0 évaluation0% ont trouvé ce document utile (0 vote)

5 vues45 pagespseudo static analysis for reinforced earth retaining walls

GI-V2-N5-Paper2

© © All Rights Reserved

PDF, TXT ou lisez en ligne sur Scribd

pseudo static analysis for reinforced earth retaining walls

© All Rights Reserved

0 évaluation0% ont trouvé ce document utile (0 vote)

5 vues45 pagesGI-V2-N5-Paper2

pseudo static analysis for reinforced earth retaining walls

© All Rights Reserved

Vous êtes sur la page 1sur 45

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/254560182

ARTICLE in GEOSYNTHETICS INTERNATIONAL JANUARY 1995

Impact Factor: 1.68 DOI: 10.1680/gein.2.0037

CITATIONS

READS

37

336

2 AUTHORS, INCLUDING:

Richard J. Bathurst

Royal Military College of Canada

190 PUBLICATIONS 3,606 CITATIONS

SEE PROFILE

letting you access and read them immediately.

Retrieved on: 11 January 2016

GEOSYNTHETIC-REINFORCED SEGMENTAL

RETAINING WALLS

ABSTRACT: The paper examines seismic stability analyses of geosynthetic-reinforced segmental retaining walls (modular block walls). Stability analyses are developed within the framework of a pseudo-static approach that gives factors of safety

against collapse mechanisms or rupture of component materials. The MononobeOkabe method is used to estimate dynamic earth pressures. Parametric analyses of

forces and factors of safety related to external, internal and facing failure modes for

walls constructed on competent foundations are presented. Shear interfaces between

facing units are considered as possible planes of failure in facing stability analyses. The

potential for local toppling of the facing column is also investigated. The results of analyses demonstrate that there is a limiting value of the horizontal seismic coefficient

above which the margin of safety against base sliding and overturning may be unacceptably low during a seismic event for segmental retaining walls designed to just satisfy

minimum factors of safety under static loading conditions. Pseudo-static seismic analyses of the performance of two geosynthetic-reinforced segmental walls during the

Northridge Earthquake in Los Angeles in 1994 are demonstrated to be consistent with

visual observation of tension cracks in the soil backfill. Limitations of pseudo-static

methods are discussed and recommendations for further research are made.

KEYWORDS: Pseudo-static analysis, Segmental retaining walls, Geosynthetic reinforcement, Seismic analysis, Modular block walls.

AUTHORS: R.J. Bathurst, Professor, and Z. Cai, Research Associate, Department of

Civil Engineering, Royal Military College of Canada, Kingston, Ontario, K7K 5L0,

Canada, Telephone: 1/613-541-6000 Ext. 6479, Telefax: 1/613-541-6599, E-mail:

bathurst@rmc.ca.

PUBLICATION: Geosynthetics International is published by the Industrial Fabrics

Association International, 345 Cedar St., Suite 800, St. Paul, MN 55101, USA,

Telephone: 1/612-222-2508, Telefax: 1/612-222-8215. Geosynthetics International is

registered under ISSN 1072-6349.

DATES: Original manuscript received 1 May 1995, revised manuscript received and

accepted 3 August 1995. Discussion open until 1 May 1996.

REFERENCE: Bathurst, R.J. and Cai, Z., 1995, Pseudo-Static Seismic Analysis of

Geosynthetic-Reinforced Segmental Retaining Walls, Geosynthetics International,

Vol. 2, No. 5, pp. 787-830.

GEOSYNTHETICS INTERNATIONAL

787

INTRODUCTION

The use of segmental retaining walls that include dry-stacked concrete block units

as the facia system together with extensible sheets of polymeric materials (geosynthetics) that internally reinforce the retained soils and anchor the facia has gained wide popularity in North America (Bathurst and Simac 1994). These structures have also been

reported in Europe, Scandinavia and Australia in recent years (Cazzuffi and Rimoldi

1994; Gourc et al. 1990; Knutson 1990; Won 1994). An example of a reinforced segmental retaining wall structure is illustrated in Figure 1. The distinguishing feature of

these structures is the facing column that is constructed using mortarless modular concrete block units that are stacked to form a wall batter into the retained soils (typically

3 to 15_ from vertical). The economic benefits of these systems over conventional reinforced concrete gravity wall structures and mechanically stabilized soil retaining walls

that use inextensible (steel) reinforcement and select backfills have been demonstrated

in several of the references cited in an earlier paper by Bathurst and Simac (1994).

12kPa surcharge

Compacted

drainage fill

Masonry

concrete

facing units

Geogrid type ll

Geogrid type ll

5.5 m

20

Compacted drainage

fill with geotextile

Geogrid type ll

6.1 m

Compacted native soil

for reinforced soil zone

Geogrid type ll

gravity flow to outlet

0.3 m (typical)

Slope 1:6 (typical)

Excavation limits

4.4 m

Figure 1. Typical geosynthetic-reinforced soil segmental retaining wall cross-section

(after Simac et al. 1991).

788

GEOSYNTHETICS INTERNATIONAL

Stability analyses for geosynthetic-reinforced soil walls under static loading conditions (including segmental retaining wall systems) involve separate calculations to establish factors of safety against external modes of failure and internal modes of failure

(Figure 2). External stability calculations consider the reinforced soil zone and the facing column as a monolithic gravity structure. The evaluation of factors of safety against

base sliding, overturning about the toe, and foundation bearing capacity is analytically

identical to that used for conventional gravity structures. Internal stability analyses for

geosynthetic-reinforced soil walls are carried out to ensure that the structural integrity

of the geosynthetic-reinforced soil mass is preserved with respect to reinforcement

over-stressing and pullout of geosynthetic reinforcement layers from the anchorage

zone.

A comprehensive design methodology has been recently proposed by the National

Concrete Masonry Association (NCMA) for the static analysis of segmental retaining

walls (Simac et al. 1993; Bathurst et al. 1993). The NCMA guidelines address potential

(b) overturning

(excessive settlement)

(e) pullout

(bulging)

(toppling)

Figure 2. Modes of failure: external (top row); internal (middle row); and facing (bottom

row) (Simac et al. 1993).

GEOSYNTHETICS INTERNATIONAL

789

failure mechanisms not found in other geosynthetic-reinforced soil wall systems as illustrated in Figures 2f, 2g, 2h and 2i. The dry-stacked (mortarless) concrete blocks are

discrete units that transmit shear through concrete keys, interface friction, mechanical

connectors, or a combination of these methods. The stacked facing units result in potential failure planes through the facing column and this requires that additional stability

calculations be carried out to estimate interface shear forces and to compare these

forces with available shear capacity. In addition, the connection between the reinforcement layers and the facia is typically formed by extending the reinforcing layers along

the interface between facing units to the front of the wall. The connection detail must

also be evaluated for satisfactory design capacity (Bathurst and Simac 1993).

The NCMA method proposes a consistent approach to calculate earth pressures for

both external and internal stability calculations that is based on Coulomb earth pressure

theory. The advantage of Coulomb earth pressure theory over Rankine theory, which

has been adopted in earlier guidelines for geosynthetic-reinforced soil walls (AASHTO

1990, 1992; Christopher et al. (FHWA) 1989), is that the former explicitly accounts for

the influence of wall batter and wall-soil friction on the development of earth pressures

and hence is less conservative. A review of the essential features of the Coulomb earth

pressure approach as it applies to segmental retaining wall structures and comparisons

of the NCMA design methodology with earlier limit-equilibrium methods of design can

be found in the paper by Bathurst et al. (1993).

The scope of the NCMA guidelines is currently restricted to design of routine segmental retaining walls under static loading conditions. Questions related to the performance of the discrete facia system and the connections between the facia units and geosynthetic reinforcement layers during a seismic event have been raised (Allen 1993).

Nevertheless, the satisfactory performance of a number of geosynthetic-reinforced segmental walls during the Loma Prieta Earthquake of 1989 (Eliahu and Watt 1991; Collin

et al. 1992) and the Northridge Earthquake of 1994 in California (Sandri 1994) has been

qualitatively demonstrated.

The present paper investigates the stability of geosynthetic-reinforced segmental retaining walls under dynamic loading due to seismic excitation (earthquake). The study

is restricted to structures built on competent foundations for which foundation collapse

or excessive settlement is not a potential source of instability. A pseudo-static rigid

body approach that uses the well-known Mononobe-Okabe (M-O) method to calculate

dynamic earth forces (Okabe 1924) is outlined in this study. The method is restricted

to a limit-equilibrium approach in which factors of safety against collapse or rupture

mechanisms are calculated. The method developed in the current study should not be

confused with displacement methods which have been used for the design of conventional walls and explicitly incorporate permanent displacement criteria in stability

analyses (Richards and Elms 1979; Whitman 1990). The current approach is a logical

extension of the Coulomb wedge theory adopted by the NCMA guidelines for structures

under static loading. The M-O method has been used to calculate earth forces for seismic stability analyses of conventional gravity wall structures (Seed and Whitman 1970;

Richards and Elms 1979). Some of the questions raised in earlier papers concerned with

the implementation of pseudo-static methods for conventional gravity retaining walls

must also be addressed for the special class of structure that is the focus of this paper.

790

GEOSYNTHETICS INTERNATIONAL

2.1

The Mononobe-Okabe method is used to calculate dynamic active earth forces acting

on a planar surface that is inclined at an angle, , into an unsaturated, homogeneous,

cohesionless soil mass (Figure 3). In Figure 3, W refers to the static weight of the active

wedge of soil acting behind the wall and Ww refers to the static weight of the facing column. Quantities kh and kv are horizontal and vertical seismic coefficients, respectively,

expressed as fractions of the gravitational constant, g . In the current study, horizontal

inertial forces are assumed to act outwards (+kh ) to be consistent with the notion of active earth pressure conditions. The convention adopted in this paper is that a positive

vertical seismic coefficient, +kv , corresponds to a seismic inertial force that acts downward and a negative seismic coefficient, - kv , corresponds to a seismic inertial force that

acts upward. The total dynamic active earth force, PAE , imparted by the backfill soil

is calculated as (Seed and Whitman 1970):

P AE = 1 (1 kv)K AEH 2

2

(1)

where: = unit weight of the soil; and H = height of the wall. The dynamic earth pressure

coefficient, KAE , can be calculated as follows:

+kh W

PAE

H

kh Ww

(1 kv )W

(1 kv )Ww

AE

GEOSYNTHETICS INTERNATIONAL

791

K AE =

cos 2( + )

sin(+)sin()

cos(+)cos(+)

(2)

where: = peak soil friction angle; = total wall inclination (positive in a clockwise

direction from the vertical); = mobilized interface friction angle assumed to act at the

back of the wall; = backslope angle (from horizontal); and = seismic inertia angle

given by:

= tan 1

1 k k

h

(3)

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 1 to 3 are an exact analytical solution to the classical Coulomb

wedge problem that is modified to include the inertial forces kh W and kv W. Examination

of the trigonometric terms in Equation 2 shows that solutions are only possible for

. Hence, the maximum value of horizontal seismic coefficient for which there

are solutions to Equation 2 is restricted to kh (1kv )tan(-- ) .

