the performance of different geosynthetics;
(e) be sensitive under low displacements; and (f) be
easy to conduct. The PST approach was developed
keeping these features in mind, and it appears promising for design of geosynthetic-reinforced pavements.

(12)

A comprehensive field monitoring program is under

way (Figure 12) to relate the field performance to
laboratory PST results for a number of geosynthetic
reinforcements. While ongoing field monitoring is
still in progress, good agreement has been obtained
so far between the field performance and the properties defined from PST testing. Thus, a new performance-based test method in the form of a pullout
stiffness test is promising as a performance-based
test to evaluate the soil-geosynthetic confinement.

4.3 Numerical Studies

The design of flexible pavements involves understanding the behavior and the interaction between
various materials (asphalt, base course, subgrade,
geosynthetic reinforcements). Current design methods are empirical in nature. This is partly because
of the inability of available analytical tools to predict
the time-dependent behavior of pavements under actual traffic loads. However, numerical methods can
be used to provide insight into the mechanics of
pavement systems. The most commonly used methods are the finite elements method (FEM) and the
discrete elements method (DEM).

An overall assessment of the various tests developed

so far in North America for geosynthetic-reinforced
pavements indicates that unconfined tests are simple,
economical and expeditious, although they do not
capture the important aspects associated with confinement and the type of soil. Also, unconfined tests
have provided only index measures of the actual
mechanisms, requiring subsequent correlations with
field performance. It should be noted that field studies sometimes led to performance trends that contradicted the trends obtained using properties from
393

Table 1 Features of confined tests

Test Type
References
Loading type
Design property

Cyclic plate Cyclic triaxial

Cyclic pullout test
load test
test
Perkins
Perkins et al. Cuelho and Perkins
(1999)
(2004)
(2005)

Sprague et al.
(2004)

Modified asphalt
pavement analyzer
Han et al.
(2008)

Bending stiffness test

Pullout Stiffness Test

Gupta
(2009)

Cyclic

Cyclic

Cyclic

Cyclic

Moving wheel

Monotonic

TBR

Mr

Gi

BS

RRR

KSGI

Suitable design method

AASHTO

M-E

M-E

AASHTO

AASHTO

M-E

Ease of running test

Difficult

Difficult

Moderate

Moderate

Easy

Moderate

Yes

Yes

No

Yes

Yes

No

Repeatability of test results

No

No

No

Yes

Yes

Ability to distinguish among

various geosynthetics

No

No

No

Yes

Yes

Control section

Finite elements have been used in several studies to

simulate the behavior of geosynthetics used to reinforce flexible pavements. Many of these studies
were performed in combination with laboratory or
field test studies so that comparisons between model
predictions and experimental results could be made.
A summary of important features of these studies is
presented in Table 2.

soil layer above the geogrid. However, additional research is still required to establish a correlation between the results of DEM pullout test simulations
and actual field performance.
(a)

Results from the finite element studies have been

generally reported in the form of surface deformation of the given system under the applied load.
Comparisons were generally made between the
magnitudes of surface deformation for unreinforced
and reinforced pavements. Finite element modeling
of flexible pavements was conducted by Perkins
(2001) using representative sections such as that
shown in Figure 23a. The numerical results indicated a reduction in lateral strain at the bottom of the
base and a reduction in shear at the top of the subgrade due to the presence of the reinforcement (Figure 23b).

(b)

The Discrete Element Method (DEM) has also been

recently used to model soil-geosynthetic interaction,
with particular emphasis on assessment of the interlocking of geogrid with base course material. This
method is expected to capture the interaction between geogrid and soil in terms of load transfer mechanism and deformation behavior. Pullout test results were simulated by Konietzky et al. (2004) to
model interlocking effect of geogrids under static
and cyclic loading. Also, McDowell et al. (2006)
simulated the use of biaxial geogrids to determine
the optimum ratio of geogrid aperture size to soil
particle size (Figure 24).

