You are on page 1of 136

National University of Singapore

Department of Civil Engineering

CE 5112
Structural design and construction of
deep basements &
cut & cover structures
Lecture 4/5
1
1

Words of wisdom
1. The state of mind which enables a man to do work of this kind ... is akin to
that of the religious worshipper or the lover; the daily effort comes from no
deliberate intention or program, but straight from the heart. "Principles of Research"
2. The ordinary adult never gives a thought to space-time problems.... I, on the
contrary, developed so slowly that I did not begin to wonder about space and
time until I was an adult. I then delved more deeply into the problem than
any other adult or child would have done.
3. The important thing is not to stop questioning. Curiosity has its own reason
for existing. One cannot help but be in awe when he contemplates the
mysteries of eternity, of life, of the marvelous structure of reality. It is
enough if one tries merely to comprehend a little of this mystery every day.
4. My interest in science was always essentially limited to the study of
principles.... That I have published so little is due to this same circumstance,
as the great need to grasp principles has caused me to spend most of my time
on fruitless pursuits.
5. One thing I have learned in a long life: that all our science, measured against
reality, is primitive and childlikeand yet it is the most precious thing we
have.
Albert Einstein

2
2

Words of wisdom
1. A hundred times every day I remind myself that my inner and outer
life depend on the labors of other men, living and dead, and that I
must exert myself in order to give in the same measure as I have
received.
2. There are two ways to live your life. One is as though nothing is a
miracle. The other is as though everything is a miracle.
3. Before God we are all equally wise - equally foolish
4. Everything should be made as simple as possible, but not simpler.
5. The search for truth is more precious than its possession.
6. I have never belonged wholeheartedly to a country, a state, nor to a
circle of friends, nor even to my own family. When I was still a rather
precocious young man, I already realized most vividly the futility of
the hopes and aspirations that most men pursue throughout their
lives. Well-being and happiness never appeared to me as an absolute
aim. I am even inclined to compare such moral aims to the ambitions
of a pig. (Written in old age?)
Albert Einstein
3
3

Practical Design Considerations


1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
13)
14)

Introduction sharing of structural engineer perspectives


General requirements clients, builders & designers
Ground, soil profile & gases
Concept of effective stress vis--vis total stress
Groundwater control
Movements caused by excavation activities
Methods of construction
Types of earth retaining system
Influence of foundations type adopted
Site Investigation
Geotechnical & structural analysis, soil-structure interaction
Protective measures
Durability and waterproofing
Safety, legal and contractual issues & risk communications4
4

Design and analysis of retaining system


Shortcoming of Current (UK) Practice
There is a lack of clear authoritative guidance on appropriate
design standards or code of practice for the design temporary
retaining system. Consequently there is the absence of an
industry-wide approach.
The design of the temporary retaining system within a limit
state framework need to set up to meet both geotechnical and
structural considerations. In the limit state approach ultimate
failure (ULS) & failure caused by loss of serviceability (SLS)
(e.g. excessive deformations) are treated separately & different
factor of safety apply to each.
The absence of a standard approach to design has led engineers
to apply design guidance for permanent works (e.g. BD42/94)
to the design of temporary retaining systems. It is never the
requirement that the temporary works be designed to the same
standards as the permanent works and this misuse led to overconservative design.
5

DESIGN STANDARDS (UK)


Two design standards in use in the UK that cover the derivation of loads for the
design of retaining systems for deep excavations:
BS8002 (1994): Code of practice for earth retaining structures
Eurocode 7 (EC7) (1995): Geotechnical design (A pre-standard to replace
BS8002 by 2010).
BS8002 aims to be a limit state code, but its approach is unclear and so not been
widely adopted. For singly propped walls, the code is clear about the serviceability
limit state (SLS) but unclear about the partial safety factor for the ultimate limit
state (ULS).
For multi-propped walls, BS8002 recommends the use of the Peck envelopes to
obtain a prop load, but it does not provide guidance on how this load should be used
in SLS and ULS calculations. This leaves some gaps at the interface where prop
loads derived from the geotechnical design are used in the structural design.
Only method that gives characteristic prop and waling loads can there be a proper
interface between the geotechnical and structural designs. These characteristic loads
can be used with any of the limit states codes (e.g. steel, concrete) and be factored
appropriately to give SLS and ULS prop loads.
CIRIA Report 104 (1984) was adopted as an unofficial design standard before the
publication of BS8002 and is still widely used, due to familiarity and also concerns
about BS8002. C104 did not address multi-propped walls, but its principle of
factoring soil strength has been used in analyses of such walls by deformation
methods. In a similar way, Eurocode 7 includes the principle of factoring soil
6
strength as ULS Design Case C.
6

EUROCODE 7: GEOTECHNICAL DESIGN


EC7 philosophy applies to the design of temporary
support systems for deep excavations in general,
requires a design be verified for three separate cases
A, B and C. In each case both ultimate and
serviceability limit states are specified.
Case A is not relevant to the design of temporary
propping systems.
In Case B the actions (i.e. loads and imposed
displacements) that act on the retaining system are
increased and characteristic values are taken for the
properties of the soil.
While in Case C lower factors apply to the actions
but the soil properties are reduced (factored).
7
7

EUROCODE 7: GEOTECHNICAL DESIGN


The relevant features of the EC7 philosophy are:
1. EC7 states that Case B is often critical when
determining the strength of the structural elements
of retaining walls while Case C is generally critical
in cases where the strength of structural elements is
not involved. The partial factors for ULS for these
two cases are CIRIA 517:

1.25
8

Table EC7 partial factors for ultimate limit states in permanent and transient situations
8

EUROCODE 7: GEOTECHNICAL DESIGN


Partial factors for ultimate limit states Geo/Str DA1
Footing, Walls and Slopes

9
9

Plaxis and EuroCode 7 - issue 16 / Oct 2004


Considering the safety of slopes and excavations, distinction is made in EC7 in three different
design approaches: DA1, DA2 and DA3, whereas in DA1 two sets of partial factors have to be
considered (DA1/1 and DA1/2). Moreover, distinction is made between Actions, Soil
Properties and Resistances.
With the current option of -C reduction in Plaxis it is, to a certain extend, possible to prove
that situations comply with DA1 or DA3. DA2 involves an increase of unfavourable
permanent action. This means for a situation of an excavation that the active soil pressure
behind a wall (= unfavourable permanent action) needs to be increased by a factor 1.35.

10
10

EUROCODE 7: GEOTECHNICAL DESIGN


Partial factors for ultimate limit states for persistent
and transient situations (Japan)

* Partial factors not relevant, and hence not provided, for Case A.

11
11

EUROCODE 7: GEOTECHNICAL DESIGN


Partial factors for ultimate limit states for persistent
and transient situations.
Actions
Case

Permanent

Ground Properties
Variable

Unfavourable

Favourable

Unfavourable

tan

c'

cu

qu #

Case A

1.1

0.90

1.50

1.1

1.3

1.2

1.2

Case B

1.35

1.00

1.50

1.0

1.0

1.0

1.0

Case C

1.00

1.00

1.30

1.25

1.6

1.4

1.4

Compressive strength of soil or rock.

12
12

EUROCODE 7: GEOTECHNICAL DESIGN


2. Permanent actions include pressures caused by ground,
groundwater and free water. Variable actions may alter
with time and include surcharges and temperature effects
on the prop loads.
3. Design pressure due to ground and groundwater may be
derived using the partial factors in Table or by other
methods. The partial factors in the Table indicate the level
of safety appropriate for conventional design in most
circumstances and are to be used as a guide to the required
level of safety when the method of partial factors is not
used.
Where design values for ultimate limit state calculations
are assessed directly, they are selected such that a more
adverse value is extremely unlikely to govern the
occurrence of the limit state.
4. The characteristic value of a parameter is one that is a
cautious estimate of the value governing the occurrence of
limit state. If statistical methods are used the probability of
13
a worse value is not greater than 5%.
13

EUROCODE 7: GEOTECHNICAL DESIGN


5. The Table indicates that for the method of partial factors
all permanent characteristic earth pressures on both sides of
the wall are multiplied by 1.35 if the total resulting action
is unfavorable, and by 1 if the total resulting action effect is
favorable. Variable characteristic earth pressures are
multiplied by 1.50. However, it also permits the partial
factors to be applied to the action effects derived from the
characteristic earth pressures (i.e. multiply prop loads from
permanent actions by 1.35 and from variable action by
1.50). EC7 states that this latter method should be used for
the design of the structural elements of a retaining wall
system.
6. For ULS, the design water pressures should be the most
unfavorable values occur in extreme circumstances. For
SLS the design water pressures should be the most
unfavorable which could occur in normal circumstances.
7. For ULS calculations, the excavation depth should be
increased by 10% of the height beneath the lowest support,
14
up to a maximum of 0.5 m.
14

EUROCODE 7: GEOTECHNICAL DESIGN


Design water pressures are affected by Tide and Rainfall

15
15

EUROCODE 7: GEOTECHNICAL DESIGN


& Distributed Prop Loads (DPL)
Analyses of propped excavations in soft clay, stiff clay and dry
sand have been undertaken to establish that Case B is likely to
give the higher ultimate prop load in most situations (Case C
only gave significantly higher loads for 3 out of the 20 props
for excavations in dry sand). Case C aims to address
uncertainty in the ground. Where Case C gives the higher prop
load, the distributed prop load method will usually account for
this because it is based on actual field data. The recommended
characteristic DPL diagrams have been assessed conservatively
and it is reasonable for designers in conventional situations to
conclude that only Case B has to be considered.
The distributed prop loads (DPL) are the action effects of
ground and water pressures. Following EC7 Case B
philosophy, characteristic values of the DPL can be multiplied
by 1 to give the serviceability limit state (SLS) design values,
and by 1.35 (permanent) or 1.50 (variable) to give the ultimate
limit state (ULS) design values.
16

16

EUROCODE 7: GEOTECHNICAL DESIGN


& Distributed Prop Loads (DPL)
METHOD OF DESIGN
The method of establishing the SLS and ULS prop
and waling loads from the DPL diagrams is
straightforward. The SLS prop and waling loads are
calculated from the characteristic distributed prop
load diagram recommended for the relevant soil class
with the partial factor of 1.0. The ULS prop and
waling loads are obtained by multiplying the
characteristic prop loads by a factor of at least 1.35,
except for the load contributed by variable actions
such as surcharges that should be multiplied by 1.50.

