Vous êtes sur la page 1sur 13

DFI Journal - The Journal of the Deep Foundations

Institute

ISSN: 1937-5247 (Print) 1937-5255 (Online) Journal homepage: http://www.tandfonline.com/loi/ydfi20

Slope Stabilisation of a Section of the Thirlmere


Aqueduct by the Use of Spaced Micropiles and
Permeation Grouting
David Preece
To cite this article: David Preece (2011) Slope Stabilisation of a Section of the Thirlmere
Aqueduct by the Use of Spaced Micropiles and Permeation Grouting, DFI Journal - The Journal
of the Deep Foundations Institute, 5:1, 27-38
To link to this article: http://dx.doi.org/10.1179/dfi.2011.003

Published online: 24 Jan 2014.

Submit your article to this journal

Article views: 14

View related articles

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=ydfi20
Download by: [University of California, San Diego]

Date: 05 March 2016, At: 10:06

Slope Stabilisation of a Section of the Thirlmere Aqueduct


by the Use of Spaced Micropiles and Permeation Grouting
David Preece, BEng, MSc, CEng, MICE; Bachy Soletanche Burscough UK; tel 01704 895686;
david.preece@bacsol.co.uk
ABSTRACT

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

The Thirlmere aqueduct was constructed over 110 years ago to provide water from the Lake District (in
the north-west of England) to the city of Manchester some 60 miles (100 km) away. The aqueduct remains,
to this day, a key piece of infrastructure in the distribution of raw water in the north-west of England.
A short section of the aqueduct was identified as being in poor condition during routine inspections
by the owner. Subsequent investigations attributed the conduit damage to the movement of unstable
ground over the conduit. Bachy Soletanche Ltd was subsequently appointed to carry out the design
and construction of the slope stabilisation works and to undertake repairs inside the damaged
conduit. The works were undertaken on an existing slope which, in places stands at a gradient of up
to 40 to the horizontal. Strict limitations on plant and equipment were imposed by the client (United
Utilities) to ensure that slope and conduit loading and associated vibrations were kept to an absolute
minimum during construction. An intensive instrumentation system was installed to monitor surface
movements, ground movements at depth, conduit movement and strains across existing planes of
weaknesses within the conduit. The solution comprised the installation of spaced piles up-slope and
down-slope of the conduit. Restrictions were imposed on the minimum stand-off distance between
the outside face of the conduit and the piling works so targeted permeation grouting was undertaken
to improve the connection between the conduit and the down-slope piles. The presence of existing
sub conduit groundwater flows was suspected and the permeation grouting was focused in areas to
minimize disruption to the groundwater regime.
The paper presents the basis of the design and highlights the combined construction and instrumentation
methods adopted to minimise damage to this valuable asset.

INTRODUCTION
The Thirlmere Aqueduct supplies water
from the Lake District to Manchester. It was
constructed in the 19th century and consists
of single line tunnel, conduit sections and large
diameter cast iron and steel pipe sections. Pipe
siphons convey the aqueduct across numerous
valleys, Hill (1896).
Nab Scar conduit is a 50m (164 ft) length
of aqueduct conduit linking two aqueduct
rock tunnels. It is situated on a steep hillside
above Rydal Village, north of Ambleside. A
photograph taken during the construction of
a rock tunnel portal in the area of Nab Scar is
shown in Fig. 1 and view of the existing Nab
Scar hillside is shown in Fig. 2.
The purpose of the project was to stabilise the
conduit section of the Thirlmere Aqueduct at
Nab Scar and to undertake internal concrete
repairs during a 4-week Outage period during
the autumn of 2009.

[FIG. 1] Rock tunnel portal during construction near Nab Scar


- Source United Utilities web site

[FIG. 2] Looking towards Nab Scar, with the approximate


extent of the conduit section of the Thirlmere Aqueduct shown

DFI JOURNAL Vol. 5 No. 1 June 2011 [27]

HISTORY

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

The 50m (164 ft) long section of Thirlmere


Aqueduct at Nab Scar extends, in conduit,
between two sections of rock tunnel. This
section of conduit runs in a general southeasterly direction, before turning left through
an angle of 55, to enter the Nab Scar Tunnel
portal. A desk study suggested that the
conduit was originally built across a natural
gulley in the hillside, and was then covered
over with rock fill on completion. Gallagher
et al (2009). Extracts from the original
construction drawings are included in the desk
study which has been reproduced below in
Figs. 3 and 4.

