Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/335274359

A 16m Deep Basement Excavation in KL Karst: Mixed Retention Solutions

Conference Paper · March 2017

CITATIONS READS

0 393

3 authors, including:

Mikias Yohannes
Keller Holding GmbH
7 PUBLICATIONS 3 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Deep Vibro Techniques View project

All content following this page was uploaded by Mikias Yohannes on 20 August 2019.

The user has requested enhancement of the downloaded file.


A 16 M DEEP BASEMENT EXCAVATION IN KUALA LUMPUR KARST: MIXED
RETENTION SOLUTION

Prasad PVSR, Mikias Yohannes, Serhat Baycan


Keller (M) Sdn. Bhd.

ABSTRACT

Kuala Lumpur Karst is a complex limestone formation that is characterized by presence of cavities,
solution channels, pinnacles and floaters. Basement excavations in such geological formations are
usually challenging, mainly owing to the unpredictable rock head depths and potential for groundwater
inflow. Any excavation support system should have adequate robustness and water tightness.
Traditionally, Secant Bored Pile walls (SBP) with struts are the preferred option for karst ground. This
paper discusses a cost effective and innovative alternative solution for a particular site – using Deep
Soil Mixing (DSM) and using Contiguous Bored Piles (CBP) together with Ground Anchors (GA). The
site, located around one of the busiest areas of Kuala Lumpur, is a 5,700 m2 commercial and residential
development. A 16.0 m deep excavation was needed for four levels of basement and excavation support
along the 275 m perimeter of the development. Within the site, the depth to rock head varied from 6 m
to 19 m. The basis for each solution and conditions for its suitability are discussed in detail. Some design
considerations and construction challenges are also presented. The performance of the two retention
systems during excavation are presented, together with comparisons versus design predictions.
Suggestions for future development are also given. Advantages of this approach compared to other
traditional approaches are highlighted.

Keywords: Deep soil mixing, DSM, KL Limestone, karst, basement excavation, mixed solution

1. INTRODUCTION

The geology of Kuala Lumpur is among the most complicated and geotechnically challenging
formations in South East Asia. The majority of the city falls into the Kuala Lumpur Limestone
formation; which is a thick, highly-weathered and cavernous limestone. This formation is pure calcitic
to dolomitic in composition, having infrequent schist and phyllite interbedding (Gobbett et al., 1973).

The KL limestone has a highly karstic topography dotted with cavities, ravines, interconnected
subterranean channels, pinnacles, overhangs, floaters and slump zones (Tan, 2006). This results in
erratic bedrock profiles even within short distances. It is not uncommon to see a precipitous rock head
incidence within short distances in the same site. In most parts of the City, the bedrock is typically
encountered ranging from very shallow depths to about 30.0 m. The exception to this is at the heart of
the City, where Kenny Hill formation prevails above the KL limestone. In such cases the limestone
bedrock can be encountered much deeper. The erratic rock head profile and the highly karstic nature of
the KL limestone makes it a geotechnical challenge requiring careful consideration of suitable
engineering solutions.

1.1 – Site Description

The site, located in downtown Kuala Lumpur, is abutted by a busy street on the North. Before
development, the site previously served as a car park. To the west an access road separates the site from
an empty lot; while on the east, the lot is bounded by light weight residential and commercial buildings.
On the south side, the site is bordered by a road that separates it from a prominent Shopping Mall in the
City. A layout plan and an artist’s impression of the finished development is shown in Figure 1.

The site has an estimated area of 5,700 m2 and a total perimeter length of about 275 m. The development
is a mixed type consisting of a 56-storey commercial/SOHO tower with 4-levels of basement. The tower
foundation was designed as a 3.0 m thick raft resting on the bedrock. Hence, construction of the raft

215 DFI-PFSF 2017 Conference, Melbourne


and 4 levels of basement necessitated excavation up to maximum of about 16.0 m deep. Due to the
presence of limestone at depths shallower than the final excavation level across the larger part of the
site, the excavation works have to be carried out mostly in rock.

