8 Soil Movement Due To Deep Excavations PDF

Basement construction and design
8
Engineering an A number of factors control the relative difficulty of basement construction.
excavation Very often these factors cannot be changed by the designer, and include the
location of the building, the proximity of existing buildings and services,
previous site use, and the proposed use of the basement together with soil
and groundwater conditions. The basement structure will be designed to over-
come these constraints to transfer the loads from the superstructure to the
subsoil. The method of basement construction and the type of peripheral
basement wall will be selected to support soils and groundwater at the
curtilage of the basement as economically as possible. The permitted soil
deformation around the basement construction has to be assessed and
complied with. The process was itemized by Lambe1 , as shown in Table 8.1.
Increasingly, clients and architects are demanding larger and deeper base-
ments. This chapter reviews the development of construction methods and
the range of basement walling methods available, and describes the design
problems that arise.
Construction methods Seldom does the location of a basement allow open battered excavations.
for soil support Particularly on urban sites, space is limited and insufficient to accommodate
the cut slopes of battered excavations. Land is expensive and basement con-
structions inevitably occupy as much of the site as possible. The use of open
excavations was reviewed in Chapter 3, although mention will be made here
of the need to review soil strength parameters critically for temporary cut
slopes.
In certain soils, over-consolidated clays such as London clay for example,
the soil strength characteristics are time-dependent, so the period for which
the excavation is to be kept open must be carefully assessed. Where space
allows the use of battered slopes, the cost penalty of a slope failure should
be weighed against the cost of a full soil retention system using temporary
walling. It is possible that a compromise solution, using soil nailing or similar
ground improvement methods incorporating reinforced soil, may be econom-
ically attractive where some horizontal working space is available at the rear
of the basement construction but is not sufficient to accommodate a full
battered slope. Where some space exists behind the permanent basement
wall the choice of method will be determined in permeable soils or granular
soils by the extent of groundwater flow and the feasibility and cost of control-
ling groundwater during basement construction.
An example of a battered basement excavation with a slurry trench cut-off
to control groundwater inflow and cut slopes designed on a cost against risk
basis was given by Wakeling2 . The excavation was 130 m 80 m in plan, to a
depth of 5.8 m in soft clay and gravel, extending into stiff fissured silty London
clay to a maximum depth of 14.5 m. A bentonite slurry cut-off wall into the
London clay contained groundwater in the upper gravels. Groundwater
flow from the gravel and the underlying silty sands was controlled by
Downloaded by [ UNIV OF ILLINOIS] on [13/10/16]. Copyright © ICE Publishing, all rights reserved.
330 Deep excavations
Table 8.1 Engineering an

excavation: a checklist1 Step Activity Considerations
1 Explore and test subsoil

2 Select dimensions of excavation Structure size and grade requirements,
depth to good soil, depth to floor
requirements; stability requirements
3 Survey adjacent structures and Size, type, age, location, condition
utilities
4 Establish permissible movements
5 Select bracing, if needed, and Local experience, cost, time available,
construction scheme depth of wall, type of wall, type and
spacing of braces, dewater excavation
sequence, pre-stress
6 Predict movements caused by
excavation and dewatering
7 Compare predicted with
permissive movements
8 Alter bracing and construction
scheme if needed
9 Instrument – monitor construction
and alter bracing and construction
as needed
gravel-filled counterfort drains dug down the slope during bulk excavation.
The excavation was battered with side slopes of 1:1 with an intermediate
berm at the top of the London clay. A plan and cross-section of the excavation
are shown in Fig. 8.1. The method was successful and demonstrates the use of
soil parameters based on partially drained soil conditions. Wakeling reported
that three slips occurred, all shallow-seated, in the batters within the London
clay, reaching their greatest depth between two and five months after excava-
tion. From back analysis on these slopes and using shallow-seated slides
reported by Skempton and La Rochelle3 , Wakeling reached a tentative
conclusion: for excavations in stiff fissured clays, short-term shallow-seated
slips are likely to occur when the computed failure strength exceeds the
measured undrained shear strength in the clay by approximately 20%. The
point of interest in this example is the cost-effectiveness of a relatively steep
slope batter of 1:1 with some risk of minor failure accepted during a relatively
short construction time. The use of fully drained parameters in the slope
analysis would have led to flatter slopes, albeit with less risk of slippage.
Where compromise solutions to peripheral soil support are required, that is
where some space exists at the rear of the permanent basement structure but is
insufficient to accommodate a battered slope, crib walls and anchored crib
walls can be used. A further solution is the use of soil nailing. These methods
were discussed in general terms in Chapter 4, but the relevance of soil nailing
to basement construction deserves discussion here.
The soil nailing method developed from the use of fully bonded rock bolts
for tunnel support in the 1950s and 1960s. Using the same principles of
ground support its use progressed from weaker rocks, such as marls and
weak sandstones, into cemented sands, strong clays and, later, to a wider
range of granular soils and middle-strength cohesive soils. The range of
soils in which nailing can be used is relatively wide (weak clays and loose
silts are probably precluded, and the presence of groundwater limits its
application in any soil). Although finding widespread application for general
soil support in basement schemes in France, Austria, Germany and North
America, its application in the UK has been slow.
Basement construction and design 331
Fig. 8.1. Plan and cross-section of a battered deep excavation into London clay2
Soil nailing is in fact a soil-reinforcing technique and uses short tendons,

driven or inserted into short bores in the excavated soil face, to improve the
shear strength of subsoils. The exposed face is retained and protected by a
gunite layer reinforced with a wire mesh. The technique is described by
Gässler and Gudehus4 and Banyai5 , but a complete description of its develop-
ment – the soil–tendon interaction, design, construction and specification – is
given in the report of the French Clouterre project6 . The UK code of practice
BS 8006 Strengthened/Reinforced Soils applies to soil nailing and the draft
European code on soil nailing is in course of preparation (Pr EN 14490 at
enquiry stage, July 2003).
Typical cross-sections of soil-nailed excavations were reported by Barley7
and are shown in Fig. 8.2. A soil-nailed excavation support in Pocking,
Bavaria, is shown in Fig. 8.3.
Fig. 8.2. Cross-sections of typical nailed slopes in the UK: (a) temporary soil-nailed slope; (b) soil-nailed slope; (c) rock-bolted
The application of soil nailing to retain excavated slopes at the periphery of

basements is referred to here as temporary support, but soil-nailed slopes can
also be used for permanent works. Tendon durability and protection was
discussed by Barley7 and in the Clouterre report6 . The use of alternative
reinforcing materials such as stainless steel, carbon fibre and glass-reinforced
plastics may lead to an increase in permanently retained soil-nailed excava-
tions. The cost effectiveness of any scheme, the material used and the installa-
tion method is much dependent on job size.
The extent of working space at the rear of the permanent retaining wall
is likely to reduce the nearer the basement site is to a city centre. In the
remainder of this chapter, it is assumed that such space is limited. Soil support
systems which incorporate both temporary and permanent support are likely
to prove most efficient in minimizing the total width of soil support wall.
Alternatively, the construction of sloping sheeting can lead to economies in
construction cost where limited working space is available. Schnabel8
reported that where sheeting sloped at an angle of about 10% from the
Fig. 8.3. Soil-nailed slope at Pocking, Bavaria Fig. 8.4. Plate method being used
(courtesy of Bauer) for a basement extension to the
Technical University, Zurich
(courtesy of Bauer)
vertical, the measured anchor strut loads were consistently less than two-
thirds of the computed anchor loads for vertical sheeting in the same soil. A
further study9 of sloped sheeting supported by ground anchors presented
model tests results in sand, confirming Schnabel’s recommendations for soil
pressures on an inclined anchored wall. It was noted from these model tests
that inclined walls require a considerable base width if they are not to suffer
a bearing capacity failure.
Underpinning in short lengths may prove necessary to avoid settlement of
adjacent structures during basement construction. In dry soil conditions,
where the water table lies beneath basement formation level, it may be
sufficient, and expedient, to rely on the underpinning to provide horizontal
soil support during basement construction in addition to its main purpose
of vertical load transfer to depths below the new basement construction.
Unless the underpinning is braced or propped from the excavation side its
depth will be restricted in either concrete or grouted soil because of horizontal
soil pressure at the rear of the underpinning.
Where ground conditions allow successive excavation in the dry, an
anchored reinforced concrete plate can be used to provide a continuous
reinforcement wall at the periphery of the new basement. These ground con-
ditions may be obtained by grouting in certain soil conditions, given legal
consents. Figure 8.4 shows the plate method being used in a base extension
in Zurich. Each cast in situ element was retained by ground anchors,
excavated alternately at each level. The subsoil was a cohesive silty sandy
gravel.
The provision of lateral support to a deep excavation thus turns on six
factors: neighbours’ rights (reference to the Party Wall Act 1996 in the UK
is made in Chapter 1), neighbouring construction, subsoil and groundwater
conditions, neighbouring services, and the proposed construction depth and
optimization of site area to give the best financial return.
In complying with these factors the majority of urban basements sites will
not allow battered open excavations due to space limitations. Vertical periph-
eral soil support is therefore required, temporarily during construction and as
a permanent retaining wall. During construction the simplest form for either
sheeting or walling is to cantilever without propping. In typical basement
excavations in London the maximum height of cantilever is generally of the
order of 5.5 m from formation level. The extent of soil movement during
and after bulk excavation, and the presence of delicate services or important
highways at the rear of the wall, mitigate against the use of high cantilevers.
Temporary berms at the front of the cantilevered wall reduce soil movements
effectively but are often uneconomical because of the need to remove the berm
successively in short lengths and small volumes. Although propped cantilevers
provide more security against excessive wall movement, the cost penalties of
providing this support and the impedance to bulk excavation may prove
unacceptable. The economical use of peripheral steel sheet pile propped by
steel raker tubes from a completed central raft construction with a temporary
edge berm was shown in Fig. 6.1. A cost comparison on that particular site
showed little difference between propped steel sheet piling with in situ perma-
nent retaining walls and propped diaphragm wall construction.
Figure 8.5 shows why the use of temporary soil berms to reduce cantilever
wall movement is unpopular with contractors. The basement, in West
London, was large in plan area but limited in depth to 5.625 m from existing
ground level. Maximum horizontal wall deformation was specified as 25 mm
and ground conditions were medium dense sands and gravels overlying
London clay. Two schemes were prepared by a specialist diaphragm contrac-
tor, the first with a temporary soil berm and minimum wall depth, the second
Fig. 8.5. Cantilever wall, West

London: (a) wall cross-section;
(b) soil profile; (c) calculated
maximum deformation with and
without soil berm
with a free-standing wall of greater depth. Although the analysis showed a

significant beneficial effect on wall deformation of providing a relatively
small berm, the main contractor preferred the deeper wall without the
berm, shown in Fig. 8.5(c). The cost of the additional walling was significant
but avoided later excavation of small soil volumes and impedance to the base
raft construction programme. During construction the maximum deforma-
tion of the cantilever wall without the berm was 19 mm.
Using finite element methods and assuming linear elastic perfectly plastic
soil material (with the effects of enhanced stiffness at small strains), Potts
et al.10 reached a number of conclusions on the effectiveness of berms: for
berms between 2.5 and 5.0 m high it is the volume of the berm, not its specific
geometry, that dictates soil movements adjacent to the excavation wall defor-
mation and bending movements; as the height of the berm reduces below
2.5 m, berms of equal volume, but varying geometry result in different wall
deflections and moments – deflections increase and the berm becomes less
efficient.
To avoid the obstruction of temporary berms and rakers during construc-
tion, soil anchors may be used for soil support to basements of moderate
depth. This solution depends on the suitability of soil conditions for anchor-
ing and the legal and practical implications of founding anchors outside the
curtilage of the site. The popularity of anchoring in the UK appears to be
less than in France and this may depend on cost which in turn may depend
on design safety factors. In France a value of 1.5 is usual, whereas in the
UK a minimum value of 2.0 has been specified by BS 8001 for temporary
works.
Water-resisting A complete review of the methods of safeguarding basements against water

basement construction and dampness was given in the CIRIA report 13911 in 1995. Later comment
on the durability and water resistance of basements is made in the Institution
of Structural Engineers report on basements and cut-and-cover construction12
in draft form in 2003. The British Standard BS 810213 , published in 1990,
serves as a code of practice on the subject.
The whole matter of waterproofness standards is a matter of potential
controversy and originates from a false belief that a basement construction
in water-bearing ground can be made watertight and completely dry. The
requirements of water resistance, a more realistic term, are described in both
the CIRIA report 139 and BS 8102 in similar classification. A summary
chart of the basic grades of water resistance as defined in both documents is
given in Table 8.2. Four performance grades are specified as follows: grade 1
basic utility, grade 2 better utility, grade 3 habitable and grade 4 special.
The water-resisting methods to be adopted to address these performance
standards are given in CIRIA report 139 as being one (A, B or C) or a
Table 8.2 Basic grades of water resistance11
From Table 1 of BS 8102: 1990 Abbreviated commentary given

by CIRIA Report 39
Grade Basement usage Performance Form of protection
levela
Grade 1 Car parking plant Some seepage Type B with RC design to Visible water and BS 8110 crack
Basic rooms (excluding and damp BS 8110 width may not be acceptable.
utility electrical equipment); patches tolerable May not meet Building
workshops Regulations for workshops.
Beware chemicals in
groundwater
Grade 2 Workshops and plant No water Type A or Type B with Membranes in multiple layers
Better rooms requiring drier penetration but RC design to BS 8007 with well lapped joints. Requires
utility environment; retail moisture vapour no serious defects and higher
storage tolerable grade of supervision. Beware
chemicals in groundwater
Grade 3 Ventilated residential Dry environment Type A or Type B with As Grade 2. In highly permeable
Habitable and working, RC design to BS 8007, plus ground, multi-element systems
including offices, Type C with wall and floor (possibly including active
restaurants, leisure cavities and DPM precautions, and/or permanent
centres and maintainable under-
drainage) probably necessary
Grade 4 Archives and stores Totally dry Type A or Type B with As Grade 3
Special requiring controlled environment RC design to BS 8007 and a
environment vapour-proof membrane,
plus Type C with ventilated
wall cavity and vapour
barrier to inner skin and
floor cavity with DPM
a
See CIRIA Report 13911 for limits on environmental parameters.
combination of two (C þ A or C þ B); the types being as follows: A, structure

