Based on notes prepared by Dr Charles MacRobert

Lateral Earth Pressure

Introduction

Lateral earth pressure theory is used to analyse the forces acting on near-vertical walls that
retain soil, of which there are two main categories; rigid and flexible walls. Flexible walls are used as
temporary support for deep excavations. Rigid walls are used as permanent support, however they
are becoming obsolete with soil reinforcement or mechanically stabilised earth (MSE) used instead.

Box 1: Types of retaining walls

Rigid Walls

These are walls that consist of concrete or masonry walls and rely on gravity for stability. Four
different retaining walls are shown. The design of these walls requires taking into account; a) Sliding
or translation failure, b) Rotational failure, c) Deep seated failure, and d) Structural failure, as
illustrated:

Flexible Walls

Flexible retaining walls consist of slender members of steel, concrete, or wood and rely on passive
soil resistance, props or anchors for stability. The design of these walls requires taking into account;
a) Deep seated failure, b) Rotation about the prop or anchor, c) Bearing failure (Rotation) about the
base, d) Failure of anchor/prop, and e) failure by bending, as illustrated:

1
 Mechanically Stabilised Earth

The concept of reinforcing soil was accidentally discovered by Henri Vidal in 1967, when he
disocovered that, with his son, they could build steeper sand castles when incorporating strips of
palm leaves. This works because the reinforcement mobilises friction at the interface with the sand,
and therefore increases the lateral effective confining stress. This effect is shown using a Mohr’s
circle below. For the MSE the Mohr’s circle of stress is smaller and further away from the failure
line and is therefore able to resist larger shear stresses than ordinary soil.

The design of retaining walls requires the analysis of all possible failure modes and is therefore
so vast that it cannot be fully covered in a single course. This course will therefore concentrate on
general concepts and some of the design aspects associated with flexible walls.

Lateral Pressures

Up to now we have only considered the vertical stresses in soils, very simplistically the lateral
stress (i.e. horizontal stresses) are determined by multiplying the vertical stress by a suitable
coefficient:

𝜎𝜎ℎ = 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 × 𝜎𝜎𝑣𝑣

The need to be able to calculate lateral earth pressures arises particularly in the design of
vertical or near-vertical retaining walls. Owing to the difficulties of quantifying the detailed stress-strain
behaviour of a soil, the approach to the calculation of lateral earth pressures in geotechnical
engineering has traditionally been based on the concepts of plasticity. By considering the stress states
of soil elements on the verge of failure, it is possible to deduce the limits between which the lateral
earth pressures must lie. The three stress states considered are:

1. At Rest State, K0. At rest stress conditions occur when the lateral strain is zero.

2
 2. Active State, Ka. Active conditions correspond to failure by lateral expansion of the soil
such that the major principal stress is vertical, and represents the minimum possible lateral
stress for a given vertical stress – usually associated with conditions behind a retaining wall.
3. Passive State, Kp. Passive conditions correspond to failure by lateral compression of the
soil such that the major principal stress is horizontal, and represent the maximum possible
lateral stress for a given vertical stress – usually associated with conditions in front of a
retaining wall.

Figure 1 below illustrates a retaining wall initially vertical, with soil on either side in an at rest
state. As the wall rotates about its base the soil to the right of the wall expands and at failure will be
in the active state. Soil on the left of the wall compresses and at failure will be in the passive state.

Figure 1: Relationship between lateral strain and lateral pressure coefficient

At rest state

If a wall is static – that is, if it does not move either to the right or to the left of its initial position
– the soil mass will be in a state of elastic equilibrium; that is, the horizontal strain is 0. The ratio of
the effective horizontal stress to the vertical stress is called the coefficient of earth pressure at rest¸
K0, or:
𝜎𝜎ℎ′
𝐾𝐾0 = ′
𝜎𝜎𝑣𝑣

Box 2: Coefficient of earth pressure at rest

Coarse Grained Soils

The coefficient of earth pressure at rest can be

estimated from the empirical relationship proposed K 0 =1- sin φ'
by Jacky, 1944:

