11/7
Pilot Test for Soil Stabilisation at Changi Airport.
V. CHOA Executive Engineer, Port of Singapore Authority, Singapore
G.P. KARUNARATNE Lecturer, University of Singapore, Singapore
S.D. RAMASWAMY Senior Lecturer, University of Singapore, Singapore
A. VIJIARATNAM Engineering Consultant, Port of Singapore Authority, Singapore
S.L. LEE Professor of Civil Engineering, University of Singapore, Singapore
SYNOPSIS The second runway of the new international airport at Changi in Singapore is situated on land reclaimed
by sand dredged from the sea. Soil investigation prior to the reclamation indicated that the runway is underlain
by several deep channels of soft marine clay up to a depth of about 40 metres. In order to minimise the problem
of uneven settlements, soil treatment was necessary for the runway and its associated high-speed turnoffs and
taxiways. A pilot test was carried out to compare the performance and feasibility of vertical sand drains,
flexible drains, surcharge and dynamic consolidation. This paper describes the different techniques tested and
reports the results of the insitu testing, laboratory work and instrumentation.
INTRODUCTION clay except that the new marine clay has numerous sand
A major hydraulic fill reclamation project was lenses and shell fragments whereas the old marine clay
undertaken by the Port of Singapore Authority acting has fine silt partings. Both clays are laminated.
as agents to the Government of Singapore in 1976 in Fig. l summarises the basic soil properties of the
order to reclaim 645 hectares of land for the construc- marine clay in a typical borehole. The coefficient of
tion of the new international airport at Changi in the consolidation Cv of the marine cla2 as determined in
Republic of Singapore. The 3.4 km long second runway the 75mm diameter oedometer is l m /yr for ~n effec-
is situated on the reclaimed land. Soil investigations tive over-burden pressure range of 100 kN/m to 200
prior to the reclamation indicated that the runway is kN;m2.
partly on hard ground of sandstone and shale and part-
Water Undrained
ly on thick deposits of soft marine clay. The marine c Soil Vertical Prossurt ____££___
clay exists in depths up to about 40m below the seabed (J Contont Shtar Strength
~ KN/m2 1 • eo
and is deposited mainly in deep channels. Long term Type % KN/m2
E 2 40 60 80 20 40 100 2 0 0.2 0.4
settlements caused by the 6.5m of hydraulic fill and
the runway pavement load were estimated to be up to f-2.- ;:< MN~ 0
2m. In order to minimise the problem of uneven settle- -- ~ARINE
I~ ~~
~ \
ments and heavy maintenance work on the runway soil p Pc.::!. 1.2
0p
treatment was necessary for the runway and its asso- CLAY
•= t- Po
ciated high speed turnoffs and taxiways. A pilot test ""~~
.:!.Q
;.:~ ~ (i)C:
was carried out to compare the performance and feasi- i 18i
bility of several soil improvement techniques such as 0
-- ~= :! (i)
\ ··~
vertical sand drains, flexible drains, surcharge and
dynamic consolidation. On the basis of the pilot test MARINE ~
.:Zll -- ~~ .......
