Pile Performance in Weathered Meta-Sedimentary Formation and

KL Limestone
S.S. Liew
G&P Geotechnics Sdn Bhd, Kuala Lumpur, Malaysia
C.M. Khoo
G&P Geotechnics Sdn Bhd, Kuala Lumpur, Malaysia
S.T. Tan
G&P Geotechnics Sdn Bhd, Kuala Lumpur, Malaysia
Y.E. Loh
G&P Geotechnics Sdn Bhd, Kuala Lumpur, Malaysia

ABSTRACT: This paper presents the performance of three preliminary instrumented bored cast-in-situ test
piles socketted in weathered meta-sedimentary formation and limestone formation in Kuala Lumpur, Malaysia.
The test piles were instrumented with Global Strain Extensometer technology and were load tested under
quick maintained load test with two piles under compressive axial loading and one under tensile loading. For
the compression bored pile embedded in soil by wet hole construction, about 8.7% of the test loads was
transferred down to the pile base, whereas the short rock socketted pile constructed with soft toe in wet hole
condition has about 38% of the test loads transferred to rock socket. Unconfined compressive strength tests
and point load tests on both the collected rock cores during boreholes exploration and the rock fragments
recovered during pile construction were correlated with mobilised rock socket resistances. However, the
correlations are scattered and do not have a clear trend.

1. INTRODUCTION

Three test piles presented in this paper are all mobilisation and load distribution for the similar
bored cast-in-situ piles of 1500mm and 1800mm geological materials with similar pile construction
diameter for compression piles and 1200mm practice and use of bentonite as stabilising medium
diameter for tension pile. For the two compression for the test piles construction.
load tests, a total of 2494 nos. concrete blocks have
been stacked-up for maximum test load of 4560
Tons with the height of approximately 20m above
ground level. Two rows of 32 temporary steel pipe
piles were driven to support the dead weight of the
kentledge system and provide stable platform for
pile testing (Fig. 1).
The required reaction load for tension test pile
was provided by four reaction piles of 750mm
diameter installed to 21m below testing platform. Fig. 1 4560 Tons Maintained Static Pile Load Test &
The reaction piles were installed at a distance of 1120 Tons Tension Pile Load Test
approximately 3.5 times test pile diameter away
from the centre of the test pile. The maximum test
load of 1120 Tons was applied to the tension pile 2. GEOLOGICAL FORMATION
using four 1000 Tons capacity hydraulic jacks.
This paper is aimed to present the test results of The subsurface investigation revealed that the site
these three preliminary instrumented bored piles is located at the geological contact of two
installed in the weathered meta-sedimentary formations, namely weathered meta-sedimentary
formation underlain by Kuala Lumpur Limestone. formation, locally known as Kenny Hill formation,
The results from these test piles provide some overlying Kuala Lumpur Limestone formation.