Equations 1 and 2 can be modified to account for additional loads due to a uniformly

distributed surcharge acting behind the wall (Okabe 1924; Motta 1994). However, the

influence of any surface distributed surcharge loading on the stability of segmental retaining walls is not investigated in the current study. A closed-form solution for the calculation of dynamic earth force for c-- soils in retaining wall design is reported by Prakash (1981); however, this solution is restricted to the special case of = 0 and kv = 0.

For more complicated wall geometries and cases with surface loadings, trial single failure plane geometries, or two-part wedge failure plane geometries, can be evaluated to

find the critical geometry giving the maximum value of PAE . However, while these solutions are more general, they do not offer the designer the convenience of the closedform solutions adopted in the current study.

In the discussions to follow, it is convenient to decompose the total dynamic active

earth force, PAE , calculated according to Equations 1 and 2 into two components representing the static earth force component, PA , and the incremental dynamic earth force

due to inertial seismic effects, Pdyn (Seed and Whitman 1970). Hence:

P AE = P A + P dyn

(4)

(1 k v)KAE = K A + K dyn

(5)

or

where: KA = static active earth pressure coefficient; and Kdyn = incremental dynamic

active earth pressure coefficient. For brevity in the following text, the quantity PAE will

be called the dynamic earth force.

792

GEOSYNTHETICS INTERNATIONAL

2.2

The position of the dynamic earth force, PAE , acting against gravity retaining walls

has been shown to be variable and to depend on the magnitude of ground acceleration.

A general range for the point of application of the dynamic force increment (Pdyn in

Equation 4) has been reported to be H=0.4H to 0.7H above the toe of the wall (Seed

and Whitman 1970). ( is the distance of the dynamic load increment above the toe of

the wall normalized with respect to wall height, H.) Seed and Whitman suggest that a

value of = 0.6 is reasonable for practical design purposes and this value is consistent

with the results of small-scale shake table tests reported by Ishibashi and Fang (1987).

Based on experience with conventional gravity retaining walls the earth pressure distributions illustrated in Figure 4 for = 0.6 are assumed to be applicable to segmental

retaining wall structures in this paper. The parameter m in Figure 4 denotes the normal-

0.8Kdyn H

0.8Kdyn H

Pdyn

PA

PAE = PA + Pdyn

H

mH

H/3

KA H

(a) static component

0.2Kdyn H

(b) dynamic increment

(KA +0.2Kdyn )H

(c) dynamic (total) pressure

distribution

(Note: =0.6 .)

GEOSYNTHETICS INTERNATIONAL

793

ized point of application of the dynamic earth force and is limited to the range

1/3m0.6. This range compares favorably to measured values of m ranging from 0.3

to 0.5 reported by Ichihara and Matsuzawa (1973) from the results of shake table tests

on small-scale gravity wall models. The distributions for static and dynamic increment

of active earth pressure illustrated in Figure 4 are also identical to those recommended

for the design of flexible anchored sheet pile walls (Ebling and Morrison 1993).

Finally, to simplify all stability calculations in this paper, and to be consistent with

the convention adopted in the NCMA guidelines, only the horizontal component of PAE

is used in stability calculations, i.e. PAE cos(-- ). This assumption results in a conservative (i.e. safe) design by ignoring the stabilizing benefit of the vertical component of

PAE .

2.3

Closed-form solutions for the orientation of the critical planar surface from the horizontal, AE , have been reported by Okabe (1924) and Zarrabi (1979). These solutions

are rewritten here as:

a AE = + tan1

A E + D

AE

AE

AE

(6)

where

A AE = tan( )

D AE = A AEA AE + B AE B AECAE + 1

E AE = 1 + C AE A AE + B AE

B AE = 1tan( + )

C AE = tan( + )

Equation 6 can be used to calculate the orientation of the assumed active failure plane

within the reinforced soil mass and in the retained soil.

2.4

In the theoretical developments and parametric analyses to follow, the friction angle,

, of the cohesionless backfill soils is assumed to be the peak value determined from

conventional laboratory practice and its magnitude is assumed to be unchanged under

seismic excitation due to an earthquake. The choice of peak friction angle for seismic

794

GEOSYNTHETICS INTERNATIONAL

design is consistent with AASHTO (1990, 1992), FHWA (Christopher et al. 1989) and

NCMA (Simac et al. 1993) guidelines for static design of geosynthetic-reinforced soil

walls. Under rapid loading conditions the strength of compacted unsaturated cohesionless backfills can be expected to be at least as great as the static value.

The admissible range of the interface friction angle, , is 0 in Coulomb wedge

analyses. In static stability analyses, is assumed to be equal to 2/3 for internal stability analyses (facing column-reinforced soil interface) and = for external stability

analyses (reinforced soil-retained soil interface). A value of 2/3 has been shown to be

applicable for wall-soil interface friction based on small-scale shake table tests of conventional gravity wall structures (Ishibashi and Fang 1987) and is assumed to be also

applicable for geosynthetic-reinforced retaining wall structures in the current study.

Parametric analyses are restricted to the case to avoid the complication that results from vertical components of earth forces that act upward. The condition is

valid for typical segmental retaining walls since fully-mobilized interface friction

angles, , taken at the facing column-reinforced soil interface and the reinforced soilretained soil interface are generally greater than the wall inclination angle, .

2.4.2 Seismic Coefficients

The general solution to the M-O method of analysis admits both vertical and horizontal components of seismic-induced inertial forces. The range of horizontal seismic coefficients used in the parametric analyses to follow is restricted to kh < 0.5 .

The choice of positive or negative kv values will influence the magnitude of dynamic

earth forces calculated using Equations 1 and 2. In addition, the resistance terms in factor of safety expressions introduced later in the paper will be influenced by the choice

of sign for kv . An implicit assumption in many of the papers on pseudo-static design

of conventional gravity wall structures reviewed by the writers is that the vertical component of seismic body forces acts upward. However, the designer must evaluate both

positive and negative values of kv to ensure that the most critical condition is considered

in dynamic stability analyses if non-zero values of kv are assumed to apply. For example,

Fang and Chen (1995) have demonstrated in a series of example calculations that the

magnitude of PAE may be 12% higher for the case when the vertical seismic force acts

downward (+kv ) compared to the case when it acts upward (-- kv ). Nevertheless, selection

of a non-zero value of kv implies that peak horizontal and vertical accelerations are time

coincident which is an unlikely occurrence in practice. The assumption that peak vertical accelerations do not occur simultaneously with peak horizontal accelerations is

made in the current FHWA guidelines for the seismic design of mechanically stabilized

soil retaining walls (Christopher et al. 1989). Indeed, Seed and Whitman (1970) have

suggested that kv = 0 is a reasonable assumption for practical design of conventional

gravity structures using pseudo-static methods. Wolfe et al. (1978) studied the effect

of combined horizontal and vertical ground acceleration on the seismic stability of reduced-scale model Reinforced Earth walls using shake table tests. They concluded that

the vertical component of the seismic motion may be disregarded in terms of practical

seismic stability design. Their conclusion can also be argued to apply to geosyntheticreinforced segmental walls. Nevertheless, significant vertical accelerations may occur

at sites located at short epicentral distances and engineering judgement must be exer-

GEOSYNTHETICS INTERNATIONAL

795

cised in the selection of vertical and horizontal seismic coefficients to be used in pseudo-static seismic analyses.

However, in order to address specific concerns raised by Allen (1993) related to facing stability of geosynthetic-reinforced segmental retaining walls during a seismic

event that includes vertical ground accelerations, parametric analyses were carried out

in the current study for the range kv = - 2kh /3 to + 2kh /3. The upper limit on the ratio kv

to kh is equal to the calculated ratio of peak vertical ground acceleration to peak horizontal ground acceleration from seismic data recorded in the Los Angeles area

(UCB/EERC, 1994).

In conventional pseudo-static M-O methods of analysis the choice of horizontal seismic coefficient, kh , for design is related to a specified horizontal peak ground acceleration for the site, ah . On the west coast of British Columbia (the most seismically active

area of Canada), the typical maximum design horizontal ground acceleration on rock,

based on a 10% probability of exceedance in 50 years, is ah = 0.32g (CFEM 1992). The

relationship between peak ground acceleration for the site, ah , and a representative value of kh is nevertheless complex and there does not appear to be a general consensus

in the literature on how to relate these parameters. For example, Whitman (1990) reports that values of kh from 0.05 to 0.15 are typical values for the design of conventional

gravity wall structures and these values correspond to 1/3 to 1/2 of the peak acceleration

of the design earthquake. Bonaparte et al. (1986) used kh = 0.85ah /g to generate design

charts for geosynthetic-reinforced slopes under seismic loading using the M-O method

of analysis. However, the results of finite element (FE) modelling of reinforced soil

walls by Segrestin and Bastick (1988), Cai and Bathurst (1995) and limited 1/2 scale

experimental work (Chida et al. 1982) has shown that the average acceleration of the

composite soil mass may be equal to or greater than ah depending on a number of factors

such as: magnitude of peak ground acceleration; predominant modal frequency of

ground motion; duration of motion; height of wall; and stiffness of the composite mass.

Current FHWA guidelines use an equation proposed by Segrestin and Bastick (1988)

that relates kh to ah according to:

k h = (1.45 a hg)(ahg)

(7)

This formula results in kh > ah /g for ah < 0.45g . However, as clearly stated by Segrestin

and Bastick, their equation should be used with caution because it is based on the results

of FE modeling of steel reinforced soil walls up to 10.5 m high that were subjected to

ground motions with a very high predominant frequency of 8 Hz. The results of FE modeling reported by Cai and Bathurst (1995) for a 3.2 m high geosynthetic-reinforced segmental retaining wall with ah = 0.25g and a predominant frequency range of 0.5 to 2

Hz gave a distribution of peak horizontal acceleration through the height of the composite mass and retained soil that was for practical purposes uniform and equal to the base

peak input acceleration. These observations are consistent with the results of Chida et

al. (1982) who constructed 4.4 m high steel reinforced soil wall models and showed that

the average peak horizontal acceleration in the soil behind the walls was equal to the

peak ground acceleration for ground motion frequencies less than 3 Hz. In practice, the

final choice of kh may be based on local experience, and/or prescribed by local building

codes or other regulations.