o--1
a combination of flexible drains and surcharge was CLAY ......,
chosen as the most appropriate method of soil treat-
ment. The soil improvement works included compaction
of 460,000 sq m of hydraulic fill of 6.5m thickness
--
-- ~~~ - p G
and consolidation of 275,000 sq m of marine clay with --
... I~f:: ~~ (
an average depth of l8.5m. The contract price for the ~ ~:':DE~~
improvement works including the pilot test was S$11.7 ·:;.lck~ND '\
million (approximately US$5.2m). Wp w WL A Vano Tost Po: offtctivt ovtrburdon
(undisturbed) prossurt ( botoro till)
Wp Plastic Limit
SOIL PROPERTIES c Vant Tost
v Critical Prtssurct Pc
w~ Liquid Limit
Cc
The central portion of the sand fill is underlain by w Waltr Con !tnt
(romoldtd) Moan ~= 0.36
about one metre of sandy marine clay followed by a Fig. l Soil Properties of Marine Clay
dense sandstone and shale. The northern and southern
ends of the hydraulic fill on either side Qf this hard SPECIFICATION OF TREATMENT
central zone are underlain by thick deposits of marine
clay. The marine clay exists in two layers which are .The specificat~on requires 100% primary c?nsolida-
separated by a transition zone which consists of either tlon under a des1gn load of 100 kN;mz. An 1ncrease in
a soft sandy peaty clay or a firm mottled red and brown shear strength equal to 30% of the increase in vertical
fissured silty clay which was probably the result of effective stress and a differential settlement of not
dessication of the old marine clay when the sea level more than Scm within any distance of lOOm along the
fell. There does not appear to be a great difference length and breadth of the treated area over a period
between the properties of the two layers of marine of one year commencing from the completion of the works
141
11/7
were specified. The contractor was to be held res- Table 1 Details of Dynamic Consolidation Pilot Test
ponsible for the design of the soil improvement works
based on certain design guidelines. The following
soil parameters were specified: Ch = 3Cv(lab) = 3 Zones Del OC2
m2;yr for non-displacement drain, Ch = 1.5Cv(lab) = Phases 1st 2nd 1st 2nd
1.5 m2;yr for displacement drains, where Ch = coeffi-
cient of consolidation with the flow of pore fluid in Pounder weight (ton) 35 35 35 12
the horizontal direction and Cv(lab) = coefficient of Height of drop (m) 28 28 28 24
consolidation with the flow of pore fluid in the ver- Nunt>er of blows 8 to 10 10 to 15 6 to8 15 or 20
tical direction as determined in oedometer tests. Print grid (m x m) 7 .5xl3.0 7.5xl3.0 lO.OxlO.O 6.0x6.0
The contract also contained the provision that Total energy{tmtmZJ 90 125 70 120 or 160
should any contractor wish to offer an alternative
proposal to vertical non displacement sand drain
method he must demonstrate his method by carrying out
a pilot test to compare the results of his proposed
alternative with that of vertical sand drains. A
period of four months was allowed for the pilot test. in the removal and disposal of the spoil from the
boreholes were an additional disadvantage. Because
PILOT TEST of the above reasons and particularly the slow rate
Dynamic Consolidation of progress, the sand drain test area was reduced and
it was considered necessary to test an alternative
One of the alternative methods proposed was Dynamic type of vertical drain.
Consolidation. This method consisted of tamping the
ground with very heavy pounders of up to 40 ton with Flexible Drains
a free fall up to 40m. Piezometers, horizontal pres-
sure cells, inclinometers, settlement gauges and in- After considering the many available varieties of
situ strength tests were used to check the perfor- flexible drains in the market it was decided to test
mance of this method. Two areas designated DC 1 and the flexible drain known as Geodrains. This drain
DC 2 were tested using the dynamic consolidation consists of a corrugated plastic core surrounded by a
method. The details of the tests are given in Table specially designed filter paper sleeve. The cross
l. section of the drain is approximately lOOmm x 4mm
equivalent to a circular drain of 50mm diameter for
After the lst phase of compaction in DC 1, no signi- design purposes (Hansbo and Tortensson, 1970). The
ficant effect of dynamic consolidation was noticed. main reason for its choice was the availability of
Geodrains at 6.5m x 7.5m grid were then installed to good case histories on the performance of this type
assist the technique. There was however no signifi- of drains at depths similar to those at Changi. The
cant difference to the pore pressure response and pilot test area for the Geodrains is shown in Fi·g. 2.