For overconsolidated coarse grained soils this

equation is modified as proposed by Mayne and K 0 =(1- sin φ' )(OCR)sin φ'
Kulhway, 1982:

3
 Fine Grained Soils

The coefficient of earth pressure at rest can be

estimated from the empirical relationship proposed 𝑃𝑃𝑃𝑃(%)
by Massarsch, 1979: 𝐾𝐾0 = 0.44 + 0.42 � �
100

For over consolidated clays, the coefficient of earth K 0(overconsolidated) =K 0(normally consolidated) √OCR
pressure at rest can be approximated as:

OCR is the overconsolidation ratio which is defined as:

preconsolidation pressure
OCR=
present effective overburden pressure

Upper and lower bounds

The active and passive forces at collapse can be determined by either the upper or lower bound
theorem of plasticity.

The lower bound theorem states that if a system of stresses within the soil mass can be found
which is in equilibrium with the external loads and body forces (i.e. self-weight), and nowhere violates
the failure criterion for the soil, then the external loads and body forces represent a lower bound to
those which will actually cause collapse. Any error will be on the safe side, since it may be that a
different system of stresses will carry even higher loads without violating the failure criterion. In short,
if it can be shown that the soil can carry the loads, then it will. For a stress field based solution to be
a true lower bound the stress field needs to extend to infinity, however it is common practice to focus
on the stress acting at the wall. This leads to a limit equilibrium method, such as was originally used
by Rankine (1857) to determine the lower bound to the failure of a retaining wall. This is the approach
that will be followed in these notes.

The upper bound theorem states that if a mechanism can be found such that the work done by
the external loads and body forces is equal to the energy dissipated within the soil mass as it deforms,
then the external loads and body forces represent an upper bound to those which will cause collapse.
Any error will be on the unsafe side, since it may be that a different mechanism will already have been
responsible for collapse at a lower load. In short, if it can be shown that the soil can fail, then it will. If
the mechanism consists entirely of non-distorting rigid blocks, then a similar answer can be obtained
by considering static equilibrium. However in many cases some of the blocks assumed to be rigid
may dissipate energy in distortion which would not be taken into account. This would therefore not be
a true upper bound and would also be termed a limit equilibrium solution. Coulomb (1776) originally
used this method to analyse retaining walls by assuming a mechanism of collapse. This solution is
analogous to how slope stability is analysed and will therefore not be covered at this point.

Mohr circle of stress

A state of plane stress within a soil (or indeed any solid material) can be characterised by means
of the Mohr circle of stress, plotted on a diagram with axes representing the normal stress, σ (in the
x-direction) and the shear stress τ (in the y-direction). The Mohr circle shows how the combination of
τ and σ on a given line within the plane varies with the orientation of that line. The Mohr circle is in
effect a locus defining all possible combinations of τ and σ, acting on lines at different orientations,

4
that make up a given state of stress within a plane. In the current context, the plane is usually the
cross-sectional plane of a long retaining wall. The circle is symmetric about the σ-axis: the two points
at which it crosses the σ-axis represent principal directions, on which the shear stress τ is zero. The
corresponding normal stresses are termed the major (greater) and minor (lesser) principal stresses.
The stress state on a line at an angle θ anticlockwise from the line on which the major principal stress
σ1 acts is given by the point on the circle whose radius subtends an angle 2θ at the centre of the circle
with the radius through the major principal stress point. This is illustrated in Figure 2. The angle 2θ
between the radii corresponding to the major and the minor principal stresses is 180°, which means
that in reality their direction (or the directions of the lines on which they act) are at 90°.

Figure 2: Mohr circle representation of the stress states within a principal plane

In soils, the Mohr circle may be plotted in terms of either total stress (σ, τ) or effective stress (σ’,
τ), depending on the analysis being used. The choice of whether to consider total stresses or effective
stresses is dependent on the permeability of the material and the rate at which the load is applied.
For fine grained soils due to the low permeability during usual load application pore water pressures
cannot escape. Effective stresses therefore cannot change and remain constant, making it necessary
to consider total stresses. For coarse grained non-cohesive soils the higher permeability means that
pore pressures can dissipate during normal load application. In this case effective stresses change
with total stresses and it is necessary to consider effective stresses.

