DEEP SOIL MIXING COLUMNS AS

RETAINING STRUCTURE

K. K. Soh1,a , W. K. Leong1,b , L. Yeo1,c , S. H. Chew2 , K. W. Leong3,d and Z. W. He3,e

1 AGS Consultants Private Limited, Singapore. E-mail: a [email protected],
b [email protected], c [email protected]

2 National University of Singapore, Singapore. E-mail: [email protected]

3 Keller
Foundations (SE Asia) Private Limited, Singapore.
E-mail: d [email protected], e [email protected]

As part of the new water front town in Punggol, Singapore, a 4.2 km long water-
way has been constructed to connect two existing river reservoirs. The waterway is
about 4 m deep, with a varying width of 24 m to 80 m. The geological history record
shows that some branches of ancient streams and rivers run through part of the
site. Grid pattern of Deep Soil Mixing (DSM) columns extend beneath the waterway
and connect to the opposite bank was proposed when approximately 6 m of very
soft marine clay with qc < 0.25 MPa was identified by additional Cone Penetration
Test (CPT) along the CH3790-CH3900 (∼110 m). This proposed configuration not
only fully utilise the advantage of compressive strength but also isolates the devel-
opment of tensile stress at the DSM columns and serves as retaining structure to
the design cut-slope. 3D Finite Element (FE) Analysis is adopted to discern and
study the interaction of DSM columns with the confined very soft marine clay dur-
ing excavation. The analysis shows acceptable shear and isolated tensile stresses
developed within the longitudinal DSM columns. This grid pattern of DSM columns
with 0.44 m overlapped width will act as a retaining structure. On site, the DSM
columns performed satisfactorily with only a maximum 6 mm deflection registered
by the inclinometers. No sign of distress was observed on the slopes and surfaces
of the DSM columns as well.

1. INTRODUCTION

The Housing & Development Board of Singapore is constantly improving the environment
of their public housing to provide quality lifestyle to the residents. One such vision is to
transform Punggol Town in the northeast part of Singapore into the waterfront under the
“Realising the Vision for New Estates”. As part of the rejuvenation of the heartland’s pro-
posal, a 4.2 km long waterway meandering through the town was constructed connecting
two existing river reservoirs: Sungei Punggol in the west and Sungei Serangoon in the
east. This paper presents the principle and design of a robust soil treatment using DSM
columns where maximum thickness 6 m of very soft marine clay with qc < 0.25 MPa was
found along the CH3790-CH3900 (∼110 m) at South Bank.

Advances in Geotechnical Infrastructure

Edited by C. F. Leung, S. H. Goh & R. F. Shen
Copyright © 2013 Geotechnical Society of Singapore (GeoSS). Published by Research Publishing.
ISBN: 978-981-07-4948-4 :: doi:10.3850/978-981-07-4948-4 253 157
158 Advances in Geotechnical Infrastructure

North Bank

A A

South Bank

Figure 1. Proposed DSM columns at CH3790-3900.

1m
6m 1m
1m
2m

The very soft

marine clay layer
with qc<0.25MPa

Figure 2. Thickness of very soft marine clay with qc < 0.25 MPa at South Bank (Section A-A).

2. SUBSOIL CONDITION

Figure 1 shows the location of conducted boreholes and CPTs. The subsoil profile gener-
ally consists 5–7 m backfilled material (silty CLAY or loose clayey SAND) overlying by
varying thickness of very soft marine clay. Typical Singapore Old Alluvium with varying
degree of weathering is underlying the very soft marine clay layer. A maximum thickness
of 13 m and 8 m of very soft marine clay with qc < 0.5 MPa is detected along the CH3790-
CH3900 at South Bank and North Bank respectively by the CPTs. However, a layer with
qc < 0.25 MPa of very soft marine clay is about 6 m at the South Bank as shown in Figure 2.