understanding of the pile behaviour in terms of The thickness of Kenny Hill formation at this site
shaft friction mobilisation, end bearing ranges from 15m to more than 60m. It consists of
mostly highly decomposed Grade IV Table 1 Summary of Pile Installation Records
metamorphosed sandstone/shale and completely Test Pile Ref. A B C
weathered Grade V residual soils. The overburden Pile Diameter 1500mm 1200mm 1800mm
alluvial materials generally consist of silty or Drilling System BG22/31 CMV BG45
gravelly SAND. Stabilising Fluid Bentonite Bentonite Bentonite
Temporary Casing 7m 7m 14m
Length
3. PILE INSTALLATION AND Drilling Speed (m/hr) in Respective to Geological
INSTRUMENTATION Formations
Kenny Hill 3.7m/hr 6.7m/hr 9.3m/hr
Figs. 2 to 4 show the nearest borelogs together Formation
with the details of three instrumented Test Piles A, Limestone N.A. N.A. 0.5m/hr
B and C. Test Pile A has a diameter of 1500mm Formation
and was tested with excessive pile top settlement. Duration of 1 day 1 day 1.5 days
Test Pile B is 1200mm diameter tension pile Drilling
having a penetration depth of 24m from piling Concrete Casting Tremie 2 Tremie 2 Tremie 2
platform level. The test piles were embedded Method
within the upper top layer of alluvium material of
sandy soils and the lower weathered meta-
sedimentary materials consisting of very stiff
sandy silt with SPT-N more than 50. Test Pile C
of 1800mm diameter has a penetration depth of
31m from piling platform level with 4m rock
socket length into limestone of Rock Quality
Designation (RQD) from 20% to 85%. A
“polystyrene foam soft toe” was installed at the
pile base for transferring load to rock socket with
minimum load interference from pile base.
All test piles were formed by auger drilling
through the overburden soils with bentonite slurry
for hole stability and concrete casting using tremie
method. The load tests were performed after the
piles had achieved minimum 28 days of designed Fig. 2 Test Pile A (φ1500mm)
concrete strength. Temporary steel casing were
driven to prevent borehole collapse for top 7m to
14m of loose alluvial soils. Drilling speeds in the
installation of test piles are summarised in Table 1.
The test piles were instrumented with
proprietary Global Strain Extensometer technology
(Glostrext method) using the access to the pile
shaft from the sonic logging tubes as presented by
Hanifah & Lee (2006). This system uses advanced
retrievable pneumatically-anchored extensometers
coupled with high-precision spring-loaded
vibrating-wire sensor with a simple analytical
technique to monitor loads transferred down the
shaft and the toe of test piles. It is a post-
installation instrument which can accurately Fig. 3 Test Pile B (φ1200mm)
measure the relative deformations of anchored
segments across the entire pile length.
 Fig. 7 Double Zinc Fig. 8 Mobilised Unit End
Sheets wrapped Bearing for Test Pile A
around Test Pile A