17
17

Characteristic Distributed Prop Load diagrams for


Class A, B & C soils

18
18

Characteristic DPL diagrams for Class A soils

19
19

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall

Simplified soil profile and


strutting/excavation sequence
of the Swiss Tower project,
Taipei. (Chang & Wong)

20
20

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall

Soil profile and


geometry adopted for
parametric study using
computer program
EXCAV95

21
21

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall
When the Ei/cu ratio falls between 200 to 500, except for the lowest strut, the
reference APD underestimates strut forces by as much as 100%. As the Ei/cu
ratio increases to 1000, the reference APD becomes more applicable.

(a) Strut force intensity vs. soils initial


tangent modulus (Ei/cu10-1)

(b) strut force exceedance ratio vs.


22
soils initial tangent modulus.
22

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall
Strut forces is dependent on diaphragm wall stiffness. The greater the wall
stiffness, the larger are the strut forces. This phenomenon is believed to be
linked to the arching effect induced by wall displacement. Stronger arching
effect associating the larger displacement of thinner walls reduces the
corresponding strut forces.

(a) Strut force intensity vs. wall


stiffness

23
(b) Strut force exceedance ratio vs.
wall stiffness
23

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall
Strut forces increase with higher preloading levels. The strut force exceedance
ratio appears to increase linearly with the preloading level when the
preloading level exceeds 20%.

(a) Strut force intensity vs. preloading

(b) Strut force exceedance ratio vs.


24
preloading
24

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall
Effect of wall penetration (D)
As the normalized wall penetration, D/T, increases beyond a ratio of 0.5, the
strut forces do not vary significantly.
But when D/T drops below 0.5, the effect of wail penetration becomes
noticeable with an increase in the exceedance ratio. An exceedance ratio of 2
seems to be sufficient in enveloping the observed strut force variations.
Effect of excavation width (B)
Varying excavation widths appear to have no significant effect on the strut
force as long the excavation width is about 3 times the excavation depth for a
constant B/T ratio of 1.
For a narrower excavation, more passive resistance below the excavation level
is mobilized due to the interaction from the side walls, which results in
reduced strut forces, especially at the lower strut levels. It appears that, when
the B/H ratio is below 1, the reference APD is able to compute strut forces
satisfactorily.
25
25

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall
Effect of thickness to hard stratum (T)
The thickness of soft clay below the excavation level is seen to impose a
strong effect on the strut force. When the T/B ratio drops below 1, the
restraining effect from the presence of hard stratum at shallow depth reduces
the strut force. A maximum of 80% reduction is noted at the lowest strut level
as the T/B ratio scales down from 1 to 0.25.
Conversely, as the T/B ratio increases above 1, the effect of clay thickness
becomes negligible. It is, therefore, postulated that, for a braced excavation
with a T/B ratio more than 1, the layer of soil below a depth of l.0B from
the excavation level could be neglected from the strut force analysis.
Effect of number of strut levels
Regardless of the number of struts, the shape and the magnitude of the
apparent pressure diagram tend to remain the same, provided that no
buckling develops in any of the strut. A strut force exceedance ratio of 2 is
sufficient in encompassing all variations of the strut forces.
26
26

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall
By adopting the amended APD and a factor of safety of 1.5 for temporary
work, the factored strut forces are sufficient in encompassing the maximum
strut force exceedance ratio, regardless of any variation to the configuration.
The amended APD is derived from cases with T/B ratio greater than 1. When
the T/B ratio is less than 1, the strong restraining effect from the underlying
hard stratum reduces the strut force.

Proposed amended
Apparent Pressure
Diagram
is the strut force
exceedance ratio
27
27

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall
Correlation among strut force exceedance ratio, Ei/cu value and cu*/H value.

28
28

Apparent Pressure Diagram for braced excavations in


soft clay with diaphragm wall (Chang & Wong)
CONCLUSION
1. For braced excavation in soft clay with diaphragm wall, the Terzaghi-Peck
Apparent Pressure Diagram tends to underestimate the strut forces when
the ratios of Ei/cu and cu*/H fall below 500 and 1.5 respectively.
2. The forces at the top and the lowest strut levels tend to be near or below
the value computed from the reference APD, regardless of the excavation
configuration and shear strength variations.
3. The shape and the magnitude of the Apparent Pressure Diagram are not
affected by the number of strut levels.
4. For T/B ratio > 1, a value of 2 appears sufficient in enveloping all possible
variations in the strut forces.
5. For T/B ratio < 1, the restraining effect imposed by the underlying hard
stratum reduces the strut force. The shallower the soft clay deposit, the
smaller are the strut forces.
29
29

Condition For the Use of DPL Method


The distributed prop load method is based on
empirical relationships and its selection for use on
any project should be considered carefully. It is
advisable to use one or more alternative methods as
well and to compare the results obtained. When
considering whether the DPL method is appropriate
the engineer should consider:
1. Is the specific site stratigraphy covered by the data set?
2. If the answer to (1) is No, do the site specific soils
behave in a similar way to the soils in the data set, i.e. do
the specific soils behave differently from the general Class
A, B or C soils? Is it reasonable to apply DPL
recommendations to the site?
30
30

Condition For the Use of DPL Method


3. Is the geometry of the excavation and propping system within the
range represented by the data set? This should be considered
particularly in regard to:
width of excavation
depth of excavation
number of props and their horizontal and vertical spacing
duration of propping
installing props before excavating below the prop level
4. Do the limitations stated for each soil class apply, e.g. T/B < 0.5 &
T/H < 0.8 for excavations in soft clay where the wall enhances base
stability?
5. Are there any other unusual features of the project, e.g. very high
surface surcharges, which might make the DPL recommendations
inapplicable?
31
31

Condition For the Use of DPL Method


Sufficient Toe Embedment
The construction of the DPL diagram for each case
history assumes that the bottom of the excavation is a
prop. The soil in front of the toe of the wall is
assumed to support the wall between the base of the
excavation and halfway towards the lowest prop, as
well as the earth pressures from the retained ground
over the embedment length. The engineer should
check that the wall embedment is sufficient to satisfy
this assumption, with a factor of safety on the passive
pressures or soil strength appropriate to the allowable
movement of the toe of the wall.
32
32

Condition For the Use of DPL Method


Surcharges
The characteristic DPL diagrams include for the
nominal surcharges associated with general
construction activities and adjoining roads. These
will not generally exceed a distributed surcharge of 10
kPa. Identifiable additional loads such as tower
cranes, mobile cranes, material storage and loads
from adjacent buildings should be treated separately.
These extra surcharges should be multiplied by the
Rankine active earth pressure coefficient and added to
the characteristic DPL diagram, provided they are a
second order contribution to the load diagram. If the
surcharge effect contributes to a significant portion of
the earth pressure, the resulting prop loads should be
33
corroborated by another method.
33

Condition For the Use of DPL Method


Preload
The characteristic DPL diagram can be applied to
situations where preloading is used to remove slack in
the support system. Preload applied to remove slack
should not exceed 15% of the characteristic prop load.
Higher values are not usually required and may result
in prop loads that exceed the characteristic value.
Some authority asked for 50% preload!