requiring the protection of an impervious membrane (i.e. tanked); B, structure
without a membrane but with structurally integral protection; C, drained
cavity (for use with type A or type B or alone).
Overall, water-resisting reinforced concrete should be used for all the grades
together with appropriate design, detailing and construction. Somerville14 ,
reviewing the durability of R.C. structures referred to the importance of the
four ‘Cs’ in terms of durability; constituents of the mix, cover, compaction
and curing. In terms of water resistance it should be recognized that reinforced
concrete inevitably cracks as the cement hydrates, through thermal movement
and later drying shrinkage. Perhaps, in terms of reinforced concrete base-
ments, the risk of leakage and the management of that leakage to ensure
compliance with the required performance standard should be based on
four ‘Ss’: simplicity in structural detailing to take into account the intended
construction methods and buildability; location of construction joints to
receive special attention; services: make design provision at an early stage;
spacing of wall reinforcement and supervision of concreting to ensure well
compacted dense concrete.
It is inevitable that reinforced concrete walls and rafts will crack, initially
due to thermal shrinkage and thence with time due to drying shrinkage.
Early age strains due to thermal action are likely to be far greater and more
important than strain due to drying shrinkage. The hydration temperatures
generated in thick reinforced concrete sections are high and any change to
thin sections is likely to generate differential movement and resulting cracking.
Similarly, restraint provided by cross-walls or similar structural restraint is
likely to cause cracks. These early cracks are important in terms of water resis-
tance because they are likely to pass right through the section. The maximum
crack width for early age thermal effects should be limited to 0.2 mm. Cracks
due to flexure under load do not generally pass through the section, compres-
sive stresses progressively reducing strains in this zone of the structural
section. The maximum flexural crack widths at the surface of the concrete
section may be greater than the maximum of 0.2 mm for thermal cracks and
typically would be of the order of 0.3 mm. The avoidance of steel congestion
in basement R.C. sections to obtain stringent crack control and only secure
poorly compacted concrete with risks of bad durability and water leakage
cannot be over-emphasized.
A recent paper by Boikan15 summarized the pitfalls in waterproofing of
basements and their prevention. As an introduction, Boikan quoted a legal
case in England, Outwing Construction vs. Thomas Weatherald, when the
High Court held that the designer takes responsibility not only for the speci-
fication of a waterproofing product, but must also assess the compatibility of
that product in conjunction with other parts of the basement design, and the
impact of inadequate workmanship and site conditions on the integrity of the
design overall. Boikan recommended that particular matters such as the use of
incompatible discontinuous products, use of products such as low tolerance to
poor workmanship, use of waterproofing products from various sources all
required special attention. He further commented that physical barriers, pre-
viously asphalt tanking but nowadays polyethylene cast in membranes, or
bentonite clay bound into a geotextile required particular attention. These
membranes should be assessed in terms of their performance in hard and
soft water and contaminants, their gas resistance, their means of bonding to
the concrete and the details of bonding one sheet to its neighbour. Boikan
included a review of the performance of hydrophilic waterstop systems
which are now much preferred to PVC waterstops in in situ R.C. basement
construction.
Progressive As basement excavations increase in depth, excavation methods have become

development of more complex, leading to top-downwards techniques which allow simulta-
construction methods neous basement construction and superstructure erection. These techniques
for deep basements became popular in major international city centres in the 1980s and 1990s.
This review of the development of these methods in London is largely based
on a paper by Zinn16 and shows the increasing dependence of basement
construction methods on progressively larger and more powerful piling and
diaphragm wall equipment.
Trench construction
At the beginning of the last century, basement construction in London was
restricted to major buildings. Basement walls were built as gravity structures
in deep, heavily-timbered trenches at the basement periphery. In later years
the walls were built as cantilever reinforced concrete walls in trench. Wall
dimensions that resulted from this construction method can be judged from
Fig. 8.6, a cross-section of the basement wall to the Shell Centre17 . The
wall, cantilevering for a depth of four basement storeys, was built in pre-
stressed concrete within a sheet piled trench with reinforced concrete wallings
and struts. Two separate excavation operations were required, the first by
Fig. 8.6. Cross-section of

peripheral cantilever
prestressed concrete retaining
wall in trench, Shell Centre
basement, London17
grab within the peripheral trench, the second by excavation of the central
dumpling after wall completion. The wall construction operations were
obstructed by the temporary trench bracing but before it could be removed
the inner and outer sheeting to the trench had to be re-strutted against the
completed wall; all frame levels, reinforcement splicing levels, concrete lift
heights and re-strutting levels were interdependent.
Peripheral walls propped by floor and raft

In the late 1950s the trench method began to be replaced by more efficient
basement excavation methods. The changes exploited large-diameter bored
piles which were introduced into the UK at that time together with the
simple innovation of using the horizontal strength of floors and raft sections
as deep beams to span the length of breadth of the excavation. The Fu Centre,
Hong Kong (Fig. 8.7) was built with a pile wall, but the base of the cantilever
reinforced concrete wall was designed to span horizontally and resist all
horizontal earth pressure, allowing the removal of temporary support without
inducing purely cantilever moments into the wall.
In the tower block basement of the Hilton Hotel, London (Fig. 8.8) two
waling beams, each forming part of a structural floor, were used to tempora-
rily support the outer contiguous bored pile wall. The walings were designed
as frames and the upper waling was supported at the bored pile wall and by
bored piles inside the wall.
In the third phase of the development of this method, the Royal Garden
Hotel site (Fig. 8.9) used diagonal struts, also supported on piles, to reduce
the 72 m span of the longest side of the basement. Later, peripheral diaphragm
walls propped by successive floors, designed to span as a horizontal frame,
were used at Gardiner’s Corner, London (Fig. 8.10).
Fig. 8.7. Stages of basement

construction, Fu Centre, Hong
Kong16

construction, Hilton Hotel,
London16

construction, Royal Garden
Hotel, London16
Fig. 8.10. Top-downwards

basement construction with
floors used as horizontal frame,
Gardiners Corner, London
(courtesy of Cementation)
Fig. 8.11. Stages of car park

basement construction,
Leicester Square, London16
Top-downwards construction
The fourth and fifth phases of the development of the basement excavation
method are illustrated by a car park at Leicester Square and the Winter
Gardens Theatre, both in London.
In the car park works (Fig. 8.11) a 458 berm was used to support the outer
bored pile wall while the central area substructure was completed. Floors
acting as walings as the berm was removed were supported by reinforced
concrete columns, cast in advance of the berm excavation, within pre-bored
shafts. Early construction of the central part of the substructure allowed
superstructure works to begin before berm excavation and completion of
the peripheral wall and slabs.
At the theatre site (Fig. 8.12) the disadvantage of removing the berm subsoil
from within the existing work was overcome with temporary steel lattice
columns to support the floor sections used as walings at each level.
The top-downwards method was, therefore, at that time almost in place; to
achieve maximum economy three criteria had to be achieved by the excava-
tion method:
(a) the retaining wall which supports the width of the excavation should
obtain support at each floor level

construction, Winter Gardens
Theatre, London16
Fig. 8.13. Stages of basement construction, House of Commons underground car park, London: (a) site plan; (b) vertical
cross-section showing soil profile; (c) vertical section and plan of car park19
(b) the ground should act as a temporary soffit ‘shutter’ to floor construction
(c) removal of the excavation must be rapid and continuous.
With the development of diaphragm walling during the 1970s it was logical
that bored pile peripheral walls would be replaced with diaphragm walls to
act as both temporary and permanent soil support. Fenoux18 described the
construction of a nine-storey basement for car parking in which the super-
structure was opened for use before the completion of the substructure. In
1972 the House of Commons underground car park in London was built
with peripheral diaphragm walls with temporary support from floor slabs
cast successively with continuing excavation. The basement, shown in section
in Fig. 8.13, reached a maximum depth of 18.5 m with diaphragm wall 30 m
deep. The prime concerns at design stage were to minimize soil movements
due to the bulk excavation and limit the effect on nearby historic buildings.
The risk of soil heave was more acute since there was no superstructure
above the basement. The solution, to build a relatively stiff wall (1 m thick
diaphragm) propped at relatively small centres (storey heights) with relatively
stiff propping from in situ reinforced concrete floors was, therefore, designed
to reduce the risk of settlement of existing structures rather than to reduce
construction time. A detailed description was given by Burland and
Hancock19 ; details of soil movements, which caused only minor cracking
and movement to the adjacent buildings, were given by Burland et al.20
A more recent deep basement construction in London was described by
Marchand21;22 . This basement, constructed by top-downwards techniques to
a depth of 23.9 m from ground level to the lowest basement formation level,
was built for car parking below an eight-storey office block superstructure.
The basement is one of the deepest in London. Two details are worthy of
note: precast concrete stop ends were used in the 1 m thick diaphragm wall
construction and, although generally successful, Marchand commented:
‘Some of the joints between the precast stop ends and in situ concrete
leaked and this was dealt with by grout injection. The sealant used has been
specially developed for the mining industry and is pumped in as a fluid
which changes to a flexible mass of matted rubber particles. This material
can then flex without cracking. In a few places at low level clay had adhered
to the stop end, leaving a strip of clay up to 70 mm wide between adjacent
panels. This was raked out to a depth of 150 mm and made good in order
to provide a waterproof joint.’
Waterproofness of diaphragm wall basements is discussed later in this
chapter. Where basement walls built by pile or diaphragm wall techniques
remain unlined, the longevity of remedial measures to ensure acceptable
waterproofness remains a matter of concern.
The second noteworthy innovation was the use of five rows of pin piles
installed in front of the basement wall to stiffen the London clay and prevent
softening with time. With increasing groundwater levels the construction
would otherwise have required a substantial ground slab to prop the wall
and prevent passive failure of the wall. The stiffened soil approach allowed
the wall toe to be raised 5 m from the original design, a significant reduction
in a very deep wall, originally almost 38 m deep from ground level.
In the 1960s top-downwards construction techniques required the setting of
steel columns as part of the superstructure support within pile heads or pile
caps at basement formation level. This operation required personnel to trim
pile and set precast caps or make bases in situ for the column installation at
final basement level. On this contract, a scheme for setting the steel columns
directly into the wet pile concrete was investigated but rejected because of the
risk of inaccurate placing of the columns. Liners, 21 m long, were used to gain
access to each pile head. The operation was costly and time-consuming and
was frequently underestimated in terms of construction time.
More recently, specialist firms have developed jigs to enable the steel plunge
column to be placed very accurately both in verticality and position within an
unlined box supported by bentonite slurry. The development of this jig there-
fore permits the use of slurry support and allows the steel column to be placed
in the wet concrete of the recently concreted pile. The general accuracy of
placing plunge columns is of the same order as bored pile construction of
the order of 1:75 but with the use of specialist jigs and good site control
placing accuracy in the range 1:200 to 1:400 can be maintained. The rolling
tolerance of the steel column itself may be critical and should be taken into
account in determining likely in situ verticality.
The top-downwards method has obvious advantages in terms of soil
movement and completion time, but important disadvantages include the
additional cost of excavation and removal of soil from beneath floor slabs
in cramped conditions compared with conventional open excavation
methods. Also, there is the congestion caused on site by superstructure and
substructure contractors working within the same programme period.
A successful application of top-downwards construction was recently made
in London on a congested site for the redevelopment of a site in Knightsbridge
for use by Harrods department store. The new building consisted of seven
storeys above ground and a seven-storey 25 m deep basement. The works
are described by Slade and Darling et al.23 , and the control of ground
movements and the application of compensation grouting by Fernie24 and
Kenwright et al.25
The new structure, built at the rear of the existing Harrods building and
connected to it by a new access tunnel, incorporated an existing façade on
one elevation. A site plan is shown in Fig. 8.14 and the construction sequence
is shown in Fig. 8.15.
The ground conditions on the site are made ground of 3.5 m thickness
overlying Terrace Gravel 6.0 m thick which in turn overlie London clay of
proven thickness greater than 50 m. The groundwater conditions comprise
Fig. 8.14. Harrods,

Knightsbridge, London.
Site plan23
Fig. 8.15. Harrods,

Knightsbridge, London.
Construction sequence23
an upper perched water table in the Terrace gravels and a lower aquifer in the
Thanet sands and Chalk which underlie the London clay. A hydrostatic pore
water pressure distribution was shown by piezometers within the London
clay.
The basement construction for parking, plant rooms and workshop use was
required to comply with grade 1 of BS 8102 allowing some seepage and damp
patches. Due to the demands for basement space the use of a lining wall and
drained cavity was not viable and the 800 mm thick diaphragm wall remained
unlined. Top-down construction with the ground floor initially built and then
excavation in two-storey level increments was possible and shown to be so by
finite element analysis. This procedure saved construction time and was
monitored during construction by comprehensive instrumentation, collecting
data from precise level points, electrolevel beams, in-place inclinometers,
water levels, survey targets and base traverse stations. Readings were taken
on a 24-hour basis and any change above program supervision trigger levels
was sent to the Engineers’ terminals. In the event, the trigger levels were not
exceeded.
The initial finite element analysis predicted both the deflected shape of
the basement walls and the resulting settlement profile outside the basement
perimeter. In turn, using these deformation contours, the settlement behaviour

of nearby existing buildings was assessed and the classification due to
Boscardin and Cording26 used to make a risk prediction. Maximum settle-
ments due to basement construction were limited to 10 mm and angular
distortions to 1 in 750.
The overall stability of the basement in the long term was checked to ensure
a minimum factor of safety of 1.4 for the whole basement, taking into account
the uplift force due to soil heave and the groundwater pressure resisted by the
combination of building weight, the tensile resistance of the piles and the skin
friction on the faces of the diaphragm walls. (Skin friction to the lowest base-
ment level was used.) The total assessed long-term uplift pressure over the
whole basement area was 340 kN/m2 .
In the USA, the top-downwards method has been adopted more recently.
American practice is described by Fletcher et al.27 The excavation of a large
four-storey deep basement for the Milwaukee Centre, close to historic
structures and within 3.5 m of the Milwaukee River, demanded a cut-off and
control of groundwater, minimum soil movement and early completion. A
major bracing or raker system was judged to be too cumbersome and
costly, and a temporary freeze wall system was dismissed because a permanent
ground water cut-off would have been necessary. The top-downwards method
with a deep diaphragm wall as a cut-off was adopted and proved successful.
Scope for innovation remains. In 1993, a contract for an opera house in
Paris used the technique for a 28 m deep basement with anchored support
in lieu of lateral support from basement floors as excavation proceeded.
Large barrette sections were constructed for superstructure support. A typical
cross-section of the basement is shown in Fig. 8.16. Due to the planned con-
struction, after completion of the basement, of a Metro running tunnel on one
side of the basement and in close proximity to it, anchor tendons constructed
from glass fibre were adopted to avoid obstruction to the Metro tunnelling
machine. The brittle, low shear strength of the glass tendons ensured that
they would be easily removed by the tunnelling machine.
In Hong Kong, land values have increased demand for larger, deeper base-
ments frequently in unfavourable soil conditions with stringent settlement
criteria applied to the basement peripheral subsoils. This market demand
has brought about almost an exclusive use of top-downwards construction
generally using diaphragm walls. A recent example of the construction of
the deep basement for the Dragon Centre in the heart of the Western Kow-
loon Peninsula was given by Lui and Yau28 . The site was located on old
reclaimed land and the development, for a new retail building, comprised a
nine-storey reinforced concrete structure over a five-storey basement for
parking use. Historical records showed that the foreshore originally fronted
the site. The existing MTR tunnels are within 100 m of the site; the site
reclamation was made in 1924. Adjacent existing buildings are either sup-
ported on pad footings or driven precast piles. The ground conditions are
loose granular fill underlain by marine deposits, generally loose silty clayey
fine sand and highly developed granite (dense silty fine to coarse sand) and
granite bedrock. Groundwater level is 1.5 m below ground level. A geological
section is shown in Fig. 8.17.
The basement structure, 107 m 67 m in plan was formed from a dia-
phragm wall box, 1200 mm thick up to 40 m deep installed by cutter with
CWS joints to 30 m depth and a cut joint below 30 m. The top-downwards
method was used with support of basement floors by steel box stanchions
filled with sand–cement grout. Each internal column was supported on a
single large-diameter bored pile founded on the bedrock with an allowable
bearing pressure of 5 MPa at depths between 45 and 65 m below ground level.
Fig. 8.16. Paris basement construction sequence, Provence Opera, Paris: (a) barrette sections installed; (b) ground floor
construction; (c) superstructure construction and excavation below ground floor slab; (d) superstructure and substructure
The basement construction proceeded as the basement floors were cast on

