CHAPTER 5

DESIGN OF ANCHORED SYSTEMS

5.1 INTRODUCTION

The purpose of this chapter is to present analysis procedures for the design of ground anchors and
anchored systems. As previously discussed in chapter 1, design concepts for permanent systems are
also applicable to critical temporary systems. The emphasis of this chapter is on the design of
permanently anchored soldier beam and lagging walls, as these systems are most commonly used for
highway applications. Other systems discussed include temporary support of excavation (SOE)
anchored walls (either sheet-pile or soldier beam and lagging), landslide and slope stabilization
systems, and tiedown structures. For consistency, when concepts and procedures are generally

A general flow chart for anchored wall design (temporary and permanent) is shown in table 4. This
table was developed assuming that an anchored wall has been judged to be a suitable support system
for a specific application. Information on selecting a particular wall type (e.g., anchored walls, soil
nailed walls, mechanically stabilized walls, etc.) can be found in FHWA-SA-96-038 (FHWA, 1997).
Step (1) involves establishing overall geometric requirements for the anchored system and
identifying project requirements and constraints. This step involves developing the wall profile,
locating wall appurtenances such as traffic barriers, utilities, and drainage systems, establishing right-
of-way (ROW) limitations, and construction sequencing requirements. Project requirements and
constraints may significantly affect design, construction, and cost of the wall system and should
therefore be identified during the early stages of the project. Since the information in Step (1) is
required prior to actual wall selection and design, further discussion is limited in this chapter
regarding these issues. Steps (2) through (13) address specific geotechnical and structural
requirements that are addressed when designing an anchored wall.

This chapter focuses on procedures that should be addressed in designing specific components of an
anchored wall. As part of the overall design, the relationship between type of ground, selection of
ground anchors, type of soldier beam, connections (ground anchor/soldier beam, soldier
beam/permanent facing), and type of facing must be considered. Detailed information on these
considerations is not included in this document as decisions related to these considerations are
typically made by the contractor. The engineer, however, should ensure that the specific components
and combinations of components used for the anchored system are consistent with all performance
requirements.

46
Table 4. Typical design steps for an anchored wall (modified after FHWA-RD-81-150, 1982).

Step 1. Establish project requirements including all geometry, external loading

conditions (temporary and/or permanent, seismic, etc.), performance criteria,
and construction constraints.
Step 2. Evaluate site subsurface conditions and relevant properties of in situ soil and
rock.
Step 3. Evaluate design properties, establish design factors of safety, and select level
of corrosion protection.
Step 4. Select lateral earth pressure distribution acting on back of wall for final wall
height. Add appropriate water, surcharge, and seismic pressures and evaluate
total lateral pressure. A staged construction analysis may be required for
walls constructed in marginal soils.
Step 5. Calculate horizontal ground anchor loads and wall bending moments. Adjust
vertical anchor locations until an optimum wall bending moment distribution
is achieved.
Step 6. Evaluate required anchor inclination based on right-of-way limitations,
location of appropriate anchoring strata, and location of underground
structures.
Step 7. Resolve each horizontal anchor load into a vertical force component and a
force along the anchor.
Step 8. Evaluate horizontal spacing of anchors based on wall type. Calculate
individual anchor loads.
Step 9. Select type of ground anchor.
Step 10. Evaluate vertical and lateral capacity of wall below excavation subgrade.
Revise wall section if necessary.
Step 11. Evaluate internal and external stability of anchored system. Revise ground
anchor geometry if necessary.
Step 12. Estimate maximum lateral wall movements and ground surface settlements.
Revise design if necessary.
Step 13. Select lagging. Design walers, facing drainage systems, and connection
devices.

5.2 EVALUATION OF EARTH PRESSURES FOR WALL DESIGN

5.2.1 Introduction

The earth pressure distribution that develops on an anchored wall depends on the magnitude and
distribution of lateral wall deformations. Some relatively flexible nongravity cantilevered walls
(e.g., sheet-pile or soldier beam and lagging walls which are not anchored) can be expected to
undergo lateral deformations sufficiently large to induce active earth pressures for the entire wall
height. For design of these systems, theoretical active earth pressure diagrams using either Rankine
or Coulomb analysis methods can be used.

47
complex and not consistent with the development of a theoretical Rankine or Coulomb earth
pressure distribution. Soil shear strength, wall stiffness, anchor inclination, vertical spacing of the
anchors, and anchor lock-off loads directly influence the wall deformation pattern and the resulting
earth pressures acting on these types of walls. For example, higher than active earth pressures
develop at the upper anchor location since the upper anchor restrains the wall from moving outward
sufficiently to locally cause a reduction of earth pressures to the active state.

This section presents information on and makes recommendations for evaluating earth pressure
distributions used in design of temporary SOE and permanent anchored walls with flexible wall
elements. Methods for evaluating earth pressures for these types of anchored walls include the use
of apparent earth pressure, sliding wedge-type, and limit equilibrium calculations.

5.2.2 Background

Apparent earth pressure diagrams are semi-empirical diagrams that were originally developed by
Terzaghi and Peck (1967) and Peck (1969) to provide loadings for conservative design of struts in
internally braced excavations. Diagrams were developed for homogeneous profiles representing: (1)
drained loadings in sands; (2) undrained loadings in stiff to hard fissured clays; and (3) undrained
loadings in soft to medium clays. Since 1969, modifications to the original diagrams have been
proposed. Two notable modifications that have been incorporated in this manual are described
below:

Henkel (1971) modified the equation used to calculate the maximum earth pressure ordinate
for the Terzaghi and Peck soft to medium clay apparent earth pressure diagram. Henkel
assumed a failure mechanism consistent with deep-seated movements for excavations in soft
to medium clay that had not been previously used by Peck (1969). Backcalculated values of
the active earth pressure coefficient for excavations in which deep-seated movements

In FHWA-RD-97-130 (1998), a variation in the distribution of earth pressure calculated

fissured clay is proposed. The earth pressures for anchored walls with flexible wall elements
are greatly influenced by the prestressing and lock-off procedure used for each anchor (see
section 4.4.4). Earth pressures concentrate at the anchor locations. The apparent earth
pressure diagram for anchored walls in sands and stiff to hard fissured clays requires that the
location of the uppermost and lowermost anchor be known. The distribution of earth
pressure is therefore, in addition to being influenced by excavation depth (as is the case for
the Terzaghi and Peck diagrams), also influenced by the location of the anchors.

The use of apparent earth pressure envelopes has resulted in reasonable estimates of ground anchor
loads and conservative estimates of wall bending moments between anchors for flexible walls
constructed in competent soils. The apparent earth pressure diagrams recommended herein are based
on the Terzaghi and Peck (1967) diagrams for internally braced excavations and research results
obtained from full-scale and model-scale instrumented anchored soldier beam and lagging walls.

48
loads and wall bending moments. They represent an envelope that can be used to develop an
adequate anchored system for the entire history of the excavation, but they do not provide actual
loads that might exist on the wall at any time. Where an assessment of the actual loads on the wall is
required, staged construction analyses such as soil-structure interaction analyses (e.g., beam on
elastic foundation) may be used. Staged construction analyses may also be required where: (1) the
wall is influenced by loadings from nearby foundations; (2) large surcharge loadings need to be
resisted by the wall; or (3) there are preexisting instabilities or planes of weakness in the retained
soil.

Limit equilibrium calculations may be used for evaluating the total load required to stabilize a slope
or excavation in highly stratified soils, profiles for which the potential failure surface is deep-seated
or occurs along weak, well-defined interfaces, and where complicated surcharges are present. Limit
equilibrium calculations may be performed using hand calculation methods such as trial wedge or
using slope stability analysis computer programs. Limit equilibrium calculations are equally as valid
as apparent earth pressure diagrams for evaluating required loads for walls constructed in relatively
homogeneous soil profiles. However, apparent earth pressure diagrams are more expedient to use
and are therefore recommended herein over limit equilibrium calculations. Comparisons between
these two methods are provided in this chapter.

5.2.3 Terzaghi and Peck Apparent Earth Pressure Diagrams

The apparent earth pressure diagrams developed by Terzaghi and Peck (1967) and Peck (1969),
although not recommended herein in their original form, provide the framework for the diagrams that
will be recommended in subsequent sections. These diagrams represent the envelope of pressures
back-calculated from field measurements of strut loads in internally braced excavations. These
diagrams produce conservative design loads, implying that if a strut load would be equivalent to the
calculated load from the apparent pressure diagram at that location, the other strut loads would
necessarily be less than that calculated from the apparent pressure diagram.

The Terzaghi and Peck apparent earth pressure envelopes are rectangular or trapezoidal in shape.
These diagrams are summarized in figure 23. The maximum ordinate of the apparent earth pressure
diagrams in figure 23 is denoted by p. The Terzaghi and Peck envelopes were developed based on
the following factors:

The excavation is assumed to be greater than 6 m deep and relatively wide. Wall
movements are assumed to be large enough so that the full value of the soil shear strength
may be mobilized.

Groundwater is assumed to be below the base of the excavation for sands, and for clays, its
position is not considered important. Specifically, loading due to water pressure was not
considered in these analyses.

The soil mass is assumed to be homogeneous and soil behavior during shearing is assumed
to be drained for sands and undrained for clays, i.e., only short-term loadings are considered.

49
Figure 23. Terzaghi and Peck apparent pressure envelopes (after Terzaghi and Peck, 1967, Soil
Mechanics in Engineering Practice, Reprinted by permission of John Wiley & Sons, Inc.).

The loading diagrams apply only to the exposed portion of the wall and not the portion of the
wall embedded below the bottom of the excavation.

For clays, the apparent earth pressure is related to the stability number, Ns, which is defined as

H
Ns (Equation 9)
Su

where is the total unit weight of the clay soil, Su is the average undrained shear strength of the clay
soil below the base of the excavation, and H is the excavation depth. Standard SI units are:
(kN/m3), Su (kPa), and H (m). As shown in figure 23, two apparent earth pressure envelopes were
developed for clays to account for differences in earth pressures for clays with relatively low Ns
values (i.e., stiff to hard clays) and relatively high Ns values (i.e., soft to medium clays). Using these
diagrams for initial reference, specific recommendations for anchored walls are provided in
subsequent sections.

5.2.4 Recommended Apparent Earth Pressure Diagram for Sands

For sands, the value for KA in figure 23a is given as:

'
KA tan 2 45 (Equation 10a)
2

and the maximum earth pressure ordinate is:

50
p 0.65 K A H (Equation 10b)

where is the effective stress friction angle of the sand. Using this value of lateral earth pressure,
the total lateral earth load from the rectangular apparent earth pressure diagram (figure 23a) for sands
is 0.65 Ka H2. The recommended apparent earth pressure envelope for single level anchored walls
and walls with two or more levels of ground anchors is trapezoidal and is shown in figure 24.

Figure 24. Recommended apparent earth pressure diagram for sands.

51
Unlike the Terzaghi and Peck envelopes, the diagrams recommended herein require that the location
of the upper and lower ground anchors are known in order to construct the apparent earth pressure
diagram. The trapezoidal diagram is more appropriate than the rectangular diagram for the following
reasons:

earth pressures are concentrated at the anchor locations resulting from arching;

earth pressure of zero at the ground surface is appropriate for sands (provided no surcharge
loading is present);

earth pressures increase from the ground surface to the upper ground anchor location; and

for medium dense to very dense sands, earth pressures reduce below the location of the
lowest anchor owing to the passive resistance that is developed below the base of the
excavation.

This diagram is appropriate for both short-term (temporary) and long-term (permanent) loadings in
sands. Water pressures and surcharge pressures should be added explicitly to the diagram to evaluate
the total lateral load acting on the wall.

5.2.5 Recommended Apparent Earth Pressure Diagram for Stiff to Hard Fissured Clays

Temporary Conditions

Although apparent earth pressure diagrams for temporary conditions in stiff to hard fissured clays are
proposed herein, the selection of an earth pressure diagram for design should be based on previous
successful experience with excavations constructed in similar soils. This reliance on previous
experience is particularly important for designing excavation support systems in stiff to hard fissured
clays. Earth pressures in these soils are most influenced by degree of fissuring or jointing in the clay
and the potential reduction in strength with time, not necessarily the shear strength of the intact clay.

Table 5 provides a summary of empirical apparent earth pressure envelopes for stiff to hard clays.
Although several variations of the stiff to hard fissured clay envelope have been used in practice, a
comparison of the envelopes which can be developed using the information in table 5 indicates that
the range of total load is similar for each of the envelopes. The most important observation is that
twice as much load must be resisted by systems that are designed using an envelope based on an
upper range value of the maximum pressure ordinate as compared to systems designed using a lower
range value of the maximum pressure ordinate. The selection of the maximum ordinate value should
therefore be based on previous experience with excavations constructed in similar deposits.

52
Table 5. Summary of trapezoidal apparent pressure envelopes for
temporary excavations in stiff to hard clays.

Reference b1 b2 b3 Range of Total load

maximum pressure
ordinate, p
Terzaghi and Peck 0.25 0.50 0.25 0.2 H - 0.4 H 0.15 H2 - 0.30 H2
(1967)
Schnabel (1982) 0.20 0.60 0.20 0.2 H (1) 0.16 H2
Winter (1990) 0.20 0.60 0.20 0.2 H - 0.32 H(1) 0.16 H2 - 0.26 H2
Ulrich (1989) 0.25 0.50 0.25 0.2 H - 0.4 H 0.15 H2 - 0.30 H2
FHWA-RD-75-130 0 1.0 0 0.15 H - 0.30 H 0.15 H2 - 0.30 H2
(1976)
This work (2) 0.17 (3) 0.66 0.17(4) 0.2 H - 0.4 H 0.17 H2 - 0.33 H2

Notes: (1) Assumes = 19.6 kN/m3

(2) Diagram for multiple levels of ground anchors
(3) Assumes H1 = H/4 (see figure 27)
(4) Assumes Hn+1 = H/4 (see figure 27)

For the Terzaghi and Peck apparent earth pressure diagram for temporary loadings in stiff to hard
fissured clays, (see figure 23b), the maximum ordinate, p, of the diagram ranges from 0.2 H to
0.4 H. The total load for this diagram is therefore 0.15 H2 to 0.30 H2. For an assumed total unit
weight for stiff to hard clay equal to 20 kN/m3, the total load for this diagram is 3H2 to 6H2 where H
is in meters and total load is in kN/m per meter of wall. Ulrich (1989) presented ground anchor load
measurements for seven temporary excavation support walls, five of which were constructed in
overconsolidated soils in the Houston, Texas area. For each excavation, the stability number, Ns,
was less than 4. Measured loads are plotted on figure 25. In all but one case (i.e., Site 3), the
maximum apparent earth pressure ordinate ranges from approximately 0.1 H to 0.25 H. These
maximum pressure ordinates correspond to a total load using the Terzaghi and Peck diagram of
1.5H2 to 3.75H2, respectively. For Site 3, the apparent earth pressure was between 0.25 H and
0.35 H. These maximum pressure ordinates correspond to a total load using the Terzaghi and Peck
diagram of 3.75H2 to 5.25H2, respectively.

53
Reprinted by permission of ASCE).

Winter (1990) presented measured ground anchor loads for a 23-m deep excavation in Seattle.
Primary excavation support soils were heavily overconsolidated silts and clays. Ground anchor
design loads were calculated using a trapezoidal envelope in which the pressure increased linearly
from zero at the top of the wall to the maximum pressure over the upper 20 percent of the wall, and
decreased linearly over the lower 20 percent of the wall back to zero at the base of the excavation.
To evaluate the actual wall pressures, the anchors were locked-off at 50 percent of the design load as
compared to a more common level of 100 percent of the design load. The purpose of the lower lock-
off loads was to create a condition whereby the anchors would be required to resist actual loads, not
just the prestressed loads. Without lower lock-off loads, the actual loads would have to be higher
than the design values to register on the load measuring devices. Figure 26 shows the design
pressure envelope and the recorded loads as a percentage of the design values. The design pressure
envelope has a maximum pressure ordinate equal to 30H psf (H in feet). The actual pressure
envelopes for two test sections were 19H psf and 22H psf (H in feet) indicating that the actual
pressures were 65 to 75 percent of the design values. For the design pressure (i.e., 30H) envelope,
the total load is 24H2 lb/ft per ft of wall (3.77H2 kN/m per meter of wall). The total load from the
19H and 22H actual pressure envelope is 15.2H2 to 17.6H2 lb/ft per ft of wall (2.38 H2 to 2.76 H2
kN/m per meter of wall), respectively.

54
Test Section W11 Test Section E11
0 20 40 60 80 100% 0 20 40 60 80 100%
8
4
0 0
Highest

-10 -10
Highest

-20 -20

-30 -30

-40 -40
Final

-50 -50
Final
-60 -60
Design 30H
19H Design 30H
-70 -70 22H

-80 -80 1Hft psf = .1575 Hm kPa

Publication No. 25, Reprinted by permission of ASCE).