5
Total stress failure criterion

If the soil is in a state of limiting equilibrium, there will be some orientation within the plane for
which the combination of stresses (σ, τ) lies on the line defining states of stress at failure, or the failure
envelope. The total stress failure criterion specifies the maximum possible shear stress anywhere
within the plane, τmax as the undrained shear strength of the soil su. This only applies to clays that are
sheared rapidly to failure. The undrained shear strength su is not a soil property, but depends on the
water content of the clay (i.e. a state parameter). The undrained shear strength criterion is written as:

𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑠𝑠𝑢𝑢

Box 3: Total stress failure criterion

A Mohr circle of total stress which touches
the undrained shear strength failure
criterion is shown. Consideration of the
geometry of the Mohr circle shows that the
difference between the major and minor
principal total stresses, σ1 - σ3 is equal to
twice the radius of the circle, and the
radius is equal to su. Thus at failure:

𝜎𝜎1 − 𝜎𝜎3 = 2𝑠𝑠𝑢𝑢

(su and cu are the same)

Assuming the wall is frictionless then the principal total stress directions are vertical and horizontal,
there are two ways in which the soil can fail:

Active State Passive State

In the first, the major principal stress is
vertical, and the minor principal stress is
horizontal. This could correspond to the
withdrawal of lateral support at constant
vertical stress until failure is reached. In
active conditions, the (minimum)
horizontal total stress is given by:
In the second mode of failure, the major principal total
stress is horizontal, and the minor principal stress is
vertical. This could correspond to an increase in lateral
stress and a constant or possibly reducing vertical
stress, as might occur in the soil in front of an
embedded retaining wall. In passive conditions, the
(maximum) horizontal total stress is given by:

𝜎𝜎ℎ,𝑚𝑚𝑚𝑚𝑚𝑚 = 𝜎𝜎𝑣𝑣 − 2𝑠𝑠𝑢𝑢 𝜎𝜎ℎ,𝑚𝑚𝑚𝑚𝑚𝑚 = 𝜎𝜎𝑣𝑣 + 2𝑠𝑠𝑢𝑢

σv is the total vertical stress (should include effect of any surcharge q)
When calculating active earth pressures behind a wall two provisos apply:
a) Tensions stresses are not permitted. Negative values of σh,min result in a tension crack
penetrating to a depth ztc given by:
(2𝑠𝑠 −𝑞𝑞)
𝜎𝜎ℎ,𝑚𝑚𝑚𝑚𝑚𝑚 = (𝑞𝑞 + 𝛾𝛾𝛾𝛾𝑡𝑡𝑡𝑡 − 2𝑠𝑠𝑢𝑢 ) = 0 or 𝑧𝑧𝑡𝑡𝑡𝑡 = 𝑢𝑢𝛾𝛾
b) Flooded tension crack. If there is any tendency for the crack to become flooded the lateral
stress in the active zone becomes equal to hydrostatic pressure resulting in a deeper tension
crack:
(2𝑠𝑠 −𝑞𝑞)
𝜎𝜎ℎ,𝑚𝑚𝑚𝑚𝑚𝑚 = �𝑞𝑞 + 𝛾𝛾𝛾𝛾𝑓𝑓𝑓𝑓𝑓𝑓 − 2𝑠𝑠𝑢𝑢 � = 𝛾𝛾𝑤𝑤 𝑧𝑧𝑓𝑓𝑓𝑓𝑓𝑓 or 𝑧𝑧𝑓𝑓𝑓𝑓𝑓𝑓 = 𝑢𝑢 )
(𝛾𝛾−𝛾𝛾𝑤𝑤
As γw ≈ 0.5γ, if q = 0 the depth to which a flooded tension crack may remain open (and, more
importantly, hydrostatic pressures are exerted on the retaining wall), is zftc ≈ 4su/γ. The tendency
for su to increase with depth means a flooded tension crack can be more than twice as deep as a
dry tension crack. This will require an increased depth of embedment.

