Loose Fill Slopes

Johnny Cheuk

CIVL6079 – Slope Engineering

 Origin of the problem

The presence of loose fill slopes in a unique problem in Hong Kong, which
has not seen in other parts of the world. The formation of these loose fill
slopes is related to the rapid development in Hong Kong in the early
1970s. Due to the pressure in building development, a lot of embankments
(or slopes) were formed by dumping fill materials, mainly, completely
decomposed granitic or volcanic soil (CDG or CDV), on gentle sloping
ground. There was no proper compaction and very limited engineering
control. The soils were deposited in layers parallel to the slope surface.
The average relative compaction is about 70-80%. There are about 6,000
old loose fill slopes in Hong Kong.

1. Natural slopes (natural terrain)

2. Man-made slopes

Fill slopes Cut slopes

CIVL6079 – Slope Engineering 2

 Overview

The majority of the loose fill slopes in Hong Kong are very gentle, and not
very high in height. Therefore conventional LE calculations may reveal a
satisfaction FoS.

Sun (1999)

CIVL6079 – Slope Engineering 3

 Overview

Statistics shows that the old loose fill slopes has a very low dry density
(relative compaction), with an average of about 70%. How would this low dry
density affect the engineering behaviour?

Sun (1999)

CIVL6079 – Slope Engineering 4

 Sau Mau Ping failures

The Hong Kong government employed an Independent Review

Panel to look at the cause of the Sau Mau Ping failures (Hong Kong
Government, 1977). They concluded that:

“The failure was caused by the infiltration of rainwater into the slope,
leading to a loss of strength in the fill, envisioning the instantaneous
conversion of the slope into a mud avalanche (Knill et al., 1976).”

Since then, static liquefaction was attributed to be the governing

mechanism of loose fill slope failures.

CIVL6079 – Slope Engineering 5

 Behaviour of loosely compacted residual soil

Loose fill materials in HK can

strain-soften. The mobilised Mins=1.12(28°)
friction angle at the onset of M=1.58(38.7°) Instability Line (Lade, 1992)
softening can be well lower
150
the critical state friction angle
(this is the so-called

Deviator Stress (kPa)

instability line)
100
However, when strain-
softening is triggered by
wetting along a constant q- 50
reducing p’ stress path, the
mobilised friction angle will
be higher (i.e. conventional
0
CU is conservative)
0 50 100 150 200 250 300

CDV is not as contractive Mean Effective Stress (kPa)

CDG from Shau Kei Wan
(Cheuk, 2001)

CIVL6079 – Slope Engineering 6

 Behaviour of loosely compacted residual soil

Ng et al. (2004)
CIVL6079 – Slope Engineering 7
 Behaviour of loosely compacted residual soil

CDV is not as contractive. A phase transformation is normally observed.

Ng and Chiu (2001)

CIVL6079 – Slope Engineering 8

 Behaviour of loosely compacted residual soil

A mobilised friction angle of 26° appears to be the lower bound for peak
shear resistance

HKIE (2003)

CIVL6079 – Slope Engineering 9

 Behaviour of loose fill slopes in numerical model
Maximum : 5.2m
Numerical modelling was carried
5m
out the examine the progressive Original Vector Scale
Position
failure mechanism in loose fill
slopes using FLAC. A pre-defined
post-peak strain softening curve is
programme into the soil model,
and the behavior of the slope
0 10m
under the effects of rainfall Scale
Toe Wall

infiltration was examined.

Cheuk et al. (2005)

CIVL6079 – Slope Engineering 10

 Behaviour of loosely fill slopes in numerical model

Progressive failure mechanism

0.2 second
0.1 second

0.4 second
0.3 second

Remain Elastic
0 10m
Scale Liquefaction Initiated

Cheuk (2001)
CIVL6079 – Slope Engineering 11
 Behaviour of loosely fill slopes in numerical model

Progressive failure mechanism

70

60
Dveiator Stress, q(kPa)

50

40

30 Element A
A
Element B
20 Element C
B
At 1 second
10 At 2 seconds
C At 3 seconds
0
0 1 2 3 4 5 6
Shear Strain, e s(%) Cheuk (2001)
CIVL6079 – Slope Engineering 12
 Behaviour of loose fill slopes in centrifuge

Although significant strain softening was observed in triaxial tests, whether

loose residual soil can liquefy has been a controversial topic. In an attempt
to answer these questions, a lot of centrifuge model tests were conducted.
For example, some centrifuge model tests were carried out at HKUST in
which a loose fill embankment subjected to uni-axial and bi-axial
earthquakes (Ng et al. 2004). No liquefaction failure was observed in
model tests.