796

GEOSYNTHETICS INTERNATIONAL

In the current analyses, kh and kv are assumed to be uniform and constant throughout

the facing column, the reinforced soil mass and in the retained soils. This assumption

simplifies the analysis for geosynthetic-reinforced soil walls but may not be true for

walls higher than (say) 7 m, or walls with complex geometries, surface loadings, and/or

difficult foundation conditions. For these structures, and/or structures subjected to high

frequency ground motions, more sophisticated analyses may be warranted.

2.4.3 Other

In order to simplify stability analyses in this study, the facing units and backfill soils

are assumed to have the same unit weight, . This assumption introduces negligible

error in the example calculations to follow. It can be noted that the majority of the segmental retaining wall units on the market today are hollow soil infilled masonry units.

For these systems the facing unit weights and typical backfill soil unit weights are very

similar.

2.5

A distinguishing feature of the geometry of segmental retaining walls is that the surfaces against which active earth pressures are assumed to act are oriented at > 0 from

the vertical (Figure 3). Hence, the wall-soil interface and reinforced soil-retained soil

interface are rotated in the opposite direction to that of many conventional gravity wall

structures. Typically, varies from 3 to 15_ from the vertical (Simac et al. 1993) depending on the setback of the stacked modular units and the initial base unit inclination.

The influence of horizontal seismic coefficient, kh , and wall inclination angle, , on

the dynamic earth force, PAE , is illustrated in Figure 5. The data shows that the effect

of positive wall inclination is to reduce dynamic earth forces to levels less than those

developed against conventional gravity wall structures of the same height and retaining

the same frictional soil.

The backslope angle, , also influences the magnitude of dynamic earth force. As the

backslope angle becomes larger, the magnitude of dynamic earth force increases as illustrated in Figure 6. The effect of increasing positive wall inclination is seen to reduce

the magnitude of dynamic earth force for a given backslope angle.

The influence of wall inclination angle on the orientation of the critical failure surface, AE , through the reinforced soil mass is shown in Figure 7. The figure illustrates

that for a given wall inclination angle, the size of the active soil wedge behind the wall

facing increases with increasing magnitude of the horizontal seismic coefficient, kh . A

similar result has been reported by Vrymoed (1989). The implication of the results

shown in Figure 7 to internal stability design is that the length of reinforcement layers,

particularly those near the top of the reinforced soil zone, may have to be extended in

order to capture potential failure surfaces propagating from the heel of the facing column (inset diagram in Figure 7). A similar conclusion was made by Bonaparte et al.

(1986) for the design of reinforced slopes under seismic loading. The same calculation

for AE can be carried out to determine the critical failure plane through the retained soil

mass that is assumed to propagate from the heel of the reinforced soil mass. The implication of such calculations to external stability analysis is that the size of the failure

GEOSYNTHETICS INTERNATIONAL

797

1.2

---15_

1.1

0.9

PAE

0.8

2P AE

H 2

---10_

1.0

kv

= 35_

= 2/3

= 0_

=0

--- 5_

0_

+ 5_

0.7

0.6

+10_

+15_

Conventional walls

0.5

0.4

0.3

0.2

0.1

0.0

0.0

0.1

0.2

0.3

0.4

0.5

kh

Figure 5. Influence of seismic coefficient, kh , and wall inclination angle, , on dynamic

earth force, PAE .

zone behind the reinforced soil zone will also increase with increasing horizontal acceleration.

The combined influence of horizontal and vertical accelerations on dynamic earth

force, PAE , is illustrated in Figure 8 for two wall inclination values. For values of horizontal seismic coefficient, kh , less than about 0.35 and kv = 2kh /3, downward vertical

components of seismic earth force (kv > 0) give the largest dynamic earth forces. However, the value of PAE using kv = +2kh /3 is only 7% greater than the value calculated using

kv = 0 for kh < 0.35. Hence, over a wide range of horizontal seismic coefficient values

the assumption that kv = 0 is reasonably accurate and, in fact, results in a slightly more

conservative value of PAE than values calculated assuming that the vertical component

of seismic earth force acts upward (kv < 0).

798

GEOSYNTHETICS INTERNATIONAL

1.4

= 10_

1.3

1.1

0_

1.2

PAE

1.0

5_

= 20_

0.9

0_

0.8

5_

10_

15_

= 0_

0_

5_

2P AE 0.7

H 2

10_

0.6

15_

10_

15_

0.5

0.4

= 35_

= 2/3

kv = 0

0.3

0.2

0.1

0.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

kh

Figure 6. Influence of seismic coefficient, kh , backslope angle, , and wall inclination angle,

, on dynamic earth force, PAE .

The influence of backslope angle, , and wall inclination angle, , on calculated values of dynamic earth force can be argued to have a more significant effect than the magnitude of kv in the example calculations. The requirement that peak horizontal and vertical accelerations must be time coincident in order to generate even the small calculated

differences in dynamic earth forces shown here gives support to recommendations in

current design guidelines for conventional gravity structures that vertical accelerations

can be safely ignored in many seismically active areas.

GEOSYNTHETICS INTERNATIONAL

799

70

kv

---15_ 65

---10_

--- 5_

60

0_

+ 5_

=

=

=

=

35_

2/3

0_

0

Conventional walls

AE (degrees)

+10_ 55

+15_

50

45

40

Segmental

retaining walls

+

35

kh = 0

kh > 0

30

25

20

0.0

Increased

reinforced soil mass

for kh > 0

Increased reinforcement

lengths

AE

0.1

0.2

0.3

0.4

0.5

kh

Figure 7. Influence of seismic coefficient, kh , and wall inclination angle, , on orientation

of internal failure plane, AE .

3.1

General

Potential external modes of failure are illustrated in Figures 2a, 2b and 2c. The analyses described in this section assume that the foundation provides a competent base such

that potential modes of failure are restricted to translational sliding along the base and

800

GEOSYNTHETICS INTERNATIONAL

1.0

0.9

PAE

0.8

H

0.7

0.6

2P AE 0.5

H 2

= 0_

kv = +2kh /3

kv = 0

kv = ---2kh /3

0.4

0.3

= 10_

kv = +2kh /3

0.2

kv = 0

0.1

0.0

0.0

= 35_

= 2/3

= 0_

0.1

kv = ---2kh /3

0.2

0.3

0.4

0.5

0.6

kh

Figure 8. Influence of seismic coefficients, kh and kv , and wall inclination angle, , on

dynamic earth force, PAE .

overturning about the toe of the gravity mass (reinforced soil zone plus facing column).

The dynamic earth force, PAE , calculated according to Equation 1 is used in external

stability calculations to estimate destabilizing active earth forces. In the external stability analyses to follow, the dynamic earth force imparted by the cohesionless retained

soil on the gravity mass is assumed to act along a surface that is parallel to the wall face

(i.e. at angle from the vertical) and at a constant distance L from the front face of the

wall (i.e. L is the minimum width of the gravity mass). According to NCMA guidelines,

GEOSYNTHETICS INTERNATIONAL

801

the minimum width of the gravity mass is L = 0.6H for critical structures and L = 0.5H

for non-critical structures. Dimension L is used to locate the heel of the assumed minimum gravity mass for external stability calculations and to control the minimum length

of all reinforcement layers (Figure 9).

3.2

The dynamic factor of safety against base sliding for purely frictional soils can be expressed as:

12

11

FS bsl (static)

kv = ---2kh /3

10

kv = ---kh /2

kh WR

WR (1kv )

Lw

PAE cos()

kv =0

6

5

4

3

= 20_

kv = ---2kh /3

kv = ---kh /2

kv = ---kh /4

kv = 0

= 0_

2

1.5

kv = ---kh /4

R = WR (1kv )tan

1

0

0.0

0.1

0.06

0.2

0.3

=

=

=

=

=

0.6

35_

0_

0.5H

0.4

0.5

0.6

kh

Figure 9. Static factor of safety against base sliding to give a minimum dynamic factor of

safety of 1.125 against base sliding for a range of seismic coefficients, kh and kv , and

backslope angle, .

(Notes: WR = weight of reinforced zone plus weight of facing column; and R = base sliding resistance.)

802

GEOSYNTHETICS INTERNATIONAL

L

LL

a + (1 k ) tan

H

H

LL

L

(1 k )a cos( ) + k

a +

H

H

w

FS bsl =

1K

2 AE

2

1

(8)

where:

a1 = 1 +

L Lw

tan

H

a2 = 1 +

L Lw

tan

2H

Here, Lw is the width of the facing column. Equation 8 can be simplified by setting Lw

= 0 for Lw << L. This assumption introduces no error for = 0 and negligible error for

moderate values of > 0. Parameter is an empirical constant that is used to artificially

reduce the inertial force of the gravity mass and applies only to the inertial part of Equation 8. A value of = 0.6 has been used for design purposes for both geosynthetic-reinforced soil walls (Christopher et al. 1989) and for Reinforced Earth walls that use steel

reinforcement strips (Segrestin and Bastick 1988). Parameter is assumed to be less

than unity to account for the transient nature of the peak accelerations in the gravity

mass and retained soils and the expectation that the inertial forces induced in the gravity

mass and the retained soil zone will not reach peak values at the same time during a

seismic event. The terms a1 and a2 in Equation 8 are geometric constants that account

for the effect of the backslope angle on the calculation of the mass of the reinforced soil

zone. The factor of safety for static loading conditions (FSbsl (static)) can be recovered

from Equation 8 by setting kh = kv = 0 .

The FHWA (Christopher et al. 1989) recommends that the minimum factor of safety

against base sliding under dynamic loading be no less than 75% of the minimum allowable static value. If this rule is applied to a minimum allowable static factor of safety

against sliding of 1.5 then, the data in Figure 9 shows that the maximum permissible

ground acceleration is 0.06g. Larger static factors of safety will be required to ensure

FSbsl (dynamic) (0.75)(1.5) = 1.125 as illustrated in the figure. Nevertheless, there

are limiting seismic coefficient values for any set of wall parameters for which there

is no solution using Equations 1 and 2 (i.e. < 0 ). The analytical results described here are consistent with conclusions made by Richards and Elms (1979) who

showed that there is very little margin of safety against base sliding of gravity structures

designed to meet minimum conventional static factors of safety. Figure 9 also demonstrates that the required static factor of safety against base sliding for a structure with

a horizontal backslope is relatively insensitive to the magnitude of vertical acceleration

for a horizontal seismic coefficient kh < 0.3. The vertical seismic coefficient used to generate the curves in Figure 9 was taken as kv 0 since negative values of kv gave the

most conservative results (i.e. largest required FSsld (static) values). The data in Figure

9 do show that even a modest backslope angle will require large increases in the static

factor of safety against base sliding to ensure an acceptable margin of safety under seismic loading.