6
Effective stress failure criterion

The second and more fundamental failure criterion is in terms of effective stresses. All soils
derive their strength from interparticle friction. In natural soils, cementing between particles may also
be present, but such bonds will be brittle and enhance the peak strength, but not the strength at critical
state. In terms of effective stresses, most soils thus obey the purely frictional failure criterion:

𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚 = 𝜎𝜎 ′ tan ∅ ′

Box 4: Effective stress failure criterion

The Mohr circle of effective stress which touches the effective stress envelope (i.e. on the verge
of failure) is shown below.

From the geometry the distances s’ and t can be determined:

(𝜎𝜎1′ − 𝜎𝜎3′ ) (𝜎𝜎1′ + 𝜎𝜎3′ )

Radius: 𝑡𝑡 = Abscissa of Centre 𝑠𝑠 =
2 2

Further consideration of the Mohr circle geometry shows that 𝑡𝑡 = 𝑠𝑠 ′ sin ∅′. Substitution of the
expressions for t and s’ in terms of the major and minor principal stresses 𝜎𝜎1′ and 𝜎𝜎3′ into this
equation, followed by a little manipulation, gives the ratio of the major and minor principal stresses
𝜎𝜎1′ /𝜎𝜎3′, at failure, as:

σ1′ (1 + sin ∅′ )
=
σ′3 (1 − sin ∅′ )

If we assume the wall is frictionless then we can assume that at failure the principal effective stress
directions are vertical and horizontal and there are two ways in which the soil can fail.

Active State Passive State

In the first, the major principal stress is
vertical, and the minor principal stress is
horizontal. This could correspond to the
withdrawal of lateral support at constant
vertical stress until failure is reached. In
active conditions, the (minimum) horizontal
effective stress is given by:
In the second mode of failure, the major principal
stress is horizontal, and the minor principal stress is
vertical. This could correspond to an increase in
lateral stress and a constant or possibly reducing
vertical stress, as might occur in the soil in front of an
embedded retaining wall. In passive conditions, the
(maximum) horizontal effective stress is given by:

σ′h (1 − sin ∅′ ) σ′h (1 + sin ∅′ )

= = 𝐾𝐾𝑎𝑎 = = 𝐾𝐾𝑝𝑝
𝜎𝜎𝑣𝑣′ (1 + sin ∅′ ) 𝜎𝜎𝑣𝑣′ (1 − sin ∅′ )

7
 The friction angle used could be taken as either the peak or the critical (a.k.a. ultimate or
residual) value. For a given soil the critical value is a constant, whereas the peak value depends on
the potential for dilation, in turn dependent on soil density and the average effective stress at failure.
The peak value can only be maintained while the soil dilates, however as the soil continues to deform
the peak value drops as the soil strain softens. Theorems of plasticity require material to exhibit a
plastic plateau, which is essentially the case when the critical value is reached. It is therefore
reasonable to use the critical value when analysing embedded walls at failure.

It is important to remember that the failure of the soil depends on the effective stress state. The
earth pressure coefficient Ka and Kp are therefore applied to the vertical effective stress. The
horizontal effective stress at a given depth, z is equal to Ka or Kp times the vertical effective stress. In
the absence of wall friction, the vertical effective stress is equal to γz – u, where u is the local pore
water pressure.

The wall, however, experiences the effect of both the pore pressures and the effective stresses,
and is unable to distinguish between them. The equilibrium of the wall is therefore maintained by the
total horizontal stress distributions – i.e. effective stresses and the pore water pressures acting
together.

Design Considerations

Embedment

Unpropped embedded walls rely entirely for their stability on an adequate depth of embedment:
they are not supported in any other way. They will tend to fail by rotation about a pivot point below the
toe. The idealised stress distribution at failure is shown in Figure 3 below. These conditions are
sometimes known as fixed earth support, because the depth of embedment must be large enough to
prevent free rotational movement about the toe. The wall acts as an unpropped cantilever, built into
the ground.

