6.

RETAINING WALLS
Retaining walls can be broadly split into two categories. Those that rely on their weight for the
stability of the wall (Gravity walls), and those that mobilise earth pressures in the ground to
provide resistance (Embedded walls). Within each category there are a variety of wall types.
The selection of the appropriate wall type depends on many factors that include:
Soil and groundwater conditions
Height and ground topography
Availability of suitable fill material
Construction constraints (space, access, equipment, specialist techniques available)
Environment appearance and impact during construction
Ground movements and their affects on adjacent structures
Underground obstructions and services
Design life and maintenance requirements
Cost
These notes are primarily concerned with the general design methods used for the two types of
wall. Further details of individual wall types and their advantages and disadvantages may be
found in many texts on retaining wall and foundation design. For retaining walls used to support
excavations particular attention should always be given to the effects of groundwater. Failure to
consider this can lead to failure of the soil-wall system by mechanisms not always considered in
standard design calculations. For instance, groundwater lowering will lead to settlements which
may damage adjacent services and structures, groundwater flow may lead to erosion and piping
at the base of excavations, and groundwater pressures may cause heave into an excavation.
6.1 Gravity Walls
Gravity walls are generally used to retain soil above the existing ground level. The simplest
walls rely on the mass of the wall for stability. These include walls made of mass concrete,
concrete with masonry facing, unreinforced masonry (bricks and stone), gabions (wire baskets
filled with stone), and crib walls (hollow crib formwork filled with soil). These types of wall are
common for small retained heights up to 3 m, and are rare for heights greater than 8 m. For
walls between 3 and 8 m precast reinforced concrete (cantilever) walls are very common. These
walls are usually in the shape of an L or inverted T. Reinforced soil walls are also widely used.
These use strips of steel or plastic placed in the soil connected to facing elements that retain the
soil. Friction between the reinforcing strips and the soil provides the resistance to hold up the
facing elements. To mobilise the soil resistance some movement must occur and reinforced soil
walls are therefore more flexible and require relatively large tolerances to ground movement.
Soil nailed walls are similar to reinforced soil but are used to support the soil face during
excavation.
There are four principal modes of failure that need to be analysed for any gravity wall. These
are
Translation
Overturning
Bearing capacity
Overall failure of the soil and wall
In addition it is generally necessary to check that the wall deformations and the ground
movements will not be excessive.
6.1.1 Translation

Translation is the mode of failure where the wall slides because the frictional force, F, is less
than the force due to the difference in the active and passive pressures. The active and passive
pressures can be determined from either Rankines method or from a limit equilibrium method.
It is found that the factor of safety is very dependent on any passive pressures developed in front
of the wall. Because of this it is normal to ignore the upper 0.5 to 1 m of soil contributing to the
passive pressures. This reduces the possibility of inadvertent excavation leading to failure.
6.1.2 Overturning
If the wall height becomes large then there will be a significant moment due to the active earth
pressures. In the limit the wall will topple about the toe, point A in the diagram above. At this
limit the overturning moment due to the earth pressures must be balanced by the restoring
moment due to the weight of the wall.
6.1.3 Bearing capacity
If the stress due to the weight of the wall is large there is the possibility that the underlying soil
will not be able to support it. This is known as a bearing capacity failure. Section 7 of these
notes discusses the bearing capacity in more detail. It should be noted that due to the earth
pressures acting on the wall there will be a moment (eccentricity of the normal load) and
horizontal force acting on the base of the wall. This moment and horizontal load will
significantly reduce the bearing capacity (vertical stress) that the soil can support. One method
of allowing for these loads is given in the Soil Mechanics Data Sheets (p74, 75). The general
bearing capacity formula includes reduction factors that account for the load inclination
(horizontal loads) and load eccentricity. The moment is allowed for by using an effective
foundation width B (= B 2e), where e is the eccentricity of the load, in the correction factors
for load inclination.
Active Earth
Pressures
Passive Earth
Pressures
F
N
A
W
6.1.4 Overall failure
A check is required on the overall stability of the soil and wall combined to check that a failure
surface will not occur in the soil. This may be analysed using the methods discussed previously
for assessing slope stability. This may include checking a rotational failure mechanism as shown
above, and possibly a wedge mechanism if there are weak layers at some depth beneath the
wall.
6.2 Embedded retaining walls
Embedded walls are generally used for construction from the ground level down. They can be
partly driven and then backfilled, or fully driven or constructed in-situ followed by excavation.
There are four main construction methods: walls constructed of sheets of timber, steel or
concrete; soldier or king piles with sheeting placed between the piles; bored pile walls; and
diaphragm walls. Each wall type may act as a cantilever or be supported by one or more rows of
anchors or props. They can be used either as temporary supports during construction, or for
permanent structures such as quay or basement walls. The walls range from relatively flexible
steel sheet piles to relatively stiff diaphragm walls. These walls are generally more expensive
than gravity walls but their cost is balanced by the speed of construction and lack of temporary
support. Cantilever walls are only suitable for moderate retained heights, typically less than 5 m,
but if a stiff reinforced concrete wall is formed may be suitable to about 10 m. Significant
ground movements can occur behind cantilever walls, and they are generally unsuitable if
services or foundations of adjacent buildings are close. The use of anchors or props can reduce
the required penetration length, the ground deformation and the bending moments in the walls
From the design viewpoint we can split these sheet pile walls into three groups
1. Cantilever Walls
2. Walls with a single anchor or prop
3. Walls with multiple props
For any wall type we need to consider:
The overall stability of the soil/wall system
The structural strength of the wall
The possibility of damage to adjacent structures, and services in the ground, due to wall
construction
6.1.4 Overall failure
6.2.1 Cantilever walls

