A7 Design example

A7.1 WORKED EXAMPLE

A7.1.1 Introduction
This appendix presents a complete worked example for a typical three level basement construction
in an urban environment. For the purpose of consistency and to allow direct comparison, the same
worked example presented in Appendix A11 in Gaba et al (2003) has been used, although the parameter
selection and design approach will follow EC7 rather than the Gaba et al (2003) design method.

To demonstrate the advantages in economic design that can be realised by undertaking a thorough site
investigation and implementing a project under good site management and control, the same example is
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

analysed twice – once using information from a minimal site investigation, and a second time with a well-
designed appropriate site investigation and good on-site management during construction.

It should be noted that this example is not intended to give recommendations for site investigation scope
or the interpretation of parameters, as these are project specific. The intention is merely to demonstrate,
using real site investigation data, that significant economic advantage can be achieved in design and
construction by appointing skilled designers and investing in appropriate site investigation and good on-
site management during construction.

A7.1.2 Proposed basement

The geometry of the basement and the stratigraphy is the same as that presented in Appendix A11 of
Gaba et al (2003). The total depth of the basement excavation is 16 m and the construction sequence is
fully top-down with support provided to the wall by three levels of permanent slab (denoted P1, P2 and
P3). For each excavation stage, formation level will be 0.5 m below the soffit of the permanent slab.

In the permanent case, a 1.0 m thick basement slab will also support the wall. The basement is 80 m long
and 40 m wide.

The stratigraphy at the site comprises 4.0 m of made ground, overlying 7.0 m of gravel which in turn
overlies a clay stratum with a thickness of 50 m. Chalk underlies the clay.

The following is a summary of the main requirements and assumptions:

zz The retaining wall comprises a 900 mm hard/hard secant pile wall (the external casing diameter is
880 mm although the resultant pile diameter is 900 mm due to the presence of the outside driving
teeth, the auger diameter is 750 mm).
zz The retaining wall is a load bearing element in the short and long term with a permanent line load
of 150 kN/m and a variable line load of 50 kN/m at all construction stages.
zz Permanent drainage will be provided beneath the base slab to prevent long term build-up of water pressure.
zz An access road is to be located behind the wall offset at more than one metre. The road will be
subject to normal traffic loading, which will be modelled as a uniformly distributed load of 10 kPa
across the width of the model.

To resist long-term swelling pressures from the underlying clay, a grid of tension piles are installed
before the basement is excavated. These piles also have plunge columns installed to support the
permanent slabs as the excavation proceeds.

382 CIRIA, C760

 The permanent structure is shown in Figure A7.1.
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.1 Geometry of proposed basement

An outline of the proposed construction sequence is shown in Figure A7.2.

Figure A7.2 Proposed construction sequence

Guidance on embedded retaining wall design 383

 A7.1.3 Structural parameters
Retaining wall
The basement wall is required to provide a barrier to groundwater flow in the permanent condition.
To achieve a water-retaining structure, interlock between the primary and secondary piles is required
to a depth of at least 1.0 m below formation level, in this example 17.5 m below ground level including
an allowance of 0.5 m for unplanned excavation. The interlock is achieved by specifying installation
tolerances on position and verticality and by choosing the spacing between the piles in accordance with
the principles described in Chapter 3.

For the example presented here, the best achievable tolerances discussed in Chapter
3 have been adopted, ie 25 mm on position and 1 in 200 on verticality. To determine
the spacing between secondary piles, the nomogram shown in Figure A7.3 has been
prepared using the principles described in Appendix A3.

The proposed spacings between secondary piles as shown in Figure A7.3 is 1350 mm,
which results in a spacing between primary and secondary piles of 675 mm (1350/2).
The resultant cut into the primary pile (225 mm) leaves half of the primary pile
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

intact and so is acceptable.

The flexural stiffness of a line of secondary circular piles is calculated from the pile
diameters (Dsec) and pile spacing (s) as follows:

Isec = πDsec4/64.s (A7.1)

Figure A7.3 Nomogram for interlock of segmental cased piled walls (verticality 1:200, position tolerance ±25 mm)

Figure A7.4 900 mm diameter secondary pile through casing

384 CIRIA, C760

 The stiffness of the cased section of the secondary piles (Isec) where D = 900 mm and s = 1350 mm is:

Isec = 0.024 m4/m (A7.2)

The cutting of the primary piles of diameter Dprim by the secondary piles, results in a complex shape that
is not easily calculated. However, its contribution to the wall stiffness (Iprim) can be estimated by treating
the primary piles as an equivalent rectangle of width w and depth d given by the following equations:

(A7.3)

Iprim = 0.013 m4/m (A7.4)

So, the total stiffness per metre run of the cased section of the retaining wall is given by:

Iwall = Isec + Iprim = 0.037 m4/m (A7.5)

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Below the cased section of the wall (at a depth of 17.5 m below ground level), the wall stiffness is derived from
the secondary piles only with a reduced diameter assumed to be equal to the auger diameter (750 mm) at
1350 mm spacings (see Figure A7.6).

Figure A7.5 900 mm primary pile Figure A7.6 750 mm diameter secondary pile below
casing

Wall
Iwall = πD4/64.s
Iwall = 0.0115 m4/m

Young’s modulus of concrete, E0 = 30 MPa (C32/40 concrete, see Table 3.1 of EC2-1).
Wall EI during construction = 0.7 E0I
Wall EI in the long term = 0.5 E0I

A summary of the derived wall stiffness parameters is given in Table A7.1.

Table A7.1 Summary of flexural stiffness values for retaining wall

Wall section I (m4/m) Short-term EI (kNm2/m) Long-term EI (kNm2/m)

Cased section (ground level to 17.5 m) 0.037 7.77 × 105 5.55 × 105

Uncased section (17.5 m to toe level) 0.0115 2.42 × 105 1.73 × 105

Guidance on embedded retaining wall design 385

 As the beam elements used in the finite element mesh do not have any width, the weight of the
element should be equal to the difference between the weight of the soil that would be removed during
construction and the concrete used to construct the wall. Assuming a unit weight of 20 kN/m3 for soil
and 24 kN/m3 for concrete gives an effective weight of 4 kN/m3 for the beam elements. A typical Poisson’s
ratio of 0.15 for concrete is adopted.

Props
The props (slabs) are assumed to act axially across the width of the basement and have a stiffness (k) given by:
Prop stiffness, k = EA/leff s

where
leff is the effective length of the prop, which is defined as the point of zero axial movement. For the
symmetric basement presented here the effective length of the prop is half the basement width (20 m)
s is the prop (slab) spacing. The slabs are continuous across the basement and so the prop spacing is unity
Young’s modulus of the concrete E0 is taken to be 30 MPa (C32/40). During construction, E is taken to
be 0.7E0 and in the long term E is taken to be 0.5E0 .
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

The values of k used in the retaining wall analysis are summarised in Table A7.2.

Table A7.2 Summary of prop stiffness for design example

Slab thickness Stiffness k during construction Stiffness k in long term

500 mm (P1, P2 and P3) 525 000 kN/m/m 375 000 kN/m/m

The base slab is modelled with solid elements, which represent the actual depth of the structure and so
requires elastic material properties that are typical for concrete:

Young’s modulus of concrete, E0 = 30 MPa (C32/40)

During construction, E = 0.7E0 = 21 MPa
In the long term, E = 0.5E0 = 15 MPa
Poisson’s ratio, v = 0.15
Unit weight, γ = 24 kN/m3

Tension piles
The tension piles are assumed to be installed on a regular 10 m grid and each has a diameter of 1.5 m
and a length of 24 m below the soffit of the base slab. The piles are modelled as beam elements in the
numerical model and so their structural properties are smeared in the out-of-plane dimension to mimic
the spacing of the piles. The properties are summarised in Table A7.3.