GEOSYNTHETICS INTERNATIONAL

803

3.3

The moment arm, Ydyn , of the dynamic force normalized with respect to the wall

height (Figure 4) can be calculated as follows:

1 K + [K (1 k ) K ]

v

A

AE

A

Y dyn

m=

=3

H

KAE(1 k v)

(9)

The relationship between normalized moment arm, m, and horizontal seismic coefficient, kh , is shown in Figure 10a for values of ranging from 0.4 to 0.7 and = and

= 2/3 . The location of the dynamic force moment arm, m, is sensitive to the assumed

value of but is relatively insensitive to the range of interface friction angle, . For

0.6, the point of application of the dynamic force is generally at 1/3 to 1/2 of the wall

= 35_

0.70

0.65

=

= 2/3

0.7

0.60

0.55

m

= 0_

0.70

0.6

0.50

kv = 0

0.65

0.60

0.55

15_

0_

---15_

0.50

0.5

0.45

0.40

0.4

0.35

0.45

0.40

0.35

0.30

0.30

0.0 0.1 0.2 0.3 0.4 0.5 0.6

0.0 0.1 0.2 0.3 0.4 0.5 0.6

kh

kh

(a) = 0_

location, , wall inclination angle, , and wall-soil interface friction angle, , on location of

normalized dynamic moment arm, m.

804

GEOSYNTHETICS INTERNATIONAL

height above the toe. Figure 10b illustrates the influence of horizontal seismic coefficient, kh , and magnitude of wall inclination angle, , on parameter m for = 0.6. The

point of application of the dynamic earth force is shown to be only slightly dependent

on the magnitude of wall inclination angle associated with typical segmental retaining

walls.

The dynamic factor of safety against overturning about the toe of the free body comprising the reinforced soil mass and the facing column can be calculated as follows:

LL w

H

FS bot =

L

2L

b + LL

a + (1 k )

H

2

LL w

L

mKAE(1 k v)a 31 cos( ) + k h

b1 + w

H

H

(10)

where

b 1 = a1 + 1 (a1 1)2

3

b 2 = 1 + 2 (a 1 1)

3

and terms a1 and a2 are defined in Equation 8. Equation 10 is applicable to the case of

a vertical wall ( = 0) and will give a slightly conservative estimate of FSbot for inclined

wall facing columns. The static factor of safety, FSbot (static), required to satisfy FSbot

(dynamic) = (0.75)(2.0) = 1.5 is plotted in Figure 11 for the case of two different reinforcement length to wall height ratios, L/H . The vertical component of seismic force

has been taken as upward (-- kv ) in order to calculate results for the most critical orientation. If the conventional rule that the dynamic factor of safety be not less than 75% of

the minimum allowable static factor of safety against overturning (i.e. 2) is applied,

then the horizontal seismic coefficient is limited to kh = 0.04 in the example calculations. The data in Figure 11 illustrates that even a modest backslope angle leads to a

dramatic increase in the required static factor of safety for a given design acceleration.

Indeed, for the case of = 20_ and kh > 0.25 an acceptable margin of safety against overturning may not be possible. Figure 11 also shows that the effect of the L/H ratio on the

required value of FSbot (static) is relatively insignificant compared to the effect of backslope inclination for typical structures designed for static loading environments.

4

INTERNAL STABILITY

4.1

General

The influence of the magnitude of seismic coefficients on lateral earth forces has been

demonstrated earlier in the paper. If the calculation of dynamic reinforcement loads is

carried out in the same manner as for conventional static wall design then, the effect

of seismic loading can be shown to increase the magnitude of the net horizontal force

carried by the geosynthetic reinforcement layers. In addition, the change in the distribu-

GEOSYNTHETICS INTERNATIONAL

805

kv = ---2kh /3

15

13

12

Lw

14

kv = 0

kh WR

H

11

ma2 H

WR (1kv )

10

FSbot (static)

= 0_

PAE cos()

kv = ---2kh /3

kv = 0

= 20_

L/H = 0.5

L/H = 0.7

6

5

4

3

2

1

0

0.0

0.1

0.04

0.2

0.3

=

=

=

=

=

0.4

0.6

35_

0_

0.6

0.5

kh

Figure 11. Minimum static factor of safety against overturning required to give a factor of

safety of 1.5 against dynamic overturning for a range of seismic coefficients, kh and kv ,

backslope angle, , and length to height ratio, L/H.

(Note: WR = weight of reinforced zone plus weight of facing column.)

tion of the lateral earth pressure (Figure 4) means that the percentage of total lateral

force to be carried by the reinforcing elements in the upper portions of the wall will increase. Finally, the influence of ground accelerations on the volume of the internal potential failure wedge leads to an increase in the length of the reinforcement layers as

discussed in Section 2.5 and illustrated in Figure 7.

806

GEOSYNTHETICS INTERNATIONAL

4.2

The contributory area approach used for the static stability analysis of segmental retaining walls is extended to the dynamic loading case. In this method the reinforcement

layers are modelled as tie-backs with the dynamic tensile force, Fdyn , in each layer equal

to the dynamic earth pressure integrated over the contributory area, Sv , at the back of

the facing column plus the corresponding wall inertial force increment, kh Ww . The

contributory area for the topmost reinforcement layer is taken from the top of the crest

to mid-elevation between the first and second reinforcement layers from the crest. For

the simple geometry illustrated in Figure 12, the dynamic factor of safety, FSos , against

over-stressing of a reinforcement layer at depth z below the crest of the wall is given

by:

FS os =

=

T allow

F dyn

(11)

T allow

L

0.8K dyn cos( ) + (K A 0.6K dyn) cos( ) z + k h W HS v

H

H

Here, Tallow is the allowable tensile load for the reinforcement under seismic loading.

Dynamic

earth pressure

distribution

0.8Kdyn Hcos()

H

kh Ww

LW

Sv

Fdyn

Reinforcement

layer (typical)

Figure 12. Calculation of tensile load, Fdyn , in a reinforcement layer due to dynamic earth

pressure and wall inertia.

(Note: = 0.6.)

GEOSYNTHETICS INTERNATIONAL

807

loads can be explored by computing a magnification factor, rF , that is the ratio of dynamic tensile force, Fdyn , to static tensile force, Fsta , for a reinforcement layer at depth

z below the wall crest. Results of this calculation for reinforcement layers at five different depths below the wall crest are presented in Figure 13. The data illustrates that the

largest increases in reinforcement force occur in the shallowest layers in a reinforced

soil wall. This result is not unexpected due to the change in the active earth pressure

distribution that results from dynamic loading as discussed in Section 2.2 and illustrated

10

9

8

7

= 35_

0.2

F dyn

rF =

F sta

= 0_

= 0.6

= 2/3

= 0_

0.3

LW /H= 0.1

kv = ---2kh /3

kv = 0

kv = +2kh /3

rF

0.4

5

4

0.6

3

0.8

2

z/H

1

0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

kh

Figure 13. Influence of seismic coefficients, kh and kv , and normalized depth below crest

of wall, z/H, on dynamic reinforcement force amplification factor, rF .

808

GEOSYNTHETICS INTERNATIONAL

in Figures 4 and 12. The plotted data in Figure 13 also reveals that rF is sensibly independent of the magnitude of kv for kh 0.35, and hence solutions using kv = 0 are sufficiently accurate for design over this range and even slightly conservative at the shallowest

depth investigated (z/H=0.2).

An implication of these results to design is that the number of reinforcement layers

may have to be increased at the top of the wall in order to keep tensile loads within allowable limits. However, it should be noted that the allowable design tensile load under

dynamic loading is routinely taken as a greater percentage of the index strength of the

reinforcement than the percentage used for static loading design because of the short

duration of peak tensile loading during a seismic event. AASHTO (1992) guidelines can

be interpreted to permit the value of Tallow used for static loading designs to be increased

by 33% for the seismic loading condition. Rapid in-isolation wide-width strip tensile

loading of a typical high density polyethylene (HDPE) geogrid reinforcement reported

by Bathurst and Cai (1994) has demonstrated a potentially large increase in reinforcement stiffness of the material when compared to conventional rates of loading. This observation suggests that HDPE geogrids may be designed for much greater strengths under seismic loading than those values that result from the interpretation of AASHTO

recommendations.

4.3

The dynamic reinforcement tensile load must be carried by the reinforcement anchorage length which is located between the internal active failure plane (oriented at AE

from horizontal) and the reinforcement free end (Figure 7). A common approach for

anchorage capacity design is to use a simple Coulomb-type interface model in which

anchorage capacity is linearly proportional to anchorage length, overburden pressure

and soil shear strength (AASHTO 1990). An implication of this model for dynamic anchorage design is that the required anchorage length will increase in proportion to the

magnification factor, rF , introduced in the previous section. Owing to the short duration

of the anchorage force during a seismic event, the factor of safety against anchorage

pullout may be taken as 75% of the static value of 1.5 according to AASHTO (1992).

However, the principal effect of dynamic loading on anchorage design is the requirement to increase the length of reinforcement layers in order to capture the larger active

wedge of soil that occurs as a result of seismic loading, i.e. the internal failure plane

angle, AE , decreases as kh increases. This effect is illustrated in Figure 14. For example,

this figure shows that for a horizontal seismic coefficient, kh = 0.25, reinforcement

lengths would have to be increased by 60 to 100% of the lengths required under static

loading conditions. The requirement that the lengths of the uppermost reinforcement

layers may need to be increased for reinforced slopes subject to seismic loading has

been noted by Bonaparte et al. (1986) and is based on similar arguments.

4.4

walls requires that the designer check for internal sliding along horizontal planes that

pass along the reinforcement-soil interface and through the facing column between facing units. The results of large-scale shear tests reported by Bathurst and Simac (1994)

GEOSYNTHETICS INTERNATIONAL

809

5.0

30_

kv = 0

kv = ---2kh /3

4.5

+

4.0

kh = 0

kh > 0

L sta

3.5

L dyn

35_

AE

L dyn 3.0

Lsta

40_

45_

2.5

2.0

1.6

= 0_

1.5

= 0_

= 2/3

0.25

1.0

0.0

0.1

0.2

0.3

0.4

0.5

kh

Figure 14. Influence of seismic coefficients, kh and kv , and soil friction angle, , on ratio of

minimum reinforcement lengths, Ldyn /Lsta , to capture the internal failure wedge in

pseudo-static Coulomb wedge analyses.

have shown that the static shearing resistance, Vu , available at a horizontal interface

in the facing column can be described by a Coulomb-type failure law. This failure criterion can be modified to account for the dynamic loading case as follows:

810

GEOSYNTHETICS INTERNATIONAL

V u = au + Ww(1 k v) tan u

(12)

Parameters, au and u represent minimum interface shear capacity and equivalent interface friction angle, respectively, and are not expected to vary between static and seismic

loading conditions. The analysis required to calculate the factor of safety against internal sliding is similar to that described for external sliding. The dynamic factor of safety

against internal sliding along a horizontal surface at depth z below the crest of the wall

is calculated as:

FS isl =

Vu

LLw

+

a 2 (1 k v) tan ds

z

z2

LLz a + Lz

1 K (1 k )a 2 cos( ) + k

v 1

h

2 AE

(13)

Coefficients a1 and a2 are the same as those reported for Equation 8 substituting H = z.