settlement in the 2nd phase of dynamic consolidation Section A-A shows the profile across the pilot test
carried out with Geodrains. In both areas the maxi- area in a typical marine clay valley with new and old
mum rise in the pore pressure was only between 15 kN/ marine clay sandwiching a silty clay transition zone
m2 to 25 kN/m2 at a depth of about 5m below the sur- and overlying cemented sand. Numerous sandy lenses
face of the marine clay. The geocells indicated very in the young clay and large pockets of sand in the
little change in horizontal pressures and the few form of old river terraces exist on either side of the
checks of compaction next to inclinometers indicated old marine clay valley. A surcharge of 5.0m of sand
virtually no horizontal displacement of the marine was placed on the 6.5m of hydraulic sand fill after
clay. Surface settlements measured on the fill were: the installation of the drains. Based on the assump-
20cm after the lst Phase in both zones and 12cm and tion of Ch = 3Cv(lab), it was decided to adopt a
20cm after the 2nd Phase in DC 1 and DC 2 respectively. square grid with a drain spacing of around 2.5m.
These settlements were obviously due to the compaction Geodrains were then installed with three different
of the sand and partial consolidation of the marine spacings 3.2m, 2.6m and 2.lm on square grids in areas
clay. It was not possible to separate the two pro- GD 1, GD 2 and GD 3 respectively. An area ND was
cesses due to excessive disturbance caused by the deliberately left without any drains. Pneumatic pie-
impact. Insitu vane tests carried out before and zometers, surface settlement plates, deep settlement
after dynamic consolidation showed slight increase in gauges and inclinometers were used to monitor the
shear strength to depths of 5m only below the surface performance of marine clay with Geodrains as shown in
of the marine clay and virtually no change at depths Fig. 2. The drains were installed within a 17cm dia-
below 12m. Reviewing all the data gathered from the meter mandrel which was introduced into the ground by
DC Pilot test it was concluded that Dynamic Consolida- a vibrator the whole system being supported by a crane
tion was not effective for consolidation of the marine mounted leader. The whole system was capable of in-
clay. stalling drains to depths of up to 43m and an average
production of 2200m of drains per rig per 10 hour day
Sand Drains was achieved.
Sand drains were installed by a non displacement The field instruments in various zones of the pilot
method using a continuous flight auger to bore a hole test area were monitored for about 200 days at the
within a casing of 40cm diameter, filling the hole with time of writing this paper. Fig. 3 illustrates the
sand and simultaneously extracting the casing care- settlement records and pore pressure readings of few
fully. This proved to be a laborious and time con- piezometers installed in areas GD 1, GD 2, GD 3 and
suming method especially when lengths of drains of 20m ND. The settlement observed in GD 3 occurs at a
to 40m were involved. A crane mounted auger rig with faster rate than the settlements in GD 1 and GD 2,
a telescopic kelly bar was also tried and found to be which shows the influence of the drains at closer
equally inefficient at greater depths. Difficulties spacings. However, it is interesting to note that ND,
142
11/7
fev 0 10
Scale
20 30m
NO 5
Ill
'
GD3 ~
...
·-&:.
~ .~
B
A
u
0
.._:0
~ a...
fil ~ 12
+ 10.40m A CD
0
Ill Sottlement Plate GO Goodrains
A Deep Sell lement S0 Sand Drains 16
Gauge
NO No Drains 0
0 Piezometer
II)3 Sand
u 14
<{
' Inc I inomotor 8 Clay e
~ Silty Clay 12
"'
0
....
<{
> 10
w
...J
w
8
~
a:
....
w
::IE G03(Yl
0
N
w 4 WL
Q. LOCATION OF PIEZOMETERS ( mACD l
G01(Y) -5.23 GD2tOI-16.25 NO(Y) -4.00
SECTION A- A
G D 1 ( 01 -16.56 GO 3 (Y)- 6.70 N 0 (0) -16.90
GD2IYI -4.46 GDJ 101-21.62
DRAINS 0
20
E
" 40 NO
.... WL Water Lovol
"'
w 60
::IE
GO Geed rains NO
GD1
w NO No Drains G 02
...J
....
....