grade successively downwards from the ground floor slab. The prime design
requirement was to minimize ground movements outside the site. In order
to do this, the cut-off effectiveness of the diaphragm wall was improved by
grout injection a further 10 m below the wall. A full-scale pumping test was
Fig. 8.16. (e) Load transfer of temporary support loads to floors; (f) geological conditions; (g) site plan; (h) anchor head detail
for composite soil anchor (courtesy of Soletanche–Bachy)
made and then back analysed. The results were used as design parameters for
prediction of ground movements due to full-scale excavation. A further
multiple-well pumping test was made after the completion of the diaphragm
wall box to simulate the construction dewatering. An array of pumping
wells at about 25 m centres was used to lower the water table over the site
Fig. 8.17. Dragon Centre, Hong Kong. Geological section.
area to the lowest basement level which is about 27 m below ground level. The
test was maintained for eleven days until steady-state flow was obtained and
then the pumps were switched off and the groundwater allowed to recover to
its initial level. The results showed the following.
(a) The total steady rate of flow from the wells was 25 m3 per hour with
measured hydrostatic water pressure on both sides of the wall at
steady state.
(b) Readings from piezometers within 5 m distance from the wells showed
that the groundwater was lowered to 23.5–29.5 m depth.
(c) With the exception of the eastern corner of the site, the drawdown
outside the site was an average 0.5 m at standpipes and 1.5 m at
piezometers.
(d ) Ground surface settlements were recorded in the range 4 to 16 mm with
negligible settlement at adjacent buildings.
(e) The maximum lateral deflections recorded by three inclinometers were
in the range 50 to 80 mm at the top of the diaphragm walls, with general
wall rotation from the toe.
Following the cessation of pumping, there was negligible recovery to the
ground surface settlements and the lateral wall deflections recovered to
about half of the maximum wall deflections during the pumping test.
Back analysis using the finite element computer program SEEP was made
to vary the rock mass permeability so that calculated pore-water pressures
matched those at steady state in the pumping test. The matched calculated
value was 1 108 m/s.
Further back analyses were made with the computer program FREW,
which assumes a linear elastic continuum between active and passive limits
on both sides of the wall. With some difficulty the observed wall displacements
Fig. 8.18. Brittanic House, London. Construction sequence30
were used to estimate a value of K0 (0.3 to 0.4) and a relationship between soil
modulus EI and SPT value N.
Using these back analysed values, the wall deflections due to the subsequent
excavation were analysed and the ground settlements estimated.
Overall, the pumping test and the seepage analysis made thereafter high-
lighted the importance of achieving a good seal by grouting between the toe
of the diaphragm wall and the intact rock. Comparative analyses could be
made between a grouted and ungrouted box. The predicted wall deflections
reasonably agreed with observed ground movements after basement excava-
tion. In general the maximum ground surface settlement as predicted was
approximately 30 to 50% of the maximum wall deflection caused by the
pumping test and basement excavation respectively. The vital importance of
carrying out pumping tests for deep basements in water-bearing soils was
established on this job and continues as standard practice in Hong Kong.
Recharge wells were installed at the Dragon Centre but were not used
extensively following observations of settlements outside the basement as
excavation progressed.
Semi top-down construction

The use of very large openings in floor slabs that are designed as a frame to
provide lateral support to the external walls together with the use of excava-
tors with long dipper arms enables excavation to proceed below the top slab
support for considerable depths without intermediate floor support. This
method, with only a skeletal structure for the working platform and one inter-
mediate floor was used for a 25 m deep excavation for station construction on
the Singapore MRT29 . Following the construction of the roof, the skeletal
plan shape of the roof was used to excavate down to concourse level using
backhoes and long arm excavation. The concourse slab was then built with
similar large openings to allow further deeper excavation whilst propping
the external walls.
Combination of bottom-up and top-down methods

Basements of large plan area can benefit from a combination of both bottom-
up and top-down methods. This was used at a very early stage of top-
downwards construction in the UK in 1962–1963 at the site of an 18 m deep
excavation at Brittanic House in central London. The work, reported by
Cole and Burland30 , involved an external diaphragm wall box supported
temporarily by an earth berm during construction of the central raft followed
by bottom-upwards construction of the central core and tower columns. The
outer floors were then constructed top-downwards below an upper strutting
floor. The sequence is shown in Fig. 8.18 taken from Cole and Burland’s
paper. A similar combination of bottom-upwards and top-downwards
construction was used for the basement of the Main Tower in Frankfurt31 .
A piled raft was used to support the 198 m high tower on Frankfurt clay,
an over-consolidated but rather weak soil. The core of this tower was built
bottom-upwards in a conventional four-frame supported excavation in
advance of construction, top-downwards of the floors around the core
within a peripheral secant pile wall. This combination, shown in sequence
in Fig. 8.19, altered superstructure core construction to advance in parallel
with basement construction, to the advantage of the overall programme.
Peripheral sheeting or The system adopted to sheet or wall the periphery of the excavation will be
walling influenced by the choice of basement construction method, the suitability of
ground and groundwater conditions, the need to build close to site boundaries
Fig. 8.19. Main Tower, Frankfurt. Construction sequence31
and minimize wall thicknesses and, not least, by the local availability of
materials and specialist plant and equipment. Peripheral sheeting methods
are generally the following:
. anchored underpinning: reinforced concrete plates and grouted soil
. king post wall: vertical soldiers and horizontal laggings or reinforced
concrete skin wall
. sheet piling
. contiguous bored piling
. secant piling
. soldier pile tremie concrete method (SPTC) – as used previously in the
USA.
. diaphragm walls
k reinforced concrete cast in situ
k precast reinforced concrete
k post-tensioned.
The general features of each method were covered in Chapter 4, but their
particular application to basement works is reviewed below.
Anchored underpinning
Where the total excavation depth of basement work is typically in the range
8 to 12 m and ground conditions are dry and capable of supporting a face
1.5 to 2 m deep and of similar length, the anchored plate method provides
an economical temporary wall support if permission to install anchors outside
the curtilage of the site is forthcoming. In conditions where soils lack the
strength to stand unsupported to these modest depths, pre-grouting may
prove worthwhile in granular soils. Where foundation loads from adjacent

structures are such to necessitate transfer of load below the proposed excava-
tion depth, pre-injection of the subsoil below the existing foundations, with
anchorage to avoid lateral movement of the grouted soil mass, may prove
an economical alternative to conventional mass concrete underpinning.
King post wall

The king post or soldier pile and horizontal timbered wall, previously widely
used in North America, has become increasingly popular for basement con-
struction in Europe in recent years. For use in shallow excavations, the king
posts may be cantilevered or propped by raking shores or anchored in succes-
sive layers as bulk excavation proceeds in deeper basement works. The king
posts may be double joist or channel units battened together to allow the
anchor to conveniently pass between. Figure 8.20 shows soldier pile walls
supported by anchors constructed in basement works in Saudi Arabia to
depths exceeding 20 m. Subsoil conditions were layered washdown silts and
silty sands, with a groundwater table at formation level or a small height above.
The method requires moderately dry ground conditions with soil of suffi-
cient strength to maintain a vertical face prior to support from the horizontal
lagging being placed. King post centres vary from 1.5 to 3.5 m, depending on
soil strength, depth of excavation and surcharge loads. A popular innovation
is the use of in situ reinforced concrete skin walls cast against the exposed
soil face, with thicknesses between 150 and 200 mm. The walls, which span
horizontally between king posts, are cast in lifts between 1 and 1.5 m high,
depending on the ability of the soil to stand without support.
The king post excavation may be bored by auger rig or, where headroom is
limited, low-headroom rigs may be necessary. The toe of the king post is
usually concreted to basement formation level, although deeper king posts
surrounded by sandy gravel washed in may be preferred if the king posts
are to be subsequently extracted. It is economical to use the lined face of
the timber laggings or the face of the skin wall as a back shutter to the perma-
nent basement wall, but allowance must be made for tolerances in the king
post wall construction.
Due to the width of the king post wall and the permanent wall construction
it may be necessary to drill the king post bores close to the site boundary.
Where an existing structure is close to this boundary the minimum distance
between king post bore and site boundary will be determined by the minimum
Fig. 8.20. Anchored soldier pile walls used in dry layered sand and clayey silt soil for deep basement construction.
Note the unimpaired access for plant and site operations, Medinah, Saudi Arabia (courtesy of NCF)
Table 8.3 Minimum distances between soil support system and site boundary for various types of installation plant
Support system Installation plant Distancea
Underpinning Conventional bulk excavation plant, e.g. Nil

hydraulic excavator with hydraulic grab
Steel sheet piling Crane and piling hammer/tracked hydraulic 500 mm, rear of sheeters
piling machine to face of boundary wall
Contiguous bored pile wall Bored pile:
tripod equipment typical 600 diameter pile 150 mm
Large-diameter rig:
Hughes CEZ 300 typical 740 pile 385 mm
Hughes CEZ 450 typical 750 pile 385 mm
Hughes KCA 100/130 typical 900 pile 450 mm
Contiguous bored pile wall CFA rig:
and hard–soft secant wall Soilmec CM 45 typical 750 pile 350 mm
Soilmec CM 48E typical 750 pile 350 mm
Rotary rig CFA:
Bauer BG 11 typical 500 pile 400 mm
Hard–hard secant wall Bauer BG 7 FOW method Nil
254, 273, 305, 343, and 406 mm diameter
Diaphragm wall Rope suspended grab 200 mm
City Cutter 150 mm
Hydrofraise 150 mm
Berlin walls: soldier piles Rotary piling equipment 600 diameter bore 400 mm
and horizontal lagging
Manual excavation: 200 mm up to 6 m depth
hydraulic excavator and trench box
a
Minimum distance between outer face of support system and site boundary. Distances quoted are those at ground level;
consideration must be given to verticality tolerance of support system.
Fig. 8.21. Vertical cross-section

through part of a substructure
basement, Marylebone,
London32
overhang of the auger rig from the rear face of the pile bore. Table 8.3 shows
the minimum dimension between an existing structure and the outer face of
the new wall for various wall types.
It is usual for king post wall construction to be used only as temporary soil
support. An exception was described by Mair32 . In Marylebone, London, a
king post wall was used as the permanent peripheral retaining wall in a top-
downwards type two-storey basement construction. A cross-section of the
construction is shown in Fig. 8.21. King post wall construction was feasible
because the whole depth of the basement, to 8 m, was accommodated
within dry sands and gravels 11.5 m deep, below which was London clay.
The water table in the sands and gravels was below the basement formation
level, at a depth of 9.5 m. The king post centres were 1.5 to 1.8 m, and mass
concrete infill was placed in 1 m lifts as excavation proceeded. As usual in
top-downwards construction, the king post wall was successively propped
by ground, lower ground and basement floors.
Sheet piling
The use of sheet piling for temporary soil support to basement construction in
urban areas has declined as environmental controls on noise and vibration
progressively strengthen. Only where sheet piles can be installed by hydraulic
means, particularly in cohesive soils, can the effects of these controls on noise
and vibration be avoided. The use of Giken hydraulic press equipment has
steadily increased in the UK in a wider range of soils. Typically used in city
centres, penetration through stiff clays such as London clay and Gault clay
is restricted to a pile length of 16 to 20 m. Water jetting or lubrication of
the sheeter surface with water may be necessary to achieve penetration.
Where noise and vibration are critical, as in most city centres, sheet piles
for deep basements can be installed by the combined use of slurry trench
and sheet piling methods. The sheet piles are pitched into slurry trenches
filled with cementitious self-hardening slurry and the toes of the piles con-
creted in by tremie pipe up to basement formation level. The technique,
although uneconomic at first sight, is environmentally friendly. The sheet
pile section can be selected on the basis of flexural stress without consideration
of driving stresses, and considerable accuracy can be achieved in pitching the
sheeters into the slurry trench. The sheeters can obtain support from ground
anchors with conventional steel walings or from bracing or raking shores.
Interlocks in sheet piles cannot be assumed to be completely watertight in
water-bearing ground unless provision is made for sealing them. Sheet piles
sealed by welding at the exposed surface after excavation or prior to installa-
tion by a steel contaminant section to allow a bituminous or polymer material
seal of the interlock can remedy lack of watertightness. The use of welding
may be necessary to augment the sealant compound. This system, now
often the subject of contractual guarantees of maximum wall permeability,
is becoming more popular in the UK and is further promoting the use of
sheet piles for permanent works. Figure 8.22 shows typical joint details
between permanent sheet piles and the basement slab.
Contiguous bored piling

