You are on page 1of 22

PROF.

WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Outline
Background: infrastructure slopes and
discrete pile stabilization
Discrete passive piles for Limiting lateral stresses
infrastructure slope Pile spacing
Field studies
stabilization - Hildenborough
- Mill Hill
William Powrie and Joel Smethurst - Grange Hill
Concluding comments

Infrastructure slopes

Often railway embankments


> 100 years old
Clay fill
Background: infrastructure slopes End tipped
Hidden defects
shear surfaces
high permeability lenses
clod and matrix structure

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

End tipping End tipping

Image source: http://www.transportarchive.org.uk/ Image source: http://www.transportarchive.org.uk/

Clod and matrix structure arising


from construction process
Matrix Matrix Tree root

Softer Clod
Weaker
More permeable
Partly saturated
Controls strength &
permeability

Clod
Stiffer Clod and
Stronger matrix
Less permeable
Fully saturated structure
Controls shear stiffness & compressibility in CT scans
Photo: Graham Birch
Tony OBrien, Mott MacDonald

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Strain softening of clays Major slip: Stonegate, 2014


leading to deep seated slope
failure (Potts et al)
Development of plastic
shear forms deep seated
mechanism
Peak

Post rupture
Residual

Ellis and OBrien (2007)


Photo: Network Rail

Major slip: Harbury, 2015

Failure surface
observed in a
trial pit near the
toe of an
embankment in
a high PI clay

Photo: Tony OBrien,


Mott MacDonald

Photo: Network Rail



PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Seasonal cyclic changes in pore ..can be an operational nuisance..


water pressure..
Rainfall and potential Soil moisture deficit (SMD)
evapotranspiration, South- = amount of water required to
East England (Southampton) recharge the soil profile

Jul
May

Nov
Jan

Jun
Mar

Dec
Feb

Aug

Oct
Sep
Apr
Monthly rainfall and PET (mm/month)

120
Rainfall 0
100 PET 20
40
80
60
SMD (mm)

80
60
Photo: Graham Birch, Network Rail
100
40
120
20 140
160
0 180
Jul
May

Nov
Jan

Mar

Jun

Oct

Dec
Feb

Apr

Aug
Sep

and in numerical analyses..


and lead to gradual
downslope movement as a
continuum as seen in Shrink (summer) Swell (winter)

centrifuge model tests Lateral component Lateral component


at mid-slope at mid-slope
(Take and Bolton,
2011)..
At the surface, the summer
shrinkage does have a
horizontal component,
although this is lost deeper
into the slope where shrinkage
is predominantly vertical
Lateral seasonal clay
movements ~ 25 mm

Mott MacDonald FLAC analyses, Queensbury embankment



PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

or as a result of successive opening and shallow instability due to elevated


and closing of tension cracks. winter pore water pressures

Summer cracking
(upslope surface moves downhill) Positive pwps

Elevated near-surface
winter pwps

Summer dry
Winter swelling
condition
(downslope surface moves downhill)

Stabilization of infrastructure
slopes
Infrastructure slopes may be stabilised using a
variety of methods
These range from contiguous bored pile walls to
Slope stabilisation using discrete piles softer solutions such as nailing or lime stabilisation
Spaced discrete piles is an intermediate solution

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Discrete pile stabilisation of Stabilization of infrastructure


infrastructure slopes slopes using discrete piles
Strain gauges
and Uncertainty over behaviour mechanisms and
inclinometer
tubes in piles
appropriate method of analysis
May or may not be an obvious (incipient) failure
surface
Also:
How far apart?
Inclinometer
tubes at toe and Where in the slope?
crest of slope

Inclinometer Piezometers Depth, bending strength and stiffness?


tubes in
between piles Raingauge

Undrained shear strength failure


criterion

Limiting lateral stresses on side-bearing


piles

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Effective stress failure criterion Finite element analyses by Pan


(2013)
Carried out using FLAC3D
Pile is fixed at base, and
soil pushed laterally from Pile

end boundaries left to


right

Interface
elements Soil

Pan (2013): surface-affected


depth increases with

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Centrifuge model tests

Scale 1:60
Pile depth 9 m
Geotechnical centrifuge model tests to Pile diameter 0.95 m
investigate maximum effective pile EI equivalent to 1 m dia reinforced concrete
spacing (Hayward et al, 2000)
Pile spacing 3.2, 4.2 and 6.3 pile diameters
Speswhite kaolin clay
Increase to 60g then allow pore pressures to rise
until either failure occurs or stability is reached