CIVL6079 – Slope Engineering 13

 Behaviour of loose fill slopes in centrifuge

Centrifuge model tests were also conducted at Cambridge University

Engineering Department (CUED) using their 10m Tuner beam centrifuge.
Three studies – Yeung (2002), Wong (2003) and Lee (2005) – were
conducted with an aim originally set to find out the conditions for
liquefaction to occur, but later changed to find out the possible failure
mechanisms in loose fill slopes.

CIVL6079 – Slope Engineering 14

 Behaviour of loose fill slopes in centrifuge

Yeung (2002) investigated the behaviour of loose fill slopes under intense
rainfall in an attempt to model flow failure. An environmental chamber was
used to mimic rainfall.

Take et al. (2004)

CIVL6079 – Slope Engineering 15

 Behaviour of loose fill slopes in centrifuge

The technique of Particle Image Velocimetry (PIV) was used to observe

soil deformation. Excessive settlement was observed, but a flow failure
(or liquefaction) was not triggered.

Take et al. (2004)

CIVL6079 – Slope Engineering 16

 Behaviour of loose fill slopes in centrifuge
Conclusions from Yeung (2002):
- CDG is too permeable
- The air void in the unsaturated fill is highly compressible, the
model fill slope may exhibit “wetting collapse”
At Slope Crest:

Take et al. (2004)

CIVL6079 – Slope Engineering 17
 Effects of Wetting and Drying

Cyclic dead load test

Data logging system

LVDT

Automatic pressure
controllers
Load cell

Soil specimen

Dead weight

CIVL6079 – Slope Engineering

 Effects of Wetting and Drying

Ø CDG collected from Beacon Hill

Ø saturated sample
Ø initially loosely compacted
Ø load-controlled test
Ø 5 cycles of wetting and drying before failure is mobilised

CIVL6079 – Slope Engineering

 Effects of Wetting and Drying
Ø the volume reduced a lot during the 1st cycle due
to structural collapse
Volumetric
Ø the amount of volume change reduced in stage Rc (%)
behaviour
subsequent cycles
1st wet 87.5 Contraction
Ø drying is accompanied by contraction
Ø wetting is accompanied by dilation 1st dry 88.6 Contraction
Ø the behaviour is similar to over-consolidated clay
Ø in the last wetting, a lot of dilation was observed 2nd wet 88.5 Dilation

2nd dry 88.9 Contraction

3rd wet 88.8 Dilation

3rd dry 89.3 Contraction

4th wet 89.3 Dilation

4th dry 89.4 Contraction

5th wet 89.2 Dilation

5th dry 89.6 Contraction

CIVL6079 – Slope Engineering

 Effects of Wetting and Drying

Ø shear deformation is irrecoverable

Ø plastic shear strain is accumulated
Ø amount of shear deformation reduced with number of cycles
Ø very large deformation was observed in the last wetting path

CIVL6079 – Slope Engineering

 Behaviour of loose fill slopes in centrifuge

Wong (2003) performed a series of centrifuge tests to examine a proposed

alternative hypothesis.

CIVL6079 – Slope Engineering 22

 Behaviour of loose fill slopes in centrifuge

A tongue of permeability layer underneath the fill material (due to soil

heterogeneity) was included to mimic layering effects. Both loose and
dense fill slopes were tested.