These results indicate that although the total load from apparent earth pressure envelopes using a
maximum ordinate of 0.2 H represents a lower bound value for the Terzaghi and Peck envelope,
measured loads from these actual projects are in reasonable agreement with this lower bound value.
Total loads from apparent earth pressure envelopes using 0.4 H for a maximum pressure ordinate
value are conservative. The recommended apparent earth pressure diagram for temporary
excavations in stiff to hard fissured clays (i.e., Ns < 4) is shown in figure 27. The maximum
ordinate, p, should be consistent with a total load from the diagram of approximately 3H2 to 6H2
kN/m per meter of wall. Some designers have developed similarly shaped, but alternative, apparent
earth pressure diagrams for stiff to hard clays. These alternative apparent earth pressure diagrams
may also be used provided that the total load for the diagram is at least 3H2 kN/m per meter of wall
A minimum total load of 3H2 kN/m per meter of wall is recommended for all cases. If an apparent
earth pressure envelope with a total load of less than or approximately equal to 3H2 to 4H2 kN/m per
meter of wall is proposed for a temporary wall used for a critical application, the owner or engineer
should require that the contractor provide performance data that demonstrate that such an envelope
has been successfully used for anchored systems constructed in similar ground subject to similar
performance requirements.

The apparent earth pressure diagram for stiff to hard clays under temporary conditions should only
be used when the temporary condition is of a controlled short duration and there is no available free
55
water. If these conditions are not met, an apparent earth pressure diagram for long-term (i.e.,
permanent) conditions using drained strength parameters should be evaluated. The permanent
conditions apparent earth pressure diagram for stiff to hard clays is described subsequently.

Figure 27. Recommended apparent earth pressure envelope for stiff to hard clays.

56
Permanent Condition

The original Terzaghi and Peck apparent earth pressure diagram for stiff to hard fissured clays was
developed for temporary loading conditions. This diagram and ones developed based on information
in table 5 have also been used for designing permanent anchored wall systems. There are
difficulties in using earth pressures associated with temporary conditions in stiff to hard fissured
clays for designing permanent walls. Specifically, excavation induces negative excess porewater
pressures in the soil which temporarily cause the soil to possess a greater shear strength than is
available in the long term. Soil behind the wall and in front of the wall (i.e., at the base of the
excavation) experience unloading to which the soil responds by drawing in water, resulting in
softening (i.e., weakening) of the soil with time. Softening in some areas around the wall to a state
of long-term (i.e., drained) equilibrium may occur rapidly after construction. The development of
tension cracks at the surface and the possible presence of sandy or silty layers or cracks and fissures
serve to increase the rate at which soil softening may occur.

Based on the above discussion, earth pressures associated with long-term drained conditions for
excavations in stiff to hard fissured clays may be greater than those computed based on envelopes for
temporary conditions. The total resultant force calculated using a diagram for temporary conditions
can be compared to the total resultant force associated with the recommended apparent earth pressure
envelope for stiff to hard clays using a total resultant force of 0.65KA H2, where KA is based on the
drained friction angle of the clay soil. For most anchored wall applications, the drained friction
angle should correspond to the fully softened friction angle. The larger of the resultant forces from
the two diagrams should be used for design. For example, a fully softened drained friction angle of
approximately 39 results in an equivalent total force to the Terzaghi and Peck envelope using 0.2 H
for the maximum pressure ordinate. A drained friction angle of approximately 22 results in an
equivalent total force to the Terzaghi and Peck envelope with a maximum pressure ordinate of
0.4 H.

5.2.6 Recommended Apparent Earth Pressure Diagram for Soft to Medium Clays

Temporary and permanent anchored walls may be constructed in soft to medium clays (i.e., Ns>4) if
a competent layer for forming the anchor bond zone is within a reasonable depth below the
excavation. Permanently anchored walls are seldom used where soft clay extends significantly
below the base of the excavation.

For soft to medium clays and for deep excavations, the Terzaghi and Peck diagram shown in figure
23c has been used to evaluate apparent earth pressures for design of temporary walls in soft to
medium clays. For this diagram, a total stress active earth pressure coefficient is used:

4Su
KA 1 m (Equation 11)
H

where m is an empirical factor that accounts for potential base instability effects in deep excavations
in soft clays. When the excavation is underlain by deep soft clay and Ns exceeds 6, m is set equal to
0.4. Otherwise, m is taken as 1.0 (Peck, 1969). As will be shown, using the Terzaghi and Peck
diagram with m equal to 0.4 for cases where Ns > 6 may result in an underestimation of loads on the
wall and is therefore not conservative.

57
The Terzaghi and Peck (1967) diagrams did not account for the development of soil failure below
the bottom of the excavation. Observations and finite element studies have demonstrated that soil
failure below the bottom of the excavation can lead to very large movements for temporary retaining
walls in these soft clays. For Ns values greater than 6, relatively large areas of the retained soil near
the base of the excavation are expected to yield significantly as the excavation progresses resulting in
large movements below the excavation, increased support loads on the exposed portion of the wall,
and potential instability of the excavation base. Instead of using m=0.4 in equation 11, an equation
developed by Henkel (1971) should be used directly to obtain KA for use in evaluating the maximum
pressure ordinate for the soft to medium clay apparent earth pressure diagram (figure 23c).

4 Su d 5.14 Sub
KA 1 2 2 1 (Equation 12)
H H H

where d is the depth of the failure surface below the cut, Su is the undrained shear strength of the soil
through which the excavation extends, and Sub is the strength of the soil providing bearing resistance
(figure 28). For the more general case in which there is unloading at the ground surface, Henkel
provided the following solution:

4Su 2 2d H H H Sub 2Su H

KA 1 1 1 5.14 (Equation 13)
H H H (2 x ) 2 d H 2Sub d

Standard SI units are: d(m), Sub (kPa), H (m), and x (m).

Figure 29 shows values of KA calculated using Henkel's method for various d/H ratios. For results
shown in this figure, Su = Sub. Figure 29 indicates that for 4<Ns<6, the Terzaghi and Peck envelope
with m=0.4 is overly conservative relative to Henkel. Also, for Ns<5.14, the Henkel equation is not
valid and apparent earth pressures calculated using m=1.0 in the Terzaghi and Peck envelope are
58
unrealistically low. For the range 4<Ns<5.14, a constant value of KA equal to 0.22 should be used to
evaluate the maximum pressure ordinate for the soft to medium clay apparent earth pressure
envelope (figure 23c). At the transition value between stiff to hard clay and soft to medium clay, i.e.,
Ns=4, the total load using the soft to medium clay apparent earth pressure diagram with KA=0.22 is
0.193 H2. For a total load of 0.193 H2, the maximum pressure ordinate of the Terzaghi and Peck
stiff to hard fissured clay apparent earth pressure diagram is 0.26 H. Information presented on
figures 25 and 26 indicates that a value of 0.26 H for the maximum pressure ordinate results in a
calculated apparent earth pressure diagram that is consistent with measured values. The use of
KA=0.22 for 4<Ns<5.14 for calculating apparent earth pressures therefore represents a rational
transition value between apparent earth pressures for stiff to hard clays (i.e., Ns<4) and for soft to
medium clays where the Henkel solution is valid (i.e., Ns>5.14).

Henkel's method is limited to cases where the clay soils on the retained side of the excavation and
below the excavation subgrade can each be reasonably characterized using a constant value for
undrained shear strength. Where a more detailed shear strength profile is required, limit equilibrium
methods may be used to evaluate the earth pressure loadings on the wall. The use of limit
equilibrium methods to develop earth pressure loadings for walls is described in section 5.7.3.

2.0
Recommended for 4 < Ns < 5.14
1.8 Terzaghi and Peck (m=1.0)

Terzaghi and Peck (m=0.4)

1.6
Henkel d/H=0.1
1.4 Henkel d/H=0.25

Henkel d/H=0.5
1.2
Henkel d/H=1.0
1.0

0.8

0.6

0.4

0.2

0.0
4 5 6 7 8 9 10
STABILITY NUMBER, NS

Figure 29. Value of KA

59
5.2.7 Loading Diagrams for Stratified Soil Profiles

The apparent earth pressure diagrams described above were developed for reasonably homogeneous
soil profiles and may therefore be difficult to adapt for use in designing walls in stratified soil
deposits. A method based on redistributing calculated active earth pressures may be used for
stratified soils. This method should not be used for soil profiles in which the critical potential failure
surface extends below the base of the excavation or where surcharge loading is irregular. This
method is summarized as follows:

Evaluate the active earth pressure acting over the excavation height and evaluate the total
load imposed by these active earth pressures using conventional geotechnical engineering
analysis methods for evaluating the total active earth pressure diagram assuming full
mobilization of soil shear strength. For complicated stratification, irregular ground surface,
or irregular surcharge loading, the lateral force may be evaluated using a trial wedge stability
analysis.

Increase the total load determined above by a factor of 1.3 for anchored soldier beam or
sheet-pile walls. A larger value may be used where strict deformation control is required.
Distribute the factored total force into an apparent pressure diagram using the trapezoidal
distribution shown in figure 24.

Where irregular surcharges or ground surfaces are present or where potential failure surfaces are
deep-seated, limit equilibrium methods usping slope stability computer programs may be used to
calculate earth pressure loadings. These limit equilibrium methods are described in Section 5.7.3.

5.2.8 Sliding Wedge Analysis Method

A sliding wedge force equilibrium method may be used to evaluate the total horizontal load required
to provide stability to a vertical cut. An example failure surface, free body diagram, and force vector
diagram are shown in figure 30 for a wall of height H with a soil behind and in front of the wall
characterized by an effective stress friction angle,
surface passes in front of the anchor bond zone such that full anchor loads contribute to wall
stability. The shear strength is factored by the target factor of safety such that mob=tan-1(tan
Passive resistance is assumed to develop over the wall embedment depth, d. For the assumed failure
surface, an interface friction angle equal to mob may be used to calculate the passive earth
pressure coefficient.

In the analysis, PREQ represents the external horizontal force required to provide stability to the
vertical cut. This force represents the combined resistance provided by the horizontal component of
the anchor forces, T cos i, and the lateral resistance provided by the embedded portion of the wall,
SPH. The assumption that PREQ is horizontal implies that the vertical resistance provided by the
soldier beam, SPV, is equal in magnitude and opposite in sign to the vertical component of the
ground anchor loads, T sin i. This assumption should be verified using the procedures described in

60
section 5.6 to evaluate axial capacity of the wall. The required resisting force, PREQ, is then
calculated as:
2
1 2 1 2 cos( )
PREQ H Kp sin( ) tan( ) (Equation 14)
2 tan( ) tan( )

Figure 30. Force equilibrium method for anchored walls (after FHWA-RD-98-065, 1998).

where all terms are defined in figure 30. The solution is found iteratively by adjusting the angle of
the potential failure surface, , and the wall embedment depth, d, until the greatest PREQ is found.
This load (PREQ) should then be redistributed into an apparent pressure envelope for calculating
ground anchor loads and bending moments in the exposed portion of the wall. Detailed discussion
on the use of this simplified method is provided in FHWA-RD-98-065 (1998).

61
5.2.9 Water Pressures

Permanent anchored soldier beam and lagging walls are typically not designed to resist large water
loads. For these wall systems, drainage from the surface of the retained soil is collected in ditches at
the top of the wall while subsurface water is collected using prefabricated drainage elements placed
between the wall and the permanent facing. Additional information on drainage systems for
anchored walls and slopes is provided in section 5.11.2. For temporary systems, it may be necessary
to resist water forces associated with seepage behind and beneath the wall. A typical flownet for a
retaining wall in homogeneous soil is shown in figure 31. The calculation of porewater pressure may
be simplified by assuming that the head difference (H+i-j) is dissipated uniformly along the shortest
potential flowpath (2d+H-i-j) which runs down the back of the wall and up the front. The porewater
pressure calculated in this manner results in pressures greater than hydrostatic in front of the wall and
less than hydrostatic behind the wall (figure 32).

Figure 31. Flow net for a retaining wall (after CIRIA, 1984).

In figure 32, the porewater pressure at the bottom of the wall, Uf, is equal on either side of the wall.
The value for Uf is given by the following:

2 (d H j) (d i)
Uf w (Equation 15)
2d H i j

The net water pressure acting on the wall is shown on figure 32b. The largest net water pressure
occurs at the level of the water table within the excavation:

2 (d i)
U c (H i j) w (Equation 16)
2d H i j
62
For comparison, the net water pressure for the condition in which there is no seepage is also shown
on figure 32b. In that case, the net pressure is given by:

U n (H i j) w (Equation 17)

Procedures to calculate effective horizontal earth pressures including the effects of seepage are
provided in CIRIA (1984) and FHWA-HI-97-021 (1997).

The effect of special drainage conditions on porewater pressures can only be assessed by using
appropriate seepage flownets. If, for example, the wall acts as a drain, the porewater pressures will
vary significantly with distance behind the wall. The simplified methods shown in figure 32 cannot
be used for this case to calculate the pressures on the back of the wall. However, for normal designs,
it is usually sufficient to use the simple flownet shown in figure 31 and procedures to calculate
porewater pressures shown in figure 32.

Figure 32. Gross and net water pressures across a retaining wall (modified after CIRIA, 1984).

63
5.2.10 Earth Pressures Due To Surface Loads

5.2.10.1 Uniform Surcharge Loads

Surcharge loads are vertical loads applied at the ground surface which are assumed to result in an
assumed uniform increase in lateral stress over the entire height of the wall. The increase in lateral
stress for uniform surcharge loading can be written as:

h Kq s (Equation 18)

where: h is the increase in lateral earth pressure due to the vertical surcharge load, q s is the
vertical surcharge stress applied at the ground surface, and K is an appropriate earth pressure
coefficient. Standard SI units are: h (kPa), K (dimensionless), and qs (kPa). Examples of
surcharge loads for highway wall system applications include: (1) dead load surcharges such as that
resulting from the weight of a bridge approach slab or concrete pavement; (2) live load surcharges
such as that due to traffic loadings; and (3) surcharges due to equipment or material storage during
construction of the wall system. When traffic is expected to come to within a distance from the wall
face equivalent to one-half the wall height, the wall should be designed for a live load surcharge
pressure of approximately 12 kPa (AASHTO, 1996). For temporary walls that are not considered
critical, actual surcharge loads may be evaluated and considered in the design as compared to using
this prescriptive value. Both temporary and permanent wall designs should account for unusual
surcharges such as large material stockpiles and heavy cranes. Calculated lateral pressures resulting
from these surcharges should be added explicitly to the design earth pressure envelope. Loads from
existing buildings need to be considered if they are within a horizontal distance from the wall equal
to the wall height.

5.2.10.2 Point Loads, Line Loads, and Strip Loads

Point loads, line loads, and strip loads are vertical surface loadings which are applied over limited
areas as compared to surcharge loads. As a result, the increase in lateral earth pressure used for wall
system design is not constant with depth as is the case for uniform surcharge loadings. These
loadings are typically calculated using equations based on elasticity theory for lateral stress
distribution with depth (NAVFAC, 1982). Lateral pressures resulting from these surcharges should
be added explicitly to the design earth pressure envelope.

64
5.3 GROUND ANCHOR DESIGN

5.3.1 Introduction

Ground anchors are used for temporary and permanent excavation wall support, slope and landslide
stabilization, and tiedown systems. This section presents procedures that are commonly used to
design a ground anchor and includes a brief discussion on analysis procedures to locate the critical
potential failure surface, calculation of ground anchor loads from apparent earth pressure diagrams,
design of the unbonded and bond lengths of the anchor, allowable load requirements for the
prestressing steel element, and horizontal and vertical spacing and inclination of the anchor.

5.3.2 Location of Critical Potential Failure Surface

The location of the critical potential failure surface must be evaluated since the anchor bond zone
must be located sufficiently behind the critical potential failure surface so that load is not transferred

between the critical potential failure surface and the wall, and is also referred to as the unbonded
length. The unbonded length is typically extended either a minimum distance of H/5, where H is the
height of the wall, or 1.5 m behind the critical potential failure surface. Minimum requirements on
the unbonded length of the anchor and location of the anchor bond zone are described in section
5.3.4.

For walls constructed in cohesionless soils, the critical potential failure surface can be assumed to
extend up from the corner of the excavation at an angle of 45 + 
active wedge). The sliding wedge force equilibrium method presented in section 5.2.8 may also be
used to more accurately evaluate the location of the critical potential surface. Limit equilibrium
methods (see section 5.7.3) and trial wedge analyses may be used for general ground conditions and
can incorporate surcharge loadings and variable soil stratigraphies.

5.3.3 Calculation of Ground Anchor Loads from Apparent Earth Pressure Diagrams

Ground anchor loads for flexible anchored wall applications can be estimated from apparent earth
pressure envelopes. Methods commonly used include the tributary area method and the hinge
method which were developed to enable hand calculations to be made for statically indeterminate
systems. Both methods, when used with appropriate apparent earth pressure diagrams, have
provided reasonable estimates of ground anchor loads and wall bending moments for anchored
systems constructed in competent soils.

The calculations for horizontal ground anchor loads using the tributary area method and the hinge
method are shown in figure 33 for a one-level wall and in figure 34 for a two-level anchored wall.
Both methods, as shown, assume that a hinge (i.e., zero bending moment) develops at the excavation
subgrade and that the excavation subgrade acts as a strut support. This latter assumption is
reasonable for walls that penetrate into competent materials. The maximum bending moment that
controls the design of the wall typically occurs in the exposed portion of the wall, i.e., above the
excavation subgrade.