Schematic cross section of Centrifuge model


centrifuge model

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Centrifuge model: typical Downslope displacements


equalised water table

3DP 4DP 6DP


Spacing of 3.2 Spacing of 4.2 Spacing of 6.3
pile diameters pile diameters pile diameters

Bending moment, shear force and Centrifuge test results


net lateral stress Test no. Pile Mmax, Mmax, Force Force
Bending moment per pile (kNm) Shear force per pile (kN) spacing, kNm / kNm / m resisted, resisted,
m pile kN / pile kN/m
3DP
4DP 3DP 3 146 48 42 14
6DP
4DP 4 242 61 74 19
Depth (m)
Depth (m)

6DP 6 376 63 91 15

Force resisted is over the uppermost 2.1 m of the pile or


slope
Pressure per pile (kN/m2)
The unreinforced slope and test 6DP failed; 3DP and 4DP
did not
Failure was by global slip, not flow-through between the
piles
Depth (m)

Slope failure occurred at inferred lateral stresses much less


than the limiting at-depth values

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Our study sites

Field studies Grange Hill


Magnolia Road
Mill Hill

Newbury
Pound Green

Hildenborough

Monitored sites

Year of Type of structure Location Geology


instrumentation

2001 Pile stabilised Hildenborough, Weald Clay


embankment Kent (NR) Hildenborough
2004 Pile stabilised Mill Hill, Anglian Till
embankment London (LUL
Northern Line)
2006 Pile stabilised Grange Hill, London Clay
cutting slope London (LUL
Central Line)

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Hildenborough: slope geometry


Hildenborough during pile installation
Upper slope partly cut back during

near Tonbridge, Kent Ballast


construction of rockfill berm

Weald Clay railway embankment 4.5m


4m 1.5m 4.5m

33

Ongoing problems with track geometry 3.5m Weald Clay


embankment fill
Imported
rockfill
38
89.5m AOD
Discrete piles and rockfill stabilisation Weathered Weald Clay

Inclinometers to monitor slope and pile 6.5m Intact Weald Clay

movements, strain gauges to monitor pile


bending moments

Hildenborough during pile installation Hildenborough: final slope geometry

1.0m
cess 10.0m 6.0m

Ballast

4.5m

Rockfill
24

3.5m Weald Clay


embankment fill 30
89.5m AOD
Weathered Weald Clay

Intact Weald Clay



PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Hildenborough: final slope geometry Hildenborough

Piles contain
inclinometer tubes Inclinometer tubes
and 6 pairs of strain midway between piles
gauges
A
B
C

Strain gauges fixed to pile Calculation of bending moments from


reinforcement cage strain measurements

Bending moment is calculated from M =


EI(1-2)/y
Stiffness of the pile varies from uncracked
EI to fully cracked EIc, where Ic/I = ~0.5
Changes in E with time / creep? Creep is
due to moisture loss and consequent
shrinkage, less of an issue for concrete
when buried than when exposed to air

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Cracking Calculation of bending moments from


inclinometer profiles
Reinforcement Reinforcement
in tension and concrete Gauges between
major cracks - net
concrete cracks carry
compressive stress and strain relief, Bending moment can be determined by twice
under-reading of
stress
average tensile strain differentiating a curve fitted to bent shape of
Strain gauges
pile and multiplying by EI (M = EI.d2y/dx2)
embedded in
concrete Good curve fitting is tricky but some success
with curves made up of a series of polynomials

Fully cracked neutral


Still need a value of E to obtain bending
Gauges spanning
axis forms ~ of the major cracks moments
way through section possible over-reading
of tensile strain

Hildenborough: slope movements Hildenborough: pile displacements


Day 42 after Change in Displacement (mm ) Displacem ent (mm)

pile -5
-1
0 5 10 15 20 25 30 Day 42 after -5
-1
0 5 10 15 20 25 30 35 40
construction 0
pile
0