Fig. 7 shows the double-layer zinc sheets with

lubricant grease in-between wrapped around the
pile perimeter at the last 1.2m from the pile toe of
Test Pile A for base resistance measurement by
Fig. 4 Test Pile C (φ1800mm) eliminating the influence of the pile shaft friction.
The high pile base movement as shown in Fig.
8 has led to gradual mobilisation of the end-
4. INTERPRETATION OF TEST PILE bearing but without attaining failure. Pile base
RESULTS resistance attained 890kN/m2 or 8.7% of the total
test load with settlement of 50mm at the pile base
4.1. Test Pile A – Compression Pile implies possible existence of soft toe by deposition
of bentonite cake or base softening as the test pile
Figs. 5 and 6 show the load displacement curve was constructed with only normal base cleaning
and mobilised unit shaft friction for Test Pile A. procedures.
The maximum top pile settlement at the load of
10100kN at the first cycle was 12.04mm. In second 4.2. Test Pile B – Tension Pile
cycle, excessive movement of 54.77mm was
observed at the test load of 18071kN and Test Pile B is a tension pile reinforced with 26 nos.
evidenced that the pile has almost fully mobilised 40mm diameter high yield steel bars for taking the
attaining the ultimate shaft friction. tensile test load. The load displacement curve and
The interpreted pile test results show strain mobilised unit shaft friction are shown in Figs. 9
softening at the upper alluvial soil layers between and 10. In first cycle, the pile top displacement at
ground level and Level B but not at the lower tension load of 4500kN was 4.02mm. For second
alluvial soils. The large displacement of pile shaft and third cycles, the maximum displacements are
of about 11mm has led to reduction of shaft 17.15mm and 28.98mm at tension load of 9253kN
friction at the upper portion of pile. Strain and 11227kN respectively.
softening of pile-soil interface may affect the Although Test Piles A and B were located at the
distribution of skin friction down to the pile toe same meta-sedimentary formation, the test results
and cause the reduction in shear transfer between are varied significantly. Performance in term of
the soil and pile interface, from peak value to shaft friction for Test Pile B was less satisfactory
residual value. as compared to Test Pile A. Ultimate mobilised
shaft friction for the pile shaft was also scattered.
Maximum ultimate shaft friction interpreted for
test pile was 180kN/m2 with the displacement of
15mm. The average ultimate shaft friction was
about 120kN/m2. The reduction in shaft resistance
could be attributed to the pile radial shrinkage from
the Poisson ratio effect.
Similar to Test Pile A, strain softening at upper
portion of alluvial soils was observed from the test
results and there is relatively large displacement at
Fig. 5 Load Displacement Fig. 6 Mobilised Unit
the upper soil. Subsequent strain softening was
for Test Pile A Shaft Frictions for Test
observed at Level B to C and Level D to E. The
Pile A
result shows the tendency of progressive failure of
the pile. Low pile axial stiffness was observed
under tensile loading as most of the tensile test
load was taken by the steel reinforcement at higher
axial strain, probably cracking the concrete. Sonic
logging results showed signal delay in first arrival transferred to the pile toe, in which only about 5%
time and decrease of signal energy during tension of the total test load reached the pile base.
load test at top 15m of test pile as compared to the From the results of mobilised pile shaft friction,
results before testing, which further confirm tensile about 38% of the applied test load was transferred
cracking of concrete at the pile body. down in the form of rock socket friction. Fig. 12
shows that the maximum mobilised socket friction
was 1670kN/m2 between Level F and Level G
where the RQD value of limestone is about 20%,
whereas the maximum mobilised socket friction
was 1400kN/m2 for lower rock socket with RQD
value of 85% from Level G to Level H. The
relatively large displacement at the soil layer as
compared to the rock socket length suggests that
probably a slip along the pile shaft might have
occurred during the load test. When the slip
Fig. 9 Load Fig. 10 Mobilised Unit occurred, soil friction along pile shaft will most
Displacement for Test Shaft Frictions for Test probably maintain unchanged if not reduced
Pile B Pile B slightly due to strain softening effect and the
subsequent imposed test load will be transferred to
the rock socket friction, which tally well with the
4.3. Test Pile C – Compression Pile measured test results as demonstrated in this case.
Fig. 11 shows the load displacement curve for Test Lower value of ultimate mobilised shaft friction
Pile C which is a rock socketted pile with an of 63kPa was observed at the slump zone layer
artificial ‘soft toe’. A typical soft slump zone layer with average SPT-N of 17 blow counts (RL10m to
was observed at 4m above the limestone bedrock RL16m) as shown in Fig. 12. This result reveals
interface, probably due to the natural leaching that more careful assessment is required during
process and migration of fine soils from upper design and construction of bored pile foundation in
overburden soils into the lower relatively porous limestone formation with evidenced existence of
limestone formation. slump zone.

