You are on page 1of 89

H23GGE

Reinforced soil: walls and slopes



Dr E A Ellis
School of Civil Engineering, University of Nottingham
Updated 23/08/07




i
1. Introduction to reinforced soil 1
1.1. Why reinforce soil? 1
1.2. Materials and methods 3
1.2.1. Soil 3
1.2.2. Reinforcement - introduction 5
1.2.3. Reinforcement steel strips 6
1.2.4. Reinforcement geogrids 7
1.2.5. Facing systems for walls hard 12
1.2.6. Facing systems for walls soft 16
1.2.7. Other applications 17
1.2.8. Methods of construction 18
2. External stability 20
2.1. Sliding 21
2.2. Overturning 24
2.3. Slip failure 29
3. Internal stability - introduction 30
3.1. Tension in reinforcement 32

ii
3.2. Tensile strength 38
3.3. Pull out 40
4. Internal stability - refinements 45
4.1. Surcharge at soil surface 45
4.2. Distribution of vertical stress 47
4.2.1. Schlosser method 47
4.2.2. Bolton method 52
4.3. Compaction stresses on wall 54
4.4. Determination of maximum tension 57
4.4.1. Envelope of T
r,max
57
4.4.2. Determination of T
max
59
5. Reinforced slopes 61
5.1. Design of reinforced slopes 63
5.1.1. General principles 63
5.1.2. The 2 part wedge 65
6. Design example for wall 73

iii
6.1. Preliminary design 76
6.2. Detailed design 78
6.3. Full results 81
6.4. Notes on equations in exam 83


iv
Main notation

b
r
= width of discrete (non-continuous) reinforcing element
e = eccentricity (from centre of reinforced block) of force representing vertical
stress at a given depth in the soil
h
w
= height of reinforced soil wall
h
s
= height of reinforced soil slope
l
r
= length of reinforcement (in direction normal to wall)
l
r,a
= length of reinforcement in active zone
l
r,po
= length of reinforcement providing resistance to pull out (in the resistant
zone, and taking account of e as necessary for rectangular stress distribution)
q = bearing pressure (vertical stress)
s
h
= separation of reinforcing elements in horizontal direction
s
v
= separation of reinforcing elements in vertical direction
t
r
= thickness of reinforcing element

v

A
f
= area of face corresponding to a reinforcement element
F
s
= generic FoS
T
r
= tension in individual reinforcement element
T
R
= total tension in all reinforcement elements (2 part wedge). See Figure 19 for
definition of T
R,0
and T
R,max.

T
s
= shearing force resisting sliding of reinforced soil block (external equilibrium)

= slope angle (2 part wedge)

r
= interface friction angle for reinforcement and soil

s
= interface friction angle for sliding of reinforced soil block (external
equilibrium)



1
1. Introduction to reinforced soil
1.1. Why reinforce soil?
As you will be aware from the 2
nd
year course soil is actually widely used as a
construction material, primarily in embankments.
You should also be aware that we generally consider its shear strength, which is
frictional.
Consequently it has no tensile strength, and will only stand in slopes at or
shallower than its internal angle of friction (). [An exception to this is the short-
term undrained behaviour of clays.]
If we wish to retain soil with a vertical face a retaining wall is generally required
such structures are generally quite substantial, and can be expensive and
consume significant amounts of primary resources.
Whilst significantly stronger than soil, concrete also has little or no tensile
strength. However, its performance can be transformed by the inclusion of steel
(which has significant tensile strength).

2
In the 1960s the concept of (tensile) reinforcement in soil emerged in Civil
Engineering, and it is now increasingly widespread, with reinforced soil walls
commonly up to 10 m high, and sometimes higher.


3
1.2. Materials and methods
1.2.1. Soil
In principle any kind of soil can be reinforced.
However, as you may also remember from the 2
nd
year, granular soils (ie. sand
or gravel) tend to be favoured in the construction of retaining structures. This is
largely because they are free-draining, which means that the behaviour is
relatively predictable, and there is less risk of water pressure acting on the
retaining structure.
Granular soils also tend to predominate in construction of reinforced soil walls.
Clay soils are sometimes used, and this can bring significant economic and
environmental benefits.
However, here we will only consider the use of granular fill, which we will assume
to be adequately drained so that:
- u = 0, and hence
-
v
=
v