3. DESIGN REQUIREMENTS

It is prudent to review the design requirements of treated composite ground subjected to

lateral loading from the past projects and research works.
Deformation and stability for treated composite ground are generally very complex,
except for only vertical or nearly vertical loading where the horizontal forces is less than 3%
with regular patterns of individual columns (Topolnicki [1]) is adopted. Numerical analy-
ses such as FE method should be used since the analysis can capture a wider range of fail-
ure modes such as composite shearing, columns bending and tilting (George M. Filz [2]).
Soil improvement by individual DSM columns are only performed satisfactorily in the
principle to improve the composite vertical stiffness of the treated soft ground (i.e.,
 Deep Soil Mixing Columns as Retaining Structure 159

embankment) but it is not favourable to stabilise the design cut-slope or to be used as

retaining structure. It is because high bending stress is expected to be imposed on the indi-
vidual columns and triggers progressive slope failure subsequently as reported at the trial
construction of drainage canal from the Samrong Canal in Samutprakarn Province to the
Gulf of Thailand [3]. This progressive failure is due to the brittle behaviour of the DSM
columns as observed by Karastanev et al. [4] where single columns and column blocks in
soft normally consolidated clay failed one by one at his centrifuge tests. Terashi et al. [5]
also points out the shear resistance of a DSM column block can be reduced by progressive
failure due to low failure strain of the columns.
Yao Shao et al. [6] has reported Unconfined Compressive Strength (UCS) of a soil-cement
mixture varies with water content, ranging from 500 to 2000 kPa, but its tensile strength
is much lower, typically about 15% to 25% of compressive strength. Topolnicki [1] also
mentions that the typical correlation of tensile strength of DSM columns using wet method
of soil mixing is between 8% to 15% of UCS but it is unlikely to be higher than 200 kPa.
Based on the findings by the researchers as mentioned above, the compressive stress of
the DSM columns must not be exceeded and the tensile stress must be isolated to control
the onset of progressive failure.

4. SLOPE STABILITY ANALYSES AT TEMPORARY CONDITION FOR WATERWAY

CONSTRUCTION

There is a potential of deep seated failure through the very soft marine clay with the qc <
0.25 MPa at the stretch of CH3790-3900. The thickness is about 6 m at the South Bank.
Slope stability for the waterway construction at the temporary stage is a critical condi-
tion. The height is about 5.5 m for the concrete drain construction and the overall height
from the existing slope is 9 m as shown in Figure 3. Limit Equilibrium (LE) Analysis shows
very low safety factor of = 0.52 for the localised cutting and = 0.94 for the deep seated
failure which is extending 80 m from the edge of waterway. Progressive failure is likely to
occur if no soil improvement or other means is carried out. Hence, a robust soil treatment
or a structural system is necessary due to this high unbalanced force.
Preliminary analysis by 2D FE Method is carried out for the individual DSM columns.
The equivalent average shear strength of the DSM columns is idealised by area method and
smeared across the column as shown in Figure 4. This 2D FE Analysis may not
represent the local variation of stress as accurately as 3D FE Analysis and it will also

0.949
0.908
0.869

~80m

RL106 Existing Ground Level

6m of silty clay RL102.5

5.5m 9m

6m of very soft marine clay (qc<0.25MPa) RL97

Figure 3. Deep seated failure plane extending up to 80 m from the edge of waterway.
160 Advances in Geotechnical Infrastructure

Tensile points across

individual GSC/DSM
columns
Individual GSC/DSM Columns

B B

Area Method = (ʌ/4Ø2)/(W×Ø)

Smeared strength across pile section Section B-B

Figure 4. Observed tensile cut-off points by FE analysis with phi/c reduction = 1.09 at temporary
condition.

underestimate the onset of progressive failure. Nevertheless, the development of tensile

points in the individual DSM columns will be indicated. Excessive tensile points were
observed in the equivalent 2D DSM columns and are likely to be developed right across
the column section with geotechnical safety factor = 1.09. The compressive failure strain of
the Singapore Marine Clay is generally 2–5% and the DSM column is 0.5–1% as reported
by Topolnicki [1]. The tensile strain of the DSM is very low percentage of the compressive
failure strain. By using the average shear strength in 2D FE Analysis under lateral loading,
the individual DSM columns and the surrounding marine clay are assumed to be com-
patible of strains across the smeared area and this may lead to an overestimation of the
geotechnical safety factor. The individual DSM columns which surrounded by marine clay
and subjected to lateral bending individually and may not behave as a composite material
with the marine clay.
Hence, a robust DSM columns design should be used such that the high lateral active
forces can be effectively transferred along the longitudinal DSM columns and the devel-
oped tensile and compressive stresses are within the acceptable strength limit.

