10 1 1 703 8121 PDF

Keynote Lecture reprint: Proceedings of the Fifth International Conference on
Geotextiles, Geomembranes and Related Products, Singapore, September 1994
Geosynthetic Reinforced Segmental Retaining Wall

Structures in North America
Richard J. Bathurst
Department of Civil Engineering
Royal Military College of Canada
Kingston, Ontario, Canada K7K 7B4
Michael R. Simac
Earth Improvement Technologies
100 Mayflower Court
Cramerton, NC 28032 USA
ABSTRACT: The Paper reviews recent research on the analysis, design and construction of
geosynthetic reinforced soil retaining walls that use dry-stacked modular concrete units as the
facing system (geosynthetic reinforced segmental retaining walls). These systems have gained
wide popularity in North America for reasons of performance, aesthetics, cost and expediency
of construction. However, the discrete nature of these modular block systems requires that
special attention be paid to the design and construction of the facing elements. Some
consequences of the extension of limit-equilibrium (pseudo-static) methods to the stability of
segmental retaining wall structures are reviewed. Design methodologies, construction and
specification recommendations described in this paper for routine structures have been recently
adopted by the National Concrete Masonry Association (NCMA).
1 INTRODUCTION
The use of dry-stacked columns of interlocking modular concrete units as the facing for
geosynthetic reinforced soil retaining wall structures has increased dramatically in North
America since their first appearance in the mid-1980s (e.g. Crowe et al. 1989, Anderson et al.
1991, Berg 1991, Simac et al. 1991, Hill and Berg 1993, Kemp et al. 1993). Examples of
completed projects are illustrated in Figure 1, 2 and 3. The National Concrete Masonry
Association (NCMA) in the USA has recently adopted the term Soil-reinforced Segmental
Retaining Wall (SRW) to identify this type of retaining wall system.
Reinforced segmental retaining wall systems offer advantages to the architect, engineer and
contractor as described below. The walls are constructed with segmental retaining wall units
(modular concrete block units) that have a wide range of aesthetically pleasing finishes and
provide flexibility with respect to layout of curves, corners and tiered wall construction. The base
course of modular units is typically seated on a granular bearing pad which offers cost advantages
over conventional poured-in-place concrete walls and some types of reinforced concrete panel
wall systems that routinely require a concrete bearing pad. For some large wet-cast units or
transportation related projects, concrete levelling pads may be used to maintain wall alignment
and batter. The mortarless modular concrete units are easily transportable and therefore facilitate
construction in difficult access locations. The mortarless construction and typically small
segmental retaining wall unit size and weight allows installation to proceed rapidly. An
experienced installation crew of three or four persons can typically erect 2040 square metres
Bathurst and Simac 1

Fig. 1 First major geosynthetic reinforced soil segmental retaining wall structure built in
Canada (7m height) (Crowe et al. 1989).
of wall face per day. The economic benefit due to these features is that reinforced segmental
retaining walls in excess of 1m in height typically offer a 25 to 45% cost saving over comparable
conventional cast-in-place concrete retaining walls (Berg 1991, Simac et al. 1991, Anderson et
al. 1991, Geotechnical Fabrics Report 1994). At the time of writing, the majority of reinforced
segmental retaining wall structures have been built using polymeric geogrid materials as the
geosynthetic reinforcement. Nevertheless, the design methodologies reviewed in this paper do
not preclude the use of some woven geotextiles which may introduce further economies for these
systems.
However, the discrete nature of the dry-stacked column of modular concrete units that is the
distinguishing feature of reinforced segmental retaining walls introduces additional and unique
design considerations. Conventional engineering practice for geosynthetic reinforced soil
retaining walls prior to 1993 did not fully address all performance issues for modular systems
since they were developed largely with the use of precast concrete panel systems in mind. The
paper reviews recent research on the analysis, design, construction and specification of
geosynthetic reinforced soil retaining walls that use dry-stacked modular concrete units as the
facing system. Design methodologies and construction recommendations described in this paper
for routine structures have been recently adopted by the National Concrete Masonry Association
(NCMA) in the USA (Simac et al. 1993a). The NCMA is an umbrella organization whose
mandate is to support and advance the common interest of its North American members in the
manufacture, marketing, research, and application of concrete masonry products.
2 SEGMENTAL RETAINING WALL UNITS
Modular concrete facing units are produced using machine molded or wet-casting methods and
are available in a wide range of shapes, sizes and finishes. Examples of some commercially
available segmental units are illustrated in Figure 4. Most proprietary units are 80 to 600mm in

Fig. 2 Tiered geosynthetic reinforced soil segmental retaining wall with 0.75m x 0.75m x
1.5m solid concrete units (Bowden 1991).
12kPa Surcharge
compacted 1.2m long PET geogrid l

drainage fill geogrid ll
masonry geogrid ll
concrete
facing units compacted drainage
1 fill with geotextile
5.5m
20
geogrid ll
6.1m
compacted native soil
for reinforced soil zone
geogrid ll
drainage collection pipe

0.3m (typ) gravity flow to outlet
slope 1:6 (typ) excavation limits

4.4m
total geogrid embedment length (all layers)
Fig. 3 Typical geosynthetic reinforced soil segmental retaining wall cross-section (after
Simac et al. 1991).

Not to scale
Fig. 4 Examples of segmental retaining wall units.
height, 150 to 800mm in width (toe to heel), and 150 to 1800mm in length. The modular units
typically vary from 14 to 48kg each. The modular concrete units may be solid, hollow, or hollow
and soil infilled. The units may be cast with a positive mechanical interlock in the form of
concrete shear keys or leading/trailing edges. Alternatively, interlocking between layers may be
developed by essentially flat frictional interfaces that may include mechanical connectors such
as pins, clips or wedges. The principal purpose of mechanical connectors is to assist with unit
alignment and to control wall facing batter during construction. Segmental retaining walls are
usually constructed with a stepped face that results in a facing batter that ranges from 3 to 15
degrees. The majority of facing systems are between 7 and 12 degrees. Shear transfer between
unit layers is developed primarily through shear keys and interface friction. However, for
interface layers under low normal pressures (e.g. close to the wall crest) a significant portion of
shear transfer may be developed by mechanical connectors.
Specification limits for dry-cast masonry concrete blocks in retaining wall applications are given
in Table 1. The physical requirements with respect to mix design can be found in separate
publications by the NCMA (1991) and the American Society for Testing and Materials (ASTM)
standards (e.g. ASTM C 90). Methods to sample and test concrete masonry units for compressive
strength, absorption, unit weight (density), moisture content and dimensions are provided in
ASTM C 140. The compressive strength of masonry concrete units meeting the minimum value
in Table 1 is more than adequate from a structural point of view. However, a minimum
compressive strength of 40MPa has been required by at least one state department of
transportation in the USA and this strength has been achieved by adjusting the mix design and

Table 1: Specifications for segmental retaining wall

masonry units (NCMA 1991, ASTM C 55).
Minimum Maximum Dimensional

Compressive Water Tolerances
Strength Absorption (mm)
(MPa) (%)
20.7 67* 3
* Minnesota Department of Transport (Hill and Berg

1993)
manufacturing process. Reinforcement steel is not used in dry-cast masonry units or wet-cast
units for reinforced segmental retaining wall applications in North America.
3 ANALYSIS AND DESIGN
Methodologies for the analysis and design of segmental retaining walls in the United States can
be found in guidelines published by three different organizations: the Federal Highway
Administration (FHWA) (Christopher et al. 1989); the American Association of State Highway
and Transportation Officials (AASHTO 1990a,b; 1992); and the National Concrete Masonry
Association (NCMA) (Simac et al. 1993). In Canada, guidelines based on the FHWA and
AASHTO documents are found in the 3rd Edition of the Canadian Foundation Engineering
Manual (1992). In order to be consistent with conventional North American practice for the
design and analysis of retaining wall structures, stability calculations in these guidelines adopt
a limit-equilibrium approach together with the assumption of c soils. The cited references all
adopt a gravity structure approach for external stability calculations and variations of the
tied-back wedge approach for internal stability calculations.
A critical review of the design and analysis methodologies found in NCMA guidelines and the
earlier FHWA and AASHTO references can be found in papers by Bathurst et al. 1993a and
Simac et al. 1993b. The NCMA guidelines are essentially a refinement of earlier FHWA and
AASHTO guidelines. The FHWA and AASHTO guidelines were developed based on experience
with geosynthetic reinforced soil retaining wall systems that used primarily precast concrete
panels. The NCMA methodology has the advantage that the designer can quantify the influence
of different candidate facing units on the stability of otherwise identical geosynthetic reinforced
soil walls. The NCMA manual also includes an integrated design and analysis approach for
conventional (gravity) structures that use unreinforced backfills. Hence the NCMA guidelines
offer a unified approach for unreinforced and reinforced segmental retaining wall systems
consistent with the notion that both types are essentially gravity structures. Finally, the NCMA
document reduces some conservatism found in the FHWA and AASHTO guidelines with respect
to the choice of earth pressure theory, base eccentricity criteria and minimum reinforcement
lengths. Based on the comments made above, the analytical approach described in this paper for
routine structures is based on the NCMA guidelines which were prepared by the authors and
co-workers (Simac et al. 1993a,b; Bathurst et al. 1993a).