(10 May 1 1
construction
2001) Day 1345 after 2
2
(10 May Day 1345
Depth below ground level (m)

shortly after pile 3


3
2001) after pile
Depth from g.l. (m)

rockfill construction 4

regrading
4
shortly after construction
(01 Dec 2004) 5
5
rockfill (01 Dec 2004)
6
6
Tube betw een piles A
and B, day 42
regrading Pile C, day 42
Pile C, day 1345
7
7 Tube betw een piles B Pile B, day 42
and C, day 42
8 Pile B, day 1345
8 Tube betw een piles A
and B, day 1345 Pile A, day 42
Tube betw een piles B 9
9 Pile A, day 1345
and C, day 1345
10
10

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Hildenborough: bending moments Hildenborough: bending moments


(pile C, day 42) (Pile C, Day 1345)
Bending Moment (kNm) Bending Moment (kNm)
Measured -200 -150 -100 -50 0 50 100 150 Measured -100 0 100 200 300
bending 0
bending 0
Measured BM data (average Measured BM data (average
moment 1
of two sets of gauges)
moment of two sets of gauges)
1
average of Range of readings (given by
average of Range of readings (given by

CRACKING
MOMENT
2 two sets of gauges) two sets of gauges)
strain gauge Spline fitted through strain gauge
2
Spline fitted through
pairs 3 measured BM
pairs 3 measured BM data
BM required to give BM required to give
4 inclinometer deflection 4 inclinometer deflection
Depth (m)

Depth (m)
5 Cracking moment
Bending moment Deflection (mm) 5
-1
-1 0 1 2 3 4 5 6

6 obtained by 0

1 6
Bending -10
-1
0
Deflection (mm)

10 20 30 40

differentiating 2
Difference moment
0

1
7 3

spline fitted to 7
probably obtained by 2
Depth (m)

Pile
3

8 the measured 5

explained differentiating

Depth (m)
4
6
8
displacements 7 bending by pile spline fitted to 5

6
9
EI =
8

9
Fitted spline

Inclinomter measurement
moment 9
cracking the measured 7

uncracked capacity displacements


10
Fitted spline
10 9
Inclinomter measurement

10
= 250
10

kNm EI = uncracked

Hildenborough: net lateral pressures Hildenborough: explanation of pile


(pile C, day 1345) lateral stresses and relative movements
Pressure (kPa)
Obtained by 200 100 0 -100 -200
differentiating 0
Pressure distribution
strain gauge 1 from spline fitted to
measured BM data
bending
2
moments Pressure distribution
3 required to give
inclinometer deflections

4
Depth (m)

Obtained by
5 differentiating
spline fitted to the
6
measured
7 displacements
8 Net pressure over the (a) (b) (c)
9
top 4m of pile
= 127 kN /per pile
10

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Hildenborough pile and slope


displacements
Displacem ent (mm )
-1
-5 0 5 10 15 20 25 30 35 40
0

1
Rockfill
2 Day 42
Difference
Depth below ground level (m)

EFC
3
between pile Mill Hill East
4 and soil
displacements Weathered
5 Day 1345
and intact
6
WC
7 Average slope displacement, day 42

Average slope displacement, day 1345


8
Average pile displacement, day 42
9 Average pile displacement, day 1345

10

Mill Hill East: Cross section through the


Mill Hill East embankment
Mill Hill East:
21.0 m 11.5 m
discrete pile
stabilised Anglian Cable run
Till embankment,
Ash/Ballast
13 m high 4m

21.5
Boundary
fence
24
Anglian Glacial Till
Embankment Fill
11 m long 17
stabilising piles

Berm benched into


embankment slope
London Clay

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Mill Hill: instrumentation

Piezometers and
moisture content
measurement

Inclinometer
tube in the slope

Inclinometer Two piles contain


tubes and strain gauges and an
piezometers inclinometer tube

Pile and soil displacement - inclinometer Pile bending moment from strain gauges
Displacement (mm) -
Displacement (mm)
-4 -3 -2 -1 0 1 2 3 4 5 6 7 8 -4 -2 0 2 4 6 8 10 12 14
0 0

1
1
2
2 3
Bending moment (kNm)