Figure 1. Site layout plan (left), and artist impression of the intended development (right)

1.2 – Subsurface Investigation and Soil Parameters for Design

A total of 23 boreholes were available within the site. From the soil investigation data, the subsoil
conditions are generally loose to medium dense, light grey silty Sand up to about approximately 7.0 m
from the ground surface, save for a few localized pockets of decayed material, construction debris and
soft slimes. Further below and overlying the limestone is a medium dense, silty Sand layer of grey
appearance and having traces of gravel. Figure 2 shows the scatter plot of field SPT-N blow count
values versus reduced levels for all the boreholes in the site. The idealized soil properties that was
adopted for design is also shown in the same plot as a line, following a moderately conservative
approach. The ground water table as indicated by the boreholes was approximately at 2.5 m below
original ground level.

Figure 2. Scatter plot of field SPT-N versus reduced levels for all the boreholes in the site

From the borehole data that was available, the rock head incidence was seen to range from as shallow
as 7.0 m up to as deep as 20.0 m below the ground surface. However, given the erratic nature of the

216 DFI-PFSF 2017 Conference, Melbourne


karst limestone profile, additional probing was necessary to be carried-out to determine the depth of the
rock more accurately for detailed design purposes. Based on the borehole information, the limestone in
the site varied from highly weathered limestone to sound limestone, exhibiting a wide range of rock
quality designation (RQD). Based on unconfined compressive strength (UCS) tests the bedrock can be
classified as medium strong to strong limestone, as the results ranged from about 20 MPa to 75 MPa.
Figure 3 shows the scatter plot of UCS values versus depth, and RQD versus UCS strength of the rock
samples that were tested.

Figure 3. Scatter plot of rock samples’ UCS value vs. reduced level (left), and plot of RQD vs. UCS of
rock samples (right)

2. DESIGN AND ANALYSIS OF RETENTION SYSTEM


2.1 – Design Considerations

Keller carried-out rock probing along the site perimeter to check the depth of the rock and establish a
rock head profile for design and construction purposes. A total of 65 numbers of probing points were
carried out along the four walls of the site. On the West and South walls, probing points staggered in
two rows were carried-out. While on the North and East walls points in single rows were considered.

The rock was found to be as shallow as 7.0 m and as deep as 18.0 m. On the North wall the rock profile
showed some ravines or depressions were the rock was deeper. On the East wall the rock was found to
be at an average depth of 9.0 m below the ground level. On the South wall the depth of the rock was
found to be more erratic, ranging in depth from 8.0 m to 16.0 m within short wall lengths. In addition,
based on the two rows of probing, it was concluded that the rock profile was slanting towards the
excavation. In contrast, the two rows of probing on the West wall showed almost uniform rock profile,
with an average of 8.0 m deep rock head. Figure 4 shows the rock probing results along the West and
South walls.

2.2 – Conforming Design

The conforming design during tendering stage was a soldier pile system with timber lagging. The
proposed soldier piles were UC steel profiles (UC 305×305×118 kg/m up to UC 305×305×283 kg/m)
with 2.0 m socketing into the limestone rock and placed at 2.0 m centre-to-centre. The proposed wall
would be restrained by a combination of removable ground anchors and corner strutting. The ground
anchors proposed were also at 2.0 m centre-to-centre and of temporary, removable type. These were
provided along the North, West and South walls; whereas, the corner strutting system was proposed
along the East wall (buttressed against North and South walls). The reason to adopt strutting for this
wall was due to inability to install ground anchors into the neighboring lot.

217 DFI-PFSF 2017 Conference, Melbourne


Some shortcomings in the conforming design permitted consideration of an alternative retention
solution. Timber lagging and soldier pile system would not allow a dry excavation, as it is not
watertight. This is especially crucial where the ground water table is high, such as the site in discussion.
Moreover, the system would not be able to retain the soil properly since the laggings cannot form good
contact or seal with the erratic rock head profile of karst formations. The strutting system proposed in
the conforming design also included 38 king posts of UC profiles (UC 254×254×73 kg/m) to support
the struts and cross members. This added more complications to the proposed system, as this would
mean that excavation below the strutting would be delayed due to the high numbers of king posts.