Unreinforced
Geosynthetic reinforcements
(with different properties)
Rigid reinforcement

Figure 23: Flexible pavement (a) Finite element model

(b) Horizontal stress vs. strain profile for various cases
(Perkins and Edens 2002)

DEM modeling of a low volume road reinforced

with geogrids was conducted by Kwon et al. (2008).
The results indicated that the use of geogrids led to
locked-in stresses during placement, compaction,
and in-service loading, which resulted in a stiffer
394

Table 2 FEM studies for geosynthetic-reinforced flexible pavement design (adapted from Perkins 2001)
References

Type of
analysis

Geosynthetic constitutive
model

Geosynthetic element
type

Interface element
type

Load type

Validation

Burd and Houlsby

(1986)

Plane strain

Isotropic linear-elastic

Membrane

None

Monotonic

None

Barksdale et al.
(1989)

Axi-symmetric

Isotropic linear elastic

Membrane

Linear elastic perfectly

plastic

Monotonic

Field results

Burd and Brocklehurst

(1990)

Plane strain

Isotropic linear- elastic

Membrane

None

Monotonic

None

Miura et al.
(1990)

Axi-symmetric

Isotropic linear elastic

Truss

Linear elastic joint

element

Monotonic

Field results

Dondi
(1994)

Threedimensional

Isotropic linear- elastic

Membrane

Elasto plastic MohrCoulomb

Monotonic

None

Wathugala et al.
(1996)

Axi-symmetric

Isotropic elasto-plastic

Solid Continuum

None

Single cycle

None

Perkins
(2001)

Threedimensional

Anisotropic elasto-plastic

Membrane

Mohr-Coulomb

Multiple
cycles

Lab and test

tracks

Kwon et al.
(2005)

Axi-symmetric

Isotropic linear-elastic

Membrane

Linear-elastic

Monotonic

Test tracks

tude greater than those obtained from laboratory

tests. The use of results from DEM simulations has
been recently suggested as an alternative to represent
better the interaction characteristics of soilgeosynthetic interface. The results obtained using
this approach account for geosynthetic mechanisms
such as the locked-in stresses experienced during the
service life of pavement. The results have indicated
that addition of geosynthetic-reinforcement to the
pavement structure leads to an overall stiffening of
the base course layer.
Mechanistic response models have been proposed by
Perkins and Svan (2004) to account separately for
compaction and traffic loading stages in pavement
design. These models represent rational but complex
ways of incorporating the benefit of geogrid reinforcement and evaluating their performance in basereinforced pavements (Kwon et al. 2008). Studies
are under way in order to improve the use of FEMand DEM-based models for implementation into ME design methods.

Figure 24: Discrete element model (a) pullout test with

cubic clumps (b) pullout test with embedded geogrid (c)
aperture of geogrid (d) detailed view at node (McDowell
et al. 2006)

The use of results from FEM simulations has been

recently proposed in combination with M-E design
of geosynthetic-reinforced pavements. However, the
developed models use geosynthetics as a structural
element directly included within the materials but
without simulating the interface conditions. Also,
due to the significant computational effort required
to simulate realistic traffic loads, simulations have
often been conducted using a single static load cycle.
These simplifications have led to results that may
underestimate the benefit of geosynthetic reinforcement when compared with the field performance
(Perkins and Ismeik 1997). To account for this deficiency, finite element simulations have often used
increased values of geosynthetic stiffness. These
stiffness values have been over an order of magni-

5 CONCLUSIONS
The results of field, laboratory and numerical studies
have demonstrated the benefits of using geosynthetics to improve the performance of pavements. However, selection criteria for geosynthetics to be used
in reinforced pavements are not well established yet.
The purpose of this paper was to summarize information generated so far in North America to quantify the improvement of geosynthetics when used as
reinforcement inclusions in flexible pavement
projects.
395

pears promising to quantify relevant mechanisms in

pavement reinforcement design.