4.4. Summary
A summary of mobilised pile shaft frictions and
critical movements at pile-to-soil interface for
every soil strata is tabulated in Table 2. From the
test results, a similar trend of skin friction is
observed for all the test piles. The ratio of
mobilised unit skin friction, fs,mob to average SPT-
Fig. 11 Load Displacement N reduces exponentially with the increasing depth
for Test Pile C as well as with increasing SPT-N values as shown
in Fig. 13. All the instrumentation results clearly
show higher fs,mob/N ratio for alluvium soil with
lower average SPT-N values as compared to
Kenny Hill formation with higher SPT-N values.
Note:
unit shaft This may probably due to the installation of
friction not
fully mobilised temporary steel casing to prevent collapse of loose
sandy material at upper alluvial soil layer, in which
the actions of vibro-installation may provide
Fig. 13 fs,mob/N ratio Fig. 12 Mobilised Unit
sufficient compaction effort in densifying the
vs. SPT-N Shaft Frictions for Test upper loose alluvial soils. In weathered Kenny Hill
Pile C formation of stiff to very hard soil consistency, the
fs,mob/N ratio for shaft friction under compressive
load is fairly consistent, in which the ratio ranges
The maximum pile top settlement at the first between 1.5 and 2.2.
cycle test load of 22400kN was 8.17mm. In second
cycle, the movement of 20.39mm was observed at
the maximum test load of 45630kN. The ‘soft toe’
was seen to be performing well as noted from the
low strain readings implying minimum load
Table 2 Summary of Mobilised Unit Shaft Frictions and especially for anisotropic rocks. Hence, errors up
Critical Movements at Pile-to-Soil Interface to 100% are not uncommon if an arbitrary ratio
Critical Remarks value within the typical range is chosen to predict
f s,mob f b,mob
N Movemen UCS value from point load tests.
Test Soil N N t, Zc (mm) As such, 13 out of 51 rock core samples were
Pile Stratum (%)* sent for point load tests (PLT) to establish the
correlated ratio between the UCS and Is(50) for the
7 11.4 - 4 (0.3%) # limestone found at this site. The UCS tests and
Alluvium PLT results for the collected rock cores during the
15 7.3 - 6 (0.4%)
A 130 2.2 - 12 borehole exploration are shown in Table 3, in
(c) Kenny (0.8%) which the range of calibrated conversion factor is
Hill 150 1.8 5.3 6.5 13.3 ± 7.7 with considerable scatter of data. For
(0.4%) design purpose, a linear relationship of 13.3 was
Alluvium 7 10.7 - 3.5 # adopted for the correlation of UCS with Is(50) for
(0.3%) limestone bedrock as per Eq. (1).
48 3.3 - 6 (0.5%) #
B
60 3.0 - 9.5 UCS = 13.3 Is(50) (1)
(t) Kenny
(0.8%)
Hill
84 1.4 - 3.5 where UCS = unconfined compressive strength
(0.3%) for rock; Is(50) = point load index for 50mm
Alluvium 7 14 - 9 (0.5 %) Pile-soil diameter core.
72 1.6 - >10 interface &
(0.6%) socket not Table 3 Summary of Unconfined Compressive Strength
C fully
Kenny Tests and Point Load Tests Results
(c) mobilised.
Hill
Is(50) Conversion
Slump RQD UCS
BH (N/m Factor,
17 3.7 - 5 (0.3%) zone # (%) (N/mm2)
m2) K = UCS/Is(50)
Notes: (c) represent compression pile; (t) represent
tension pile; * represent percentage of critical P3/BH4 85 1.94 39.6 20.4
movement as compared to pile diameter; # strain P5/BH1 30 4.24 46.4 10.9
softening observed. P5/BH1 85 3.97 61.0 15.4
P5/BH2 14 5.41 51.9 9.6
However, for tension pile, the fs,mob/N ratio in P5/BH3 37 8.40 53.5 6.4
Kenny Hill formation ranges from 1.4 to 3.3. The P5/BH3 63 5.20 81.9 15.8
critical movement at pile-to-soil interface is P5/BH4 24 4.93 45.6 9.2
recorded at displacement of 0.3% to 0.8% of pile P5/BH6 45 2.00 56.7 28.4
diameter, which is relatively small. P5/BH6 8 4.93 30.2 6.1
For non-rock socketted bored pile, the critical P5/BH7 44 7.70 43.3 5.6
movement at pile-to-soil interface in between 0.4% P5/BH8 12 1.72 42.1 24.5
and 0.8% of the pile diameter is recorded in Kenny P5/BH9 40 3.50 59.0 16.9
Hill formation. For rock socketted pile with brittle P5/BH11 51 9.55 35.8 3.7
socket movement restricting the mobilisation of Sample Mean (MPa) x= 13.3
pile-soil interface resistance, hence there appears Standard Deviation (MPa) = 7.7
no sign of obvious yielding observed in the Coefficient of Variation
resistance mobilisation except for the slump zone, (COV) = 60%
in which the yielding occurs at 4 to 5mm Standard Deviation Range Max = 21.0
movement as shown in Fig. 12. for Conversion Factor, K Min = 5.6