[FIG. 3] Historical drawing of the proposed route of the


Thirlmere Aqueduct in the area of the site with Nab Scar
conduit location shown

Downward movement of fill


Existing slope ground level
Approximately 5m of backfilled
tunnel waste rock

Approximately 3m
of weathered rock

2170mm
Approximately 5m of backfilled
tunnel waste rock
2170mm

Approximately 3m
of weathered rock
Volcanic Tuff rock

[FIG. 5] Typical cross section through mass concrete conduit at


Nab Scar

A series of cracks were evident within the


conduit structure and the most marked of
them comprised of a spiral crack (Fig. 6), which
ran from the wall/floor joint on the outside of
the bend (right hand side) in the conduit and
around the barrel of the arch, terminating at
the tunnel portal. The crack was first noted
in 2005 during the first outage in the current
programme. Remote reading movement gauges
were installed across significant cracks to
provide early warning of any future movement.
The gauges are read frequently and, to date,
have not shown any further movement.
However, the removal of the manganese
deposits, as part of previous internal repair
works, revealed other cracks, which were not
visible to the 2005 inspectors, particularly an
opening of the joint between the base and the
wall on the down-slope side of the conduit.
The orientation of the cracking, together with
the opening of the base wall joint, indicates that
the conduit structure has undergone a torsional
movement, with the down-slope section twisting
outwards relative to the up-slope section.

[FIG. 4] Historical drawing of the proposed route showing the


low area assumed to be the natural gully at the site

The conduit is a mass unreinforced concrete


structure with continuous vertical walls and
arched roof and an infill concrete floor slab.
There is no structural connection between the
walls of the conduit and floor slab and the
construction joint is able to rotate and displace.
See Fig. 5.
[28] DFI JOURNAL Vol. 5 No. 1 June 2011

[FIG. 6] Photograph taken inside of the conduit showing spiral


cracking

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

During subsequent outages, a series of surveys


and intrusive investigations has established that
the floor of the conduit is founded directly on a
platform of rock and so the potential for global
movement of the conduit was discounted.
Instead, it is considered that the displacement
is limited to a section of outer wall and roof on
the outside of the spiral crack.
It is understood that the conduit at Nab Scar
was constructed and then buried beneath waste
materials from the tunnelling operations. The
tunnelling waste primarily comprises cobbles
and boulders of tuff, within a secondary matrix
of sand and gravel. It appears that the waste
material was end tipped within the gully and so
stands at or close to its natural angle of repose.
There is no evidence to suggest that the material
was compacted and there are clear indications
that that waste on the down-slope side of the
conduit has slumped downhill, potentially
removing lateral support from the downhill side
of conduit structure. Waste material on the
upslope side of the conduit also shows signs
of movement and the passage of material over
the conduit would drag across the roof of the
structure. Accordingly, it was concluded in the
report by Gallagher et al (2009) that the torsional
displacement of the conduit was caused by
external forces due to movement of the overlying
fill material, as shown schematically on Fig. 5.

The following restrictions were imposed by


United Utilities to protect the conduit from any
further damage:

No direct loading of conduit during


construction;

Construction methods had to adopt


techniques that imposed minimal vibrations
on to the conduit, the use of percussive
drilling techniques and down the hole
hammers were expressly prohibited;

All slope loading due to the weight of access


platforms, construction plant and materials
to be tied back to stable rock outcrops;

Conduit and slope movements to be


monitored through out the works with
works ceasing if trigger levels reached.

GROUND CONDITIONS
The ground conditions comprised a variable
thickness of Made Ground of up to 5m
thickness over completely weathered rock (up
to 3m (10 ft) thickness) over moderately strong
Volcanic Tuff. See Fig. 5.

In order to protect the conduit from the effects


of further movement of the tunnelling waste,
reinstatement of lateral support was required
to the down-slope wall of the conduit. A
secondary form of protection was also required
to shield the upslope side of the conduit
from lateral pressure due to movement of the
upslope body of ground/tunnelling waste. The
investigations had revealed the presence of
significant voids within the tunnelling waste
material, and the proposed solution required
these voids to be reduced to improve the
stability of this material in the area immediately
downhill of the conduit.

Groundwater levels were monitored within


standpipe piezometers installed up-slope and
down-slope of the conduit. Groundwater was
not encountered during the drilling of any of the
site investigation boreholes but was measured at
depths of 0.72m to 1.1m (2.4 ft to 3.6 ft) below
existing ground levels in piezometers installed
within the Made Ground. Piezometers sealed
within the rock recorded groundwater levels in
the range of 2.1m to 7.6m (6.9 ft to 24.9 ft) below
ground levels. The difference in groundwater
levels recorded within the shallow and deep
piezometers indicates the upper groundwater
to be perched within the Made Ground. The
groundwater monitoring was undertaken over a
period of less than one year. Observations of the
standpipes during inclement weather did not show
groundwater rising to the surface of the slope.
The apparent groundwater gradients within the
Made Ground and rock were broadly parallel to the
slope and their levels varied between the ranges
stated above depending on preceding rainfall.

RESTRICTIONS

SELECTED SOLUTION

Restrictions on drilling close to the conduit


were imposed to minimise the potential for
further damaging the structure. No works were
allowed to be undertaken within 2m (6.6 ft) of
the outside face of the conduit.