Figure 4. Probing point results of the rock head level at the West (top) and South (bottom) walls
2.3 – Selection of Retention Method

Due to the drawbacks in the conforming design, an alternative proposal was devised. A design using
deep soil mixing (DSM) for shallow rock head areas and contiguous bored piles (CBP) was proposed.
DSM block, as a gravity retention system, has been successfully used to support excavations in karst
ground up to 10 m retained height previously. Where DSM would not be applicable (deeper rock head
and/or space unavailability) it was proposed to use CBP and retrievable ground anchors. Cost
comparisons were made to ensure that the proposal would be competitive.

Before the designs were finalized, physical rock probing was decided to be carried out, as described in
Section 2.1. Due to space unavailability, DSM was ruled out for North and East walls (only 2.3 m was
available from boundary line to excavation line). Hence, for these walls only CBP option was explored.
Where space was available (West and South walls), the applicability of DSM was evaluated. Adequate
rock probing was carried out to map out the rock profile at these walls. Since the rock profile was more
consistent and shallower than 10 m, it was decided that DSM is an ideal application for the West wall.
For South Wall, since the rock depths ranged from 8.0 m to 19.0 m, it was deemed that DSM would not
be a suitable choice. Moreover, the rock profile showed that at certain sections the rock was dipping
towards the excavation; hence necessitating a wider DSM block. For these reasons CBP option was
preferred.

To address the water tightness issue of the conforming design, the proposed solutions provided suitable
means to prevent water seepage into the excavation. DSM treatment effectively makes the ground
impermeable as the cementation seals the voids in the soil matrix. The block treatment also cuts off
potential water seepage across the rock–soil interface, as the individual columns would have variable

218 DFI-PFSF 2017 Conference, Melbourne


depths due to the karst formation. The gap between contiguous piles was proposed to be sealed using
jet grout or jet mixed columns (JMC) having larger diameter than the opening. The jetting would ensure
the formation of the columns against the CBPs, plugging any potential water seepage. The size of the
JMC was selected such that the gap is adequately sealed after allowing for verticality deviations of the
CBPs.

Figure 5. Typical DSM cross-section and details (West Wall)


To allow a faster and less congested excavation, it was also proposed to use only retrievable ground
anchors as a means of retaining the CBP walls and avoiding the use of struts all together. As the adjacent
lot was semi-vacant with few light-weight barracks, it was proposed to the Main Contractor to rent the
land for the duration of the construction and install retrievable ground anchors into the lot. This was
more favorable to the Client as it afforded a time saving by allowing faster and unobstructed excavation.

Figure 6. Typical CBP cross-section and details (North and South Walls)

2.4 – Design and Analysis of Alternative Proposal

For the alternative proposal, preliminary design were carried-out by hand calculations – with the help
of in-house spreadsheets. For the detailed design stage, a number of plane strain finite element analyses
were done using the geotechnical engineering program Plaxis 2D.

2.4.1 – Soil Parameters

Based on the soil investigation data, the soil layers were modelled by using the Hardening Soil
constitutive model (HSM). Idealized soil profile as shown in Figure 2 was considered for design
purpose. Since the soil on the site is predominantly sandy, all soils were modelled as drained. Drained

219 DFI-PFSF 2017 Conference, Melbourne


angles of shearing resistance as shown in Table 1 were adopted for each soil layer, whereas the effective
cohesion was considered to be insignificant (taken as 0.5 kPa for modelling purposes).

Table 1. Summarized soil and rock properties


Soil Description c’ref ϕ' ψ' E50ref Eur/E m Rf
3 3
(HSM) [kN/m ] [kN/m ] [kPa] [deg] [deg] [MPa] [-] [-] [-]
Firm sandy SILT/
16.0 18.3 0.5 27 0 5.5 3.0 0.75 0.9
loose Silty SAND
Loose to medium
17.0 19.0 0.5 29 0 16.0 3.0 0.65 0.9
dense Silty SAND
M/dense SAND with
17.3 19.3 0.5 32 2 24.0 3.0 0.65 0.9
Silt/Gravel
Rock Description E' ψmax v' mi GSI D
(HB Model) [kN/m3] [kN/m3] [MPa] [MPa] [deg] [-] [-] [-] [-]
Limestone 23.0 24.0 36 30,000 0 0.25 9 50 0.7
= bulk unit weight of material, = saturated unit weight of material, c’ref = effective cohesion, ϕ’
= effective angle of shearing resistance, ψ’ = angle of dilation, E’ = Young’s modulus of material, v’ =
Poisson’s ratio, E50ref = triaxial secant stiffness for soils, Eur/E = moduli ratio of unloading–reloading,
m = HSM model power for stress level dependency of stiffness, Rf = limiting failure ratio, =
uniaxial compressive strength of rock, E’ = Young’s modulus of material, ψmax = dilatancy at zero
stress level, mi = Hoek-Brown material constant for the type of rock (intact rock), GSI = geological
strength index for the rock, D = disturbance factor for the bedrock as a result of excavation.