4
We also will consider the strength of the material to be entirely frictional as
characterised by a friction angle . Use of a cohesion intercept (c) will not be
considered.
We will generally be dealing with the active earth pressure in the soil:

h
= K
a

v

where
) 2 / 45 ( tan
sin 1
sin 1
2

' =
' +
'
=
a
K
[we will not consider the effect of any friction on the back of the wall or back of
the reinforced soil block]

5
1.2.2. Reinforcement - introduction
Fundamentally the reinforcement must have 2 important qualities:
- Good tensile strength
- A good interface with the soil, so that stresses can be transmitted between
the 2 materials
To continue the analogy with reinforced concrete, the steel which is normally
used in this application
- Has exceptionally high tensile strength
- Is deformed with ribs on the surface of the bars to ensure good bond with
the concrete

6
1.2.3. Reinforcement steel strips
Strips of mild steel were first used to reinforce soil in the 1960s. They are long
and thin, with ribs to increase friction with the soil.
A typical section for the strip is shown below.







The corresponding tensile capacity is:
T
r
=
r
b
r
t
r
= 2501005 = 100 kN
It is worth reflecting that this is approximately equal to the force due to the vertical
stress on a 1 m square at 5 m depth in soil.

7
1.2.4. Reinforcement geogrids
Polymer based geogrids have become increasingly widespread in the last 20
years. Compared to steel there are advantages:
- Used in sheets, so inclusion over significantly larger area in soil
- Nature of grids designed to promote interaction with soil
- Corrosion not an issue
and disadvantages:
- Less stiff/ strong
- Behaviour is time-dependent
- Material is more susceptible to other damage (eg. physical, chemical,
biological, UV)


8


9



10


Uniaxial grid strength mainly in one direction


Biaxial grid equal strength in either direction

11
The tensile strength is measured per metre width of the grid, and may typically be
in the range 50 to 200 kN/ m when measured for short-term load. Very strong
products may approach 1000 kN/m.
However, the strength which can be relied upon for design in the long-term is
typically only about half this. This is mainly because the material is very prone to
creep (ie. to strain increasingly at constant load) unless the load is limited to well
below the short-term strength.
Other factors which affect the safe design strength include:
- How much strain can be tolerated (even excluding significant creep)
- The temperature in the soil(!)
- How much damage may be caused when the soil is placed on the grid during
construction (see later)


12
1.2.5. Facing systems for walls hard
Dunkirk flyover


13



14
Lightweight precast concrete facings are widely used, particularly with steel
reinforcement.
Their purpose is purely to maintain local stability of the vertical face, a concrete
facing should not be confused with a concrete retaining wall, which would need to
be considerably stronger.
The connection detail to the reinforcement is extremely important to ensure that
load is transmitted.


15
Masonry facings can also be used, especially for smaller walls with geogrid
reinforcement.



16
1.2.6. Facing systems for walls soft
Where a geogrid is used a wraparound can be used to form a vertical or near-
vertical face. Vegetation may be able to grow on the face, improving aesthetics.

Figure 1 wrap around facing

17
1.2.7. Other applications
Here we will focus on the use of reinforced soil for vertical walls, and steep
slopes.
However, geogrid reinforcement has many other applications in geotechnical
engineering including:
- Reinforcement of roads, railways or working platforms on construction sites
- Embankments on soft ground
- Facings for other steep slope systems (eg. soil nails)


18
1.2.8. Methods of construction
In reinforced concrete the reinforcement cage is constructed and then the
concrete is poured around it. In reinforced soil the reinforcement must be added
as necessary as the embankment is constructed.
Generally fill is in any case placed in layers about 300 mm thick so that it can be
compacted. Generally the reinforcement is placed between such layers as
construction proceeds


19
The end product is thus generally a rectangular block of reinforced soil with
horizontal reinforcement at regular intervals, and some form of facing system.


20
2. External stability
The external stability of a reinforced soil wall refers to the external forces which
act on it. At this stage we treat the block of reinforced soil like a (potentially very
large) gravity retaining structure so we will check stability against sliding and
overturning.
To keep things simple and focus on the fundamentals of behaviour we will initially
assume that:
- The soil in the reinforced block and immediately behind it is uniform, with unit
weight and active earth pressure coefficient K
a
. The soil is free-draining so
that the pore pressure u = 0.
- Loading on the back of the block is purely horizontal (ie. there is no shear
stress on the vertical face)
- There is no surcharge loading on the surface of the soil
As is usual with any retaining structure we will assume that the situation is 2-
dimensional (ie. as if it was infinitely long into the page). We will assume that
the reinforced soil block is rectangular in side elevation.