Previous research has led to a reasonably good understanding of the benefits achieved with the use of
geosynthetics in pavement design but, for the most
part, only from the empirical point of view. That is,
while methods have been developed for designing
geosynthetic-reinforced flexible pavements, quantification of the reinforcement mechanisms, identification of properties governing the pavement performance and, ultimately, acceptable design guidelines
are yet unavailable.

Overall, it may be concluded that significant advances have been made in the area of geosynthetic
reinforcement of pavements. While the state of practice is rapidly improving, further research is still
needed to provide a better theoretical basis to the
currently available empirical design approaches.
REFERENCES

Efforts are currently under way in the North America to develop design models consistent with the
AASHTO and M-E approaches. The TBR and BCR
ratios have been used in the AASHTO approach but
are limited because the approaches are specific to
the products and test conditions under which these
ratios have been calibrated. Thus, M-E methods are
considered more generic and, consequently, more
promising as framework to incorporate the use of
geosynthetics in current pavement design. However,
due to the complex nature of flexible pavements, research to identify and quantify the properties governing the performance of reinforced pavements
and its incorporation into M-E design is still under
way.

Abusaid, A.H. 2006. Development of a method to test geogrid

reinforcement of base materials in flexible pavements.
M.S. Thesis, Department of Civil Engineering, The University of Texas at Austin.
Al-Qadi, I.L., Brandon, T.L., Valentine, R.J., Lacina, B.A. and
Smith, T.E. 1994. Laboratory evaluation of geosyntheticreinforced pavement sections. Transportation Research
Record, Volume 1439, pp. 647-662.
Al-Qadi, I.L 2006. Pavement interlayer system mechanisms:
separation reinforcement and reflective cracking control.
Lecture, Chinese Society of Pavement Engineering, Taipei, Taiwan, June 2, 2006.
Al-Qadi, I.L., Dessouky, S.H., Kwon J. and Tutumluer, E.
2008. Geogrids in flexible pavements: validated mechanisms. Journal of the Transportation Research Board, No.
2045, Transportation Research Board of the National
Academies, Washington, D.C., 2008. pp. 102-109
American Association of State Highway and Transportation
Officials 1993. AASHTO Guide for design of pavement
structures. Washington, DC, USA.
Anderson, P. and Killeavy, M. 1989. Geotextiles and Geogrids:
cost effective alternate materials for pavement design and
construction. Proceedings of Geosynthetics 89, IFAI,
Vol. 2, Sand Diego, California, USA, February 1989, pp.
353-360.
ASTM D4595 2001. Standard test method for tensile properties
of geotextiles by the Wide-Width strip method. American
Society for Testing and Materials, West Conshohocken,
PA.
ASTM D4632 2008. Standard test method for grab breaking
load and elongation of Geotextiles. American Society for
Testing and Materials, West Conshohocken, PA.
ASTM D5311 2004. Standard test method for load controlled
cyclic triaxial strength of soil. American Society for Testing and Materials, West Conshohocken, PA.
ASTM D5617 2004. Standard test method for multi-axial tension test for geosynthetics. American Society for Testing
and Materials, West Conshohocken, PA.
ASTM D6706 2003. Standard test method for measuring geosynthetic pullout resistance in soil. American Society for
Testing and Materials, West Conshohocken, PA.
Barker, W.R. 1987. Open-Graded based for airfield pavements.
Technical report GL-87-16, U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg, Mississippi, USA, 76p.
Barksdale, R.D., Brown, S.F. and Chan, F. 1989. Potential
benefits of geosynthetics in flexible pavement system.
National Cooperative Highway Research Program, Report No. 315, Transportation Research Board, Nation Research Council, Washington, DC.
Bay, J.A., and Stokoe, K.H. 1998. Development of a Rolling
Dynamic Deflectometer for continuous deflection mea-