The limestone bedrock was modelled using Hoek-Brown (HB) rock model by assuming an isotropic
rock with no preferred jointing direction. The necessary input parameters were evaluated following
Hoek et al. (2002), Marinos et al. (2007) and Marinos and Hoek (2000). Intact rock properties were
evaluated first, and the laboratory or intact rock properties were reduced by incorporating the geological
strength index (GSI) to reflect actual or in-situ rock characteristics. Additional relevant factors were
also considered, such as the damage and stress relaxation effects due to rock excavation. Since the rock
excavation was to be carried out by a third party, a no-blasting zone of 5.0 m from the walls was
prescribed as a safeguard. Rock bolts were also prescribed if any unstable rock was noted on site during
excavation. The idealized soil and rock input parameters adopted for detailed analyses are provided in
Table 1.

The initial in-situ locked in stresses for the soils were considered based on Ko approach, where Ko was
taken as (1- sin ϕ’). In addition to these, the ground water table was considered to be 2.0 m below the
ground level. A surcharge load of 15 kPa was also considered on the retention side to account for the
live load on existing roads. Moreover, over-excavation for up to 0.5 m below the final excavation level
was considered. Seismic action was not considered in the design since Kuala Lumpur is not located in
earthquake prone area.

2.4.2 – Design for DSM

The width of the DSM block was decided based on rock head depths from rock probing results.
Attention was given to areas where the rock showed a slanting profile towards the excavation. For DSM
wall, no rock benching was considered in order to allow the direct casting of permanent wall against
the DSM wall, allowing cost saving of formwork erection and removal on the West wall.

The DSM block was designed for a target UCS of minimum 1.0 MPa. Such a value was specified based
on internal stability and stiffness requirements. Using a relationship of UCS to Young’s modulus of
(250 × ), a design stiffness of 250 MPa was used (where is the UCS target value of DSM samples,
in MPa). For lack of a better constitutive model, DSM was modelled as a non-porous MC model having
a cohesion value only. The DSM cohesion value is taken as = ⁄2. An allowance for tensile stress
up to 5% of the UCS strength was accounted for in the model. Detailed DSM properties that were
adopted for design are as shown in Table 2.

220 DFI-PFSF 2017 Conference, Melbourne


Table 2. DSM properties adopted for detailed design
Soil Description UCS cref ϕ v' E ,
(MC Model) [kN/m3] [MPa] [kPa] [deg] [-] [MPa] [kPa]
DSM (non-porous) 21.0 1.0 500 0 0.30 250 50
= bulk unit weight of material, UCS = unconfined compressive strength of material, cref
= cohesion, ϕ = angle of shearing resistance, E = Young’s modulus of material, v’ = Poisson’s
ratio, , = maximum allowable tensile stress.

2.4.3 – Design for CBP

Design of the CBP wall was carried out by categorizing the wall stretches based on rock head profiles
from the rock probing results. Based on the UCS strength values for the site, competent rock was
specified to be as rock having UCS strength greater than or equal to 36.0 MPa (which is a representative
value adopted on the moderately conservative side, refer Figure 3).

The CBPs were designed to be socketed minimum 2.0 m into competent rock. The side shear resistance
of the rock was determined based on Horvath and Kenney (1979) approach. If no rock was encountered,
the CBPs were continued for 5.0 m into the medium dense Silty Sand below the final excavation level.
Where rock head was encountered shallower than final excavation depth, rock benching of 1.0 m was
considered from CBP face. This allowed termination of piles encountering rock shallower than the final
excavation depth. The benching was provided so as to provide the piles with enough rock to develop
geotechnical capacity within the socket length and augment pile toe stability. It also acts as a buffer
during rock hacking, so that the CBPs are not compromised.