When designing sheet retaining walls it is normal to assume that the effective lateral stresses
acting on the wall are given by simple RANKINE active and passive zones. Friction on the
wall is usually ignored as this leads to conservative (safe) designs.
6.2.2 Rankine Active and Passive Pressures
The earth pressures acting on the wall are strongly dependent on the deformations in the
surrounding soil. When the wall moves away from the soil the stress on the wall drops reaching
a minimum, the ACTIVE pressure, with the soil deforming plastically. When the wall moves
into the soil the stress increases, finally reaching a maximum, the PASSIVE pressure, when
again the soil is deforming plastically.
Excavation
Direction of
wall movement
Active pressures
Passive
pressures
Direction of
wall movement
Active
Passive

h
Wall assumed frictionless - Then vertical and
horizontal stresses are principal stresses

c + tan

hmin

hmax

1 3
= N + 2 c N
For most retaining walls the long term, fully drained, situation usually governs the wall
stability. For the analysis of fully drained conditions the Mohr-Coulomb criterion needs to be
expressed in terms of effective stress using the effective strength parameters c and . For
design it is also conservative to use the critical state strength parameters, that is c = 0 and =

cs
. The effective lateral stresses on the wall are then
ACTIVE

h
v
v a v
N
K
1
1
sin
sin
PASSIVE

+

h v v p v
N K
1
1
sin
sin
K
a
and K
p
are known as the active and passive earth pressure coefficients. For soil at failure the
earth pressure coefficients are simply related by
K
K
a
p

1
.
For any vertical wall it is possible to relate the horizontal effective stress to the vertical
effective stress, determined from the vertical overburden, by an earth pressure coefficient. The
coefficient will depend on the slope of the soil surface and the wall roughness. Published
values are available for many situations.
6.2.3 Stability - Limiting Equilibrium
When assessing the stability it is normal to assume triangular pressure distributions, and this is
in fact quite realistic if the wall is rigid. For a cantilever wall the stresses acting at failure will
then be as shown below, with the wall rotating about a point just above the toe of the wall. The
stability of the wall depends mainly on the passive force developed below the excavation.
Geometry Pressure Diagram
For design we need to determine the required depth of penetration for stability and then to size
the wall to resist the maximum moment. To determine the depth of penetration required for a
given height H we need to consider both moment and force equilibrium:
F = 0
M = 0
If the soil is dry the pressures and forces are as shown below
Where
P K x H
A a d 1
2
1
2
+ ( )
P K x
P P d 1
2
1
2

P K x d x K d x
A a d a d 2
2
1
2
+ ( ) ( )
H
x
d
Point of
Rotation
Active
Passive
Active
Passive