The selected solution was to use drilled spaced


piles uphill and down-hill of the conduit to
stabilise the hillside and conduit. Spaced piles
comprise of separate single piles or single
groups of piles that are not connected together

REQUIREMENTS

DFI JOURNAL Vol. 5 No. 1 June 2011 [29]

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

with a continuous pile cap/ground beam. The


spaced piles rely on soil arching for the transfer
of load on to them.

The micropiles were provided in pairs, tied


at their head with a pile cap. The pairs of
micropiles were termed A-frames. The
A-frames act as a passive restraint system
whose resistance is mobilized following minor
downward movement of the ground up-slope
of the piled A-frames. A typical section
through of the design is shown in Fig. 8.

Direct structural connection to the conduit


was not permitted and so permeation
grouting via a grid of tube--manchette (TaM)
grouting boreholes were installed between
the conduit and the down-slope spaced piles.
The permeation grouting would back-fill
open voids and provide an indirect structural
connection between down-slope piles and
the conduit. The permeation grouting was
designed to be installed in isolated blocks to
enable the passage of existing groundwater
flows to continue unhindered. The works could
be undertaken with light weight drilling rigs
working off a temporary scaffold platform,
which itself, was anchored back to existing
stable rock outcrops. See Fig. 7.

The following steps were undertaken in the


design:

[FIG. 7] Light-weight drilling operating of temporary scaffold


platform

The magnitude of the required hillside


stabilising force, to be provided by the
A-frames, was determined by undertaking
static slope stability analysis to give a
minimum factor of safety against slope
failure of 1.3. Two methods of determining
this force were considered; circular
slip surface analysis using commercial
software, and a simple infinite slope
method calculation.

The required hillside stabilising force was


applied to the A-frame arrangement to
determine individual pile loads within each
A-frame. Two methods of determining
the individual pile loads were considered;
simple structural frame model based on the
triangle of forces method, and a pile group
analysis using commercially available
software.

A soil movement calculation was


undertaken to check that the A-frames
were spaced sufficiently close together so
that soil arching would develop up-slope
of the A-frames and that soil would not
subsequently flow between A-frames.

DESIGN OF SPACED PILES


General
The selected slope stabilisation design
comprised the use of two rows of
spaced piles. One row of spaced piles
being provided up-slope of the conduit
and one row of spaced piles being
provided down-slope of the conduit.
Each group of spaced piles comprised
of two micropiles as follows:

one vertical micropile acting in


compression
one raked micropile acting in
tension (raked 40 degrees from
horizontal into the slope).

175mOD
Low flow
cement grout

170mOD
Grout injection
pipes and
permeation
grout zone

165mOD

Rock

postulated
'infinite' slope
failure plane

160mOD
3m
155mOD

6.5m
rock
socket

varies

150mOD

40

6.5m

40

3m

145mOD
0m

10m

20m

30m

[FIG. 8] Typical section through slope stabilisation works

[30] DFI JOURNAL Vol. 5 No. 1 June 2011

2.5m

40m

50m

The infinite slope equation shown above


can be re-written for finite slopes to allow
incorporation of a restoring force term. The
resulting equation is shown in Fig. 10.

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

Slope stability

The slope, pre-stabilisation, was considered


to have a low factor of safety against failure
(approaching unity). Back analysis calculations
The micropile A-frames were designed to
of the slope where undertaken considering
provide the calculated slope stabilising force
a factor of safety of 1 to back calculate soil
and to thereby increase the slope factor of
parameters for the previously determined
safety to at least 1.3. The up-slope A-frames
geometry and groundwater conditions. The
were designed to provide the required restoring
back-calculated parameters are provided in
force to achieve a post stabilisation slope
the calculations shown in Fig. 9 and Fig. 10.
stability factor of safety of 1.3. The down slope
Slope stability calculations were undertaken
row of A-frames resists significantly smaller
using the slope stability software package
forces than the up slope row only the section
SLOPE developed by Geosolve. Circular and
of slope between the two rows of A-frames.
non-circular slip failure mechanisms were
Nevertheless identical piles, albeit at slightly
considered. Following the determination of
wider spacings, were designed within the
the back-calculated soil parameters, critical
slope failure slips were repeated with
inputted restoring forces to establish
The equation in Figure 10 can be re-written to include a restoring force H term. Where H
is the restoring force parallel to the slope in kN per m run of slope.
the required restoring resistance to
The inclusion of this term requires the slope length, l, being considered to be included in
provide an engineered slope with a
the calculation of the factor of safety, F.
factor of safety of at least 1.3. The
The factor of safety, F, against translation slip failure is given by the equation below for a
required up-hill restoring force was
'specific' slope with seepage parallel to ground surface and where ground water is not at
calculated as 250kN per m (17.13
the surface of the slope.
2
2
(sat.l.z.cos .m.z.l.cos ).tan + H
F=
c.l
+
kips/ft) length of slope, acting in a
sat.l.z.sin.cos
sat.l.z.sin.cos
direction parallel to the slope.
Infinite slope stability calculations were
also undertaken to provide comparison
with the SLOPE software results. These
calculations provided good agreement
with the computed factors of safety for
the existing condition (FoS = 1.0) and
the strengthened condition (FoS =1.3).
See Figs. 9 and 10 below. A sensitivity
analysis was undertaken to consider
different slip depths and different
angles of frictional for the Made
Ground/weathered bedrock.