The cheapest type of concrete piled wall is the contiguous piled wall. The use
of modern continuous flight auger (CFA) rigs allows high output in a wide
variety of soils. The depth of pile is limited to the order of 18 to 20 m by
the difficulty of inserting reinforcement cages to greater depths through wet
concrete and the lack of water resistance in water bearing ground due to
the gaps between piles. A structural facing may be applied as sprayed concrete
or an in situ reinforced concrete lining.
Fig. 8.22. Typical joint details

between permanent sheet piles
and the basement slab
Prior to the advent of CFA rigs in these jobs only a short length of top
casing was necessary, separating the piles by approximately 50 mm. The
piled wall depth was limited to some extent by the verticality tolerance that
could be obtained by the augers, typically 1% with depth. In the UK many
basement walls were constructed in this way in the 1960s to the 1980s. The
walls were anchored temporarily using steel walings or were braced with
strutting or rakers. Grouting was used in permeable soils where groundwater
entered in the gap between piles. In some instances the intended use of the
basement allowed the bored pile wall constructed in this way to remain
unlined, while for high-grade basements the piled walls were lined with
reinforced concrete or an independent, non-load-bearing blockwork wall.
The advent of CFA piling rigs in the early 1980s, with their ability to
operate without casings (even without a top casing), their high output and,
for smaller low-torque machines, their ease of transport and erection on
site, produced economies which allowed them to replace conventional
augers in most soil conditions. CFA piles for wall construction are typically
300, 450, 600, 900, 1000 and 1200 mm in diameter.
Hydraulic auger cleaners, introduced to prevent soil falling on personnel
also avoid contamination of new concrete with soil. Other innovations include
a projecting tremie pipe from the base of the auger to pump concrete to a
lower level than the core of soil progressively lifted by the auger. Standards
of quality control for CFA piles are now much improved by in-cab monitor-
ing. Data referring to auger depth, torque applied during pile excavation, rate
of withdrawal and concrete pressure are relayed to the rig operator and a more
recent IT development enables this data to be transmitted to an off-site
terminal such as the contractor’s or engineer’s office. Further progress is
needed to measure concrete pressure near the point of discharge.
CFA rigs operate in a wide range of soil and soft rock conditions, but hard
rocks, rock chalk and strong mudstones cause obstruction and make the rigs
uneconomic. Minimum distances for rig operation from existing wall bound-
aries are shown for a range of CFA and rotary rigs in Table 8.3. Some stated
dimensions may be reduced by modifying the standard equipment. Bauer,
in particular, has introduced a purpose-made rig to operate with reduced
minimum distance.
Secant piling
Improvements in rotary rig and equipment design have, as with contiguous
pile walls, changed construction methods for secant piles in recent years.
Until the 1980s the Benoto rig was the primary method of installing secant
pile walls, cutting the concrete of female piles to interlock male piles between
them. A heavy-duty hammer grab was used with twin-wall lockable tempor-
ary casing equipped with a cutting edge for excavation, the casing being
Fig. 8.23. Configuration of

reinforcement in contiguous and
secant pile walls62
oscillated under crowd to achieve penetration. The introduction of powerful

rotary machines equipped with CFAs or casing oscillators has enabled
much higher output rates than the traditional Benoto rig, which had been
used for almost 50 years in Europe and the UK.
Benoto rigs used bores of 880, 1080 and 1180 mm diameter. At present,
secant piles installed by CFAs have diameters in the range 450 to 750 mm
diameter, whereas cased secant piles typically range from 750 to 1180 mm
diameter. Reinforcement, usually limited to male piles, may be from cages
or, where required for shear or flexural strength, joist sections may be used.
A comparison between contiguous pile and secant pile configuration and re-
inforcement is shown in Fig. 8.23. The method provides a near-waterproof
wall for both temporary and permanent soil support and, with CFA, rotary
or hammer grab excavation, secant walls can penetrate most soils and rock
obstructions to maximum depths of 30 to 40 m. The secant pile basement
wall is capable of supporting both lateral load from soil and groundwater
and vertical load from the curtilage of the superstructure. Vertical loads
may be transmitted to the piles by a reinforced concrete capping beam or,
for lesser loads, by shear on the vertical contact face between adjacent piles.
Since 1985, the use in the UK of hard–soft and hard–firm secant piles, for
depths to 20 m by CFA rigs and for greater depths with high-torque rigs with
casing oscillators, has improved production, reduced cost and to some extent
reduced the use of diaphragm walling. The unreinforced female piles, cast in
the bentonite–cement or bentonite–cement–Pulverized Fuel Ash mix, are cut
by the male piles which are reinforced and concreted in the normal way. This
type of hard–soft secant wall may not be satisfactory for walls requiring high
standards of waterproofness or long-term durability, although internal
reinforced concrete lining walls may remedy this situation. Pile rigs used to
install secant piles before and after the 1980s are shown in Fig. 8.24.
Other innovations in recent years include mix-in-place piles. This process
utilizes cement slurry, pumped through the hollow stem of the auger, mixed
with sandy soils during boring and extraction. The unreinforced female
piles constructed in this way vary with the sand–cement ratio (with compres-
sive strength of the order of 15 N/mm2 ) and can be alternatively spaced with
male piles constructed with concrete and reinforced in the normal way.
Fig. 8.24. Piling rigs used for

secant pile installation: (a) BG
26 rig (courtesy of Bauer); (b)
Benoto rig with hammer grab
(courtesy of Lilley)
Fig. 8.25. CDSM piles:

suggested pile spacing35
The use of cement deep soil mixing (CDSM) to produce blocks of over-
lapping piles for nailed or self-supporting gravity-retaining structures has
recently progressed rapidly in the USA after early development in Japan.
Yang33 summarized the use of CDSM for cut-off walls and excavation
support in addition to its use in ground stabilization. A later review of deep
mixing technology was given by Porbaha et al. in 2001.34 The method, devel-
oped in the 1970s from soil–lime mixing methods uses a triple auger machine
to produce pile sections in the range 550 to 990 mm in stiffer soils. Steel H
sections are installed as flexural reinforcement in retaining walls prior to the
hardening of the soil–cement mixture. The soil–cement is designed to arch
between adjacent steel H sections. Taki and Yang35 suggested a spacing of
H sections based on empiricism, as shown in Fig. 8.25. Cement deep soil
mixed piles reinforced with steel joists for soil support and groundwater
control are shown in Fig. 8.26. A soil–cement wall for both groundwater
control and for soil retention in highly permeable coralline conditions was
used for a two-storey basement at the Marin Tower project in Hawaii. The
excavation was only 30 m from the harbour and a high groundwater level.
The partial cut-off scheme was achieved by a soil–cement wall of average
depth 14 m, using 55 cm soil mix piles installed by triple auger. The coral
limestone was ground down to gravel size by the augers without pre-drilling
for thorough mixing with the cement grout. Mix designs with cement dosages
of 300 to 500 kg/m3 of in situ soil were used.
The relatively slow take up of soil mixing processes for excavation support
in Continental Europe and the UK may be associated with the lack of
adequate QA methods. Bruce et al.36 reported current methods in 2000.
Support for piled basement walls by ground and rock anchors

Secant pile walls may be supported by strutting and walings, rakers with
walings or ground anchors. Where used, anchors may be taken axially
through the piles or through the contact face between the piles. It may be
sufficient to use anchors at alternate male piles or every fourth pile, depending
on the extent of lateral load, anchor capacity, and the available shear resis-
tance mobilized on the contact face between adjacent piles.
The decision to use ground anchors as a temporary wall support for any
wall system will be based on practicality, cost and installation time, which
are all influenced by:
. depth of the basement
. groundwater conditions during anchor installation and, thereafter,
during basement construction
. subsoil conditions and their suitability for accommodating anchors of
adequate capacity economically
. maximum permissible soil and basement wall movements and the plan
shape of the basement; the susceptibility of adjacent existing structures
to soil movement caused by the basement excavation
Fig. 8.26. Soil–cement wall for

excavation support and
groundwater control, Tokyo,
Japan (courtesy of Raito Inc.)
. the basement construction programme

. the aggressiveness, if any, of groundwater
. the location of existing services
. the location of neighbouring substructures and/or basements
. legal permissions to accommodate anchorages outside the curtilage of
the construction site
. the risk of obstruction of future works within the construction site by
the presence of anchors.
This list, in no order of priority, may not be exhaustive on any particular site,
but indicates those items requiring earliest consideration. Some items are self-
explanatory. Subsoil conditions will indicate likely anchor capacities, compact
granular soils generally being preferred to cohesive soils in the fixed zone of
each tier of anchors. Subsoil and groundwater conditions will dictate drilling
costs for anchor installation, and the aggressiveness of groundwater and the
period of use of the anchors will dictate the need for corrosion protection.
The location of adjacent substructures and services will determine the

practicality of installing anchors at the required elevations, and the plan
shape of the basement may determine any difficulties caused by obstructing
the drilling of anchors from an adjacent re-entrant basement wall. Above
all, legal permissions and licences must be available from owners of adjacent
land or highway authorities to allow anchor installation outside the site area.
Where anchors are likely to obstruct future construction, a removable-type
anchor may be necessary.
Wall movement is likely to be reduced by the use of anchors that are
stressed after installation, particularly when fixed-length anchors are founded
in competent medium-dense or dense granular soil. When the anchors are
founded in stiff cohesive soils only short-term benefit may be gained.
The programme implications of anchor installation also require exam-
ination and depend on the timing of bulk excavation following anchor
installation. The sequence of drilling, tendon installation, grouting, grout
strengthening and stressing for each bank of anchors has to be phased
within the overall excavation programme. Comparison with an overall con-
struction programme using alternative forms of wall support may be neces-
sary, taking into account the improved construction outputs obtained by
unobtained work areas achieved by anchoring.
The design and construction of ground anchors, described in Chapter 7, is
explained in detail in BS 808137 , which contains an extensive bibliography on
ground and rock anchors. Littlejohn and Bruce38 reviewed the state-of-the-art in
rock anchoring and Barley39 updated this, in particular giving observed bond
stress values of both straight shafted and under-reamed anchors in chalk,
mudstones, siltstones, shales, marls and sandstones.
Soldier pile tremie concrete method

This North American practice, popular in the 1960s, of modifying the king
post wall method by excavating a panel by grab under bentonite slurry
between the king posts and filling the panel excavation with unreinforced
concrete by tremie, has rarely been used in Europe. In Germany, however,
a modification of this method, using mesh-reinforced gunite sprayed between
and over the king posts successively as built excavation proceeds, has gained
acceptance. Although the construction provides only temporary soil support
it is particularly economical in dry granular soils which can be excavated to a
vertical face for 2 to 3 m without collapse in shallow to medium-depth
basements. The overall thicknesses of temporary and permanent walls and
working space at the site boundary are usually not excessive.
Structural diaphragm walls

It may be argued that the most significant advance in recent years in basement
construction has been the introduction of the structural diaphragm wall. The
principal advantages of this form of construction, introduced into Europe
by Icos in the 1950s and 1960s, are: the dual use of the wall to provide
both temporary and permanent soil support; the efficiency in bending of the
rectangular wall section compared with the circular pile cross-section
used previously; the reduction in noise and vibration during installation
compared with percussive drilling of sheet piling; the ease of installation of
propping, strutting and anchoring against the wall face; the ease of applying
finishes to the flat wall face; the ability of the walls to transfer vertical loads;
and its use to depths generally in excess of other forms of wall construction.
The dual support provided by the diaphragm wall at construction stage and
then during the basement life was often sufficiently economical to justify it on
financial savings alone compared with other walling methods. Advantages
such as the minimum thickness of construction required by the diaphragm

wall for both temporary and permanent soil support were a bonus.
The principal disadvantages of diaphragm walling are the risk of loss or
spillage of bentonite slurry, the relatively high cost of cleaning and the dis-
posal of the slurry, the site space needed for large reinforcement cages and
the large cranes needed to handle them. Above all, the need for continuity
in the construction process from excavation through concreting to removal
of temporary stop end formers, is a disadvantage of the method. Structural
diaphragm walls still remain the preferred method of walling for deep base-
ments and concrete piled walls, particularly secant walls, have only tended
to replace diaphragm walling in basement works of medium depth.
Icos40 gave details of a wide range of basement constructions. These base-
ments were generally of moderate depth, perhaps two or three basement
storeys. The diaphragm walls were all excavated by cable grab mounted on
tripods on rails. The wall depth attainable by this equipment was consider-
able, however – up to 28 m in one example in Paris. The panels used were
straight, L- and T-shaped and corrugated in plan. Figure 8.27 shows the
diaphragm wall options for basement excavations recommended by Icos.
As patent protection on the Icos wall waned, specialist firms in Europe
introduced alternative excavation equipment, although in later years Icos
persevered with rope grabs mounted on heavy tracked cranes. In Europe,
kelly bar mounted hydraulic grabs became popular in the 1970s and the
early 1980s. These grabs, which were capable of excavating wall widths
between 500 and 1500 mm, were mounted on single or telescopic kelly bars
to maximum depths of approximately 25 m. Panels were dug in a series of
grab ‘bites’ which were typically each 2.8 m long.
Excavation in medium-strong to strong rocks for diaphragm wall works
was difficult for all specialist firms at this time. The usual method from the
1950s to the early 1980s was to use a drop chisel progressively along the
Fig. 8.27. Alternative methods

of wall support and plan forms
of diaphragm walls for
basement construction
(courtesy of Icos)
(a)
Fig. 8.28. Development of diaphragm wall excavation rigs 1950s–1960s: (a) (i) Icos tripod rig; (ii) action of Else bucket scraper;
(iii) excavation with bucket scraper; (iv) hydraulic grab, Kelly mounted; (v) rock chisel
panel under the slurry and alternatively grab the arisings and chisel again.
Progress was slow and vibration often quite severe in strong rock. In excep-
tional cases rotary core barrettes were drilled successively into the rock
along each panel under the slurry, and both Icos and Soletanche developed
rigs incorporating percussive chiselling with direct or reverse slurry circulation
to remove cuttings.
The introduction of a rail-mounted reverse circulation rig for excavation in
soil and soft rock by the Tone Boring Company of Japan in the late 1970s was
followed by the development of reverse circulation equipment known as the
Hydrofraise by Soletanche. In the mid 1980s Bauer and Casagrande
produced similar equipment. The Bauer Trenchcutter has been developed
into smaller, more manoeuvrable rigs known as City Cutters and more
recently, Mini Cutters. In turn, the Hydrofraise has been made more compact
for city sites in joint development between Soletanche and Rodio to produce
the HL 4000 track-mounted rig. Diaphragm wall rigs developed between the
1950s and 1990s are illustrated in Fig. 8.28. The improved manoeuvrability of
the Bauer City Cutter rig is shown in Fig. 8.29.
(b)
Fig. 8.28. (b) (i) and (ii) Crane-mounted grabs; mechanism of submersible motor drill; (iii), (iv) and (v) Tone Long Wall Drill;
(vi) Tone Long Wall Drill: vertical section through cutter
Fig. 8.28. (c) (i) Tone

Electro-Mill Drill; (ii) Bauer BC
30 Trench Cutter rig;
(iii) Bauer MBC 30 Trench
Cutter rig
Fig. 8.29. Bauer City Cutter rig:

use of alignment device on
small base machine to improve
manoeuvrability in limited
space
Excavation for diaphragm walls is currently made by a variety of rigs

depending on job size, wall depth and soil and rock properties. The reverse
circulation machines have the added qualities of silent and vibration-free
excavation through a wide range of soils and rocks but, even so, conventional
crane-mounted rope grabs are frequently used for walls of moderate depth on
small to medium sized jobs. Heavy mechanical grabs up to 9 tonnes in weight
are favoured for smaller, shallower basements particularly in cohesive soils.
The recent re-introduction of hydraulic grabs by Bauer and others, rope
mounted, has accompanied the use of electronic monitoring to improve
installation tolerances. The usefulness of rams to allow steerage of these
rope suspended hydraulic grabs is less certain.
Cutter reverse circulation rigs saw two significant developments in the early
1990s: the use of rock roller bits on the vertical cutter wheels to allow more
efficient excavation in moderately strong and strong rocks, and the construc-
tion of a compact, low-headroom rig known as the MBC 30 Trench Cutter.
This rig (Fig. 8.30) has its own carrier system mounted on a railway bogie
or on crawlers, so a conventional crane is not needed to carry the cutter.
The overall dimensions of the complete unit are reduced to 4.7 m long,
4.1 m wide and 5.0 m high (6.0 m high when crawler mounted). Trench
widths vary from 640 to 1500 mm; the standard trench length is 2790 mm;
and the maximum cutting depth is 55 m.
Excavation in rock for diaphragm walls still remains arduous and expen-
sive. Mention should be made of a practice, typically in Hong Kong, of
Fig. 8.30. Bauer MBC 30 Trench