P
A1
P
P1
P
P2
P
A2
Pressures
Forces

P K x H d x K d x
P p d p d 2
2
1
2
+ + ( ) ( ) ( )
From equilibrium
F = 0 : P
A1
+ P
P2
- P
P1
- P
A2
= 0
This gives a quadratic equation involving terms in x
2
and d
2
M = 0: Taking moments about the point of rotation
P
x H
P
d x
P
x
P
d x
A A P P 1 2 1 2
3 2 3 2
+

1
]
1
+

1
]
1
+

1
]
1
This gives a cubic equation involving terms in x
3
and d
3
.
We have 2 equations with 2 unknowns, x and d, and hence we can determine the required depth
of penetration for the wall. The equations can be solved graphically or by computer.
Alternatively simplifying assumptions about the forces below the pivot can be made to enable
analytical solutions to be obtained as described in many text books.

As an illustration consider a wall with H = 1.8 m placed in dry soil with
d
= 19 kN/m
3
and =
30
o
. For = 30
o
K
p
= 3, K
a
= 0.3333 and the required depth of penetration d = 1.767 m.
6.2.4 Serviceability - Design requirements
By considering the stability we can obtain the limiting stresses on the wall, but the wall would
have been considered to have failed from a serviceability viewpoint well before this, owing to
large settlements in the supported soil. The design approach is to factor the earth pressures.
There are two main design approaches which are both based on the knowledge that the earth
pressures acting on the wall are strongly dependent on the deformations in the surrounding soil.
The movements required to reach the active and passive conditions depend on the soil type and
can be quite different. For example, for retaining walls of height H the movements required are
approximately:
SAND Active 0.001H
Passive 0.05H - 0.1H
CLAY Normally Consolidated Active 0.004H
Passive large
Over-Consolidated Active 0.025H
Passive 0.025H
Method 1 - Sands and normally consolidated clay
Assume that sufficient movement occurs to allow active (minimum) pressures to develop, then
factor the effective passive pressures by 2. Note that where insufficient movement of the wall
occurs the active pressures will not reach a minimum and higher pressures will act on the wall.
These must be allowed for in design as they can influence the required structural strength.
Force Equilibrium
Moment Equilibrium
Consider the same wall as above with H = 1.8,
d
= 19 kN/m
3
, = 30
o
The pressure diagram looks identical but the passive pressures are reduced by using a reduced
value of the passive earth pressure coefficient, K
*
p
.
Where K
*
p
= K
p
/2 = 1.5 and K
*
a
= K
a
= 0.3333 as before.
Hence d = 2.94 m
The total depth of sheet pile required = 1.8 + 2.94 = 4.74 m
Some texts recommend increasing the depth of penetration by a further 10-20% to allow for
uncertainties in the analysis. Alternatively some design codes recommend assuming the top 0.5
m of the soil beneath the excavation provides no restraining effect.
Method 2 - Over-consolidated clays
Here both active and passive pressures are developed for similar movements and both are
factored. This is achieved by dividing tan by a Factor F

, and using the reduced angle of

friction when calculating the earth pressure coefficients K
*
a, K
*
p
. The factor F

can be in the
range from 1.2 to 1.5, depending on the allowable settlement and soil type, but is usually taken
as 1.3.
As for method 1 it is assumed that the shape of the pressure diagram is similar to that at limiting
equilibrium, but in this case the passive pressures are reduced and the active pressures increased.
Using the same parameters as previously H = 1.8,
d
= 19 kN/m
3
, = 30
o
Calculate
*
from
tan
tan tan ( )
.
*

F
30
13
Hence
*
= 23.95
o
and K
*
a
= 0.423, K
*
p
= 1/K
*
a
= 2.366
Then the required depth of penetration becomes d = 2.46 m
6.2.4 Structural strength
Having determined the required depth of penetration, the next stage in design is to calculate the
maximum moment in the wall so that an appropriate wall thickness and strength can be selected.
The position down the wall of the maximum moment can be found by determining where the
shear stress in the wall is zero. ( F
dM
dz
)
Consider a free body diagram of a section of the wall