The available literature involving field and laboratory test results is conclusive in that the mechanical
properties of the geosynthetics used for pavement
applications are improved under the confinement
provided by the soil. Field test sections showed improved performance in the reinforced sections over
the unreinforced sections in terms of reduced surface
deflections. Overall, available experimental evidence indicates that the improved performance of
geosynthetic-reinforced pavements can be attributed
to lateral restraint mechanisms. Attempts have been
made to quantify the lateral restraint in terms of the
interface shear stiffness property of the soilgeosynthetic system.
A number of confined laboratory tests have been recently developed with the objective of quantifying
the interface shear stiffness of the soil-geosynthetic
system. Several of these tests have applied cyclic
loads to the soil-geosynthetic system in an attempt to
simulate the dynamic nature of traffic-induced loading. However, probably due to the fact that measurements are sensitive to small changes in displacements, currently available methods have
resulted in significant scatter in test results. This has
compromised the repeatability of the approaches and
has made it difficult to differentiate the performance
among different geosynthetics. Ongoing research focusing on confined testing under low displacements
using monotonic loading pullout stiffness test ap396

surements of pavements. Center of Transportation Research, Report 1422-3F, University of Texas, Austin,
TX.
Bender, D.A. and Barnberg, E.J 1978. Design of soil-fabricaggregate systems. Transportation Research Record 671,
pp 64-75.
Benjamin, C.V.S., Bueno, B., Zornberg, J.G. 2007. Field monitoring evaluation of geotextile-reinforced soil retaining
walls. Geosynthetics International Journal, April, Vol.
14, No. 1.
Berg, R.R., Christopher, B.R. and Perkins, S.W. 2000. Geosynthetic reinforcement of the aggregate base/subbase
courses of flexible pavement structures-GMA white paper II. Geosynthetic Materials Association, Roseville,
MN, USA, 176p.
Bray, J.D. and Merry, S.M. 1999. A comparison of the response of geosynthetics in the multi-axial and uniaxial
test devices. Geosynthetics International, Vol.6, No. 1,
pp. 19-40.
Brown, S.F., Jones, C.P.D. and Brodrick, B.V. 1982. Use of
Non-Woven fabrics in permanent road pavements. Proceedings of the Institution of Civil Engineers, part 2, Vol.
73, pp. 541-563.
Bueno, B.S., Benjamim, C.V., and Zornberg, J.G. 2005. Field
performance of a full-scale retaining wall reinforced with
non-woven geotextiles. Slopes and Retaining Structures
under Seismic and Static Conditions, ASCE Geotechnical
Special Publication No. 140, January 2005, Austin, Texas
(CD-ROM).
Bueno, B.S., Costanzia, M.A., and Zornberg, J.G. 2005. Conventional and accelerated creep tests on nonwoven
needle-punched geotextiles. Geosynthetics International,
December, Vol. 12, No. 6, pp. 276-287.
Burd, H.J. and Brocklehurst, C.J. 1990. Finite element studies
of the mechanics of reinforced unpaved roads. Proceedings of the 4th International conference on Geotextiles,
Geomembranes and Related Products, The Hague, Netherlands, pp. 217-221.
Burd, H.J and Houlsby, G.T. 1986. A large strain finite element
formulation for one dimensional membrane elements.
Computers and Geotechnics, Vol. 2, pp. 3-22.
Cancelli, A., Montanelli, F., Rimoldi, P. and Zhao, A. 1996.
Full scale laboratory testing on Geosynthetic-reinforced
paved roads. Earth Reinforcement, Ochiai, H., Yasufuku,
N., and Omine, K., Editors, Balkema, Proceeding of the
International Symposium on Earth Reinforcement, Fukuoka, Kyushu, Japan, November 1996, pp. 573-578.
Collin, J.G., Kinney, T.C. and Fu, X. 1996. Full scale highway
load test of flexible pavement systems with geogrid reinforced base courses. Geosynthetics International, Vol. 3,
No. 4, pp. 537-549.
Cuelho, E.L. and Perkins, S.W. 2005. Resilient interface shear
modulus from short-strip cyclic pullout tests. GSP-140,
Slopes and retaining structures under seismic and static
conditions, Geofrontiers, Austin, TX.
Dondi, G. 1994. Three-Dimensional finite element analysis of
a reinforced paved road. Proceedings of the Fifth International Conference on Geotextiles, Geomembranes and
Related Products, Singapore, pp. 95-100.
Dougan, Charles 2007. Mechanistic-Empirical Pavement Design Guide: project level pavement management. Lecture
Session 1a: PMS to support New MEPDG Norfolk, VA,
May 7, 2007.
Elias, V., Zehong, Y., Swan, R.H. and Bachus, R.C. 1998. Development of protocols for confined extension and creep
testing of geosynthetics for highway applications. FHWARD-97-143, Final report, 201 p.