Table A7.3 Tension pile structural properties for analysis

Axial stiffness Bending stiffness

Diameter Spacing Toe level
EA EI

1.5 m 10.0 m -40 m 4.95 × 106 kN/m 7.0 × 105 kNm2/m

A7.1.4 Analysis assumptions

General
As discussed in Section 4.2.3, using the finite element method for checking the ULS of a geotechnical
structure is not straightforward as there is no definitive, agreed technique for applying partial factors to
soil properties.

386 CIRIA, C760

 Simpson (2012) discussed the following two options:
zz Strategy 1 – applying partial factors to soil parameters at the start of the analysis to model the whole
construction sequence with design soil parameters.
zz Strategy 2 – starting the analysis with characteristic parameters and reducing the strength to the
design values at critical stages to check no ULS occurs.

For the reasons discussed in Section 4.2.3, Strategy 2 has been adopted together with the commercially
available geotechnical finite element software PLAXIS (see Websites).
Websites
PLAXIS: www.plaxis.nl
Geometry
For the purposes of undertaking the design of the embedded retaining wall, it is reasonable to assume
that the basement retaining wall at the centre of the excavation will behave approximately in a plane
strain manner with no significant movement (or strain) in the out-of-plane direction. So, this situation is
well suited to 2D plane strain analysis using a finite element program.

The chalk stratum is significantly stronger and stiffer than the overlying strata and can be reasonably
assumed to represent a rigid boundary (where movement is prevented in the vertical and horizontal
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

direction) for the finite element mesh. The lateral boundaries of the mesh should be placed at a sufficient
distance from the retaining wall so as to not affect its behaviour, but not so far away as to substantially
increase the computational time and memory requirements. The appropriate position for the lateral
boundaries is a function of the excavation depth and width. For the example presented here the lateral
boundary is set at a distance equal to 2.5 times the excavation width from the retaining wall. On the
horizontal boundaries, horizontal movements are restrained, but the mesh is free to move vertically.

Figure A7.7 Finite element mesh

To increase the accuracy of the analysis, a zone of mesh refinement close to the retaining wall has been
included. The result is a mesh with 2328 elements with an average size of 1.773 m. The elements in the
area of interest are less than 0.5 m.

Material behaviour
The clay is homogeneous with a coefficient of permeability, k < 10 -8 m/s. Based on experience
and the guidance provided in Chapter 5, it is assumed to behave in an undrained manner during
construction (excavation) and in a drained manner in the long term. As previously noted, the wall will
be impermeable and will provide a groundwater cut-off in the clay. There is no source of groundwater
recharge at or below excavation level, ie there are no water-bearing permeable horizons that could

Guidance on embedded retaining wall design 387

 provide drainage paths within the clay. This means there will only be excavation disturbance to a depth
of 0.5 m below formation level (see Chapter 5). No further softening of the clay is assumed during
excavation on the restraining side.

The behaviour of all strata will be represented by a linear elastic perfectly plastic soil model. The
short-term behaviour of the clay layer will be represented by a Tresca model. The short- and long-term
behaviour of the made ground and gravel layers, and the long-term behaviour of the clay stratum are
represented by a Mohr-Coulomb model. All structural elements are modelled as linear elastic.

Interface friction
The reduced interface friction between the wall and the ground has been modelled by introducing zero
thickness interface elements into the mesh along these boundaries. To avoid stress discontinuities and
numerical problems, the interface elements are extended one metre beyond the toe of the wall. The
frictional strength of the interface elements (δ) are as follows:
zz made ground and gravel layers and clay stratum (long term only): δ = ϕ
zz clay stratum (short term only): cw = 0.5cu .
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

It should be noted that while the wall is load bearing in the short and long term, there is no need to
modify the value or direction of the interface friction as this will automatically be dealt with by the finite
element analysis (see Chapters 4 and 7).

A7.2 SCENARIO 1: MINIMAL SITE INVESTIGATION

A7.2.1 Description of site investigation

In an attempt to save initial expenditure on the proposed basement, the client for the proposed
development tendered the site investigation to several contractors without a minimum scope. As the
contractors were in a competitive situation, the winning contractor had chosen to undertake a single
cable percussion borehole to the base of the clay stratum. Classification testing was undertaken on
disturbed samples from all layers. No installations were included in the borehole for determining the
groundwater level.

A7.2.2 Determination of characteristic parameters

Made ground
A few SPTs were undertaken in the made ground layer, although due to the highly variable nature of this
material, a conservative view has to be taken on its engineering properties. The borehole logs described
the material as predominately granular in nature (comprising reworked sands and gravels as well as
varying quantities of crushed concrete and bricks), so the following characteristic parameters were
ascribed based on experience and typical parameters suggested in Chapter 5:

Bulk unit Young’s Angle of shearing Effective Coefficient of Poisson’s

Stratum weight modulus resistance cohesion permeability ratio
(gb) (E′) (ϕcv ′/ϕpk ′) (c′) (k) (v′)

Made ground 18 5 MPa 25°/25° 0 kPa 1 × 10 -6 m/s 0.3

The in situ stress was calculated based on the Jaky formula:

K0 = 1 – sin ϕ = 0.58

388 CIRIA, C760

 Gravel deposit
In order to estimate the properties of the
gravel layer, the SPT N60 values recorded in
the single borehole have been used, using
the correlations presented in Chapter 5. The
recorded SPT N60 value for the gravel layer are
shown in Figure A7.8.

The particles are described in the borehole

log as being sub-angular to sub-rounded and
the particle size distribution curves result in a
uniformity coefficient of about 3.0. Due to the
limited data, a characteristic N60 value of 10 is
taken for the gravel layer, resulting in a density
index of about 40 per cent according to Table 5.4.
So from Section 5.5.4 and Table 5.14:
ϕ′cv,k = 30 + ϕ′ang,k + ϕ′PSD,k
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

ϕ′pk,k = 0 + ϕ′ang,k + ϕ′PSD,k + ϕ′dil,k

where
ϕ′cv,k is the characteristic constant volume
angle of shearing resistance
ϕ′pk,k is the characteristic peak angle of Figure A7.8 SPT N60 values in gravel layer
shearing resistance
ϕ′ang,k is the contribution from the angularity of the particles = 2°
ϕ′PSD,k is the contribution from the soil’s particle size distribution = 2°
ϕ′dil,k is the contribution from soil dilation = 2°

Based on this approximate method, the characteristic constant volume and peak angles of shearing
resistance are 34° and 36° respectively.

To estimate the drained Young’s modulus of the gravel layer, the relationship proposed by Stroud (1989)
as reported in Section 5.5.5 is used:
E′ = 2.0 N60 (MPa)

This gives a characteristic E′ value of 20 MPa for the gravel layer. Other design parameters for the gravel
layer, such as bulk unit weight, permeability and Poisson’s ratio were ascribed based on experience and
typical parameters suggested in Chapter 5:

Bulk unit Young’s Angle of shearing Effective Coefficient of Poisson’s

Stratum weight modulus resistance cohesion permeability ratio
(gb) (E′) (ϕcv ′/ϕpk ′) (c′) (k) (v′)

Gravel 19 20 MPa 34°/36° 0 kPa 1 × 10 -4 m/s 0.25

The in situ stress was calculated based on the Jaky formula:

K0 = 1 – sin ϕ = 0.41

Clay stratum
To characterise the clay stratum and estimate its strength and stiffness, the single borehole has provided
index testing and SPT N values. The plasticity index for the material was determined to be 45 to 50 per
cent. The plot of SPT N against elevation is shown in Figure A7.9.