21
2.1. Sliding
We will assume that the length of the reinforcement is l
r
and the height of the
vertical wall is h
w
. The angle of friction (= tan
-1
) between the base of the wall
and the underlying ground is
s
.










Figure 2 Forces for sliding equilibrium

22
Hence, the reaction from the underlying ground is equal to the weight of the
block (vertical equilibrium):
N = W = l
r
h
w
(kN/m width)
The maximum resistance to sliding is
T
s,max
= N tan
s


The horizontal load on the block is (from the triangular distribution of stress):
H = (1/2) K
a
h
w
2


Thus F
s
(the factor of safety against sliding) is:
w
r
a
s
w a
s w r
s
s
h
l
K
h K
h l
H
T
F
o

o tan
2
tan 2
2
max ,
= = =
Equation 1

23
As we would expect, F
s
increases as
s
increases, and reduces as K
a
increases.
It increases as l
r
increases relative to h
w
, ie. the wider the reinforced block.
However, the cost also increases with l
r
.
If the uniform soil has a friction angle = 30
o
then we find that K
a
= 0.33.

s
, so we could assume (say) 27
o
.
If F
s
is required to be 2.0 (although it could potentially be lower) then we find that
(l
r
/ h
w
) 0.67, ie. the length of the reinforcement is 2/3 the height of the wall. In
fact we will find that it is normally necessary for it to be longer than this.


24
2.2. Overturning
We will assume that the reaction from the underlying ground has an eccentricity e
relative to the centreline. As in the 2
nd
year course this can be taken to
correspond to a rectangular distribution of stress, acting over an effective width
of (l
r
2e)



25













Figure 3 Forces for overturning equilibrium, eccentricity of vertical
reaction and rectangular distribution of vertical stress
e
N
H
W
2e

26
Thus taking moments about any point on the base of the wall:
W e = H h
H
Hence
2
2
6
1
3
1
2
1
|
|
.
|

\
|
= = =
r
w
a
r
w
w
r
w a
r
H
r
l
h
K
l
h
h l
h K
l
h
W
H
l
e




Hopefully you recall from the 2
nd
year that it is normal to ensure that the reaction
is within the middle third of the base, ie. (e/l
r
) (1/6). Notice that e increases as
the square of (h
w
/l
r
).
If (e/l
r
) = (1/6) then
a
w
r
K
h
l
=
Equation 2
So using K
a
= 0.33 gives (l
r
/ h
w
) 0.6 similar to sliding.

27
Adding the effect of (for instance) a surcharge at the surface of the fill
complicates the formulae somewhat, and tends to increase (l
r
/ h
w
) since the point
of action for the surcharge is at h
w
/2.
Typically the actual ratio for a real design might be 0.7 or 0.8 so our simple
estimates have given good results.
The bearing pressure on the underlying ground for a rectangular distribution will
be:
e l
h l
e l
W
q
r
w r
r
b
2 2
=

=


hence
) / 2 ( 1
1
r w
b
l e h
q


Equation 3

28
ie. the bearing pressure is increased relative to the nominal vertical overburden
stress (eg. if e/l
r
= 1/6 the increase is a factor of 1.5).
We will generally consider here that the underlying ground is sufficiently
competent to carry this load. However in the design of real structures this
aspect may well require careful consideration.

29
2.3. Slip failure
We will also need to consider this possibility (eg. using the method of slices),
particularly if the retained soil slopes upwards behind the wall.

Figure 4 Slip failure

30
3. Internal stability - introduction
We must now consider whether the block of reinforced soil itself will be stable.
We will again begin with simple assumptions to demonstrate the basic principles.
We will assume that an active wedge forms behind the vertical face within the
reinforced soil.
Again assuming that = 30
o
, we find that the angle of the wedge to the
horizontal is (45+30/2) = 60
o
. Hence the width of the block at the ground surface
is 0.58h
w
, so based on the geometry dictated by external equilibrium the active
wedge will be entirely within the reinforced soil.
Referring to the 2
nd
year course this material is failing behind the wall and
loading it. However, outside this zone we can rely upon it to anchor the
reinforcement against pullout.