Among the walls where CBP was proposed, the space at the East Wall was more restricted. This was
due to the roof eave boards of the barracks at adjacent lot impinging into the site. Installation of bigger
CBP diameter would result in potential damaging of these quarters by the piling rig’s drilling tool. Thus,
a smaller diameter of CBPs was selected for this wall (i.e. 0.8 m vs. 1.1 m for North and South Walls).

FEM modelling of CBPs was done in 2D plane strain condition considering 5-noded isotropic and end-
bearing enabled plate elements. The input properties considered for the design are as shown in Table 3.

Table 3. Summarized CBP properties for FEM modelling


Dia. EA* EI* w v'
Type
[m] [kN/m] [kN.m2/m] [kN/m/m] [-]
CBP-T1 1.10 1.90×107 1.44×106 18.24 0.20
7
CBP-T2 0.80 1.32×10 5.29×105 12.70 0.20
Dia. = diameter of piles, EA = axial rigidity, EI = flexural rigidity,
w = weight per meter width per meter run, v’ = Poisson’s ratio.
* An effective long term stiffness of 80% was considered

Based on the member forces envelop from Plaxis, the structural design of the CBPs was carried out.
Axial loads on the CBPs were predominantly from the vertical component of ground anchor forces. The
resulting axial forces ranged from 800 kN to 1,500 kN per pile, depending on ground anchor levels and
retention height. The bending moments on the CBPs varied from 550 kN.m to 1,700 kN.m per pile.
Based on these, the CBPs were designed to have reinforcement contents of 0.8% to 1.8%. The CBPs
were designed for 40 MPa characteristic strength concrete.

2.4.4 – Ground Anchors

The CBP walls were designed to be supported by grouted ground anchors. The ground anchors were
designed using retrievable, 15.2 mm dia. high tensile steel (Y1860C) strands. As the overburden soil is
loose to medium dense sands, the fixed length of the anchors was designed to be socketed into the rock
to get the necessary pull-out resistance. To shorten the length, the anchors were designed with a 45o
inclination. The anchors were chosen to be spaced at alternate gap between CBPs (i.e. 2.50 m at North

221 DFI-PFSF 2017 Conference, Melbourne


and South Walls, 1.90 m at East Wall). Cement grout of 30 MPa characteristic compressive strength
was chosen for the ground anchors. Material properties of the different ground anchors that were used
is given in Table 4.

Table 4. Summarized Ground Anchor properties for FEM modelling


Free Length Fixed Length
Type Strands’ EA Spacing* Dia. EA EI v' fs
[kN] [m] [m] [kN/m] [kN.m2/m] [-] [kN/m]
GA6a 1.72×105 2.50 0.15 1.77×105 2.49×102 0.20 141.4
GA8a 2.30×105 2.50
0.18 2.54×105 5.15×102 0.20 169.6
GA10a 2.87×105 2.50
Dia. = diameter of ground anchor, EA = axial rigidity, EI = flexural rigidity, fs = limiting skin friction,
v’ = Poisson’s ratio.
* Data shown only for 2.5m spacing ground anchors.

2.4.5 – Rock Fissure Grouting

To prevent seepage of water through exposed rock face, drilling and grouting was necessary. A row of
curtain grouting was done in the rock at 3.75 m spacing around the perimeter of the excavation;
secondary and tertiary points were added at intermediary locations as necessary when volume and
pressure requirements were not met. The curtain wall was designed to be done from rock incidence up
to 10.0 m below the excavation depth. Base grouting was specified at the excavation area to prevent
seepage from the bottom of the pit. These points were designed to be carried out from final excavation
up to a maximum depth of 5.0 m.

3. CONSTRUCTION OF RETENTION SYSTEM


3.1 – Construction and Excavation Process

The drilling and grouting works were carried out before the other products. This allowed some
understanding on the nature of the bedrock in terms of depth and quality. The installation of the CBP
was facilitated by the depth information obtained from the drilling work. The nature and quality of
bedrock was also assessed based on the drilling information and grout intake. Figure 7 shows the spatial
grout intake plot of all drilling and grouting works. In total, about 300 m3 of grouting was carried out
for the project.