Equation 13 assumes that the critical internal sliding mass is initiated at the free end

of the reinforcement layer. Parameter ds is the interface (direct sliding) friction angle

between the geosynthetic-reinforcement and the cohesionless reinforced soil. In general, ds < but the reduction in sliding resistance is typically more than compensated

for by the large value of shear interlock that is available in many block systems. However, the combined effect of a low interface friction value, ds , and facing units with low

shear capacity, Vu , can result in unacceptably low factors of safety against interface

shear and this failure mechanism must be checked as a matter of routine.

A parametric analysis involving Equation 13 was not performed. However, the results

plotted in Figure 9 are applicable for internal sliding for the case au = 0, u = ds = and

setting H = z.

5

FACING STABILITY

5.1

The following potential failure mechanisms must be examined in pseudo-static seismic analysis of the facing stability of segmental retaining walls: interface shear failure;

connection failure; toppling (local overturning) (Figure 2).

5.2

The influence of interface shear transmission on facing column stability can be analyzed by treating the facing column as a beam in which the integrated lateral pressure

(i.e. distributed load) must equal the sum of the reactions (forces in reinforcement layers). The calculation of interface shear force under dynamic loading must include the

effect of wall facing inertia. The general approach is illustrated in Figure 15. The total

force carried by reinforcement layers located above facing unit j is calculated as the area

ABDC of the lateral earth pressure distribution, plus the facing column inertial force

GEOSYNTHETICS INTERNATIONAL

811

Horizontal

component of

dynamic

earth pressure

distribution

z

kh

W

j+1

j

w

i

F dyn

i+1

Svi+1

k h Wjw

S vi

j

S dyn

C

F

D

E

i

F dyn

j

S dyn

= k h Wjw + AREA CDEF

elevation.

(Notes: N = total number of reinforcement layers; and M = total number of facing units.)

over the same height. The out-of-balance force to be carried through shear at the bottom

of facing unit j is simply the sum of the incremental column inertial force k hWjw plus

the force due to area CDEF in the figure. The partitioning of forces illustrated in the

figure is a direct result of the contributory area approach introduced earlier to assign

tensile loads to reinforcement layers. The locally maximum interface shear forces will

occur at reinforcement elevations. A general expression for the factor of safety against

dynamic interface shear failure (FSsc (dynamic)) at a reinforcement layer is:

FS sc =

=

Vu

S dyn

0.8K

(14)

Vu

dyn cos(

S

) + (K A 0.6K dyn) cos( ) z v

+ k LH HS2

4H

where: Sdyn = interface shear force; and Vu = shear capacity. For facing columns

constructed at large inclination angles the magnitude of the normal force, Ww , transmitted between facing units may be less than the sum of the weights of the individual

812

GEOSYNTHETICS INTERNATIONAL

facing units above the interface elevation. The loss of normal load is due to the effect

of the facing column units leaning into the reinforced soil mass. The maximum height

of column units that will transmit all of the facing weight to a target interface is called

the hinge height. Its calculation is described in detail by Bathurst et al. (1993) and

Simac et al. (1993).

The calculation of factor of safety against interface shear failure under static loading

conditions is carried out using Equation 14 with kv = kh = 0 . Figure 16 shows the ratio

1.0

= 35_

0.9

Sv /H = 0.2

= 0.6

0.8

LW /H = 0.1

= 0_

= 2/3

kv = ---2kh /3

= 0_

kv = 0

0.7

0.6

z/H

0.5

0.9

0.4

0.7

0.3

0.9

0.2

0.7 0.5

0.1

0.5

0.0

0.1

0.3

0.3

0.0

0.1

0.2

0.3

0.4

0.1

0.5

0.6

kh

Figure 16. Influence of seismic coefficients, kh and kv , and normalized depth below crest

of wall, z/H, on the ratio of dynamic to static interface shear factor of safety.

GEOSYNTHETICS INTERNATIONAL

813

of dynamic factor of safety to static factor of safety against interface shear failure for

a range of horizontal and vertical seismic coefficient values applied to an example wall

with five evenly spaced reinforcement layers (i.e. Sv /H=0.2). The data shows that the

potential for interface shear failure under seismic loading increases with proximity of

the shear interface to the crest of the wall. However, the effect of vertical acceleration

on calculated dynamic factors of safety diminishes with height of interface for the worst

case situation of vertical seismic force components acting upward. The curves for

z/H=0.1 correspond to sliding stability of the top unreinforced portion of the wall facing. These curves appear to support the argument that narrow unreinforced heights of

segmental facing units are susceptible to sliding failure. However, in practice, large values of interface shear capacity are possible with many modular block systems that are

constructed with shear keys, or other forms of positive interlock. These systems, as opposed to systems that rely solely on frictional sliding resistance, are the preferred choice

in order to achieve an adequate margin of safety against interface sliding. Furthermore,

Bathurst and Simac (1994) have demonstrated that the shear capacity at an interface

layer may be substantially reduced by the presence of a reinforcement inclusion if interface shear capacity is developed primarily through frictional resistance.

A minimum factor of safety against interface shear failure, for critical structures under static loading, is 1.5 according to NCMA guidelines. Reducing this criterion by 25%

for the dynamic loading condition is recommended in order to be consistent with the

approach adopted for base sliding.

5.3

The influence of increased dynamic forces on connection load is identical to the analysis described for reinforcement over-stressing. Peak connection load capacities under

static loading conditions have been described using bi-linear failure envelopes based

on the results of full-scale connection tests carried out at rates of loading matching the

10% strain/min rate used in the ASTM D 4595 method of test (Bathurst and Simac

1993). A Coulomb-type law with a maximum connection load cut-off has been used by

the first writer and co-workers to characterize a large number of (static) test results. Modified for the dynamic loading condition, the peak connection load envelope becomes:

F c = acs + Ww (1 k v) tan cs F c(max)

(15)

where the parameters acs and cs represent the minimum connection capacity and the

slope of the connection strength envelope, respectively. The dynamic factor of safety

against connection failure, FScn , is expressed by Equation 11 by replacing Tallow with

Fc . Depending on the connection type, it is possible that the maximum reinforcement

load, Fdyn , may be limited by the facing connection capacity. The results of connection

tests carried out at different rates of loading have demonstrated that peak connection

capacities may be sensitive to rate of loading (Bathurst and Simac 1993). At the time

of writing there is no data available that can be used to quantify changes in connection

capacity that may develop as a result of repeated application of load and the rapid loading rates anticipated during a seismic event.

814

GEOSYNTHETICS INTERNATIONAL

5.4

using the beam analog described in Section 5.2 for interface shear stability calculations.

Internal moments that cause net outward moment at the toe of a facing unit provide a

possible failure mechanism for which an adequate factor of safety should be checked.

Local peak destabilizing moments will occur at reinforcement elevations. The factor

of safety expression adopted by the NCMA (Simac et al. 1993), for local overturning

at a reinforcement layer i under static loading conditions, can be modified for the dynamic loading case as follows:

N

M R(1 k v) +

FS lot =

i

c

Y ic

i+1

dyn

H

H

2 hH

6 a

(16)

where: MR = resistance to static overturning due to facing column self-weight above the

toe of the target facing unit; and N = number of reinforcement layers. The summation

with F ic Yic terms denotes the resisting moment due to the connection capacities of reinforcement layers, F ic , and their corresponding moment arms, Y ic , from the target point

of rotation.

The ratio of the dynamic to static factor of safety against local overturning is plotted

in Figure 17 for five different interface elevations ranging from z/H = 0.2 to 1. To simplify example calculations the connection capacities, F ic , have been assumed to be purely

frictional (acs = 0). The largest reductions in factor of safety under seismic loading were

found to occur when the vertical component of the seismic force acts upward (kv =

- 2kh /3). The figure clearly shows that for the seismic coefficient values investigated,

the shallow interface layers (small z/H values) require a higher static factor of safety

against overturning to maintain a dynamic factor of safety equal to or greater than unity.

The unsupported height of the facing column at the top of the structure is the most critical portion of the wall in these calculations (i.e. z/H=0.2 in this example). However, the

effect of the magnitude and orientation of the vertical component of seismic force on

the upper portions of the wall (z/H < 0.4) in the example used to produce Figure 17 is

only significant for horizontal seismic coefficient values greater than 0.3.

The experience of the writers is that designers typically try to maximize the unreinforced height of wall at the crest in order to reduce reinforcement quantities in segmental retaining wall structures. This strategy will result in unacceptably low margins

of safety against toppling at the top of the structure under dynamic loading conditions

as illustrated in Figure 17. The only strategy to minimize the potential for this failure

mechanism to occur is to introduce reinforcement layers close to the wall crest and to

ensure that these layers have adequate facing connection capacity. Although the results

presented in Figure 17 suggest that toppling of the facing column is a potential problem,

the results of the analyses presented in the next section for two walls that survived the

Northridge Earthquake in 1994 show that this problem did not develop in practice because reinforcement layers were placed close to the top of the structure and static factors

of safety against local overturning were very high.

GEOSYNTHETICS INTERNATIONAL

815

1.0

= 35_

= 0_

= 0.6

0.9

= 2/3

= 0_

0.8

0.7

FSlot (dynamic) / FS lot (static)

F Nc

PAE cos()

Fi+1c

(1kv )Ww

kh Ww

Layer i

MR

0.6

LW /H = 0.1

LW

0.5

z/H

0.4

1.0

0.8

0.3

1.0

0.8

0.2

0.6

kv = ---2kh /3

0.1

0.4

0.2

kv = 0

0.0

0.0

0.1

0.2

0.3

kh

0.4

0.5

0.6

0.4

0.2

0.6

Figure 17. Influence of seismic coefficients, kh and kv , and normalized depth below crest of

wall, z/H, on the ratio of dynamic to static local overturning factor of safety.

CASE STUDIES

6.1

1994

than 4.5 m in height in the Los Angeles area immediately after the Northridge Earth-

816

GEOSYNTHETICS INTERNATIONAL

quake of 17 January 1994 (moment magnitude = 6.7). The results of the survey showed

no evidence of visual damage to 9 of 11 structures located within 23 to 113 km of the

earthquake epicenter. Two structures (Valencia and Gould Walls) showed tension

cracks within and behind the reinforced soil mass that were clearly attributable to the

results of seismic loading. Data supplied by Sandri and from other sources can be used

to estimate factors of safety against external, internal and facing stability modes of failure for these two structures. The results of these analyses allow a preliminary assessment to be made of the applicability of the pseudo-static approach described in this paper to actual field performance.