Finnefrock, L.T. 2008. Bending stiffness index test for geogrid

reinforcement of pavement base course material. M.S.
Thesis, Department of Civil Engineering, The University
of Texas at Austin.
Giroud, J.P. and Han, J. 2004. Design method for geogridreinforced unpaved roadsI. Journal of Geotechnical
and Geoenvironmental Engineering, Volume 130,
Issue 8, pp. 775-786.
Giroud, J.P., Ah-Line, C., and Bonaparte, R. 1984. Design of
unpaved roads and trafficked areas with geogrids. Polymer Grid Reinforcement, A conference sponsored by
SERC and Netlon, Ltd., Thomas Telford, London, England, pp. 116-127.
Giroud, J.P. and Noiray, L. 1981. Geotextile-reinforced unpaved roads. Journal of Geotechnical Engineering Division, American Society of Civil Engineers, Vol. 107, No
GT9, pp. 1233-1254.
Gupta, R., McCartney, J.S., Nogueira, C.L. and Zornberg, J.G.
2008. Moisture migration in geogrid reinforced expansive
subgrades. Geoamericas 2008, Cancun, Mexico.
Gupta, R. 2009. A study of geosynthetic reinforced flexible
pavement system. Ph.D. Dissertation, submitted to The
University of Texas at Austin, Texas, USA.
Halliday, A.R. and Potter, J.F. 1984. The performance of a
flexible pavement constructed on a strong fabric. Transport and Road Research Laboratory, Report 1123, Crowthorne, Berkshire, United Kingdom, 15p.
Han, J., Zhang, Y., and Parsons, R.L. 2008. Development of a
performance-based laboratory test method for evaluating
geosynthetic-soil confinement. Geosynthetics Committee
(AFS70) TRB 2008 Annual meeting, Washington DC.
Haas R., Walls, J. and Carroll, R.G. 1988. Geogrid reinforcement of granular bases in flexible pavements. Transportation Research Record 1188, Washington DC, pp. 19-27.
Holtz, R.D, Christopher, B.R. and Berg, R.R. 1998. Geosynthetic design and construction guidelines. U.S. Department of Transportation, Federal Highway Administration, Washington, DC, FHWA-HI-98-038, 460 p.
Hsieh, C. and Mao, L. 2005. A bench-scale performance test
for evaluation of the geosynthetic reinforcement effects
on granular base courses. GRI-18 Geosynthetics Research
and Development in Progress, Geofrontiers, Austin, TX.
Huang, Y.H. 1993. Pavement analysis and design, PrenticeHall, Englewood Cliff, NJ.
Hugo, F., McCullough B.F., and Vander Walt B. 1991. Fullscale accelerated pavement testing for the Texas State
Department of Highways and Public Transportation.
Transportation Research Record 1293, Transportation
Research Board, National Research Council, Washington,
D.C., 1991, pp. 52-60.
Jewell, R.A. 1981. Some effects of reinforcement on soils.
PhD. Thesis, University of Cambridge, UK.
Kinney, T.C. and Yuan, X. 1995. Geogrid aperture rigidity by
in-plane rotation. Proceedings of Geosynthetics 1995, pp
525-537.
Koerner, R.M. 2005. Designing With Geosynthetics, 5th Edition, Prentice-Hall Inc., Englewood Cliffs, NJ, 1998, 796
p.
Konietzky, H., Kamp L., Groeger, T. and Jenner, C. 2004. Use
of DEM to model the interlocking effect of geogrids under static and cyclic loading. International PFC symposium 28-29 October, 2004, Kyoto, Japan, pp. 3-12.
Kupec, J. and McGown, A. 2004. The load-strain behavior of
biaxial geogrids. Proceedings of 3rd Asian Regional Conference on Geosynthetics, Seoul, South Korea, pp. 349356