34
34

Condition For the Use of DPL Method


TEMPERATURE EFFECTS (Not thermal)
The characteristic DPL diagrams do not allow for changes in
temperature but it is not necessary to increase the calculated
loads to allow for temperature increases. It is instead
recommended that the temporary support system should be
designed for the characteristic prop loads. The resulting
structural members should then be checked using the simple
serviceability and ultimate limit state criteria. These criteria
will often be satisfied.
Prop removal
Available data indicate prop removal can increase the prop
load by up to 30%, but may equally have very little effect
depending on the specific circumstances of the site.
The characteristic DPL diagrams do not include for prop
removal. Prop loads should be established from DPL diagrams
for both excavation to final level & the subsequent sequence of
removal of props. However, the higher loads so calculated
need not control the structural design, as it may be possible35to
adopt lower partial safety factors.
35

Condition For the Use of DPL Method


Mixed support systems
The case histories on which the method is based do not permit
an assessment to be made of the influence of ground anchors
on prop load distributions. The potential for the stiffer
propping system to attract a disproportionate amount of the
load relative to the prestressed, but less stiff ground
anchorages. The high prop loads is the result of lower than
expected stiffnesses for the ground anchorages.
Where props are used as part of a combined support system of
props and anchors the DPL method of calculating prop loads is
not applicable. (FEM)
Frost effects
Case history in Norway showed frost effects on the ground can
increase the prop loads dramatically (e.g. 800%). Such frost
effects are not included in the characteristic DPL diagrams. 36
36

Structural considerations
Eccentric Axial Loading
No eccentricity of axial loading should be assumed in the
design if the end plate is grouted/concreted to a concrete
waling or the connection to a steel waling has been designed to
eliminate eccentric loading e.g. by spherical bearings.
For other situations, CIRIA Special Publication 95 gives the
following advice on the eccentricity of axial load to be used for
the prop design:
for walings made from a single section (UC or UB), the
eccentricity should be approximately 10% of the overall
dimension of the prop in the vertical plane
where the walings are constructed from twin beams, the
eccentricity in the vertical plane should be the distance
between the webs of the two beams.
37
37

Structural considerations
Accidental Loading
The provision in the design for accidental loading, & possible
loss, of a prop depends on the risk and consequences of failure.
These are matters of judgement for the designer and the project
team, which should always be given thorough consideration
and evaluation.
It is recommended that this loading condition should be
considered in the design unless positive steps are taken in the
management and operation of the site to eliminate effectively
the risk of accidental loading or loss of a prop.
CIRIA Special Publication 95 suggests accidental loading be
considered as a load of 10 to 50 kN applied normal to the prop
at any point in any direction.
38
38

Structural considerations
WALINGS
Tue design and construction details of walings are covered in CIRIA Spcial
Publication 95, to which reference should be made. Some of the salient
points are mentioned here.
While the waling will be designed for a uniform loading, the actual load
will vary considerably depending on the variation of the ground and its
movement, any arching effect, the construction methods, quality of
packings between the wall and the waling, etc. It is therefore normal to use
a simplified approach to design.
Goldberg et al (1976) recommended using 80% of the design prop load
determined from the Peck envelope for design of the waling to the
American permissible stress code (AISC). The reduction was an allowance
for arching of the soil resulting from deflection of the waling between the
props.
The waling deflection depends on the stiffness of the wall and waling, and
the spacing of the props. It is likely to be small for stiff walls, especially in
stiff ground conditions, e.g. Class B and C soil profiles. Consequently, it is
recommended that the waling is designed for 100% of the prop design load
unless the effects of arching are assessed.
39
39

Structural considerations
The walings should be continuous over two or more supports and be
designed for a max. bending moment of wL/l0, where w is the waling
load per unit length and L is the horizontal prop spacing. If continuity is
not possible, it should be designed for wL/8. Similarly the end of a
continuous waling acts as cantilever about the last support and should be
designed for wL/2.
The waling design should consider the effects of load increases from
temperature rise in the same way as for props.
Where a waling acts as a prop to another waling or the arrangement
involves diagonal props, the waling has to resist both the axial (in-plane)
load and the bending moments and shears due to the out-of-plane load. If
there is an imbalance in the axial force in the complete waling system, the
load is transferred into the ground via the wall. Sufficient shear connection
between wall and waling is needed and the wall must provide bending &
shear capacity for these in-plane forces in combination with those out-ofplane. It is also necessary as part of the wall design to consider how the
wall will act to transfer the in-plane loads into the ground (e.g. as a
diaphragm or as individual elements).
Where raking props are used, the waling and wall should be designed to
40
support the vertical component of load with minimal deflection.
40

Structural considerations - PROGRESSIVE COLLAPSE


The design of individual props should be robust but in addition the designer
should also consider the implications of the accidental loss of a prop. This
may be done in one of two ways:
1. Incorporating the loss of a prop into the design of the support system.
This design case could be combined with reduced partial safety
factors, reflecting the accidental nature of the loading. Collapse of the
excavation would be prevented, but there could be large wall and
ground movements close to an ultimate limit state. These movements
could damage adjoining property and impair the watertightness of the
retaining wall and its subsequent serviceability.
2. A risk assessment and management strategy to eliminate the risk of
accidentally damaging/removing a prop.
This aspect of the design of the temporary propping system is of
interest to the client and the engineer as well as to the main
contractor, and is often interpreted differently. Widely differing
viewpoints were expressed. Any requirements of the client/engineer
should be specified in the tender documents.
Some of the engineers contacted during the study considered it good
practice to design the props to have greater capacity than the waling. The
difference in capacities is chosen such that the waling exhibits signs of
overloading before the props become overloaded and hence provides an
41
early warning of impending prop failure.
41

Structural considerations
PROP REMOVAL
Prop removal is often the most critical stage of
construction for both the prop and walls. It can be the
worst design case and is easily overlooked. The
removal of props may cause the largest prop loads and
waling spans.
It is important that the general principles of the prop
removal sequence are agreed between the temporary
works designer and the site staff (and that any
constraints arising from the permanent works are
identified). A situation in which the constructed
support system is not sufficient to permit the site
staffs preferred method and sequence of removal
should be avoided.
42
42

Structural considerations
It may be appropriate to adopt reduced partial safety factors for the
elements of the support system during prop removal. This will depend
primarily on:
the load increase from removal of other props
duration of increased loading
whether the increased loading is combined with the maximum
temperature rises
the amount of support offered by the constructed permanent works,
and hence the consequences of potentially excessive prop
deformations.
Where walings carry axial (in-plane) loads, the props will be acting as
intermediate supports, so reducing the effective length of the waling. It is
necessary to check that the waling will not buckle when the props are
removed.
Methods of prop removal can increase the prop and waling loads. Props
may be unloaded by jackmg the wall back but the movement required to do
this can give rise to very large increases in load. The structural capacity of
the prop, connection, waling and wall may be exceeded unless such jacking
43
was allowed for in the original design of the support system.
43

CIRIA 580 Geotechnical characterization of


retaining walls
CIRIA 580 establishes design requirements
for geotechnical categories 1, 2 and 3. Prior
to the geotechnical investigations, the
designer should assign a geotechnical
category to the earth retaining structure. The
category indicates the degree of effort
required for site investigation & design. This
should be reviewed and changed (if necessary)
at each stage of the design and construction
process.
44
44

Geotechnical categorization (Simpson and Driscoll, 1998)


Small & relatively simple?

No

Yes
Ground conditions known from
comparable experience to be
straightforward, routine design &
construction methods?

No

No

Yes

CATEGORY 1
Small & relatively simple
Earth retaining system less
than 2m in depth

Unusual & exceptionally


difficult ground?

Yes

Abnormal risks?

Yes

No
No

Yes

Negligible risk to life &


property?

Yes

No

Yes
Site free of abnormal risks e.g.
unusual loading, seismic risk?

Structure very large &


unusual?
No

Yes
Excavation below water table &
comparable experience indicate
straightforward solution?

No

Loading conditions
unusual or abnormal?

Yes

No
No

High seismic area?

Yes

No

Yes

CATEGORY 2

CATEGORY 3

Conventional
Retaining system supporting
soil & water
Bridge piers & abutment

All other earth retaining


45
systems
45

CIRIA 580 Geotechnical categories EC7 (1995)


Category 1
Walls are small and relatively simple structures with the
following characteristics:
retained height does not exceed 2 m
ground condition are known from comparable experience
to be straightforward enough to allow routine methods of
design and construction to be used
previous experience indicates that a site-specific
geotechnical investigation will not be required
there is negligible risk to property or life.
Comparable experience is defined as:
documented or other clearly established information related
to the ground being considered in design, involving the same
type of soil & for which similar geotechnical behavior is
expected, & involving similar structures. Information gained
46
locally is considered to be particularly relevant.
46

CIRIA 580 Geotechnical categories EC7 (1995)


Category 2
walls comprise conventional structures with no abnormal
risks or unusual or exceptionally difficult ground or loading
conditions. These walls require site-specific geotechnical
data (e.g. a desk study and ground investigation) to be
obtained and analyses to be carried out.
The majority of embedded retaining walls fall into
geotechnical category 2.

Category 3
walls are structures or parts thereof that do not fall within
the limits of geotechnical categories 1 & 2. These include
large or unusual structures, structures involving abnormal
risks, or unusual or exceptionally difficult ground or loading
conditions & structures in highly seismic areas.
Specialist advice should be sought to deal with special circumstances adequately.
47
47

CIRIA 580 ON ANALYSIS


5 major elements of geotechnical design:
understanding the geological and
hydrogeological setting of the situ and its
environs & the historical development of
the site
determination of ground stratigraphy &
groundwater conditions
understanding soil behavior
undertaking calculations & analyses
applying empiricism based on sound
judgment & experience.

48
48

CIRIA 580 ON ANALYSIS


Retaining walls with a stabilizing base
In some circumstances, a wall with a
stabilizing base (i.e. a platform extending a
short distance in front of the wall with a rigid
connection at formation level) can represent a
more economic solution than either a rigidly
propped wall or an unpropped wall of deeper
embedment.
Finite element analyses by Powrie & Chandler
(1998) suggest an optimum stabilizing base
49
width of about the retained height
49

CIRIA 580 ON ANALYSIS


Forces acting on a stabilizing base retaining wall

Relief Platform

50
50

CIRIA 580 ON ANALYSIS


Retaining walls with a stress-relieving platform
If some excavation and/or fill is needed on the retained side of
the wall, there maybe an advantage in constructing a stress
relieving platform, attached rigidly to the wall stem some
distance below the top and protruding horizontally into the
retained soil. The relieving platform will reduce bending
moments in the wall by:
(a) applying a reverse moment at platform level, due to the
weight of the soil on top of it, &
(b) reducing vertical stresses in the retained soil below
platform level. For maximum efficiency, the platform
should extend far enough into the retained soil to reduce
vertical stresses adjacent to the wall, & there may need to
be a void below it.
51

51

CIRIA 580 ON ANALYSIS


Pressure redistribution - arching
Local variations in wall movement and rotation can, for
propped or anchored walls, lead to non-linearities in lateral
stress distribution. This redistribution of stress away from the
linear-with-depth variations assumed in simple limit
equilibrium analyses can be exploited to reduce design bending
moments and wall depth if a soil-structure interaction analysis
is carried out.