4 MODES OF FAILURE
Potential failure modes for reinforced segmental retaining wall structures are illustrated in
Figure 5. External failure mechanisms consider the stability of an equivalent gravity structure
comprising the facing units, geosynthetic reinforcement and reinforced soil fill. Not included
in Figure 5 is global instability which involves failure mechanisms passing through or beyond
the reinforced soil mass. Conventional slope stability methods of analysis that have been
modified to include the stabilizing influence of horizontal layers of geosynthetic reinforcement
can be used for this purpose (e.g. Christopher et al. 1989). Modes of failure that require special
considerations in reinforced segmental retaining wall design and analysis are illustrated in the
last four diagrams of Figure 5. The reinforcement layers are placed between the masonry units
to form an essentially frictional connection. The modular unit-geosynthetic reinforcement
connection capacity can control the spacing and the selection of polymeric reinforcement type.
Similarly, adequate unit to unit interface shear capacity is required to prevent internal sliding
mechanisms that propagate through to the face of the structure and/or to prevent local bulging
of the facing units.
5 APPLICATION OF EARTH PRESSURE THEORY
Limit-equilibrium approaches are routinely adopted for the design and analysis of reinforced
segmental retaining walls. Earth pressure distributions and important wall geometry parameters
are illustrated in Figure 6. In addition to the wall batter () that is generated by the built-in
a) base sliding b) overturning c) bearing capacity

(excessive settlement)
d) pullout e) tensile over-stress f) internal sliding
g) connection failure h) shear failure i) toppling

(bulging)
Fig. 5 Modes of failure: external (top row); internal (middle row); facing (bottom row)
(Simac et al. 1993a).

anchorage length for pullout calculations (Le)

inclined surface Coulomb inclined
for internal stability internal surface for
calculations failure
reinforced soil external
plane stability
segmental calculations
retaining
wall units
+ +

Ka v(z)
z
z
H H
Ka v(z)
Sv
ib retained soil
i e
foundation soil
base reinforcement length
granular
bearing pad
Fig. 6 Principal components, geometry and earth pressures assumed in NCMA method
(Bathurst et al. 1993a).
setback of the units, the base course may be inclined at some angle ib which results in a further
net wall face inclination () from the vertical ( = + ib). The choice of Rankine or Coulomb
earth pressure theory varies between guidelines. The Coulomb approach has been adopted by
the authors for all stability calculations because it can explicitly accommodate the contribution
to lateral earth pressure of wall inclination angle (), backslope angle () and shear mobilized
at the interfaces between the reinforced soil and retained soil zones (interface friction angle e),
and the facing column and reinforced soil zone (interface friction angle i). In addition, other
research has demonstrated that Rankine theory over-estimates lateral pressures acting within
instrumented geosynthetic reinforced soil retaining wall structures (e.g. Berg et al. 1986,
Bathurst and Simac 1991). Outward movement of the facing and settlement of the reinforced soil
mass is assumed to generate positive interface shear at the back of the facing units (+i). For
internal stability calculations the interface shear angle acting between the inclined surface ()
and the reinforced soil is taken as i=2/3. This general approach is consistent with conventional
retaining wall design but is slightly more conservative than in AASHTO (1992) guidelines that
recommend i=3/4 for granular soil in contact with prefabricated modular concrete panels.
Interface friction is assumed to be fully mobilized at the back of the reinforced soil zone (i.e. e
= where is the lesser of the peak friction angle for the retained soil and reinforced soil
materials).
In the NCMA manual only the horizontal component of lateral earth pressure due to soil
self-weight and any uniformly distributed surcharge loading is considered for external and
internal stability calculations. This assumption simplifies calculations and results in the
conservative assumption that the vertical component of earth pressures does not contribute to
resisting forces in stability calculations. The influence of the choice of Rankine theory (equation

1) or Coulomb theory (equation 2) on the magnitude of horizontal earth pressure is illustrated

in Figure 7. Coulomb earth pressures are 15% to 45% lower than those calculated by Rankine
theory for typical segmental retaining wall geometries. Boussinesq solutions are used to
calculate any additional lateral stresses due to line loads and other finite distributed surface loads.
Lateral earth pressures are integrated over the contributory area of each reinforcement layer (Sv
in Figure 6) to calculate the maximum tensile load (Tmax) to be carried by each layer: e.g. Tmax
= SvKa cos() v(z). This tensile load is used in the calculation of factors of safety against
reinforcement pullout, tensile over-stress, and connection failure. Generally, the value of Tmax
will increase with depth of the layer below the crest of the wall, particularly if uniform layer
spacings are adopted.
To be consistent with Coulomb theory, the orientation of potential failure planes (Figure 6)
cos cos 2 cos 2

K a = cos (1)
cos + cos 2 cos 2
cos 2( + )
Ka =
[ cos()cos(+)
cos 2() cos( ) 1 +
sin(+)sin ()
] 2 (2)

(2)
R= h Coulomb
h (1) Rankine
h
=2/3
=
1.1
= 35o
1.0

0.9
0.8 0o
R 3o
0.7 5o
typical
10o range
0.6
15o
0.5
20o
0.4
0 5 10 15 20 25 30 35 40 45
(degrees)
Fig. 7 Ratio of Coulomb solution to Rankine solution for the calculation of horizontal
earth pressures (Bathurst et al. 1993a).

through the reinforced soil zone are calculated as = f(,,,i). The closed-form solution can
be found in geotechnical engineering textbooks. The internal failure plane is used to locate the
active wedge that must be restrained by the anchorage zone in pullout capacity calculations. For
design purposes the internal failure plane is assumed to propagate from the heel of the lowermost
facing column unit. An implication of the Coulomb approach to internal stability calculations
is that internal failure planes are shallower than those calculated using the Rankine solution as
illustrated in Figure 8. In order to satisfy pullout criteria, some reinforcement layer lengths close
to the crest of the wall may have to be longer than those required at the base of the wall. However,
the NCMA method does not require that all reinforcement layers have the same length as
required in the AASHTO documents. NCMA requires that the minimum length of all
reinforcement layers be at least equal to the base length of the reinforced mass required to satisfy
all external stability requirements but not less than 0.6H for critical structures or 0.5H for
non-critical structures (see Table 2). The designer is permitted to locally increase the width of
=0
internal
failure
plane
L
70
Rankine theory 0o
= 45o + /2 5o
65
10o
15o
60
20o
(degrees)
55
50
Coulomb theory
45 = f(, , , i)
i = 2/3
=0
40
20 30 40 50
friction angle (degrees)
Fig. 8 Comparison of internal failure plane orientation based on Rankine and Coulomb
earth pressure theory (Bathurst et al. 1993a).

the reinforced soil zone and the length of individual layers near the crest of the wall as required
to satisfy pullout criteria.
6 CALCULATION OF ALLOWABLE TENSILE LOAD AND ANCHORAGE CAPACITY
The calculation of long-term design strength (LTDS) in North America for reinforced soil struc-
tures is based on the application of partial factors to a reference index strength. Strategies used
to select appropriate values for partial factors and factor of safety expressions for tensile over-
stress and anchorage failure for other geosynthetic reinforced soil retaining wall structures are
equally applicable to segmental wall design. A brief description of internal stability calculations
follows for completeness. The reader is referred to the references at the end of the paper for a
more thorough treatment of the topic and for details of the variations to the general approach
illustrated here.
The following expressions can be used to calculate the maximum allowable strength (Ta) for
a reinforcement layer and are adapted from FHWA guidelines (Christopher et al. 1989, Berg
1992) and Geosynthetic Research Institute (GRI) standards:
T index
LTDS = (6)
FS CR FS ID FS CD FS BD FS MU
Ta = LTDS/FSOS > Tmax (7)
Here Tindex refers to the minimum average roll value from wide-width strip tests (ASTM D
4595). Partial factors of safety are defined as: FSCR = partial factor of safety for creep deforma-
tion (ratio of Tindex to creep limiting strength taken at 10% of in-isolation strain); FSID = partial
factor of safety for installation damage; FSCD = partial factor of safety for chemical degradation;
FSBD = partial factor of safety for biological degradation; FSMU = partial factor of safety for ma-
terial uncertainty (1.5 minimum).
Pullout resistance (RPO) and the factor of safety against anchorage failure (FSPO) are calculated
as:
RPO = 2 Le Ci v tan (8)
FSPO = RPO/Tmax (9)
where: Le = anchorage length beyond the internal failure plane (Figure 6); Ci = coefficient of
shear stress interaction; v = average vertical stress acting over the geosynthetic in the anchorage
zone. An alternative but more complicated equation for geosynthetic reinforcement pullout
capacity has been reported by Christopher et al. (1989). However, equation (8) offers simplicity
with no proven loss in accuracy. Minimum factors of safety against tensile over-stress (FSOS)
and anchorage failure (FSPO) are given in Table 2 based on both peak load and deformation
criteria.
7 FACING STABILITY
The discrete nature of the dry-stacked modular column construction in reinforced segmental
retaining wall design requires performance data and analytical models that are unique to these
systems. Calculation of interface shear capacity and connection capacity developed between the

modular units and the polymeric reinforcement requires that an estimate of the normal stress
transmitted between stacked units be made. The magnitude of normal stress will depend on: the
height of wall above any interface; the wall inclination; and the magnitude of vertical load
transmitted to the column by wall-soil shear over the design life of the structure. A lower bound
estimate of normal load can be made by calculating the hinge-height of the unsupported facing
column. The calculation is related to the maximum number of units to satisfy moment
equilibrium about the heel of the column. The general expression for hinge height He is:
He = 2[(WuGu0.5Hu tanib) cos ib]/tan(+ib) (10)
where: Wu is the width of the modular unit; Gu is the distance to the centre of gravity of the unit
measured from the front face; Hu is the height of the unit; ib is the base inclination angle; and
is the wall batter. The concept is illustrated in Figure 9 which shows that for a typical solid
unit with a block width to height ratio of two, the number of units corresponding to the hinge
weight of
facing units
falling outside of
heel
hinge
height
He retained soil
WL
ib WR
heel
WR + WL
50
40
N =2
cos i b
tan( + i b)
1tan i b
2

number of units (N)
example facing
30 unit geometry
B/2
20
B
10 ib
10o 5o 0o
15o
0
0 5 10 15 20
(degrees)
Fig. 9 Influence of wall inclination on number of facing units within hinge height
(after Bathurst et al. 1993a).