sat

saturated soil unit weight

w
unit weight of water
angle of friction

slope angle to horizontal

cohesion
c
depth to slip surface, z
depth to groundwater surface, zw
slope length, l
Restoring force parallel to slope, H

For, F =

w
unit weight of water
angle of friction

slope angle to horizontal

cohesion
c
depth to slip surface, z
depth to groundwater surface, zw

20 kN/m
3
10 kN/m
45
34
0
2.5
1

kPa
m
m

45

2.5

m = (z-zw)/z
0.6
m = fraction from 0 to 1 (=1 for ground water at slope surface)
F=

1.04

kPa
m
m
m
kN per m

0.785 radians
0.593 radians

1.32, H = 250kN/m of slope

[FIG. 10] Finite slope equation re-written to include slope restoring force term

where:
sat

45
34
0
2.5
1
38
250

m = (z-zw)/z
0.6
m = fraction from 0 to 1 (=1 for ground water at slope surface)

The factor of safety, F, against translation slip failure is given by the equation below for an
infinite slope with seepage parallel to ground surface and where ground water is not at
the surface of slope
F = [c/(sat.z.sin.cos)]+[(sat-w.m)/sat].[tan/tan]

saturated soil unit weight

20 kN/m
3
10 kN/m

down slope row, to provide down-slope


support to the conduit. The calculation
of pile loads and associated pile
deflections are discussed below.
The slope restoring force of 250kN per m
(17.13 kip/ft) run of slope, acting parallel
to the slope, was determined as the force
required to increase the slopes factor of
safety against slope failure to 1.3. This
restoring force was to be provided by
A-frames spaced at 2.0m (6.6 ft) centres
and were therefore designed to carry safe
working loads of 500kN (112.4 kips). The
second line of raking piles down-slope of

[FIG. 9] Infinite slope calculation

DFI JOURNAL Vol. 5 No. 1 June 2011 [31]

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

the conduit were identical to the up-slope pairs


other than they were spaced at slightly wider
centres (2.25m or 7.4 ft) along the slope.
The lower line of A-frames provided support
to the conduit and are expected to be lightly
loaded once the upslope piles have been
installed and resistance mobilised. One of
the reasons for maintaining a similar pile
arrangement down-slope of the conduit was
in order that stability could be built into the
slope during the progression of the works.
United Utilities would not have permitted the
erection of temporary scaffolding and operation
of equipment up-slope of the conduit without
stabilisation measures being in place downslope of the conduit first. In the long-term,
the full safe working micropile load is only
expected to be mobilised on the up-slope pairs
of piles.
Calculation of pile loads and deflections
The calculation of A-frame loads and
deflections were calculated by two methods:

Structural frame model using the


triangle of forces method

Elastic continuum model using


commercially available software Piglet.

Both of these methods are reported as suitable


methods by Elson (1984).
Structural frame model
The structural frame model is based on
resolving the horizontal force into two
components, producing an axial compressive
force in the far pile and tensile force in the near
pile. The restraint offered by the pile cap is
ignored, and the magnitude of each component
is obtained from a simple triangle of forces
as shown on Fig. 11. Estimates of deflections
were made on the basis of elastic compression/
extension of the pile section resisting the
compression/tension forces.
The calculated lateral deflection for these
piles, under the full safe working load, was
estimated to be of the order of 10mm (0.4 in)
using the structural frame model. This model
estimates deflections by ignoring any resistance
contribution from the soil. Lateral deflections
were estimated by resolving the individual
structural members deflections under the
action of the calculated loads assuming that
the members behave elastically as a frame, with
fixity in the rock.
[32] DFI JOURNAL Vol. 5 No. 1 June 2011

Elastic continuum model


The Piglet software, published by Randolph,
was used to undertake the elastic continuum
model calculations. The pile group analyses to
determine pile loads and group performance
were undertaken using Piglet version 5.1.
The solution for laterally loaded pile groups
adopted in the program is one developed by
Randolph (1981a) by curve fitting the results
of finite element analyses of laterally loaded
piles embedded in elastic soil. It was found
that, for piles which behave flexibly under
lateral load, simple power law relationships
could be developed giving the lateral
deflection, u, and the rotation, , of the pile at
the soil surface, in terms of the pile stiffness
and the soil properties. The relationships
are similar in form to those arising from
considering the soil as a Winkler material
characterised by a coefficient of subgrade
reaction (e.g. Reese and Matlock, 1956;
Matlock and Reese, 1960). Horizontal loads
and bending moments at the head of each pile
may then be calculated from lateral deflection,
u, and the rotation, , by equations that take
account of the appropriate shear modulus
of the soil, pile stiffness and the critical
pile length (depth to which the pile deforms
appreciably). Please refer to Randolph (2004)
for a full account of the analysis method.
The pile is taken to be fixed within the pile
cap such that rotation of the pile cap is not
permitted. This is in accordance with the
guidance in Chapter 5 of the US Army Corps
of Engineers (2008) document with criteria for
Plane
Lateral H