Reduction of lateral stress


in the retained soil due to
arching on to a rigid prop
52
52

CIRIA 580 ON ANALYSIS

For stiff wall, where the deflection at the level of the excavated soil surface
was of the same order an the deflection at the toe, the stress distribution in
front of the wall under working conditions is approx. triangular. Measured
bending moment were in agreement with those from a limit equilibrium
calculation based on a fully active triangular stress distribution behind the
wall and a smaller-than-passive (is factored) triangular stress distribution in
front.
For flexible wall, so that the deflection at excavation level was significantly
greater than at the toe, the centroid of the stress distribution in front of the
wall under working conditions was raised. This led to smaller anchor loads
and bending moments than those given by the factored limit equilibrium
calculation.

Components of wall
displacement & definition
of a stiff wall
A stiff wall has e t

Deflection @
excavated soil
surface e

53

Deflection @ toe, t
53

CIRIA 580 ON ANALYSIS

Changing wall EI to allow for cracking, creep of concrete 54


54

CIRIA 580 ON ANALYSIS


The high short-term stiffness on OA is required to drop to the
lower long-term stiffness on line OBC. Consider an element of
structure that in the short-term baa been stressed to point A. In
the course of time, its state will move to be somewhere on line
BC. If it is in a situation in which there is no change of strain
during this change, stresses will simply relax and it will move to
point B. If, on the other hand, the load on the clement cannot
change, it wilt creep and move to point C.
If an element is at point A and the only change made is to
change the Youngs modulus in the data, further behavior will
proceed along tine AD. This does not represent creep or
relaxation. The soil-structure interaction analysis should ensure
that even if nothing moves, stresses will change from point A to
point B. If these new stresses are no longer in equilibrium, the
analysts should then indicate further strains such that the stress
55
state will move up line BC.
55

CIRIA 580 ON ANALYSIS


Wall flexural stiffness
Appropriate values of the flexural stiffness of the wall, EI,
should be used at each stage of the analysis to model wall
stiffness during construction & in the long term. Where E, is
the uncracked short-term Youngs modulus of concrete
(typically, E0 = 28 GPa) & I is the 2nd moment of area of the
section.
The calculated load effects & wall deflection will depend upon
the magnitude of the wall flexural stiffness adopted in
analysis. The value of EI assumed should be appropriate for
each construction & long term stage. For reinforced concrete
walls, this should allow for the effects of flexural crack &
concrete creep.
In subgrade reaction & pseudo-finite element analyses, it is
necessary to input explicitly the wall flexural stiffness EI for
56
all stages.
56

CIRIA 580 ON ANALYSIS


Reinforced concrete walls
For reinforced concrete, the value of EI should strictly be
determined for the section with the value changes overtime &
with long term creep, equal to 50% of the short-term uncracked
value at infinity. EI should therefore be calculated at each
construction & long term stages.
It is appropriate to adopt 0.7EI & 0.5EI during the construction
& long-term stages respectively. The way in which the reduction
in EI is applied in the analysis should be considered carefully in
a soil-structure interaction analysis. This approach is required
in moat available computer programs in which stiffness
represents response to load increments only. The same approach
may be used to model corrosion of steel sheet piles, in which I
reduces with time.
57

57

CIRIA 580 on steel sheet pile walls


Values of I for steel Larssen (U-profile), Frodingham (Zprofile), box and high-modulus piles are given by
manufacturers. The development of full section modulus in a
sheet pile wall is based on the assumption that any 2 adjacent
flanges are able to work together in bending (composite).
Z-profile steel sheet piles have their interlocks in the flanges to
develop the full section modulus of the combined wall
(BS8002). It should be noted that with Z-profile piles, the
effective section modulus will be reduced if the piles are
allowed to rotate about a vertical axis during driving: as a
rough guide, 5 of rotation will result in a 15% reduction in
the combined sectional modulus. The construction tolerances
compatible with the design assumptions must be specified in
this respect.
58

58

CIRIA 580 on steel sheet pile walls


U-profile steel sheet piles incorporates an interlock
which is located on the centre line or neutral axis of
the wall. If the two piles are able to displace relative to
one another along the interlock, then the full modulus
of the combined sections will not be realized. These
piles rely on the transfer of longitudinal shear stress
between adjacent piles through friction at the
interlocks or clutches. It is likely that shear will be
generated by surface irregularities, rusting, lack of
initial straightness & soil particle migration into the
interlocks during driving.
59
59

CIRIA 580 on steel sheet pile walls


For U-profile sheet pile walls, it is common for to
assume the full combined modulus, except in
circumstances where shear transfer may not be fully
effective, e.g.:
piles forming cantilever walls
piles cantilevering a significant distance above or
below walings
piles driven into and supporting silts and/or soft clay
piles retaining free water over a part of their length
piles that are prevented (e.g. by rock or obstructions)
from penetrating to their required toe level.
60
60

CIRIA 580 on steel sheet pile walls


For the earlier circumstances, it is common by
welding, pressing or other means, to connect the Uprofile sheet pile sections together to develop the
necessary shear resistance so that the full combined
section modulus can be relied upon in design.
Friction between the interlocks probably contribute
at least 40% of the full section modulus
development.
Little is known about the effect of clutch slippage in
sheet pile walls; significant further research is
required in this area to improve understanding.
61
61

CIRIA 580 on axial stiffness of supports


In subgrade reaction & pseudo-finite element analyses, it is
necessary to input explicitly the axial stiffness, k (in kN/m
/m run) of any temporary or permanent props calculated as
follows:

k =AE(cos2)/Ls
Where
F = Youngs modulus of the material comprising the prop
A = cross-sectional area of the prop
L = effective length of the prop (typically the half-width of
the excavation that the prop spans)
s = prop spacing

= angle of inclination of the prop from the horizontal


62
62

CIRIA 580 on axial stiffness of supports


If concrete slabs are used to support the wall (e.g. in
a top-down construction sequence), the calculated
axial stiffness of the slab should be reduced to allow
for any openings. For concrete stabs and props, the
Youngs modulus should be reduced to allow for the
effects of creep as described earlier.

63
63

CIRIA 580 EFFECT OF METHOD OF ANALYSIS


Four generic retaining wall design & commercially
available software:
limit equilibrium methods: STAWAL; ReWaRD
subgrade reaction & pseudo-finite
methods: FREW; WALLAP

element

finite element & finite difference methods:


SAFE; PLAC
The problems analysed are defined in the following
figures.
64
64

CIRIA 580 EFFECT OF METHOD OF ANALYSIS

Cantilever Wall - effective stress analysis

65
65

CIRIA 580 EFFECT OF METHOD OF ANALYSIS

Propped Wall - effective stress analysis

66
66

CIRIA 580 EFFECT OF METHOD OF ANALYSIS

Cantilever Wall - total stress analysis

67
67

CIRIA 580 EFFECT OF METHOD OF ANALYSIS

68

Propped Wall - total stress analysis


68

CIRIA 580 EFFECT OF METHOD OF ANALYSIS


The main conclusions:
in circumstances where there is little or no stress
redistribution, e.g. cantilever walls, simple limit
equilibrium calculations & soil-structure interaction
analyses (subgrade reaction or pseudo-finite element
methods & finite element or finite difference methods) are
likely to give similar wall embedment depth & wall
bending moments
for propped or anchored walls where stress redistribution
will occur, design by limit equilibrium calculations will
result in deeper walls with higher wall bending moments
compared with those obtained from soil-structure
interaction analyses. Use of soil-structure interaction
analyses may result in significant savings in wall material
69
costs, depending upon project & site-specific details
69

CIRIA 580 EFFECT OF METHOD OF ANALYSIS


where stress redistribution occurs, prop or anchor loads
calculated from limit equilibrium methods will be smaller
than those obtained from soil-structure interaction
analyses. Prop or anchor loads solely obtained from limit
equilibrium
calculations
may
be
significantly
underestimated & should be treated with caution in design
in situations where the calculated prop loads differ
significantly from those derived from experience of
comparable construction (e.g. from the APD or DPL
method), the designer should carefully investigate &
understand the reasons for the calculated values. This
typically involves a detailed review of the assumptions
made in the calculations & the carrying out of sensitivity
analyses. Such investigations should enable the designer
to adopt appropriate design values
70
70

CIRIA 580 EFFECT OF METHOD OF ANALYSIS


for walls embedded in soils where the total
horizontal pressures near the base of the wall on
the retained side are similar in magnitude to those
on the restraining side (soft clay), the results of
calculations will be very sensitive to relatively
small changes in pressures around the wall. The
results of such calculations will also be
influenced by node spacing in beam spring &
pseudo-finite element models, & mesh details in
finite element & finite difference models. The
designer should carry out sensitivity checks on
the effects of such variations in the models
71
adopted.
71