Cutter rig
underpinning deep diaphragm walls where formation level is below rock head,
using shear piles drilled through the diaphragm wall, often bundles of T50
rebar, to support the wall in the temporary condition. Figure 8.31 shows a
cross section.
The cutters of a modern Soletanche rig, in this instance for a 1500 mm wide
wall, are shown in Fig. 8.32.
Fig. 8.31. Underpinning of

diaphragm wall for excavation
of formation level below
rockhead; typical Hong Kong
practice
Fig. 8.32. Cutters for a modern

Soletanche cutter rig (courtesy
Soletanche–Bachy)
In a recent paper, Guillaud and Hamelin41 reviewed the development of

diaphragm wall excavation plant during the 1990s and beyond, to 2002.
They describe the early sites as cluttered areas, with noisy machines, a liberal
covering of slurry over ground surfaces and nearby streets, and long road
closures. Such sites should no longer remain; modern, compact low-noise
machines and similar slurry treatment equipment have allowed environ-
mentally friendly diaphragm wall excavation to greater depths with greater
accuracy. Guillaud and Hamelin review excavation under two headings,
cutters for continuous excavation and hydraulic or mechanical grabs, rope
or Kelly suspended (or a combination of both). For cutters they refer to
recent Soletanche/Bachy cutter developments, referring to the HC03 machine,
particularly for confined, city jobs requiring less than 5 m headroom. Despite
its 90 tonne weight it can excavate to 50 m depth (see Fig. 8.33(a)). The leading
details of this machine are as follows:
. diesel engine delivery 370 kW at 2400 rpm
. max hydraulic pressure 32 MPa
. max depth 50 m
. max torque 80 kNm at 32 MPa
. suction pump: 450 m3 /hr
. max pressure on tool 25 tonnes
. total weight 93 tonnes
. cutter drums 650, 800, 1000 and 1200 mm
. excavated length 2400 mm or 2800 mm.
Regarding noise emissions, the noise from the HC03 is no more than 72 dB in
the cab and 80 dB within a hemisphere of 16 m radius.
(a)
Fig. 8.33. (a) Modern

Hydrofraise rig, type HC03
(courtesy Soletanche–Bachy);
(b) modern Evolution
Hydrofraise rig (courtesy
Soletanche–Bachy) (b)
For deep diaphragm walls, the Evolution Hydrofraise rig (see Fig. 8.33(b))
digs to 70 m depth. Other rigs, capable of excavation to more than 100 m
display the same accuracy, reliability and ease of erection as the smaller
machines. (The Hydrofraise used in 1997 by Obayasti Corp. at Nagoya to
build a buried LNG tank inside a circular diaphragm wall, 1.8 m thick
weighed 245 tonnes.)
Fig. 8.34. Swivelling hydraulic

grab, type KS 3000 (courtesy
Soletanche–Bachy)
Developments to grabs are shown in the Soletanche–Bachy KS 3000 (see

Fig. 8.34), a swivelling hydraulic grab mounted on a hydraulic crane which
also powers the grab. The rig is very compact and can operate very close to
existing walls. Steerable grabs are used on the latest variants.
Both grab and Hydrofraise machines benefit from an automatic control and
reporting system called SAKSO. This system has three stages of automation:
manual (operator in full control); teaching (operator tells the system what
movements to make); and automatic (system repeats movements learnt).
The movements are fourfold: jib orientation, jib angle, grab orientation,
hose winder retract. The real-time monitoring systems of cutters and grabs
now include features to include not only deflection on the XX and YY axes,
but also rotation about the ZZ axis (corkscrewing) and deviation from the
vertical (drift).
Practical design and construction of diaphragm walls

A number of items require consideration in the pre-planning and design of
diaphragm wall works.
(a) Panel size. The panel length will vary from a minimum of one grab bite
to a multiple of grab bites typically 6 to 7 m preferably made up of a
number of whole bites with smaller widths between them. Grab bites
vary between 2.3 and 2.8 m. The panel length will include two stop
ends for the initial (primary) panels, or one stop end for mixed panels
dug next to a completed panel. Secondary panels are those dug between
two previously concreted panels. Cut joints can be used with cutter rigs,
and are particularly advantageous for deep walls. Views of the surfaces
of a cut joint from a test panel are shown in Fig. 8.35. The length of the
panels must first be assessed on the basis of panel stability (DIN
Standard 412642 gives methods of assessing panel stability for varying
subsoils and surcharge loadings). It is necessary to limit panel length,
and hence panel volume, to ensure that concrete outputs are sufficient
to fill the panel within a reasonable period taking maximum daily
working hours into account. Panels of modest depth can often be
dimensioned to ensure excavation of one panel each day (say a 20 m
deep panel 4.5 m long dug in stiff clay with no obstructions at an
average of 5.5 m2 per hour, rope grab excavation for 10 hr, one daily
Fig. 8.35. Surfaces of cut joint

from test panel
shift). Such an arrangement gives output a rhythm. Panel size is there-

fore a decision for the specialist contractor.
(b) Excavation sequence. The sequence of excavation to the basement walls
is planned to minimize rig movement and avoid moving pipework from
panel to panel. Where stage completions of the peripheral walls are
agreed, or where top-downwards construction encourages simultaneous
use of the site by both superstructure and substructure contractors,
detailed programming must allow access to panels for rigs, muck-away
trucks or slurry removal vehicles, service cranes and concrete trucks,
and allow curing of concrete in complete panels prior to adjacent
mixed or secondary panels.
(c) Panel stability. Working platform levels must be selected with an aware-
ness of the minimum differential between the head of bentonite slurry
in the panel excavation and groundwater level next to the panel. The
minimum value to ensure stability is 1.0 m, and many specialists
would specify a preference for 1.5 m, especially where groundwater
flows in permeable strata.
(d ) Guide trench construction. The standard of diaphragm wall construction
is itself influenced by standards of temporary guide wall construction.
The guide walls must be sufficiently robust to avoid movement due to
loads from excavation rigs, service cranes or reinforcement cages and
reaction from stop end jacking systems. Reusable precast concrete
guide wall sections have been successfully used on T-shaped panels
and cellular walls but each precast unit must be interlocked by a
bolted mechanical joint to ensure the same standard of rigidity as in
situ concrete walls. It is essential to maintain continuity between in
situ guide wall pours by reinforcement passing through the construction
joints.
(e) Wall–slab construction joints. Joints between basement rafts and
diaphragm wall, and between intermediate basement floors and wall,
can transmit vertical shear or, where necessary, bending moment. Alter-
native forms of joint construction are shown in Fig. 8.36. Threaded-end
couplers (such as Lenton couplers) can be used to develop the full
strength of reinforcement bars from within a recess in the face of the
Fig. 8.36. Vertical cross-

sections of typical horizontal
joints between diaphragm wall
and basement floor slab:
(a) to achieve resistance to
groundwater ingress and
transmission of vertical shear;
(b) moment connection by
couplers
diaphragm wall at the junction with the slab. Bend out bars can be used
instead of couplers, although the closeness of the bar spacing and the
diameter of these bars may be limited by the ability to house them in
the face of the joint.
( f ) Box-outs. The depth of box-out will normally be limited to the concrete
cover on the main reinforcement. Although polystyrene is used as a
Fig. 8.36. (c) Variation for a structural hinge; (d) Variation for structural continuity
box-out material there is a risk that the box-out will be displaced unless
a timber frame with a thin plywood cover secured to the front to
ensure its rigidity is used to contain the polystyrene. It is usually a
disadvantage to extend the depth of the box-out behind the vertical
reinforcement due in part to the difficulty of adequate preparation of
this face of the joint but particularly because of the risk of entrapment
of heavy slurry below the box out during concreting.
(g) Reinforcement cage and density. It is unwise to allow the requirements
of calculated shear, moment or crack width to make the spacing of
horizontal binders and vertical main steel too small. DIN Standard
412642 advises that to ensure no slurry inclusions remain in the con-
crete, the differences between the flow resistances in adjoining zones
in plan in the panel should be kept as small as possible. Minimum
spacing of bars for both single- and double-layer vertical steel and
horizontal binders is reproduced in Fig. 8.37.
The increasing depth of walls and the higher calculated flexural
strengths required from them has led to cages of considerable tonnage.
These cages are usually joined in sections over the panel using couplers
to join bars in each section. A cage awaiting the arrival of the next
section is shown in Fig. 8.38. The vertical bars are staggered to avoid
couplers forming a block to concrete flow at one level. Cages are some-
times fabricated off-site in sections and transported to site overnight (see
Fig. 8.39). The maximum panel length may be restricted because of
transport although multiple cages may be used to allow an economical
panel length. Total cage weights of jointed sections for deep walls may
demand very large craneage for the final lift into the panel. A 90 tonne
weight cage is shown during final insertion into the panel in Fig. 8.40.
(h) Tremie operation. Reinforcement cage detailing must allow sufficient
vertical access for tremie tubes. Very small single-grab bite panels can
only accommodate a single tremie tube, and even with almost continu-
ous concrete supply from truck mixers it is difficult to obtain an average
pour rate greater than 40 m3 per hour. For larger panels a second tremie
pipe can be used and with continuous concrete delivers a concreting rate
between 60 and 80 m3 per hour. In deep, large, T-shaped panels it may be
necessary to use three tremies to maintain a uniform upper surface to the
concrete as the pour continues, but generally two tremies are sufficient.
Fig. 8.37. Minimum spacing and concrete cover to reinforcement in diaphragm walls as recommended by DIN standard 412642
Fig. 8.38. Reinforcement cage

in panel excavation awaiting
jointing with a further cage
length
Fig. 8.39. Cage transportation
(i) Concreting rate. In the Author’s view there is considerable risk of poor
panel concrete when concreting rates drop below 15 to 20 m3 per hour
per tremie. Even when concrete mix quality varies from the optimum, a
high rate of concreting may be sufficient to displace the slurry, scour the
surface of the reinforcement bars and flow between the reinforcement
and around the box-outs. Where the density of reinforcement bar is
high, and especially with multi-layers of bars, a high concreting rate
and a very workable cohesive concrete are essential.
(j) Slurry reuse. The earliest diaphragm wall jobs by Icos did not have the
equipment to clean bentonite slurry after excavation and concreting.
The slurry was used once and, after storage for much of the pour
volume, was carted from site in road tankers or, on the earliest jobs,
pumped into public sewers. With the advent of reverse circulation
Fig. 8.40. Heavy cage during

final lowering into panel. In this
panel, two cages have been
used in separate vertical
sections
rigs which used large quantities of slurry to excavate each panel, clean-
ing technology from the oil industry was used to design and build
shaker screens, hydrocyclones and at a later date, centrifuges and
presses, to clean the slurry. Compact, transportable units were made
which could be brought to site and quickly mobilized for use. Nowa-
days, slurry is cleaned and reused as a matter of routine on virtually
all diaphragm wall works to clean bentonite slurry where either grab
or reverse circulation rigs are used.
(k) Stop ends and extraction. Initially, tubular steel stop ends were used to
form semi-circular joints between diaphragm wall panels. This practice,
popular with Icos, the originator, and other European firms, gradually
changed as rectangular formers gained acceptance. Both types of stop
end were extracted as the concrete at the bottom of the panel started
to set and gain strength. In very large pours, it was necessary to use
high dosages of retarder additive within the concrete to ensure that
the set was delayed. Even so, extraction of the stop ends often began
before the concrete pour had been completed. Types of vertical panel
joints are shown in Fig. 8.41. With the advent of reverse circulation
rigs with vertically mounted cutter wheels, it became possible to cut
the vertical surface at the end of a concreted panel during excavation
of the adjacent panel. This cutting back of the concrete surface avoided
the use of temporary formers, and the new concrete in the second panel
could be poured against a true vertical surface on the first panel. This
procedure has since become less favoured because of the risk of
heavy, calcium-contaminated slurry remaining near the cut surface,
Fig. 8.41. Types of vertical

panel joint used in diaphragm
waling, shown in horizontal
section
which could contaminate the concrete within the second panel, near
the vertical joint. This slurry-contaminated concrete often proved to
be porous and led to leakage of groundwater into the basement after
excavation. The CWS-type stop end former developed by Bachy
incorporates single, twin or triple water bars cast into the vertical joint
and is released from the vertical surface of the concrete pour after the
concrete has hardened during excavation of the second panel. This
type of joint (and its derivatives by competing specialist firms) has now
gained widespread useage where resistance to groundwater penetration
is needed. The CWS joint details are shown in Fig. 8.42. Examples of
CWS joint construction for a 1.5 m wide wall, 55 m deep with a double
water bar are shown in Fig. 8.43. A temporary plywood former was
used on the CWS stop end to facilitate its removal.
Recent developments in diaphragm wall works