31












Figure 5 Active wedge causes active earth pressure because of
tendency to slide down and outwards. This is resisted by shearing in soil
and tension in reinforcing elements which tie it back
T
r,max

Tendency
to slide
t
soil


32
3.1. Tension in reinforcement
We must then calculate the maximum tension in the reinforcement T
r,max
, which
we will initially assume to occur at the back of the active wedge, and to be
generated by a combination of:
- normal stress on the face which is transferred to the reinforcing element by
the connection to the face
- shearing stress on the reinforcing element itself within the active wedge
(both are generated by and within the active wedge)
T
r,max
= K
a

v
A
f
Equation 4a
where
A
f
= the area of the vertical face corresponding to an individual reinforcing
element

v
= vertical effective stress at the depth of the reinforcing element (=
v
since we
will assume that u = 0).

33













Figure 6 Streses acting on facing and reinforcing element within active
wedge, resisted by T
r,max
at back of wedge
T
r,max

Area of face
corresponding to
reinforcement
element, A
f

Active zone

34













Figure 7 Distribution of tension in reinforcing element
T
r,max

Resistant
zone
Active zone
?
?
Tension
transmitted by
connection to
face

35

For now we will assume that

v
= z
where z is the vertical distance from the surface of the retained soil.
T
r,max
= K
a
z A
f

Equation 4b
Hence the tension in reinforcing elements tends to increase with depth if A
f
does
not vary.



36
If
s
h
= separation of reinforcing elements in horizontal direction along the structure
(parallel to the face)
s
v
= separation of reinforcing elements in vertical direction along the structure

Then
A
f
= s
h
s
v
for discrete (non-continuous) reinforcing elements


37
Where geogrid is used, which may well be continuous in the plane strain
direction) then effectively
A
f
= s
v
and has units (m
2
/m width). Correspondingly T
r,max
will have units kN/m as
previously discussed for geogrid.


38
3.2. Tensile strength
Now that we have established T
r,max
the first concern is that the reinforcing
element should be able to carry this load.
The capacity of the reinforcement itself, T
r,des
is generally certified by the
manufacturer, so this is generally relatively straightforward. We have mentioned
previously that various factors need to be considered when selecting a design
strength for a geogrid. A factor of safety will also generally be incorporated
irrespective of the reinforcing material.
We then require that
T
r,max
T
r,des

Equation 5

39
Referring to Equation 4 we have noted that T
r,max
will increase with depth.
Generally a number of approaches may be adopted (and combined) to cater for
this:
- Use elements with constant strength and spacing elements near top will
have excess capacity, but this may be tolerated for standardisation of
product and method of construction.
- Use stronger elements at greater depth
- Reduce spacing of elements with depth


40
3.3. Pull out
Every force will have an equal and opposite reaction, and our second concern will
be: can the reinforcing element resist this load somewhere?
We have assumed that T
r,max
is generated by a combination of load at the face
and within the active wedge. Hence it must be resisted outside this region. We
will refer to the length of the reinforcement outside the active wedge as l
r,po


41













Figure 8 T
r,max
is resisted by pull out (governed by shear at the soil-
reinforcement interface) in the resistant zone over length l
r,po
T
r,max

Resistant
zone

42
po denotes pull-out, since we are concerned that the grid should not pull out of
the soil outside the active wedge, which it is anchored in.
We will assume that
r
(= tan
-1
) is the friction angle at the interface of the
reinforcing element and the soil. Hence if the interface mobilises a certain
fraction of the soil strength (say ) then
tan
r
= tan
The area of this interface is
A
r,po
= 2 b
r
l
r,po

where the factor 2 arises since both faces of the reinforcement are in contact with
the soil.
The normal stress acting on these faces is
v
. Hence
T
r,po
= 2 b
r
l
r,po

v
tan
r

Equation 6

43
Combining Equation 6 with Equation 4ba the FoS against pullout, F
s,po
is
a
r
f
po r r
r
po r
po s
K A
l b
T
T
F
o tan
2
,
max ,
,
,
= =
Equation 7
Using assumed values of = 30
o
so that K
a
= 0.33, and
r
= 27
o
, and assuming
that F
s,po
is required to be 1.5
b
r
l
r,po
0.5A
f



44
For discrete steel strips A
f
= s
h
s
v
, where s
h
is typically perhaps 10b
r
. Hence
typically
(l
r,po
/s
v
) 0.5(s
h
/b
r
) 5
So if s
v
is 1.0 m l
r,po
will need to be about 5 m

For continuous geogrid A
f
= s
v
where b
r
is effectively 1.0 m. Hence typically
(l
r,po
/s
v
) 0.5
and if s
v
is 1.0 m l
r,po
will need to be about 0.5 m

This demonstrates the advantage of larger area of interaction for the geogrid.