CIRIA 580 DESIGN PARAMETERS


Design parameters required:
soil and groundwater conditions
loading condition, e.g. surcharge adjacent to the wall
geometry, e.g. unplanned excavation allowance (0.5m)
factors to ensure safety, to allow for uncertainty in material & soil
properties, loading & analysis models, & to ensure acceptable
deformations.
The selection of parameters for use in design should be based on careful
assessment of the range of values determined for each parameter that
might govern the performance of the retaining wall during construction &
during its design life. It requires an understanding of the input parameters
for the envisaged design calculations & should specify an appropriate site
investigation to obtain them. At times, site investigation is carried out
before designer is appointed. In this ease, the designer may specify any
additional site investigation required to achieve an economic design.
72
72

CIRIA 580 DESIGN PARAMETERS


Soil parameters required for various design approach

73

Note: 1. Special input parameters required depending upon analytical medal adopted
73

CIRIA 580 DESIGN PARAMETERS


Knowledge of the soil density (unit weight) & shear strength is
essential in the design of an embedded retaining wall & also general
appreciation of the following soil properties:
classification and index properties, e.g. particle size
distribution, moisture content, plasticity indices (for finegrained soils)
soil permeability.
Knowledge of in situ stress conditions, particularly the value of the
in situ earth pressure coefficient Ko, & soil stiffness is essential in
soil-structure interaction analyses.
Stiff over-consolidated soils have several soil strengths: peak,
critical state, residual, and drained or undrained. There is also a
range of soil stiffnesses, depending on shear strain.
Backfill materials may require parameters for the determination of
74
compaction and swelling pressures.
74

CIRIA 580 DESIGN PARAMETERS


The 1st step is to decide which soil parameters are appropriate
for a particular analysis. Then to consider other issues such as
reliability, selection of values for design & factors of safety.
It may be appropriate to adopt different selected values for a
parameter in different limit states & design situations. E.g., in
total stress analysis, the selected value of the undrained shear
strength of the clay should consider the mechanisms or modes
of deformation being considered for the wall. Different
strengths will be required for a shear failure in fissured
material depending upon whether the shear surface in free to
follow the fissures or is constrained to intersect intact
material. A range of values should be considered. These
values should also allow for any softening due to potential
changes in moisture content and the effect of excavation
disturbance.
75
75

CIRIA 580 DESIGN PARAMETERS


Many soil parameters are not true constants but depend upon
factors such as stress & strain levels, mode of deformation,
type of analysis, etc. Under working conditions while
deformations are comparatively small, some or all of the soil
will operate at below peak strength. Under ULS where
deformations are comparatively large, the soil may operate
beyond peak strength conditions & may dilate to approach
critical state values (BS 8002, 1994).
The designer of an embedded wall in a stiff over-consolidated
soil should decide on the appropriate strength to use in a
particular circumstance. The residual strength might be
appropriate where sliding along a pre-existing polished
rupture surface represents a potential failure mechanism, but
it will in general be far too conservative in other situations.
76
76

CIRIA 580 DESIGN PARAMETERS


The choice is therefore usually between the peak & the critical
state strength & the following points should be borne in mind:
for a given soil, the critical state angle of shearing
resistance, crit, is a constant over the range of stresses
normally encountered in geotechnical engineering.
Conversely, the development of a peak angle of shearing
resistance, peak, depends on soil-structure & potential for
dilation. The latter depends in turn on the soil density &
the average effective stress during shear
failure at the peak angle of shearing resistance is brittle.
With continued post-peak deformation the soil strains &
softens, leading to the possibility of progressive failure.
The factor of safety adopted in design should therefore
ensure that displacements & strains will not be large
77
enough to take the material into the post-peak range
77

CIRIA 580 DESIGN PARAMETERS


the onset of large deformations tends to occur when about
80% of peak strength is mobilized. This applies to a wide
range of soils
in an over-consolidated soil that fails by rapture, the peak
strength is easier to identity than the critical state
at a given effective stress, denser soils (of a particular type)
have both a higher stiffness & a higher peak strength. This
is particularly relevant when retaining walls are designed
by the application of a factor of safety to the soil strength.
If critical state strengths are used in the collapse
calculation, a higher factor of safety would be needed for a
retaining wall in a loose soil than for an identical retaining
wall in a dense soil, for the wall movements under working
conditions to be the same.
78

78

CIRIA 580 DETERMINATION OF SOIL PARAMETERS


Soil behavior is influenced by the following:
initial soil state: porosity, fabric, current effective stress
state
stress, time and chemical history, which are embodied in
the yield characteristics
stress or strain path
rate of shearing and drainage conditions.
The influences of initial stress state & of history diminish
with increasing strains post-yield. Thus, sample quality to
measure small-strain stiffness differs from that required to
measure high-strain drained shear strength. Laboratory
measurement of pre-yield behavior or undrained shear
strength, the soil state & effects of history should be
preserved. This requires no change in volume or effective
79
stress. This ideal cannot be achieved.
79

CIRIA 580 DETERMINATION OF SOIL PARAMETERS


Certain laboratory tests bear the full imprint of
disturbance, The unconsolidated undrained triaxial
test is a good example since no attempt is made to
reimpose the in situ stresses, hence it is particularly
prone to sample disturbance.
The effects of sample disturbance & limitations of
many laboratory tests have contributed to poor field
predictions. This has contributed to greater use of in
situ testing, or at least of integrated laboratory & in
situ testing. A balanced view should be taken of the
advantages & limitations of both types of tests so
that they are included appropriately in a ground
80
investigation.
80

CIRIA 580 DETERMINATION OF SOIL PARAMETERS


There is no reason why a shear strength derived from an
in sits vane, pressuremeter or cone test should coincide
with that measured in a laboratory triaxial compression
or simple shear test. For a soil of a given composition,
deposition & post-deposition history, peak shear strength
will be influenced by the initial effective stress state, by
drainage during shear, by the stress path and the rate &
direction of shear. These will vary between the different
types of in situ & laboratory test and also the measured
strength. The small-strain stiffness behavior will also be
affected by the recent stress or strain history, imposed by
the sampling process. In view of stress-strain nonlinearity, comparisons are only meaningful if they are
made at corresponding levels of strain.
81
81

CIRIA 580 DETERMINATION OF SOIL PARAMETERS


Soil parameters should
independent sources:

be

determined

from

several

directly from the results of in situ & laboratory tests


from established empirical correlations between different
types of in situ & laboratory tests and with the soils
grading & index properties
from relevant published data, and local & general
experience
from back analysis of measurements taken from
comparable full-scale construction in similar ground
conditions.
The selected soil parameters should encapsulate the designers
expertise & understanding of the ground and be based on both
site-specific information & a wider body of geotechnical
82
knowledge and experience.
82

CIRIA 580 Classification properties


Classification tests
The results of classification testing are essential in
understanding material characteristics & behavior and are
necessary in the interpretation of in situ & laboratory testing.
The index properties below should be routinely determined for
fine-grained & coarse-grained soils:

83
83

CIRIA 580 Classification Permeability


The coefficient of permeability of soil, k, varies over a very
wide range of values from about 10-10m/s for practically
impervious clays to about 1 m/s for clean gravels. A range of
values for various soils is presented in BS 8004. This is
reproduced below which shows the mass permeability of
fissured clays can vary over a wide range of values.

84
84

CIRIA 580 Soil Stiffness


It is good practice to determine soil stiffness using
several different approaches. Current UK practice
includes specialist in situ self-boring preesuremeter
testing, geophysical testing & specialist sampling &
laboratory small-strain stiffness measurement.
The self-boring pressuremeter is probably the most
robust means of determining soil stiffness at strains
relevant for wall design across a broad range of overconsolidated clays & very weak rocks in the UK.
The stress-strain behavior of soil is highly non-linear
and soil stiffness decays with strain by orders of
magnitude.
85

85

CIRIA 580 Soil Stiffness


At very small strains of about 0.001%, the stiffness is large; at
strains close to failure, the stiffness is small. Atkinson and
Sllfors (1991) identity 3 regions of a typical stiffness strain
curve for soil:

86
86

CIRIA 580 SELECTION OF DESIGN PARAMETERS


The following should be considered in selecting appropriate
parameters for use in design calculations:
geological & other background information, such as data
from previous projects
the variability of the determined values, including
differences between the in situ conditions & the properties
measured by field and laboratory tests
the extent of the zone of ground governing the behavior of
the wall at the limit state being considered
the effect of construction activities on the properties of in
situ ground
changes that may occur in the field due to variation in the
environment or weather.
87

87

CIRIA 580 SELECTION OF DESIGN PARAMETERS


Uncertainty in the selection of soil strength, stiffness, loads and geometric
parameters are of particular importance in retaining wall design. The risks
of soil strength and stiffness being less or greater than assumed, or
surcharge loads being greater, or of over-excavation or a rise in
groundwater pressures occurring, influence the factor of safety appropriate
for design. Three design approaches A, B & C are discussed in C580:

88
88

CIRIA 580 SELECTION OF DESIGN PARAMETERS

Process for obtaining design values from test results.