Since the conception by Veder of reinforced concrete walls cast into trenches
dug under bentonite slurry, and development of the process by Icos and
others, there have been many improvements in both application and
Fig. 8.42. CWS joint detail: (a) pulling out stop end sideways after excavation of the adjacent panel; (b) stop end blades
installed in the CWS joint; (c) CWS joint before concreting; (d) CWS joint with water stop installed (courtesy of
Soletanche–Bachy)
mechanical plant. Developments reviewed by Puller and Puller43 in 1992 and

since in terms of plant innovation by Guillaud and Hamelin41 are as follows:
(a) the use of polymeric slurries for excavation support, avoiding effects of
some contaminants, reducing pumping energy, avoiding wall cake
thickness, reducing slurry disposal costs
(b) the development of structurally efficient diaphragm wall plan shapes
based on improved joint efficiency between panels
(c) the use of post-tensioned diaphragm walls, either precast or cast in situ
(d ) the use of precast concrete diaphragm walls
(e) the use of reverse circulation excavation equipment such as the
Hydrofraise and Trenchcutter rigs
( f ) the development of excavation equipment for work on congested sites
(g) improved design of mechanical grabs
(h) the development and use of electronic monitoring in grabs and cutters
to improve panel excavation tolerances and overall quality control
standards
(i) the development of improved stearage of cutters and hydraulic grabs
Fig. 8.43. CWS joint fabrication for a 55 m deep panel joint
(j) the development of improved stop end design; use of CWS joints
(k) improved standards of slurry cleaning and quality control of slurry
during use.
Items (a) and (b) are reviewed in more detail below; the remainder are referred
to elsewhere in this chapter and in Chapters 4 and 9.
While the original use of bentonite slurry to support diaphragm wall
excavations has generally persisted for excavations made by grab, the larger
quantities of slurry required for circulation purposes with Hydrofraise and
Trenchcutter equipment have brought innovation to slurry design. Using
experience from the oil drilling industry, polymeric slurries and mixes of poly-
meric and bentonite slurries have been used successfully on larger diaphragm
wall jobs where the economies of scale have proved beneficial.
Polymeric slurry behaves as a pseudo-plastic fluid and, unlike bentonite
slurry, acts in trench support without forming a filter cake. Within the
polymeric slurry a molecular lattice structure, which is different from the
thixotropic gel structure of a natural clay slurry, is built up after mixing.
The polymeric structure is more efficient in the suspension and transportation

of soil particles and leads to reduced energy costs in pumping from excavation
to slurry station on the job site. As fluid loss from polymeric slurries is less
than that from bentonite slurries, the polymeric slurry can be used with advan-
tage in weak soils where an increase in soil moisture content would cause risk
of panel instability. Polymeric slurries cost considerably more than bentonite
slurry, but since, with care, the slurry is reusable many times, disposal costs
partly compensate for the high initial cost. The rheology of bentonite slurries
was discussed by Rogers44 , and the merits of polymeric slurries were reviewed
by Beresford et al.45
Icos began innovations in joint design to enable wall panels to be incorpo-
rated into rectilinear and arch plan shapes in the late 1960s. At Redcar, UK,
diaphragm wall panels were joined by tension connections to form cellular
structures to a new harbour wall. In plan, each of the cells measured 30 m
by 15 m. In plan, an inverted arch of diaphragm wall panels was restrained
by the tensile resistance of cross-walls, also in panels which, in turn, were
anchored to a rear arch of panels. The weight of the soil enclosed within
the cells was not used to restrain the wall in overturning because the panel
joints could not transmit vertical shear, only tension46 .
Developments by Bachy have led to patented joint forms that develop either
full flexural continuity from panel to panel (known as the Teba system) or
shear between adjacent panels. The first option, to develop a continuous
wall, is based on the use of hydraulic jacks cast within the panel concrete
Fig. 8.44. Development of continuity between diaphragm wall panels to achieve improved, structurally efficient wall plan
forms43
and near the panel joint. After concreting, these jacks are actuated and rods
are thrust horizontally from the cast concrete to provide continuity with the
reinforcement of the adjacent panel subsequent to excavation and concreting.
The method is rarely used and only finds application on special projects
because of the cost implications of the jack system.
The use of joints to transmit shear, however, has wider application and
more scope for development with methods of interactive soil-structure analy-
sis. The construction of walls from large T-shaped and H-shaped diaphragm
wall panels joined by shear connection enables the full flexural strength of
the multi-panel plan shape to be realized and, in addition, can utilize the
stabilizing effect of the weight of soil encapsulated between the legs of the
T-units or within the enclosed cellular areas formed by a series of H-panels.
The use of such plan shapes to form semi-gravity structures, shown in
Fig. 8.44, may only be justified in terms of economy where bracing or
anchoring of a basement wall is not possible and single-wall construction
has insufficient flexural strength to cope with high imposed cantilever
moments and shears. The wall area in T-shaped or cellular panel construction
is not cost-effective in terms of the linear wall unless special support con-
straints apply. Figure 8.45(a) shows development in joint construction to
transmit full continuity, shears and tension. The use of shear joints, however,
to form a diaphragm wall of varied castellated plan shape can be very cost-
effective as a deep unpropped cantilever for a very large basement providing
the plan area occupied by this wall can be accommodated in the site space. The
flexural efficiency of the castellated shape is shown in Fig. 8.45(b).
The use of semi-gravity structures may be justified where soil and wall
movement is critical. Such movements can be effectively minimized by the
use of a stiff wall structure and the gravity effect of the weight of soil retained
within the cells. One of the largest structural diaphragm wall contracts
completed to date utilized semi-gravity cellular plan form jointed diaphragm
wall panels for a large underground car park in Medinah, Saudi Arabia, in the
Fig. 8.45. Diaphragm wall built

to castellated plan shape: (a)
typical layout; (b) plot of section
modulus of castellated section
against overall depth of section
showing improvements to
flexural strength compared with
a straight wall
Fig. 8.46. Medinah car park: views of unobstructed excavation supported by cellular diaphragm wall
early 1990s. The total area of structural wall exceeded 320 000 m2 to provide
the exterior walls to an excavation approximately 18 m deep from the ground
surface, 100 m wide and more than 1.5 km in total length. Views of the com-
pleted diaphragm walls are shown in Fig. 8.46. The site for the car park lies in
the centre of a bowl of washdown soils, mainly silty sands and clayey silts
from surrounding mountains. The depth to basaltic bedrock over the site
area varied from 23 to 55 m. The groundwater level had been monitored
over a period of two years prior to construction and showed some variation
with time. The average level during construction was 2 m above final excava-
tion level.
The client’s brief was to provide a basement construction with a design life
of 120 years, which could be excavated without the impediment of cross-
bracing, raking shores or temporary berms, and would give the minimum
of soil movement behind the walls and, in the long term, below the lowest
basement level in the car park. Precious historic structures were sited 20 m
from the excavated face of the basement wall and neither noise nor vibration
could be tolerated. A cellular diaphragm wall construction was adopted,
propped by two basement floors in the permanent condition but acting as a
cantilever in the two-year construction period. The cellular wall shown in
plan in Fig. 8.47 was adopted using vertical rock anchors into the basalt at
the rear of the wall to achieve stability during construction where bedrock
was shallow, but where bedrock exceeded 35 m in depth the cellular wall
was allowed to cantilever during the construction period.
The use of the cellular wall was justified on the basis of its stiffness to reduce
wall and soil movement and the opportunity it gave to undertake unimpeded
bulk excavation to the basement with staged handovers of large working
areas. Alternative designs using ground anchors, bored piles and T-shaped
diaphragm wall panels were considered but were precluded by the predicted
soil movement and the relative inefficiency of ground anchoring in the
Medinah silts and clays.
To construct the H-shaped cellular diaphragm wall units, three separate
panels were excavated and cast using precast concrete permanent stop ends
in the construction of the central web panel. Concrete was poured into the
central web excavation enclosed within a bag of plastic sheeting to avoid
leakage around the stop end which would reduce the effectiveness of the
shear and tension bond to the outer panels. The use of plastic sheeting as a
temporary measure proved effective in retaining all the concrete within the
web panel. The use of projecting reinforcement from one panel to the next
had been developed in France and Japan, but its application had been limited;
Fig. 8.47. Medinah car park,

cellular wall construction:
(a) plan of cellular wall
construction; (b) construction
procedure for single wall unit;
(c) wall unit construction
adjacent to completed unit
at Medinah, however, the precast joints were used successfully more than 1000
times with only five minor collapses. Construction of the central web panel
cage is shown in Figs. 8.48 and 8.49.
The rate of diaphragm wall production on this contract was impressive. The
soil conditions, which varied from medium strength silts and clays to stiff and
hard clayey silts, were conducive to grab excavation but were less economic-
ally excavated by cutter rigs until excavation depths increased below 20 m or
so. Three cutter rigs and up to five rope grabs were used on a 24-hour, 6-day
week basis. Concreted panel production averaged more than 4000 m2 per
week over much of the two-year wall construction period and reached
7000 m2 per week over several weeks at peak production.
The Medinah cellular wall was designed using a two-dimensional analysis
of the soil–wall structure taking into account non-linear elastic–plastic soil
conditions. Soil movement and stress levels in the surrounding soils were
predicted for the modelled excavation stages by finite element analysis,
taking into account the dissipation of negative pore-water pressure with
time for varying depths of rockhead and groundwater. The analysis methods
were those described by Jardine et al.47 The structure deformation results
from the two-dimensional analysis were then used with a three-dimensional
structural program to predict stress levels and design reinforcement within
the cellular wall. This work was incomplete because the analysis did not
Fig. 8.48. Medinah car park:

cellular wall detail showing
reinforcement and permanent
stop ends to central web panel
include the deformations within the soil and the resulting changes in soil stress
levels caused by excavation of the diaphragm wall panel itself. It had been
realized prior to the Medinah design that accurate prediction of soil move-
ment and stress adjacent to a completed diaphragm wall basement excavation
had to include installation effects of the diaphragm wall panel during panel
excavation and concreting. The effects of concreting on in situ soil stress
had been shown48 in measurements which concluded that induced stress
caused by the pressure of wet concrete within the diaphragm panel, and the
resulting soil deformation, discouraged the use of low earth pressure coefficients
in wall design. More recently Lings et al.49 referred to placing temperatures
within the concrete as an important influence on wet concrete pressure.
A number of published results of soil movement caused by panel excavation
tend to show small horizontal soil movements near the panel, rapidly
reducing at short distances equal to the panel length, say, from the panel.
Measurements made at Medinah confirm this. (The exception to these
generally small movements were those observed during diaphragm wall con-
struction for the Hong Kong Metro. The decomposed granite residual soil
conditions in Hong Kong are quite different from those where the other
measurements of soil deformation adjacent to panel excavation were made.
It was concluded that the cause of the large soil movements in Hong Kong
Fig. 8.49. Medinah car park:

(a) fabrication of web panel with
permanent precast concrete
stop-ends to cellular wall;
was the presence of a soil with high swell potential, high permeability and a
high groundwater table.) Finite element modelling of installation effects by
Gunn et al.50 showed promise in predicting soil movement and stress during
panel excavation.
The Medinah soil–structure analysis to predict deformed wall shape and
soil movement was only partly successful because of the difficulty of modelling
the shear stiffness of a three-dimensional structure in plane strain and the
oversimplification of ignoring panel installation movement and stress in the
Fig. 8.49. (b) Key plan of one

element; (c) isometric of web
panel
subsoil. The maximum predicted horizontal soil movement of 27.5 mm

after bulk excavation for the deepest cantilever walls was not reached, the
total observed maximum horizontal deformation for both panel and bulk
excavation for the cellular wall being less than 20 mm, allowing for dissipation
of pore pressure with time.
Composite walls and grouting techniques

Mention should be made of the use of diaphragm walls, either precast or in
situ concrete construction, as part of a soil retention and groundwater control
protection system incorporating non-structural slurry wall cut-offs and hori-
zontal grout plugs over the plan area of the basement. Figure 8.50 shows
examples of composite precast diaphragm walls incorporating temporary
king post walls cast in the upper section of one wall. The use of jet grouting
and intrusion grouting to form a horizontal grout plug to control the inflow
of groundwater to excavations was discussed in Chapter 2.
Water resistance of structural diaphragm walls

The difference between expectation and the actual performance of diaphragm
walls regarding water resistance has caused disappointment and dispute since
the earliest structural walls in the 1950s and 1960s. Then, structural walls
generally required the minimum of surface treatment to produce a dry face.
Many of the earliest diaphragm wall basements in Paris and London were
used for car parking and were either left without finishes or, at most, with
an applied sand–cement render. Where leaks occurred these were sealed by
application of surface chemical renderings such as Vandex or Xypex to
make a crystalline waterproof coating to the wall. These walls, generally
600 or 800 mm thick, were excavated by rope grab or hydraulic grab (rope
or Kelly mounted), and temporary tubular steel stop ends were used through-
out. Only in basement construction where storage of perishable goods was
Fig. 8.50. Composite

diaphragm wall
construction using precast
concrete wall sections:
(a) anchored Panosol wall
with slurry cut-off wall into
an impermeable stratum;
(b) anchored Panosol wall
with temporary Berlin wall
above and slurry cut-off
wall below;
(c) load-bearing anchored
Panosol wall with concrete
wall below, penetrating a
bearing stratum at depth;
(d) table of flexural strength
of typical Panosol panel
sections (courtesy of
Soletanche–Bachy)
planned, in shopping areas or office facilities, was a separate lining wall