Guidance on embedded retaining wall design 389

 The characteristic design line has been fitted
through the data points with the following
equation:
N = 10 + 1.45z

where z is the depth below the top of the clay

stratum.

Due to the limited data, the characteristic

design line is necessarily close to the lower
bound. Using the relationship proposed by
Stroud (1989) as reported in Section 5.5.4, the
undrained shear strength can be estimated using
the SPT N value by the following relationship:
cu = f1 N60

The value of f1 is dependent on the plasticity

index. With no triaxial testing to support the
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

correlation, a conservative value of f1 = 4.5 is

chosen for a plasticity index of around 50 per
cent (see Figure 5.9). This gives the following
characteristic undrained shear strength profile:
cu = 45 + 6.53 z
Figure A7.9 SPT N values in clay stratum

where z is the depth below the top of the clay stratum.

Based on experience relating to the back analysis of case studies in clay, the following relationship is
proposed between undrained shear strength and the Young’s modulus:
Eu = 1000cu = 45 + 6.53 MPa

where z is the depth below the top of the clay stratum

E′ = 750cu = 33.75 + 4.9 MPa

where z is the depth below the top of the clay layer

Other design parameters for the clay stratum, such as bulk unit weight, permeability and Poisson’s ratio
were ascribed based on experience and typical parameters suggested in Chapter 5:

Bulk unit Angle of shearing Effective Coefficient of Poisson’s

Stratum weight resistance cohesion permeability ratio
(gb) (ϕcv′) (c′) (k) (v′/vu)

Clay 20 25° 5 kPa 1 × 10 -9 m/s 0.25/0.5

Groundwater level
No groundwater monitoring installations were included in the single borehole and so no reliable
measurement of groundwater level was available. The driller’s logs indicated a water strike in the gravel
at a depth of 7.5 m below ground level, rising to 7.1 m below ground level over a duration of 20 minutes.
No further groundwater strikes were noted during the drilling of the borehole.

Due to the limited information relating to possible spatial and temporal variations in the groundwater
level, an appropriately conservative assumption has to be made with regard to the design value. Due
to the degree of uncertainty, and following the guidance provided in Section 7.3, a design level of two
metres below ground level has been judiciously chosen.

390 CIRIA, C760

 A7.2.3 Site control
No particular requirements are specified in the project-specific technical documentation or employer’s
requirements for special controls during construction and so the designer is required to assume
unplanned excavation at each excavation stage amounting to that stated in Section 5.8, ie the lesser of:
zz 0.5 m
zz 10 per cent of the total height retained for cantilever walls, or the height retained below the lowest
support level for propped or anchored walls.

A7.3 SCENARIO 2: APPROPRIATE WELL

CONSIDERED SITE INVESTIGATION

A7.3.1 Description of site investigation

Constructive dialogue between the client and the designer at the start of the project led to a reasonable
budget being allocated for an appropriate site investigation. The designer involved the client in
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

discussions regarding the scope of the investigation and how this early expenditure should be seen
as part of the project risk management as well as providing a better understanding of the ground
conditions and engineering behaviour of the ground.

The site investigation scope and specification was defined by the designer following the guidelines of
EC7 as set out in Chapter 5. The site investigation contract was competitively tendered based on the site
investigation specification and contract documents produced by the designer on behalf of the client. The
site investigation comprised:
zz four boreholes, including two that proved the depth of the clay stratum
zz in situ SPT testing in all boreholes
zz triaxial testing of thin-walled U100 clay samples
zz large shear box testing of gravel samples to directly assess angle of friction
zz classification and index testing of all strata
zz installation of standpipe piezometers in the gravel layer in three of the boreholes, including pressure
transducers and a data-logger to get continuous readings for an adequate period of time.

A representative from the designer’s organisation was on site for the duration of the site investigation
to ensure the technical objectives were achieved and the site investigation contractor followed the
requirements of the specification. By having a representative on site from the designer’s organisation, it
was possible to respond to events immediately, which minimised delays and the risk of missing important
information as well as getting a ‘hands on’ feel for the ground conditions as they were encountered for a
full understanding of how these should be accommodated in design.

A7.3.2 Determination of characteristic parameters

Made ground
Due to the inherent variability of the made ground layer, and the difficulty in undertaking
representative laboratory testing in this stratum, no enhancement to the engineering properties assumed
in Section A7.2.2 was possible. The engineering parameters required for the embedded retaining wall
analysis is given here:

Guidance on embedded retaining wall design 391

 Bulk unit Young’s Angle of shearing Effective Coefficient of Poisson’s
Stratum weight modulus resistance cohesion permeability ratio
(gb) (E′) (ϕcv ′/ϕpk ′) (c′) (k) (v′)

Made ground 18 5 MPa 25°/25° 0 kPa 1 × 10 -6 m/s 0.3

The in situ stress was calculated based on the Jaky formula:

K0 = 1 – sin ϕ = 0.58

Gravel deposit
The angle of shearing resistance for the gravel was assessed directly by undertaking a total of six large
shear box tests on representative samples of the deposit in accordance with the guidance provided in
Section 5.5.4 for coarse grained soils. The measured peak angle of shearing resistance was as follows:

Sample 1 2 3 4 5 6 Average

Angle of shearing resistance

43.7° 38.4° 42.9° 41.0° 40.6° 38.1° 40.8°
ϕ′pk
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

To augment the directly assessed angle of shearing resistance presented above, an estimate based on the
correlation proposed in Section 5.5.4 and Table 5.14, which relies on the SPT N60 value was undertaken.
The recorded SPT N60 value for the gravel layer are shown below in Figure A7.10. It should be noted that
N60 values are obtained from all four boreholes.

The particles are described in the borehole

logs as being sub-angular to sub-rounded and
the particle size distribution curves result in
a uniformity coefficient of about 3.0. Due to
the extra SPT data from the additional three
boreholes, a characteristic N60 value of 20 is
taken for the gravel layer, resulting in a density
index of about 55 per cent according to Table
5.4. From Section 5.5.4 and Table 5.14:
ϕ′cv,k = 30 + ϕ′ang,k + ϕ′PSD,k

ϕ′pk = 30 + ϕ′ang,k + ϕ′PSD,k + ϕ′dil,k

where
ϕ′cv,k is the characteristic constant volume
angle of shearing resistance
ϕ′pk,k is the characteristic peak angle of
shearing resistance
ϕ′ang,k is the contribution from the angularity
of the particles = 2°
ϕ′PSD,k is the contribution from the soils
particle size distribution = 2°
ϕ′dil,k is the contribution from soil dilation = 2°
Figure A7.10 SPT N60 values in gravel layer

Based on this approximate method, the characteristic constant volume and peak angles of shearing
resistance are 36° and 38° respectively.

Using the measured average value of peak angle of shearing resistance (40.8°) and the estimate based on
the approach presented in Section 5.5.4, a characteristic design angle of ϕpk of 39° was chosen.

392 CIRIA, C760

 To estimate the drained Young’s modulus of the gravel layer, the relationship proposed by Stroud (1989)
as reported in Section 5.5.5 is used:
E′ = 2.0 N60 (MPa)

This gives a characteristic E′ value of 40 MPa for the gravel layer. Other design parameters for the gravel
layer, such as bulk unit weight, permeability and Poisson’s ratio were ascribed based on experience and
typical parameters suggested in Chapter 5:

Bulk unit Young’s Angle of shearing Effective Coefficient of Poisson’s

Stratum weight modulus resistance cohesion permeability ratio
(gb) (E′) (ϕcv ′/ϕpk ′) (c′) (k) (v′)

Gravel 19 40 MPa 36°/39° 0 kPa 1 × 10 m/s

-4
0.25

The in situ stress was calculated based on the Jaky formula:

K0 = 1 – sin ϕ = 0.37

Clay stratum
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

The undrained shear strength of the clay layer has been determined by undertaking triaxial tests on
thin-walled U100 samples. The resultant undrained shear strength profile with depth from the four
boreholes is shown in Figure A7.11.