Wong (2003)

CIVL6079 – Slope Engineering 23

 Behaviour of loose fill slopes in centrifuge

Video from Wong (2003):

CIVL6079 – Slope Engineering 24

 Behaviour of loose fill slopes in centrifuge

Conclusions from Wong (2003):

- Both dense and loose fill slopes failed as a fast flow slide under the
model conditions
- Dense slope exhibited a much stiffer response to the build up of pore
water pressures

CIVL6079 – Slope Engineering 25

 Behaviour of loose fill slopes in centrifuge

Lee (2005) tried to shorten the tongue of permeability layer and to make it
thinner than that in Wong (2003). Again, both loose and dense fill slopes
were tested.

CIVL6079 – Slope Engineering 26

 Behaviour of loose fill slopes in centrifuge

The centrifuge model from Lee (2005):

Monsoon High
Delivery Speed
System Camera

Light 1

Mirror

Light 2 Web
cam

CIVL6079 – Slope Engineering 27

 Behaviour of loose fill slopes in centrifuge

Video from Lee (2005):

CIVL6079 – Slope Engineering 28

 Behaviour of loose fill slopes in centrifuge

Results from Lee (2005):

- Rate of water infiltration into the
coarse layer exceeded the rate of
dissipation at the interface

- Transient pore water pressure

was again observed at the
location of seepage impediment

- Global slope failure was

triggered in both loose and dense
fill slopes

CIVL6079 – Slope Engineering 29

 Behaviour of loose fill slopes in centrifuge

Conclusions from Lee (2005):

- Deformation was progressive in layered loose fill slope
- Dense fill slope exhibited a stiffer response to the building up of pore
water pressure, but the ultimate failure was much more brittle than that
exhibited in the loose fill model.
25
PIV1
PIV2
PIV3

20 PIV1 PIV2
PIV3

Slip surface

15

10
loose dense
5

0
0 5 10 15
Start of water Time (s)
percolation

CIVL6079 – Slope Engineering 30

 Upgrading method - 3m re-compaction

The independent review panel has recommended some remedial measures,

which were then set out in the Geotechnical Manual for Slopes (GCO, 1984).

- excavate and
compact the top 3m
loose material to a
density of not less
than 95% of British
standard maximum
dry density

- provide an effective HKIE (2003)

drainage system

CIVL6079 – Slope Engineering 31

 Soil nailing in loose fill slopes

Maximum Vector: 6.87 m

Maximum Vector: 5.22 m
(Failed at 22 Seconds)

a. Loose Fill Slope b. Nailed Loose Fill Slope without Grid Facing

Maximum Vector: 0.38 m Maximum Vector: 0.33 m

c. Nailed Loose Fill Slope with Grid Facing d. Nailed Loose Fill Slope with Grid Facing
(EI=7 MPa-m4/m) (EI=14 MPa-m4/m)

Cheuk (2001)
CIVL6079 – Slope Engineering 32
 Soil nailing in loose fill slopes

0%
0% 25 %
25 % 50 %
50 % 75 %
75 % 100 %
100 % 125 %
125 % 150 %

a. Loose Fill Slope b. Nailed Loose Fill Slope without Grid Facing
(
0% 0%
2% 2%
4% 4%
6% 6%
8% 8%
10 % 10 %
12 %

c. Nailed Loose Fill Slope with Grid Facing d. Nailed Loose Fill Slope with Grid Facing
4 4
(EI=7 MPa-m /m) (EI=14 MPa-m /m)

Cheuk (2001)
CIVL6079 – Slope Engineering 33
 Design of soil nails in loose fill slopes

CIVL6079 – Slope Engineering 34

 Soil nails in loose fill slopes - design scheme

Grillage structure to transfer earth pressure through soil nails to

in-situ ground

Saturated loose fill

Soil nail inclination nearly normal to

slope face
In-situ
ground
Soil nails bonded into
competent ground

Vertical nails
Possible construction deviation

CIVL6079 – Slope Engineering 35

 Design steps

Step 1: Determine if slope upgrading is required (LE using f'col)

Step 2: Determine stabilising surface pressure with FoS = 1.1, assuming
(i) liquefied fill reaching steady state shear strength
(ii) triangular pressure distribution with basal shear resistance

CIVL6079 – Slope Engineering 36

 Design steps

Step 3: Determine inclination of the soil nails and bond lengths

Consider force equilibrium of grillage:

Basal shear
resistance Earth pressure
from loose fill
(i.e. Pmin from limit equilibrium analysis)

Other requirements:
Weight of the (i) Vertical nails at toe
grillage (ii) Grillage embedment of
0.5m
(iii) Maximum grillage
opening = 50%
Required nail
force resultant

CIVL6079 – Slope Engineering 37

 Design of soil nails in loose fill slopes

CIVL6079 – Slope Engineering 38

 New design recommendations

Question: Why do we have to adopt a hybrid nail arrangement?