6.2

Valencia Wall

The Valencia wall has a maximum height of 6.5 m and is located at a distance of 23

km from the epicenter of the Northridge Earthquake. This wall had the smallest epicentral distance of all 11 structures surveyed by Sandri. Three landslides were reported to

have occurred in the general area of the Valencia wall as a result of the earthquake. The

foundation soils at the Valencia wall site are composed of a deep deposit of silty sand

and clay. A cross-section of the wall is shown in Figure 18. The width of the reinforced

mass measured from the toe of the wall is 5.5 m with the exception of the top 2.2 m of

the structure where the reinforcement lengths were shortened to facilitate placement of

subsurface utilities. Hence, this upper portion of the wall has a reinforced mass that is

1.8 m wide measured from the wall face. Design data for the wall is limited but it appears that the wall was designed for kh = 0.3 and kv = 0 and the effect of the horizontal

acceleration was treated as an additional uniform horizontal earth pressure distribution

equal to 104 kPa. No data is available to show how this distribution was used (if at all)

in stability calculations related to internal and local facing modes of failure.

The estimated peak horizontal ground accelerations at this site range from 0.19g and

0.5g based on data from UCB/EERC (1994). The lower value is based on a mean estimate taken from a peak horizontal ground acceleration-epicentral distance attenuation

curve for the Northridge earthquake. The maximum value is based on peak ground acceleration contours reported for the Northridge earthquake in the Los Angeles area. A

simplifying assumption made by the writers is that the range of peak ground accelerations estimated for the site can be used as a conservative estimate of the range of kh values for the backfill soils (i.e. would be conservative for design). This range of values

also includes the value of kh that was used in the original design. Application of Equation 7 to this range of peak ground acceleration values according to current FHWA

guidelines would result in a mean value of kh = 0.23 and a maximum value of kh = 0.48.

However, these adjustments do not influence the general conclusions that follow, and

to simplify the analyses, values of kh = 0.19 and kh = 0.5 are used for demonstration purposes only. No site-specific vertical acceleration data is available and kv = 0 was assumed by the writers. Soil and reinforcement properties used in the stability analysis

of the structure are given in Table 1.

The Coulomb failure wedge geometries calculated using Equation 6 for static, mean

and maximum kh values are illustrated in Figure 18. Due to the shortened length of reinforcement at the top portion of the wall, the top 2.2 m of wall was analyzed as a separate

structure (Figure 18a). The results of static and dynamic stability analyses are given in

Table 2 for the top 2.2 m height of wall and Table 3 for the entire wall. All of the calcu-

GEOSYNTHETICS INTERNATIONAL

817

(a)

1 2

H=2.2 m

Wedge

1

2

3

12

11

10

L=1.8 m

(b)

kh

0.0 Static

0.19 Mean

0.50 Max

AE

56.6_

48_

26_

1

12

Type 4

11

10

Type 3

Layer number

9

8

H=6.5 m

Type 2

7

6

5

4

3

2

L=5.5 m

Type 1 geogrid

1

AE

Figure 18. Cross-section view of Valencia Wall showing location and orientation of

internal failure planes under static and dynamic loading conditions: (a) top portion of wall;

(b) entire wall.

lated factors of safety for the static loading condition are larger than minimum recommended values reported in the NCMA guidelines for critical structures (Simac et al.

1993).

Analytical results in Table 2 corresponding to mean and maximum kh values show that

factors of safety for the top 2.2 m of wall are within acceptable limits (i.e. >

0.75 minimum allowable static values) with the exception of base overturning (1.18)

and reinforcement over-stressing (0.75). However, over-stressing may not be a problem

since the peak seismic loading is transient and the long-term design strength of the reinforcement based on conventional static design is very conservative for seismic design.

A value of 1.18 for dynamic overturning is still well above unity and in the opinion of

the writers appears unacceptably low only because the default minimum static factor

of safety (2.0) is unreasonably high (i.e. FSdyn (dynamic) = (0.75)(2.0) = 1.5). The pre-

818

GEOSYNTHETICS INTERNATIONAL

dicted dynamic failure wedges are seen to extend well beyond the topmost reinforcement layer. The shortening of these layers over the top section of the wall is a design

shortcoming and is consistent with visual evidence of distress to this structure reported

by Sandri (i.e. 50 mm wide surface cracks directly at the back of the shortened reinforcement length portion).

Table 1. Material properties for Valencia Wall and Gould Wall.

Properties

Values

Valencia Wall

Gould Wall

Soil1

, Friction angle (o)

33

33

19.8

19.8

Type 1

125

Type 2

100

Type 3

49.5

Type 4

35.5

35.5

Geosynthetic properties2

Index strength (ASTM D 4595) (kN/m)

Type 1 and 2

27

Type 3 and 4

8.3

8.3

600

600

Height (mm)

200

200

Length (mm)

450

450

117

117

32.7

32.7

15

15

23.5

32

acs (kN/m)

17

17

cs (o)

Fc(max) (kN/m)

17

17

Interface

shear3

au (kN/m)

u (o)

Connection

strength4

Type 1 and 2

acs (kN/m)

cs (o)

Fc(max) (kN/m)

Type 3 and 4

Notes: (1) Sandri (1994); (2) Manufacturers recommended values; (3) unpublished data; and (4) GeoSyntec

Consultants (1991).

GEOSYNTHETICS INTERNATIONAL

819

Static

Mechanism

Factor

of

safety

Dynamic

Default

minimum

(NCMA)

Calculated

minimum

Default

minimum

Calculated minimum

kh = 0.19

(mean)

kh = 0.5

(max)

Base sliding

FSbsl

1.5

5.21

1.125

2.94

1.16

Base

overturning

FSbot

2.0

11.30

1.5

2.68

1.18

Reinforcement

over-stressing

FSos

1.2

1.26 (10)

1.14 (10)

0.75 (10)*

Pullout

FSpo

1.5

1.75 (12)

1.125

Internal sliding

FSisl

1.5

11.52 (10)

1.125

4.15 (10)

1.63 (10)

Local

overturning

FSlot

2.0

8.05 (9)

1.5

4.94 (12)

1.81 (12)

Facing shear

FSsc

1.5

25.47 (11)

1.125

13.8 (11)

5.94 (11)

Connection

FScn

1.5

2.64 (10)

1.125

2.38 (10)

1.57 (10)

Notes: Values in parentheses () refer to reinforcement layer number counted from the bottom of the wall; X

denotes internal failure plane extends beyond free end of reinforcement; and * factor of safety less than unity.

Static

Mechanism

Factor

of

safety

Dynamic

Default

minimum

(NCMA)

Calculated

minimum

Default

minimum

Calculated minimum

kh = 0.19

(mean)

kh = 0.5

(max)

Base sliding

FSbsl

1.5

5.30

1.125

2.07

1.01*

Base

overturning

FSbot

2.0

11.17

1.5

2.90

1.10

Reinforcement

over-stressing

FSos

1.2

2.14 (5)

1.88 (5)

1.12 (5)*

Pullout

FSpo

1.5

17.88 (8)

1.125

15.70 (8)

Internal sliding

FSisl

1.5

5.26 (1)

1.125

2.15 (1)

1.02 (1)*

Local

overturning

FSlot

2.0

6.42 (1)

1.5

3.80 (8)

1.61 (8)

Facing shear

FSsc

1.5

7.27 (4)

1.125

5.83 (4)

3.26 (4)

Connection

FScn

1.5

4.78 (5)

1.125

4.20 (5)

2.50 (5)

Notes: Values in parentheses () refer to reinforcement layer number counted from the bottom of the wall; X

denotes internal failure plane extends beyond free end of reinforcement; and * marginal factor of safety.

820

GEOSYNTHETICS INTERNATIONAL

Analytical results summarized in Table 3 for the entire structure (i.e. assuming a reinforced soil mass with a base width of 5.5 m) show that all dynamic factors of safety are

greater than unity. However, the dynamic factors of safety against base sliding (1.01)

and internal sliding (1.02) are marginal with respect to collapse for kh = 0.5. The predicted internal Coulomb wedge failure planes plotted in Figure 18b can be seen to intersect the soil surface beyond the back of the reinforced soil zone for both kh = 0.19 and

kh = 0.5 and is consistent with surface cracking reported by Sandri.

Inspection of dried mud that had accumulated on the face of the retaining wall prior

to the seismic event was observed to be undisturbed and intact at the time of post-earthquake inspection supporting the analytical results that predict no facing instability.

6.3

Gould Wall

The Gould Wall has a maximum height of 4.6 m and is located at a distance of 35 km

from the epicenter of the Northridge Earthquake. The structure is founded on rock. A

cross-section of the wall is shown in Figure 19. A single geogrid reinforcement type was

used in this structure and each layer was extended to a uniform length of 3.6 m from

the face of the wall. Soil and reinforcement properties assumed for the structure are given in Table 1. The wall was not designed for seismic loading. The peak mean and maxi-

1

Layer number

11

10

9

8

7

6

5

4

3

2

1

H=4.6 m

Wedge

1

kh

0.0

Static

AE

56.6_

0.12 Mean

53_

0.30 Max

42_

AE

L=3.6 m

Figure 19. Cross-section view of Gould Wall showing location and orientation of internal

failure planes under static and dynamic loading conditions.

GEOSYNTHETICS INTERNATIONAL

821

mum horizontal ground acceleration values at this site are estimated by the writers to

range from 0.12g (mean) and 0.3g (maximum), based on interpretation of data from

UCB/EERC (1994) as described for the Valencia Wall. The vertical seismic coefficient

value, kv , has been taken as zero. As in the Valencia Wall case, the writers have assumed

that peak ground accelerations estimated at the site can be used as a first approximation

of the range of kh for demonstration purposes. The results of stability analyses under

static and dynamic loading conditions are given in Table 4. The data in Table 4 for the

Gould Wall show that factors of safety for the static loading condition are sufficiently

great that reduced factors of safety using kh = 0.3 at the site do not fall below minimum

acceptable levels (i.e. 0.75 minimum allowable static values) with the exception of

reinforcement over-stressing (0.69) and local overturning at the bottom of the unreinforced portion of the wall (1.35). Comparison of factors of safety for internal and facing

modes of failure for static and dynamic loading conditions show clearly how the most

critical elevations are higher in the wall under seismic loading conditions than under

static loading conditions. The internal failure plane which is approximately at the location of the back of the reinforced soil mass under static conditions can be seen to extend

beyond the reinforced soil mass under dynamic loading. This analytical result is consistent with the observed surface cracks in this zone. Close inspection of the wall face did

not reveal any evidence of relative movement of the facing units which is consistent

with estimated minimum values of dynamic factors of safety against facing modes of

failure which are well above unity (Table 4).

Table 4. Stability analysis results of Gould Wall.