It can be seen that (without a surcharge) the probability of pull out is independent
of depth. Since l
r,po
increases with depth (for a rectangular block of
reinforcement), pull out is generally of most significant concern at limited depth.

45
4. Internal stability - refinements
We will now consider the effect of introducing new factors, and some more
complex assumptions of behaviour. It should be noted that there is no single
completely prescriptive method for design of reinforced soil retaining walls. BS
8006 contains much information on the range of approaches, but is not
particularly prescriptive. Hence the decision regarding the details of the method
to be used in a given circumstance may be subjective.

4.1. Surcharge at soil surface
If a surcharge q acts on the soil surface then:

v
= z + q
and the previously triangular distribution of horizontal stress is increased by an
amount K
a
q, which is constant with depth (a rectangular distribution).
The equations above are modified correspondingly.


46














Figure 9 Effect of surcharge on vertical and horizontal stress

47
4.2. Distribution of vertical stress
4.2.1. Schlosser method
We saw in Section 2.2 (Equation 3) that the reaction from the ground may be
idealised as a rectangular distribution of stress which does not act over the full
width of the reinforced soil block, and consequently exceeds the nominal vertical
stress
v
= z.
The previous results can be stated for a general depth z in the soil:
) / 2 ( 1
1
) / 2 ( 1
1
r
v
r w
b
l e z l e h
q

=
'


where
2 2
6
1
6
1
|
|
.
|

\
|
=
|
|
.
|

\
|
=
r
a
r r
w
a
r
l
z
K
l
e
l
h
K
l
e


48
If we consider moment equilibrium for a general depth z in the reinforced soil
block (Figure 10) rather than the full depth (h
w
) we derive the equations for the
Schlosser method, (3) in the data sheet which used to accompany the exam
paper for this module.
Assuming K
a
= 0.33 we can plot the expressions showing variation with z as
below (Figure 11).
Essentially the effects increase with z
2
. Hence over the top half of the wall they
have very limited effect.

49


Figure 10 Rectangular distribution of stress at depth z

50


Figure 11 Effect of Schlosser method for K
a
=0.33 (q =0)

51
How would this approach affect a design?

T
r,max
:
Equation 4b (T
r,max
= K
a

v
A
f
) predicts increased loading on the reinforcement
since the increased value of
v
acts immediately behind the face and hence
would be used in place of Equation 4 (where it is assumed that
v
= z).

T
r,po
:
It is effectively assumed that
v
= 0 over a width 2e at the back of the reinforced
soil block. Hence this length would be ignored when calculating l
r,po
and T
r,po
,
since if
v
is zero then the interface shear stress will be zero (Equation 6).
However, note that this effect is most significant near the bottom of the reinforced
fill, where pull out is any case unlikely to be an issue.

52
4.2.2. Bolton method
Clearly it is not completely realistic to assume that the vertical stress drops to
zero over some (relatively small) portion of the width at the back of the reinforced
soil block. However, the method is pragmatic.
You may recall from the 2
nd
year course that we may alternatively assume a
trapezoidal distribution of vertical stress beneath a gravity retaining wall. This
can be used to derive the Bolton method in the data sheet.
It was shown in the 2
nd
year course that for a gravity retaining wall of width B the
trapezoidal bearing pressure was given by:
B
e
q
q
av
6
=
A

Here B = l
r
. Hence substituting (e/l
r
) from above and using
v
in place of q,
where q
av
is the nominal value z:
2
|
|
.
|

\
|
=
' A
r
a
v
l
z
K
z
o


53
Note that again the effect will be limited for small z.
In fact the actual distribution of stress beneath the wall will perhaps be something
like this.