There is considerable scatter in the undrained shear strength data. To augment the undrained strength
data the SPT data can be used with the correlation proposed by Stroud (1989) as reported in Section
5.5.4. The SPT N60 data from the four boreholes is shown in Figure A7.12.

Figure A7.11 Undrained shear strength profile with depth Figure A7.12 SPT N60 values in clay stratum
from triaxial testing

Guidance on embedded retaining wall design 393

 The characteristic design line has been fitted through the data points with the following equation:
N = 15 + 1.715z

where z is the depth below the top of the clay stratum.

Due to the additional data, the characteristic design line can be representative of a cautious estimate
of the mean, rather than a lower bound as in the previous example shown in Figure A7.9. Using the
relationship proposed by Stroud (1989) as reported in Section 5.5.4, the undrained shear strength can
be estimated using the SPT N60 value by the following relationship:
cu = f1 N60

The value of f1 is dependent on the plasticity index. A value of f1 = 5.0 is chosen for the plasticity index
determined from the results of index testing on clay samples recovered from the site investigation. This
gives the following characteristic undrained shear strength profile:
cu = 75 + 8.575z

where z is the depth below the top of the clay stratum.

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

The proposed design line is compared to the undrained shear strength from triaxial tests and the
factored SPT N60 values (using an f1 value of 5.0) in Figure A7.13.

It can be seen that the proposed design line

represents a cautious estimate of the mean
through the SPT N60 data and the undrained
shear strength data from triaxial testing.

Based on experience relating to the back

analysis of case studies in clay, the following
relationship is proposed between undrained
shear strength and the Young’s modulus:
Eu = 1000cu = 75 + 8.575 MPa

where z is the depth below top of clay stratum

E′ = 750cu = 56.25 + 6.43 MPa

where z is the depth below top of clay stratum.

Other design parameters for the clay stratum,

such as bulk unit weight, permeability and
Poisson’s ratio were ascribed based on
experience and typical parameters suggested
in Chapter 5.

Figure A7.13 Undrained shear strengths from triaxial testing

and factored SPT

Bulk unit Angle of shearing Effective Coefficient of Poisson’s

Stratum weight resistance cohesion permeability ratio
(gb) (ϕcv ′/ϕpk ′) (c′) (k) (v′/vu)

Clay 20 25° 5 kPa 1 × 10 m/s

-9
0.25/0.5

394 CIRIA, C760

 Groundwater level
Standpipe piezometers were installed in three of the four boreholes. The response zones were located
close to the bottom of the gravel deposit where the groundwater level was expected, based on the
designer’s hydrogeological understanding of the site. An automatic pressure transducer was installed
in each of the boreholes and connected to a data-logger, which took readings every 15 minutes. The
pressure transducers were left in the boreholes for 14 days. The recorded groundwater level data are
shown in Figure A7.14.
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.14 Water level recorded in gravel layer over 14 day period

Over the 14 day period the recorded water level varied between 7.6 m and 8.05 m below ground level
with small fluctuations at each location. There is no clear diurnal cycle and no source of water identified
in the area of the wall that could artificially raise the groundwater level in the short term. So for the
construction stages of the analysis a design water level of seven metres below ground level has been
adopted. This is based on the highest measurements shown in Figure A7.14 plus an allowance of 0.5 m
(in accordance with guidance provided in Section 7.3). In the long-term condition, once the permanent
structure has been completed, the design with a water level of two metres below ground surface is
checked to account for uncertainty regarding the future use of the area surrounding the basement
during its design life.

A7.3.3 Site control

The earthworks and piling specification produced by the wall designer on behalf of the client stipulate
that extra caution shall be exercised and appropriately tight on-site management and demonstrable
control measures should be put in place to ensure that unplanned excavation does not occur. So
a practical minimum tolerance for over-excavation of 100 mm at each excavation stage has been
considered and allowed for in the analysis.

Guidance on embedded retaining wall design 395

 A7.4 DETERMINATION OF DESIGN PARAMETERS

A7.4.1 Soil parameters

The characteristic parameters for each of the strata have been derived in Sections A7.2.2 and A7.3.2 for
the minimal and appropriate well-considered site investigation conditions respectively.

To undertake the design of the embedded retaining wall in accordance with EC7 requirements and
in-line with the guidance provided in Chapter 7, the characteristic values (X k) are converted to design
parameters (Xd) by applying the appropriate partial factors (γM) for DA1C1 and DA1C2 calculations, as
stated in Chapter 5, such that:

(A7.7)

The strength of the coarse-grained layers (made ground and gravel deposit) and the clay stratum in the
long term have been assumed to be given in terms of effective strength parameters as follows:

Effective stress:
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

tanϕ′d = tanϕ′k /γϕ′

c′d = ck′/γc′

The short-term behaviour of the fine-grained clay stratum has been assumed to be given in terms of
total stress strength parameters as follows.

Total stress:
cu,d = cu,k /γcu

According to the UK NA to EC7-1, the partial factors given in Table 7.1 are adopted. For convenience,
these are restated in Table A7.4.

Table A7.4 Material partial factors

Partial factor DA1C1 DA1C2

γϕ′ 1.0 1.25

γc′ 1.0 1.25

γcu 1.0 1.4

For Scenario 1 (minimal site investigation) and Scenario 2 (appropriate well-considered site
investigation), the design parameters for the DA1C1 and DA1C2 calculations are given in Table A7.5.

396 CIRIA, C760

 Table A7.5 Summary of design parameters adopted in calculations

Scenario 1 Scenario 2
Parameter
DA1C1 DA1C2 DA1C1 DA1C2
gb 18 kN/m3 18 kN/m3

ϕ′ 25° 20.5° 25° 20.5°

Made ground

c′ 0 kPa 0 kPa 0 kPa 0 kPa

E′ 5 MPa 5 MPa 5 MPa 5 MPa

v′ 0.3 0.3

k 1 × 10 -6 m/s 1 × 10 -4 m/s

K0 0.58 0.65 0.58 0.65

gb 19 kN/m3 19 kN/m3

ϕ′ 36° 30.2° 39° 32.9°

c′ 0 kPa 0 kPa 0 kPa 0 kPa

Gravel

E′ 20 MPa 20 MPa 40 MPa 40 MPa

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

v′ 0.25 0.25

k 1 × 10 -4 m/s 1 × 10 -4 m/s

K0 0.41 0.50 0.37 0.46

gb 20 kN/m 3
20 kN/m 3

cu 45 + 6.53z kPa 32 + 4.66z kPa 75 + 8.575z kPa 53.6 + 6.13z kPa

Eu 45 + 6.53z MPa 45 + 6.53z MPa 75 + 8.575z MPa 75 + 8.575z MPa

vu 0.5 0.5 0.5 0.5

cw 22.5 + 3.27z kPa 16 + 2.34z kPa 37.5 + 4.29z kPa 26.8 + 3.06z kPa
Clay

ϕ′ 25° 20.5° 25° 20.5°

c′ 5 kPa 4 kPa 5 kPa 4 kPa

E′ 33.75 + 4.9z MPa 33.75 + 4.9z MPa 56.25 + 6.43z MPa 56.25 + 6.43z MPa

v′ 0.25 0.25

k 1 × 10 -9 m/s 1 × 10 -9 m/s

K0 1.0 1.0 1.0 1.0

The SLS calculation was undertaken using the DA1C1 soil design parameters as set out in Table A7.5.
No partial factors applied to unfavourable variable actions (loadings) and no allowance for unplanned
excavation and for softening was given, in-line with the guidance in Section 7.3.