Compression C

Tension T

Lateral H

[FIG. 11] Triangle of forces used to estimate micropile axial loads

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

the assessment of pile head fixity based on


the penetration of the pile into the cap and
the relative stiffness between the piles and the
pile cap. The software then calculates the pile
cap fixing moments to ensure zero rotation of
the cap.

Comparison of Structural frame model results


and elastic continuum model

The Piglet input parameter values are given below:

The deflections and axial pile loads calculated


using Piglet (Elastic continuum model) are
consistent with those calculated by the
Structural frame model.

Pile layout: based on two pairs of


A-framesat 2.0m (6.6 ft) centres for the upslope piles and 2.25m (7.4 ft) centres for the
down-slope piles.

Shear modulus stiffness parameters have


been applied as follows:
a. Surface shear modulus = 0kN/m2
(0 kips/ft2)
b. Shear modulus gradient = 4000kN/m2
per m (25.5 kips/ft2/ft)
c.

Shear modulus below base =


150,000kN/m2 (3,133 kips/ft2)

Scope of problem = 3, i.e. all 6 degrees of


freedom.

Poissons ratio for soils and rock = 0.25.

Free standing length of pile is set at


1.0m (3.3 ft) (i.e. allowing for a nominal
disturbance of ground).

Piles Youngs Modulus = 2.8*107 kN/m2


(4,061 ksi) assuming C32/40 grout.

Piles are modelled as fixed to the pile cap.

Pile shaft diameter = 225mm (8.9 in), pile


base diameter = 196mm (7.72 in).

The load is applied to a rigid pile cap


that is fixed against rotation but able to
translate laterally. This is a reasonable
model given the fixity of the pile within the
pile cap and the relative stiffness between
pile and pile cap.

A summary of the pile loads and performance


outputs from structural frame model and
elastic continuum model are shown below for
the up-slope piles.

Arching check & development of A-frame load


in response to hillside movement
The discrete micropile A-frames were spaced
to retain slope material up-slope of the A-frames.
The A-frames rely on soil arching, following
minor ground movements, to transfer the slope
loads on to them. The spacing of the A-frames is
consistent with the recommendations of Carder
and Temporal (2000).
The downward movement of the hillside will
load the A-frames. The forces on the A-frame
pile caps and micropiles were estimated based
on Ito and Matsui (1975) and Ito et al (1981)
as reported in TRL Report 466: A review of the
use of spaced piles to stabilise embankment
and cutting slopes. This method allows the
calculation of lateral applied forces per
unit length of pile or pile cap. A number of
equations and design charts were developed for
different soil strengths which enabled the force
acting on the pile to be determined. For example
the equation for the lateral force (p) acting on a
pile per unit thickness of a layer is as follows:

[TABLE 1] Comparison of structural frame and elastic continuum model for 2m spacing (500kN slope
resistance)
Tension pile
force

Compression pile
force

Calculated
resultant lateral
deflection

Individual pile
bending moment
/ & shear force

500kN Slope restoring force:


Structural frame model

541kN

628kN

10mm

Not available

500kN Slope restoring force:


elastic continuum model

531kN

635kN

11.1mm

27kNm / 20kN

Analysis ref

DFI JOURNAL Vol. 5 No. 1 June 2011 [33]

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

Slope movement forces on pile caps


The calculations initially consider the forces
applied to the pile caps caused by the movement
of the surrounding ground. For the up-slope
piles, where 600mm (24 in) wide pile caps are
positioned at 2.0m (6.6 ft) centres the calculated
lateral force per metre length (i.e. equal to
force on the pile cap for 1m (3.3 ft) deep piles
caps is 477.4kN/m (32.7 kips/ft). This force is
resisted by the A-frame piles and the associated
tension and compression leg forces, which can
be determined by resolution of forces in the
direction of the piles, are 623kN and 400kN (140
kips and 90 kips) respectively. Calculations were
also undertaken based on the application of this
force on to the Piglet model. The Piglet input
parameters are consistent with those stated above
other than the applied lateral load of 477.4 x 2 =
954.8kN (315 kips) is used. The resulting outputs
from the Piglet calculation are given below.
[TABLE 2] Peak pile deflections and loads for
up-slope piles
Deflection / Load