65
Figure 33. Calculation of anchor loads for one-level wall.

For walls constructed in competent materials, a reaction force, R, is assumed to be supported by the
passive resistance of the soil below the excavation subgrade. The wall must be embedded
sufficiently deep to develop this passive resistance. In this case, the lowest anchor carries only the
tributary area of the apparent pressure diagram, and the reaction force is equivalent to the load from
the apparent pressure diagram from the base of the excavation to the midheight between the base of
the excavation and the lowest anchor. For walls that penetrate weak materials, sufficient passive
capacity below the base of the excavation may not be available to resist the reaction force regardless
of the wall embedment depth. For that case, the lowest anchor may be designed to carry the same
load as defined above for the lowest anchor plus the load corresponding to the reaction force.
Alternatively, soil-structure interaction analyses (e.g., beam on elastic foundation) may be used to
design continuous beams with small toe reactions as it may be overly conservative to assume that all
load is carried by the lowest anchor.

The values calculated using figures 33 and 34 for the anchor loads are the horizontal component of
the anchor load per unit width of wall, Thi. The total horizontal anchor load, Th, is calculated as:

Th Thi s (Equation 19)

where s is the horizontal spacing between adjacent anchors. The anchor load, T, to be used in
designing the anchor bond zone (i.e., the design load) is calculated as:

66
Th
T (Equation 20)
cos

where is the angle of inclination of the anchor below the horizontal. The vertical component of the
total anchor load, Tv, is calculated as:

Tv T sin (Equation 21)

Figure 34. Calculation of anchor loads for multi-level wall.

5.3.4 Design of the Unbonded Length

The minimum unbonded length for rock and soil ground anchors is 4.5 m for strand tendons and 3 m
for bar tendons. These minimum values are intended to prevent significant reductions in load
resulting from seating losses during transfer of load to the structure following anchor load testing.
67
Longer unbonded lengths may be required to: (1) locate the bond length a minimum distance behind
the critical potential failure surface; (2) locate the anchor bond zone in appropriate ground for
anchoring; (3) ensure overall stability of the anchored system; and (4) accommodate long term
movements. In general, the unbonded length is extended a minimum distance of H/5 or 1.5 m
behind the critical potential failure surface (see section 5.3.2) to accommodate minor load transfer to
the grout column above the top of the anchor bond zone.

As a general rule, the anchor bond zone and unbonded zone should be grouted in one stage to
maintain hole stability and to create a continuous grout cover for corrosion protection. However, for
large diameter anchors in which the unbonded length of the anchor extends just behind the critical
potential failure surface, significant strains at the top of the anchor bond zone may cause load
transfer into the grout column above the anchor bond zone. Large diameter anchors have been
grouted in two stages (two stage grouting). With two stage grouting, the anchor bond length is
grouted (Stage 1) and the anchor is tested. The unbonded length portion of the drill hole is then
grouted (Stage 2) after the anchor is tested. The two-stage procedure is not recommended since local
collapse of the ground can occur which will compromise the corrosion protection provided by the
grout.

5.3.5 Compression Anchors

Compression anchors are anchors in which the grout body in the bond length is, at least partially,
loaded in compression. For a typical tension ground anchor (see figure 1), the anchor bond length
and tendon bond length coincide. For these anchors, load is transferred first at the top of the anchor
and, with continued loading, progresses downward to the bottom of the anchor. For single-stage
grouted tension anchors, because load is first transferred to the top of the anchor bond zone, there is

structure and the assumed failure plane. This is especially a concern for large diameter anchors
installed in some cohesive soils for which relatively large residual movements are required to
develop bond at the grout/ground interface.

Two types of compression anchors have been used. These include: (1) a ground anchor fitted with
an end plate (figure 35a); and (2) a composite design where the top of the tendon bond length is
extended a certain distance into the anchor bond length (figure 35b). During stressing, the entire
grout column for the endplate compression anchor is loaded in compression whereas for the
composite design, the portion of the anchor grout located above the top of the tendon bond length is
loaded in compression. The use of compression type anchors minimizes the load transferred above

levels of compressive strain in the grout body. Strains should be within tolerable limits to minimize
the potential for the grout to fail due to compression loading. Where compression anchors are to be
used for a permanent application, a predesign test program may be warranted unless the behavior and
satisfactory performance of the proposed compression anchor has been verified through prior
experience or research results.

Compression anchors are not commonly used for small diameter anchors in cohesionless deposits,
but may be used for large diameter anchors in cohesive soils. In cohesive soils, composite design
compression anchors are typically designed with a tendon bond length equal to 50 to 100 percent of
the anchor bond length.

68
Figure 35. Types of compression anchors.

5.3.6 Design of the Anchor Bond Length

Estimates of load transfer capacity in the anchor bond length are typically based on previous field
experience. When estimating capacities using previous field results, potential variations in capacity
due to differing installation and grouting methods must be considered. In a given soil deposit, the
69
actual capacity achieved in the field will depend on the method of drilling including quality of drill
hole cleaning and period of time that the drill hole is left open, the diameter of the drill hole, the
method and pressure used in grouting, and the length of the anchor bond zone. Except for certain
minimum values, the selection of these items should be left to the discretion of the specialty anchor
contractor. The main responsibility for the designer is to define a minimum anchor capacity that can
be achieved in a given ground type. Therefore, estimation of anchor capacity should be based on the
simplest commonly installed anchor, i.e., the straight shaft gravity-grouted anchor. Estimates made
assuming this anchor will be installed will produce a design capacity which may confidently be
achieved while allowing specialty contractors to use more effective and/or economical anchoring
methods to achieve the specific capacity. The design capacity of each anchor will be verified by
testing before accepting the anchor.

Many projects have been completed using small diameter, straight shaft gravity-grouted anchors.
Because of the similarity of many projects, some fairly typical anchor characteristics can be
summarized. These are intended to provide a range of typical design values to engineers who are
unfamiliar with anchor design.

Design Load Between 260 kN and 1160 kN: Anchor tendons of this capacity can be handled
without the need for unusually heavy or specialized equipment. In addition, stressing
equipment can be handled by one or two workers without the aid of mechanical lifting
equipment. The drill hole diameter is generally less than 150 mm, except for hollow stem
augered anchors that are typically approximately 300 mm in diameter.

Total Anchor Length Between 9 and 18 m: Because of geotechnical or geometrical

requirements, few anchors for walls or for tiedown structures are less than 9 m long. A
minimum unbonded length of 3 m for bar tendons and 4.5 m for strand tendons should be
adopted. These minimum unbonded lengths are required to avoid unacceptable load
reduction resulting from seating losses during load transfer and prestress losses due to creep
in the prestressing steel or the soil.

Ground Anchor Inclination Between 10 and 45 degrees: Ground anchors are commonly
installed at angles of 15 to 30 degrees below the horizontal although angles of 10 to 45
degrees are within the capabilities of most contractors. Regardless of the anchor inclination,
the anchor bond zone must be developed behind potential slip surfaces and in soil or rock
layers that can develop the necessary design load. Steep inclinations may be necessary to
avoid underground utilities, adjacent foundations, right-of-way constraints, or weak soil or
rock layers. Anchors should be installed as close to horizontal as possible to minimize
vertical loads resulting from anchor lock-off loads, however grouting of anchors installed at
angles less than 10 degrees is not common unless special grouting techniques are used.

For a specific project, the first step in estimating the minimum allowable capacity is to assume a
maximum anchor bond length. In the case of a site with no restrictions on right-of-way, a 15-degree
inclination of the anchor should be assumed with a bond length of 12 m in soil or 7.5 m in rock.
Anchors founded in soil and rock should be designed assuming the entire embedment is in soil, i.e.,
assume a bond length equal to 12 m. The bond lengths at sites with more restricted right-of-way may
be evaluated assuming an anchor inclination of 30 degrees and that the bond length is equal to the
distance from the end of the unbonded length to within 0.6 m of the right-of-way line. When using

70
these assumptions to develop a preliminary estimate of the anchor bond length, it must be verified
that for the required excavation height the minimum unbonded length can be developed.

Soil Anchors

For the purposes of preliminary design, the ultimate load transferred from the bond length to the soil
may be estimated for a small diameter, straight shaft gravity-grouted anchor from the soil type and
density (or SPT blowcount value) (table 6). The maximum allowable anchor design load in soil may
be determined by multiplying the bond length by the ultimate transfer load and dividing by a factor
of safety of 2.0.

Table 6. Presumptive ultimate values of load transfer for preliminary design of small diameter
straight shaft gravity-grouted ground anchors in soil.

Soil type Relative density/Consistency Estimated ultimate transfer load

(SPT range)(1) (kN/m)

Loose (4-10) 145

Sand and Gravel Medium dense (11-30) 220
Dense (31-50) 290
Sand Loose (4-10) 100
Medium dense (11-30) 145
Dense (31-50) 190
Sand and Silt Loose (4-10) 70
Medium dense (11-30) 100
Dense (31-50) 130
Silt-clay mixture with low Stiff (10-20) 30
plasticity or fine micaceous sand Hard (21-40) 60
or silt mixtures

Note: (1) SPT values are corrected for overburden pressure.

Anchor bond lengths for gravity-grouted, pressure-grouted, and post-grouted soil anchors are
typically 4.5 to 12 m since significant increases in capacity for bond lengths greater than
approximately 12 m cannot be achieved unless specialized methods are used to transfer load from the
top of the anchor bond zone towards the end of the anchor. For anchor bond zones that function in
tension, initial load increments transferred to the anchor bond zone are resisted by the soil near the
top of the anchor bond zone as strains occur in the upper grout body (figure 36). As additional
increments of load are transferred to the anchor bond zone, the strains in the top of the anchor bond
zone may exceed the peak strain for strain sensitive soils. In that case, the bond stress begins to
decrease at the top and the peak strain shifts down the anchor body. In strain sensitive soils, the
shape of the stress-strain diagram will determine the actual bond length where significant load is
mobilized. Attempts to mobilize larger portions of the bond length will result in small increases in

71
capacity as residual load transfer values develop at the top and the peak value shifts towards the
bottom.

Figure 36. Mobilization of bond stress for a tension anchor.

Pressure grouting in cohesionless soils significantly increases the normal stresses acting on the grout
body (i.e., increases confinement). Small increases may also be observed in the effective diameter of
the anchor bond zone, but capacity estimates should be based on the as-drilled hole diameter. A
range of ultimate bond stress values that have been measured for gravity-grouted and pressure-
grouted soil anchors is provided in table 7 to show variation in field measured ultimate values.
When reviewing ultimate bond stress values such as those presented in table 7, it is important to
recognize that large bond stress values do not necessarily imply a correspondingly large anchor
capacity per unit length. For example, a hollow stem augered anchor can develop more capacity per
unit length than a small diameter, post-grouted anchor due primarily to the anchor diameter not the
bond stress value.

Pressure grouting can be effective in increasing capacity in cohesive soils, however, post-grouting is
a more effective means of increasing capacity in cohesive soils. Post grouting increases the radial
stresses acting on the grout body and causes an irregular surface to be developed around the bond
length that tends to interlock the grout and the ground. It is difficult to predict load capacity in post-
grouted anchors owing to the complexity of the grouting procedures used, however, post-grouting of
ground anchors in cohesive soil can increase load capacity of a straight shaft anchor by 20 to 50
percent or more per phase of post-grouting with three phases being the common limit (PTI, 1996).

72
Table 7. Presumptive average ultimate bond stress for ground/grout interface along anchor bond zone (after PTI, 1996).

Rock Cohesive Soil Cohesionless Soil

Rock type Average ultimate Anchor type Average ultimate Anchor type Average ultimate
bond stress bond stress bond stress
(MPa) (MPa) (MPa)
Granite and basalt 1.7 - 3.1 Gravity-grouted anchors 0.03 - 0.07 Gravity-grouted anchors 0.07 - 0.14
(straight shaft) (straight shaft)
Dolomitic limestone 1.4 - 2.1 Pressure-grouted anchors Pressure-grouted anchors
(straight shaft) (straight shaft)
Soft limestone 1.0 - 1.4 Soft silty clay 0.03 - 0.07 Fine-med. sand, 0.08 - 0.38

Slates and hard shales 0.8 - 1.4 Silty clay 0.03 - 0.07 0.11 - 0.66
(w/gravel), med.
dense
Soft shales 0.2 - 0.8 Stiff clay, med. to 0.03 - 0.10 0.25 - 0.97
high plasticity (w/gravel), dense -
very dense
Sandstones 0.8 - 1.7 Very stiff clay, med. 0.07 - 0.17 Silty sands 0.17 - 0.41
to high plasticity
Weathered 0.7 - 0.8 Stiff clay, med. 0.10 - 0.25 Dense glacial till 0.30 - 0.52
Sandstones plasticity
Chalk 0.2 - 1.1 Very stiff clay, med. 0.14 - 0.35 Sandy gravel, med. 0.21 - 1.38
plasticity dense-dense
Weathered Marl 0.15 - 0.25 Very stiff sandy silt, 0.28 - 0.38 Sandy gravel, dense- 0.28 - 1.38
med. plasticity very dense
Concrete 1.4 - 2.8

Note: Actual values for pressure-grouted anchors depend on the ability to develop pressures in each soil type.
73
Rock Anchors

For rock anchors, typical bond lengths range from 3 to 10 m with a minimum of 3 m. The ultimate
load transferred from the bond length to competent sound rock may be estimated from the rock type
(table 8). Lower values may be recommended after input from a geologist especially if the rock mass
strength is controlled by discontinuities. The maximum allowable anchor design load in competent
rock may be determined by multiplying the bond length by the ultimate transfer load and dividing by
a factor of safety of 3.0. This relatively high value of the factor of safety (compared to that for soil)
is used to account for uncertainties associated with potential discontinuities in the rock mass such as
joints, fractures, and clay-filled fissures. In weak rocks such as clay shales, bond stress transfer is
relatively uniform as compared to bond stress transfer in more competent rock. These weak rocks

varying from 0.5 to 5.0 MPa. Design values for evaluating anchor bond lengths in these materials
should use a factor of safety of 2.0 on the ultimate load transfer value.

Table 8. Presumptive ultimate values of load transfer for

preliminary design of ground anchors in rock.

Rock type Estimated ultimate transfer load

(kN/m)
Granite or Basalt 730

Dolomitic Limestone 580

Soft Limestone 440

Sandstone 440

Slates and Hard Shales 360

Soft Shales 150

Typical ranges of ultimate bond stress values for the rock/grout interface which have been measured
are provided in table 7. Alternatively, PTI (1996) suggests that the ultimate bond stress between
rock and grout can be approximated as 10 percent of the unconfined compressive strength of the rock
up to a maximum value for ultimate bond stress of 3.1 MPa.

In the calculation of bond length, the implicit assumption is that the bond at the rock-grout interface
is mobilized uniformly. This is unlikely to be the case unless the anchor bond zone is formed in soft
or weak rock. For conditions where the ratio of the elastic modulus of the grout to the elastic
modulus of the rock is less than one (e.g., in competent rock), load is transferred from the tendon to
the rock only in the upper 1.5 to 3 m of the anchor bond zone; any additional length of anchor bond
zone may be considered to provide an additional margin of safety. Therefore, use of average bond
stress values such as those provided in table 7 may result in calculated bond lengths significantly
greater than that which is required to resist the design load.

74
Predesign and Preproduction Load Testing

Predesign load tests are occasionally performed to evaluate ultimate anchor load-carrying capacity
and/or creep behavior of anchors installed in creep susceptible soils. When the capacity of
individual anchors is critical to the design, it may be desirable to install and test several test anchors.
Predesign load tests may be performed for cases where the required capacity of the anchors exceeds
local experience or the required construction method is unusual. In general, predesign load tests are
not commonly used and when they are conducted they are performed as part of a separate contract
that is paid for by the owner.

Anchors used for predesign load tests are generally not incorporated into the final structure as load
carrying elements because of the damage that may be induced by the high testing loads required to
evaluate ultimate anchor capacity. If possible, the anchors should be fabricated and installed exactly
as planned for the production anchors. If testing loads will exceed 80 percent of SMTS of the
production anchors, additional tendon capacity should be provided (i.e., increase the number of
strands or use larger bar diameter). Procedures used for a predesign testing program are provided in
appendix D. The objective of most predesign test programs is to establish the anchor load at which
the creep rate becomes unacceptable. Complete documentation of a predesign test program for the
I90 project in Seattle, Washington is contained in FHWA-DP-90-068-003 (1990). In general,
however, predesign load testing test programs are rarely executed due to time and cost factors.

Preproduction anchor testing programs, which can provide similar information concerning
acceptable anchor loads, are commonly performed. With a preproduction testing program, the
contractor performs performance tests on several anchors. Performance tests (see section 7.3.2)
involve incremental loading and unloading of an anchor in progressively increasing load increments
to a maximum test load equal to 133 percent of the design load. Extended creep tests (see section
7.3.4) are commonly used in the preproduction testing program to evaluate the creep behavior of the
anchor at all test loads from 25 to 133 percent of the design load. The advantages of preproduction
load testing as compared to predesign load testing includes: (1) less expensive since contractor only
mobilizes to site on one occasion; (2) less time consuming (e.g., one day) when compared to
predesign testing (e.g., five days); and (3) ability to duplicate ground conditions for production
anchors. The results of the early-on performance tests carried out as part of a preproduction load
testing program may be used to verify anchor bond zone load transfer rates or as a means to optimize
wall design through use of a higher load transfer rate as compared to the load transfer rate used to
develop the original design.