Figure 12 Assumptions of bearing stress and possible actual distribution
(BS8006)

54
4.3. Compaction stresses on wall



55
Compaction of the fill is generally achieved by applying a significant surcharge
(often in combination with vibration) to the soil surface. When this vertical stress
is removed the soil can relax (strain) vertically, but it is restrained horizontally,
and thus the corresponding horizontal stresses tend to be locked in.
At significant depths in the soil the vertical stress continues to increase as more
fill is added. A point is then reached when even the active horizontal stress
associated with the large vertical stress exceeds the horizontal stress which was
initially locked in. Thus the horizontal stress becomes the active stress.
Hence a distribution of earth pressure such as shown below may be used instead
of K
a
throughout - Equation 4ba is modified accordingly. In principle such a
distribution should be specifically derived taking account of:
- the compactive effort
- the stiffness of the wall
Generally speaking reinforced soil walls do not allow the use of large compaction
plant near the face of the wall (which is relatively insubstantial), and thus the
compaction induced pressures may not prove to be a significant issue.

56
0
1
2
3
4
5
6
7
8
9
10
0.0 0.1 0.2 0.3 0.4 0.5 0.6

Figure 13 Potential distribution of K for = 30
o
and h
w
=10 m (based on
Fig 2 in data sheet)

57
4.4. Determination of maximum tension
There are also refinements which may be applied to this part of the design.
4.4.1. Envelope of T
r,max

Based on physical and numerical studies of reinforced soil walls, the location of
T
r,max
in the reinforcement may be empirically modified, for instance as shown
below and in Fig 1 of the module data sheet.
Compared to the active wedge in Figure 6 it can be seen that the width of the
active zone is assumed to be narrower near the top of the wall , implying that
the effect of the wall on the reinforcement is more localised than implied by the
active wedge. Generally speaking approximately the top half of the mechanism
is affected.
This is most likely to impact (beneficially) on a design with steel reinforcement,
where l
r,po
may be relatively large, potentially requiring l
r
to be increased so that
pullout does not occur for the shallower elements if the full active wedge is
assumed.


58

0
2
4
6
8
10
0 2 4 6 8

Figure 14 Potential active zone for = 30
o
and h
w
=10 m (based on Fig 1
in data sheet)

59
4.4.2. Determination of T
max

Finally, T
r,max
may be estimated by making allowance for additional specific load
on the reinforcement in the active zone. This can be compared to Error!
Reference source not found. where the earth pressure and area of face is
assumed to give the full load on the reinforcement irrespective of whether this is
generated at the face or directly on the reinforcement.
This approach adds further complexity, and is fundamentally at odds with the
concept that the active earth pressure results not just from consideration of the
stress state at the face of a retaining structure, but also (with numerically identical
results) from consideration of the active wedge. Thus there is an argument that
the combination of load on the face and within the wedge should not exceed the
active earth pressure.
You should be aware that some design methods may require you to account for
this (in a similar way to calculation of the pull-out resistance). However, we will
not consider it further here, and Eq. 5 of previous data sheets would not consider
the L
A
term.

60













Figure 15 Alternative definition of maximum tension in an element of
reinforcement (see comments in text above)

61
5. Reinforced slopes
The reason for constructing a wall is to achieve a change in height with minimum
horizontal distance. Slopes steeper than 70
o
are generally considered as vertical
walls.
If more distance is available a sloping face may be feasible. As the slope
becomes less steep the amount of reinforcement can be reduced. Ultimately the
slope can be shallow enough that no reinforcement is need at all.
It may also be necessary to steepen an existing slope (for instance for highway
widening on an embankment or in a cutting) by adding reinforcement to a
previously unreinforced slope.
As slopes become less steep than this there is increasing possibility of growing
vegetation on the face, particularly if it is South facing. This is normally done in
conjunction with a soft facing, which may be purposely designed to retain top soil.
The desire to sustain vegetation the slope will be aesthetic, but also to prevent
erosion of soil from the face.

62



63
5.1. Design of reinforced slopes
5.1.1. General principles
The principles are similar to a wall:
- The external design requires that mechanisms outside the reinforced zone
should be stable in their own right. A circular mechanism is more likely to be
considered for a slope than sliding and bearing specifically.
- The internal design requires that sufficient total tensile capacity is available
to ensure stability of mechanisms within the block. Many method of slices
slope stability software packages allow the beneficial effects of the tensile
reinforcement forces to be incorporated in assessment of moment
equilibrium. As for a wall the maximum tensile force available will be limited
by the minimum of the tensile capacity of the reinforcement element and the
pullout resistance.


64


Figure 16 Slope stability including the effect of reinforcement (BS 8006)

65
5.1.2. The 2 part wedge
Although it is generally considered to be rather conservative, one design method
which we will (briefly) consider specifically is the 2-part wedge, as used in the
Highways Agency design code HA68-94.