A7.4.2 DESIGN WATER PRESSURES

As described in Sections A7.2.2 and A7.3.2, the design groundwater levels for Scenarios 1 and 2 are:
zz Scenario 1 – two metres below ground level due to the unreliable and insufficient available data,
both during construction and in the long term.
zz Scenario 2 – seven metres below ground level during construction and two metres below ground in
the long term.

For both cases, water pressures are assumed to be hydrostatic from the groundwater level.

Guidance on embedded retaining wall design 397

 A7.4.3 Design actions (surcharges and loads)
In accordance with the guidance in Section 7.3, the partial factors shown in Table A7.6 should be
applied to unfavourable variable actions (loading) acting on or behind the wall:

Table A7.6 Partial factors applied to unfavourable actions (surcharges)

Combination Variable Permanent

DA1C1 1.11* 1.0*

DA1C2 1.3 1.0

Note
* These values are modified to take account of the 1.35 partial factor that is applied to the effect of actions for DA1C1 as discussed in
Section 7.3.

The basement retaining wall is subject to a variable surcharge of 10 kPa at ground level on its retained
side. The partial factors summarised in Table A7.6 result in design actions (surcharge values) of:
zz 11 kPa for the DA1C1 calculation
zz 13 kPa for the DA1C2 calculation
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

zz 10 kPa for the SLS calculation.

The wall is also subject to 150 kN/m of permanent load and 50 kN/m of variable load applied vertically at
the top of the wall throughout the construction sequence.

Using the partial factors given in Table A7.6, the following design actions (vertical loads) are applied to
the wall at its top:
zz 205 kN/m for the DA1C1 calculation
zz 215 kN/m for the DA1C2 calculation
zz 200 kN/m for the SLS calculation.

A7.5 SET UP OF FINITE ELEMENT MODEL

A7.5.1 DA1C2
As discussed in Chapter 7, the DA1C2 analysis is undertaken first in order to determine the required
embedment depth for the wall. By a separate calculation, a minimum wall embedment depth of five
metres has been determined for the vertical load bearing requirements of the wall. So, the analysis will
be set up with five metre embedment and a check on lateral stability undertaken.

The stages of the PLAXIS DA1C2 analysis are summarised in Table A7.7.

398 CIRIA, C760

 Table A7.7 Stages for DA1C2 PLAXIS analysis

Behaviour of
Stage Description Comment
clay layer
Initialise stress with characteristic soil
1 Drained
parameters

2 Install wall and tension piles Undrained

Apply 13 kPa surcharge behind wall and 215 kN/m

3 Undrained
line load to top of wall
Unplanned excavation:
Excavate to 2.5 m depth plus allowance for
4 Undrained zz 0.25 m for Scenario 1 (2.75 m)
over-excavation
zz 0.1 m for Scenario 2 (2.6 m).

4a Apply partial factors to soil parameters Undrained

P1 slab modelled as fixed end anchor.

Install P1 slab at depth of 1.75 m and excavate Unplanned excavation:
5 Undrained
to 7.0 m plus allowance for over-excavation zz 0.5 m for Scenario 1 (7.5 m)
zz 0.1 m for Scenario 2 (7.1 m).
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

5a Apply partial factors to soil parameters Undrained

P2 slab modelled as fixed end anchor.

Unplanned excavation:
Install P2 slab at depth of 6.25 m and excavate
6 Undrained zz 0.5 m for Scenario 1 (12.0 m)
to 11.5 m plus allowance for over-excavation
zz 0.1 m for Scenario 2 (11.6 m).
Apply softening in clay to 0.5 m depth.

6a Apply partial factors to soil parameters Undrained

P3 slab modelled as fixed end anchor.

Unplanned excavation:
Install P3 slab at depth of 10.75 m and excavate
7 Undrained zz 0.5 m for Scenario 1 (16.5 m)
to 16.0 m plus allowance for over-excavation
zz 0.1 m for Scenario 2 (16.1 m).
Apply softening in clay to 0.5 m depth.

7a Apply partial factors to soil parameters Undrained

Unplanned excavated material replaced with clay.

Replace over-excavated material and construct
8 Undrained Base slab modelled with solid elements with
base slab
properties of concrete (linear elastic).

9 Switch clay parameters to drained Drained

Fix pore water pressure on boundary of mesh to
Consolidate for 100 years to dissipate excess represent source.
10 pore pressures in clay layer and reach steady Drained
state seepage under the wall Set zero flow boundary on base of mesh and
along the line of symmetry.
10a Apply partial factors to soil parameters Drained

11 Switch clay parameters to undrained Undrained

Check for water pressure 2 m below ground Raise groundwater level to check for future rise in
12 Undrained
level groundwater level.
12a Apply partial factors to soil parameters Undrained

Key

Highlighted stages are ULS checks at key stages

A7.5.2 DA1C1
As discussed in Chapter 7, the DA1C1 analysis is undertaken after the required wall embedment depth is
confirmed by the DA1C2 analysis described in Section A7.5.1.

Guidance on embedded retaining wall design 399

 The steps in the DA1C1 analysis are described in Table A7.8.

Table A7.8 Stages for DA1C1 PLAXIS analysis

Behaviour of
Stage Description Comment
clay layer
Initialise stress with characteristic soil
1 Drained
parameters
2 Install wall and tension piles Undrained
Apply 11 kPa surcharge behind wall and 205 kN/m
3 Undrained
line load to top of wall
Unplanned excavation:
Excavate to 2.5m depth plus allowance for
4 Undrained zz 0.25 m for Scenario 1 (2.75 m)
over-excavation
zz 0.1 m for Scenario 2 (2.6 m).
P1 slab modelled as fixed end anchor.
Install P1 slab at depth of 1.75 m and excavate Unplanned excavation:
5 Undrained
to 7.0 m plus allowance for over-excavation zz 0.5 m for Scenario 1 (7.5 m)
zz 0.1 m for Scenario 2 (7.1 m).
P2 slab modelled as fixed end anchor.
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Unplanned excavation:
Install P2 slab at depth of 6.25 m and excavate
6 Undrained zz 0.5 m for Scenario 1 (12.0 m)
to 11.5 m plus allowance for over-excavation
zz 0.1 m for Scenario 2 (11.6 m).
Apply softening in clay to 0.5 m depth.

P3 slab modelled as fixed end anchor.

Install P3 slab at depth of 10.75 m and Unplanned excavation:
7 excavate to 16.0 m plus allowance for over- Undrained zz 0.5 m for Scenario 1 (16.5 m)
excavation zz 0.1 m for Scenario 2 (16.1 m).
Apply softening in clay to 0.5 m depth.

Unplanned excavated material replaced with clay.

Replace over-excavated material and construct
8 Undrained Base slab modelled with solid elements with
base slab
properties of concrete (linear elastic).

9 Switch clay parameters to drained Drained

Fix pore water pressure on boundary of mesh to
Consolidate for 100 years to dissipate excess represent source.
10 pore pressures in clay layer and reach steady Drained
state seepage under the wall Set zero flow boundary on base of mesh and
along the line of symmetry.
11 Switch clay parameters to undrained Undrained
Raise groundwater level to check for future rise in
12 Check for water pressure 2 m below ground level Undrained
groundwater level.

The resulting effects of actions (wall bending moments, shear forces and prop forces) from the DA1C1
analysis are then multiplied by 1.35 to obtain the design effects of actions, as discussed in Chapter 7.

A7.5.3 SLS analysis

For the basement considered in this design example, the close proximity of sensitive buildings requires
the direct assessment of the serviceability state for the retaining wall as described in Section 7.3.2.