height diminishes rapidly with increasing wall inclination. Clearly, as the wall batter = +
ib ! 0 the hinge height will be restricted to the height of the column above the interface.
7.1 Interface shear capacity
In order to satisfy minimum factors of safety against block to block sliding, the shear capacity
of any interface between units must be determined. The calculation of required shear capacity
is based on a continuously supported beam analog in which the lateral earth pressure is taken as
the distributed load and the reinforcement layers as the supports. Unit to unit interface sliding
also contributes to the resistance to internal sliding as illustrated in Figure 5f. The magnitude of
shear capacity available at the interface of concrete units can only be established by full-scale
direct shear testing. The authors have developed a large-scale test apparatus to quantify interface
shear under varying normal load (Figure 10). Peak interface shear capacity test data is illustrated
in Figure 11 for six different modular units using a method of test developed by the authors. The
data in Figure 11 has been grouped into four categories. The hollow units were infilled with a
compacted angular stone with a top size of 18mm. The toe to heel dimension of the units ranged
from 300 to 600mm. For a given connection type the relationship between peak interface shear
capacity (SSU) and normal load (Nn) can be described by a simple Coulomb failure criterion:
SSU = asu + n tan su (11)
In fact, interface shear capacity may be developed by shear key interlock and to a lesser extent
by mechanical connectors that are part of some systems. The authors have had some success in
predicting the shear capacity of units that have a continuous concrete shear key or concrete
trailing edge, based on the concrete strength, failure mechanism and dimensions of the unit. The
data set in Figure 11 is limited but does illustrate that: a) there is a wide range of interface shear
capacity between different types of units; and b) concrete shear keys or concrete trailing edges
provide a more efficient shear connection than purely frictional interfaces. However, the
1 loading frame 2 horizontal actuator

3 horizontal load cell 4 reaction beam
5 vertical load cell 6 air bag
7 platform 8 spacers
9 modular block 10 geosynthetic
dimensions in mm 4
5
6
8
10 1200
2 3 9
7
1
1500
Fig. 10 Schematic of shear connection test apparatus (with geosynthetic inclusion).

200
Nn
180
160 SSU
peak interface shear (kN/m)

140
solid (shear key)
hollow
120 (infilled with shear key)
solid (trailing concrete edge)
100 hollow (infilled)
80 [1]
SSU = 5 + Nn tan 37o
60
40 [2]
20

0
0 20 40 60 80 100
normal load (kN/m)
[1] lower bound limit on units with

concrete shear key or concrete trailing
edge (solid and hollow infilled)
[2] hollow infilled unit with flat interface
Fig. 11 Example peak interface shear versus normal load data.
fully-mobilized interface shear capacity can be expected to be reduced by the presence of any
geosynthetic reinforcement inclusion that is present as part of the wall facing-geosynthetic
reinforcement connection detail. Figure 12 demonstrates the reduction in peak interface shear
capacity for a solid unit that results from the presence of a polyester geogrid inclusion. The data
in Figures 11 and 12 demonstrates that shear tests should be carried out with and without a
geosynthetic between course layers. The design interface shear capacity (VU) is calculated as:
VU = SSU/FSSC (12)
with the choice of factor of safety (FSSC) based on either a peak or deformation (serviceability)
criteria (Table 2). The choice of deformation criterion can be based on maximum permissible
lateral wall deformations during wall construction. The test protocol developed to generate the
data presented in this section has been adopted by the NCMA (Simac et al. 1993a).

200
Nn
geogrid
180
SSU
160
140
peak interface shear (kN/m)
(concrete to
120 concrete interface)
100
80
60
40

20 (with polyester geogrid
within interface)
0
0 20 40 60 80 100
normal load (kN/m)
Fig. 12 Influence of geosynthetic layer inclusion on peak interface shear capacity (solid unit
with continuous concrete shear key).
7.2 Facing connection performance
The load-deformation properties of the connection formed by placing the reinforcement between
block courses have been quantified by the results of full-scale laboratory connection testing as
described by Bathurst and Simac (1993a). The tensile load-deformation properties are
influenced by: a) geometry and type of geosynthetic-facing unit interface (i.e. continuous keys,
trailing edges, dowels or pins); b) quality of the concrete; c) whether the facing units are hollow
or solid; d) whether the hollow core is left empty or infilled with a granular soil; e) tolerances
on block dimensions, f) quality of construction; and g) thickness, structure and polymer of the
geosynthetic. A test apparatus developed at the Royal Military College of Canada (RMC) to
carry out prototype-scale connection tests is illustrated in Figure 13. A similar device has been
developed by researchers at the University of Wisconsin-Platteville (UWP) (Buttry et al. 1993).
A protocol for carrying out these tests has also been developed by the authors and has been
adopted by the NCMA (Simac et al. 1993b). A slightly modified version of this procedure is
under ballot as an ASTM Standard Test Method at the time of writing. Typical test results are
illustrated in Figure 14. Connection strengths have been normalized with respect to the ultimate

1 masonry concrete block 11 spacers

2 geosynthetic 12 platform
3 loading platen 13wire-line extensometer
4 gum rubber mat/air bag 14computer-controlled
5 roller clamp hydraulic actuator
6 lateral restraining system 15 load cell
7 guide rail
8 extensometer clamp
9 surcharge actuator
10 loading frame
10
200mm
2
5 9
1.5m 8
13
15
connection detail 3
13 14 5 4
15 1
2
7 11
12
6
3m
initial 200mm free length
wire-line extensometer
1000 mm wide
sample
350
extending to front
mm of connection
interface
Fig. 13 Schematic of RMC connection test apparatus showing typical masonry concrete
block units and geosynthetic reinforcement (Bathurst and Simac 1993a).
(index) strength of the product determined according to the ASTM D 4595 (Wide-Width Strip)
Test Method. In fact, the rate of loading, free length of the reinforcement, laboratory ambient
temperature and relative humidity required in the NCMA method of test have been selected to
match the ASTM D 4595 method of test. Essentially, the facing connection test method
developed at RMC can be thought of as the standard in-isolation tensile test performed with one
set of poor clamps (i.e. the block-geosynthetic connection). In this way meaningful comparisons
between connection configurations can be made and their connection capacities related to a
widely used index strength value. The maximum connection strength may be due to: a) pullout
of the reinforcement through the interface; b) rupture of the reinforcement; c) failure of the
concrete units; or d) combinations of these mechanisms.
Connection strength envelopes based on peak capacity and a 20mm displacement criterion
measured at the connection point at the back of the facing units are plotted in Figure 15. The

1.0
normalized tensile load Tpull /T index

normal normal stress
stress
0.8 275 kPa 192 kPa
Tpull
0.6
rupture
0.4 137 kPa

pullout
0.2
27 kPa
0.0
0 10 20 30 40 50 60
horizontal displacement (mm)
Strong uniaxial polyethylene geogrid in com-
bination with solid concrete units (2m long)
constructed with continuous concrete key
Fig. 14 Example normalized connection load-displacement curves (Bathurst and Simac

1993a).
peak tensile load
0.4 pullout rupture

Tconn / Tult
0.2 20mm displacement

criterion
equivalent depth of connection (m)
2 4 6 8 10
0
0 50 100 150 200
normal stress (kPa)
medium strength woven polyester geogrid-
300mm wide by 450mm long hollow core
masonry concrete block (granular infill)
Fig. 15 Example normalized connection strength curves (Bathurst and Simac 1993a).
20mm displacement criterion has been used by the authors to be consistent with the displacement
criterion for pullout testing of geosynthetics in soil recommended by AASHTO (1990b). The
bi-linear shape of the connection envelopes plotted in the figure are typical of most modular
block-geosynthetic reinforcement material combinations investigated by the authors. An
explanation for the bi-linear trend in the data is that the polymeric reinforcement inclusion

suffers mechanical damage when placed between concrete block units, particularly after large
surcharge pressures have been applied. Mechanical damage to the reinforcement occurs as a
result of (typically): rough concrete surfaces; irregularities in unit dimensions; joints between
the blocks; and sharp edges. These conditions lead to pinching and crushing of the reinforcement
during construction and result in uneven load transfer within the connection. The initial
mechanical damage increases at higher confining pressures and prevents the further increase in
connection capacity with depth that might otherwise be anticipated for an essentially frictional
connection. The results of more than 1350 tests from RMC and UWP are summarized in Figure
16. The data represents 29 different segmental unit types, 20 different polymeric reinforcement
materials and 131 different combinations of segmental unit and geosynthetic reinforcement. The
reinforcement types include polypropylene geotextiles, and polyester and HDPE geogrids. The
data in Figure 16 shows that there can be a significant reduction in connection capacity when
compared to the index strength of the reinforcement for the reasons stated above. The data
highlights the need to perform product-specific testing to determine accurate project data.
Based on connection capacity envelopes of the type shown in Figure 15, a design capacity
envelope can be developed for critical structures that is the lesser of the peak load envelope,
divided by a factor of safety value (FSCS), or the deformation envelope. The maximum tensile
load in any reinforcement layer (Tmax), calculated according to the contributory area approach,
must fall below this design envelope.
8 SUMMARY OF DESIGN CRITERIA
100
90
general range
80
Tconn / Tult (percent)
70
60
50
40
30
20 RMC
10 UWP
0
0 20 40 60 80 100 120 140
normal load (kPa)
Fig. 16 Summary of connection test data.