Given the retained height and the soil strength there are two unknowns which must be
calculated in an investigation of conditions at collapse. These are the depth of embedment, required
just to prevent collapse, and the depth of the pivot point (at which moments are taken about – i.e. the
point the wall rotates about). The equations of horizontal and moment equilibrium can be used to find
these two unknowns, so the system is statically determinate.

This results in two simultaneous equations that are quadratic in the unknowns and so an
iterative solution technique is required. This is inconvenient to apply by hand and is best suited to
computerised goal seeking solutions.

8
 Figure 3: Idealised stress distribution for an unpropped embedded cantilever wall at failure

Prop or anchor forces

If the possibility of a structural failure of the wall or the prop is neglected, an embedded wall
propped at the crest (or anchored just below the crest) will tend to fail by rigid body rotation about the
position of the prop. (In reality the possibility of a structural failure of the prop or the wall itself would
need to be taken into account, as would the bending of the wall, which here is neglected due to the
assumption of rigid body failure). The idealised effective stress distribution at failure is reproduced in
Figure 4.

The conditions giving rise to the effective stress distributions shown in Figure 4 are sometimes
known as free earth support, because the toe of the wall is relatively free to move laterally. In other
words, no fixity is developed at the toe. In this case, the two unknowns are the prop force and the
depth of embedment required just to prevent failure. The depth of embedment can be calculated by
taking moments about the prop, and then the prop force follows from the condition of horizontal force
equilibrium.

Figure 4: Idealised stress distributions at failure for a stiff wall propped rigidly at the crest

Bearing failure

Retaining walls must be designed with due consideration to bearing capacity at the bottom of
the structure. In essence, the base pressure applied by the wall must not exceed the ultimate bearing
capacity of the soil. For procedures used to check for adequate bearing capacity, refer to your notes
"Foundation Design: Bearing Capacity." In practice it is also necessary to embed walls to some depth

9
below formation level in order to prevent the soil below the toe of the wall undergoing bearing failure,
due to the weight of the soil outside the excavation. This results in large settlements behind the wall,
while the soil in front of the wall will be pushed upward in the excavation. The depth of embedment
must therefore be greater than the depth required to prevent bearing failure.

Figure 5: Stiff, frictionless embedded cantilever retaining wall propped at the crest

Long term effects – Total stress analysis

An important point concerning an analysis based on undrained shear strengths is that it is only
valid for the period during which the water content of the clay does not change. In problems involving
the excavation of clay, for example an in situ retaining wall, the long term effective stresses are
generally smaller than those in the undisturbed ground. This means that the clay will tend to take in
water and soften, leading to a reduction in undrained shear strength with time. As the undrained shear
strength decreases the active lateral stresses will increase. The rate of softening in practice will
depend on:

• The bulk permeability of the soil (which may be increased by the effect of fissures which
open up on unloading)
• The soil stiffness (which might be quite high following a change in the direction of the
stress path)
• The availability of water (potential sources include rainfall, surface run-off, natural water
courses, leaking pipes, gravel aquifers (gravel layers) and sand lenses, waste water
from hosing down construction plant, etc.)

In assessing the applicability of the undrained shear strength model, all these factors must be
considered in relation to the time scale over which the excavation is expected to remain open.

Long term effects – Effective stress analysis

The long term stability of permanent structures must be investigated by means of an effective
stress analysis. It is necessary therefore at the design stage to estimate the long term pore water
pressure distribution around a retaining wall. In the case of an embedded wall in a uniform soil, where
the wall is effectively impermeable and weepholes or other drainage measures behind the wall are
not installed, the long term equilibrium pore water pressure distribution will correspond to steady state
seepage from a high water table behind the wall to a reduced ground level in front. The pore water
pressure distribution may be obtained by sketching a flownet, but an approximation in which the fall

10
in hydraulic head is assumed to be distributed linearly around the wall is often close enough for design
purposes.