Figure 17 Two part wedge

66
The variables X,
1
and
2
can be varied independently to consider a variety of
mechanisms, which may take a number of forms compared to the geometry of
the slope. However, the mechanism is constrained to pass through the toe of the
slope. Note that as we have also assumed, the underlying soil is competent.
It is generally assumed that there is no shear stress on the interface between the
two wedges since this considerably simplifies solution (although it is
conservative).
The output of the calculation is the total horizontal force (tension) from all layers
of geogrid, T
R
, required to give equilibrium for the mechanism. The factor of
safety is generally assumed to be inherent due to the conservatism in the
method, but also use of dependable strength parameters for the soil (eg. critical
state rather than peak). Some allowance is also made for sliding on a geogrid-
soil (rather than soil-soil) interface at the base of the mechanism when it is
horizontal.

67


Figure 18 Examples of two part wedge mechanisms

68
Two mechanisms are considered, corresponding directly to external and internal
equilibrium:
- The T
R,0
mechanism which has a flat base and is otherwise the smallest
mechanism to give T
R
= 0. Since no additional force from reinforcement is
required to give stability this mechanism is used to determine an initial
estimate of the extent of reinforced block.
- The T
R,max
mechanism, which is smaller, generally has a sloping base, and
gives the maximum value of T
R
for the slope geometry and soil strength.
Typical such mechanisms are shown below.

69



Figure 19 Examples of T
R,0
and T
R,max
mechanisms

70
Results for T
0
and T
max
mechanisms have been tabulated for a range of soil
strengths () and slope angles (). The results can then be expressed in a form
that is independent of the height of the slope (h
s
):
2 /
2
s
R
h
T
K

=
So K is entirely equivalent to the earth pressure coefficient for a vertical face.
If K is known then the corresponding total amount of reinforcement can be
determined
2
2
s
R
h K
T

=
[We will not consider here specifically how this reinforcement is distributed
throughout the height of the slope]
Note that the total amount of reinforcement required increases as the square of
the height of the slope.
The graph below shows a plot of typical data for the T
R,max
mechanism.

71

0.0
0.1
0.2
0.3
0.4
0.5
0.6
15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90


Figure 20 Data for T
R,max
from HA86-94 (u =0 and factor for base sliding =0.8)

72
Actual data from HA68-94 is shown joined by solid lines. The dotted lines are
extrapolated to show that the data is consistent with an unreinforced slope, or a
vertical wall.
Selecting a value of and following the corresponding line from left to right we
see that:
- K = 0 when = . This corresponds to an unreinforced slope, which can
stand without any reinforcement.
- As increases K increases steadily, indicating that more reinforcement will
be needed.
- When reaches 90
o
K = K
a
and it is as if we were designing a wall.
We can also see that (not surprisingly) it is slightly conservative to design a 70
o

slope as if it was a vertical wall, most particularly if the soil has a high friction
angle.


73
6. Design example for wall
A 6.0 m high geogrid reinforced soil vertical wall using soil with = 30
o
, =
18 kN/m
2
(and no surcharge load) is to be constructed using light concrete facing
panels with dimension of 1.0 m x 0.6 m (length height). Each layer of concrete
panel is reinforced at mid-height by one layer of geogrid which is 1.0 m wide and
4.0 m long. Two types of geogrid are available, with design strength 20 and 30
kN/m.
You are only required to consider the internal stability of the reinforced soil wall.
Your solution must specify the type of geogrid to be employed at each level, as
well as the FoS against tensile failure and pullout at each level . The values given
for geogrid strength are design values, and hence F
s
1.0 will be adequate.
However, F
s
1.5 is required for pullout. You may assume that the geogrid-soil
interface angle is 80% of the soil friction angle (using tan).


74
You are required to submit 2 designs.

The preliminary design should assume that:
An active wedge forms behind the facing, and the maximum tension in the
geogrid occurs at the back of this wedge.
Fully active pressures determine the reinforcement tension throughout the
full height: K
a
= (1-sin)/(1+sin).
The distribution of vertical stress in the reinforced soil is not affected by
rotational equilibrium of the wall.


75
The detailed design should assume that:
The potential failure wedge is bounded by two surfaces: a vertical one at
0.3h
w
behind the facing, and the active failure surface passing through the front
toe of the facing.
That fully active pressure occurs at more than 6 m depth in the soil. At the
ground surface the at rest pressure (K
0
= 1sin) occurs due to compaction
stresses. Variation of K between 0 and 6 m depth is linear.
For the vertical effective stress distribution within the reinforced soil, an
eccentric rectangular distribution is used.