5.3.7 Spacing Requirements for Ground Anchors

Each ground anchor in an anchored system is commonly designed assuming that the anchor carries a
tributary area of load based on the horizontal and vertical spacing between adjacent anchors. The
size and strength of the anchor tendon, drilling and grouting procedures, and diameter and length of
the anchor are selected to ensure that the ground anchor can carry this load throughout its service
life. The horizontal and vertical spacing of the ground anchors will vary depending on project
specific requirements and constraints, which may include: (1) necessity for a very stiff system (i.e.,
closely spaced anchors) to control lateral wall movements; (2) existing underground structures that
may affect the positioning and inclination of the anchors; and (3) type of vertical wall elements
selected for the design.

75
The vertical position of the uppermost ground anchor (i.e., the ground anchor closest to the ground
surface) should be evaluated considering the allowable cantilever deformations of the wall. The
vertical position of the uppermost anchor must also be selected to minimize the potential for
exceeding the passive capacity of the retained soil during anchor proof and performance load testing.
During load testing, permanent anchors are typically loaded to 133 percent of the design load
resulting in movement of the wall into the retained ground. If the design load for the uppermost
ground anchor is relatively large, as is the case where large surcharge or landslide loads must be
resisted, or if the soils are disturbed or relatively weak, the passive capacity of the soil may be
exceeded during load testing. If the passive capacity is exceeded, the soldier beams or sheet pile will
move excessively into the retained ground; for soldier beam wall systems, the timber lagging may
bend and crack excessively. A method to check the passive capacity of the soil at the location of the
uppermost anchor is provided in section 5.11.4.

For ground anchors installed in soil, a minimum overburden of 4.5 m over the center of the anchor
bond zone is required (figure 37a). This is required to prevent grout leakage during installation of
pressure grouted anchors and to prevent heave at the ground surface resulting from large grouting
pressures. For gravity-grouted anchors, the minimum overburden criterion is required to provide the
necessary soil overburden pressure to develop anchor capacity.

Figure 37. Vertical and horizontal spacing requirements for ground anchors.

The maximum horizontal spacing between anchors is based on allowable individual ground anchor
loads and flexural capacity of individual soldier beams or sheet pile sections. Typical horizontal
spacing for soldier beams is 1.5 m to 3 m for driven soldier beams and up to 3 m for drilled-in
soldier beams. The minimum horizontal spacing between anchors shown in figure 37b ensures that
group effects between adjacent ground anchors are minimized and that anchor intersection due to
drilling deviations is avoided. Group effects reduce the load carrying capacity of individual ground
anchors.

76
5.3.8 Selection of Prestressing Steel Element

The prestressing steel element of the tendon (i.e., strand or bar) must be capable of safely
transmitting load in the anchor bond zone to the structure without tendon breakage. For the design
load and the lock-off load, separate factors of safety are applied with respect to the potential failure
mechanism of tendon breakage. The design load shall not exceed 60 percent of the specified
minimum tensile strength (SMTS) of the prestressing steel. The lock-off load shall not exceed 70
percent of the SMTS and the maximum test load shall not exceed 80 percent of the SMTS.

For example, if the maximum test load is 133 percent of the design load, then the ground anchor
should be selected based on a maximum allowable design load of (0.8/1.33) SMTS or 0.6 SMTS. If
the maximum test load is 150 percent of the design load, then the maximum allowable design load is
(0.8/1.5) SMTS or 0.53 SMTS.

Dimensions and strengths of bars and strands commonly used in the U.S. for highway applications
are provided in table 9 and table 10, respectively. Larger size strand tendons (i.e., strand tendons
with more strands than those shown in table 10) are available for applications requiring greater
ground anchor design loads. Information on 13-mm diameter strand or Grade 250 (metric 1725)
strand can be found in ASTM A416.

Table 9. Properties of prestressing steel bars (ASTM A722).

Steel Nominal Ultimate Nominal cross Ultimate Prestressing force

grade diameter stress section area strength
fpu Aps fpu Aps
0.8 fpu Aps 0.7 fpu Aps 0.6 fpu Aps
(ksi) (in.) (ksi) (in.2) (kips) (kips) (kips) (kips)
1 150 0.85 127.5 102.0 89.3 76.5
1-1/4 150 1.25 187.5 150.0 131.3 112.5
150 1-3/8 150 1.58 237.0 189.6 165.9 142.2
1-3/4 150 2.66 400.0 320.0 280.0 240.0
2-1/2 150 5.19 778.0 622.4 435.7 466.8
1 160 0.85 136.0 108.8 95.2 81.6
160 1-1/4 160 1.25 200.0 160.0 140.0 120.0
1-3/8 160 1.58 252.8 202.3 177.0 151.7
(ksi) (mm) (N/mm2) (mm2) (kN) (kN) (kN) (kN)
26 1035 548 568 454 398 341
32 1035 806 835 668 585 501
150 36 1035 1019 1055 844 739 633
45 1035 1716 1779 1423 1246 1068
64 1035 3348 3461 2769 2423 2077
26 1104 548 605 484 424 363
160 32 1104 806 890 712 623 534
36 1104 1019 1125 900 788 675

77
Table 10. Properties of 15-mm diameter prestressing steel strands
(ASTM A416, Grade 270 (metric 1860)).

Number of 15-mm Cross section area Ultimate Prestressing force

diameter strands strength
0.8 fpuAps 0.7 fpuAps 0.6 fpuAps
2 2
(in. ) (mm ) (kips) (kN) (kips) (kN) (kips) (kN) (kips) (kN)
1 0.217 140 58.6 260.7 46.9 209 41.0 182 35.2 156
3 0.651 420 175.8 782.1 140.6 626 123.1 547 105.5 469
4 0.868 560 234.4 1043 187.5 834 164.1 730 140.6 626
5 1.085 700 293.0 1304 234.4 1043 205.1 912 175.8 782
7 1.519 980 410.2 1825 328.2 1460 287.1 1277 246.1 1095
9 1.953 1260 527.4 2346 421.9 1877 369.2 1642 316.4 1408
12 2.604 1680 703.2 3128 562.6 2503 492.2 2190 421.9 1877
15 3.255 2100 879.0 3911 703.2 3128 615.3 2737 527.4 2346
19 4.123 2660 1113.4 4953 890.7 3963 779.4 3467 668.0 2972

The type and size of the anchors should be evaluated prior to design of the anchor bond zone because
the required hole diameter varies as a function of the tendon size. Table 11 can be used to estimate
the minimum trumpet opening for strand or bar tendons.

Table 11. Guidance relationship between tendon size and trumpet opening size.

Tendon type Minimum suggested trumpet opening size (mm)

Class II corrosion protection Class I corrosion protection
Number of 15-mm diameter strands
4 102 150
7 115 165
9 127 178
11 140 191
13 153 203
17 165 216
Bar diameter (mm)
26 64 89
32 70 95
36 76 102

5.4 WALL DESIGN BASED ON LATERAL PRESSURES

5.4.1 Design of Soldier Beams and Sheet-Piles

Anchored soldier beam and lagging walls and sheet-pile walls are designed to resist lateral loads
resulting from apparent pressure envelopes including appropriate surcharges, water forces, and
seismic forces. Figure 38 illustrates the method used to calculate wall bending moments for single-
level and multi-level walls for the exposed portion of the wall using the hinge method. The exposed
portion of the wall refers to the height of wall between the ground surface and the bottom of the
excavation. Figure 39 shows the equations that may be used to calculate wall bending moments for
78
single-level and multi-level walls using the tributary area method. For walls constructed in
competent soils such as most sands and stiff clays, the maximum bending moment, Mmax, occurs in
the exposed portion of the wall. For walls that penetrate deep deposits of weak material, the
maximum bending moment may occur in the embedded portion of the wall. The embedded portion
of the wall refers to the length of wall that is below the base of the excavation. Bending moment
calculation for the embedded portion of the wall is provided in section 5.5.

Figure 38. Calculation of wall bending moments using hinge method.

Selection of an appropriate wall section is based on the calculated maximum bending moment in the
vertical wall element. The negative bending moment calculated at the location of the first anchor is
evaluated by summing moments about the first anchor location. The vertical wall elements are
commonly assumed to be continuous between each support location. The maximum positive
bending moment between each ground anchor is, for the tributary area method, assumed equal to
1/10 p 2 where p is the maximum ordinate of the apparent pressure envelope and  is the vertical
spacing between adjacent anchors. For the hinge method, the maximum positive bending moment
between each ground anchor corresponds to the point of zero shear. These methods provide
conservative estimates of the calculated bending moments, but may not accurately predict the
specific location. For continuous sheet-pile walls, the bending moment per unit of wall is used to
select an appropriate sheet-pile section. To evaluate the maximum bending moment for design of a

79
soldier beam, the maximum bending moment per unit of wall calculated from figure 38 and 39 is
multiplied by the center-to-center spacing of the soldier beams.

Figure 39. Calculation of wall bending moments using tributary area method.

80
For permanent walls and temporary walls that are considered critical, an allowable bending stress in
the soldier beam, Fb, of 0.55 Fy, where Fy is the yield stress of the steel, is recommended. Steel
sheet-pile and soldier beams are commonly either Grade 36 (Fy = 248 MPa) or Grade 50 (Fy = 345
MPa). For temporary SOE walls, a 20 percent increase in the allowable stress may be allowed for
positive wall bending moments between anchor locations; no allowable stress increase is
recommended for negative wall bending movements at the anchor locations. The required section
modulus Sreq, is calculated as:

M max
Sreq (Equation 22)
Fb

Standard SI units are S(mm3), Mmax (kN-m), and Fb (MPa). In most cases, several available steel
sections will typically meet this requirement. The actual wall section selected will be based on
contractor/owner preference, cost, constructability, and details of the anchor/wall connection.

When designing permanent anchored walls in relatively uniform competent materials, it is usually
only necessary to check the final stage of construction provided that: (1) the ground can develop
adequate passive resistance below the excavation to support the wall; (2) apparent earth pressure
diagrams are used to assess the loading on the wall; and (3) there is minimal over excavation below
each anchor level (FHWA-RD-97-130, 1998). For cases where there are large concentrated
surcharges or berms at the ground surface, it is prudent to check wall bending moments for the initial
cantilever stage (i.e., stage just prior to installation and lock-off of uppermost anchor).

Where the final excavation height is not the most critical condition, designers commonly use a
staged construction analysis where the maximum wall bending moment, wall deflections, and wall
embedment depth are evaluated for several stages of construction. An analysis is required for this
case since the maximum bending moment may occur at an intermediate stage of construction (i.e.,
before the final excavation depth is reached). Intermediate construction stages may be critical when:
(1) triangular earth pressure diagrams are used to design the wall; (2) the excavation extends
significantly below an anchor level prior to stressing that anchor; (3) a cutoff wall is used to maintain
the water level behind the wall; (4) the soil below the bottom of the excavation is weak resulting in
active earth pressures that are greater than available resistance provided by the toe of the wall; and
(5) structures are located near the wall.

5.4.2 Design of Lagging for Temporary Support

The thickness of temporary timber lagging for soldier beam and lagging walls is based primarily on
experience or semi-empirical rules. Table 12 presents recommended thicknesses of construction
grade lumber for temporary timber lagging. For temporary SOE walls, contractors may use other
lagging thicknesses provided they can demonstrate good performance of the lagging thickness for
walls constructed in similar ground.

Permanent timber lagging has been used in lieu of a concrete face to carry permanent wall loads. For
permanent applications, the timber grade and dimensions should be designed according to structural
guidelines. Several problems may exist for permanent timber lagging including: (1) need to provide
fire protection for the lagging; (2) limited service life for timber; and (3) difficulty in providing

81
Table 12. Recommended thickness of temporary timber lagging (after FHWA-RD-75-130, 1976)

Soil Description Unified Soil Depth Recommended thickness of lagging (roughcut) for clear spans of:
Classification (m)
1.5 m 1.8 m 2.1 m 2.4 m 2.7 m 3.0 m
COMPETENT
Silt or fine sand and silt ML, SM-ML
SOILS
above water table
Sands and gravels GW, GP, GM, 0-8 50 mm 75 mm 75 mm 75 mm 100 mm 100 mm
(medium dense to dense) GS, SW, SP,
SM
Clays (stiff to very stiff); CL, CH 8 - 18 75 mm 75 mm 75 mm 100 mm 100 mm 125 mm
non-fissured

Clays, medium CL, CH

consistency and H <5
Su
DIFFICULT
Sand and silty sand (loose) SW, SP, SM
SOILS
Clayey sands (medium SC 0-8 75 mm 75 mm 75 mm 100 mm 100 mm 125 mm
dense to dense) below
water table
Clay, heavily CL, CH 8 - 18 75 mm 75 mm 100 mm 100 mm 125 mm 125 mm
overconsolidated, fissured

Cohesionless silt or fine ML, SM-SL

sand and silt below water
table
POTENTIALLY
Soft clays H >5 CL, CH 0-5 75 mm 75 mm 100 mm 125 mm --------- --------
DANGEROUS
SOILS Su

Slightly plastic silts below ML 5-8 75 mm 100 mm 125 mm 150 mm --------- --------
water table
Clayey Sands (loose), SC 8-11 100 mm 125 mm 150 mm --------- --------- ---------
below water table
Notes: 1) In the category of "potentially dangerous soils", use of soldier beam and lagging wall systems is questionable.
2) The values shown are based on construction grade lumber.
3) Local experience may take precedence over recommended values in this table.

82
corrosion protection to the ground anchor. Additional information on design of timber lagging for
permanent facings is provided in section 5.6.6 of AASHTO (1996). As previously mentioned,
concrete lagging is not recommended for anchored walls due to construction difficulties in top-down
placement of the lagging.

5.4.3 Design of Wales and Permanent Facing

For anchored walls, wales and permanent facing should be designed to resist apparent earth
pressures, surcharges, water pressures, and seismic pressures. Maximum bending moments in wales
and permanent facings can be estimated using table 13.

Table 13. Maximum design bending moments for wales and

permanent facing (after AASHTO, 1996).
Support and soil condition Maximum moment
in a 1 m height

Simple span p 2
/8
No soil arching (e.g., soft cohesive soils; rigid concrete facing
placed tightly against soil)
Simple span p 2
/12
Soil arching (e.g., granular soil or stiff cohesive soil with flexible
facing; rigid facing where space is available to allow in place soil
to arch)
Continuous facing p 2
/10
No soil arching (e.g., soft cohesive soils; rigid concrete facing
placed tightly against soil)
Continuous facing p 2
/12
Soil arching (e.g., granular soil or stiff cohesive soil with flexible
facing; rigid facing where space is available to allow in place soil
to arch)

Note: p = maximum ordinate of the total pressure envelope along span

= span between supports

Permanent facings that are cast-in-place (CIP) are typically 200 to 300 mm thick. This thickness will
typically ensure that the wall is structurally sound and allow for some deviations in soldier beam
placement. Significant deviations, however, in soldier beam alignment may require that additional
concrete in excess of that required for the nominal thickness of the wall be used so that the finished
wall face is properly aligned. Precast concrete facing may be cost-effective if there is a local
fabricator and if there is adequate on-site storage. Precast panels are designed as simple spans
between the soldier beams.

Wale design bending moments will depend on the fixity of the wale/soldier beam connection (i.e.,
shear or full moment connection). Bending moments in wales that extend over less than three spans
should be calculated as p 2/8. Three spans or more should be considered continuous and should be
designed using a maximum bending moment of p 2/10. External wales are not commonly used in
permanent applications due to aesthetics, corrosion protection requirements, and other factors
associated with the protrusion of the anchors. Internal wales (i.e., between the flanges of adjacent
soldier beams) have been used in situations where replacement anchors were required.

83
5.5 LATERAL CAPACITY OF EMBEDDED PORTION OF WALL

5.5.1 General

Anchored walls derive support from the ground anchors installed above the final excavation grade
and from passive soil resistance provided over the depth of the embedded portion of the wall. The
wall elements are subjected to various lateral loading conditions depending on the stage of
construction. Prior to stressing the first anchor, the wall acts as a nongravity cantilevered wall and
all resistance is provided by passive soil resistance along the embedded portion of the wall. After
installation of the first anchor and during subsequent excavations for lower anchors, the wall
embedment provides temporary resistance for the unanchored height. At final height, the anchors
carry most of the load from above the base of the excavation while the embedded portion of the wall
is designed to carry loads associated with the lower portion of the apparent pressure envelope (i.e.,
the reaction force R on previous figures) and active earth pressure loads acting along the back side of
the embedded portion of the wall.

The overall stability of an anchored wall system and the stress level developed within the wall
elements depend on the relative stiffness of the wall, the depth of wall penetration, and the soil
strength and stiffness. Figure 40 shows the general relationship between depth of penetration, lateral
earth pressure distribution, and deflected wall shape for an anchored wall. Case (a) refers to a

insufficient to prevent lateral deflection and rotation at point D. Designs based on free earth support
conditions assume that the soil in front of the wall is incapable of producing effective restraint to the
extent necessary to induce negative bending moments. The wall element is extended just deep
enough to assure stability.