Static

Mechanism

Factor

of

safety

Dynamic

Default

minimum

(NCMA)

Calculated

minimum

Default

minimum

Calculated minimum

kh = 0.12

(mean)

kh = 0.3

(max)

Base sliding

FSbsl

1.5

4.85

1.125

2.55

1.38

Base

overturning

FSbot

2.0

9.44

1.5

3.45

1.71

Reinforcement

over-stressing

FSos

1.2

1.38 (1)

1.20 (9)

0.69 (9)*

Pullout

FSpo

1.5

3.63 (11)

1.125

2.02 (11)

Internal sliding

FSisl

1.5

5.13 (1)

1.125

2.76 (1)

1.49 (1)

Local

overturning

FSlot

2.0

4.58 [

1.5

2.96 (11)

1.35 (11)

Facing shear

FSsc

1.5

14.03 (5)

1.125

6.15 (11)

2.47 (11)

Connection

FScn

1.5

2.95 (1)

1.125

2.56 (11)

1.48 (11)

Notes: Values in parentheses () refer to reinforcement layer number counted from the bottom of the wall; X

denotes internal failure plane extends beyond free end of reinforcement; [ overturning about the toe of the

bottom facing unit; and * factor of safety less than unity.

822

GEOSYNTHETICS INTERNATIONAL

6.4

Seismic stability analyses of the Valencia Wall and Gould Wall using the pseudo-static approach introduced in this paper are based on a number of assumptions. The writers

have used properties for backfill soils reported in the original designs and reinforcement

properties suggested by the manufacturer of the reinforcement materials for static loading conditions. In addition, connection performance and interface sliding data has been

inferred from laboratory test results under static loading conditions.

Perhaps the most important assumptions are related to the selection of seismic coefficient values that are implemented in the stability calculations. As discussed in Section

2.4.2, strategies to calculate a representative seismic coefficient for wall design based

on site peak ground acceleration values vary widely. Nevertheless, the calculations consistently show that for the range of kh values assumed, the inadequacy of reinforcement

lengths in the upper portions of each wall is apparent. The design methodology proposed herein would have led to their increase. It is interesting to note that tilting of a

geogrid-reinforced soil wall that was observed after the Great Hanshin Earthquake of

17 January 1995 in Japan has been attributed to inadequate reinforcement lengths at the

top of the structure (Tatsuoka et al. 1995).

Finally, with respect to Valencia and Gould Walls, the writers wish to emphasize that

while tension cracks were observed in these structures the function of the structures was

not compromised in any practical way. In fact, these structures can be judged to have

performed satisfactorily despite potentially large dynamic loadings.

7

CONCLUSIONS

The paper has presented a pseudo-static Mononobe-Okabe (M-O) approach for limit

equilibrium stability analysis of geosynthetic-reinforced segmental retaining walls.

The approach extends the Coulomb wedge method that is currently recommended for

the static stability analysis of these types of structures. Calculations are reasonably simple and are framed within the conventional limit-equilibrium approach used by geotechnical engineers for geosynthetic-reinforced soil walls.

The following implications to geosynthetic-reinforced segmental retaining wall design can be made based on the results of a number of parametric analyses presented in

the paper:

1. The method proposed in this paper to calculate internal dynamic earth pressure distributions results in a redistribution of tensile load to reinforcement layers and facing connections located close to the top of the wall. The number of reinforcement

layers at the top of the wall may have to be increased to compensate for the combined effect of larger earth forces and redistribution of earth pressures.

2. The progressive inclination of the internal active failure plane with increasing magnitude of horizontal seismic coefficient can lead to the requirement for greater reinforcement lengths at the top of geosynthetic-reinforced soil walls than those lengths

calculated based on static loading conditions.

3. Example calculations for base sliding and overturning about the toe of the gravity

mass assumed in geosynthetic-reinforced soil wall design demonstrate that there is

GEOSYNTHETICS INTERNATIONAL

823

4.

5.

6.

7.

little margin of safety against these modes of failure under seismic loading conditions for walls designed to just satisfy minimum factors of safety under static loading conditions.

The unreinforced portion of the facing column above the uppermost reinforcement

layer was demonstrated to be the most critical portion of a geosynthetic-reinforced

segmental wall with respect to local shear failure and toppling. However, reductions

in factors of safety against local shear and toppling failure were relatively insensitive to the magnitude of vertical seismic coefficient assumed in the example calculations when compared to the influence of the magnitude of the horizontal seismic

coefficient. Hence, for the unreinforced top portions of these walls, negligible error

results from assuming kv = 0.

Minimizing the height of the top unreinforced portion of the wall is an important

strategy to ensure adequate factors of safety against local shear failure and toppling

of the facing column under seismic loading. Segmental facing units that have positive shear interlock in the form of concrete keys, pins, or other forms of mechanical

connectors are the preferred choice in segmental retaining wall design to ensure that

these systems have adequate interface shear capacity.

Pseudo-static seismic analysis of two walls that experienced significant ground accelerations during the 1994 Northridge Earthquake predicted failure surfaces exiting beyond the reinforced soil zone. Observed cracks in the backfill soils can be considered to be consistent with the predicted range of internal failure plane

geometries.

The observed good performance of the facing column of two walls during the Northridge Earthquake is predicted by pseudo-static seismic analysis and demonstrates

that geosynthetic-reinforced segmental retaining walls can be designed to withstand

a significant earthquake event provided that facing units are able to develop adequate interface shear capacity and reinforcement layers are placed close to the crest

of the wall.

The results of this investigation have identified the following research needs related

to stability analyses of geosynthetic-reinforced soil retaining walls in general and segmental retaining walls in particular:

1. The properties of geosynthetic reinforcement materials under rapid loading are not

well understood and new methods to select allowable design loads under seismic

loading are required to reduce likely conservativeness in current seismic design

methods.

2. Laboratory testing of the connection formed between the reinforcement and modular facing units is required to provide connection capacity data that is applicable to

seismic loading conditions.

824

GEOSYNTHETICS INTERNATIONAL

3. The effect of rapid cyclic loading on load transfer in the reinforcement anchorage

zone is not well understood. Current anchorage models for static loading conditions

should be investigated and modified to account for the seismic loading condition.

A shortcoming of the pseudo-static method outlined in this paper is that there appears

to be very little guidance on how to select seismic coefficient values based on site-specific ground motion data. Recommendations that are available in the literature vary

widely. Another shortcoming of the pseudo-static seismic method of design proposed

by the writers is that it can only provide the designer with an estimate of the margins

of safety against collapse of segmental retaining walls, or failure of their components,

and does not provide any direct estimate of anticipated wall deformations. This is a deficiency that is common to all limit-equilibrium methods of design in geotechnical engineering. In practice, geosynthetic-reinforced segmental retaining walls may fail because of unacceptable deformations.

More sophisticated analytical techniques are available to the designer that can be

used to predict the time-deformation response of these systems. For example, dynamic

finite element analyses have been carried out on reinforced soil structures (Yogendrakumar et al. 1992; Bachus et al. 1993; Cai and Bathurst 1995). However, the experience

of the writers is that finite element model techniques require material properties that

are seldom available to designers and the interpretation of results by inexperienced users of finite element programs is always a concern.

An alternative strategy for the design of gravity retaining walls is a displacement

method approach (Richards and Elms 1979; Whitman 1990) which can explicitly incorporate horizontal wall movements in stability analyses. This approach, adapted to

reinforced segmental retaining walls, is currently under development by the writers and

will be described in a forthcoming publication.

ACKNOWLEDGEMENTS

The writers would like to thank Mr. M.R. Simac of Earth Improvement Technologies

and Mr. D. Sandri with the Nicolon Mirafi Group for their review of the original manuscript. The writers also acknowledge the efforts of two anonymous reviewers whose

comments materially improved the revised manuscript. The funding for the work reported in the paper was provided by the Department of National Defence (DND, Canada) through an Academic Research Program (ARP) grant to the senior writer and by the

Director of Architecture (DArch) (DND, Canada).

REFERENCES

AASHTO-AGC-ARTBA, Design Guidelines for Use of Extensible Reinforcements

(Geosynthetic) for Mechanically Stabilized Earth Walls in Permanent Applications,

In Situ Soil Improvement Techniques, Task Force 27 Report, American Association

of State and Highway Transportation Officials, Washington, D.C., USA, August

1990, 38 p.

GEOSYNTHETICS INTERNATIONAL

825

AASHTO, 1992, Standard Specifications for Highway Bridges, 15th Edition, American Association of State Highway and Transportation Officials, Washington, DC,

USA,686 p.

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 BlockFaced Geosynthetic Walls, Transportation Research Record 1414, Washington, DC,

USA, pp. 6-11.

Bachus, R.C., Fragaszy, R.J., Jaber, M., Olen, K.L., Yuan, Z. and Jewell, R., 1993, Dynamic Response of Reinforced Soil Systems, Report ESL-TR-92-47, Engineering

Research Division, US Department of the Air Force Civil Engineering Support

Agency, March 1993, Vol. 1, 230 p., Vol. 2, 227 p.

Bathurst, R.J., and Cai, Z., 1994, In-isolation Cyclic Load-Extension Behavior of Two

Geogrids, Geosynthetics International, Vol. 1, No. 1, pp. 1-19.

Bathurst, R.J. and Simac, M.R., 1993, Laboratory Testing of Modular Unit-Geogrid

Facing Connections, Geosynthetic Soil Reinforcement Testing Procedures, Cheng,

S.C.J., Ed., ASTM STP 1190, Proceedings of a symposium held in San Antonio, TX,

USA, January 1993, pp. 32-48.

Bathurst, R.J. and Simac, M.R., 1994, Geosynthetic Reinforced Segmental Retaining

Wall Structures in North America, Proceedings of the Fifth International Conference on Geotextiles, Geomembranes and Related Products, Singapore, September

1994, 24 p.

Bathurst, R.J., Simac, M.R., and Berg, R.R., 1993, Review of the NCMA Segmental

Retaining Wall Design Manual for Geosynthetic-Reinforced Structures, Transportation Research Record 1414, Washington, DC, USA, pp. 16-25.

Bonaparte, R., Schmertmann, G.R. and Williams, N.D., 1986, Seismic Design of

Slopes Reinforced with Geogrids and Geotextiles, Proceedings of the Third International Conference on Geotextiles, Vol. 1, Vienna, Austria, pp. 273-278.

Cai, Z. and Bathurst, R.J., 1995, Seismic Response Analysis of Geosynthetic Reinforced Soil Segmental Retaining Walls by Finite Element Method, Computers and

Geotechnics, Vol. 17, No. 4, pp. 523-546.

Canadian Geotechnical Society, 1992, Canadian Foundation Engineering Manual,

3rd Edition, 512 p.