The steps in the serviceability analysis are described in Table A7.9.

400 CIRIA, C760

 Table A7.9 Stages for serviceability PLAXIS analysis

Behaviour of
Stage Description Comment
clay layer
Initialise stress with characteristic soil
1 Drained
parameters
2 Install wall and tension piles Undrained
Apply 10 kPa surcharge behind wall and 200 kN/m
3 Undrained
line load to top of wall
4 Excavate to 2.5 m depth Undrained No unplanned excavation.

Install P1 slab at depth of 1.75 m and excavate P1 slab modelled as fixed end anchor.
5 Undrained
to 7.0 m No unplanned excavation.

Install P2 slab at depth of 6.25 m and excavate P2 slab modelled as fixed end anchor.
6 Undrained
to 11.5 m No unplanned excavation.

P3 slab modelled as fixed end anchor.

Install P3 slab at depth of 10.75 m and
7 Undrained No unplanned excavation or allowance for
excavate to 16.0 m
softening.
Base slab modelled with solid elements with
8 Construct base slab Undrained
properties of concrete (linear elastic).
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

9 Switch clay parameters to drained Drained

Fix pore water pressure on boundary of mesh to
Consolidate for 100 years to dissipate excess represent source.
10 pore pressures in clay layer and reach steady Drained
state seepage under the wall Set zero flow boundary on base of mesh and
along the line of symmetry.
11 Switch clay parameters to undrained Undrained

The pertinent results from the SLS analysis are:

zz the bending moments and shear forces in the wall that are used in the assessment of crack widths
(if required, see Section 7.5.2)
zz serviceability prop forces for the design of props (see Section 8.1.6)
zz structural deflections (principally the wall, props and base slab)
zz movement of the ground behind the wall.

A7.6 RESULTS

A7.6.1 Scenario 1: minimal site investigation

The results from the Scenario 1 sets of analyses are summarised in this section. The computed effects
of actions (bending moments and shear forces) in the retaining wall for each construction stage for the
DA1C1 and DA1C2 analyses are shown in Figures A7.15 and A7.16 respectively.

The wall bending moment and shear force envelopes for the DA1C1 and DA1C2 analyses together with
the results of the SLS analyses are shown in Figure A7.17. As discussed in Chapter 7, the maximum
values shown by these analyses represent the values to be adopted in the structural design of the wall.

The maximum computed axial forces in each of the floor slabs are presented in Table A7.10.

Guidance on embedded retaining wall design 401

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.15 Scenario 1 – DA1C1 effects of actions (wall bending moments and shear forces)

Figure A7.16 Scenario 1 – DA1C2 effects of actions (wall bending moments and shear forces)

402 CIRIA, C760

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.17 Scenario 1 – Design effects of actions (wall bending moments and shear forces) from ULS and SLS analyses

Table A7.10 Scenario 1 – Summary of effects of actions (prop/slab forces)

Prop DA1C1 DA1C2 SLS

P1 295 kN/m 263 kN/m 202 kN/m

P2 737 kN/m 832 kN/m 506 kN/m

P3 1150 kN/m 1616 kN/m 831 kN/m

The computed horizontal deflection of the retaining wall at each construction stage in the SLS analysis,
and the associated settlement at ground level behind the retaining wall for the Scenario 1 SLS analysis,
are shown in Figure A7.18. For comparison, the upper bound movement line for high stiffness support
systems shown in Figure 6.15b is also shown.

Guidance on embedded retaining wall design 403

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.18 Scenario 1 – SLS wall deflections and ground surface settlement behind retaining wall

A7.6.2 Scenario 2: appropriate well considered site

investigation
The results from the Scenario 2 sets of analyses are summarised in this section. The computed effects
of actions (bending moments and shear forces) in the retaining wall for each construction stage for the
DA1C1 and DA1C2 analyses are shown in Figures A7.19 to A7.20 respectively.

The wall bending moment and shear force envelopes for the DA1C1 and DA1C2 analyses together with
the SLS analyses are shown in Figure A7.21. As discussed in Chapter 7, the maximum values shown by
these analyses represents the design values to be adopted in the structural design of the wall.

The maximum axial forces in each the floor slabs are presented in Table A7.11.

404 CIRIA, C760

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.19 Scenario 2 – DA1C1 effects of actions (wall bending moments and shear forces)

Figure A7.20 Scenario 2 – DA1C2 effects of actions (wall bending moments and shear forces)

Guidance on embedded retaining wall design 405

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.21 Scenario 2 – Design effects of actions (wall bending moments and shear forces) from ULS and SLS analyses

Table A7.11 Scenario 2 – Summary of effects of actions (prop/slab forces)

Prop DA1C1 DA1C2 SLS

P1 160 kN/m 136 kN/m 115 kN/m

P2 513 kN/m 441 kN/m 376 kN/m

P3 950 kN/m 794 kN/m 672 kN/m

Figure A7.22 shows the computed horizontal deflection of the retaining wall at each construction stage
in the SLS analysis, and the associated ground surface settlement behind the retaining wall for the
Scenario 2 SLS analysis. For comparison, the upper bound movement line for high stiffness support
systems in Figure 6.15b is also shown.

406 CIRIA, C760

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A7.22 Scenario 2 – SLS wall deflections and ground surface settlement behind retaining wall

A7.6.3 Comparison of Scenario 1 and 2

The input parameters for Scenario 1 and Scenario 2 differ due to the quality and scope of the site
investigation undertaken in preparation for the project, and also due to the quality of site control
exercised during construction. The maximum design bending moments and shear forces predicted in
the retaining wall and the maximum prop/slab forces are compared for Scenario 1 and Scenario 2 in
Table A7.12.

Table A7.12 Comparison of results from Scenario 1 and 2 analyses

Maximum wall Maximum Maximum computed

Maximum prop/ Maximum computed
Scenario bending moment wall shear ground surface settlement
slab force (kN/m) wall deflection (mm)
(kNm/m) force (kN/m) behind wall (mm)
1 -1200/1340 1050 1616 33 10

2 -770/350 673 950 20 5

There are significant potential benefits of undertaking an appropriate and well-considered site
investigation together with good site management and control during construction. This is evident
when shown in the comparison of structural effects of actions, computed wall deflections and associated
ground surface settlements presented in Table A7.12.

A7.6.4 Discussion of predicted movements

The different magnitude of wall and ground movement predicted by the analyses for Scenarios 1 and 2 can
be attributed to the significantly improved soil parameters adopted in Scenario 2. However, it is worth
noting the comparisons with the normalised case study data discussed in Chapter 6.

Guidance on embedded retaining wall design 407

 Wall displacements
As discussed in Section 6.1.2, a first estimate of the wall movement due to excavation can be made using
the normalised plots proposed by Clough et al (1989). This method requires the assessment of a system
stiffness, which is related to the flexural stiffness of the wall and the prop spacing:
System stiffness = EI/δwh4

where
EI is the Young’s modulus of the wall multiplied by its second moment of area (7.77 × 105 kNm2/m for
the cased section of the hard/hard secant wall)
γw is the unit weight of water (generally taken to be 10 kN/m3)
h is the average spacing between props (4.5 m for the proposed basement).

For the basement considered here, the resultant system stiffness is 189. The simplified FoS against basal
failure (Fb) given by Bjerrum and Eide (1956) is given by:

(A7.8)
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

where
cu,1 is the average undrained shear strength above formation level
cu,2 is the average undrained shear strength below formation level, within the postulated failure
mechanism (assumed to extend to a depth of (√2/2)B below formation, see Figure 6.2)
Nc is the bearing capacity factor, is a function of the ratio of depth to breadth (taken to be 6.5 for the
current example)
γ is the bulk unit weight of soil
H is the height of the excavation
B is the width of the excavation.