constructed.
In London, the earliest diaphragm walls were built by Icos from 1961
onwards. By 1974 a sufficient number of contracts had been completed to
hold a keynote conference on diaphragm walling. At the conference, Sliwinski
and Fleming51 addressed the problem of water resistance:
It is therefore evident that the concrete used for diaphragm walls can for
practical purposes be considered impermeable. However, in practice the
permeability of a panel must also depend on the formation of cracks and
on any local defects in the concrete such as may result from segregation
but with normal concrete control and care, such occurrences should be
limited to isolated cases. The simple butt joint between panels cannot
be claimed to be proof against water entries but significant leakages are
rare due to the presence of soil impregnated with bentonite behind the
joint, and to the presence of some thin layer of contaminated bentonite
at the edges of the joint. Where leaks occur they can usually be ascribed
to differential deflexions between wall panels, and these differentials are
worse near corners. The whole matter of deflexion differentials (and
thereby risk of leakage) between wall panels depends on panel shape in
plan, wall height, the use of anchors, excavation procedures and other
factors. The present practice for dealing with damp joints is to allow
the leak to appear, for the differential wall deflexion, for the most part,
to take place, and then to inject cement or chemical grout into the soil
at the back of the joint, either vertically or horizontally through drillings
depending upon access. Alternatively, steel or other suitable plate,
bedded on epoxy-resin mortar, can be bolted to the concrete over the
internal face of the joint.
Sliwinski and Fleming continued to refer to panel deviations caused by
obstructions due to concrete passing beyond the tubular stop ends, but did
not refer to this as a risk of bad water resistance of the wall below basement
excavation level. No reference was made to the occurrence of leaks at the junc-
tion of diaphragm wall and basement slab in basement or other construction.
The situation described by Sliwinski and Fleming provides a reasonable
summary of the opinion of UK specialist firms in the early 1970s. BS 200452
stated:
These joints are usually watertight but minor seepage through leaking
joints can be dealt with by grouting or may even be tolerable in certain
classes of structure.
Generally the attitude was optimistic, and risk of leakage was only con-
sidered after it had happened. The earliest model specification in the UK53
did not refer to water resistance and typical paragraphs in tender letters by
specialist contractors stated:
The diaphragm walls will be constructed so as to be substantially water-
tight on initial exposure (free from running leaks but not damp proof )
and we only accept responsibility for repairing leaks, within the exposed
height, caused by faulty workmanship and/or materials. It should be
noted that possible ingress of water into the excavation from below
formation level is not prevented by the diaphragm walls.
Overall, specialist firms were relatively optimistic about the likely occurrence
of leaks on jobs throughout the UK in a variety of soil conditions and for a
range of basement uses. The risk of overbreak, panel collapse, loss of bento-
nite, displaced box-outs, inadequate excavation rates, etc. were all critical
tender risk assessments for the specialist contractor, and wall waterproofness
and the cost of associated remedial works were not considered as important as
they are today. In the UK in the 1960s and 1970s, specialist firms were usually
awarded diaphragm wall contracts on the basis of design-and-construct after
technical discussion with a consulting engineer or architect. The contractual
risk for waterproofness (apart from damp patches) generally remained
firmly with the specialist contractor. By the end of the 1970s most major
consultancy firms were designing and specifying diaphragm wall schemes
themselves and the contracts were let on a construct-only basis. At this
stage, the overall use of the underground structure was clear to the designer,
who could incorporate measures such as non-load-bearing blockwork walls
and drainage channels to hide any persistent ingress of groundwater.
The use of basement lining walls, cut-and-cover works and underpasses has
continued in the UK, France and Germany. In some instances (Lyon Metro
1981, Eastbourne Pumping Station 1993) an in situ reinforced concrete lining
has been specified by the engineer or owner to be capable of withstanding the
full groundwater pressure acting on the wall.
As referred to previously, the guidance for water resisting design of
basements is BS 810254 and CIRIA report 13911 . The use of diaphragm
walls in basements can comply with the water-resisting methods defined in
documents in terms of method B structural integral protection and method
C drained cavity construction. To address method A (an external waterproof
membrane) specialist firms have sought over a period of some years to develop
a diaphragm wall system to include an outer plastic liner to completely
encapsulate the diaphragm wall. The method has had very little application
on site and doubts remain regarding its efficiency.
The requirements of BS 8102 to comply with type C (drained protection)
have become standard design principles for basement diaphragm wall work
in the UK. Most basements are designed with a drained cavity construction
and only those used for car parking have the option of exposed unclad
diaphragm wall surfaces. Building owners are not prepared to allow unlined
diaphragm walls (say for a basement for storage use) where there is reasonable
chance of a change of use during the life of the basement. Current practice is
therefore to make a drained cavity with a permanent pump provided at a
sump at the lowest level. To make this provision, the design volume of the
basement is reduced by the volume of the drainage cavity, the volume of
the blockwork lining and the volume occupied by the verticality tolerance
of the diaphragm walls (say for grabs 1:80 or 1:100), irrespective of whether
this tolerance is used by the wall or not. It is widely acknowledged that this
solution is uneconomic and often leads to a reduction in car parking spaces
in basements where lining is used.
The use of lining walls does not automatically produce protected construc-
tion because faulty drainage cavities and drainage between basement floors
often lead, over time, to damp blemishes on the exposed face of the lining
because of leaks in the hidden diaphragm wall.
The current situation in the UK is this: the specialist contractor generally
contracts to leave the exposed diaphragm wall free from running leaks (but
not damp patches) and probably has half of the full retention money held
against this for, say, twelve months from the end of the main contractor’s
contract. When leaks arise during this maintenance period the specialist con-
tractor seals them by grouting and trusts that a final inspection at the end of
the maintenance period will be the end of leakage responsibility.
The use of non-load-bearing lining walls in basements is similar in both
Germany and the USA, although a slightly more optimistic view regarding
the water-resisting efficiency of diaphragm walls may remain in France. The
water resistance of structural diaphragm walls has been reviewed in some
detail55 .
In the Author’s experience, basement wall leakages occur at any of five
locations:
. in the panel itself
. at vertical panel joints
. at horizontal bottom slab/wall joints
. at the top horizontal joint between panel and capping beam
. below formation level.
Each is now discussed in more detail.
(a) The panel. Leaks and damp areas in the panel are caused by soil or
slurry inclusions, random cracking perhaps due to shrinkage, or poor
quality concrete. Xanthakos56 concluded:
It is not difficult to produce walls with permeability of the order of
1010 cm/s. Suppose a wall is 60 cm thick, retains a hydraulic head
of 10 m and has a porosity of 15%; if we take into account a
suction pressure of 1 atmosphere to assist water flow, the quantity
of water percolating through the wall is close to 0.3 litre for
100 m2 of wall surface over a period of 24 hours.
So if this estimate is correct and the concrete is sound and homo-
geneous, only very limited dampness should occur on the exposed
surface due to concrete permeability.
If soil or slurry inclusions penetrate the full thickness of the wall in
water-bearing ground their removal can be difficult and expensive.
This matter is a particular risk in water-bearing silts of low strength.
The density of reinforcement, the depth of wall, the size of box-outs
and the thickness of the wall, concreting continuity and concrete
fluidity will all influence the risk of slurry inclusions. The risk is
increased by any obstruction to flow or tremied concrete and any dis-

ruption, time-wise, to the flow of fluid concrete. These risks are evident
in thin, highly reinforced walls with box-outs. Entrapment of heavy
slurry below congested, large-diameter reinforcement at slab/wall
junctions are particular risk areas for slurry-contaminated concrete
and leakage. Random cracking within panels seldom appears, and
cracks are not necessarily continuous nor to full wall depth, but where
they occur the cracks can cause running leaks. There is a possibility that
the size of vertical reinforcing bars may influence shrinkage cracking in
relatively thin walls.
There is risk of soil inclusion in weak silts and silty clays and in highly
fissured over-consolidated clays, and panel length should be minimized
in such soil conditions. Overall, poor water resistance resulting from soil
or slurry inclusion may prove to be a high-cost risk because removal of
the inclusion throughout the whole wall thickness may be essential for
strength, durability and waterproofness requirements. This operation
may prove to be costly both in remedial expense and contract delays.
Fig. 8.51. Leakage path of

groundwater between water
bars at basement slab level,
diaphragm wall construction
(b) Vertical panel joints. This is the area at greatest risk of leakage. The
most likely causes of leakage are panel movement caused by bulk exca-
vation or application of vertical load, or pre-stress of ground anchors,
honeycombed concrete and poor concrete near the stop end due to
inclusions. Leakages can also occur through slurry contaminated with
calcium from concrete cut from the end of the adjacent panels, and
through inclusions in joints where precast concrete stop ends are
inadequately cleaned before concreting. The use of CWS joints to
reduce the risk of leakage at vertical joints by the inclusion of one or
more water bars has been shown to be effective. Nevertheless some
leakage may occur due to the passage of water between the water
bars as shown in Fig. 8.51.
(c) Horizontal bottom slab/wall joints. The next highest risk of leakage is the
horizontal joint at the basement slab. Although such leaks are possibly
a split responsibility between wall and slab constructors, precautions to
avoid leakage are necessary in the wall design. The length of the leakage
path from the underside to the top of the basement slab will be
increased if the underside of the slab is haunched to an increased thick-
ness adjacent to the junction with the wall. Special provision can be
made by securing a horizontal L-section flexible water bar to the vertical
wall surface within the slab box-out recess and connecting this to the
slab water bar system. This precaution is not successful, however,
unless the L-section water bar is connected to a water bar within the
vertical panel joint. A continuous water bar system between wall and
slab, in all horizontal and vertical joints, is expensive and success is
not guaranteed. The provision of a continuous water bar system is
more easily achieved in precast diaphragm wall construction than
conventional in situ walls. The risk of heavy slurry entrapment referred
to previously should be noted.
Where a groundwater head exists below the basement floor/dia-
phragm wall joint, it is essential that design provision is made to
resist penetration. The practice of pouring slab concrete into a recess
in the wall to receive the floor slab is simply not sufficient to ensure
water resistance even if the vertical concrete floor slab surface within
the recess is correctly prepared.
(d ) Top horizontal joint between panel and capping beam. It is not unusual to
find leakages in the horizontal joint at the top of a diaphragm wall with
a capping beam or other in situ reinforced concrete where a high
groundwater occurs or where rain water is allowed to pond in porous
backfill to the capping beam/guide wall excavations. Such leakages
can occur even when the top of the diaphragm wall is adequately cut
down to remove porous concrete and the surface is correctly prepared.
Although such work is essential, the provision of a Hydrotite water bar
strip in the horizontal joint on the earth face would overcome the risk of
water leakage.
(e) Below formation level. This risk is often ignored, although the financial
consequences of wall leakage or even loss of ground where ‘blow’
symptoms occur are likely to be very significant. The standard tender
clause used by some diaphragm wall specialists in the past has stated:
We only accept responsibility for repairing leaks, within the
exposed height, caused by faulty workmanship and/or materials.
The implication must be that the specialist contractor would not be
responsible for faulty workmanship and/or materials below formation level
in the area of highest risk. The clause was rarely queried by main contractors
or their clients and presumably relieved the specialist contractor of consider-

able risk.
The obvious preventative measures for leakages all refer to standards of
workmanship and quality control during diaphragm wall construction. The
importance of design decisions, however, regarding wall thickness, permanent
stop end construction and provision of water bars in panel joints, should not
be ignored when risk of leakage below formation level is assessed.
Up to the 1990s the general level of acceptance of wall water resistance
in the UK was based on the criterion that damp patches on the exposed
wall surface would be accepted but running leaks would not, and grout or
surface treatment would be accepted as a remedy. Documents such as the
DIN standard 412642 , the British Department of Transport’s Specification
for Highway Works57 , and the European Code Execution of Special
Geotechnical Works: Diaphragm Walls58 , failed to make any reference to
acceptable standards of water resistance.
The ICE Specification for Piling and Embedded Walls59 , a widely used
standard specification together with a Particular Specification written for
the individual contract, makes reference to water resistance, defined as
‘water retention’ but fails to define unacceptable leakage volumes or areas
of dampness. The wording is as follows:
The Contractor shall be responsible for the repair of any joint, defect or
panel where on exposure of the wall visible running water leaks are found
which would result in leakage per individual square metre in excess of
that stated in the Particular Specifications. Any leak which results in a
flow emanating from the surface of the retaining wall shall be sealed.
In France DTU 14.160 sets more definitive limits:
The wall shall comply with the watertightness criteria of a lining wall to a
relatively waterproof structure.
The watertightness of the wall will be such that the flow of seepage and
leaks is limited to
For the whole of the outside walls in their entirety:
0.5 litre/m2 /day on yearly average
1.0 litre/m2 /day on weekly average
For all areas of 10 m2 of wall:
2.0 litre/m2 /day on weekly average.
These flows take account of the seepage flows at the joint of the raft to the
wall. The limited exactness of the ICE Standard Specification is replaced by
precise legal definitions of leakage standards in the French specification,
although some difficulty may be experienced in measuring such quantities
and applying them accurately.
A useful clause in use in the Middle East appears to strike a realistic
specification:
The diaphragm wall shall be watertight and substantially dry. Remedial
measures shall be carried out as directed by the Engineer in areas that do
not comply with this requirement. A panel will be considered acceptable
if within any area 1 m square the total damp surface does not exceed 10%
of that area. The Diaphragm Wall Contractor shall be responsible for the
repair of any joint, where, on full exposure of the wall, leaks are found.
Running leaks will not be accepted at any location.
It is the Author’s opinion that specification clauses for water resistance of
diaphragm wall works in general use are inadequate because they do not
specify acceptance to adequate standards, they do not specify the method of

repair and they do not approach the subject of waterproofness with the
considerable importance and detail that it deserves. Several matters are
ignored in the latest specifications. As an example, the building owner or
tenant’s interests are not served by the lack of consideration for leakages
occurring after the contractual maintenance period by, say, rising ground-
water, the failure of repairs to previous leaks, or movements between panels
or basement floor joints due to long-term soil movement such as heave.
Current specifications do not relieve the specialist contractor for leakage
due to panel movements caused by application of superstructure loading or
anchor stressing, matters that are frequently outside his control. The water
bar system in the diaphragm wall should be specified by the designer. The
wall system should be connected efficiently to the water bar system in the
base slab, and should also be specified by the basement designer.
Overall, early optimism among specialist contractors in Italy, France and
the UK has now been replaced by a realism that acknowledges that concrete
tremied into a slurry in a series of panels will not automatically produce a dry
basement. In Europe, internal lining walls are frequently used to avoid the effect
of the groundwater ingress. Unfortunately, these internal walls cover up a
continuing risk to the building owner and mask the occurrence of further
leaks or deterioration in the repairs to the original leaks.
Overall stability: Where hydrostatic groundwater pressures, during construction and within the
design for uplift design life of the structure, are at a higher elevation than the underside of the
lowest basement floor level, it is necessary to examine the overall stability of
the basement. Failure due to the lack of buoyant self-weight of the basement,
and insufficiency of frictional forces to avoid upward displacement of the
basement substructure, are fortunately infrequent, but not unknown. Vertical
anchoring of the basement raft to rock strata or strong soils below the base-
ment becomes necessary where hydrostatic uplift is severe and such strata
exist at economical depth. Drainage of a granular blanket or porous no-
fines concrete with permanent pumping may prove necessary where anchoring
is not feasible. The rise of groundwater within the design life of the structure
should be carefully assessed after consideration of the dead weight of the
structure at successive stages of superstructure construction, including final
completion; a factor of safety of the order of 1.4 should be obtained for the
sum of downward dead weight, total vertical downward anchoring force
and frictional resistance to the basement walls compared with upward
hydrostatic force on the basement underside. Where most of the downward
force consists of dead weight a modified criterion may be used:
Dead weight Friction
Upward hydrostatic force þ
1:1 3:0
Where diaphragm walls are used for basement construction, the frictional
resistance in cohesionless material or the wall adhesion due to clay strata,
should be calculated in the same manner as frictional resistance to bored
piles in similar soils, restricting the diaphragm wall surface used in the
calculation to the inner and outer surface of the wall below formation level.
Further discussion on the overall stability of underground structure is
included in Chapter 9.
Construction The relative costs of secant pile and diaphragm walls given by Sherwood
economics et al.,61 and quoted in Chapter 4, serve only as a comparison between wall
costs and do not consider the relative costs of alternative propping systems or
the costs of linings to the inside face of the wall. To make a logical choice of
wall, propping and lining system on cost grounds, a detailed comparison of
the cost of systems should be made for each job site.
A broad cost comparison is reproduced in Fig. 8.52. This comparison
includes completed wall construction for two- and three-storey basements
Fig. 8.52. Cost comparison

of basement walls for
temporary and permanent
works, two- and three-storey
deep cross-sections: (a), (b)
for soil profiles shown at
boreholes A and B for the
constructions described in
(c)–(f). A base index of 100
has been used for the lowest
cost construction shown in
(c).
in two particular soil and groundwater conditions. It would be unwise to draw