Cazzuffi, D. and Rimoldi, P., 1994, The Italian Experience in Geosynthetics Reinforced Soil Retaining Wall with Vegetated and Concrete Facing, Recent Case Histories of Permanent Geosynthetic-Reinforced Soil Retaining Walls, Tatsuoka, F. and

Leshchinsky, D. Eds., Balkema, 1994, Proceedings of Seiken Symposium No. 11, Tokyo, Japan, November 1992, pp. 21-43.

Chida, S., Minami, K. and Adach, K., 1982, Test de stabilit de remblais en Terre Arme. (translated from Japanese)

826

GEOSYNTHETICS INTERNATIONAL

Christopher, B.R., Gill, S.A., Giroud, J.P., 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, Washington, DC, USA, November

1989, 287 p.

Collin, J.G., Chouery-Curtis, V.E. and Berg, R.R., 1992, Field Observations of Reinforced Soil Structures Under Seismic Loading, Earth Reinforcement Practice,

Ochiai, Hayashi and Otani, Eds., Balkema, 1992, Proceedings of the International

Symposium on Earth Reinforcement Practice, Fukuoka, Kyushu, Japan, Vol. 1, November 1992, pp. 223-228.

Ebling, R.M. and Morrison, E.E., 1993, The Seismic Design of Waterfront Retaining

Structures, Naval Civil Engineering Laboratory Technical Report ITL-92-11 NCEL

TR-939, Port Huenene, CA, USA, 329 p.

Eliahu, U. and Watt, S., 1991, Geogrid-Reinforced Wall Withstands Earthquake,

Geotechnical Fabrics Report, IFAI, St. Paul, MN, USA, Vol. 9, No. 2, pp. 8-13.

Fang, Y.-S. and Chen, T.-J., 1995, Modification of Mononobe-Okabe Theory, Gotechnique, Vol. 45, No. 1, pp. 165-167.

GeoSyntec Consultants, 1991, Masonry Block Wall Connection Evaluation with Select Geogrid Products, Report prepared by Geosyntec Consultants for Tensar Earth

Technologies, Inc., 12 p.

Gourc, J.P., Gotteland, P. and Wilson-Jones, H., 1990, Cellular Retaining Walls Reinforced by Geosynthetics: Behaviour and Design, Performance of Reinforced Soil

Structures, McGown, A., Yeo, K., and Andrawes, K.Z., Eds., Thomas Telford, 1991,

Proceedings of the International Reinforced Soil Conference held in Glasgow, Scotland, September 1990, pp. 41-45.

Ichihara, M. and Matsuzawa, H., 1973, Earth Pressure During Earthquake, Soils and

Foundations, JSSMFE, Vol. 13, No. 4, pp. 75-86.

Ishibashi, I. and Fang, Y.-S., 1987, Dynamic Earth Pressures with Different Wall

Movement Modes, Soils and Foundations, JSSMFE, Vol. 27, No. 4, pp. 11-22.

Knutson, A.F., 1990, Reinforced Soil Retaining Structures, Norwegian Experiences,

Proceedings of the Fourth International Conference on Geotextiles, Geomembranes

and Related Products, Vol. 1, Balkema, May 1990, The Hague, Netherlands, pp.

87-91.

Motta, E., 1994, Generalized Coulomb Active Earth Pressure for Distanced Surcharge, Journal of the Geotechnical Engineering Division, ASCE, Vol. 120, No. 6,

pp. 1072-1079.

Okabe, S., 1924, General Theory on Earth Pressure and Seismic Stability of Retaining

Wall and Dam, Doboku Gakkaishi - Journal of the Japan Society of Civil Engineers,

Vol. 10, No. 6, pp. 1277-1323.

Prakash, S., 1981, Soil Dynamics, McGraw-Hill Book Company, New York, NY,

USA, 426 p.

Preliminary Report on the Principal Geotechnical Aspects of the January 17, 1994

Northridge Earthquake, Report No. UCB/EERC-94/08, University of California at

Berkeley, Earthquake Engineering Research Center, Stewart, J.P., Bray, J.D., Seed,

R.B. and Sitar, N., Eds., June 1994, 245 p.

GEOSYNTHETICS INTERNATIONAL

827

Richards, R. and Elms, D.G., 1979, Seismic Behavior of Gravity Retaining Walls,

Journal of the Geotechnical Engineering Division, ASCE, Vol. 105, No. GT4, pp.

449-464.

Sandri, D., 1994, 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, pp. 103-147.

Segrestin, P. and Bastick, M.J., 1988, Seismic Design of Reinforced Earth Retaining

Walls - The Contribution of Finite Element Analysis, Theory and Practice of Earth

Reinforcement, Yamanouchi, T., Miura, N. and Ochiai, H., Eds., Balkema, 1988, Proceedings of the International Geotechnical Symposium on Theory and Practice of

Earth Reinforcement, Fukuoka, Kyushu, Japan, October 1988, pp. 577-582.

Simac, M.R., Bathurst, R.J., Berg, R.R. and Lothspeich, S.E., 1993, National Concrete Masonry Association Segmental Retaining Wall Design Manual, National

Concrete and Masonry Association, Herdon, VA, USA, March 1993, 250 p.

Simac, M.R., Bathurst, R.J. and Goodrum, R.A., 1991, Design and Analysis of Three

Reinforced Soil Retaining Walls, Proceedings of Geosynthetics 91, IFAI, Vol. 2,

Atlanta, GA, USA, February 1991, pp. 781-798.

Tatsuoka, F., Tateyama, M. and Koseki, J. 1995, Performance of Geogrid-Reinforced

Soil Retaining Walls During the Great Hanshin-Awaji Earthquake, January 17,

1995, to appear in Proceedings of the First International Conference on Earthquake

Geotechnical Engineering, IS-Tokyo 95, Tokyo, Japan, November 1995, 8 p.

Vrymoed, J., 1989, Dynamic Stability of Soil-Reinforced Walls, Transportation Research Record 1242, Washington, DC, USA, pp. 29-38.

Whitman, R.V., 1990, Seismic Design and Behavior of Gravity Retaining Walls, Design and Performance of Earth Retaining Structures, Geotechnical Special Publication No. 25, Lambe, P.C. and Hansen, L.A., Eds., ASCE, 1990, Cornell University,

NY, USA, pp. 817-842.

Wolfe, W.E., Lee, K.L., Rea, D. and Yourman, A.M., 1978, The Effect of Vertical Motion on the Seismic Stability of Reinforced Earth Walls, Proceedings of ASCE Symposium on Earth Reinforcement, Pittsburgh, PA, USA, April 1978, pp. 856-879.

Won, G.W., 1994, Use of Geosynthetic Reinforced Structures in Highway Engineering

by the Roads and Traffic Authority (NSW), Ground Modification Seminar No 3 Geosynthetics in Road Engineering, University of Technology, Sydney, Australia,

September 1994, 26 p.

Yogendrakumar, M., Bathurst, R.J. and Finn, W.D.L., 1992, Dynamic Response Analysis of a Reinforced Soil Retaining Wall, Journal of Geotechnical Engineering,

ASCE, Vol. 118, No. 8, pp. 1158-1167.

Zarrabi, K., 1979, Sliding of Gravity Retaining Wall During Earthquakes Considering

Vertical Acceleration and Changing Inclination of Failure Surface, Master of Science thesis, Department of Civil Engineering, Massachusetts Institute of Technology,

Cambridge, MA, USA, 140 p.

828

GEOSYNTHETICS INTERNATIONAL

NOTATIONS

Basic SI units are given in parentheses.

acs

ah

au

av

Fc

Fc(max)

Fdyn

Fsta

FScn

=

=

=

=

=

=

=

=

=

FSbot

FSlot

FSpo

FSsc

FSbsl

FSisl

g

H

KA

=

=

=

=

=

=

=

=

=

KAE

kh

kv

L

Lw

MR

m

=

=

=

=

=

=

PA

PAE

rF

=

=

=

=

Sdyn

Sv

=

=

peak horizontal ground acceleration (m/s2)

minimum interface shear strength (N/m)

peak vertical ground acceleration (m/s2)

wall/reinforcement connection capacity (N/m)

maximum wall/reinforcement connection capacity (N/m)

dynamic reinforcement force (N/m)

static reinforcement force (N/m)FS

factor of safety against connection failure between facing units

(dimensionless)

factor of safety against base overturning (dimensionless)

factor of safety against local overturning (dimensionless)

factor of safety against pullout (dimensionless)

factor of safety against interface shear failure (dimensionless)

factor of safety against base sliding (dimensionless)

factor of safety against internal sliding (dimensionless)

gravitational constant (m/s2)

wall height (m)

static earth pressure coefficient calculated using Coulomb earth pressure

theory (dimensionless)

dynamic earth pressure coefficient calculated using Mononobe-Okabe

method (dimensionless)

horizontal seismic coefficient (dimensionless)

vertical seismic coefficient (dimensionless)

base width of reinforced soil zone plus facing column (m)

width of facing column (m)

resisting moment due to weight of facing column (N-m/m)

ratio of moment arm of dynamic active earth force to wall height

(dimensionless)

static active earth force (N/m)

dynamic active earth force (N/m)

dynamic reinforcement force magnification factor

ratio of dynamic reinforcement force to static reinforcement force

(dimensionless)

dynamic interface shear force (N/m)

contributory area (m2/m)

GEOSYNTHETICS INTERNATIONAL

829

Tallow

Yc

Ydyn

Ww

z

Vu

AE

=

=

=

=

=

=

=

Kdyn

Pdyn

Ww

ds

=

=

=

=

=

=

=

=

=

cs

u

=

=

=

=

moment arm of reinforcement force (m)

moment arm of dynamic force from wall base (m)

weight of facing column (N/m)

depth from crest of wall (m)

interface shear capacity (N/m)

orientation of active failure plane from horizontal under dynamic loading

(_)

wall backslope angle (_)

dynamic earth pressure coefficient increment (dimensionless)

dynamic force increment (N/m)

incremental weight of facing column (N/m)

interface friction angle (_)

peak friction angle of soil (_)

angle of geosynthetic-soil interface friction (_)

soil or facing column unit weight (N/m3)

ratio of moment arm of dynamic force increment to wall height

(dimensionless)

inertial force reduction factor for gravity mass in external stability

calculations (dimensionless)

slope of connection strength failure envelope (_)

interface friction angle between facing units (_)

inertia angle (_)

wall inclination angle from vertical (_)

ABBREVIATIONS

AASHTO-AGC-ARTBA:

American Association of State Highway and Transportation OfficialsAssociation of General Contractors-American Road and Transportation

Builders Association

CFEM:

Canadian Foundation Engineering Manual

FHWA:

Federal Highway Administration

NCMA:

National Concrete Masonry Association

830

GEOSYNTHETICS INTERNATIONAL