Box 5: Approximate steady state seepage pore water pressure distribution by linear approximation
Steps:

1. Chose a datum from which to measure the elevation

head.
2. The total head at A and C are plotted against distance
from A around the wall to C and distributed linearly.
3. The pore pressure at the base of the wall (or any
other position along the wall) is given by:
Pore Pressure= γw (Total head - Elevation head)

Soil/Wall friction and adhesion

The earth pressure coefficients derived for frictionless walls will lead to uneconomical designs
when the wall is rough. This can be taken into account by considering the effect on the average
effective or total stress of the rotation in the direction of the principal stresses that must occur between:
i) a zone of soil near the free surface (within which the principal stresses are horizontal and vertical,
as assumed earlier), and ii) the soil adjacent to the wall (in which the presence of shear stresses on
the vertical surface of the wall means that this cannot be a principal direction). Details of these
calculations can be found in Powrie (1997) however Table 1 and Table 2 are provided from which
values for effective stress parameters can read off. For total stress analyses refer to Box 6.

The key decisions facing the designer concern the magnitude of the soil/wall adhesion or friction
(usually expressed as a proportion of the design strength, su or φ’), and the direction in which the
adhesion or friction acts. Literature is littered with advice on what magnitude should be used, but
generally consensus is that limiting values of soil/wall friction δ are, δ = φ’crit for rough (cast in situ)
concrete, and δ = ⅔ φ’crit for smooth (precast) concrete or sheet piling supporting sand or gravel. With
regards to values of soil/wall adhesion it is recommended a maximum value of half the design
undrained shear strength be used.

If the soil behind a retaining wall settles relative to the wall and the soil in front heaves, then the
directions of soil/wall adhesion or friction act so as to reduce the active earth pressures and increase
the passive pressures. Both of these tend to enhance the stability of the wall. However, there are
circumstances in which these conventional directions of relative soil/wall movement (and hence
adhesion and friction) could be reversed. For example, if the wall carries a vertical load – such as a
perimeter wall for a building basement – then the wall may settle relative to the retained soil and the
direction of active side adhesion or friction will be reversed. Dewatering inside the excavation could
cause relative settlement of the soil in front of the wall, but this is rather less likely.

11
Table 1: Active earth coefficients, Ka
Ka with tanδ =
φ' Ka with δ = 0 Ka with δ = ½ φ' Ka with δ = ⅔ φ’ Ka with δ = φ’
0.75 x tan φ’
12 0.656 0.613 0.604 0.600 0.593
13 0.633 0.589 0.580 0.575 0.568
14 0.610 0.566 0.556 0.552 0.545
15 0.589 0.544 0.534 0.529 0.522
16 0.568 0.522 0.512 0.508 0.500
17 0.548 0.502 0.492 0.487 0.479
18 0.528 0.482 0.472 0.467 0.459
19 0.509 0.463 0.453 0.448 0.440
20 0.490 0.445 0.434 0.429 0.422
21 0.472 0.427 0.417 0.412 0.404
22 0.455 0.410 0.400 0.395 0.387
23 0.438 0.394 0.383 0.378 0.371
24 0.422 0.378 0.368 0.363 0.355
25 0.406 0.363 0.353 0.348 0.340
26 0.390 0.348 0.338 0.333 0.326
27 0.376 0.334 0.324 0.319 0.312
28 0.361 0.320 0.311 0.305 0.298
29 0.347 0.307 0.298 0.292 0.286
30 0.333 0.294 0.285 0.280 0.273
31 0.320 0.282 0.273 0.268 0.261
32 0.307 0.270 0.262 0.256 0.250
33 0.295 0.259 0.250 0.245 0.239
34 0.283 0.248 0.240 0.234 0.228
35 0.271 0.237 0.229 0.224 0.218
36 0.260 0.227 0.219 0.214 0.208
37 0.249 0.217 0.209 0.204 0.198
38 0.238 0.208 0.200 0.195 0.189
39 0.228 0.198 0.191 0.186 0.180
40 0.217 0.189 0.182 0.177 0.172