Cases (b), (c), and (d) in figure 40 show the effect of increasing the depth of wall penetration. In
case (b) and (c), the passive pressures are sufficient to prevent lateral deflection at point D, but
rotation at the bottom of the wall still occurs. For case (d), passive pressures have developed
sufficiently on both sides of the wall to prevent both lateral deflection and rotation at point D. Case

5.5.2 Evaluation of Ultimate Passive Resistance

5.5.2.1 Soldier Beam and Lagging Walls

The passive side of the embedded portion of an anchored wall (i.e., the excavation side) must resist
the lateral load resulting from the reaction force at the base of the excavation, R, with an adequate
factor of safety. The passive resistance for walls with discrete elements (i.e., soldier beams) below
excavation subgrade has been typically evaluated using relationships developed by Broms (1965) for
laterally loaded piles (figure 41). In cohesionless soils and for drained conditions in cohesive soils,
passive resistance at depth is assumed to be developed over three times the soldier beam width, b,
with a magnitude determined using the Rankine passive earth pressure coefficient. In cohesive soils,
passive resistance is assumed to develop over one soldier beam width and to be constant over most
of the beam depth with a magnitude of nine times the soil undrained shear strength. As shown in

84
figure 41c, no passive resistance is assumed to develop over a depth of 1.5 times the soldier beam
width.

Figure 40. Relationship between lateral earth pressure, wall deflection,

and depth of wall embedment.

85
Figure 41. Broms method for evaluating ultimate passive resistance.

Back calculated passive resistances for laterally loaded piles and soldier beam and lagging walls also
compared favorably with passive resistance calculations developed by Wang and Reese (1986) for
cantilevered drilled shaft walls. The Wang-Reese equations consider several potential failure
mechanisms for laterally loaded piles in sands and clays. The effects of spacing between adjacent
piles and the potential for soil to squeeze between adjacent piles are also considered. These
mechanisms and the calculations developed for evaluating ultimate passive resistance for
cohesionless and cohesive soils are provided in appendix B. Comparisons of the Broms method and
the Wang-Reese method are provided in subsequent sections.

For driven soldier beams, the flange width of the soldier beam should be used for lateral resistance
calculations. For drilled-in soldier beams backfilled with structural concrete, the full diameter of the
soldier beam should be used for lateral resistance calculations. For drilled-in soldier beams
backfilled with lean-mix concrete, the full diameter of the beam may be used for lateral resistance
calculations provided the lean-mix concrete backfill has a compressive strength of no less than 0.35
MPa.

5.5.2.2 Continuous Walls

The evaluation of passive resistance for walls with continuous elements involves calculating the
passive soil resistance according to the methods described in section 4.4.2. When evaluating the
passive earth pressure coefficient for cohesionless soils (see figure 16 and 17), an interface friction
angle, , varying from 0.5
construction, type of wall element (i.e., steel sheet-pile, tangent/secant pile, slurry wall), and axial
load transfer in the embedded portion of the wall.

5.5.3 Depth of Penetration below Excavation

Competent Soil Conditions

The depth of penetration of vertical wall elements based on lateral capacity is generally calculated
using a factor of safety with respect to lateral capacity of 1.5. When using the Wang-Reese
86
equations, the required embedment depth corresponds to the depth at which the ratio of ultimate
passive resistance to the reaction force, R, is greater than or equal to 1.5. When calculating passive
resistance using the Broms method or for the analysis of continuous sheet-pile walls, the passive
earth pressure coefficient is reduced by a factor of safety of 1.5 to calculate the passive resistance for
comparison directly to the reaction force. For the Broms method, the Rankine passive earth pressure
coefficient with = 0 is used. When using the Wang-Reese equations to calculate the ultimate
passive resistance in competent sands and clays, the envelope of minimum resistance calculated from
the various failure mechanisms should be used to evaluate the required embedment depth.

The reaction force was previously defined as being computed from the area of the apparent pressure
diagram from the base of the excavation to the midheight between the base of the excavation and the
lowermost anchor (see section 5.3.3). For the calculations of required embedment for cohesionless
soils, the active earth pressure that acts below the bottom of the excavation over the width of the
soldier beam or over a unit-width of sheet-pile is also considered to be a driving force. For
competent cohesive soils, the active earth pressure may be negative and is therefore neglected in the
embedment calculation.

Weak Underlying Layer

Lateral load capacity is limited below the base of the excavation in very loose to loose granular soils
or soft to medium clays. In very loose to loose granular soils, the wall elements must experience
relatively large movements to fully develop passive resistance. The vertical wall element may
become overstressed prior to achieving these movements. When undrained conditions are assumed
in soft to medium clays with stability number NS > 4, the net pressure remains on the active side of
the excavation regardless of the embedment depth, thus no passive resistance can be developed.
Penetration of the vertical wall elements (i.e., sheet-pile or soldier beam) should be limited to a
nominal minimal depth of approximately 20 percent of the excavation depth unless deeper
embedments are necessary to develop sufficient capacity to resist vertical loads (see section 5.6),
provide basal stability (see section 5.8.2), or limit ground movements.

For these cases, the embedded portion of the wall should be designed as a cantilever fixed at the
lowest anchor. An example calculation for a wall in a cohesive deposit is shown in appendix C. For
design, the wall section should be selected based on the maximum bending moment evaluated, i.e.,
either the maximum bending moment in the exposed portion of the wall above the lowermost anchor
or the calculated cantilever bending moment about the lowermost anchor as shown in appendix C.

5.5.4 Comparison of Wang-Reese and Broms Method for Competent Soils

Calculated required embedment depths based on the Wang-Reese method and the Broms method are
compared in figure 42 for an example wall constructed in sand and figure 43 for an example wall
constructed in clay. The reaction force was calculated based on apparent pressure envelopes for
sands and stiff to hard clays. Spreadsheet calculations for these examples are included in appendix
B. For the examples shown, for a factor of safety of 1.5, the Wang-Reese method predicts less
embedment as compared to the Broms method. For design of anchored walls, either method may be
used.

The computer program COM624P (FHWA-SA-91-048, 1993) can be used to verify fixity conditions
(i.e., free earth support or fixed earth support) for the embedment depth calculated using either the
87
Wang-Reese or the Broms analysis. A reversal of curvature in the deflected shape of the wall along
the embedded portion of the wall indicates fixed-earth support conditions.

PASSIVE PRESSURE (kN/m)

0 200 400 600 800 1000
0
5.0
Wang-Reese Passive Resistance
1 4.5
Broms Passive Resistance
4.0
FS (Wang-Reese)
2 3.5
FS (Broms)
3.0
3 2.5
2.0
4 1.5
1.0

5 0.5
0.0

6 0 1 2 3 4 5 6
TOE DEPTH (m)

Figure 42. Comparison of Broms and Wang-Reese Method for wall in sand.

PASSIVE PRESSURE (kN/m)

0 100 200 300 400 500 600
0

5.0
1 4.5
4.0
2 3.5
3.0

3 2.5
2.0
FS (Wang-Reese)
4 1.5
Wang-Reese Passive Resistance
1.0 FS (Broms)
Broms Passive Resistance
0.5
5
0.0
0 1 2 3 4 5
6 TOE DEPTH (m)

Figure 43. Comparison of Broms and Wang-Reese Method for wall in clay.

5.6 AXIAL CAPACITY OF WALL

5.6.1 Introduction

The sum of the vertical component of each ground anchor load and other vertical loads (e.g., dead
weight of wall, permanent live loads) must be considered in the design of the wall elements to
minimize potential of a bearing failure and/or excessive vertical wall movement. Soldier beams for

88
anchored walls are either driven or they are placed in pre-drilled holes that are subsequently
backfilled with lean-mix or structural concrete. Conventional analyses of axial load capacity for
driven piles and drilled shafts may be used to design the vertical wall elements of anchored walls.
Analysis methods presented herein are described in greater detail in FHWA-HI-97-013 (1996) for
driven piles and FHWA-SA-99-019 (1999) for drilled-in piles.

5.6.2 Axial Load Evaluation

External vertical loads on an anchored wall include: (1) vertical ground anchor forces; (2) dead
weight of the wall elements (e.g., soldier beams, lagging, concrete facing); and (3) other external
loads. Other loads that may be significant for anchored walls, but which are relatively difficult to
evaluate a priori include: (1) load transferred to the retained ground above the excavation subgrade;
and (2) downdrag loads that result when the retained ground settles relative to the wall. The method
recommended herein for designing vertical wall elements of permanent walls for axial capacity
assumes that all external vertical loads are resisted by side friction and end bearing resistance in the
embedded portion of the wall. Target factors of safety for calculating the allowable axial load are
recommended based on soil type.

Research results (see FHWA-RD-97-066, 1998) and review of limited case history information
indicate that:

axial load will be transferred from the wall to the ground above the excavation subgrade in
dense to very dense sands or stiff to hard clays, however the length of time is unknown as to
when these loads may be transferred to the embedded portion of the wall;

axial load may be minimized by installing anchors as near to horizontal as possible;

downdrag loads are reduced to zero when the wall settles approximately 2.5 mm relative to
the supported ground; and

downdrag loads will likely act on walls constructed in soft to medium clays or loose to
medium dense sands that are founded on a relatively firm stratum.

These observations support the conservative assumption that all external loads should be designed to
be supported by the embedded portion of the wall.

Table 14 presents recommended factors of safety (FS) for calculating the allowable axial capacity of
driven and drilled-in soldier beams for permanent walls. Lower factors of safety may be justified
based on the results of site-specific load testing. These factors of safety (table 14) were developed
based on the requirement that vertical wall movements are minimized. The allowable axial capacity,
Qa, of driven and drilled-in soldier beams is defined as:

Q ult
Qa (Equation 23)
FS

Methods to calculate the ultimate axial capacity, Qult, are described subsequently.

89
Table 14. Recommended factors of safety for axial capacity of driven and drilled-in soldier beams.

Soil Type Factor of Safety on Factor of Safety

Skin Friction on End Bearing

Clays 2.5 2.5

Sands 2.0 2.5

For temporary SOE applications, designs may consider the potential for axial load support above the
excavation subgrade and thus, axial capacity for SOE wall elements may be based on lower factors
of safety than are listed in table 14. Lower factors of safety can be used if the designer can provide
data or be able to demonstrate that the vertical settlement of the wall will be relatively small.
Tolerable vertical settlement is necessary to assure that the ground anchor/wall connection does not
become overstressed and that lateral movements of the wall resulting from vertical wall movement
will be acceptable.

5.6.3 Axial Capacity Design of Driven Soldier Beams

5.6.3.1 General

The following guidance on the design of driven piles for axial capacity has been excerpted from
FHWA Report No. FHWA-HI-97-013 (1996). That document should be consulted for
supplementary information.

5.6.3.2 Effective Stress Analysis for Driven Soldier Beams

The ultimate axial load carrying capacity of driven piles in cohesionless soils or for effective stress
analysis of drained loading conditions in cohesive soils is given by:

Qult = fsAs + qtAt (Equation 24)

where Qult is the ultimate pile capacity, fs is the unit skin friction resistance, As is the perimeter of the
of the pile, qt is the unit end bearing resistance, and At
section. For piles founded on predominantly gravelly soils or rock, the actual steel area of the pile
should be used for At. Standard SI units are: Qult (kN); fs (kPa); qt (kPa); As (m2); and At (m2).

The unit skin friction resistance, fs, is calculated from the following expression:

fs po (Equation 25)

90
where the beta coefficient, = Kstan , po is the average vertical effective stress along the pile shaft,
Ks is an earth pressure coefficient, and is the interface friction angle between the pile and the soil.
Standard SI units are: (dimensionless); po(kPa); Ks (dimensionless); and (degrees).

In equation 25, po is evaluated using elastic superposition of the vertical effective stress on both sides
of the midpoint of the embedded portion of the soldier beam. For uniform soil deposits, p o is
computed as 0.5 times the height of excavation times the unit weight of the soil plus 0.5 times the
embedded soldier beam length times the unit weight of the soil.

The unit end bearing resistance, qt, may be calculated from the following expression:

qt = Ntpt (Equation 26)

where Nt is the toe bearing coefficient and pt is the vertical effective stress at the pile tip which is
calculated based on the depth of the pile tip measured relative to the base of the excavation.
Standard SI units are: Nt (dimensionless) and pt (kPa).

Charts for estimating and Nt based on the drained friction angle of the soil are provided in figures
44 and 45, respectively. The designer should, if possible, confirm the selection of and Nt in a
particular soil with local correlations between static capacity calculations and static load tests.

Clay
Silt
Sand
Gravel

0.1
20 25 30 35 40 45 50
DRAINED FRICTION ANGLE (DEGREES)

Figure 44. Chart for estimating  coefficient versus soil type friction angle (after Fellenius, 1991).

91
1000

100

Clay
10 Silt
Sand
Gravel

1
20 25 30 35 40 45 50
DRAINED FRICTION ANGLE (DEGREES)

Figure 45. Chart for estimating Nt coefficient versus soil type friction angle (after Fellenius, 1991).

5.6.3.3 Total Stress Analysis for Driven Soldier Beams in Clays

For driven soldier beams in clay, a total stress analysis may be used where the ultimate capacity is
calculated from the undrained shear strength of the clay. The unit shaft resistance, fs, may be
calculated from:

fs = ca = Su (Equation 27)

where ca is the adhesion between the pile and the soil at failure, is an empirical adhesion factor for
reduction of the average undrained shear strength of the undisturbed clay along the embedded length
of the pile. Figure 46 shows recommended adhesion values for cohesive soils. For driven H-piles,

the pile.

The unit end bearing resistance in a total stress analysis for cohesive soil can be expressed as:

qt = Su Nc (Equation 28)

The term Nc is a dimensionless bearing capacity factor that depends on the pile diameter and the
depth of embedment. The bearing capacity factor, Nc, should be taken as 9 in this total stress
analysis for anchored wall applications.

92
100

80
D > 40b

60

40

D = 10b
20

0
0 50 100 150 200

Undrained Shear Strength, Su (kPa)

Concrete, Timber, Corrugated Steel Piles

Smooth Steel Piles
D = Distance from ground surface to bottom of
clay layer or pile toe, whichever is less
b = Pile Diameter

Figure 46. Adhesion values for piles in cohesive soils (after Tomlinson, 1980).

5.6.4 Axial Capacity Design Of Drilled-in Soldier Beams

5.6.4.1 General

The following guidance may be used to estimate axial capacity for drilled-in soldier beams in
cohesionless or cohesive deposits. For more complex soil types (e.g., intermediate geomaterials) or
rock, axial capacity should be computed from methods described in FHWA Report No. FHWA-SA-
99-019 (1999).

5.6.4.2 Cohesionless Soils

The axial capacity of a drilled-in soldier beam in cohesionless soils may be calculated based on
conventional design methods for drilled shafts. The ultimate axial capacity is given by equation 24
where As is the surface area of the drilled-in soldier beam and At is the cross section area of the
drilled-in soldier beam.

The unit skin friction resistance is given by the following expression:

93
fs po (Equation 29)

where: po = average vertical effective stress;

= 1.5 - 0.42z0.34, 0.25 < < 1.2, for sands with SPT N  15 blows/300 mm (z in meters);
= N/15 (1.5 - 0.42z0.34) for sands with SPT N < 15 blows/300 mm; and
= 2.0 - 0.15z0.75 for gravelly soils.

The factor in equation 29 is a factor used to account for the effects of lateral stress changes due to
open hole drilling, introduction of fluid concrete, and interface friction between granular soil and
concrete. In calculating the factor, z is the depth measured from the top of the wall to the midpoint
of the embedded soldier beam in the case of uniform soils. For uniform soil deposits, po is computed
as 0.5 times the height of excavation times the unit weight of the soil plus 0.5 times the embedded
soldier beam length times the unit weight of the soil. For cases where the soldier beam is embedded
in a multiple soil layer system, the factor should be evaluated at the midpoint of each layer.

The unit end bearing resistance, qt, may be calculated as:

q t (kPa ) 57.5 N (Equation 30)

where N is the uncorrected average SPT N value within two times the diameter of the base of the
shaft. Values calculated according to equation 30 are ultimate values corresponding to drilled shaft
settlements of approximately five percent of the base diameter.

5.6.4.3 Cohesive Soils

For drilled-in soldier beams in clay, a total stress analysis is used to evaluate the ultimate axial
capacity under undrained conditions. The unit skin friction is calculated as:

fs = Su (Equation 31)

where: = 0.29 + 0.19 Su/po , and Su = undrained shear strength determined from consolidated
undrained triaxial tests.

The unit end bearing resistance is calculated as described below.

Depth to Shaft Base (D)  5B

qt N*c Su (Equation 32)

where D is the depth of embedment measured from the base of the excavation, B is the
diameter of the shaft, and Su is the undrained shear strength determined from unconsolidated
undrained (UU) triaxial tests, and Nc* is a bearing capacity factor (table 15).

94
Table 15. Bearing capacity factors for
evaluation of end bearing in drilled shafts in clays.
Su (kPa) Nc *
24 6.5
48 8.0
96 8.7
>96 9.0

Depth to Shaft Base (D) < 5B

qt = [0.667+0.0667(D/B)] Nc* Su (Equation 33)

5.6.4.4 Design Issues for Concrete Backfill of Predrilled Soldier Beam Holes

General design recommendations for concrete backfill of predrilled holes include the use of
structural concrete from the bottom of the hole to the excavation base and lean-mix concrete for the
remainder of the hole. The design concept is to provide maximum strength and load transfer in the
permanently embedded portion of the soldier beam while providing a weak concrete fill in the upper
portion which can easily be removed and shaped to allow lagging installation. However, contractors
often propose to use lean-mix concrete backfill for the full depth of the hole to avoid the delays
associated with providing two types of concrete in relatively small quantities.