Here we will consider layer 6 of the 10 (at depth 3.3 m) in some detail results
from the other layers are summarised at the end of the handout.

76
6.1. Preliminary design
We start by calculating the active earth pressure coefficient:
=
' +
'
=

sin 1
sin 1
a
K 0.33
and the width of the active zone at 3.3 m depth:
l
r,a
= (6.0-3.3) tan 30 = 1.56 m.

The nominal vertical stress at this depth is given by:

v
= 3.318 = 60 kN/m
2


Hence
T
r,max
= K
a

v
A
f
= 0.33600.6 = 12 kN
and the lower strength reinforcement will be suitable at this level with
F
s
= 20/12 = 1.67

77
We must now consider pullout. For the preliminary design the length resisting
pullout is:
l
r,po
= l
r
- l
r,a
= 4.0-1.56 = 2.44 m

Hence
T
r,po
= 2 b
r
l
r,po

v
tan
r
= 21.02.4460tan 25 = 137 kN (for 1 m width)
(where
r
= tan
-1
(0.8tan30) = 25
o
)
F
s
= 137/12 = 11.4
So this is clearly not a problem recall that pull-out is generally more critical near
the surface of the fill.


78
6.2. Detailed design
We now calculate the earth pressure coefficient including the effect of
compaction:
= ' = sin 1
0
K 0.50
So that at 3.3 m depth
K = 0.50 - (3.3/6.0)(0.50-0.33) = 0.41

We now need to know the value of e for the assumed rectangular distribution of
vertical stress:
2
6
1
|
|
.
|

\
|
=
r
a
r
l
z
K
l
e
= 0.170.33(3.3/4.0)
2
= 0.038 0.15 m
[Note that since we are now concerned with the load on the back of the
reinforced soil block (rather than the face) we do not consider modified K here.]

79
Thus the equivalent width for the rectangular distribution is
l
r
- 2e = 4.0 20.15 = 3.7 m
and

v
= 604.0/3.7 = 65 kN/m
2


T
r,max
= K
a

v
A
f
= 0.41650.6 = 16 kN
and the lower strength reinforcement will still be suitable at this level although
now
F
s
= 20/16 = 1.25


80
The length resisting pullout is now:
l
r,po
= l
r
- l
r,a
2e = 4.0 1.56 20.15 = 2.14 m
(here l
r,a
< 0.3h
w
= 1.8, so this limit doesnt have any effect)

Hence
T
r,po
= 2 b
r
l
r,po

v
tan
r
= 21.02.1465tan 25 = 130 kN/m
F
s
= 130/16 = 8.1


81
6.3. Full results

82
nom = nominal values from preliminary design
com = more complex analysis for detailed design

Restricting the extent of the active zone according to the assumption used here
affects the top half of the wall. This has a significant beneficial effect for pullout
resistance in the detailed analysis.
Conversely the effect of rotational equilibrium on the magnitude of vertical stress
only begins to have significant effect for the bottom half of the wall.
The T
r,max
values are consistently higher for the detailed analysis. Over the top
half of the wall this is due to the effect of the compaction stresses which have
been assumed. For the bottom half of the wall the effect of compaction reduces,
but the effect of eccentricity (and thus magnitude) of the vertical stresses
increases.
Pullout is most critical near the surface. It would be more of an issue for steel
reinforcement, and here l
r
could potentially be reduced if external equilibrium
allowed.

83
6.4. Notes on equations in exam
The handout for the module exam will consist of the following:
- Diagram for locus of maximum tension force (active zone)
- Diagram of earth pressure including the effects of compaction

The following information will not be included:
Equations for eccentric rectangular or trapezoidal distributions of vertical stress
(Schlosser or Bolton methods). You will be given the relevant equation to use in
a question, but marks may also be awarded for demonstrating that you can
derive this equation yourself. In the Schlosser method you would be expected to
derive the expression for vertical stress from the expression for eccentricity.
Equations for calculation of T
r,max
, T
r,des
or T
r,po
. If you understand what you are
doing the formula is force = stressarea (and the relevant variables will be
evident), which is fundamental enough to be considered common sense. You
will be told any material or interface properties which you need.

You might also like