89
89

CIRIA 580 SELECTION OF DESIGN PARAMETERS


Characteristic Strength (fck ) MeanValue ( xm ) 1.64 Standard Deviation( )
Design Strength =

Characeristic Strength (fck )


Material Safety Factor ( m )
n

Standard Deviation

x xm
1

n 1

stress

40 N/mm

26.7 N/mm
strain

EXAMPLE:
10 concrete cubes were tested in compression at 28 days. The following
crushing strengths (N/mm) were obtained:
44.5 47.3 42.1 39.6 47.3 46.7 43.8 49.7 45.2 42.7
Mean strength xm = 448.9/10 = 44.9 N/mm
Standard deviation = [(x-xm)/(n-1)] = (80/9) = 2.98 N/mm
Characteristic strength fck= 44.9 (1.642.98) = 40.0 N/mm
90
Design strength = 40/m = 40/1.5 = 26.7 N/mm
90

CIRIA 580 SELECTION OF DESIGN PARAMETERS


To find the 95% confidence level, for soil properties, as only a small portion of
the total volume involved in a design situation is tested, it is not possible to
rely on Normal Distribution.
For a small sample size the Student t value for a 95% confidence level may be
used to determine that Xck value, given by:

t
1 tV
xck xm
xm
n
n
Some typical values of V (/xm) for different soil properties given by:
Soil Property

Range of typical V
values

Recommended V Value
if limited Test results
available

tan

0.05 0.15

0.12

0.30 0.50

0.42

cu

0.20 0.40

0.32

mv

0.20 0.70

0.42

(unit weight)

0.01 0.10

91
91

t table with right tail probabilities

degrees of freedom df = n1

Probability density function

Cumulative distribution function

df\p
1
2
3
4
5

0.40
0.324920
0.288675
0.276671
0.270722
0.267181

0.25
1.000000
0.816497
0.764892
0.740697
0.726687

0.10
3.077684
1.885618
1.637744
1.533206
1.475884

0.05
6.313752
2.919986
2.353363
2.131847
2.015048

0.025
12.70620
4.30265
3.18245
2.77645
2.57058

0.01
31.82052
6.96456
4.54070
3.74695
3.36493

0.005
63.65674
9.92484
5.84091
4.60409
4.03214

0.0005
636.6192
31.5991
12.9240
8.6103
6.8688

6
7
8
9
10

0.264835
0.263167
0.261921
0.260955
0.260185

0.717558
0.711142
0.706387
0.702722
0.699812

1.439756
1.414924
1.396815
1.383029
1.372184

1.943180
1.894579
1.859548
1.833113
1.812461

2.44691
2.36462
2.30600
2.26216
2.22814

3.14267
2.99795
2.89646
2.82144
2.76377

3.70743
3.49948
3.35539
3.24984
3.16927

5.9588
5.4079
5.0413
4.7809
4.5869

11
12
13
14
15

0.259556
0.259033
0.258591
0.258213
0.257885

0.697445
0.695483
0.693829
0.692417
0.691197

1.363430
1.356217
1.350171
1.345030
1.340606

1.795885
1.782288
1.770933
1.761310
1.753050

2.20099
2.17881
2.16037
2.14479
2.13145

2.71808
2.68100
2.65031
2.62449
2.60248

3.10581
3.05454
3.01228
2.97684
2.94671

4.4370
4.3178
4.2208
4.1405
4.0728

16
17
18
19
20

0.257599
0.257347
0.257123
0.256923
0.256743

0.690132
0.689195
0.688364
0.687621
0.686954

1.336757
1.333379
1.330391
1.327728
1.325341

1.745884
1.739607
1.734064
1.729133
1.724718

2.11991
2.10982
2.10092
2.09302
2.08596

2.58349
2.56693
2.55238
2.53948
2.52798

2.92078
2.89823
2.87844
2.86093
2.84534

4.0150
3.9651
3.9216
3.8834
3.8495

21
22
23
24
25

0.256580
0.256432
0.256297
0.256173
0.256060

0.686352
0.685805
0.685306
0.684850
0.684430

1.323188
1.321237
1.319460
1.317836
1.316345

1.720743
1.717144
1.713872
1.710882
1.708141

2.07961
2.07387
2.06866
2.06390
2.05954

2.51765
2.50832
2.49987
2.49216
2.48511

2.83136
2.81876
2.80734
2.79694
2.78744

3.8193
3.7921
3.7676
3.7454
3.7251

26
27
28
29
30

0.255955
0.255858
0.255768
0.255684
0.255605

0.684043
0.683685
0.683353
0.683044
0.682756

1.314972
1.313703
1.312527
1.311434
1.310415

1.705618
1.703288
1.701131
1.699127
1.697261

2.05553
2.05183
2.04841
2.04523
2.04227

2.47863
2.47266
2.46714
2.46202
2.45726

2.77871
2.77068
2.76326
2.75639
2.75000

3.7066
3.6896
3.6739
3.6594
3.6460

infinty 0.253347

0.674490

1.281552

1.644854

1.95996

2.32635

2.57583

92

3.2905
92

CIRIA 580 SELECTION OF DESIGN PARAMETERS


The Characteristic Value of the angle of shearing resistance ck is required for
a 10m depth of ground consisting of sand for which the following values
were determined from 10 triaxial tests: 33, 35, 33.5, 32.5, 37.5, 34.5,
36.0, 31.5, 37, 33.5
Average angle of shearing resistance m = 34.44
Standard Deviation = 2.91
Coefficient of variation V = 0.0509/0.6858 = 0.0742
Student t for a 95% confidence level with 10 test results = 1.833
tanck = 0.6858 0.05091.833 / 10 = 0.6563 (ck = 33.27)
Design Value XD = Xck/m & Applying m = 1.25 for Case C
D = arc tan [(tan ck ) / 1.25] = 27.70
Average angle of shearing resistance m = 34.4
Standard Deviation = 1.97
Coefficient of variation V = 1.97/34.4 = 0.057
Student t for a 95% confidence level with 10 test results = 2.26 (wrong!)
ck = 34.4 - 1.972.26 / 10 = 33.0
Design Value XD = Xck/m & Applying m = 1.25 for Case C
93
D = arctan[(tan ck ) / 1.25] = 27.8
93

CIRIA 580 SELECTION OF DESIGN PARAMETERS


The Characteristic Value of the angle of shearing resistance ck is required for
a 10m depth of ground consisting of sand for which the following values
were determined from 10 triaxial tests: 33, 35, 33.5, 32.5, 37.5, 34.5,
36.0, 31.5, 37, 33.5 (Exel has built-in function to calculate these values)

wrong method
(X-Xm)

1.96

0.36

0.81

3.61

9.61

0.01

2.56

8.41

6.76

1.969207

in '

0.81

34.9

in '

Average

'

33

35

33.5

32.5

37.5

34.5

36

31.5

37

33.5

34.4

in '

X = Tan' 0.649408 0.700208 0.661886 0.63707 0.767327 0.687281 0.726543 0.612801 0.753554 0.661886 0.685796
'm

34.44218

Xm = tan
0.685796
'm
(X-Xm)

0.001324 0.000208 0.000572 0.002374 0.006647 2.2E-06

Correct method

0.00166 0.005328 0.004591 0.000572 0.023279

0.050858 in Tan'

2.91143

in '

94
94

CIRIA 580 Design Approach A


Moderately
conservative
soil
parameters,
groundwater pressures, loads & geometry are
selected and safety factors are applied. Moderately
conservative in a cautious estimate of the value
relevant to the occurrence of the limit state. It is
considered to be equivalent to representative values
as defined in BS 8002 & to characteristic values as
defined in EC7 (1995). This should not be confused
with characteristic values (5% fractile) adopted in
structural engineering for materiel properties.
95
95

CIRIA 580 Design Approach B


Worst credible soil parameters, groundwater
pressures, loads & geometry are selected and safety
factors lower than those in Design Approach A are
applied. This value is the worst that the designer
reasonably believes might occur - a value that is
very unlikely. As a guide, it may be regarded as the
0.1% fractile. Design Approach B is not
appropriate for SLS calculations.

96
96

EUROCODE 7: Singapore Technical Reference for


Deep excavation
The term moderately conservative is taken to mean the cautious
estimate of the value relevant to the occurrence of the limit state as in
CIRIA C580. It is also considered to be equivalent to the representative
value as in BS 8002 and to the characteristic value as in EC7.
Worst credible value is the worst value which is reasonably believed
might occur a value that is very unlikely. It is considered to be equivalent
to the conservative value as in BS8002.
The ULS design shall be based on the most onerous of:
(a) Approach 1: Earth pressures derived from design values as defined in
this Section in which the reduction factors m in Tables 3.1a or 3.1b are
appropriately applied to the moderately conservative parameters.
(b) Approach 2: Earth pressures derived from the worst credible
parameters.
97
97

CIRIA 580 Design Approach C


Most probable soil parameters, groundwater pressures,
loads & geometry are selected and the safety factors of
Design Approach A are adopted. Most probable values have
a 50% probability of exceedance. Design Approach C should
only be used within an Observational Method process. It
should be used in conjunction with Design Approach B, to
enable contingency measures to be developed for rapid
implementation in the event that conditions actually
encountered are less than the most probable. Thus, it is
unacceptable to proceed solely on the basis of Design
Approach C. The construction cost saving of this approach
should be offset against the costs relating to the additional
calculations to Design Approach B & those associated with
the development of contingency measures, the additional
monitoring & measurement systems necessary for the
implementation of the Observational Method.
98
98

CIRIA 580 DESIGN PHILOSOPHY


Limit state design philosophy
Design calculations should satisfy the ultimate limit states
(ULS) of wall stability & structural strength and the
required serviceability limit states (SLS) by verifying
satisfactory performance in respect of wall deflections,
associated ground movement, wall watertightness criteria
etc. Neither ultimate or serviceability limit states should be
exceeded in the envisaged design.
The factor Fs, should be applied on soil strength. The soil
design parameters derived therefrom should be used in
conjunction with the groundwater pressures, loads and
design geometries for collapse (ULS) calculations, SLS
calculations and the accidental design situation respectively.
99
99