When using structural concrete with a minimum compressive strength of 21 MPa and appropriate
concrete placement procedures, the vertical load from the exposed portion of the wall is transferred
from the steel beam to the concrete and that the entire drilled shaft cross section is effective in
resisting the vertical load. For this case, the wall can be analyzed as a drilled shaft using the methods
presented in section 5.6.4.2 and 5.6.4.3. However, for lean-mix backfilled drilled shafts, the lean-
mix concrete may not be sufficiently strong to allow vertical load transfer from the soldier beam to

cross section will not be effective in transferring load to the surrounding soil.

When designing the embedded portion of a permanent drilled-in soldier beam wall that is backfilled
with lean-mix concrete, the following two analyses should be performed, and the analysis that results
in the greater required embedment depth should be used:

Analysis 1: Compute the required embedment depth assuming the drilled-in soldier beam can
be analyzed as a drilled shaft. Use the procedures described in section 5.6.4.2 and 5.6.4.3 and
assume the full cross section of the shaft is effective in resisting vertical load.

Analysis 2: Compute the required embedment depth assuming that the soldier beam will

(section 5.6.3.2 and 5.6.3.3) should be used as follows: (1) use equation 25 to evaluate skin
friction resistance assuming that Ks = 2, = 35
evaluate As; and (2) use equation 26 or 28, depending on the soil type at the bottom of the shaft,

95
5.7 ANCHORED SLOPES AND LANDSLIDE STABILIZATION SYSTEMS

5.7.1 General

Ground anchors may be used in combination with walls, horizontal beams, or anchor blocks to
stabilize unstable slopes and landslides. Prestressed ground anchors act against the thrust of the
potential slip surface and increase the normal stress on the potential slip surface. Both of these
actions contribute to increase stability of the slopes. Also, by compressing the soil, softening
processes that tend to weaken the soil with time are inhibited (Morgenstern, 1982). Issues related to
structural and global stability analyses of anchored systems for slopes and landslides are presented
herein.

Anchored slopes and landslide stabilization systems are designed to restrain forces associated with
unstable ground masses. Restraint forces calculated based on the apparent earth pressure envelopes
(section 5.2) may significantly underestimate the required restraint force necessary to stabilize an
unstable slope or landslide to a particular target factor of safety. For this case, limit equilibrium
analyses should be used to evaluate ground anchor and wall loads for anchored slopes and landslide
stabilization systems. Details on the use of limit equilibrium analysis methods for modeling these
anchored systems are discussed subsequently. Where failure surfaces are steep, however, calculated
required restraint forces may be greater when apparent pressure diagrams are used as compared to
limit equilibrium methods. For that case, design loadings should be based on apparent pressure
diagrams.

5.7.2 Design Concepts

The target slope stability factor of safety for slopes and landslide stabilization systems is typically 1.2
to 1.3. Higher values, although not common, may be required depending on the criticality of the
structure, requirements with respect to deformation control, and confidence in the selected shear
strength parameters. When analyzing slopes and landslides, the factor of safety should be calculated
for all potential failure surfaces since several surfaces (both planar and circular) may have factors of
safety less than the target value. Also, the stability of the downslope ground in front of the wall or
slope face must be verified. If the downslope material is unstable, potential exists for downslope
movements to occur resulting in a reduction in passive capacity of the soil in front of the wall.

Information on soil and bedrock hydraulic head (i.e., porewater pressures) is required for slope
stability analyses of anchored systems. The available piezometer and hydraulic head data for each
water-bearing zone should be evaluated, but note that these hydraulic heads are likely to change as a
result of seasonal changes in precipitation and construction activities that change or interrupt water
flow paths. For stability analyses, an envelope of maximum heads measured at different times of the
year should be conservatively assumed.

The required restraint force that must be developed by the anchors and the wall will depend on the
location of the wall relative to the sliding surface and the amount of material to be stabilized.
Required restraint forces are relatively small for steeply inclined failure surfaces and relatively large
for long, shallow failure surfaces because of the size of the mass requiring restraint. For very long,
unstable slopes, multiple anchored walls along the slope may be more cost effective than one wall.

96
For anchored walls that are used to stabilize an unstable or moving slope, a minimum design consists
of either one level of ground anchors and a wall that penetrates the critical failure surface or two
levels of ground anchors and a wall that does not penetrate the failure surface. This minimum design
ensures that there are at least two points of restraint at the wall location.

5.7.3 Limit Equilibrium Calculations

5.7.3.1 Overall Approach

Slope stability computer programs that incorporate limit equilibrium methods of slices are routinely
used for analyzing the stability of slopes and embankments. They may also be used to examine the
stability of anchored walls, slopes, and landslide stabilization systems. However, the current state-
of-the-practice does not include a generally accepted method of modeling the restraint force provided
by the prestressed ground anchors. Methods used in practice distribute the anchor forces to slices in
different ways and each slope stability computer program includes one or several of these methods.
For this reason, caution should be exercised when using limit equilibrium methods to calculate
required forces to restrain a slope. Calculated forces should be reviewed critically and compared to

equilibrium methods for the analysis of anchored systems is provided in FHWA-RD-98-065 (1998).
Two methods that may be used to model the ground anchor restraint forces are introduced below.

Method 1: Apply Surcharge or Concentrated Force to Wall or Slope Face: If a surcharge or

concentrated force equivalent to the total ground anchor restraint force is applied at the wall
or slope face, a very large vertical force component will be transmitted to the slice base on
which it acts. The calculated factor of safety for this slice will be unrealistically high. This
method seems realistic in that the large compressive forces imposed by ground anchors are
applied at the face. However, the large increase in vertical force to only one slice while
nearby slices remain unaffected seems incorrect since ground anchors presumably increase
the normal forces on the critical failure surface in a more widespread fashion.

Method 2: Apply Concentrated Force to Slice Base Where Failure Surface Crosses Anchor:
With this method, the normal stress on the slice where the failure surface and the anchor
intersect is increased while nearby slices remain unaffected. This method is commonly used
for modeling geosynthetic reinforcement. This method suffers from the same limitation as
Method 1 in that the increase in normal force on the failure surface is highly localized.

For both of these methods, the increase in normal stress on the critical potential failure surface is
highly localized and not likely to be consistent with the actual distribution of stresses imposed by the
ground anchors. For a case where the failure plane is at a constant inclination and the soil strength
along the failure plane is homogeneous, both of these methods provide similar results. For failure
surfaces that are irregular and for highly stratified soils, it is likely that these two methods will result
in different calculated factors of safety.

A reasonable approach to using limit equilibrium methods for evaluating anchored slopes and
landslides is to perform an analysis using either Method 1 or Method 2 and to compare the calculated
factor of safety for a given anchor restraint force from the analysis to the design target value
(typically 1.3). Ideally, analyses using both methods should be performed and the results from each

97
compared to the design target factor of safety. If available, a slope stability computer program that
has the capability to search for critical failure surfaces using a slice equilibrium method that satisfies
both moment and force equilibrium should be used. If this feature is not available, the search for the
critical failure surface may be performed using simpler force or moment equilibrium methods. Some
programs may not offer the capability to use a method that satisfies both force and moment
equilibrium to perform a general search for a critical potential failure surface, but does offer the
capability to use a method that satisfies both force and moment equilibrium to calculate the factor of
safety for a specific failure surface.

If the calculated factors of safety from one or both analyses exceed the target value, than the
calculated anchor restraint force can be used for design. If the calculated factors of safety are less
than the target value, then the anchor force can be increased until the target value is reached. The
user should evaluate whether the calculated restraint force required to meet the target factor of safety
is reasonable. If the calculated restraint force seems excessively large or small or if changes in
analysis parameters (e.g., inclination of failure surface) result in very large variations in calculated
factors of safety, then additional analyses should be performed.

Both of these methods for evaluating the total stabilizing load for an anchored wall or slope are
described herein. The analyses should be performed for each critical design cross section.
Noncircular (i.e., planar) failure surfaces should be used where the soils are predominantly
cohesionless or where the failure surface is located along a well-defined interface. For analysis of
temporary walls constructed in weak (i.e., soft to medium) cohesive soils, a circular failure surface
should be used. Table 16 provides a general outline for performing the analysis.

Table 16. Procedure to evaluate total lateral earth load using slope stability computer programs.

Step 1. Develop cross section geometry including subsurface stratigraphy, external surcharge
loadings, and water pressures.

Step 2. Assign shear strength and unit weight to each soil and/or rock layer.
Step 3. Select limit equilibrium method that satisfies both force and moment equilibrium and
appropriate critical surface search parameters.
Step 4. Apply surcharge or concentrated force(s) to wall or slope face (Method 1) or model
the ground anchors as reinforcements (Method 2). For vertical walls, model the wall
face with a slight batter to avoid anomalous numerical instabilities.
Step 5. Evaluate critical surface and factor of safety for the load applied in Step 4.
Step 6. Repeat steps 4 and 5, increasing the surcharge or concentrated force(s) (Method 1) or
reinforcement tension (Method 2), until the target factor of safety is obtained.

5.7.3.2 Method 1 Analysis

In using Method 1, the following cases are considered: Case 1 - the wall penetrates the potential
critical failure surface; and Case 2 - the wall does not penetrate the potential critical failure surface.
These cases are illustrated in figure 47. For Case 1, it is assumed that the vertical component of the
anchor load is transmitted below the critical failure surface at the wall location, thus only the
horizontal component of the ground anchor force is transmitted to the failure surface. The total
surcharge load must be resisted by the ground anchors and the lateral capacity of the portion of the
98
wall that extends below the failure surface. A method to model the lateral capacity of the portion of
the wall that extends below the failure surface is described in section 5.7.4. For Case 2, both the
vertical and horizontal component of the anchor load are transmitted to the failure surface.

Figure 47. Modeling the ground anchor force in limit equilibrium analysis
(after FHWA-RD-97-130, 1998).

99
For cases where homogeneous weak cohesive soil extends far below the base of the excavation (at
least approximately 20 percent of the wall height), the potential critical failure surface may likely
penetrate significantly below the bottom of the excavation. For these cases, relatively large loads
will need to be resisted by the lower anchors. When modeling the ground anchor restraint force
using Method 1 and the procedure outlined in table 16, the resultant of the surcharge or concentrated
force(s) used to model the ground anchor restraint force should be located between 0.3H and 0.5H
measured from the bottom of the excavation. A procedure for evaluating the total load required to
stabilize a cut, for which the failure surface penetrates significantly below the wall, is described in
table 17 and is illustrated on figure 48. With this procedure, the location of the resultant of the total
load required to stabilize the system to the target factor of safety will get progressively lower in the
wall as the failure surface penetrates deeper.

Figure 48. Limit equilibrium analyses used to evaluate total lateral earth load for anchored
systems constructed in weak cohesive soils (after FHWA-RD-97-130, 1998).

5.7.3.3 Method 2 Analysis

With this method, the ground anchor is considered to be a high capacity reinforcement. The axial
anchor force is modeled along the length of the anchor and the anchor bond zone. The axial force in
the reinforcement is assumed to vary linearly from the full anchor capacity for all positions in front
of the anchor bond zone, to zero force for the end of the ground anchor. This concept is similar to
stability analyses involving soil nails as described in FHWA-DP-96-69R (1998). Multiple levels of
anchors may be modeled, so the user should assume a reasonable layout of anchors and anchor
inclinations in performing the analysis. If the failure surface crosses the wall, the additional restraint
provided by the wall may be modeled (see section 5.7.4).

100
Table 17. Procedure to evaluate total lateral earth load for anchored systems
constructed in weak cohesive soils.

Step 1. Same as table 16.

Step 2. Same as table 16.

Step 3. Perform a limit equilibrium analysis wherein the failure surface intersects the bottom of
the excavation. Use a slice equilibrium method that satisfies both force and moment
equilibrium and assume a circular failure surface.
Step 4. Apply surcharge or concentrated force(s) to wall or slope face to model the restraint
force of the ground anchor(s). For vertical walls, model the wall face with a slight
batter to avoid anomalous numerical instabilities.
Step 5. Evaluate factor of safety for failure surface intersecting the bottom of the excavation
for the load applied in Step 4.

Step 6. Repeat steps 4 and 5, increasing the surcharge or concentrated force(s), until the target
factor of safety is obtained.

Step 7. Perform a second limit equilibrium analysis that searches for the most critical potential
failure surface. Apply uniform surcharge or concentrated force(s) over upper half of
wall or slope equivalent to one-half the total load calculated from Step 6. Apply
uniform surcharge or concentrated force(s) over lower half of wall and increase this
force until the target factor of safety is achieved for the critical potential failure surface.

5.7.4 Modeling Lateral Wall Resistance in Limit Equilibrium Analyses

When the critical potential failure surface intersects the embedded portion of the wall, the additional
resistance provided by the wall may be included in a limit equilibrium analysis. The resisting force
to be used in the limit equilibrium analysis is the lesser of the following: (1) the shear capacity of the
wall; or (2) the total passive force that may be developed in the soil over the length of the wall from
the failure surface to the bottom of the wall. The shear capacity of the wall is constant and is
assumed to be equal to the allowable shear capacity of the vertical wall element.

The methods presented in section 5.5 may be used to calculate the total passive force that may be
developed over the length of the wall below the failure surface. Figure 49 shows the total passive
force developed over a 6 m embedded portion of the example wall in cohesionless soil described in
section 5.5.4.

The force to be modeled in the limit equilibrium analysis, Fp, may be modeled as a one-unit wide
element with a cohesive strength equal to the minimum passive force as described above. For
analyses involving soldier beam walls, this force should be reduced by the soldier beam spacing to
provide the restraint force on a per unit basis for the limit equilibrium analysis.

101
TOTAL PASSIVE FORCE, Fp (kN)
0 200 400 600 800 1000
0

4
z
6m
5 Fp
Total Passive Force (Broms)
Total Passive Force (Wang-Reese)
6

Figure 49. Total passive force for example wall in cohesionless soil.

5.7.5 Comparison of Methods to Evaluate Required Earth Load in Homogeneous Soils

This section provides comparisons between apparent earth pressure and limit equilibrium based
calculations for evaluating required restraint forces in relatively homogeneous soils. These
comparisons were performed for vertical walls with either planar or circular failure surfaces and
where the soil strength properties ( u) were constant for the entire profile analyzed.

Cohesionless Soils

Three methods have been described for evaluating the required total earth load, PREQ, to stabilize a
cut in cohesionless soils. For design of the wall, the ground anchors and the reaction force at the
excavation subgrade carry this total earth load. In table 18, the normalized total earth load (KREQ =
PREQ/½ H2) required to stabilize a cut in cohesionless soil is compared for the following three
methods: (1) apparent earth pressure envelope for sands; (2) sliding wedge analysis (section 5.2.8);
and (3) limit equilibrium method (section 5.7.3). The apparent earth pressure envelope produces a
total earth load equal to 0.65KA H2, which is 1.3 times greater than that for active Rankine
conditions. For the sliding wedge and limit equilibrium analysis, a factor of safety of 1.3 on the
shear strength was used. For the limit equilibrium analysis, a uniform horizontal surcharge was
applied to the wall face and increased until the target factor of safety was achieved (i.e., Method 1
from section 5.7.3).

The results indicate that all three methods give similar results, especially for higher strengths. When
designing anchored walls in reasonably homogeneous cohesionless soils for which competent soils
exist below the wall excavation, any of these methods will provide reasonable results, but using the
apparent earth pressure envelope to calculate the required anchor loads is the most expedient.

102
Table 18. Values of KREQ in cohesionless soil using various methods to evaluate earth pressures.

Apparent Earth Sliding Wedge Limit Equilibrium Percent Difference

Pressure Envelope (%)(1)
25 0.53 0.58 0.59 10
30 0.43 0.46 0.46 7
35 0.35 0.38 0.37 7
40 0.28 0.31 0.29 8
Note: (1) Percent Difference = ((Maximum - Minimum)/Maximum) *100

Cohesive Soils

Limit equilibrium methods were also assessed for evaluating the total earth load for anchored
systems in purely cohesive soils. For temporary anchored systems in soft to medium clays with Ns >

and limit equilibrium solutions. These results are shown in figure 50. Limit equilibrium methods

equilibrium. Results indicate that limit equilibrium methods compare favorably to Rankine analyses
where the failure surface intersects the corner of the wall. When the failure surface extends below
the excavation (e.g., d/H = 0.2 in figure 50), Henkel's method and Bishop's method are in reasonable
agreement and are upper bounds. For cases where the critical potential failure surface extends below
the base of the excavation and where Ns>5, the Rankine analysis results are unconservative. For
those cases, either Henkel's method or limit equilibrium analysis methods should be used to evaluate
the total earth load. The total load should then be redistributed into an apparent pressure diagram
using the Terzaghi and Peck diagram for soft to medium clays (figure 23c).