Value

Lateral pile group


deflection

10.5mm

Compression

422 kN in each
vertical pile

Tension

623 kN in each
raking pile

Lateral

18 kN in all piles

Individual pile moment

25 kNm in all piles

Pile cap moment

663 kNm

The Piglet results are similar to those calculated


using the structural frame model.
The compressive force of 422kN (95 kips) is
less than the compression design force of
635kN (143 kips) and is therefore satisfactory.
However, the 623kN (140 kips) tension force
is greater than the safe working design force
of 541kN (122 kips). The micropile design was
therefore based on the greater tension force
calculated from this mechanism.
Slope movement forces on piles
The unit pile length slope movement force
was determined for 225mm (8.9 in) diameter
piles at 2.0m (6.56 ft) centres. The calculated
[34] DFI JOURNAL Vol. 5 No. 1 June 2011

force is 30.9kN per m length of pile (2.12 kips/


ft). This lateral force can be converted into an
equivalent pile bending moment for piles of
different free lengths. The free length would
be that length of the pile that is situated within
the moving ground. The bending moments have
been determined on the basis that the pile acts
as a simply supported vertical beam. This is
reasonable given that the pile is fixed at its head
(by the pile cap and accompanying pile) and at
its toe by embedment into the stable rock. A
free length of 1.5m (5 ft) is considered which
represent slope slip depths of 2.5m (8.2 ft) (i.e.
1.5m + pile cap depth of 1m). The following pile
bending moments were calculated:
[TABLE 3] Equivalent Pile Bending Moment
Pile free length (m)

Equivalent pile bending


moment (kNm)

1.5

8.7

Structural pile design


The piles structural capacity was designed
to resist the maximum forces and bending
moments from all of the mechanisms
considered. The calculated forces from the
different mechanism are shown in the Table 4
below. The worst case forces and moments are
highlighted in bold.
Moment capacity of piles
A circular hollow section (CHS) was included
within the upper section of the pile to provide
enhanced resistance to bending and shear. The
selected section was a 168.3mm (6.625 in) outside
diameter with 10mm (0.4 in) wall thickness. A
40mm (1,575in) (950/1050) full length central bar
was installed within the vertical piles and a 50mm
(1.969 in) (500/600) full length central bar was
installed within the raking piles.
The CHS was designed to be installed in
short sections to allow it to be handed on site
without the need for lifting gear. The short
sections were connected together by virtue
of threads cut within the ends of each length.
The moment capacity of the CHS at this
threaded joint was therefore reduced from
the continuous section value calculated above.
The reduced moment capacity, allowing for
the threads, was determined by considering
the same circular section but with half the wall
thickness (i.e. 5mm or 0.2 in).

Axial capacity of section

determined the rock socket length to support the


maximum compression load of 627kN (141 kips)
(6.5m or 21.3 ft rock socket required) and the
maximum tension load of 623kN (140 kips) (6.5m
or 21.3 ft rock socket required).

Checks on the structural axial capacity of


both piles within the A-frame were carried
out to ensure the piles can safely support the
designed tension and compression loads. These
checks demonstrated that the proposed piles
were suitable. The bar diameter of the tension
pile was increased to control crack widths
within the micropiles.

MONITORING AND INSTRUMENTATION


An instrumentation scheme was developed
to monitor the movements on and within the
conduit and to measure movements of the
hillside during the construction phase. Postconstruction monitoring of the conduit was
planned to run for a minimum of two years to
monitor the effectiveness of the works. The
instrumentation layout is shown on Fig. 13 and
described below:

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

Pile cap
A pile cap design was undertaken to ensure
that the piles are adequately anchored into
the pile cap and that the piles cap can resist
the calculated moments. The resulting pile cap
detail is shown in Fig. 12.

Hillside surface monitoring


A fully automated total station surveying
system was employed to continuously monitor
hillside ground movements in real-time. The
system is a fully automated, optical monitoring
system comprising of a motorised total station
that observes prismatic targets fixed to the
hillside. The monitoring prisms were installed
on a square matrix with prisms at 5.0m (16.4 ft)
centres across the hillside.

4No. 400mm dia


B16 helical at
55mm centres &
3No. 350mm dia
B16 helical at
55mm centres

7No. 350mm dia


B16 helical at
55mm centres

Main reinforcement omitted for


clarity comprises:
4No. B25 bars top and bottom
5No. B25 bars to each side
face
B12 link bars @ 130mm c/c

The total station unit was piloted by a single


industrial PC which operates the instrument,
logs the results, undertakes the survey
calculations, performs quality control checks
and then internally stores each prisms
observed co-ordinates. The PC automatically
downloads the acquired survey co-ordinate
data and imports the information into the
monitoring database for subsequent analysis.