A summary of principal design criteria for reinforced segmental retaining walls is given in Table
2. The factors of safety correspond to all potential failure mechanisms described earlier and have
been taken from the NCMA guidelines (Simac et al. 1993a). Many of the deformation limits used
in serviceability criteria in the NCMA guidelines have been selected to meet values
recommended in FHWA and AASHTO guidelines.
9 COMPUTER AIDS
The analysis and design steps described in the paper lend themselves to solution using computer
programs. A number of proprietary computer packages are available from manufacturers of
modular block systems and the suppliers of geosynthetic reinforcement materials. A generic
software package called GEOWALL (ver 2.0) has been recently released that implements many
of the recommendations found in the NCMA guidelines. This program is an upgrade of the
program described in the paper by Bathurst and Simac (1993b). The NCMA will release a
computer program called SRWall in early 1995 that will provide a full implementation of the
NCMA guidelines. The program has been developed for the Microsoft Windows operating
system. A typical graphics user interface (GUI) from the program is illustrated in Figure 17.
10 CASE STUDIES
A number of geosynthetic reinforced segmental retaining walls in North America have been
monitored. The data from these structures is useful to verify design methodologies and provide
performance data.
10.1 Algonquin Wall
A research project sponsored by the Federal Highway Administration (FHWA) was initiated in
the late 1980s in Algonquin, Illinois, USA that involved the full-scale construction and testing
of several proprietary retaining wall systems using both extensible (polymeric) and in-extensible
reinforcement with a variety of facing types (Christopher et al. 1989, Simac et al. 1990,
Bonczkiewicz et al. 1991, Bathurst et al. 1993b, Christopher 1993). The wall that is discussed
here is illustrated in Figure 18 and 19 and had a total length of 15m and a crest height of 6.1m.
The wall was heavily instrumented and was monitored for a period of 479 days following
completion. The facing comprised 47.5kg hollow (soil infilled) modular masonry concrete
facing units (Keystone Standard) with the nominal unit dimensions shown in Figure 20. A pair
of 13mm diameter pins was used to provide an automatic setback of 10mm per course. The
resulting setback gave a target facing batter of 20 vertical to 1 horizontal. The reinforcement for
this project was a high-tenacity woven polyester geogrid distributed by Nicolon/Mirafi
(Miragrid 5T). The reinforcement was placed over the masonry units and pre-tensioned by hand.
A uniform size pea gravel was used to infill the hollow portions of the modular block units and
the spaces between adjoining units. The pea gravel was also extended to a distance of 600mm
behind the facing columns to create a drainage layer which is a common recommended practice
in reinforced segmental retaining wall design. The reinforced and retained soil comprised a
cohesionless well-graded sandy gravel with a peak friction angle of 40 degrees (Christopher
1993) and this material was compacted to 95% of standard Proctor.
The wall was designed using a computer program that implements a conventional
limit-equilibrium approach together with Coulomb-type soils (Bathurst and Simac 1993b). The
design methodology was based on recommendations contained in AASHTO and FHWA

Table 2: Recommended minimum factors of safety for design of geosynthetic reinforced

soil segmental retaining walls.
APPLICATIONS
CRITICAL NON-CRITICAL
FAILURE MODE
Base Sliding FSSLD 1.5 1.5
Overturning FSOT 2.0 2.0
Bearing Capacity FSBC 2.0 2.0
Global Stability FSGL 1.5 1.3
Tensile Over-stress FSos 1.2 1.0
Pullout (peak load criterion) FSPO 1.5 1.5

Pullout (serviceability criterion) FSPO 1.0 N/A
Facing Shear (peak load criterion) FSSC 1.5 1.5

Facing Shear (serviceability criterion) FSSC 1.0 N/A
Connection (peak load criterion) FSCS 1.5 1.5

Connection (deformation criterion) FSCS 1.0 N/A
NOTES:
1. The minimum factors of safety given in this table assume that stability calculations are based on measured
site-specific soil/wall data. Measured data are defined as the results of tests carried out on actual samples
of soils and geosynthetic products for the proposed structure and actual samples of masonry concrete units
(i.e. the same molds, forms, mix design and infill material or same broad soil classification type (e.g. G,S)
if applicable).
2. The designer should use larger factors of safety than those shown in this table or conservative estimates of
parameter values when estimated data is used. Estimated data includes bulk unit weight and shear strength
properties taken from the results of ASTM methods of test (or similar protocols) carried out on samples of
soil having the same USCS classification as the project soil and the same geosynthetic product. Estimated
data for facing shear capacity and connection capacity analyses shall be based on laboratory tests carried
out on the same masonry concrete unit type under representative surcharge pressures for the project structure
(and the same broad soil classification type (e.g. G,S) if applicable).
3. For critical structures, minimum factors of safety based on serviceability and peak load criteria must be
satisfied for pullout, facing shear and facing connection failure modes.
4. Design of the maximum unreinforced wall height at the top of the structure is carried out using the stability
analyses and factors of safety recommended for conventional (gravity) segmental retaining walls.
5. Minimum wall embedment depths as a function of wall height follow recommendations given in
AASHTO/FHWA. In no case shall the minimum wall embedment depth be less than 0.45m for critical struc-
tures or 0.15m for non-critical structures.
6. A non-critical structure in NCMA manual terminology is a structure in which loss of life would not occur
as a result of wall failure nor would failure result in significant property damage or loss of necessary function
of adjacent services or structures. A critical structure is clearly the converse. Permanent structures are
usually considered critical structures and are designed for a 75 to 100 year life. Similarly,
transportation-related structures would normally be considered critical structures.
(after Simac et al. 1993a)

Fig. 17 Program SRWall.
Fig. 18 Algonquin test wall.
guidelines in effect at the time of the project (Christopher et al. 1989 and AASHTO 1990a,b).
The exceptions to AASHTO/FHWA guidelines were the use of 4.3m lengths of reinforcement
(versus 3.5m required) and the use of an overall factor of safety (FSMU) of 1.0 to calculate the
long-term allowable design strength of the reinforcement. The 4.3m lengths were chosen to be
consistent with a number of other structures at the site (Christopher et al. 1989). The number and
vertical arrangement of the geogrid reinforcement was selected to give a factor of safety of 1.1
with respect to over-stressing of the reinforcement beyond its serviceability limit load of
15.5kN/m (strain < 10%) based on in-isolation creep tensile testing of the geogrid material
together with typical partial factors of safety for construction damage and durability (Simac et
al. 1990). With the surcharge in place the same factor of safety was reduced to 0.8. The resulting
deficiency in internal design with respect to over-stressing of the reinforcement was purposely
carried out to encourage large reinforcement strains and wall movements and hence examine

inclinometer
modular concrete Phase II
facing unit INCLINED
340 2.1m SURCHARGE
6 Phase I
elevation
1 8
5
geogrid
HEIGHT IN METRES
7 well-graded
4 sand and
20
6 gravel
3 (SWGW)
5
extensometer
2 4
3
1 strain gauge
2
1
0
earth pressure cell
0.83m
2.7m
reinforcement
4.3m length
0 1 2 3 4 5
HORIZONTAL DISTANCE (m)
Fig. 19 Cross-section and instrumentation of Algonquin wall.
Fig. 20 Modular block unit: 200mm high by 600mm deep (toe to heel dimension) by
460mm wide (Algonquin wall).
conservatism in current methods of design used in North America at the time of the project.
The wall was constructed in six weeks using a three-man crew with a front end loader and
self-propelled steel drum roller. The wall was constructed with a routine standard of care using
conventional construction procedures interrupted only by installation of instrumentation. A
complete description of construction activities can be found in the paper by Simac et al. (1990).

A full description of the instrumentation program and all test data gathered from the project is
reported in the paper by Bathurst et al. (1993b).
The results of inclinometer readings are summarized in Figure 21. The data shows that
post-construction outward wall movements were generated over the bottom half of the wall.
Lateral earth pressures recorded for the bottommost pressure cell mounted against the back of
the facing column are summarized in Table 3. Only the data from this lateral earth pressure cell
17 109 surcharge
+17
9 31 0 +182 +370 days
6
5
4
3
2
elevation (m)
1
0
1
2
3
4 elapsed times are with respect
to end of construction and
5 surcharging (+)
6
10 0 10 20 30 40 50 60 70 80 90 100
deformation (mm)
Fig. 21 Post-construction front inclinometer deformations (Bathurst et al. 1993b).
Table 3: Comparison of measured and predicted lateral

earth pressures in Algonquin wall
(3.8m below wall crest).
Measured Rankine Coulomb

kPa (=0) (=0) (=2/3) (=)
end of construction
max15.2
min 10.2
avg 12.7 16.6 18.5 12.8 11.9kPa
surcharge in place
max15.8
min 9.92
avg 12.7 18.1 20.2 14.0 13.0kPa
NCMA method

was considered reliable by field technicians (Simac et al. 1990). However, all cells gave signals
indicating that lateral earth pressures were developed directly against the facing column. In
addition, strain gauges mounted directly on the reinforcement indicated load transmission within
the facing connections. Hence, the facing connections for these segmental retaining wall
structures must be designed to carry load and the horizontal interface between stacked facing
units must be designed to transmit shear. The data shows that the most accurate prediction of the
horizontal component of active lateral earth pressure is given by Coulomb solutions with the
inclination of active earth thrust taken as = or =2/3. Based on the experience with the cells
located at higher elevations, it is not possible to investigate the influence of any
compaction-induced stresses that may have developed over the top half of the wall.
Compaction-induced stresses could lead to greater pressures than those predicted by active earth
pressure theory. However, routine practice in the field is to minimize compaction stresses by
using light-weight compaction equipment directly behind the facing.
An advantage of most modular block systems is that they provide some lateral compliance
against compaction loads that may assist to minimize compaction-induced stresses. Further high
quality field monitoring of lateral pressures at the back modular block facing columns is required
to explore this issue further.
Reinforcement strains inferred from extensometer readings are summarized in Figure 22. The
2.0
Layer 8
1.5 within facing (top)
strain (%)
1.0 surcharge +0 days

surcharge
0.5 +17 days
145
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
2.0
+54 elapsed times are with
1.5 respect to end of construction
strain (%)
and surcharging (+)

+1
1.0
0.5 145
Layer 4
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
1.0
Layer 2
0.8 +54
strain (%)
0.6 +1
0.4
0.2 145
0.0
0.2
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
distance from front face (m)

Fig. 22 Summary of reinforcement strains at selected times (Bathurst et al. 1993b).

data shows that the maximum reinforcement strains in the reinforcement were not greater than
about 2%. With the exception of one reading at layer 4, all strains in the reinforcement layers
were less than 1%. The maximum strain reading of 2% is well within the creep limit of the
reinforcement. Despite being designed for strains as great as 10%, the loads in the reinforcement
did not approach design loads. The over-prediction of reinforcement loads by conventional
limit-equilibrium methods that employ active earth pressure theory has been noted by the first
author in large-scale model tests of other geosynthetic reinforced soil walls that used full-height
and flexible facing systems. In addition, significant footing restraint has been demonstrated to
occur in carefully monitored laboratory experiments and this mechanism may be responsible for
reducing the total tensile load carried by extensible reinforcement layers (Bathurst 1993). Based
on data from the Algonquin wall and other carefully monitored geosynthetic reinforced wall
structures there is increasing confidence that Coulomb earth pressure theory results in
conservative (i.e. safe) estimates of reinforcement tensile loads.
The distribution of base pressures at selected locations below the facing column and reinforced
soil zone is given in Figure 23. The peak pressure readings agree well with computed pressures
based on bulk unit weight of the reinforced soil and surcharge materials (z+q). However,
vertical pressure concentration towards the front of the wall was not observed as would be
predicted by considering eccentricity of the resultant vertical load from the reinforced zone and
a trapezoidal vertical pressure distribution. An implication of this data to design is that a base
eccentricity criterion of the type employed for rigid footings may be unwarranted for flexible
reinforced soil structures. Eccentricity criteria can be shown to control the minimum required
width of the reinforced soil zone in earlier design guidelines and hence adds to excessive
conservatism in the calculation of minimum reinforcement lengths.
below facing
surcharge q=36 kPa
180
z+q days
160
+245
140 +182
+1
120 +370
pressure (kPa)
38
6.1m
100 end of construction
3.66m
80 24
60
3.05m
20
40
z=1.42m
6
20
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
distance from wall face (m)
Fig. 23 Vertical pressure at base of reinforced soil mass (after Bathurst et al. 1993b).

