1

Temporary stability of slopescut in London Clay

Dr Nesha Kovacevic

Geotechnical Consulting Group, London

Deep excavations at Terminal 5

• Bottom-up construction

• Open cut excavations for economy

• Maximum depth 22m

• Cut slopes to be as steep as possible

• Stand-up time up to 1/2 years

2

End shunt

CTB

Launch

chamber

Launch chamber

3

Temporary stability of deep cuts in London Clay

• To predict stand-up time of deep cuts using numerical analyses and so optimise slope geometry – maximum steepness and minimum excavation and backfill?

• Only after calibration against case histories of temporary slope failures

(Bradwell, Prospect Park, Wraysbury)

Constitutive model forLondon Clay

4

(εεεεdp)p (εεεεd

p)r εεεεd

p

ϕ ϕ ϕ ϕ ', c

'

ϕϕϕϕ'r , c'r

ϕϕϕϕ'p , c'p

Soil model used

G = G (p’, εεεεd)

K’ = K’ (p’, εεεεv)

0.0001 0.001 0.01 0.1 1

Axial strain or volumetric strain (%)

0

500

1,000

1,500

2,000

2,500

3,000

Se

ca

nt

(3

.G)/

p' o

r K

'/p

'

recommended shear

modelled shear

modelled bulk

Small strain stiffness curves in extension

5

0 5 10 15 20

Shear strain, γ γ γ γ (%)

0

50

100

150

Sh

ea

r s

tre

ss

, τ τ τ τ

(kP

a)

0 0.025 0.05 0.075 0.1

Displacement across a 0.5m thick layer, ∆ ∆ ∆ ∆ (m)

σσσσ'=166.7kPa

γγγγ=∆/TT

∆∆∆∆ ττττ

Predicted behaviour in drained simple shear

-50

0

50PW

P c

ha

ng

e

0 2 4 6 8 10

Axial strain (%)

-150

-100

-50

0

Sh

ea

r s

tre

ss

(kP

a)

measured

predicted

Behaviour in undrained triaxial extension

6

0 50 100 150 200 250

Vertical effective stress (kPa)

0

1

2

3

4

5

Vo

lum

etr

ic s

tra

in (

%)

Test (depth: 13.55m)

Predicted

Swelling in oedometer test

Non-linear permeability

model:k=k0.e

-b.p’

1E-012 1E-011 1E-010 1E-009 1E-008

Horizontal permeability (m/s)

50

40

30

20

10

0

Dep

th b

elo

w g

rou

nd

leve

l (m

)

A3

A3

A3

A3

A3

A2

A2

A2

A2

A2

A3

A2

A2A2

A2

A2

A2

A2

A2

A2

A2A2

A2A2

A2 A2A3

A3

A3

A3

A3

A2

Biii

Biii

Biii

Biii

BiiBii

Bi

East of basin

Near surface

A3

cen

tra

l and w

est

A2 central and west

7

Short - term slope failure at Bradwell

3.5

m

Original ground levelClay Fill1:

1

Excavated slope at Bradwell

Marsh Clay

Weathered

London Clay

UnweatheredLondon Clay

GWL0.9m

11.3

m

2.7

m7.2

m

0.5

:1

1:1

200 40 60 80 100

Su (kPa)

8

Tension zone

Predicted rupture surface at Bradwell by FE analysis

11.3

m London Clay

Marsh Clay

Clay Fill

MoS=1.02

εεεεpD=1%

Softening starts

at εεεεpD=5%

εεεεdp

Str

en

gth

residual

peak

5% 20%

Rupture

surface

Back-analysis by LEM(Skempton, 1965)

11.3

m

0.5

:1

1:1

1:1

Predicted rupture surfaces by LE and FE method of analysis

Clay Fill

Marsh Clay

London Clay

Predicted by FEM

9

Short-term (undrained) failuresat Bradwell

• Can be analysed in terms of effective stresses using

the same constitutive model and soil parameters

as for the delayed failures of cuttings in London Clay

• Role of progressive failure is small, and concentrated

in the area around the toe of the slope

• Soil stiffness is of importance in estimating

the short-term stability

Slough

Windsor

Scale

0 1 2 3 4km

M25

Staines

StainesReservoir

HeathrowAirport

M4

Prospect Park

NM25

M4 Motorway

DatchetReservoir

WraysburyReservoir

King George VIReservoir

Site location

10

Prospect Park failure 9 weeksafter excavation

Tectonic shear surface in London Clay

at Prospect Park

11

21

Terrace Gravels

London Clay

London Clay

11

. 6m

Cut-off wall

Tectonic shear zone

Typical cross section at Prospect Park

Slip surface

0.5m

0 10mScale

With shear No shear

Movement vectors during excavation

12

0

100 kPa

50

150

-50

0 10mScale

With shear No shear

Pore pressures at the end of excavation

0

100 kPa

50

150

-50

0 2 4 6 8

Depth of excavation below top ofLondon Clay (m)

0

0.1

0.2

Ho

rizo

nta

l m

ove

me

nts

(m

)

with shear

with shear, no prior gravel removalno shear, no prior gravel removal

shearzone

no shear

Horizontal movements at the top of London Clayduring excavation

13

15% εεεεpD=2%

2%

15% εεεεpD=2%

2%

Rupture

surfaceRupture

surface

0 10mScale

With shear No shear

Plastic shear strains just prior to collapse

Ho

rizo

nta

l m

ove

me

nts

(m

)

0 0.2 0.4 0.6 0.8 1

Time since excavation (yrs)

0

0.1

0.2

with shear

with shear, no prior gravel removal

no shear, no prior gravel removal

no shear

Horizontal movements at top of London Clayafter excavation – Zero suction

14

0 0.2 0.4 0.6 0.8 1

Time since excavation (yrs)

0

0.1

0.2

0.3

0.4

0.5

Ho

rizo

nta

l m

ove

me

nts

(m

)

025

Horizontal movements at top of London Clayafter excavation – With shear zone

Surface suction in kPa

50

Key factors determining time to failure

• Presence of tectonic shear

• Previous site history

• Surface suction; 25kPa gave the best match to observed time to failure

• Permeability profile and non-linearity

• Ko – if tectonic shear zone is absent

• Progressive failure – rate of drop from peak to residual strength

15

Predicted failure mode for

5m bermed 1:1 slopes at T5

Surface suction 25kPa

Predicted failure mode

for 5m bermed 1:1 slope at T5

Surface suction ‘zero’

16

Launch

chamber

Launch

chamber

17

Lessons for the future

• Monitor movements to identify presence of tectonic shears or development of basal shear

• Take measures to maintain suctions at slope surface

• Monitor suctions• Differentiate between drying beds and

lagoons• Check sensitivity to Ko, k and surface

suction