[FIG. 12] Pile cap detail

Rock socket length calculation


The intact rock is described as moderately strong
volcanic Tuff. Rock sockets are to be drilled using
open hole rotary techniques and an ultimate
rock shaft capacity of 500kPa (72.6 psi) was
considered suitable in such rock. Calculations

[TABLE 4] Summary of peak pile forces and moments to be resisted (based on up-slope pile frames only
which are the most critical case).
Analysis mechanism

Tension pile
force

Compression pile
force

Individual pile bending


moment / and shear force

Pile
cap moment*

Slope restoring force:


Structural frame model

541kN

628kN

Slope restoring force:


elastic continuum model

531kN

635kN

27kNm / 20kN

784kNm

Slope movement induced


forces on pile caps: elastic
continuum model

623kN

422kN

25kNm / 18kN

663kNm

Slope movement induced


forces on piles

8.7kNm / 23kN

* the pile cap moment is fixing moment calculated by the Piglet program to ensure zero rotation at the head of the
piles in the fixed head condition.
DFI JOURNAL Vol. 5 No. 1 June 2011 [35]

External conduit monitoring

Instrumentation database

Five survey targets were fixed to the external


surface of the conduits roof structure to enable
movement of the structure to be monitored
whist the conduit is in operation.

All of the instrumentation data was stored


on the database. The database enables rapid
interrogation and manipulation of the data to
provide useful information. The database was
programmed with pre-determined trigger levels
for the different instruments. A warning beacon
was established on site to warn site workers of
movement and automated SMS messages were
sent to members of the team when trigger levels
were reached.

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

Hillside ground monitoring


Five inclinometers were installed across the
hillside. The inclinometer boreholes were
between 4.0m (13.1 ft) and 6.0m (19.7 ft) deep
and were socketed into competent rock by
a minimum length of 1.5m (5 ft) to provide
a suitable point of fixity. ABS inclinometer
casing, 70mm (2.75 in) in diameter was fixed
within the borehole using cement grout.
Manual baseline readings were taken before
connecting the in-place-inclinometers to the
site data logger. The in-place-inclinometer
sensors automatically recorded lateral
displacement of the slope at hourly intervals
and reported the data to the database.
10m

170mOD

Key

175mOD

Hillside surface
monitoring point
Automated
surveying total
station
Vertical in place
inclinometer

150mOD
Ground surface
contour
165mOD

Buried conduit
Rock
Outcrop

155mOD

160mOD

Strain gauges
attached inside the
conduit across
exisitng cracks
External conduit
monitoring point

150mOD

145mOD

[FIG. 13] Instrumentation layout plan

Internal monitoring of the conduit


Ten strain gauges were fitted within the
conduit across existing cracks within the
conduit. The gauges were bonded to the inside
of the conduit and surrounded in a stainless
steel shield. The space inside the shield was
filled with silicone sealant. The automatic
strain gauges recorded stain levels on an
hourly basis and feed the data back to the
instrumentation database.

[36] DFI JOURNAL Vol. 5 No. 1 June 2011

THE LAYOUT OF STABILISATION


WORKS & EFFECTIVENESS
The resulting layout of the stabilisation works is
shown in Fig. 14 below. The plan shows some of
the piles on the left hand side converging towards
the toe of the piles. Pile interaction in these areas
was considered, and the pile vertical position
was adjusted to ensure a minimum pile centre to
centre spacing of three times the pile diameter.
The lateral extent of the works was delineated by
the rock out crop on the right-hand side of the
slope and by a rise in rock level below ground
surface on the left hand side. The lower third of
the conduit is contained within a rock cut channel
just to the left hand side of the piling works.
Fig. 15 shows the results from the ten strain
gauges installed within the aqueduct conduit. A
number of the strain gauges recorded progressive
movement following their installation. However, it
is clear that the rate of increase in strain reduced
as the stabilising works were installed. Key
construction milestones are labelled on the figure
and it can be seen that the strain level across
existing cracks essentially halted following the
installation of the down slope piles, construction
of the associated pile caps and completion of
permeation grouting between the conduit and
these caps.
A reduced monitoring programme was
implemented following the completion of the
main site works. The strain gauges inside the
conduit and survey stations (extended above
ground level) are currently being read at a bimonthly frequency with readings set to continue
for a minimum period of three years following
completion of the main site works. Post site
readings are gathered manually on site to reduce
the costs associated with maintaining a power
supply and telemetry at the site. The manual
strain gauge readings are shown on Fig. 15 and

manual monitoring - post site works

1900

spurious data

gauge knocked
inside conduit

1400

reference gauge
away from works

900
400
-100
-600

-1100
17.3.10 19.4.10

-1600

18.6.10

20.8.10

15.10.10

-2100

15.12.10

25.2.11

gauge SG3

-2600

19/09/09

20/11/09

21/01/10

24/03/10

25/05/10

26/07/10

26/09/10

27/11/10

28/01/11

Date
Start on site

completion of
down slope piles
& perm. grouting

completion of up
slope piles

manual monitoring dates

The calculated vector movements


obtained from the monitoring results for [FIG. 15] Strain gauge results, measured inside the conduit, across existing
cracks
the survey points attached to the roof
of the conduit are shown in Fig. 16. The
survey
0.0040
points
vector movement fluctuates between 0mm and
fixed to
0.0030
conduit:
3.2mm (1.26 in). The survey measurements are
1
0.0020
taken using a standard total station surveying
2
3
0.0010
unit and the magnitude of the results are
4
5
considered to be similar to the accuracy of such
0.0000
30/01/10
02/04/10
03/06/10
04/08/10
05/10/10
06/12/10
06/02/11
date
equipment.
recorded movement [m]