CIRIA 580 DESIGN PHILOSOPHY


Fs factors appropriate for use in design calculations

1. Effective stress: tan d = tan / Fs & Cd = c / Fsc


Total stress: Sud = Su / Fssu
2. The design strength parameters in note 1 above are used to derive
earth pressure coefficients.
3. Not appropriate for SLS calculations.
100
100

EUROCODE 7: Singapore Technical Reference for


Deep excavation
Partial factors are to be used in the Ultimate Limit State
(ULS) design of the excavations system. The design
values of the geotechnical parameters Xd shall be derived
using:
X d = Xk / m
in which Xk is the moderately conservative estimate of the
soil parameter and m is the reduction factor for the
parameter. For designs based on EC7, the reduction
factors (which are termed as partial factors in EC7) are
shown in Table 3.1a. For designs based on BS8002, the
reduction factors (which are termed as mobilization
factors in BS8002) are shown in Table 3.1b.
101
101

EUROCODE 7: Singapore Technical Reference for


Deep excavation
Table 3.1a EC7 Partial factors for soil parameters (m)
(no case classification)

Soil parameter Symbol Value


Angle of shear resistance*

1.25

Effective cohesion (1.6)

1.25

Undrained shear strength

cu

1.4

Unconfined strength

qu

1.4

Weight density

1.0

* This factor is applied to tan

102
102

BS8002: Singapore Technical Reference for


Deep excavation
Table 3.1b BS8002 Minimum factors for soil parameters (m)

Soil parameter Symbol Value


Angle of shear resistance*

1.2

Effective cohesion

1.2

Undrained shear strength

cu

1.5

Unconfined strength

qu

1.5

Weight density

1.0

* This factor is applied to tan


103
103

EUROCODE 7: Singapore Technical Reference for


Deep excavation
This example (from CIRIA report 104, 1984) illustrates the ULS design to
determine the minimum penetration depth for a wall restrained with a strut (prop)
at a depth of 2 m below the original ground surface. The groundwater table is
assumed to be well below the tip of the wall. This worked example uses Approach
1 with moderately conservative values.
Soil and interface properties

Values

Soil unit weight (kN/m)

20

Friction angle

25o

Cohesion (kPa) c

Interface friction (active side)

= (2/3)

Interface friction (passive side)

= (1/2)
104
104

EUROCODE 7: Singapore Technical Reference for


Deep excavation

105
105

EUROCODE 7: Singapore Technical Reference for


Deep excavation
Applying a reduction factor for of 1.25,
the design d = tan-1{tan(25o)/1.25} = 20.5o
Based on Caquot and Kerisel (1948) with design d = 20.5o, the active
coefficient Ka = 0.44, and the passive coefficient Kp = 2.70.
Take moments about the strut.
For equilibrium (Factor of safety = MP/MA = 1.0), the minimum depth of
penetration of the wall, d = 4.10 m.
Force (kN/m)

Lever arm (m)

Moment (kNm/m)

PA = 0.5Ka(h+d)
= 3.6(8+d)

LA = (2/3)(h+d) - 2
= (2/3)(5+d)

MA = 2.4(8+d)(5+d)

PP = 0.5Kpd
= 34.7d

LP = (2/3)d + 8 - 2
= (2/3)d + 6

MP = 34.7d(6 +2d/3)
106
106

CIRIA 580 DESIGN PHILOSOPHY


Design Approach A, subscript is mc
For effective stress analysis, the limiting value of wall friction, max,
should be taken to be:
max k crit,mc
where:
crit,mc =

moderately conservative critical state angle of shearing


resistance

k=

1.0 for rough concrete (e.g. concrete cast directly against


soil) and for a rupture surface within the soil;

k=

0.67 for smooth concrete (e.g. precast concrete or


concrete cast against formwork) and other smooth
surfaces (e.g. steel) and for driven or jacked in walls.

The value of the design effective wall adhesion, Swd, should be taken as
zero.
107
107

CIRIA 580 DESIGN PHILOSOPHY


For total stress analysis, design su = sud = sumc
where:
sumc = modemtely conservative value of undrained shear
strength, su.
The limiting value of wall adhesion, Swmax, should be taken
as: Swmax = Sud
where:
=

0.5 in stiff clay. Smaller values of may apply in


particular circumstances, e.g. steel sheet piles
driven through overlying soft clay.

For design approach B & C, subscript is wc & mp


respectively
108
108

CIRIA 580 STRUCTURAL DESIGN OF WALL


The structural design of the wall should conform to the
relevant code of practice for the particular material, namely
ES8110 Part 1 (1997), BS 5400 Part 4 (1990) or EC2 Part 1
(DD ENV 1992-1-1: 1992) for reinforced concrete and ES
5950 Part 1 (2000), BS 449 Part 2 (1969) or EC3 Part 5 (ENV
1993-5, 1998) for structural steelwork.
The design of the structural members should allow for the
loads generated by the temporary & permanent construction
stages and the installation method.
Installation stresses are generated in pushed, driven or
vibrated sections. For concrete cast in situ, into a preformed hole, the reinforcement detailing should allow for
the method of placing the reinforcement & concrete.
109
109

CIRIA 580 STRUCTURAL DESIGN OF WALL


ULS wall bending moments and shear forces for use in the
structural design of the wall should be obtained as the
greater of:
the values obtained from limit equilibrium calculation or
soil-structure interaction analysis.
1.35 times the SLS values, where SLS calculations are
undertaken
the
values
calculated
for
accidental
situation/progressive failure check.

design

values arising from the use of the Distributed Prop Load


method for the design of temporary propping to the
wall.
110
110

CIRIA 580 Steel sheet pile walls


For driven sheet piling, the forces induced during the
driving process should out exceed the capacity of the
section.
Durability: Steel corrosion rates are generally low and steel
piling may be used for permanent works in an unpainted or
unprotected condition. The degree of corrosion and the
need for protection depends upon the working environment,
which can vary along the length & depth of the pile and
with time. Underground corrosion of steel piles driven into
undisturbed natural soils that do not comprise peat and are
not chemically contaminated is negligible. This is attributed
to the low oxygen levels present in undisturbed soils.
Corrosion rates are higher where steel piling is exposed to
atmospheric conditions, fresh water and marine
111
environments.
111

CIRIA 580 Steel sheet pile walls


Corrosion rates for steel piling in natural environments (after BS 8002 1994)

Note:
1. Fresh waters are variable. Corrosion losses in fresh water immersion zones are
112
generally lower than for seawater.
112

CIRIA 580 Steel sheet pile walls


The analysis of the sheet pile section for the bending moments & shear forces
imposed during the construction stages is carried out in accordance with the
relevant structural code of practice, e.g. BS 5950-1 (2000) Structural use of
steelwork in building (ULS), BS 449 (1969) Part 2: Specification for the use of
structural steel in building (SLS), or EC3 (1998) Design of steel structures Part 5: Piling (ULS).
Allowable bending stresses for steel sheet piling (after British Steel, 1997):
Steel Quality

Allowable permanent
stress (N/mm)

Allowable temporary 1
stress (N/mm)

EN 10248: 1996 S270GP

180

200

EN 10248: 1996 S355GP

230

260

Note:
1. Higher allowable stress permitted for a temporary wall on the basis that a
long-term corrosion allowance will not be required & that the increased
deflections will be acceptable under short-term leading.
113

113

CIRIA 580 Steel sheet pile walls


EC3 Part 5 (ENV 1993-5: 1998) allows the full plastic material properties of the steel to
be mobilized together with redistribution of earth pressures to achieve more economic
design. Sheet pile walls are divided into four classes, as described below:
Class 4 - sections that fail due to local buckling within their elastic capacity
Class 3 - sections that reach their elastic moment capacity. The stress distribution
across the section is elastic. The yield stress is reached in the extreme fibers of the
section such that the elastic moment capacity, Mel, is given by: Mel = Ze fy
where: Ze is the section modulus of the wall & fy is the yield stress
Class 2 - sections that take account of their full plastic moment resistance, Mpl,
such that: Mpl = 1.15 Mel
Class 1 - sections that are designed plastically & allow for moment redistribution
due to rotation of the section. The moment capacity of the section, Mu, will
depend on the ratio (b/t) where b & t are the width & thickness of the flange
respectively.
According to Hartmann-Linden et al (1997), the ultimate limit state design of sheet
piling can lead to considerable reductions to material use. For Class 2 design, this
reduction is 15-20%. For Class 1 plastic design, the material reductions are about 2530%.
In UK, it is often necessary to adopt sheet pile sections that are of greater thickness
to
114
withstand driving forces than one determined from analysis & design.
114

CIRIA 580 Steel sheet pile walls


Design of sheet pile walls to EC3, Part 5 (ENV 1993-5: 1998)

115
115

CIRIA 580 Steel sheet pile walls


Typical Summary sheet pile wall Plaxis Analysis

116
To include Maximum Strut Capacity
116

CIRIA 580 Steel sheet pile walls


Typical Shear Force diagram of sheet pile wall (Plaxis)

117
117

CIRIA 580 Steel sheet pile walls


Typical B.Moment diagram of sheet pile wall (Plaxis)