3
Pile C
4
Soil in front
5
4
6
of pile
Depth (m)
Depth (m)

5 07 Jan 05 7 03-Feb-05
19 Jul 05 19-Jul-05
8
6 31-Jan-06
31 Jan 06
24 Aug 06 9 24-Aug-06
7
12 Oct 06 10 12-Oct-06
27 Jun 07 27-Jun-07
8 11
22 Jan 09 22-Jan-09

9
19 May 09 12 19-May-09 1 2
29 Sept 09 29-Sep-09 Late 2004 and
13
10 28 Jan 10 28-Jan-10 2005 very dry
08 Jul 10 14
08-Jul-10
11 15

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Mill Hill: short term relative pile Mill Hill: long term relative pile and
and soil displacements soil displacements
Displacement (mm)
-4 -2 0 2 4 6 8 10 12 Displacement (mm)
0 -4 -2 0 2 4 6 8 10 12
0
1
Soil displacement Readings taken Soil displacement
1
Readings taken 08 July
2
behind pile row 24 August 2006 behind pile row 2
2010 - 66 months after
3

4
- 20 months after pile 3
Pile pile installation
Pile 5
installation 4

5
6 Soil displacement 6 Soil displacement
Depth (m)

in front of pile row in front of pile row

Depth (m)
7
7
8 8
9 9
10 10

11 11

12 12
Slope inclinometer tube b Slope inclinometer tube b
13 Slope inclinometer tube c 13 Slope inclinometer tube c
14 Pile A 14 Pile A

15 15

Bending moments from measured Net lateral pressures from measured


strains and displacements strains and displacements
Bending Moment (kNm) Pressure (kN/m)
Measured -40 -30 -20 -10 0 10 20 30 40 20 10 0 -10 -20
0 Bending moment
bending obtained by
0

moment 1
differentiating spline 1
PILE 80, 2 Obtained by
fitted to the measured Net 2
differentiating
24 August 3 displacements Differentiate downslope 3 strain gauge
06 4
the bending pressure
Deflection (mm)
bending
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
moment 4
Depth (m)

0
over the top moments
Depth (m)

5
1
diagram twice 1.5 m depth 5
CRACKING MOMENT

2
6
3
6
7 Fitted spline 4
Obtained by
Depth (m)

8 Measured bending 6
7 differentiating
Pile moment 7
Defl. from BM 8 spline fitted to the
9 Cracking moment Mcr
bending 8

9
24-Aug-06

measured
Inclo reading
9
moment 10 Spline fitted to 10
24-Aug-06
PILE1
displacements
displacements
capacity = 11
11
10
200 kNm
11

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Net pressures from spline fit to


displacements Idealised mechanisms
Pressure
Line Load (kN/m) Pressure
Line Load (kN/m)
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
0 0

1 1
~2.5 m
2 2 ~4.0 m
Upslope shrinkage
Downslope pressure
3 pressure drives pile 3
associated with
movement
shallow surface
-5 -4
Displacement (m m)
-3 -2 -1
0
0 1 2 4 -1 0
Di splacement (mm)
1 2 3 4 5 6
4 movements drives
0

2
1
pile movement Soil movement profile
Depth (m)

2
Depth (m)

4 5
3

4 Pile C dis plac ement


spline
5
Depth (m)

Depth (m)

5 5

6 Pile C inc linometer


6 dis plac ements

Pile C dis placement


7 7

6
Pile
s pline

Pile C inc linometer


8

9
6 8

9
displacements
10 10

11 11

7 7
Upslope
Pile C net pressure Pile C net pressure 8
8
spline spline Downslope
9 9

10 1 10 2 Shrinkage of Caused by shallow 2


embankment
1 downslope movement
24 August 2006 12 July 2010
11 11

Mill Hill Summary: Mill Hill


Short-term mechanism has the Initial movements of the piles occur due to seasonal
pile bending backwards into the
slope shrinkage of the clay fill
The peak bending moment so far appears to be associated
Long-term pile mechanism is a
uniform bend downslope with upslope movement of the pile due to inward shrinkage
of the embankment
Short-term: Long-term: Later pile movements are associated with shallow
Displacement Pressure Displacement Pressure
downslope surface movements (although small seasonal
reversals in pile movement still take place)
Major driving
pressure is Small seasonal ratcheting of bending moment
inward shrinkage
of embankment
Generally pile movements and bending moments remain
Pressure on the
top of the pile is
small and well below design capacities
downslope