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

The post-site monitoring strain gauge


results show greater scatter than those
measured using the automated system
during the site works. However, with
the exception of a single strain gauge
(SG3) the post-site strain gauge results
are within the magnitude of the results
measured during the site works. The
maximum recorded strain gauge value at
SG3 is 2500 microstrain which represents
a movement of 0.3mm (0.012 in) over
gauge length of 120mm (4.72 in).

automatic monitoring

Micro strain over 120mm

manual conduit survey measurements are


shown on Fig. 16.

The conduit is considered to be suitably stable


on the basis of the post-site results obtained
to date. The results also show that the works
comply with the performance specification
movement limit of 5mm (0.2 in).

[FIG. 14] Stabilisation works layout

[FIG. 16] Post site works survey results for points fixed to the
roof of the conduit and extended above ground level

CLOSING REMARKS & CONCLUSIONS


The site works were originally planned to be
conducted during a four week outage period
in October 2009 followed by another outage in
October 2010. However, following:

successful conclusion of the internal


conduit repair works;

excellent progress on the downslope that


provided increasing stability to the conduit;

completion of detailed monitoring of


internal and external installations during
the 2009 outage;

The client (United Utilities), their engineer


(MWH) and the contractor (Bachy Soletanche)
determined and agreed that works could be
confidently completed post-outage with the
aqueduct in full service and removing the need
for a subsequent visit and the cost and health &
safety risks involved in re-establishing the site.
This decision was assisted by some very delicate
piling operations in a very exposed area with a
significant proportion of the works undertaken
during the harsh weather conditions of winter
2009. The works were completed in February
2010 eight months ahead of the final outage in
DFI JOURNAL Vol. 5 No. 1 June 2011 [37]

October 2010 and also saving about a third of


a million pounds. This also added additional
security to the project as snagging work could
be undertaken later that year. However, an
internal inspection undertaken in October 2010,
determined that no further work was necessary.

Downloaded by [University of California, San Diego] at 10:06 05 March 2016

ACKNOWLEDGEMENTS
Thank you to our client, United Utilities for
letting us undertake this work for them and
subsequently allowing us to share the results
with the wider community by allowing us to
publish our work. Thanks to United Utilities
engineer, MWH, for working closely with us
during the works, and thanks to the dedicated
site team who completed this challenging
project in this challenging environment. Specific
mention goes to Ruth Webster, Andy Tucker,
Paul Doyle and Graham Cassidy.

REFERENCES
1. Carder, D. R. and Temporal, J. (2000). TRL
Report 466. A review of the use of spaced
piles to stabilise embankments and cutting
slopes.
2. Elson, W.K. (1984). Design of laterally-loaded
piles, CIRIA Report 103
3. Gallagher, E., Kitchen, R., Eddleston, M. and
Kavanagh, P. (2009). United Utilities Water
plc, Thirlmere Aqueduct, Nab Scar Conduit
Geotechnical Summary Report June 2009.
Unpublished document.

[38] DFI JOURNAL Vol. 5 No. 1 June 2011

4. Hill, G.H. (1896). The Thirlmere Works for


the water supply of Manchester; Minutes of
the Proceedings of the Institution of Civil
engineers, Vol 126, 1896.
5. Ito, T. and Matsui, T (1975). Methods to
estimate lateral force acting on stabilising
piles. Soils and Foundations, Vol 15, No 4,
pp 43-59
6. Ito, T., Matsui, T. and Hong, W.P. (1981).
Design for stabilizing piles against landslide
one row of piles. Soils and Foundations,
Vol 21, No 1, pp 21-37.
7. Matlock H. and Reese L.C. (1960).
Generalised solutions for laterally loaded
piles. Journal of Soil Mechanics and
Foundation Engineering Division, ASCE,
86(SM5).
8. Randolph, M.F. (1981). The response
of flexible piles to lateral loading.
Gotechnique, 31(2), 247-259.
9. Randolph M.F., (2004). Piglet Manual,
Analysis and Design of Pile Groups Version
5.1: Released May 2004.
10. Reese, L.C. and Matlock, H. (1956). Nondimensional solutions for laterally loaded
piles. Proceedings 8th Texas Conference on
Soil Mechanics.
11. US Army Corps of Engineers (2008).
Technical Instructions, Design of Deep
Foundations, TI 818-02, 3 August 1998.

Vous aimerez peut-être aussi