CIVL6079 – Slope Engineering 39

 Problems with steeply inclined nails

Force Directions in Original

Nail Configuration

Force Directions in Deformed

Nail Configuration

CIVL6079 – Slope Engineering 40

 Verification by numerical analysis
If the toe is free, facing movement =
147mm

Soil deformation: 412mm

Soil deformation: 350mm
Structural deformation: 72mm
Structural deformation: 35mm
Soil deformation: 160mm
Structural deformation: 72mm
Soil deformation: 169mm
Structural deformation: 27mm

Nail rotation = 1.9° - 5.1° Nail rotation = 0.4° - 1.3°

CIVL6079 – Slope Engineering 41

 References and reading assignment

1. Cheuk, C.Y., Ng, C.W.W. & Sun, H.W. (2005). Numerical experiments of soil
nails in loose fill slopes subjected to rainfall infiltration effects. Computers and
Geotechnics, 32, 290-293.
2. GEO-HKIE (2011). Design of Soil Nails for Upgrading Loose Fill Slopes.
Geotechnical Engineering Office and Hong Kong Institution of Engineers
(Geotechnical Division), 96 p.
3. Hong Kong Government (1977). Report on the Slope Failures at Sau Mau Ping
August 1976. Hong Kong Government Printer.
4. Hong Kong Institution of Engineers (HKIE) (2003). Soil nails in loose fill – A
preliminary study. Final report. The Geotechnical Division of Hong Kong
Institution of Engineers.
5. Lade, P. V. (1992). Static instability and liquefaction of loose fine sandy slopes.
ASCE Journal of Geotechnical Engineering, 118(1): 51-71.
6. Lee, Y.S. (2005). Centrifugal Modelling of Landslides Triggering Mechanism in
Layered Fill Slopes, MPhil thesis, the University of Cambridge.
7. Ng, C. W. W. and Chiu, C. F. (2001). ‘‘Behaviour of a loosely compacted
unsaturated volcanic soil.’’ J. Geotech. Geoenviron. Eng., 127(12), 1027–1036.

CIVL6079 – Slope Engineering 42

 References and reading assignment

8. Ng, C.W.W, Fung, W.T, Cheuk, C.Y. & Zhang, L. (2004). Influence of Stress
Ratio and Stress Path on Behavior of Loose Decomposed Granite. Journal of
Geotechnical and Geoenvironmental Engineering. ASCE 2004 130(1), 36-44.
9. Ng, C.W.W., Li, X.S., Van Laak, P.A. & Hou, Y.J. (2004). Centrifuge Modeling of
Loose Fill Embankment Subjected to Uni-axial and Bi-axial Earthquakes.
Journal of Soil Dynamics and Earthquake Engineering, 24(4), 305-318.
10. Sun, H.W. (1999). Review of Fill Slope Failures in Hong Kong. GEO Report No.
96. Geotechnical Engineering Office, HKSAR Government.
11. Take, W.A., Bolton, M.D., Wong, P.C.P. & Yeung, F.J. (2004). Evaluation of
landslide triggering mechanisms in model fill slopes. Landslides, 1:173-184.
12. Wai, S.K. & Siu, C.K. (2005). Methods Other Than Recompaction for Upgrading
Loose Fill Slopes. GEO Report No. 162. Geotechnical Engineering Office,
HKSAR Government.
13. Wong, P.C.P. (2003). Centrifugal modelling of liquefaction failure of loose
granitic fill slopes. MPhil thesis, the University of Cambridge.
14. Yeung, F.J. (2002). Modelling of the behaviour of saprolitic soil slopes under
severe rainfall. MPhil thesis, the University of Cambridge.

CIVL6079 – Slope Engineering 43