Using the undrained shear strength profile derived for Scenarios 1 and 2, the resultant FoS against
basal failure are 2.7 and 4.2 respectively. Using these values and the derived system stiffness, Figure 6.4
predicts maximum wall movements of 40 mm (0.25 per cent of excavation depth) and 33.6 mm (0.21 per
cent of the excavation depth) for Scenarios 1 and 2 respectively. These values compare to between 24
mm (0.15 per cent of excavation depth) and 32 mm (0.2 per cent of excavation depth) for walls with stiff
support systems presented by St John et al (1992) for excavations in London Clay.

Ground movement profile

As noted in Section 7.3.2, to make reliable predictions of ground movement with any degree of
confidence is extremely complex and typically requires sophisticated laboratory testing data, an analysis
that adopts an advanced non-linear soil model and quality reliable case study data.

The design example in this appendix has used finite element analysis with a simplified linear elastic
soil model with a Mohr-Coulomb or Tresca failure criteria. While this type of model may be suitable for
undertaking ULS design of an embedded retaining walls, the absence of a strain-dependant stiffness
model means it is unlikely to reliably predict the ground movements.

Linear elastic soil models with a single stiffness value can be calibrated using case study data to give a
reasonable match to the maximum ground movement (Grammatikopoulou et al, 2008). However, the
shape of the settlement trough and the influence zone will not be accurately represented. This can be
seen in Figures A7.15 and A7.19 where the ground movement profile from the PLAXIS analyses are
compared to the data from Chapter 6. The maximum settlement predicted for Scenario 1 and 2 by the
PLAXIS analysis is 10 mm and 5 mm respectively. The normalised case study data suggests that for a
16 metre deep excavation predominately in stiff clay would be around 12.5 mm. The data does not take
account of the subtle differences in the Scenario 1 and Scenario 2 examples, so it can be concluded that

408 CIRIA, C760

 the magnitude of the maximum movement is predicted with a reasonable degree of accuracy. The shape
and extent however of the settlement trough is not well represented, with the trough being too shallow
and too wide, extending beyond the influence zone of 3.5 times the excavation depth shown by the case
study data (56 m influence zone based on a 16 m deep excavation).

It should be noted that if the predicted ground movement profile from a finite element analysis, which
assumes linear elasticity is to be used to estimate building damage, the incorrect settlement profile
may have significant consequences as the slope of the ground and the induced distortion of any nearby
building is likely to be underestimated. Conversely, the assumption that any nearby building will distort
with the ground is likely to be a conservative assumption. If the ability of the building stiffness to
modify the ground settlement profile can be taken into account, significant reductions in the predicted
settlement and distortion can be made as discussed in Section 6.4.

A7.6.5 Comparison with results from Gaba et al (2003) study

Parameters
Although the basement considered in the design example in this publication is the same as that used in
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Gaba et al (2003), direct comparison with the results presented here cannot be made. This is because the
soil parameters derived to illustrate the difference between Scenario 1 and Scenario 2 are not the same
as those used in Gaba et al (2003). A comparison of the critical soil parameters is given in Table A7.13.
Soil parameters that are significantly different between the two examples are highlighted.

Table A7.13 Comparison of characteristic parameters used in current study and parameters used in Gaba et al
(2003) design example

EC7 characteristic Gaba et al (2003)

Parameter Moderately
Scenario 1 Scenario 2 Worst credible Serviceability
conservative
γb 18 kN/m3 18 kN/m3 18 kN/m3 18 kN/m3 18 kN/m3

ϕ′ 25° 25° 25° 23° 25°

Made ground

c′ 0 kPa 0 kPa 0 kPa 0 kPa 0 kPa

E′ 5 MPa 5 MPa 5 MPa 5 MPa 10 MPa

v′ 0.3 0.3 0.3 0.3 0.3

k 1 × 10 m/s
-6
1 × 10 m/s
-6
1 × 10 m/s
-6
1 × 10 m/s
-6
1 × 10 -6 m/s

K0 0.58 0.58 0.6 0.6 0.6

γb 19 kN/3 19 kN/m3 19 kN/m3 19 kN/m3 19 kN/m3

ϕ′ 36° 39° 35° 32° 35°

c′ 0 kPa 0 kPa 0 kPa 0 kPa 0 kPa

Gravel

E′ 20 MPa 40 MPa 25 MPa 25 MPa 50 MPa

v′ 0.25 0.25 0.25 0.25 0.25

k 1 × 10 -4 m/s 1 × 10 -4 m/s 1 × 10 -4 m/s 1 × 10 -4 m/s 1 × 10 -4 m/s

K0 0.41 0.37 0.4 0.4 0.4

γb 20 kN/m3 20 kN/m3 20 kN/m3 20 kN/m3 20 kN/m3

cu 45 + 6.53z kPa 75 + 8.575z kPa 75 + 5z kPa N/A 75 + 5z kPa

Eu 45 + 6.53z MPa 75 + 8.575z MPa 37.5 + 2.5z MPa N/A 75 + 5z MPa

Clay

vu 0.5 0.5 0.5 N/A 0.5

cw 22.5 + 3.27z kPa 37.5 + 4.29z kPa 37.5 + 2.5z kPa N/A 37.5 + 2.5z kPa

ϕ′ 25° 25° 25° 21° 25°

Guidance on embedded retaining wall design 409

 Table A7.13 Comparison of characteristic parameters used in current study and parameters used in Gaba et al
(2003) design example (contd)

EC7 characteristic Gaba et al (2003)

Parameter Moderately
Scenario 1 Scenario 2 Worst credible Serviceability
conservative
c′ 5 kPa 5 kPa 5 kPa 0 kPa 5 kPa

E′ 33.75 + 4.9z MPa 56.25 + 6.43z MPa 30 + 2z MPa 30 + 2z MPa 60 + 4z MPa

Clay

v′ 0.25 0.25 0.25 0.25 0.25

k 1 × 10 -9 m/s 1 × 10 -9 m/s 1 × 10 -9 m/s 1 × 10 -9 m/s 1 × 10 -9 m/s

K0 1.0 1.0 1.0 1.0 1.0

In the design example by Gaba et al (2003), the groundwater was assumed to be able to rise to a level of two
metres below groundwater levels, which is the same conservative assumption taken in Scenario 1. However,
overall, the soil parameters assumed in Gaba et al (2003) are closer to the Scenario 2 assumptions.

Structural design
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

The maximum design values for prop forces, bending moments and shear forces for Scenarios 1 and 2
(which follow the principles of EC7) and the Gaba et al (2003) design example are given in Table A7.14.

Table A7.14 Comparison of structural design values with those reported in Gaba et al

Parameter Scenario 1 Scenario 2 Gaba et al (2003)

Maximum prop force in P1 295 kN/m 160 kN/m 222 kN/m

Maximum prop force in P2 832 kN/m 513 kN/m 686 kN/m

Maximum prop force in P3 1616 kN/m 950 kN/m 1860 kN/m

Maximum wall shear force 1050 kN/m 673 kN/m 978 kN/m

Maximum positive wall bending moment 1340 kNm/m 350 kNm/m 543 kNm/m

Maximum negative wall bending moment -1200 kNm/m -770 kNm/m -2048 kNm/m

The maximum values presented in Table A7.14 show that for the prop forces and shear force in the
wall, the values predicted by Scenario 1 are similar to those presented for the Gaba et al (2003) case.
The predicted values for Scenario 2 are significantly less, mostly due to the lower water level assumed
for this case.