The pressures recorded at the base of the facing column provide a check on the calculation of
normal force transferred across modular block interfaces. For the Algonquin wall the steep
facing batter results in the hinge height being equal to the height of the facing column above any
interface (He=H). The measured footing loads are plotted in Figure 24 together with theoretical
values calculated as the facing column weight alone and the facing column weight plus additional
down drag force due to facing column-soil interface shear. The figure shows that the column
self-weight predicted using just the column self-weight is accurate. Additional forces assumed
to be developed by interface shear result in over-prediction of interface normal force which
would result in unconservative design for interface shear and connection capacity design. It
should be noted that the Algonquin wall was constructed on a very stiff foundation which can
be expected to minimize relative downward movement of the reinforced soil mass relative to the
facing column. For example, the maximum vertical settlement of the wall was measured as
1.5mm with distortional settlements of about 0.01% along the length of the wall.
10.2 California Wall
A temporary geosynthetic reinforced segmental retaining wall was constructed in 1990 at the site
of a deep excavation in California (Figure 25 and 26). The wall was used to retain a sand backfill
N = W +Ka sin() ( z2 +z q)
Ka = f(, , =2/3) curve [1]
Ka = f(, , =) curve [2]
q=0 q>0
120
q
==3
100
[2]
normal force N (kN/m)
6.1m
80 W [1]

z Pa
60 measured
range
40
N
20
N=column weight W
0
0 1 2 3 4 5 6 7 8
height of facing column z (m)
Fig. 24 Comparison of measured normal force at base of wall with predicted values
(Algonquin wall).

Fig. 25 California wall.

to a height of 14.4m and had a base length of 76m. The modular units were identical to the units
used in the Algonquin wall (Figure 20) but were specially fabricated to give a zero setback (i.e.
no wall batter). The modular units were founded on a compacted granular base and reinforced
with 23 layers of a polyester geogrid with a long-term design strength of 59kN/m. Once the
retaining wall was completed it was used as the formwork for the basement wall that was cast
against the modular unit facing. The wall was designed to satisfy all limit-equilibrium based
performance criteria in accordance with AASHTO and FHWA guidelines. However, the vertical
out-of-alignment that developed as wall height increased resulted in encroachment of the wall
into the excavation. In order to monitor this movement, measurements were taken at six stations
along the wall as the height of the facing column increased. The data in Figure 27 represents
outward movements at the crest of the wall during construction normalized with respect to
current wall height. The data shows that out-of-alignment movements do not increase
monotonically with increasing wall height during construction. Rather, crest movements are
largely controlled by lower portions of the wall. Bulging of the wall structure over the lower half
of the wall height at the California project is consistent with the trend of post-construction
movement observed for the Algonquin wall (Figure 21). The development of outward
movements will also be influenced by quality of construction, compaction-induced stresses
immediately adjacent to the back of the facing units, block interface shear stiffness and lateral
earth pressures developed under soil self-weight. The maximum crest movement as a result of
construction activity was approximately 1.3% of the height of the wall and it is reasonable to
assume that conditions of active earth pressure were developed behind the structure. Because of
vertical out-of-alignments as great as 130mm after the wall had reached a height of 10.4m, the
wall was not constructed to its design height of 14.4m over the entire length of the wall as
originally planned in order to minimize further movement into the excavation.
The data contained in Figure 27 may be useful to estimate how much pre-batter should be
constructed into a reinforced segmental retaining wall face in order to ensure that the
end-of-construction wall batter is not less than the target wall value. Based on the limited data
available, a value of 1.0% of the height of the wall for structures 8m or less and 1.5% of the height
of the wall for structures greater than 8m in height may be reasonable construction guidelines.

Fig. 26 Facing detail (California wall).

These recommendations are very preliminary and assume that structures are built with the same
reinforcement density and length, and the same routine standard of care that was employed at
the projects cited. The limits include an additional margin for post-construction creep equal to
that recorded for the Algonquin wall (0.25%) after 109 days. The data in Figure 27 for the
California wall are in close agreement with outward movements predicted by extrapolating an
empirically established chart contained in the FHWA guidelines (Christopher et al. 1989). The
recommendations contained in the FHWA guidelines were based on finite element modelling
and measurements recorded at the end of construction for geosynthetic reinforced structures
built with other types of facing treatments (Mitchell and Christopher 1990).
A detailed description and analysis of the performance of the California wall can be found in the
paper by Bathurst et al. (1995).
10.3 University of Wisconsin-Platteville (UWP) Wall
An instrumented segmental wall 3.5m high and 37m long was constructed at the University of
Wisconsin-Platteville in 1993. Preliminary results of this structure have been reported by Wetzel
et al. (1995). The wall is illustrated in Figure 28. The wall was divided into three sections 7.62m
in length and each section was reinforced with a different geosynthetic (HDPE geogrid, woven
polyester geogrid and a woven polyester geotextile). The wall was designed according to NCMA
guidelines with the long-term design strengths of the materials taken as 15.3, 15.3 and
13.1kN/m, respectively. The wall units were seated on a compacted granular footing. The

preliminary construction guideline

2.0
H
UWP wall
horizontal deformation (H/H %)

California wall
1.5
H 1.26
1.0
0.5
0.0
0 2 4 6 8 10 12 14
wall height during construction (m)
Fig. 27 Maximum normalized horizontal wall movement with increasing wall height during
construction.
granular footing and backfill comprised a washed crushed limestone with a top size of 19mm
and a peak friction angle of 38 degrees.
The vertical out-of-alignment of the wall due to construction activities is estimated to be about
8mm at the end of construction. Hence, construction-induced outward movements are about
0.2% of the height of the wall with essentially all construction-induced movement occurring
within the lower half of the wall height. As in the earlier cited case studies, bulging of the lower
half of the wall occurred resulting in larger relative outward movements at intermediate stages
of wall construction.
A load cell was incorporated into the facing column at 0.75m above the base of the wall. Vertical
load transmitted across the interface between units was monitored during construction and for
a period of six months thereafter. Measured normal force readings reported by Wetzel et al.
(1995) are compared against: the weight of the facing column above the point of measurement;
predicted load based on the hinge height concept; and column weight plus down drag forces
using Coulomb earth pressure theory (Figure 29). The plot shows that interface normal loads
during construction and at the end of construction are significantly underestimated by neglecting
facing-soil interface shear. However, with time there was a reduction in normal load and
post-construction normal forces are seen to approach values predicted by neglecting down drag
forces. In fact, the lowermost curve in Figure 29 (hinge height approach) provides a reasonable
lower bound on all measured values reported by Wetzel et al. (1995). The initially high wall
friction forces inferred from the field data may be due to relative downward movement of the
reinforced soil mass behind the facing units. The maximum vertical settlement of the foundation
since start of construction was 22mm. This large movement compared to the Algonquin wall,
for which down drag forces appeared to be non-existent, may explain why wall friction was
mobilized in the UWP wall. The influence of relative downward movements of the backfill on
connection loads for rigid facing systems has been clearly demonstrated by Bathurst (1991,

TOPSOIL AND
IMPERVIOUS FILL
wall batter
= 7 degrees GEOTEXTILE SEPARATOR
CRUSHED ROCK
BACKFILL
3.05 m
GEOSYNTHETIC REINFORCEMENT
PLACED IN 5 LAYERS
GRADE
PERFORATED
DRAINAGE PIPE
2.12 m
CRUSHED
ROCK BASE
Fig. 28 University of Wisconsin-Platteville segmental retaining wall (after Wetzel et al.