12
Table 2: Passive earth coefficients, Kp
Kp with tanδ =
φ' Kp with δ = 0 Kp with δ = ½ φ' Kp with δ = ⅔ φ’ Kp with δ = φ’
0.75 x tan φ’
12 1.525 1.686 1.723 1.739 1.763
13 1.580 1.766 1.808 1.826 1.855
14 1.638 1.850 1.899 1.920 1.953
15 1.698 1.939 1.995 2.020 2.057
16 1.761 2.034 2.098 2.126 2.168
17 1.826 2.135 2.207 2.240 2.287
18 1.894 2.242 2.324 2.361 2.415
19 1.965 2.356 2.449 2.492 2.552
20 2.040 2.477 2.582 2.631 2.699
21 2.117 2.607 2.726 2.782 2.858
22 2.198 2.745 2.879 2.943 3.028
23 2.283 2.893 3.044 3.117 3.212
24 2.371 3.051 3.222 3.305 3.411
25 2.464 3.222 3.413 3.509 3.627
26 2.561 3.404 3.619 3.729 3.861
27 2.663 3.601 3.843 3.969 4.116
28 2.770 3.812 4.085 4.229 4.393
29 2.882 4.041 4.347 4.512 4.695
30 3.000 4.288 4.633 4.822 5.026
31 3.124 4.555 4.944 5.162 5.389
32 3.255 4.845 5.283 5.534 5.788
33 3.392 5.160 5.655 5.944 6.227
34 3.537 5.504 6.062 6.395 6.712
35 3.690 5.879 6.510 6.895 7.250
36 3.852 6.289 7.004 7.449 7.847
37 4.023 6.738 7.549 8.066 8.512
38 4.204 7.232 8.153 8.754 9.255
39 4.395 7.777 8.825 9.525 10.088
40 4.599 8.378 9.573 10.390 11.026

Box 6: Total stress analysis taking soil/wall adhesion into account

For favourable wall/soil adhesion the following equations can be derived:

σh, active =σv - su (1+∆+ cos ∆) σh, passive =σv + su (1+∆+ cos ∆)

where sin ∆ =τw /τu, with τw =0.5τu the term in the brackets becomes numerically equal to 2.39,
compared to 2 in the case of the adhesion-less wall in Box 3. The same comments about the depth
of dry and flooded tension cracks apply. However, there is a minor intellectual complication
concerning the depth below the bottom of the tension crack at which soil/wall adhesion starts to be
mobilised.

13
Working Load/Permissible Stress Design

The design of retaining structures has traditionally been based on the specification of a factor
of safety in terms of moments, i.e. the ratio of the resisting (or restoring) moment to the disturbing (or
overturning) moment. This is known as the lumped factor of safety and is given a high enough value
to allow for all the uncertainties in the analytical method and in the values of the soil parameters. It
must be recognised that relatively large deformations are required for the mobilization of available
passive resistances and that a structure could be deemed to have failed due to excessive deformation
before reaching a condition of collapse (I.e. the working load/permissible stress design does not
necessarily take into account serviceability limits). In using this method the passive resistance is
divided by an appropriate factor of safety. In general, the higher the factor of safety, the lower will be
the deformation required to mobilise the proportion of passive resistance necessary for stability.

Limit State Design

The limit state approach is based on the application of partial factors to actions and
characteristic soil parameters. Characteristic values are conservative values of φ’, su and γ that take
into account the variation in soil conditions, the quality of the investigation and the quality of
construction likely to be achieved. Partial factors that should be used are those that correspond to the
geotechnical failure mode (see Foundation Design: Bearing Capacity notes). When considering the
design of retaining walls the design values of active thrust and passive resistance are calculated using
the design values (i.e. factored characteristic values) of the relevant shear strength parameters. The
active thrust and passive resistances due to the self-weight of the soil are considered permanent
actions (unfavourable and favourable respectively). The active thrust due to surcharge loading is
normally a variable unfavourable action (i.e. an imposed load).

Other considerations

When designing a wall to limit the potential settlements the following three modifications to the
collapse calculation should be made:

1. A reduced soil strength is used to calculate the active and passive pressures (this can
be done using either the Working Load/Permissible Stress philosophy or the limit state
philosophy)
2. The retained height is increased by at least 0.5m, or 10% of the retained height for
embedded cantilever walls, representing an unplanned excavation in front of the wall.
This profile is called the design ground.
3. An additional surcharge of 10kPa is assumed to act on the retained soil surface.