5. DEEP SOIL MIXING COLUMNS AS RETAINING STRUCTURE

The optimum arrangement of DSM columns design is shown in Figures 5 and 6. It will
form a continuous arch from the South Bank to the North Bank. The North Bank will act
as a passive resistance to resist the high lateral force induced from the South Bank. This
configuration not only utilises the high compressive strength of the DSM columns but also

Figure 5. Cross section of DSM columns for temporary condition (Section C-C).
 Deep Soil Mixing Columns as Retaining Structure 161

isolates the development of tensile stress.

The first row of cross DSM columns will
experience the lateral force from the South
Bank before it arches along the longitudi-
nal DSM columns. The cross DSM columns
between the longitudinal DSM columns will
C C
be subjected to low stress since the passive
resistance of the very soft marine clay within
the DSM grid is low. Meanwhile, it also acts
as bracing to the longitudinal DSM columns
to ensure it will not buckle below the design
compressive strength.
Zone A, B and C will form a stable con-
tinuous arch where the lateral forces are bal-
anced against each other. The self weight of
this continuous arch is supported by Zone C
and this will reduce the future settlement.

6. 3D FINITE ELEMENT ANALYSIS Figure 6. Plan View of DSM Columns for tem-
porary condition at CH3850-CH3900.
PLAXIS3D TUNNEL is used because the
shear strength and stiffness of the DSM columns and soft marine clay are considered sep-
arately. The evaluation of unacceptable tensile points and excessive compression failure
zone at the effective overlapping width of 1 m diameter DSM columns in grid pattern can
be simulated by 0.44 m thickness of predefined extruded planes. In addition, geotechnical
safety factor by shear strength reduction in 3D will be more accurate and representative
compare to the assume smeared strength in 2D modelling.
The cohesion and tensile cut-off point of DSM columns are assumed to be 400 kPa and
92 kPa respectively which is 50% and 11.5% of the design UCS = 800 kPa (Topolnicki,
2004 [1]). Construction surcharge 10 kPa and 20 kPa are assumed acting on the crest of the
design cut-slope during the temporary condition as shown in Figure 8.

Figure 7. DSM columns as a retaining structure.

162 Advances in Geotechnical Infrastructure

Longitudinal plane

0.44m

0.44m

0.44m width in grid pattern

Figure 8. 3D FE analysis at temporary condition.

Failure shear stress

Mohr-
Coulomb
points
Isolated
Tensile
points

Figure 9. Observed tensile points, Mohr-Coulomb plastic points and failure shear stress at longitu-
dinal DSM columns.

The outputs from 3D FE Analysis has shown the stress distribution along the longitudi-
nal plane of the DSM columns is just about adequate. The failure shear stress distribution
at Zone B is about 30% of the total area and isolated around 1% of the area of Zone A.
Development of shear stress at Zone C is lower compare to Zone A. Meanwhile, the plastic
Mohr-Coulomb points show that the deep seated failure plane is cut-off by the longitudi-
nal DSM columns at Zone A. The tensile points are only observed at the isolated location
as shown in Figure 9.
By assuming the DSM columns in grid pattern with effective overlapping width 0.44 m,
the 3D FE Analysis shows the longitudinal DSM columns have adequate shear and tensile
strength to function as a retaining structure. No breakage of DSM columns is anticipated
to be occured. The executed phi/c reduction giving a geotechnical safety factor about 1.32
and it is considered sufficient for the temporary condition.
 Deep Soil Mixing Columns as Retaining Structure 163

Table 1. Typical results for coring and UCS.