1995).
1993) and Bathurst and Benjamin (1990). In rigid panel facing systems the maximum load in
the reinforcement was observed to occur at the connections under working load conditions. The
data reported here indicates that an advantage of the relative flexibility of modular concrete
facing systems over rigid panel systems is that modular column faces will relieve high
connection stresses with time. Taken together, the data from the Algonquin wall and UWP wall
support the conservative NCMA approach which is to neglect any down drag force in the
calculation of interface shear and connection capacity for structures built on competent stiff
foundations.
11 SEISMIC DESIGN
11.1 Performance
The discrete nature of the facing column in segmental retaining wall design has raised questions
with respect to stability under seismic loading (Allen 1993). Quantitative data on seismic
resistance is lacking although visual inspection of a 5.5m high segmental retaining wall after the
Loma Prieta earthquake in California in 1989 did not reveal any signs of structural damage to
the facing column nor was there visual evidence of movement (Eliahu and Watt 1991, Collin et
al. 1992). Based on historical data for the California west coast and the location of the site with
respect to the epicenter of the earthquake, a reasonable upper limit on horizontal and vertical
ground acceleration at the above site is 0.1g and .07g respectively. A more recent survey by

N = W +Ka sin() ( z2 +z q)
Ka = f(, , =2/3) curve [1]
Ka = f(, , =) curve [2]
40
==7 end of
construction
z
3.5m
30 W Pa
normal force N (kN/m)
measured
range
after
construction
20 N
+12 months
hinge height
10
[1]
[2]
N=column weight W
0
0 1 2 3 4
height of facing column z (m)
Fig. 29 Comparison of normal force measured within column facing with predicted values
(UWP wall).
Sandri (1994) of reinforced segmental retaining walls greater than 4.5m in height in the Los
Angeles area after the recent Northridge earthquake in January 1994 showed no evidence of
visual damage to 9 of 11 structures located within 23 to 113km of the earthquake epicenter. Two
structures showed tension cracks behind the reinforced soil mass that were clearly attributable
to the results of seismic loading. Nevertheless, the facings of these structures remained intact
with no visible evidence of disturbance. Based on seismic data for the Los Angeles area the peak
horizontal ground acceleration for the geosynthetic reinforced segmental retaining wall (6.5m
high) closest to the epicenter was estimated to be as great as 0.5g.
11.2 Design
At the present time there are no specific guidelines in North America for design of segmental
retaining walls subject to seismic loadings. However, current limit-equilibrium based
(Coulomb) methods are easily extended to incorporate the additional inertial forces assumed to
act during ground acceleration (i.e. Mononobe-Okabe approach) (Okabe 1926, Seed and
Whitman 1970). This general approach has been adopted in a modified form in the FHWA
guidelines (Christopher et al. 1989). Potential failure mechanisms for segmental retaining walls
under seismic loading are the same as those identified in Figure 5. However, owing to the short
duration of ground excitation and the expectation that peak horizontal and vertical ground
accelerations for a gravity structure and retained soil zone will be out-of-phase during a seismic
event, factors of safety employed in seismic design are typically taken as 75% of the values used

+khW +kvW
W
H PAE
khWw
Ww(1kv)
Fig. 30 Forces and geometry used in pseudo-static seismic analysis.
in static analyses (Christopher et al. 1989). In addition, recent work on the cyclic response of
HDPE and PET geogrids shows that these materials are stiffer at rates of loading anticipated
under seismic loading (Bathurst and Cai 1994a). The implications of pseudo-static
limit-equilibrium analyses to the stability of reinforced segmental retaining walls is currently
under study at RMC and a number of analytical results from this work are described below.
11.3 Earth forces under seismic loading
The Mononobe-Okabe (MO) method to calculate active earth forces acting on a planar surface
that is inclined into the retained soil mass can be referenced to Figure 30. Quantities kh and and
kv are horizontal and vertical ground acceleration ratios expressed as fractions of the
gravitational constant. The calculation of dynamic earth thrust PAE for a homogeneous soil is
calculated as:
P AE = 1 (1 k v)K AEH 2 (13)

2
where is the unit weight of the retained soil and H is the height of the wall. The dynamic earth
pressure coefficient KAE can be calculated as follows:
cos 2( + )cos() cos 2() cos( + )

K AE = (14)
[1 + sin(+) sin ()
cos(+) cos(+)
] 2
where is the angle of internal friction of the soil; is the total wall inclination (from vertical);
is the mobilized interface friction angle assumed to act at the back of the wall; is the backslope
angle (from horizontal); and is the seismic inertia angle given by:
= tan 1 1 k k
h
v
(15)

The seismic inertia angle represents the angle through which the vectorial resultant of the gravity
force and the inertial forces (both horizontal and vertical) is rotated from vertical. Equations (13)
through (15) are an exact analytical solution to the classical Coulomb wedge problem that is
modified to include the inertial forces khW and kvW.
A distinguishing feature of the geometry of segmental retaining wall surfaces is that the surface
against which active forces are assumed to act is oriented at > 0 from the vertical. Hence wall
surfaces are rotated in the opposite direction to that of most conventional gravity wall structures.
Typically, this angle varies from 3 to 15 degrees depending on the setback of the stacked modular
units. The influence of horizontal acceleration kh, and wall inclination angle on dynamic active
earth forces is illustrated in Figure 31. The curves illustrate that the effect of positive wall batter
(according to our terminology) is to reduce dynamic earth pressures to levels less than those
developed against conventional gravity wall structures of the same height and retaining the same
frictional soil. The example data in the figure shows that by increasing the wall batter from 0 to
15 degrees the dynamic pressure is reduced by 25% during a 0.2g event.
11.4 Results of stability analyses
The active earth force calculated according to equation (13) replaces the active earth pressure
in the external and internal stability calculations described earlier for static analyses. However,
interesting limits emerge when the pseudo-static approach is extended to factor of safety
calculations for base sliding. Figure 32 shows the value of static FOS against base sliding
required to satisfy a minimum dynamic FOS against base sliding of 1.125 for a range of ground
accelerations. The value of 1.125 is based on 75% of the minimum recommended static value
of 1.5 that is specified in FHWA (Christopher et al. 1989) and NCMA guidelines. The figure
demonstrates that a structure that just satisfies a minimum factor of safety against base sliding
of 1.5 will exceed typical minimum dynamic FOS values for horizontal ground accelerations in
excess of 0.06. Sliding analyses shows that the required width of the reinforced soil mass may
become excessive in order to achieve an acceptable margin of safety under seismic loading. It
should be noted that the type of block sliding analysis reported here is independent of the type
of gravity wall structure.
Figure 33 shows the ratio of dynamic to static interface shear force calculated for a range of
ground acceleration component values. The data show that the increase in interface shear is
relatively insensitive to the magnitude of vertical acceleration for modest ground acceleration
values. However, the effect of even a modest horizontal ground acceleration is dramatic.
Consequently, the use of modular block systems with high shear capacity in seismic areas cannot
be over-emphasized. Many modular block systems have concrete keys which are capable of
developing high shear resistance. Based on the analytical results presented here, these systems
are likely the preferred choice in areas where seismic loads are a concern in design.
The prediction of modest ground accelerations to initiate mass sliding is consistent with the
performance of the two walls reported to have developed tension cracks at the back of the
reinforced zone during the recent Northridge earthquake (Sandri 1994).
The very low values of ground acceleration required to initiate base sliding was originally
pointed out by Richards and Elms (1979) who noted that there is very little margin of safety
against base sliding of gravity structures designed to meet conventional minimum static factors
of safety. In the same paper, Richard and Elms have proposed a limited displacement method
as an alternative to pseudo-static approaches of the type proposed in the FHWA guidelines.
Nevertheless, translational body sliding as predicted in Figure 32 may not be catastrophic as
reported by Sandri. The horizontal stability of segmental retaining wall structures may be
improved by embedding the facing column and using facing units with high interface shear
capacity or simply designing for a higher static factor of safety against sliding (i.e. increasing

1.2
=35,=2/3, =0, kv=0 15
+
1.1
1.0 KAE 10
0.9 5
0
0.8
conventional walls +5
0.7 +10
+15
0.6
KAE
0.5
0.4
0.3
0.2
segmental retaining walls
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5
kh
Fig. 31 Influence of horizontal ground acceleration and wall inclination on dynamic earth
pressure coefficient.
the length of the reinforced soil mass).

A closed-form solution for the orientation of the critical planar surface AE has been reported
by Zarrabi (1979). The influence of horizontal ground acceleration on inclination angle is shown
in Figure 34. The implication to internal stability design is that the length of reinforced layers,
particularly near the top of the reinforced soil zone will have to be extended in order to provide
a sufficient anchorage length under seismic loading (curves for =2/3). The implication to
external stability analysis is that the depth of the failure zone behind the reinforced soil zone
increases with increasing horizontal ground acceleration (curves for =). The use of modified
MO analyses for the design and analysis of segmental retaining wall structures is explored in
detail in a forthcoming paper by Bathurst and Cai (1994b). Research is currently underway at
RMC to investigate the validity of modified M-O methods for stability analysis of geosynthetic
reinforced soil walls using 1/6 scale models.

12
L kh/2
11 kv=2kh/3
khW PAEcos() kh/4
10 H
9 W(1kv) 0
8
R=W(1kv)tan
7
FOSsta
4
=0
3 = 0.6
= 35o
2 =
1.5 =0
1
L = 0.5H
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
0.06 kh
Fig. 32 Influence of ground acceleration on minimum static FOS against base sliding re-
quired to satisfy a minimum dynamic FOS against base sliding of 1.125.
12 CONSTRUCTION AND DURABILITY ISSUES
12.1 Drainage
Segmental retaining walls should be constructed with provision for good drainage. A drainage
layer comprising a coarse single size stone is routinely recommended behind the modular units
(Figure 3). Because of the shape of the modular units and small gaps between adjoining units
a geotextile is recommended to prevent loss of materials through the facing. In addition, a
geotextile may be required to act as a separator between the reinforced soil zone and the drainage
column behind the facing units. The drainage column should be integrated with a drainage layer
at the base of the structure. A gravity flow geotextile-wrapped pipe connected to outlets that
direct water away from the foundation should be used. The use of drainage swales behind the
wall and appropriate surface grading behind the wall crest should always be used to ensure that
surface water is prevented from infiltrating the reinforced soil mass.
12.2 Segmental units
Modular concrete units have been observed to crack in segmental retaining wall structures.
Some potential sources of cracking are illustrated in Figure 35.Vertical cracking may occur due
to tensile failure of the concrete or diagonal cracks may be generated due to shear failure,
particularly at the corners of modular units. The source of cracking may be due to differential
settlement of the foundation and/or construction-induced cracking as described by Anderson