103
Figure 50. Comparison of limit equilibrium methods for cohesive soils
(after FHWA-RD-98-065, 1998).

104
5.8 GROUND MASS STABILITY

5.8.1 Introduction

The stability analyses presented herein focus on whether the shear strength of the soil mass and the
location and magnitude of the restraint forces provided by the ground anchors and other structural
components are sufficient to provide an acceptable factor of safety with respect to several potential
ground mass instabilities. Potential ground mass instabilities that should be analyzed include: (1)
internal stability; (2) basal stability; and (3) external stability. Internal stability calculations are used
to locate the anchor bond zone behind the critical potential failure surface and have been described in
section 5.3.2.

5.8.2 Basal Stability

5.8.2.1 General

The common failure modes with respect to basal stability include bottom heave at the base of
excavations in cohesive soils and piping for excavations in cohesionless soils. Bottom heave occurs
when the soils at the base of the excavation are relatively weak compared to the overburden stresses
induced by the retained side of the excavation. Bottom heave may be a critical issue for temporary
anchored systems constructed in soft to medium clays, but is not considered critical for other soil
types. Piping occurs if there is sufficient water head to produce critical velocities at the base of the
excavation. Piping is not discussed herein because, for most soldier beam and lagging walls,
excessive water head is not a concern since the excavation typically takes place in the dry, or the
ground-water table is lowered prior to the start of excavation.

5.8.2.2 Evaluation of Bottom Heave Potential in Soft to Medium Clays

Significant basal heave and substantial increases in lateral earth pressures result when the weight of
the retained soil exceeds or approaches the soil bearing capacity at the base of the excavation.
Traditional methods for assessing the potential for bottom heave are based on the performance of
braced excavations in soft to medium clays. These braced excavation analysis methods will likely
produce conservative results for anchored walls as the mechanism of failure does not consider the
shear resistance provided by the grouted anchor body. Figure 51 shows a cut in soft clay H deep and
B wide. The block of retained soil exerts a vertical pressure qapplied on strip cd equal to its weight
minus the shear resistance of the soil along plane bd. The bearing capacity of a cohesive soil is equal
to NcSu where Nc is the bearing capacity factor. For cuts of infinite length, the factor of safety
against basal heave can be estimated as the ratio of the bearing capacity to the bearing pressure as:

N c Su
FS (Equation 34)
Su
H 
B'

105
B
D
B'

b a

Su H HB'
H
H

d c
D
B/ 2

(a) Failure planes, deep deposits of weak clay (b) Failure plane, stiff layer below bottom of excavation

10

Square and
9
circular B/L = 1
B/L = 0.5
8

H = Excavation depth
7 B = Excavation width
Nc
B/L = 0, Rectangular L = Length of excavation
6

4
0 1 2 3 4 5
H/B

(c) Bearing Capacity Factor, Nc

Figure 51. Analysis of basal stability (modified after Terzaghi et al., 1996, Soil Mechanics in
Engineering Practice, Reprinted by permission of John Wiley & Sons, Inc.).

Based on the geometry of the failure surface, B' cannot exceed B 2 . Thus, the minimum FS for
equation 34 is:

N c Su
FS (Equation 35)
Su 2
H
B

The width, B', is restricted if a stiff stratum is near the bottom of the cut (figure 51). For this case, B'
is equal to depth D. Substituting D for B' in equation 34, results in:

106
N c Su
FS (Equation 36)
Su
H
D

In relation to anchored wall designs in shallow deposits, equation 36 may be used. However in
moderate to deep soil deposits where the width of the excavation is very large, the contribution of the
shearing resistance along the exterior of the failure block is negligible and equations 34 and 35
reduce to:

Nc Nc
FS (Equation 37)
H Ns
Su

where Ns is the stability number defined as H/Su. The bearing capacity factor used in equation 37 is
affected by the height/width ratio (H/B), and the plan dimensions of the cut (B/L). Values of the
bearing capacity factor, Nc, proposed by Janbu et al. (1956) for analysis of footings may be used in
equation 37 and these values are shown on figure 51. Note from figure 51 that Nc values are greater
for excavations constructed in short lengths (e.g., slotted excavation) as compared to excavation of
the entire length of the wall. Unless the designer specifically requires staged lengths of excavation,
the design should be based on the assumption that the contractor will remove the entire length of
each lift of excavation.

Significant ground movements towards the excavation will occur when the bearing capacity of the

(1997) concluded that for excavation width to height ratios (B/H) between 0.5 and 4, factors of safety
for deep rotational stability (i.e., external stability) are likely to be less than those calculated for basal
heave. Current practice is to use a minimum factor of safety against basal heave of 2.5 for permanent
facilities and 1.5 for support of excavation facilities. As the factor of safety decreases, loads on the
lowest ground anchor increase. Factors of safety below these target values indicate that more
rigorous procedures such as limit equilibrium methods or Henkel's method should be used to
evaluate design earth pressure loadings.

5.8.3 External Stability

5.8.3.1 Introduction

Conventional limit equilibrium methods for slope stability can be used to evaluate the external
stability of an anchored system. An anchored system is externally stable if potential slip surfaces
passing behind or through the anchors have a factor of safety that exceeds the target factor of safety.
External stability analyses are particularly important in evaluating systems close to nearby structures
or for situations in which soft soil exists below the wall.

For temporary SOE anchored systems constructed in soft to medium clay soils, external stability
should be evaluated using short-term (i.e., undrained) strength parameters and temporary loading
conditions. For permanent anchored systems constructed in soils, external stability for both short-
term and long-term conditions should be checked. For systems constructed in stiff clays, external
stability for short-term conditions may not be critical, but long-term conditions, using drained shear
strength parameters, may be critical. Selection of shear strength parameters has been discussed in
107
chapter 4. External stability of walls supported by rock anchors is normally adequate; however, if
the rock mass has planes of weakness which are oriented in a direction that may affect stability,
external stability should be checked for failure surfaces passing along these weak planes.

A minimum acceptable factor of safety for external stability is 1.3. For permanent applications that
are critical, a higher factor of safety (e.g., 1.5) may be used.

5.8.3.2 Evaluation of External Stability Using Limit Equilibrium

To evaluate the external stability of an anchored system, potential failure surfaces passing behind or
through the anchors need to be checked. For walls with multiple levels of anchors, failure surfaces
should be checked that pass just behind each anchor (figure 52). In checking a failure surface that
passes behind a level of anchors, the failure surface may cross in front of or through the anchor bond
zone of another level(s) of anchors. In this case, the analysis may be amended to include a portion of
the restraint force from the other anchor(s). If the failure surface passes in front of an anchor, the full
design load can be modeled as a restraint force. If the failure surface crosses the anchor, a
proportional magnitude of load assuming that anchor bond stress is distributed uniformly over the
anchor bond length can be modeled. Where external stability requirements cannot be met, the
anchors may be lengthened or methods to improve anchor bond or load transfer mechanisms may be
used.

Denotes potential failure

surface for external
stability analyses

Figure 52. Failure surfaces for external stability evaluations.

108
5.9 TIEDOWN DESIGN

5.9.1 Introduction

Tiedowns refer to vertical or downward inclined ground anchors subjected to uplift forces.
Examples of tiedowns include foundation elements for structures subject to overturning or uplift
such as transmission towers and vertical anchors used to resist hydrostatic uplift forces in gravity
dams and underwater slabs. Tiedowns are designed to resist two possible failure mechanisms: (1)
individual anchor capacity to resist uplift pressures; and (2) overall stability of the ground mass
wherein the tiedown group geometry is sufficient to envelope a mass of ground to resist uplift forces.
The following information is presented in this section: (1) evaluation of overall ground mass stability
for individual and groups of rock and soil tiedown anchors; and (2) design of tiedown anchors for
slabs subjected to hydrostatic loads.

5.9.2 Uplift Capacity of Rock Tiedown Anchors

The potential failure mechanisms commonly considered for designing rock tiedown anchors include:
(1) overall rock mass stability where an inverted cone or wedge of rock is engaged (i.e., cone
breakout in figure 53); (2) failure in shear along the grout/rock interface; and (3) failure in shear
along the tendon/grout interface.

The uplift capacity of a rock tiedown anchor depends on the relative depth of the anchor bond zone,
defined as h/D, where h is the depth of the top of the anchor bond zone and D is the diameter of the
anchor. For values of h/D>15, the dominant failure mechanism in rock is failure at the grout/rock
interface. Test results indicate that, more specifically, failure occurs at the rock/grout interface in
weak rocks such as mudstones and shales whereas failure occurs at the tendon/grout interface in
strong rocks. For shallow anchors in weak mudstones, a combination of interface shear at the
rock/grout interface and cone breakout may occur.

For relatively deep anchors in weak rocks or where interface shear along the rock/grout interface
dominates, the uplift capacity of the rock tiedown anchor may be evaluated according to the methods
described in section 5.3.6.

For shallow anchors or where overall rock mass stability dominates, the uplift capacity of a rock
tiedown anchor is typically assumed to be equivalent to the effective weight of a cone- or wedge-
shaped failure mechanism as shown in figure 53a. In the analysis, the shear strength of the rock
mass is often ignored. If the weight of the rock within the contained cone is greater than the design
ground anchor load, the anchor is considered safe since rock shear strength has been ignored.
Designers commonly assume that the apex of the failure mechanism is located at the top, midpoint,
or bottom of the anchor bond zone and the included angle of the mechanism ranges from 60 to 90
degrees. The recommendations shown in figure 53a should be used in the absence of model or full-
scale load test results. For cases where soil overburden is above the rock anchoring strata, the failure
mechanism is assumed to be cylindrical in shape above the rock/soil interface. For anchors with
overlapping cones, the stability of the ground is analyzed as shown in figure 53b. The overlapping of
the zones of influence between adjacent anchors results in anchor uplift capacity less than that for a
single anchor.

109
A wide range of factors of safety with respect to overall rock mass stability may be calculated based
on the assumed geometry of the failure mechanism. Factors of safety for design with respect to
overall rock mass stability are typically 2 to 3 (British Standards Institution BS8081, 1989). This
factor of safety may be reduced owing to the conservative assumption that the shear strength of the
rock is neglected in the analysis, particularly for competent rocks that are not highly fissured.
However, in highly fissured or loose rock strata, an increase in the factor of safety may be required.

Figure 53. Inverted cone mechanisms for overall rock mass stability.

5.9.3 Uplift Capacity of Soil Tiedown Anchors

For tiedown anchors installed in soils, the failure mechanisms of cone breakout and interface shear
along the soil/grout interface are analyzed. Like rock anchors, the cone breakout mechanism
dominates for shallow anchors whereas interface shear dominates for relatively deep anchors. A
grouted soil anchor subjected to uplift behaves similarly to a small diameter drilled shaft subjected to
uplift.

110
Soil anchors typically used to resist uplift are relatively deep (i.e., h/D is relatively large) so that the
governing failure mechanism is the mobilization of grout/ground interface shear resistance. Uplift
resistance may be calculated as:

Qu = Qtu+ Qsu (Equation 38)

where: Qu is the uplift capacity, Qtu is the tip resistance, and Qsu is the side resistance. Tip resistance
that may develop from suction is commonly assumed to be zero for drained (long-term) uplift
capacity of drilled-in elements. Therefore, the uplift capacity of a grouted soil anchor primarily
results from interface skin resistance between the grout and the ground. Uplift capacity may be
evaluated using the procedures described in section 5.3.6 for soil anchors or may be calculated
according to (Kulhawy, 1985):
N
K
Qsu D ' K o i tan
v i i / i zi (Equation 39)
Ko i 1

where: zi = thickness of layer i, D = anchor diameter, and K/Ko = stress modification factor to
adjust for construction influences. The remaining parameters are evaluated at the mid-depth of each
layer: ´v = vertical effective stress, = effective stress angle of friction for the shear surface
interface, Ko = in-situ horizontal stress coefficient, and = effective stress friction angle for the soil.
The anchor depth and perimeter terms are computed simply from the anchor geometry, while the
vertical effective stresses are computed from the effective soil unit weight.

For soil anchors, / may be assumed equal to 1. For gravity anchors, the in situ value of Ko may be
modified based on anchor installation effects with typical values of K/Ko ranging from 2/3 to 1.
K/Ko values of 1 may be used for relatively dry installations with minimal drill hole disturbance. For
anchors installed under water or where very loose or running sands are encountered and significant
hole disturbance occurs, K/Ko values less than 2/3 may be appropriate.

For low pressure (i.e., grouting pressures less than 1 MPa) grouted anchors and gravity-grouted
anchors, no increase in K above the at-rest, Ko, value is warranted. For high pressure grouted
anchors, however, an increase in K is appropriate. The guidelines presented in table 19 are
recommended for use. Owing to the numerous factors that influence the K value including grout
pressure, construction method, and soil type, it is recommended that load tests be carried out to
confirm design values.

Table 19. Horizontal stress coefficient, K, for pressure grouted anchors (after Kulhawy et al., 1983).

Density
Soil Loose Compact Dense
Silt 1 4 10
Fine Sand 1.5 6 15
Medium Sand 5 12 20
Coarse Sand, Gravel 10 20 30

111
5.9.4 Design of Tiedown Anchors to Resist Hydrostatic Uplift

Tiedowns may be used to provide resistance to uplift forces caused by hydrostatic pressures. A
notable use of tiedown anchors in the U.S. was to resist hydrostatic uplift of a depressed roadway
section as part of the Boston Central Artery project (see Druss, 1994). The primary issues related to
the use of anchors for such tiedown applications are: (1) general stability of the enclosed ground
mass; (2) changes in anchor loads resulting from movement (i.e., surface heave, consolidation
settlements, creep deformations) in the enclosed ground mass; and (3) corrosion protection and
watertightness of the ground anchor. Corrosion protection and water tightness are discussed in
chapter 6.

General stability of a structure subjected to uplift is shown in figure 54. The system is in equilibrium
when U=W1+W2, where W1 and W2 are total weights of the structure and the enclosed ground,
respectively, and U is the total uplift resulting from the uplift pressure wh. The geometry of the soil
mass assumed to be mobilized at failure may be evaluated as shown in figure 54. Frictional
resistance that may develop between the ground and the sidewalls of the structure may be
conservatively neglected.

Structure
h

Enclosed
soil mass

30 or 45

Anchor

W1 = Weight of structure

W 2 = Weight of enclosed soil mass

U = Total uplift force = w hL

Figure 54. Stability of structure to hydrostatic uplift.

For conditions where the uplift structure is founded on relatively compressible ground, movements
associated with construction activities, fluctuations in groundwater levels, consolidation of the soil,

112
and soil creep may induce significant changes in ground anchor loads during the service life of the
structure. These movements are assumed to cause cycles of tensioning and detensioning in the
ground anchor tendon. If the tendon may be subjected to additional tensioning after lock-off, it is
important that the size of the prestressing steel be based on the maximum load that the anchor will be
subjected to during the service life.

5.10 SEISMIC DESIGN

5.10.1 Introduction

Few observations of the seismic performance of anchored walls have been made. Those
observations that are available indicate overall good performance of anchored wall systems subject
to strong ground motions in earthquakes. Most of the retaining wall failures reported during recent
earthquakes have occurred along quay wall gravity retaining wall systems and have been associated
with liquefaction of the backfill or the foundation soils. Following the Whittier, California
earthquake of 1987, Ho, et al. (1990) conducted a survey of the response of ten anchored walls in the
Los Angeles area. Only one of the ten anchored walls was designed to withstand seismic forces and
the authors concluded that the anchored walls examined performed very well and experienced little
to no loss of integrity due to the earthquake. The same conclusion was drawn following a survey of
the performance of anchored walls conducted following the 1994 Northridge earthquake (Ho,
personal communication, 1998).

Two modes of earthquake-induced failure for anchored walls are considered for design: internal
failure and external failure. Internal failure is characterized by failure of an element of the wall
system such as the tendons, ground anchors, or wall itself. External failure is characterized by a
global failure of the wall similar to that which occurs in many slope stability problems, with the
failure surface passing beyond the end of the anchors and below the toe of the wall.

To assess the internal and external seismic stability of an anchored wall, the effect of seismic loading
on the active and passive earth pressures, the resulting loads on the anchors, and force equilibrium of
potential sliding (or rotating) masses must be evaluated. The seismic loading on anchored walls is
most commonly evaluated using pseudo-static analysis, as described subsequently. Information
presented herein has been excerpted from FHWA Report No. FHWA-SA-97-076 (FHWA, 1997),
hereafter referred to as GEC No. 3. GEC No. 3 should be consulted for additional information on
seismic site characterization and design.

5.10.2 Internal Stability Using Pseudo-Static Theory

5.10.2.1 Lateral Earth Pressure

The most commonly used method for seismic design of retaining structures is the pseudo-static
method developed by Okabe (1926) and Mononobe (1929). The so-called Mononobe-Okabe method
is based on Coulomb earth pressure theory. In developing their method, Mononobe and Okabe
assumed the following:

the wall is free to move sufficiently to induce active earth pressure conditions;

113
the backfill is completely drained and cohesionless; and

the effect of the earthquake ground motion is represented by a horizontal pseudo-static

inertia force, khWs and a vertical pseudo-static inertia force kvWs if the vertical force acts
upward, or -kvWs, if the vertical force acts downward.

In figure 55, Ws is the weight of the sliding wedge and kh and kv are the horizontal and vertical
seismic coefficients, respectively. The seismic coefficient kh and kv are expressed as a fraction of the
acceleration of gravity g.