397

Kwon, J., Tutumluer, E. and Kim, M. 2005. Development of a

mechanistic model for geosynthetic-reinforce flexible
pavements. Geosynthetics International, Vol. 12, No. 6,
310-320.
Kwon, J., Tutumluer, E. and Konietzky, H. 2008. Aggregate
base residual stress affecting geogrid reinforced flexible
pavement response. International Journal of Pavement
Engineering, Vol. 9, No. 4, August, 2008, pp. 275-285.
McDowell, G.R., Harireche, O., Konietzky H., Brown, S.F.,
and Thom, N.H. 2006. Discrete element modeling of
geogrid-reinforced aggregates. Proceedings of the Institution of Civil Engineers, Geotechnical Engineering 159,
January 2006, Issue GE1, pp.35-48.
McGown, A., Kupec, J. Heerten, G. and Maubeuge K. von.
2005. Testing biaxial geogrids for specification and design purposes. GRI-18 Geosynthetics research and development in progress, ASCE, Austin, Texas.
Miura, N., Sakai, A., Taesiri, Y., Yamanouchi, T. and Yasuhara, K. 1990. Polymer grid reinforced pavement on soft
clay grounds. Geotextiles and Geomembranes, Vol. 9,
No. 1, pp. 99-123.
Moghaddas-Nejad, F. and Small, J.C. 1996. Effect of geogrid
reinforcement in model track tests on pavements. Journal
of Transportation Engineering, Vol. 122, No. 6, pp. 468474.
Muench, S. 2006. http://pavementinteractive.org/
NCHRP 2004. NCHRP Project 1-37A, Guide for MechanisticEmpirical Design of new and rehabilitated pavement
structure. Washington, D.C.
NCHRP 2000. NCHRP Project 1-28A, Harmonized Test Methods for Laboratory Determination of Resilient Modulus
for Flexible Pavement Design, Volume 1. Unbound Granular Material, 198p.
Olidis, C. and Hein, D. 2004. Guide for Mechanistic-Empirical
Design of New and Rehabilitated Pavement Structures:
Material Characterization. Annual Conference of the
Transportation Association of Canada, Quebec City,
Quebec.
Palmeira, E.M. 1987. The study of soil-reinforcement interaction by means of large scale laboratory tests. PhD. Thesis, University of Oxford, UK, 238p.
Perkins, S.W. and Ismeik, M. 1997a. A Synthesis and Evaluation of Geosynthetic-reinforced Base Course Layers in
Flexible Pavements: Part I Experimental Work. Geosynthetics International, Vol. 4, No. 6, pp. 549-604.
Perkins, S.W. and Ismeik, M. 1997b. A Synthesis and Evaluation of Geosynthetic-reinforced Base Course Layers in
Flexible Pavements: Part II Analytical Work. Geosynthetics International, Vol. 4, No. 6, pp. 605-621.
Perkins, S.W. 1999. Mechanical Response of GeosyntheticReinforced Flexible pavements. Geosynthetics International, Vol. 6, No. 5, pp. 347-382.
Perkins, S.W. 1999. Geosynthetic reinforcement of flexible
pavements: laboratory based pavement test sections. U.S.
Department of Transportation, Federal Highway Administration, Washington DC, Report No. FHWA/MT99/8106-1, 140p.
Perkins, S.W. 2001. Numerical Modeling of Geosyntheticreinforced Flexible Pavements. Final report, FHWA/MT01-003/99160-2, 97p.
Perkins, S.W 2002. Evaluation of geosynthetic-reinforced flexible pavement systems using two pavement test facilities.
Final report, FHWA/MT-02-008/20040, Federal Highway
Administration, Washington DC, 120p.
Perkins, S.W. and Edens, M.Q. 2002. Finite element and distress models for geosynthetic-reinforced pavements. In-

ternational Journal of Pavement Engineering, Vol. 3, No.