10
LW 0.2
9 kv=0
kv=2kh/3
z
8 PAEcos()
khWw
0.3
7 Ww(1kv)
S
6
0.4
rs 5 S dyn
rs =
S sta
4
0.6
3
0.8
2
z/H
1
=35o =2/3 Sv/H=0.125
=0 =0 LW/H= 0.1
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
kh
Fig. 33 Influence of ground acceleration on interface shear force.
(1993). Potential sources of construction-induced cracking are: horizontal surface irregularities

generated during casting; inadequate cleaning of the top surface of the units; and overfilling of
hollow units. These problems are overcome by strict attention to quality control during casting
of the units and field construction. The importance of achieving a level foundation pad cannot
be over-emphasized since any irregularities in the base course of modular block units will be
carried up the height of the wall. For large wet-cast units, the use of a concrete footing may be
desirable to ensure that differential settlements are minimized and the potential for large cracks
avoided. Misalignment of modular units resulting in point contacts also has implications to
available interface shear and connection capacity particularly for systems in which resistance to
interblock sliding and connection capacity is developed through interface friction. Pinching of
geosynthetic layers at point contacts and block edges can also be expected to reduce connection
capacity. The authors experience has been that polyester geogrid materials have the advantage
of being better able to conform to the interface geometry between modular unit courses.
However, stiffer and generally thicker HDPE geogrids may be less susceptible to micro-damage
caused by abrasion of the reinforcement between rough concrete surfaces. A strategy
investigated by the authors to reduce abrasion has been the use of layers of nonwoven geotextile
to cushion the geogrid reinforcement inclusion. In one project the use of a geonet to perform the
same function has been reported (Anderson et al. 1991). However, the presence of a cushion
material can be expected to influence interface sliding capacity and connection strength and
therefore the cushion material must be included in any large-scale laboratory interface shear
testing.
In many cases the hairline cracks caused by uneven loading are not easily visible because of the
broken textured appearance of the exposed face of many of the masonry units available on the

90
=35o, =0, kv=0
80 = (external) static
=2/3 (internal)
70 A
A (degrees) 60
50
40
45o
30 40o
35o
20
30o
10
20o 25o
0
0.0 0.1 0.2 0.3 0.4 0.5
kh
Fig. 34 Influence of horizontal ground acceleration and interface friction on orientation of
internal and external failure planes.
market. While the appearance of any cracking may be undesirable from an aesthetic point of
view, cracking may be considered desirable from a structural point of view since they are the
result of stress redistribution through the facing column. Long-term monitoring of foundation
settlement-induced cracks in the masonry units of a structure reported by Anderson (1993)
showed that initial cracks did not grow in width with time even though differential settlements
generated at the foundation elevation were as great as 300mm over the 158m length of wall. In
another project, a segmental wall constructed with hollow granular infilled units recorded
differential settlements of 200mm over a wall length of 10m with only minor cracking (Walls
1994). The distortional settlements in this last project were about 2%. Based on this experience,
a preliminary upper limit on distortional settlements for segmental retaining walls with 300mm
high by 300mm wide units (face dimensions) is 2% of the wall length. However, it is important
to note that structures that have performed satisfactorily with respect to cracking after large
differential settlements were constructed with units that have relatively thick concrete shells or
were solid units. There are a number of hollow units on the market with relatively thin concrete
shells. Cracking of these units has resulted in loss of the granular infill.
Durability of masonry concrete units under freeze-thaw cycles may be a concern in some
locations. The potential for masonry concrete degradation may be increased in the presence of
deicing chemicals from snow clearing operations for walls below or otherwise in proximity to
roadways. Waterproof coatings have been proposed for masonry units in these environments.
ASTM has recently published a standard for freeze-thaw testing of masonry concrete units
(ASTM C 1262). Units may be tested either in water or in a 3% saline solution depending on the
intended use of the units in actual service. Specimens are tested after each 8 to 12 freeze-thaw
cycles and the procedure repeated until the loss in weight exceeds 10% of the initial saturated
weight of the specimens. The NCMA is currently developing specifications for masonry

concrete blocks in segmental retaining wall applications that are based on the results of the
ASTM C 1262 method of test.
Specifications for the coping required at the top of concrete retaining wall structures to prevent
water penetration can be found in AASHTO guidelines (1992)
13 CONCLUDING REMARKS
The paper has focused on recent developments in the design, construction, performance and
specification of geosynthetic reinforced soil retaining walls that employ modular concrete units
as the wall facing system. The recommendations for routine structures contained in this paper
have been adopted by the NCMA. The design strategies reported here are generic in nature and
consider all potential modes of failure for these systems. Limit-equilibrium based methods of
analyses together with Coulomb active earth pressure theory are key features of stability
calculations and are consistent with the conventional approach used by geotechnical engineers
to design retaining wall structures in North America. The paper extends the general approach
used in earlier FHWA and AASHTO guidelines to examine facing instability modes of failure
crack
a) differential settlement at levelling pad
crack
b) misalignment of modular units or uneven seating

on levelling pad
crack
c) uneven modular unit dimensions

crack
d) discontinuous reinforcement layer inclusion
Fig. 35 Sources of modular block cracking.

not considered in older design methodologies. Facing stability calculations related to facing
connection performance and interface shear allow the designer to quantify performance
differences between nominally identical geosynthetic reinforced soil retaining wall systems built
with different modular facing units.
Experience with the design of structures that are fully compliant with NCMA guidelines has
shown that the combination of hinge height concept, interface shear capacity and facing
connection requirements, controls vertical reinforcement spacing and results in designs with
multiple layers of relatively low strength reinforcement as opposed to a lesser number of stronger
layers. This result is desirable from the point of view of creating a composite facing-reinforced
soil mass with redundant reinforcement elements.
To date, the majority of structures have been constructed with geogrid reinforcement materials.
However, the results of connection tests carried out by the authors using woven geotextiles shows
that their connection strength is comparable to that of geogrid materials used in similar
applications. The generally cheaper price of geotextiles will undoubtedly lead to their more
frequent use as the reinforcement material in future soil reinforced segmental retaining wall
structures.
14 ACKNOWLEDGEMENTS
The authors would like to acknowledge the contribution of Z. Cai, P. Clarabut and K. Labinaz
(Geotechnical Research Group at RMC) who assisted with the preparation of this paper and
carried out much of the research reported here. The authors would also like to thank K.E. Buttry,
E.S. McCullough and R.A.Wetzel for permission to report unpublished connection strength data
and the results of the trial walls at the University of Wisconsin-Platteville. The co-operation of
B.R. Christopher and C. Bonczkiewicz at STS Consultants in supplying field data from the
Algonquin wall is also gratefully acknowledged. D. Sandri of Mirafi/Nicolon gave permission
to report the results of the California wall project and supplied unpublished data on the
performance of segmental retaining walls after the Northridge earthquake. J. Walls of InterSol
Engineering in Milton, Ontario provided useful data on the performance of segmental walls after
large settlements. The advice and support of R. Thomas of the National Concrete Masonry
Association is also greatly appreciated. Finally, the senior author would like to acknowledge the
funding provided by the Department of National Defence (Canada) in support of
geosynthetics-related research at RMC over a period of 15 years.
REFERENCES
AASHTO/AGC/ARTBA (1990a) TASK FORCE 27 Joint Committee, In situ Soil Improvement

Techniques: Guidelines for the Design of Mechanically Stabilized Earth Walls: 418.
AASHTO/AGC/ARTBA (1990b) TASK FORCE 27 Joint Committee, In situ Soil Improvement
Techniques: Design Guidelines for Use of Extensible Reinforcements (Geosynthetic) for
Mechanically Stabilized Earth Walls in Permanent Applications: 2838.
ASTM C 55 Standard Specification for Concrete Building Brick, American Society for Testing
and Materials, Philadelphia, PA, USA.
ASTM C 90 Standard Specification for Load-Bearing Concrete Masonry Units, American
Society for Testing and Materials, Philadelphia, PA, USA.
ASTM C 140 Standard Methods of Sampling and Testing Concrete Masonry Units, American
Society for Testing and Materials, Philadelphia, PA, USA.

ASTM C 1262 Standard Test Method for Evaluating the Freeze-thaw Durability of
Manufactured Concrete Masonry Units and Related Concrete Units, American Society for
Testing and Materials, Philadelphia, PA, USA.
ASTM D 4595 Standard Test Method for Tensile Properties of Geotextiles by the Wide-Width
Strip Method, American Society for Testing and Materials, Philadelphia, PA, USA.
Allen, T.M. (1993) Issues Regarding Design and Specification of Segmental Block-Faced
Geosynthetic Walls, Transportation Research Record, 1414:611.
Anderson, R.B. (1993) Construction Considerations for Geogrid-Segmental Block
Mechanically Stabilized Earth Retaining Walls, Transportation Research Record,
1414:1215.
Anderson, R.B., Boyd, F.N. and Shaw, L. (1991) Modular Block Faced Polymer Geogrid
Reinforced Soil Walls U.S. Postal Service Combined Carrier Facility, Proceedings of
Geosynthetics91, Atlanta, GA, USA, 2:889902.
Bathurst, R.J. (1991) Case Study of a Monitored Propped Panel Wall, Proceedings of the Interna-
tional Symposium on Geosynthetic-Reinforced Soil Retaining Walls, Denver, Colorado, USA,
159166 (published by A.A.Balkema).
Bathurst, R.J. (1993) Investigation of Footing Restraint on Stability of Large-scale Reinforced
Soil Wall Tests, 46th Canadian Geotechnical Conference, Regina, Saskatchewan, Canada,
10p.
Bathurst, R.J. and Benjamin, D.J. (1990) Failure of a Geogrid-Reinforced Soil Wall, Transporta-
tion Research Record 1288:109116.
Bathurst, R.J. and Cai, Z. (1994a) In-isolation Cyclic Load-Extension Behavior of Two Geo-
grids, Geosynthetics International, 1:1:317.
Bathurst, R.J. and Cai, Z. (1994b) Pseudo-Static Seismic Analysis of GeosyntheticReinforced
Segmental Retaining Wall Structures, submitted to Geosynthetics International.
Bathurst, R.J. and Simac, M.R. (1991) Review of Three Instrumented Geogrid Reinforced Soil
Retaining Walls, Geosynthetics: Design and Performance, Vancouver Geotechnical Society
6th Annual Symposium, Vancouver, BC, Canada, 24 May, 15p.
Bathurst, R.J. and Simac, M.R. (1993a) Laboratory Testing of Modular Unit-Geogrid Facing
Connections, STP 1190 Geosynthetic Soil Reinforcement Testing Procedures (S.C.J. Cheng
editor), American Society for Testing and Materials (Special Technical Publication), 3248.
Bathurst, R.J. and Simac, M.R. (1993b) Two Computer Programs for the Design and Analysis
of Geosynthetic-Reinforced Soil Retaining Walls, Geotextiles and Geomembranes,
12:381396.
Bathurst, R.J., Simac, M.R. and Berg, R.R. (1993a) Review of the NCMA Segmental Retaining
Wall Design Manual for Geosynthetic-Reinforced Structures, Transportation Research
Record 1414:1625.
Bathurst, R.J., Simac, M.R., Christopher, B.R. and Bonczkiewicz, C. (1993b) A Database of Re-
sults from a Geosynthetic Reinforced Modular Block Soil Retaining Wall, Proceedings of Soil
Reinforcement: Full Scale Experiments of the 80s, ISSMFE/ENPC, Paris, France, 1819 No-
vember, 341365.
Bathurst, R.J., Simac, M.R., and Sandri, D. (1995) Lessons Learned from the Construction Per-
formance of a 14m High Segmental Retaining Wall, Proceedings of Geosynthetics: Lessons
Learned from Failures, Nashville, Tennessee, 20 February, (published by IFAI), 15p.