Core run Core run Length of

Core starting ending Length of sample
run depth depth coring TCR recovered UCS Remarks
(m) (m) (m) (%) (m) (MPa)

1 4.5 6.0 1.5 100 1.5 2.35 Top 1

2 6.0 7.5 1.5 100 1.5 2.12 Top 2
3 7.5 9.0 1.5 100 1.5 1.46 Top 3
4 9.0 10.5 1.5 100 1.5 2.16 & 2.72 Middle 1 & 2
5 10.5 12.1 1.6 100 1.6 2.42 Middle 3
6 12.1 13.7 1.6 100 1.6 2.93 Bottom 1
7 13.7 15.2 1.5 100 1.5 4.08 Bottom 2
8 15.2 15.7 0.5 100 0.5 6.11 Bottom 3

7. PERFORMANCES

Typical results for coring and UCS are tabulated in Table 1. It can be seen that the average
UCS of 2.93 MPa has been achieved, which is well above the design UCS 800 kPa.
The sequence of excavation was carried out from the South Bank and toward the North
Bank. Higher lateral force from the South Bank was pushing toward the North Bank ini-
tially and the DSM columns at the North Bank function more like a passive balancing force

DenseStratum
6mm

Figure 10a

Figure 10c
Figure 10b

Figure 10. (a): Inclinometers at North Bank; (b): Scene of construction at North Bank;
(c) Inclinometer A18.
164 Advances in Geotechnical Infrastructure

to the South Bank. Figure 10 shows the three numbers of inclinometers which installed at
the North Bank. The results indicate that the DSM columns are performing as a retaining
structure in an acceptable manner with minimal deflection. The actual deflection is much
less than the calculated value for the South Bank and only a maximum 6 mm was recorded
at the inclinometer A18.

8. CONCLUSION

The installation of 1 m diameter DSM columns in grid pattern with 0.44 m effective over-
lapping width has been proposed as a retaining structure for the waterway construction in
Punggol when 6 m of very soft marine clay with qc < 0.25 MPa was found at the South
Bank within CH3790-3900 by CPTs. In order for DSM columns to function effectively as a
retaining structure, the high lateral forces induced from the South Bank must be transferred
to the North Bank through the longitudinal DSM columns which act as a continuous arch.
This proposed configuration not only fully utilises the advantage of the high compressive
strength but also isolates the development of tensile stress.
3D FE Analysis is a useful tool to discern the local variation of the tensile and relative
shear stress distribution along the longitudinal DSM columns at Zone A, B and C. The
results have shown that the proposed configuration will perform satisfactorily with accept-
able relative shear stress and isolated tensile stress. The acceptable distribution of relative
shear stress and isolated tensile stress along the longitudinal DSM columns show that the
grid arrangement works well as a retaining structure.
The installed inclinometers have confirmed that the grid arrangement is performing
effectively with only maximum 6 mm of deflection was registered by the inclinometers.
No sign of distress at the slopes and DSM columns were observed during the slope exca-
vation for the concrete drain construction.

REFERENCES

1. Topolnicki, M., “In situ soil mixing”, Ground Improvement (2004), pp. 331–428.
2. George M. Filtz, “Load transfer, settlement and stability of embankments founded on columns
installed by deep mixing methods”, Geotechnical Engineering Seminar Presentation (2007), Greece,
pp. 1–35.
3. Thanadol Kongsomboon, “The rehabilitation of drainage canal embankment”, Ground Improve-
ment Technologies and Case Histories, Singapore (2009), pp. 171–176.
4. Karastanev, D., Kitazume, M., Miyajima, S. and Ikeda, T., “Bearing capacity of shallow foundation
on column type DMM improved ground”, Proceedings of the 14th International Conference on Soil
Mechanics and Foundation Engineering, Sitzungsberiche Hamburg (1997), pp. 1621–1624.
5. Terashi, M. and Tanaka, H., “Bearing Capacity and Consolidation of Improved Ground of Treated
Soil Columns”, Report of the Port and Harbour Research Institute, Yokosuka Japan (1983b),
pp. 213–266.
6. Yong Shao, Emir Jose Macari and Weiming Cai, “Compound deep soil mixing columns for retain-
ing structures in excavations”, Journal of Geotechnical and Geoenvironmental Engineering (2005),
pp. 1370–1377.
7. SLIDE5, 2D Limit Equilibrium Slope Stability for Soil and Rock Slopes.
8. PLAXIS 2D (Version 9.02) & 3D TUNNEL (Version 2).