4, pp. 239-250.
Perkins, S.W., Christopher, B.R., Cuelho, E.L., Eiksund, G.R.,
Hoff, I., Schwartz C.W., Svan, G. and Want, A. 2004.
Development of design methods for geosyntheticreinforced flexile pavements. FHWA-DTFH61-01-X00068, Final report, 263p.
Perkins, S.W. and Svan, G. 2004. Assessment of interface
shear growth from measured geosynthetic strains in a
reinforced pavement subject to repeated loads. In Compendium of Papers CD-ROM, 83rd Transportation Research Board (TRB) Annual Meeting, Washington, DC,
2004.
Perkins, S.W. and Cortez, E.R. 2005. Evaluation of basereinforced pavements using a heavy vehicle simulator.
Geosynthetic International, Vol. 12, No.2, pp. 86-98.
Reck, N.C. 2009. Mechanistic empirical design of geogrid reinforced paved flexible pavements. Jubilee symposium on
Polymer Grid Reinforcement, Institute of Civil Engineers,
London, England.
Shukla, S.K. 2002. Geosynthetics and their application. 1st edition, Thomas Telford Ltd., 425 p.
Sprague, C.J, Lothspeich, S., Chuck, F., and Goodman, R.
2004. Geogrid reinforcement of road base aggregatemeasuring the confinement benefit. Proceedings of GeoTrans 2004 Conference, Los Angeles, 2004, 996 -1005.
TRI 2001. In-plane rotational stiffness: Is this a relevant property for base reinforcement of geosynthetics? Internal report available at www.tri-env.com.
Tutumluer, E. and Dawson, A. 2004. Resilient characterization
of compacted aggregate. PowerPoint from TRB Workshop on describing aggregate behavior for todays pavements, Washington, DC.
Wartman, J., Harmanos, D., and Ibanez, P 2005. Development
of a versatile device for measuring the tensile properties
of geosynthetic. Proceeding of GRI-18 Conference,
ASCE, Austin, Texas, 4097-4102.
Wathugala, G.W., Huang, B. and Pal, S. 1996. Numerical Simulation of Geosynthetic-reinforced Flexible pavement.
Transportation Research Record 1534, TRB, National
Research Council, Washington, DC, USA, pp. 58-65.
Watts, G.R.A., and Blackman, D.I. 2009. Pavement trafficking
trials. Jubilee symposium on Polymer Grid Reinforcement, Institute of Civil Engineers, London, England.
Webster, S.L. 1993. Geogrid reinforced base courses for flexible pavements for light aircraft, test section construction,
behavior under traffic, laboratory tests, and design criteria. Technical report GL-93-6, U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg, Mississippi, USA, 86p.
WDOT 2007. Design Parameters for flexible pavementshttp://training.ce.washington.edu/wsdot/modulues/0
4_design_parameters
Yoder, E.J., and Witczak, M.W. 1975. Principles of pavement
design, 2nd edition, John Wiley and Sons, 711p.
Yuan, Z. 2005. Theoretical analysis of bending stiffness test on
geosynthetic-reinforced base layer. Proceedings of
NAGS/GRI-19 Cooperative Conference, Dec.14-16, Las
Vegas, Nevada, 2005.
Zornberg, J.G. and Gupta, R. (2009). Reinforcement of Pavements over Expansive Clay Subgrades. Proceedings of the
Seventeenth International Conference of Soil Mechanics
and Geotechnical Engineering, Alexandria, Egypt, 5-9 October, pp. 765- 768.

398