Berg, R.R. (1991) The Technique of Building Highway Retaining Walls, Geotechnical Fabrics
Report, IFAI, St. Paul, MN, USA, 9:5:3843.
Berg, R.R. (1992) Guidelines for Design, Specification, and Contracting of Geosynthetic
Mechanically Stabilized Earth Slopes on Firm Foundations, U.S. Department of
Transportation, Federal Highway Administration, Washington, D.C., USA, 98p.
Berg, R.R., Bonaparte, R., Anderson, P., and Chourey, V.E. (1986) Design, Construction and
Performance of Two Geogrid Reinforced Soil Retaining Walls, Proceedings of the 3rd
International Conference on Geotextiles, Vienna, Austria, 401406.
Bonczkiewicz, C., Christopher, B.R. and Simac, M. (1991) Load Distribution in Geogrids with
Low Junction Efficiency Proceedings of Geosynthetics91, Atlanta, GA, USA 2:643652.
Bowden, R.K. (1991) Royal Roads Military College Geogrid Wall Design, Proceedings of Geo-
synthetics: Design and Performance, Vancouver Geotechnical Society 6th Annual Sympo-
sium, Vancouver, BC, Canada, 24 May, 13p.
Buttry, K.E., McCullough, E.S. and Wetzel, R.A. (1993) Laboratory Evaluation of Connection
Strength to Geogrid Segmental Concrete Units, Transportation Research Record 1414:2631.
Canadian Foundation Engineering Manual (1992) 3rd Edition, Chapter 30 Geosynthetics: 512p.
Christopher, B.R. (1993) Deformation Response and Wall Stiffness in Relation to Reinforced
Soil Wall Design, Doctoral Thesis submitted to Purdue University, West Lafayette, Indiana,
USA, 354p.
Christopher, B.R., Gill, S.A., Giroud, JP., Juran, I., Schlosser, F., Mitchell, J.K. and Dunnicliff,
J. (1989) Reinforced Soil Structures, Volume I. Design and Construction Guidelines, Report
No. FHWA-RD-89-043, 287p.
Collin, J.G., ChoueryCurtis, V.E. and Berg, R.R. (1992) Field observations of reinforced soil
structures under seismic loading, Proceedings of the International Symposium on Earth
Reinforcement, Fukuoka, Japan, 1:223228.
Crowe, R.E., Bathurst, R.J. and Alston, C. (1989) Design and Construction of a Road Embank-
ment Using Geosynthetics, 42nd Canadian Geotechnical Conference, Winnipeg, Manitoba,
Canada, 266271.
Determination of Shear Strength between Segmental Concrete Units (1993a) NCMA Test
Method SRWU2, National Concrete Masonry Association (NCMA), Herndon, VA, USA.
Determination of Connection Strength between Geosynthetics and Segmental Concrete Units
(1993b) NCMA Test Method SRWU1, National Concrete Masonry Association (NCMA),
Herndon, VA, USA.
Eliahu, U. and Watt, S. (1991) Geogrid-Reinforced Wall Withstands Earthquake, Geotechnical
Fabrics Report, IFAI, St. Paul, MN, USA, 9:2:813.
Geotechnical Fabrics Report (1994) A SRW Improves Interstate 94, 12:3:2631.
Geotechnical Research Institute Standards, Drexel University, Philadelphia PA, USA.
Hill, J.J. and Berg, R.R. (1993) Use of Segmental Wall System by Minnesota Department of
Transportation, Transportation Research Record, 1414:15.
Kemp, S., Martin, J.S. and Stadler, A.T. (1993) The Design and Construction of
Geogrid-Reinforced Retaining Walls at the South Carolina Port Authoritys Wando Terminal,
Proceedings of Geosynthetics93, Vancouver, BC, Canada, 1:153180.
Mitchell, J.K. and Christopher, B.R. (1990) North American Practice in Reinforced Soil
Systems, Design and Performance of Earth Retaining Structures, ASCE Geotechnical Special
Publication No. 25, (P.C. Lambe and L.A. Hansen eds) 322346.

Okabe, S. (1926) General Theory of Earth Pressure, Journal of the Japanese Society of Civil En-
gineers, 12:1.
Richards, R. and Elms, D.G. (1979) Seismic Behavior of Gravity Retaining Walls, Journal of
the Geotechnical Engineering Division, ASCE, 105:GT4, 449464.
Sandri, D. (1994) Retaining walls stand up to the Northridge earthquake, Geotechnical Fabrics
Report, IFAI, St. Paul, MN, USA, 12:4:3031 (and personal communication).
Seed, H.B. and Whitman, R.V. (1970) Design of Earth Retaining Structures for Dynamic Loads,
ASCE Specialty Conference: Lateral Stresses in the Ground and Design of Earth Retaining
Structures, 103147.
Simac, M.R., Bathurst, R.J. and Goodrun, R.A. (1991) Design and Analysis of Three Reinforced
Soil Retaining Walls, Proceedings of Geosynthetics91, Atlanta, GA, USA, 2:781789.
Simac, M.R., Bathurst, R.J., Berg, R.R. and Lothspeich, S.E. (1993a) Design Manual for
Segmental Retaining Walls (Modular Concrete Block Retaining Wall Systems) First Edition,
National Concrete Masonry Association (NCMA), Herndon, VA, USA, 250p.
Simac, M.R., Bathurst, R.J. and Berg, R.R. (1993b) New design guidelines for segmental
retaining walls, Geotechnical Fabrics Report, IFAI, St. Paul, MN, USA, 11:5:1426.
Simac, M.R., Christopher, B.R. and Bonczkiewicz, C. (1990) Instrumented Field Performance
of a 6m Geogrid Soil Wall. Proceedings of the 4th Int. Conference on Geotextiles,
Geomembranes and Related Products, The Hague, Netherlands, 1:5359.
Specification for Segmental Retaining Wall Units, TEK 50A (1991) National Concrete Masonry
Association (NCMA), Herndon, VA, USA, 4p.
Standard Specifications for Highway Bridges (1992) 15th ed., American Association of State
Highway and Transportation Officials (AASHTO).
Walls, J. (1994) personal communication.
Wetzel, R.A., Buttry, K.E. and McCullough, E.S. (1995) Preliminary Results from Instrumented
Segmental Retaining Wall, Proceedings of Geosynthetics95, Nashville, TN, USA,
1:133146.
Zarrabi, K. (1979) Sliding of Gravity Retaining Wall During Earthquakes Considering Vertical
Acceleration and Changing Inclination of Failure Surface, Master of Science thesis submitted
to Massachusetts Institute of Technology, Cambridge, MA, USA.

10 1 1 703 8121 PDF

Uploaded by

Copyright:

Available Formats

10 1 1 703 8121 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 1 1 703 8121 PDF

Uploaded by

Copyright:

Available Formats

Keynote Lecture reprint: Proceedings of the Fifth International Conference on

Geotextiles, Geomembranes and Related Products, Singapore, September 1994

Geosynthetic Reinforced Segmental Retaining Wall

Bathurst and Simac 1

2 SEGMENTAL RETAINING WALL UNITS

Bathurst and Simac 2

compacted 1.2m long PET geogrid l

drainage collection pipe

slope 1:6 (typ) excavation limits

total geogrid embedment length (all layers)

Bathurst and Simac 3

Fig. 4 Examples of segmental retaining wall units.

Bathurst and Simac 4

Table 1: Specifications for segmental retaining wall

Minimum Maximum Dimensional

* Minnesota Department of Transport (Hill and Berg

3 ANALYSIS AND DESIGN

Bathurst and Simac 5

5 APPLICATION OF EARTH PRESSURE THEORY

a) base sliding b) overturning c) bearing capacity

d) pullout e) tensile over-stress f) internal sliding

g) connection failure h) shear failure i) toppling

Bathurst and Simac 6

anchorage length for pullout calculations (Le)

Bathurst and Simac 7

1) or Coulomb theory (equation 2) on the magnitude of horizontal earth pressure is illustrated

cos cos 2 cos 2

Bathurst and Simac 8

Bathurst and Simac 9

Ta = LTDS/FSOS > Tmax (7)

RPO = 2 Le Ci v tan (8)

FSPO = RPO/Tmax (9)

Bathurst and Simac 10

He = 2[(WuGu0.5Hu tanib) cos ib]/tan(+ib) (10)

Bathurst and Simac 11

7.1 Interface shear capacity

SSU = asu + n tan su (11)

1 loading frame 2 horizontal actuator

Bathurst and Simac 12

peak interface shear (kN/m)

SSU = 4 + Nn tan 20o

[1] lower bound limit on units with

Fig. 11 Example peak interface shear versus normal load data.

Bathurst and Simac 13

SSU = 21 + Nn tan 37o

7.2 Facing connection performance

Bathurst and Simac 14

1 masonry concrete block 11 spacers

initial 200mm free length

Bathurst and Simac 15

normalized tensile load Tpull /T index

0.4 137 kPa

Fig. 14 Example normalized connection load-displacement curves (Bathurst and Simac

peak tensile load

0.4 pullout rupture

0.2 20mm displacement

Bathurst and Simac 16

8 SUMMARY OF DESIGN CRITERIA

Fig. 16 Summary of connection test data.