H
z
F K z H K z
a d p d
+
1
2
1
2
0
2 2
( )
( ) K K z K z H K H
p a a a

2 2
2 0
A quadratic equation that can be solved for z using appropriate (factored) values for K
p
, K
a
.
Then taking moments M K z
z
K z H
z H
p d a d
+
+ 1
2 3
1
2 3
2 2
( )
( )
With H = 1.8,
d
= 19 kN/m
3
, = 30
o
and using K
*
p
= K
p
/2
z = 1.605 m M = 22.0 kNm/m
Note that as the factor of safety increases the maximum moment also increases.
The factor of safety can be dramatically reduced by surcharge loadings on the supported ground
next to the wall. For a uniform surcharge then the effective active pressure can be increased by
K
a

s
, while for a concentrated load from a footing the Coulomb method of trial wedges can be
used to determine the active force on the wall. In the latter situation allowance must be made
for the fact that the point of application of the load will also change.
Consideration must also be given to the water pressures acting on the wall.

s
F
+
v s d
z
M
+
h a s d
K z ( )
Q
L
P
A
The force P
A
can be estimated using the
method of wedges. The line of action
can be estimated using elastic solutions
Water
Water
Table
Effective stresses must be used in
evaluating the lateral stresses from
Rankines method

h v
K
Pore water pressures are the same on
each side of the wall so their effects
cancel when considering force and
moment equilibrium
Total vertical stress,

v sat
z
Water pressures can be determined from flow net
Hence
v v h v
u a n d K
Forces due to water pressures are different on the
two sides of the wall so their effects must be
included when considering force and moment
equilibrium.
For economic reasons cantilever walls are usually limited to excavations less than 6 m deep.
They are often used to support low banks of free draining sand and gravel soils.
They are not suitable for the long term support of soft clayey soils (clay or silt)
Corrosion can also be a problem with steel sheet piles.
6.3 Anchored Walls
6.3.1 Single anchor (or prop)
When a cantilevered sheet wall is unsuitable, for example because the height is too great or
because the deflections need to be limited. Anchors may be used to improve the stability, with
the anchors often placed close to the top of the wall.
The anchor force introduces a further unknown into our equations, so that an assumption is
required. There are two main methods of analysis:
1. Free earth support method The base of the pile is assumed to be free to rotate and move
laterally
2. Fixed earth support method The base of the pile is assumed to be fixed in position and
direction
The appropriate method depends on the relative stiffness of the wall/soil system. For a
relatively rigid system (ie. a heavy wall section in a loose sand) the earth pressure distribution
corresponds closely to the triangular active and passive conditions. If the wall is rigid the toe of
the wall will be able to move and rotate, and the free earth support method is appropriate.
As the stiffness of the system decreases the pressure distribution alters in such a way as to
reduce the bending moment in the sheet pile, and as a consequence, the wall section may be
reduced as compared with an infinitely stiff wall. The design procedure is usually to use the free
earth support method and then use empirical moment reduction methods to determine the wall
section required.
For a very flexible wall the fixed earth support method can be used. The analysis is more
complex than the free earth support method and will not be considered here.
6.3.1 The Free Earth Support Method
Steps in analysis for Design
Determine the effective vertical stresses
Determine the effective lateral stresses assuming Rankine active and passive pressures
Factor the lateral pressures to limit the deformations - either by factoring K
p
or by factoring