There is however a significant difference between the maximum predicted bending moments between
the current study and the Gaba et al (2003) example. Examination of the original FREW files has shown
that the assumption of zero friction on the back of the wall in the made ground and the gravel layer (due
to the application of the vertical load on the wall as shown in Box 7.1) was having a significant effect on
the predicted bending moment. By using a FE analysis for the current example, this assumption was no
longer required with the resultant reduction in maximum bending moment.

Serviceability
In the Gaba et al (2003) design example, the wall movement was assessed directly by the FREW analysis
and the ground settlement behind the wall was estimated using the semi-empirical method described
in Figure 6.17. The maximum wall movement and the settlement behind the wall from the analyses
presented in the current publication are compared to the results of the serviceability analyses presented
in Gaba et al (2003) and in Figures A7.23 and A7.24.

The wall movements shown in Figure A7.23 illustrated that the wall movements predicted by the
serviceability analyses carried out as part of the work by Gaba et al (2003) and this publication produce very

410 CIRIA, C760

 similar deflected shapes with only the maximum
value being affected by the stiffness assumptions. It
is interesting to note the similarity in the deflected
wall shape predicted by the two independent
analytical methods (ie FREW and PLAXIS).

Figure A7.24 shows the maximum settlement trough

from the Gaba et al (2003) analysis, which used a
semi-empirical method based on wall movement
(Figure 6.15) and the Scenarios 1 and 2 analyses
based on the serviceability PLAXIS analyses. For
comparison, the empirical data from Figure 6.15b
for a 16 m deep excavation is also shown. It is
important to note that as discussed previously, the
linear elastic model used in the PLAXIS analysis
is significantly over estimating the influence zone
of the basement wall and so under estimating the
possible distortion to any structure (when compared
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

to the semi-empirical and case study data).

The comparison of ground settlement troughs

presented in Figure A7.24 highlights the importance
of using well-documented case studies to validate
movement predictions that come from more
sophisticated numerical analyses. If a designer were
to rely on the settlement trough predicted by the
linear-elastic perfectly-plastic PLAXIS analysis, in
isolation without reference to case study data, the Figure A7.23 Comparison of wall movements from
serviceability analyses from Gaba et al (2003) and
distortion experienced by neighbouring structures Scenarios 1 and 2
could be significantly underestimated, while the
zone influence could be significantly over estimated.

Figure A7.24 Comparison of settlement behind wall from serviceability analyses from Gaba et al (2003) and Scenarios
1 and 2 and case study data

Guidance on embedded retaining wall design 411

 A8 Distributed prop load (DPL)
method

A8.1 GENERAL
Twine and Roscoe (1999) gives a detailed description of the principles and application of the distributed
prop load (DPL) method for the determination of temporary prop loads. They provide design guidance
based on extensive field measurements of prop loads for flexible and stiff walls and for the range of
ground conditions commonly encountered in the UK.

The DPL method should only be applied to excavations of similar depth and plan geometry and in
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

ground conditions that are comparable to the case study data considered in Twine and Roscoe (1999).
The designer should establish this before carrying out calculations of prop load using the DPL method.

Temporary prop loads calculated using the DPL method provide a conservative estimate of the prop
loads to be expected in the field in normal circumstances. Temporary prop loads calculated in this way:
zz require the designer to ensure that wall embedment is adequate for the application of the DPL
zz do not consider the possibility of groundwater pressures rising to most unfavourable levels over the
duration of the temporary works
zz do not explicitly check for the possibility of progressive failure – the designer should carry out a risk
assessment and/or separate calculations to show that progressive failure will not occur (Section 7.3.3).

Temporary prop loads calculated using the DPL method should be compared with SLS prop loads
calculated as previously described in accordance with the general principles applicable to temporary
works design (Section 5.11). Where the SLS prop loads calculated in this way differ significantly from
those derived from the DPL method, the designer should carefully investigate and understand the
reasons for the difference and adopt appropriate values in the design of the temporary props. Reliability
of, and control over, groundwater pressures is often particularly significant in this regard. The design of
the temporary works should be demonstrably robust (Section 5.11).

If, as a consequence of this evaluation, DPL-derived prop loads are adopted in preference to those
corresponding to the calculated values, the designer should ensure the following:
1 The wall will satisfy its design and performance criteria (Section 2.4) and will remain stable at all
times over the duration of the temporary works under the application of the adopted prop loads.
2 The maximum wall bending moments and shear forces calculated in point (1) lie within the envelope
of ULS effect of actions adopted in the structural design of the wall.
3 The maximum wall bending moments and shear forces calculated under the application of DPL
derived prop loads, assuming un-factored soil design parameters, lie within the envelope of SLS
effect of actions adopted in checking compliance with crack width criteria for reinforced concrete
walls and allowable stress criteria for steel sheet pile walls, if applicable.

The essential features of the DPL method are described in Section A8.2, adapted from Twine and
Roscoe (1999) where the treatment of additional loading from temperature effects has been adapted
from that presented to allow for the degree of restraint of the prop.

412 CIRIA, C760

 A8.2 DPL DESIGN METHOD
Details are provided in Twine and Roscoe (1999). The DPL method for calculating prop loads for
propped temporary excavations is based on the back analysis of field measurements of prop loads
relating to 81 case studies, of which 60 are for flexible walls (steel sheet pile, king post walls) and 21 are
for stiff walls (contiguous, secant, diaphragm walls). The case study data relate to excavations ranging in
depth from 4 m to 27 m, typically 5 m to 15 m in soft and firm clays (soil class A, Table A8.1), 10 m to
15 m in stiff and very stiff clays (soil class B, Table A8.1) and 10 m to 20 m in coarse-grained soils (soil
class C, Table A8.1).

The method for determining prop loads is shown in Figure A8.1.

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A8.1 Method for calculating the DPL

The data are classified based on the type of ground retained by the excavation, see Table A8.1.

Table A8.1 Classification of ground types

Soil class Description

A Normally and slightly overconsolidated clay soils (soft to firm clays).

B Heavily overconsolidated clay soils (stiff and very stiff clays).

C Coarse-grained soils.

D Mixed soils (walls retaining both fine-grained and coarse-grained soils).

The classes in Table A8.1 are subdivided according to wall stiffness, ie flexible (F) walls and stiff (S)
walls. Flexible walls retaining soft clay soil (Class AF) have been further subdivided according to base
stability conditions. Class C soils have been subdivided into dry and submerged. Figure A8.2 shows the
characteristic prop load diagrams for Class A, B and C soils.

Guidance on embedded retaining wall design 413

Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A8.2 Characteristic DPL diagrams for Class A, Class B and Class C soils

414 CIRIA, C760

 A8.3 ALTERNATIVE METHODS
Alternative methods for the assessment of prop loads for multi-propped walls based on modified earth
pressure distributions are given in DDGT (2014). Katsigiannis et al (2015) carried out parametric
studies for multi-propped walls in stiff clays, comparing the maximum prop forces predicted
between Twine and Roscoe (1999), DDGT (2014) recommendations and finite element method (FEM)
predictions for various earth pressure at-rest coefficients K0 . The results of these comparisons
show that the German recommendations give prop loads in good agreement with the numerical
analysis results (see Figures A8.3 and A8.4). The difference between the DPL and FEM predictions
is particularly evident in the upper prop level where the DPL method uses a uniform distribution of
earth pressure with depth resulting in higher prop loads.
Licensed copy:Waterman Partnership, 03/04/2017, Uncontrolled Copy, © CIRIA

Figure A8.3 Geometry of example modelled in stiff clay (deep excavation with five levels of propping)

Figure A8.4 Comparison of maximum prop forces for the geometry shown in Figure A8.3 (after Katsigiannis et al, 2015)

Guidance on embedded retaining wall design 415