14
References and Further Reading

These notes have been developed from the various texts listed below, the author admits some
sections are copied verbatim and therefore does not pass this work of as his own.

Atkinson, J. (2007). The Mechanics of Soils and Foundations, 2nd Edition. Taylor & Francis,
Oxon.

Sharma, (2002). Lateral Earth Pressure – Fine-grained Soils and MSE Walls. <
http://teacher.buet.ac.bd/sid/download/CE341/Lecture16_4on1.pdf>

Burland, J., Chapman, T., Skinner, H., and Brown, M. (Editors) (2012). ICE Manual of
Geotechnical Engineering: Volume 1 Geotechnical Engineering Principles, Problematic Soils and Site
Investigation. ICE Publishing, London.

Craig, R.F. (2004). Craig’s Soil Mechanics, 7th Edition, Taylor & Francis, Oxon.

Powrie, W. (1997). Soil Mechanics: Concepts and Applications, 1st Edition. E & FN Spon,
London.

15
Lateral Earth Pressure: Questions

Class Example 1

A cross section through an excavation braced by a retaining wall is given below. What is the minimum
thrust (in kN/m) due to the retained ground that the retaining wall must, in the short term, be able to
resist? (Assume that the retaining wall is frictionless, and that the pore water pressures in the sandy
soil are hydrostatic. Take the unit weight of water as 10 kN/m3).

a) Is your answer reliable for use in design?

a) If the middle prop were accidently removed, what would be the loads in the remaining props?
b) Why would it in reality be necessary to embed the wall to some depth below formation level?

Class Example 2

By dividing the soil into active and passive zones, calculate the depth of embedment d required just
to maintain the stability of the propped cantilever retaining wall. Calculate the corresponding prop load
F.

16
Class Example 3

Assuming that tension cracks remain dry, calculate the depth of embedment required just to prevent
rotational failure of a frictionless, rigid, embedded cantilever wall, propped at the top, in a uniform clay
3
soil of undrained shear strength su = 40 kPa and unit weight γ = 20 kN/m .The retained height is 5m,
and there is a uniform surcharge of q = 40kPa on the retained side. Calculate the corresponding prop
load F.

Class Example 4

a) The figure below shows a cross section through a rough embedded retaining wall, propped at
the crest. Stating any assumptions you make, estimate the long term pore water distribution
around the wall.
b) Assuming that the critical state angle of the soil friction of 35° is fully mobilised in the retained
soil, calculate the earth pressure coefficient (based on effective stresses) in the soil in front of
the wall required for moment equilibrium about the prop. Using Table 2, estimate the
corresponding mobilised friction angle in the soil in front of the wall.
c) Is the wall safe? Explain briefly your reasoning.

17
Tutorial Question 1

The figure below shows a cross-section through a long excavation whose sides are supported by
propped cantilever retaining walls.

a) Calculate the depth of embedment needed just to prevent undrained failure by rotation about
the prop if the groundwater level behind the wall is below formation level.
b) Describe how and why the depth of embedment would change if the groundwater level behind
the wall was at original ground level.

Neglect the effects of friction/adhesion at the soil/wall interface and take the unit weight of water as
10kN/m3.

WT

WT

18
Tutorial Question 2

Details of an anchored sheet pile are given in the figure below. The design ground and water levels
are shown. Above the water table the unit weight of the soil is 17kN/m3 and below the water table the
saturated unit weight is 20kN/m3. Characteristic soil parameters are φ’ = 36° and δ is taken as ½ φ’.
Determine:

a) The factor of safety with respect to gross passive resistance

b) Whether the design meets the limit state for the geotechnical failure mode
c) The anchor load

Guides: WL/PSD: The factor of safety is applied to the total passive resistances only
Limit State: The friction angle is appropriately factored
The surcharge force is considered a variable action (Imposed Load)
All other forces are considered as permanent actions

19