(0) Old Marino Clay
w 80 (Y) GDJ
New Marine Clay
"'
SECTION B- B
Fig. 2 Pilot Test Area Fig. 3 Settlement and Piezometer Readings
without drains, shows as much settlement as GD 3. straight lines corresponding to values of Ch ranging
This phenomenon of ND is attributed to the consolida- from 3 m2;yr to 9 m2;yr. Using the results obtained
tion settlement taking place with the flow of water from the pilot test described above the main works,
in clay vertically to the nearest horizontal drainage where applicable, was designed to have Geodrain
layers naturally occurring within the deposits. It spacings varying as follows: (a) 3.2 m square grid
is also observed that most of the settlement is con- upto 15m depth of clay or to the bottom of the new
tributed by the softer new marine clay having a low marine clay layer, whichever is the lesser depth, and
pre-consolidation pressure, as shown in Fig. 4. (b) 2.5m square grid to a depth exceeding 15m or below
The piezometric readings have been used to analyse the bottom of the new marine clay. In both cases a
the field performance of the marine clay with Gee- surcharge of 4m of sand was placed above the ground
drains driven in place. Taking the maximum recorded level taken as +5.5m ACD. The degree of consolidation
piezometric reading as the pore pres~ure at zero time, that would be achieved at the end of the construction
the average degree of consolidation U is shown in period was calculated by taking Cv = l m2;yr, the
Fig. 5. The pore pressure readings tend to fall on laboratory value for marine clay, and Ch = 3 m2;yr to
143
11/7
6 m2;yr. The degree of consolidation due to vertical
flow, Ov• was calculated on the assessment of drainage
SETTLEMENT lcml
layers observed in borehole samples. The degree of
10 20 30 40 50 60
consolidation 9ue to horizontal flow, oh. was then
combin~d with Uv, to obtain total degree of consolida-
tion, U, as shown by Carillo (lg42}.
-2
CONCLUSIONS
Dynamic Consolidation was ineffective for consoli-
dation of the marine clay at Changi. Non-displacement
type sand drains proved uneconomical due to the slow
speed of installation beyond 20m depth of clay. The
continuity of sand drains at great depths is also
questionable.
Geodrain,
Flexible drains were successfully installed to
2.6 m spacing depths exceeding 40m. The presence of permeable
GD 2
inclusions have been recognised in the marine clay
and assessed for their assistance in dissipating pore
pressure. Based on a pilot test conducted under
actual field conditions the designers were able to
adopt a very economical design for soil treatment.
The saving in cost achieved far exceeded the cost of
051 & 052 :3019177 l the pilot test which only amounted to about 4% of the
Installed Gouges [ SP61 130111177)
total cost of the soil improvement.
-18 lnstolltd Drains 6112177 to 23112177
The coefficient of consolidation Ch back-calculated
Placed Sm Surcharge 1011178 to1412178
from the field data appears to be a function of the
Sand 1 Settlement drain spacing and the fabric of clay. Estimation of
• Plato
Soft Morino Cloy Ch under actual field conditions is thought to be
Soft to Firm
Silty Cloy 1 Deep S.ttltment more realistic than that obtained from insitu per-
fa Gougo meability test method.
Cemented Sand
ACKNOWLEDGEMENT
The authors wish to acknowledge the assistance of
Fig. 4 Deep Settlement Gauges Messrs Techniques Louis Menard (S.A.), contractor for
soil improvement works, and Mr. R. Radhakrishnan,
Executive Engineer, Port of Singapore Authority, in
collecting certain laboratory and field soil data
pertaining to Changi marine clay.
REFERENCES
0.5 Carillo, N. (1942) "Simple Two and Three-dimensional
Cases in the Theory of Consolidation of Soils",
1:::::> J. Math. Phys., Vol. 21.
c Hansbo, S. and B.A. Tortensson {1977), "Geodrain and
0 03 Other Vertical Drain Behaviour", Proc. 9th Int.
c; Conf. S.M.F.E., Tokyo, Vol. 2, pp. 533-540.
~
0
Ill
c
0
u
-..
0 0.1
... Geodrain,
..
0
Cl
3. 2 m spacing
GO 1
5.0 100 200
Time, days
Fig. 5 Average Degree of Consolidation
144