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Grange Hill CTS2 prior to


remedial works
Grange Hill
London
Clay cut
slope, 11 m
Grange Hill high
Low FoS

Grange Hill
Grange Hill
Discrete
pile
stabilised
London
Clay cut
slope, 11 m
high

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Grange Hill: pile and soil movements


Grange Hill
Displacement (mm) [+ve is downslope]
-5 0 5 10 15 20 25 30 Pile single bends
0
in the downslope Displacement Pressure
1
direction under
2
Relative fairly shallow Pressure on the top
of the pile is
3 slope movements downslope
4 Soil displacement of bottom part of
Depth below ground surface (m)

5 piles and soil the pile does little


6 Pile midway between
7 No distinct
8 Average of piles B
piles for two dates: formation of a
9
and C 19-Apr-07
19 April 2007 slip surface do
10
Average of piles B
11 December 2009 the piles stop that
and C 11-Dec-09
from occurring?
11
Average of tubes b
12 and c 19-Apr-07

13 Average of tubes b
14 and c 11-Dec-09

15

Pile displacements

Site Rates of Total All show continued


movement of accumulated gradual downslope
pile top downslope movement, possibly
movement caused by either gentle
Hildenborough ~ 5 mm/yr 50 mm deterioration in
stability of the slopes
(e.g .due to pore water Concluding comments
pressure changes), or
Grange Hill ~ 4 mm/yr 20 mm downslope ratcheting
due to shrink and swell
of surface layers.

Mill Hill ~ 1 mm/yr 5 mm Loss of flexural rigidity


as the concrete in the
piles crack

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Concluding comments (1) Concluding comments (2)


1. Uncertainties remain regarding the calculation of 4. Failure in centrifuge model tests occurred by global
limiting lateral pressures on passively loaded piles movement rather than flow-through of soil between
in sloping ground characterised by the effective the piles
stress failure criterion no real consensus? 5. Pile pressures and bending moments depend on the
2. Piles can be spaced at up to 4 times the diameter relative soil/pile movement, which is qualitatively
(centre to centre) and still be effective in limiting consistent with simple mechanisms of slope and
displacements (3 pile diameter seems to be a pile displacement associated with appropriate
popular choice?) combinations of
3. Piles are commonly located about 1/3 of the way up (i) deep seated slip
the embankment, with slope reprofiling / toe (ii) shallow downslope movement
surcharging to prevent local failure downslope (iii) inward shrinkage due to vegetation effects

Concluding comments (3) Journal papers


6. Measured pile strains and displacements give Seasonal changes in pore water pressure in a grass covered cut slope in
London Clay. J A Smethurst, D Clarke and W Powrie. Gotechnique 56
broadly comparable bending moments and implied (8), 523-537, October 2006
net pressures, as long as pile EI is reduced to allow
for cracking when interpreting inclinometer data Monitoring and analysis of the bending behaviour of discrete piles used
to stabilise a railway embankment. J A Smethurst and W Powrie.
7. Bending moments in centrifuge model tests and in Gotechnique 57 (8), 663-677, October 2007
the field (and inferred lateral stresses) were always Three-dimensional numerical modelling of discrete piles used to stabilize
much lower than theoretical limiting values for a landfills. S Kanagasabai, J A Smethurst and W Powrie. Canadian
single pile Geotechnical Journal 48, 1393-1411, 2011

8. Upslope pressures on the upper part of the pile may Factors controlling the seasonal variation in soil water content and pore
occur if a stiff stratum is present or placed water pressures within a lightly vegetated clay slope. J A Smethurst, D
Clarke and W Powrie. Gotechnique 62 (5), 429-446, May 2012
(Hildenborough rockfill), or as a result of inward
shrinkage due to vegetation effects (Mill Hill) Limiting pressure on a laterally loaded pile in a frictional soil. D Pan, J A
Smethurst and W Powrie. Gotechnique Letters 2, 55-60. 2012

PROF. WILLIAM POWRIE
Discrete passive piles for infrastructure slope stabilization

Acknowledgements
EPSRC
GeoObservations
Highways Agency Network Rail
TfL
Thankyou for listening
Tubelines www.soton.ac.uk/engineering
Metronet
Mott MacDonald

You might also like