Using Mononobe-Okabe theory, the dynamic earth pressures in the active (PAE) and passive (PPE)
state are given by the following:

PAE 1 K AEH 2 (1 k v ) (Equation 40)

2

PPE 1 K PE H 2 (1 k v ) (Equation 41)

2

cos 2 ( )
K AE (Equation 42)
cos cos 2 cos( )D

1 2
2
sin( ) sin( i)
D 1
cos( ) cos(i )
(Equation 43)

cos 2 ( )
K PE
cos cos 2 cos( )D (Equation 44)

1 2
2
sin( ) sin( i )
D 1
cos( ) cos(i )
(Equation 45)

= tan-1 (kh/(1-kv)) (Equation 46)

where: = effective unit weight of the backfill;

H = height of the wall;
= angle of internal friction of the backfill;
= angle of friction of the wall/backfill interface;
i = slope of the surface of the backfill;
= slope of the back of the wall;
kh = horizontal seismic coefficient expressed as a fraction of g;
kv = vertical seismic coefficient expressed as a fraction of g; and
g = acceleration of gravity.

114
Figure 55. Forces behind a gravity wall.

Figure 56 presents values for KAE for values of from 20 to 45 degrees for vertical walls with level
backfill. The figure was derived for a wall/backfill interface friction angle set to /2. The horizontal
and vertical seismic coefficients (i.e., kh and kv) vary from 0 to 0.5 and from 0 to 0.2, respectively.
The major challenges in applying the Mononobe-Okabe theory are the selection of an appropriate
seismic coefficient to determine the magnitude of the seismic earth pressure and the distribution of
earth pressure or location of the seismic earth pressure resultant. As noted in GEC No. 3, use of a
seismic coefficient from between one-half to two-thirds of the peak horizontal ground acceleration
divided by gravity would appear to provide a wall design that will limit deformations in the design
earthquake to small values acceptable for highway facilities. Similar to slope stability analyses, the
vertical acceleration is usually ignored in practice in the design of anchored structures. Vertical
motions are not considered capable of applying significant loads to the anchors.

The total seismic active earth pressure may be assumed to be uniformly distributed over the height of
the wall, meaning that the earth pressure resultant acts at the mid-height of the wall. Therefore, place
the resultant active earth pressure calculated using the Mononobe-Okabe equations at mid-height of
wall for design analysis. The resultant passive pressure at the toe of the wall should also be placed at
mid-height of the embedded section.

115
1.0

0.8

o
= 20
o
0.6 25
o k v = 0.2
30
o 0.0
35
o -0.2
40
0.4 o
45

0.2
o
= 35
= i = kv= 0 =i=0
= /2 = /2
0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5

Horizontal Seismic Coefficient (kh )

Figure 56. Effects of seismic coefficients and friction angle on seismic active pressure coefficient
(after Lam and Martin, 1986).

5.10.2.2 Wall Design Considerations

Design of brittle elements of the system, (e.g., the grout/tendon bond (see section 4.2.2)) should be
governed by the peak force. Therefore, the peak ground acceleration (PGA), adjusted to account for
the effect of local soil conditions and the geometry of the wall, should be used with the Mononobe-
Okabe equation to calculate these peak forces. A factor of safety of 1.1 on these elements is
recommended for brittle failure modes.

Design of ductile elements (e.g., tendons, steel sheet pile and soldier beam walls, and sometimes the
grout/ground bond (this bond may also be brittle, depending on soil type)) should be governed by
cumulative permanent seismic deformation. In these cases, in lieu of a formal seismic deformation
analysis, a pseudo-static analysis with resultant forces calculated by the Mononobe-Okabe equation
using kh equal to 0.5 times the PGA should be appropriate. This recommendation is based upon the
results of numerous Newmark seismic deformation analyses for translational failures of slopes which
indicate the cumulative permanent seismic deformation for a system with a yield acceleration equal
to half the PGA is relatively small (e.g., no more than several centimeters) for earthquakes of all
magnitudes. A factor of safety of 1.1 on these elements is recommended for ductile failures.

Values of the PGA used in design should consider both the effect of local soil conditions and the
geometry of the wall. The free field PGA, including the effect of local soil conditions, may be
assumed to act at the base of the wall. The PGA at the top of the wall should be evaluated from the
free field PGA by considering the potential for amplification of the free field PGA by wall geometry.
The PGA used in the Mononobe-Okabe equation may then be assumed to be the average of the PGA
at the top and bottom of the wall.
116
5.10.2.3 Liquefaction

Where economically feasible, potentially liquefiable soils behind or in front of an anchored wall
should be stabilized to mitigate the potential for liquefaction. Stabilization techniques that may be
employed for potentially liquefiable soils include densification, either prior to wall construction (for
foundation soils) or during or after backfill placement and penetration grouting. If potentially
liquefiable soil cannot be stabilized, it should be assumed to exert an equivalent fluid pressure on the
wall based upon the saturated unit weight of the soil. The anchor bond zone should not be formed in
liquefiable soil.

5.10.3 External Stability

5.10.3.1 Pseudo-Static Analysis

The external stability of an anchored wall is evaluated by performing pseudo-static limit equilibrium
stability analysis of the wall system. The failure surfaces analyzed should pass behind the back of
the ground anchors and beneath the toe of the wall. The pseudo-static analysis will provide the
location of the critical failure surface or surfaces. The location of critical failure surface may be used
to verify the length of the proposed ground anchor. The anchor bond zone should be located outside
of the active Mononobe-Okabe wedge of soil. As the acceleration increases, the slope of the active
failure wedge flattens according to the following equation:

1
1 tan(tan a cot b )[1 tan( ) cot b ] 2 tan a
A ( ) tan
1 tan( )(tan a cot b ) (Equation 47)

where A is the inclination with respect to the horizontal of the failure surface; a = -i- ; b = - - ;
and , i, , and were defined previously.

As the slope flattens, the Mononobe-Okabe failure surface extends farther in the horizontal direction.
Figure 57 shows the variation of A and the coefficient of dynamic active and passive earth pressure
as a function of the horizontal seismic coefficient kh. Because of the extension of the Mononobe-
Okabe failure surface, the length of the ground anchors calculated in static design may need to be
increased to provide full anchorage of the ground anchors under seismic conditions.

117
Figure 57. Variation of failure surface inclination with horizontal acceleration coefficient.

AASHTO (1996) recommends using a seismic coefficient kh equal to 0.5A in pseudo-static external
stability analysis, where A is the PGA obtained from the seismic risk map published in AASHTO
Specifications. This value corresponds to an acceleration with a 10 percent probability of
exceedance in 50 years. A minimum factor of safety of 1.1 is recommended for the pseudo-static
external stability analysis.

5.10.3.2 Seismic Deformation Analysis

As an alternative to the pseudo-static design approach, external stability may be assessed using a
Newmark type seismic deformation analysis. In this approach, a pseudo-static external stability
analysis is carried out to evaluate the yield acceleration, ky, for failure surfaces passing behind the
back of the ground anchors. The yield acceleration is defined as the smallest horizontal acceleration
(seismic coefficient) that will reduce the factor of safety obtained in a pseudo static stability analysis
to 1.0. The ratio of the yield acceleration to the PGA can then be used to evaluate the earthquake
induced permanent displacement either by using design charts such as those presented in figure 58 or
by performing a formal Newmark analysis (FHWA-SA-97-076, 1997). The free field PGA should
be used in the analysis. The free field PGA considers the influence of local site conditions while the
PGA at top of the wall may be amplified. The free field PGA should be more representative of the

118
average value of ground acceleration throughout the height of the excavation than the amplified
PGA.

1000
Upper Bound

Mean +

Mean
100

10
0.01 0.1 1.0

Yield Acceleration, ky /Peak (Average) Acceleration, k max

Figure 58. Permanent seismic deformation chart (after Hynes and Franklin, 1984).

5.11 OTHER DESIGN ISSUES

5.11.1 Wall and Ground Movements

Depending on project constraints, requirements with respect to control of wall and ground
movements will vary. For example, permanent anchored walls constructed in granular soils with no
nearby structures pose little concern with respect to movements. Wall and ground movements,
however, may be the primary design issue for a temporary excavation support system located in a
major urban area. Estimates of wall and ground movements are typically made using semi-empirical
relationships developed from past performance data.

Maximum lateral wall movements for anchored walls constructed in sands and stiff clays average
approximately 0.2%H with a maximum of approximately 0.5%H where H is the height of the wall.
Maximum vertical settlements behind a wall constructed in these materials average approximately
0.15%H with a maximum of approximately 0.5%H.

To evaluate the settlement profile behind an anchored wall, the curves shown on figure 59 may be
used. Curves I and II are commonly used for permanent anchored walls. Settlements increase
rapidly for walls constructed in soft to medium clays where basal stability is marginal.

119
Figure 59. Settlement profile behind braced and anchored walls.

Several types of movement are associated with flexible anchored walls. These include: (1) cantilever
movements associated with installation of first anchor; (2) wall settlement associated with
mobilization of end bearing; (3) elastic elongation of the anchor tendon associated with a load
increase; (4) anchor yielding or load redistribution in the anchor bond zone; and (5) mass movements
behind the ground anchors. The last three components of deformation result in translation of the
wall and are relatively small for anchored walls constructed in competent soils. Excessive vertical
settlements of the wall may induce significant lateral wall movements in addition to causing high
stresses at the wall/anchor interface. Wall settlements may be minimized by installing ground
anchors at flat angles and by designing the embedded portion of the wall to carry applied axial loads.

5.11.2 Drainage Systems for Anchored Walls and Slopes

For anchored wall systems with a cast-in place (CIP) concrete wall facing, collection of subsurface
flow is usually achieved with prefabricated drainage elements placed between the wall and the
lagging. Full length elements are usually attached to the timber lagging after the design final
excavation grade is reached. Single strips can be placed at designed horizontal spacings along the
wall. Where shotcrete is used in lieu of timber lagging, special considerations are required to insure
drainage behind the shotcrete. Typically, prefabricated vertical drains are installed in segments
against the soil face with spikes. The segments are spliced by shingling the next segment over the
previously placed length after each lift is complete. An overlap length of one strip width is adequate.

Where precast concrete facings are used, the space between the temporary wall face and the
permanent facing may be backfilled with gravel. The gravel backfill acts as drainage element.
Water intercepted in a drainage element flows downward to the base of the wall where it is removed
by collector pipes or conveyed through the permanent facing in longitudinal/outlet pipes or
weepholes.

In applications where subsurface flow rates are large, horizontal drains may be used to remove water
from behind the wall. A horizontal drain is a small diameter perforated pipe that is advanced into a
120
nearly horizontal drill hole in an existing slope. For example, an anchored wall constructed on or at
the base of a steep slope will likely interfere with pre-existing natural drainage paths. This
interference may cause hydrostatic pressures resulting from trapped water to build-up against the
wall. To relieve these pressures, horizontal drains can be installed at appropriate vertical and
horizontal spacing along the wall alignment. Horizontal drains extend back from the wall face a
sufficient distance to intercept subsurface flow beyond the critical potential failure surface. Several
factors related to the construction of horizontal drains have limited their use for anchored system
applications. These factors are described below.

Horizontal drains should not be installed until the final excavated grade is reached unless a
perched water table exists above the final excavated grade. This higher drain installation
may result in water flowing into the excavation during construction.

The alignment of the drains must be carefully controlled to avoid interference with the
ground anchors. Splaying of multiple drains from a single entry point is not recommended.

Horizontal drains usually cannot achieve lowering of the water to a finished road grade as
the lowest elevation at the wall or slope face is controlled by construction equipment height
and the drains are sloped upward.

Special designs are required to collect the effluent from the drains to preserve aesthetics of
the wall face.

Surface drainage for anchored walls is usually achieved by directing water away from the wall face
either by grading or by collecting and transporting surface water in ditches or pipes. To minimize
surface water that can enter the excavation during construction and weaken the soils inside the
excavation, dikes can be constructed on the ground surface near the top of the wall or the vertical
wall element can be extended above the ground surface grade.

5.11.3 Wall System Appurtenances

Pre-existing and proposed appurtenances may have a significant effect on design, construction, and
cost of an anchored system and should therefore be identified during the early stages of project
implementation. Examples of appurtenances for wall systems associated with highway applications
include: (1) pre-existing and proposed facilities such as underground utilities and drainage systems;
(2) traffic barriers and parapet walls; and (3) noise walls.

As part of a site investigation, all pre-existing and proposed facilities that might affect wall system
design and construction need to be identified and located. Underground utilities such as telephone
cables and gas and water lines located in close proximity to the proposed wall system alignment may
become overly stressed and damaged as a result of abrupt changes in vertical and horizontal
deformation of the wall system. In such cases, it may be necessary to relocate the utilities or
incorporate protective measures during construction, either of which will increase overall
construction time and wall system cost. The location of underground utilities will influence the
inclination and spacing of anchors, and therefore the overall design and sequence of construction.

Earth pressures resulting from dead weight and impact loads from traffic barriers and parapet walls
must be accounted for in the design of a wall system. Loading requirements are provided in
121
AASHTO (1996). Noise walls are often incorporated into earth retaining system designs for urban
areas. The foundation of a noise wall is designed to resist lateral forces resulting from wind loads.
Noise walls may be integrally cast to anchored walls or they may be designed with a foundation that
is independent of the anchored wall.

5.11.4 Resisting the Upper Anchor Test Load

When the ground behind the upper portion of the wall is disturbed or the ground anchor load is high,
the soldier beam may deflect excessively during testing of the upper ground anchor. To resist the
applied test load, the ground behind the soldier beam must develop sufficient passive resistance. For
all wall designs, the passive capacity of the ground at the location of the uppermost anchor must be
checked.

The passive capacity of the soldier beam required to resist the test load applied to the upper ground
anchor may be calculated using equation 48 (FHWA-RD-97-103, 1998). For this calculation, it is
assumed that the passive resistance, FP, will be developed over a depth of 1.5 times the distance to
the upper ground anchor.
2
Fp 1.125K P h1 s (Equation 48)

In equation 48, KP is determined using either figure 16 or 17, and h1 is the depth to the upper ground
anchor. In using equation 48, a factor of safety of 1.5 is applied to the maximum capacity to obtain
the allowable resistance. The allowable resistance should be greater than the upper ground anchor
test load.

5.11.5 Anchored Walls for Fill Applications

Anchored walls for highway applications are most often constructed from the top of the wall to the
base of the excavation (i.e., top-down construction). Anchored walls have been constructed in fill
situations from the base of the excavation to the top of the wall (i.e., bottom-up construction). This
construction method only has application in rehabilitating existing walls. Examples of wall
rehabilitation with anchors are shown in FHWA-DP-90-068-003 (1990). New construction of walls
in fill is usually accomplished by employing mechanically stabilized earth techniques. Significant
differences exist with respect to the design, construction, and anchor load testing for an anchored
wall built from the bottom-up as compared to a wall built from the top-down. This section
highlights several of these differences.

The sequence of construction for a fill anchored wall with, for example, two levels of ground
anchors, can be described as follows:

Install the soldier beams or, in the case of most wall repairs, determine if the existing wall
can sustain the concentrated anchor load.

Backfill behind the wall and place lagging as required concurrently up to approximately the
midheight between the bottom level anchors and the top level anchors.

Install the bottom level of anchors.

122
Stress the bottom level ground anchors to a load that will not result in significant inward
wall movement. This load may be less than the design lock-off load.

Backfill behind the wall and place lagging as required concurrently up to a minimum of 1 m
above the level of the top anchors.

Restress the bottom level anchors to the designed lock-off load.

Install and temporarily stress the top level ground anchors.

Backfill and place lagging up to finished grade.

Restress the top level ground anchors to the designed lock-off load.

When constructing fill anchored walls, use select backfill material to permit compaction at low
energies to specified density requirements. Small compaction equipment should be used to avoid
damaging the tendons. If the wall backfill settles significantly as a result of poor backfill material or
compaction, the anchors will be subjected to bending forces at the anchor/soldier beam connection.
Anchors are not designed to carry significant bending forces.

Design loadings for fill anchored walls are based on earth pressures acting on the wall when the wall
is completely backfilled and all surcharge loadings are applied. During initial anchor installation, the
backfill may not reach the necessary height to permit the anchors to be load tested to 133 percent of
the design load at this stage. Typically, the anchors will be stressed to a small nominal load and
temporarily locked-off to remove slack from the anchors. As additional increments of backfill are
placed, the loads in the lower anchors will likely increase above the small nominal lock-off load and
the wall will deflect outward unless restressing is performed. After the backfill has been placed to
finished grade, the anchors may be able to be load tested to 133 percent of the design load if
sufficient passive resistance is available and if the wall face can sustain the test load.

With this type of incremental backfilling and staged load testing, the ground anchors will typically be
designed to carry actual earth pressure loads as compared to loads from apparent earth pressure
envelopes as may be used for anchored systems constructed from the top-down. The pattern of wall
movement for a fill anchored wall is consistent with theoretical earth pressure envelopes.

Standard anchor testing may not be possible in the case of wall rehabilitation. In that case, it is
necessary to move to an area on the site and install preproduction anchors through ground similar to
that for the production anchors. These anchors should be subjected to performance test requirements
and then loaded to 200 percent of production anchor design loads. If these preproduction anchors
pass acceptability criteria, then it is concluded that the production anchors for the fill wall would
pass acceptability criteria at 133 percent of the design load.

123