Deflected
position of
the wall
H
d
Anchor or
Prop between
sides of
excavation
T
P
A
P
P
Add in water pressures if water levels different on two sides of the wall
Take moments about the anchor/strut to determine the required depth of the wall
Use force equilibrium to determine the anchor force
Design anchor to withstand the force
Determine maximum moment in the wall and check that section is acceptable
Anchors are typically spaced 2 - 3 m apart, and the load is distributed along the wall by walings
running either behind, or in front of the sheet pile walls and bolted to them.
Accurate analysis of sheet walls is complicated by the interaction between the soil and the wall.
In practice walls are not perfectly rigid as assumed in the free earth support method and it is
important to consider the effects of wall flexibility. If the wall deforms this will influence the
pressures mobilised between the soil and the wall and consequently the anchor force and
moments in the wall.
Remember that it is important to ensure that the wall movements are compatible with the design
assumptions.
6.3.2 Multiple anchors
Where there are relatively deep temporary excavations it is common to support the walls during
construction by a system of bracing. This procedure is also used for permanent structures with
the struts forming the floors of the basement. Alternatively the walls can be supported by
multiple anchors.
A wall with several layers of struts or anchors will have increased restraint as each layer of
anchors is added. Consequently the lateral deformations are limited and the retained soil is
unlikely to attain failure. The situation is statically indeterminate and analysis is complex. The
earth pressure that acts on the wall will depend on:
relative stiffness of soil and wall
anchor/strut spacing
load-deformation response of the anchor or strut
pre-stress (if any) in the anchor/strut during construction
In practice empirical methods are used to estimate the pressures on the wall and forces in the
strut, and these methods are based on actual measurements.
Rigid wall
Flexible wall
Pressure distribution on
flexible wall
6.3.3 Anchor design
The anchor must be able to provide resistance equal to the required anchor force without
excessive displacement of the anchorage towards the wall.
There are many anchoring systems used in practice. They rely on a combination of bearing
pressures on the faces perpendicular to the anchor, and frictional forces between the anchor and
the soil. The simplest is the vertical plate anchor.

It is assumed that the resistance can be determined simply from the difference between the
passive and active pressures on the two sides of the plate. For a plate of area, A, the anchor
force is
T K K A
p v a v
( )
/m of the wall
However, to mobilise the full passive pressure significant movement of the plate would be
required. To reduce the movement the pressures should be factored as discussed above for the
wall.
If the area of the plate anchor is large it will probably be more economic to use raked pile
anchors. By installing the anchor at depth the normal stresses and hence the frictional resistance
will be much greater than at the surface.
Passive
pressures
Active
pressures
Example 1
Consider the limiting forces acting on a single strut supporting a wall retaining dry sand
Analysis requires several assumptions
Rigid wall
Rigid (unyielding) strut
Triangular active and passive pressures - no friction
sufficient wall movements
There are two possible modes of failure depending on the position of the strut.
Strut near the surface Strut near the base of the wall
Consider the limiting equilibrium of the wall. To eliminate the unknown strut force take
moments about the strut.
Strut at surface
1
2 3 2
1
2
2
3
1
2 1
1 2
2
2
2 2
K d
d
K d d
d
K d
d
p a a
+
Strut at base
1
2 3 2
1
2
2
3
1
2 1
1 2
2
2
2 2
K d
d
K d d
d
K d
d
a p p
+
Noting that K
a
= 1 / K
p
then after rearrangement we obtain for the strut near the surface
2
d
d
+ 3
d
d
-
K
= 0
3
2
1
2
2
1
2
p

_
,

_
,

Dry Sand
d
1
d
2
Strut
passive
active
active
passive
deformed
position of
wall
and for the strut near the base of the wall
2
d
d
+ 3
d
d
-
K
= 0
3
2
1
2
2
1
2
a

_
,

_
,

For ' = 30
o
we obtain solutions for d
2
/d
1
of 1.275 and 0.182, or if D is the total height of the
wall d
1
/ D of 0.44 and 0.85.
In practice several struts would probably be used because of serviceability concerns. As
discussed above it is difficult to accurately assess the loads on walls with many struts, and
failure of individual struts may occur. This simple analysis indicates where the struts could be
positioned to avoid progressive failure. For example, if 3 struts are used and placed at depths of
0.25D, 0.5D and 0.75D then either the top two or the bottom strut can be removed without the
wall failing.
In the analysis it has been assumed that the strut is unyielding, so that the wall rotates about the
strut. To determine whether this is a reasonable assumption we need to check the force in the
strut. For the strut near the surface and the wall at its limiting equilibrium
F K d K d d K d
p a a
+ +
1
2
1
2
1
2
1 2 2
2

For the example with ' = 30
o
we have found d
1
= 0.44 D, d
2
= 0.56 D and hence
F = 0.425
D
2

For a retained height of 2 m and = 18 kN/m

3
we find F 30 kN/m
Let us assume that we have a 4 m wide trench with struts every 4 m and that the bracing is
provided by steel (scaffolding) tubes 50 mm diameter by 5 mm thick.
The force in each strut will be F = 4 30 = 120 kN
and the strain =
F
AE