118
118

CIRIA 580 Steel sheet pile walls


Typical B.Moment
diagram of
Diaphragm wall
with removal stage
(Plaxis)

119
119

CIRIA 580 Cast-in-place concrete


Construction issues
The method of constructing the concrete member below the
ground can affect the structural design.
Continuous flight auger (cfa) piling (not used in Spore)
The use of cfa piling, where the pile is bored & concreted in a
single operation as the auger is drilled & extracted, restricts the
depth & reinforcement density of the pile. The reinforcing cage is
pushed into the wet concrete. Soil conditions that allow any free
water to flow out of the concrete will induce a premature set in
the concrete and prevent installation of the cage. High
reinforcement densities, particularly links, will also restrict the
installation depth. A small vibrator attached to the top of the
cage may ease the installation, the limitations of this method
120
should be recognized.
120

CIRIA 580 Cast-in-place concrete


Bored piles & diaphragm walls
The concrete used for piles and diaphragm walls is not usually
placed with the use of vibrator & is therefore self-compacting
with the ability to flow around the reinforcement cage. When a
drilling fluid is used to provide temporary support for the bore,
the concrete will displace the support fluid. Good quality
concrete should be ensured throughout the section, particularly
in the cover zone between the reinforcement and the soil. It is
important to note that the concrete has to flow out from the
centre of the section through the reinforcement cage to the cover
zone, which is for the long term protection to the reinforcement.
BS EN 1536 (2000) provides guidance on the size of tremie pipes
and annuli for different pile diameters and aggregate sizes.
121
121

CIRIA 580 Cast-in-place concrete


Reinforcement detailing
Bundles & large-diameter bars are used rather often in embedded
wall concrete sections because of large clear space is needed
between bars to allow concrete to flow into the cover zone. The
reinforcement cage is fabricated above ground, then lifted &
lowered into the bore. Long cages may need to be spliced over the
bore due to lifting restrictions.
Pile cages
ICE recommends that the minimum clear spacing between the
vertical bars, or bundles of bars, should be 100mm to ensure an
adequate flow of concrete through the reinforcement cage.
Multiple layers of reinforcement should be avoided.
122
122

CIRIA 580 Cast-in-place concrete


Diaphragm wall cages
Typically a min. clear spacing of 500mm is maintained around
the tremie pipe positions to allow the pipe to be installed &
withdrawn without snagging the reinforcement. Large panels
may require more than one set of tremie pipes to reduce the
distance that the concrete has to flow. The min. link spacing
required by the structural design codes may have to be
compromised in order to insert the pipe, by adjusting the
spacing locally to provide the 500mm minimum clearance.
There should be an unreinforced length of wall between
adjacent panels to allow for tolerances in excavating the panels
& to allow space for the joint detail and the waterstop, typically
400-550mm. The vertical bar spacing will be reduced over the
reinforced length of wall to allow for the unreinforced section,
which maybe up to 25% of the panel length for a single bite
123
panel.
123

CIRIA 580 Cast-in-place concrete


Diaphragm wall cages
The Essential guide to the ICE specification for piling & embedded retaining walls
(Federation of Piling Specialists & Institution of Civil Engineers, 1999) & BS
EN 1538 (2000) give guidance on the clear spacing between bars to ensure an
adequate flow of concrete through the diaphragm wall reinforcement cage.
The final clear horizontal distance between vertical bars in a single layer
should be 100mm. This figure may be reduced to 80mm over lap lengths
(although this should be avoided wherever possible). Where two layers of
reinforcement are required on a wall face, the bars in the inner layer should be
aligned behind those in the outer layer to allow concrete to flow between
them. The clear vertical distance between horizontal bars should be at least
200mm where the clear distance between the vertical bars is 100mm. The clear
vertical distance between horizontal bars can be reduced to 150mm if the
spacing of the vertical bars is increased to give a clear window area of at least
0.02m between the horizontal and vertical bars or 0.16 m over lap lengths.
Where links are required the clear horizontal distance between legs of the
links should be at least 150mm.
124
124

Cast-in-place concrete - Diaphragm wall

125
125

Cast-in-place concrete - Diaphragm wall

126
126

Cast-in-place concrete - Diaphragm wall

Moment Reinforcement

127
127

Cast-in-place concrete - Diaphragm wall

Shear Reinforcement

128
128

CIRIA 580 Cast-in-place concrete


EC2 (DD ENV 1992-1-1: 1992) detailing rules
Specifically for building structures & EC2 Part 6 for massive
civil engineering structures. EC2 sets out rules on the use of
bundles & large-diameter bars, restricting the use of bundles to
a max. equivalent diameter of 55mm & requiring all bars of
32mm diameter & above to be joined using mechanical
couplers rather than by lapping. 2 bundled 40mm bars =
56.6mm .
The design of RC on the Copenhagen Metro project was
carried out to EC2 requirements. On this project, it was
necessary to compromise the above criteria to the extent that
40mm bars were lapped within the pile section & bundles of up
to three 40mm bars were used to satisfy a specified 0.2mm
max. crack width criterion. This compromise was adopted for
129
piles 1050mm & diaphragm wall 1000mm thick.
129

CIRIA 580 Crack width control


The specification of a maximum crack width in a reinforced
concrete section is an SLS consideration and usually arises from
concerns about durability, watertightness and aesthetics. Cost
savings are possible if a pragmatic approach is taken to crack
width control.
The location and orientation of a crack is more important than
its size. Cracks of any size which are in line with the
reinforcement, e.g. along the lines of the links, may cause
corrosion, whereas cracks transverse to the reinforcement are
unlikely to cause corrosion. Rowe et al (1987) indicate that crack
widths of up to 0.5 mm may be acceptable from a durability
viewpoint.
Crack width control requirements & calculations principles to
meet the requirements of BS 8110 (1997), BS 5400 (1990), BD
130
42/00 (DMRB 2.1.2) and EC2 (1992) are:
130

CIRIA 580 Crackwidth control


Durability resistance is determined by cover & mix design.
BS8110 Part 1 (1997): The wall should be designed for
durability in accordance with Clauses 2.2.4, 2.4.7, 3.1.5,
4.1.5 & 6. The minimum cement content should be in
accordance with Section 3. Chemical analysis of the ground
and groundwater should be made to assess its sulphate
content & the concrete mix should be designed in accordance
with the requirements of BRE Special Digest 1 Concrete in
aggressive ground (BRE, 2001). The nominal concrete cover
should be in accordance with Table 3.3. Walls subject to
splashing or intermittent wetting by saline water should have
adequate resistance to chloride attack & may need protection
by a waterproof membrane.
131
131

CIRIA 580 Crackwidth control


BS5400 Part 4 (1990) & BD42/00 (DMRB 2.1.2): Section A11
of BD42/00 (DMRB 2.1.2) provides useful guidance on
circumstances where crackwidth control is necessary to
enhance durability. Cracks in wall sections that are
permanently embedded in undisturbed non-aggressive ground
will have little effect on reinforcement corrosion. In such
conditions, the control of crackwidth in accordance with the
requirements of Clause 5.8.8.2 of BS5400 as implemented by
BD24 (DMRB 1.3.1) should be restricted to sections of the
wall that are not classed as embedded as in Section A11.1.1 of
BD42/00. The cover to the reinforcement over the embedded
section of the wall in these conditions should be in accordance
with Table 13 of BS5400 as implemented by BD24. Where
aggressive or very aggressive environments (as defined in
Section 5 of BD42/00) exist, crackwidths should be controlled
to the requirements of Clause 5.8.8.2 of BS5400.
132
132

CIRIA 580 Crackwidth control


Watertightness
Any size of crack that passes through the section may let in
water However, wall flexure will normally cause a
compression zone (provided it is not combined with
significant tension) that will probably prevent water passage.
BS8007 (1987) provides guidance on how to limit the size of
crack in walls to tanks and reservoirs retaining water.
Aesthetics
If the main concern is the unsightliness of cracks, then a crack
width of 0.2mm (0.3mm) maybe appropriate. Salts & deposits
sometimes leak through smaller cracks and cause unsightly
marks, so a specification of crack width alone is unlikely to be
appropriate in such circumstances.
133
133

CIRIA 580 Crackwidth calculation principles


Crackwidth calculations recommended by BS8110, BS5400 & EC2 should be
interpreted as follows:
BS8ll0 Part l (1997)
For a typical pile section, the concrete cover to earth faces is 75 mm (Clause
3.3.1.4). The durability requirements are for a nominal cover e.g. 35mm for
C35 (Table 3.3) concrete in moderate exposure (Table 3.2) conditions (nonaggressive soil). The crackwidth (BS8110 Part 2; Section 3.8) should be
assessed in accordance with the distance from the point considered to the
surface of the nearest longitudinal bar (Clause 3.8.2). This is customarily
assessed at the nominal cover for durability, rather than at the 75 mm cover.

134
134

CIRIA 580 Crackwidth calculation principles


BS5400 Part 4 (1990): Guidance is provided in Annex A of
BD42/00 (DMRB 2.1.2).
EC2 Part 1 (1992)
To calculate the crack width under bending to EC2, the area of
concrete in tension, is the critical dimension (EC2 Part 1(1992;
Clause 4.4.2.2 (3)). If the concrete in the tension zone beyond
the nominal cover is ignored, the area is:

Pile cross-section showing crackwidth


calculation principles to EC2 as applied
on the Copenhagen Metro project 135
135

136
136

You might also like