conclusions from this small sample but broadly it showed that, in the UK,
contiguous bored pile walls with an internal lining are comparatively inexpen-
sive in good piling ground; secant piling is an economic choice in less condu-
cive soil conditions and for deeper basements; where sheet piling is left in
place, an expensive wall results; anchored diaphragm walling is competitive
in deeper basements in good soil conditions.
References 1. Lambe T.W. Proc. Conf. Lateral Stresses and Earth Retaining Structures, ASCE,
New York, 1970, 149–218.
2. Wakeling T.R.M. Discussion, deep excavations. Proc. 6th Euro. Conf. S.M.F.E.,
Vienna, 1976, Vol. 2.2, 29–30. Austrian National Committee, Vienna, 1976.
3. Skempton A.W. and La Rochelle P. The Bradwell slip, a short term failure in
London clay. Geótechnique, 1965, 15, Sept., 221–242.
4. Gässler G. and Gudehus G. Soil nailing, some aspects of a new technique. Proc.
10th Int. Conf. S.M.F.E., Stockholm, 1981, Vol. 3, 665–67. Balkema, Rotterdam,
1981.
5. Banyai M. Stabilization of earth walls by soil nailing. Proc. 6th Conf. SMFE,
Budapest, 1984, 459–466. Akadémiai Kiado, Budapest, 1984.
6. Project National Clouterre, recommendations 1991. Presses de l’Ecole Nationale
des Ports et Chaussées, Paris, 1992.
7. Barley A. Soil nailing case histories and developments. Proc. Conf. Retaining
Structures. Institution of Civil Engineers, London, 1992.
8. Schnabel H. Sloped sheeting. ASCE J. Civ. Engng, 1971, 41, No. 2, 48–50.
9. Hanna T.H. Anchored inclined walls – a study of behaviour. Ground Engng,
1973, 6, 24–33.
10. Potts D.M. et al. Use of soil berms for temporary support of retaining walls.
Proc. Conf. on Retaining Structures. Institution of Civil Engineers, London,
1992, 440–447.
11. CIRIA. Water resisting basement construction. CIRIA, London, 1995, Report 139.
12. Report on basement and cut and cover construction. Draft. Institute of Structural
Engineers, London, 2002.
13. BS 8102. Protection of structures against water from the ground. British Standards
Institute, London, 1990.
14. Somerville G. The design life of concrete structures. Struct. Eng., 64A(2), 1986,
60–71.
15. Boikan A. Avoiding pitfalls and risk factors in below ground waterproofing.
Struct. Engr., 80, 2002, 5 June, No. 11, 16–18.
16. Zinn W.V. Economical design of deep basements. Civ. Engng Public Works Rev.,
1968, 63, Mar., 275–280.
17. Measor E. and Williams G. Features in the design and construction of the Shell
Centre. Proc. Instn Civ. Engrs, 1962, 21, Mar., 475–502.
18. Fenoux G.Y. Le réalisation fouilles en site urbain. Travaux, Parts 437 and 438,
1971, Aug.-Sept., 18–37.
19. Burland, J.B. and Hancock R.J. Underground car park at the House of
Commons. Struct. Engr, 1977, 55, Feb. 87–100.
20. Burland J. et al. Movements around excavations in London clay. Proc. 7th Euro.
Conf. S.M.F.E., Brighton, UK, 1979, Vol. 1, 13–29. British Geotechnical Society,
London, 1979.
21. Marchand S.P. A deep basement in Aldersgate Street, London, part 1: contrac-
tor’s design and planning. Proc. Instn Civ. Engrs, 1993, 93, Feb., 19–26.
22. Marchand S.P. A deep basement in Aldersgate Street, London, part 2: construc-
tion. Proc. Instn Civ. Engrs, 1993, 97, May, 67–76.
23. Slade R.E., Darling A. and Sharratt M. Redevelopment of Knightsbridge Crown
Court for Harrods. Struct. Engr, 80, 2002, 5 June, 21–27.
24. Fernie R. Movement and deep basement provision at Knightsbridge Crown
Court, Harrods, London. Conf. Response of Buildings to Excavation-induced
Ground Movements, July 2001. CIRIA, 2002.
25. Kenwright J., Dickson R.A. and Fernie R. Structural movement and ground set-
tlement control for a deep excavation within a historic building. Deep Foundations
Institute Conference, New York, 2000. DFI, Englewood Cliffs, New Jersey, 2000.
26. Boscardin M.D. and Cording E.G. Building response to excavation induced
settlement. ASCE J. Geotech. Eng., 1986, 115, No. 1, 1–21.
27. Fletcher M.S. et al. The ‘down’ of top down. Civ. Engng, 1988, 58, Mar., 58–61.
28. Lui J.Y.H. and Yau P.K.F. The performance of the deep basement for the
Dragon Centre. Proc. of Seminar on Instrumentation in Geotechnical Engineering,
Hong Kong Institution of Engineers, 183–201. Hong Kong, 1995.
29. Mitchell A., Izumi C., Bell B. and Brunton S. Semi top-down construction
method for Singapore MRT, NEL. Proc. Int. Conf. on Tunnels and Underground
Structures, Singapore, 2000. Balkema, Rotterdam, 2000.
30. Cole K.N. and Burland J.B. Observations of retaining wall movements asso-
ciated with a large excavation. Proc. of 5th Euro. Conf. S.M.F.E., Madrid, 1972.
31. Katzenbach R. and Quick H. A new concept for the excavation of deep building
pits in inner urban areas combining top/down method and piled raft foundation.
Proc. 7th Int. Conf. and Exhibition on Piling and Deep Foundations, Vienna, 1998,
15.17.1–15.17.3. DFI, Englewood Cliffs, New Jersey, 1998.
32. Mair R.J. Developments in geotechnical engineering research: application to
tunnels and foundations. Proc. Instn Civ. Engrs, 1993, 93, Feb., 27–41.
33. Yang D.S. Deep mixing. Proceedings of the Geo-Institute Conference, ASCE,
Logan, Utah, 1997, 130–150. ASCE, New York, 1997.
34. Porbaha A. et al. State of the art in construction aspects of deep mixing technol-
ogy. Ground Improvement 5, No. 3, 2001, 123–140.
35. Taki O. and Yang D.S. Soil-cement mixed technique, Geotechnical Engineering
Congress, ASCE New York. Geotechnical Special Publication 27, Vol. 1, 298–
309. ASCE, New York.
36. Bruce D.A. et al. Deep mixing: QA/QC and verification methods. Grouting–Soil
Improvement Geosystems including Reinforcement. Editor, Hans Rathmeyer.
11–22, Finnish Geotechnical Society, Helsinki, 2000.
37. BS 8081. Code of practice for ground anchorages. British Standards Institution,
London, 1989.
38. Littlejohn G.S. and Bruce D.A. Rock anchors: state-of-the-art. Foundation
Publications, Brentwood, 1977.
39. Barley A. Ten thousand anchorages in rock. Ground Engng, 1988, 21: No. 6,
Sept., 20–29; No. 7, Oct., 24–35; No. 8, Nov. 35–39.
40. Diaphragm walls. Icos, Milan, 1968.
41. Guillaud M. and Hamelin J.P. Innovations in diaphragm wall construction plant.
Proc. D.F.I. Conf., Nice, 2002, 3–8. Presses de l’Ecole Nationale des Ponts et
Chaussées, Paris, 2002.
42. DIN Standard 4126. Cast in-situ concrete diaphragm walls. Deutsches Institut für
Normung, Berlin, 2002.
43. Puller M.J. and Puller D.J. Developments in structural slurry walls. Proc. Conf.
Retaining Structures, Institution of Civil Engineers, London, 1992, 373–384.
44. Rogers W.F. Composition and properties of oil wall drilling fluids. Gulf Publishing,
Houston, 1967.
45. Beresford J.J. et al. Merits of polymeric fluids as support slurries. Proc. Conf.
Piling and Deep Foundations, London, 1989. DFI, Englewood Cliffs, New
Jersey, 1989.
46. Fisher F.A. Diaphragm wall projects. Proc. Conf. Diaphragm Walls and
Anchorages, Institution of Civil Engineers, London, 1971, 11–18.
47. Jardine R.J. et al. Studies of the influence of non-linear stress-strain characteris-
tics in soil-structure interaction. Geótechnique, 1986, 36, Sept. 377–396.
48. Reynaud P. and Riviere P. Mesure des pressions developpees dans une paroi
moulee en cours de betonnage. Bull. Liaison Lab., Ponts et Chausée, Paris,
1981, No. 113, 135–138.
49. Lings M.L. et al. The lateral pressure of wet diaphragm wall panels cast over
bentonite. Proc. Instn Civ. Engrs, 1994, 107, 163–172.
50. Gunn, M.J. et al. Finite element modelling of installation effects. Proc. Conf.
Retaining Structures, Institution of Civil Engineers, London, 1992, 46–55.
51. Sliwinski Z. and Fleming W.G. Practical consideration affecting the construction
of diaphragm walls. Proc. Conf. Diaphragm Walls and Anchorages, Institution of
Civil Engineers, London, 1975, 1–10.
52. BS 2004. Code of practice for foundations. British Standards Institution, London,
1972.
53. Specification for cast in place concrete diaphragm walling. Federation of Piling
Specialists, London, 1985.
54. BS 8102. Code of practice for the protection of structures against water from the
ground. British Standards Institution, London, 1990.
55. Puller M.J. Waterproofness of structural diaphragm walls. Proc. Instn Civ. Engrs,
Geotech. Engng, 1994, 107, Jan., 47–57.
56. Xanthakos P. Slurry walls. McGraw-Hill, New York, 1979.
57. Specification for highway works. Department of Transport, London, 1991.
58. EN 1538. Execution of special geotechnical works: diaphragm walls. British
Standards Institution, London, 1996.
59. Specification for piling and embedded retaining walls. Institution of Civil
Engineers, London, 1996.
60. Centre Scientifique et Technique de Batiment. DTU No. 14.1. Travaux de
cuvelage. CSTB, Paris, 1987.
61. Sherwood D.E. et al. Recent developments in secant bored pile wall construction.
Proc. Piling and Deep Foundations Conf., London, 1989, 211–219. DFI,
Englewood Cliffs, New Jersey, 1989.
62. CIRIA. Embedded retaining walls: guidance for economic design. Gaba A.R. et al.
CIRIA, London, 2002.
Bibliography BS EN 12715: Grouting. British Standards Institution, London, 2000.

BS EN 12716: Jet grouting. British Standards Institution, London, 2001.
BS EN 12063: Sheet piling. British Standards Institution, London, 1999.
Pr EN 14475: Reinforcement of fills. CEN, 2002.
BS EN 1538: Execution of special geotechnical work: Diaphragm walls. British
BS EN 1537: Execution of special geotechnical work: Ground anchors. British
BS EN 1536: Execution of special geotechnical work: Bored piles. British Standards
Institution, London, 1999.
Draft EN 14679: Execution of special geotechnical work: Deep mixing. British
Davies R. and Henkel D. Geotechnical problems associated with construction of
Chater Station. Proc. Conf. Mass Transportation in Asia, Hong Kong, 1980, 1–31.
Ikuta Y. et al. Application of the observational method to a deep basement excavated
using the top-down method. Geótechnique, 1994, 44, Dec., 655–664.
Ramaswarmy S.D. Soil anchored tieback system for supporting deep excavations.
J. Instn Engrs Singapore, 1975, 14, Dec., 10–33.
Ramaswarmy S.D. and Aziz M.A. Some experiences with ground anchors for
substructure construction in Singapore. Proc. Conf. Geotech., Singapore, 1980,
170–180.
Ramaswarmy S.D. and Yong K.Y. Some case studies on problems of substructures of
high rise building. Proc. Conf. Construction Practices in Geotech. Engng, Surat,
India, 1982, 255–260. Oxford and IBH Publishing Co., New Delhi, 1982.
Smoltczyk U. Editor. Geotechnical Engineering Handbook Vol. 2. Procedures. 2.5.
Ground anchors. Ostermayer H. and Barley T. Ernst and Son, Berlin, 2003.
Washbourne J. The three dimensional stability analysis of diaphragm wall excava-
tions. Ground Engng, 1984, 17, May, 24–39.
Yandzio E. and Biddle A. Steel intensive basements. Steel Construction Institute.
Ascot 2001.

8 Soil Movement Due To Deep Excavations PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

8 Soil Movement Due To Deep Excavations PDF

Uploaded by

Copyright:

Available Formats

Basement construction and design

Table 8.1 Engineering an

1 Explore and test subsoil

Soil nailing is in fact a soil-reinforcing technique and uses short tendons,

The application of soil nailing to retain excavated slopes at the periphery of

Fig. 8.5. Cantilever wall, West

with a free-standing wall of greater depth. Although the analysis showed a

Water-resisting A complete review of the methods of safeguarding basements against water

Table 8.2 Basic grades of water resistance11

From Table 1 of BS 8102: 1990 Abbreviated commentary given

combination of two (C þ A or C þ B); the types being as follows: A, structure

Progressive As basement excavations increase in depth, excavation methods have become

Fig. 8.6. Cross-section of

Peripheral walls propped by ﬂoor and raft

Fig. 8.7. Stages of basement

Fig. 8.8. Stages of basement

Fig. 8.9. Stages of basement

Fig. 8.10. Top-downwards

Fig. 8.11. Stages of car park

Fig. 8.12. Stages of basement

Fig. 8.14. Harrods,

Fig. 8.15. Harrods,

perimeter. In turn, using these deformation contours, the settlement behaviour

The basement construction proceeded as the basement ﬂoors were cast on

Fig. 8.17. Dragon Centre, Hong Kong. Geological section.

Fig. 8.18. Brittanic House, London. Construction sequence30

Semi top-down construction

Combination of bottom-up and top-down methods

Fig. 8.19. Main Tower, Frankfurt. Construction sequence31

prove worthwhile in granular soils. Where foundation loads from adjacent

King post wall

Support system Installation plant Distancea

Underpinning Conventional bulk excavation plant, e.g. Nil

Fig. 8.21. Vertical cross-section

Contiguous bored piling

Fig. 8.22. Typical joint details

Fig. 8.23. Conﬁguration of

oscillated under crowd to achieve penetration. The introduction of powerful

Fig. 8.24. Piling rigs used for

Fig. 8.25. CDSM piles:

Support for piled basement walls by ground and rock anchors

Fig. 8.26. Soil–cement wall for

. the basement construction programme

The location of adjacent substructures and services will determine the

Soldier pile tremie concrete method

Structural diaphragm walls

such as the minimum thickness of construction required by the diaphragm

Fig. 8.27. Alternative methods

Fig. 8.28. (c) (i) Tone

Fig. 8.29. Bauer City Cutter rig:

Excavation for diaphragm walls is currently made by a variety of rigs

Fig. 8.30. Bauer MBC 30 Trench

Fig. 8.31. Underpinning of

Fig. 8.32. Cutters for a modern

In a recent paper, Guillaud and Hamelin41 reviewed the development of

Fig. 8.33. (a) Modern

Fig. 8.34. Swivelling hydraulic

Developments to grabs are shown in the Soletanche–Bachy KS 3000 (see

Practical design and construction of diaphragm walls

Fig. 8.35. Surfaces of cut joint

shift). Such an arrangement gives output a rhythm. Panel size is there-

Fig. 8.36. Vertical cross-

Fig. 8.38. Reinforcement cage