7 10
4
and the displacement = 2.8 mm
0.44 D
Wall will be stable
if strut is within
this region
0.85 D
D
Now a displacement of this magnitude is sufficient to cause the stress to drop to active
conditions at the prop. The wall movements are thus not compatible with our initial
assumption. The effect will be for the point of rotation to move up the wall and for premature
failure to occur.
This is a simple illustration of the importance of accounting for soil-structure interaction if
possible
Example 2
An underground development is planned adjacent to existing structures which can be considered
to apply a uniform stress of 100 kPa to the soil surface.
To try to minimise settlement of the surrounding structures it is proposed to construct rigid
diaphragm walls, 8 m deep in the uniform sandy soil which has properties c = 0, = 30
o
,
dry
= 18 kN/m
3
,
sat
= 20 kN/m
3
. The water table is at a depth of 5 m below the ground surface.
The proposed "top down" construction procedure is to place a slab at ground level to act as a
strut to support the wall, and then excavate a further 3 m before placing the next floor slab.
Determine the horizontal force per metre in the slab at the ground surface when the excavation
has reached 3 m, just before the next slab is placed.
Assumptions:
Rigid wall
Free earth support method
Triangular pressure distributions
Movements sufficient for active pressures to develop
actual point of rotation
position of strut
Step 1: Determine the pressures acting on the wall
Step 2 - Take moments about the strut
Moment equilibrium is required to find the factor applied to K
p
. The factor is unknown because
this is not the limiting case
( ) K
a d d sat w
100 8 4
1
2
5
2
3
5 5 3 65
1
2
3 7
2 2
+ + +

1
]
1
.
=
K
F
p
d d sat w
1
2
2
13
3
2 3 65
1
2
3 7
2 2
+ +

1
]
1
. ( )
Given
d
= 18,
w
= 10,
sat
= 20 kN/m
3
and = 30
o
, K
a
= 0.3333, K
p
= 3
2006 7 1173
1754
.
.

K
F
F
p
Step 3 - Force equilibrium to obtain strut force
S K
K
F
a
p
+ + + + + [ . . ] [ . . ] 100 8 05 18 5 18 5 3 05 10 3 05 18 2 18 2 3 05 10 3
2 2 2 2
S = 123.3 kN/m
S
100 kPa
3 m
5 m
5 m
K
a
100
K
a

d
5
K
a
(
sat
-
w
) 3
K
p
(
sat
-
w
) 3
K
p

d
2
Sheet Retaining Wall problems
1. A quay wall has been built from sheet piling and is to retain 8 m of sand which has
strength properties c' = 0, ' = 33, a bulk unit weight of 16 kN/m
3
, and a saturated unit
weight of 18 kN/m
3
. The wall is anchored 1 m below the top of the wall and has a total
length of 15 m. The water table on both sides of the wall is at a level 4 m below the top
of the wall.
Cargo is to be stored on the quay, which may be assumed to apply a uniform surcharge
to the surface of the sand. Determine the maximum magnitude of the surcharge loading
that can be applied by the cargo so that the factor of safety applied to the passive
pressures does not fall below 1.3. It may be assumed that the wall movements are
sufficient for the active pressures to be fully mobilised.
Calculate the maximum moment in the sheet pile wall when this maximum surcharge is
applied.
Explain why such a low factor of safety may give rise to problems with the quay.
2. For the quay wall described in question 1, calculate the factor of safety F

(applied to
tan affecting both active and passive pressures) if the surcharge is 25 kPa.
3. A cantilevered wall has been used to retain 2 m of a sandy soil, which has strength
properties c' = 0, ' = 35, and a dry unit weight of 18 kN/m
3
. The wall penetrates 4 m
below the base of the retained soil, into the same sandy soil.
It is proposed to raise the level of the retained soil for a new development, by adding fill
with a dry unit weight of 14 kN/m
3
. Calculate the maximum height of fill that can be
added if the factor of safety against passive failure is not to fall below 1.5. The fill may
be assumed to apply a uniform surcharge to the retained soil.