The estimated UCS values correlated to point

5. CORRELATION FOR ROCK SOCKET load test results on the collected rock cores by Eq.
FRICTION (1) during coring of Test Pile C are shown in Table
4.
A total of 51 nos. rock cores were collected for Neoh (1998) recommended the maximum
Unconfined Compressive Strength (UCS) tests. allowable rock socket friction for Kuala
According to ISRM (1985), the UCS value of rock Lumpur Limestone in Malaysia shall be
can be correlated to Is(50) using conversion factor limited to 5% of either the rock UCS value or
ranging between 20 and 25. However, it is also characteristics concrete strength of pile,
reported from the test results of many different whichever is smaller as per Eq. (2) below.
rock types, the ratio can vary between 15 and 50
fs,all = 0.05 ∗ {min(UCS, fcu)} (2)
where fs,all = allowable rock socket friction; fcu = well with the approaches mentioned above and
concrete compressive strength at 28 days (35MPa). reveals that more optimistic performance of this
limestone rock socket can be readily expected.
Table 4 Summary of Point Load Tests and Mobilised
Rock Socket Frictions for Test Pile C socketted in
Limestone 6. CONCLUSION
#
Measured fs,all Ratio of
*UCS Ratio of This paper presents three instrumented preliminary
fs ,mob
(N/mm2) fs ,mob / UCS
(kN/m2) fs,mob /
(kN/m2) fs,all load test results on bored piles installed at alluvial
1670 14.18 0.12 709 2.4 deposits underlain by the meta-sedimentary Kenny
1400 33.60 0.04 1680 0.8
Hill formation and followed by Kuala Lumpur
* UCS is average UCS value estimated using Eq. (1) Limestone formation. The following conclusions
based on PLT results carried out on rock samples are made:
collected at every 0.5m spacing a. These test piles mainly utilise the frictional
# Allowable rock shaft friction is calculated using Eq. (2) resistance to support the designed capacity of
the pile with safety factor of at least 2.0. All the
The load transfer mechanism of earlier test piles were satisfied with the settlement
mobilisation of the upper portion of rock socket requirements.
friction to much higher value has rendered b. The contribution from mobilised base resistance
relatively lower degree of socket mobilisation at is generally low within the limit of acceptable
the lower rock socket under the maximum test settlement, therefore base resistance should not
load. From the above results, it is clear that rock be considered in the bored pile design unless
socket friction estimated by Neoh’s approach proper base cleaning can be carried out.
appears to be conservative. c. The ratio of mobilised unit skin friction, fs,mob to
Another method of designing the rock socket average SPT-N reduces exponentially with
friction commonly used in Malaysia is the increasing SPT-N values
approach developed by William & Pells (1981). d. The correlation of rock test results suggest a
conversion factor of 13.3 is fairly acceptable for
fs = α*β*UCS (3) estimating unconfined compressive strength of
limestone bedrock encountered in this site.
where fs = ultimate rock socket friction e. The application of point load test at field is
α = reduction factor with respect to UCS useful to predict the UCS values of rock
β = reduction factor with respect to the encountered during pile construction.
rock mass effect f. The test results evidence that the current
Malaysia design practice of rock socket in
limestone formation is somehow conservative.
The inherent safety factor in allowable stress
design is expected much higher than 2.0 as
commonly perceived in pile design. However
one shall always remember that failure of rock
socket is normally brittle.

7. REFERENCES
Fig. 14 Back-calculated α and β factors of Test Pile C Hanifah, A. A. and Lee, S. K., Application of Global Strain
Data plotted on William & Pells Approach Extensometer (Glostrext) Method for Instrumented Bored
Piles in Malaysia, 10th International Conference on Piling
The back calculated factors of and from the and Deep Foundations, Amsterdam, June 2006.
mobilised rock socket frictions for Test Pile C are
ISRM. International Society of Rock Mechanics Commission
plotted in Fig. 14. It is observed that the mobilised on Testing Methods, Suggested Method for Determining
socket friction at the lower part of the rock socket Point Load Strength, Int. J. Rock Mech. Min. Sci. and
is still lower than the ultimate socket resistance Geomech, Abstr. 22, 1985, pp.51-60.
estimated by William and Pells’ approach (1981).
Such outcome is expected as the socket friction at Neoh, C. A., “Design & Construction of Pile Foundation in
Limestone Formation”, Journal of Institution of
this lower portion of rock socket has not been fully Engineers, Malaysia, Vol. 59, No. 1, 1998, pp.23-29.
mobilised to its ultimate resistance at the William, A. F. and Pells, P. J. N., “Side Resistance Rock
maximum test load as yielding is not even reached Sockets in Sandstone, Mudstone, and Shale” Canadian
in Fig. 12. However, the trend of test results tallies Geotechnical Journal. 1981. pp. 502-513.