Figure 7. Spatial grout intake plot of all the drilling and grouting points carried-out

Based on this plot, it can be seen that on the North Wall a stretch of about 35.0 m was highly jointed
and weathered bedrock. The grout intake was so much that secondary and tertiary points were added to

222 DFI-PFSF 2017 Conference, Melbourne


the curtain grouting (refer Figure 8a for exposed rock). This zone extended into the excavation area, as
the neighbouring base grouting also recorded relatively higher intake (refer photo in Figure 8b).
Similarly, on the South-East corner, a stretch of about 15.0 m showed highly weathered rock profile.

To act as a homogenous gravity wall, the DSM block was required to be mixed without the development
of cold joints. The columns were mixed consistently and due attention was given to creating mechanical
interlocking between each shift of DSM production. Samples at selected columns were cored and tested
for UCS of the formed DSM. The minimum UCS value that was recorded was 1.60 MPa, which is
above the target strength.

On site, point load tests for rock samples retrieved from each CBP rock coring were performed and
point load strength indices evaluated ( ( ) ). The correlation between point load index strength and
UCS strength is given by a general formula = × ( ) , where k is a correlation factor, and for
KL limestone the suggested values are 7 – 20 (eg. Liew et al., 2010). Based on the point load test on
rock samples retrieved from each pile socketed in rock, the ( ) vs. UCS correlation factor for this site
was found to be approximately 13. When the inferred UCS values were lower than the design value,
rock socket lengths were made longer. Extra socket lengths were decided based on correlated point load
indices.

a b

c d

Figure 8. Exposed rock after excavation (a) massive to thinly bedded formation with wide
burrows (North wall, where high grout intake was recorded), (b) pinnacled and fissured limestone
exposed (middle of site), (c) fractured, thinly and vertically bedded formation (South wall), (d)
solution channel in massive rock formation at DSM toe (West wall)

Figure 9. Faster excavation on DSM wall is afforded as no additional supports are needed

Installation of the ground anchors was done progressively. Since the DSM wall does not require ground
anchors, the west part of the site was excavated at a faster pace. This meant the ground anchors’
installation at the abutting walls (North and South) had to be well coordinated with the excavation
progress (refer Figure 9). Installation of second layer ground anchors at the North wall initially caused

223 DFI-PFSF 2017 Conference, Melbourne


leakage through the anchor point. This was overcome by doing pre-grouting and post installation sealing
plug. As the ground anchors were temporary and retrievable type, they were destressed and removed
progressively. When a basement floor below a row of ground anchors was cast, those ground anchors
were removed.

The excavation process was carried out reasonably fast, as the early completion of the DSM wall
provided an area where excavation works can commence. Other than a few cases of leakage from
ground anchor installation points, the seepage through the retaining walls or base of the excavation was
insignificant. As such, dewatering of the water trapped inside the pit was done and the drawdown in the
ground water table outside of the pit was not noteworthy.

3.2 – Challenges in Construction

One of the challenges in construction for this site was due to unforeseen ground conditions. Peaty soil
with high organic content and wood pieces was encountered at a depth range of 3 m to 7 m in a localized
area within the DSM treatment (refer Figure 10). Treating the area with DSM was not possible; hence,
jet mixed columns were used to drill through the peaty soil, go up to the rock incidence and form mixed
columns.

Figure 10. Peaty soil encountered at DSM treatment area (left), and rectification measure taken to
replace affected DSM columns with JMC columns (right)
As excavation was undertaken by a third party, close supervision of the works was warranted. Among
the issues encountered, one was relating to rock mining near DSM wall. It was prescribed in design
specifications that rock blasting would not be conducted within 5.0 m of retention wall, and only
mechanical hacking was expected for rock excavation close to the DSM and CBPs. The reality on site
was different. In some cases, blasting was conducted as close as 1.0 m from the wall face due to delay
in construction program. Compared to CBP walls, this caused concern particularly on the DSM wall,
since it is unreinforced and cracking of the block can potentially lead to water seepage. Based on
pertinent vibration codes (BS 7385-2), it was recommended to keep the vibrations peak particle velocity
(PPV) at the DSM to less than 15 mm/s. Save for minor deformations at DSM toe, no significant damage
was noted on the DSM wall. Vibration monitoring at the top of the DSM (approx. 15.0 m away from
blasting point) showed the resultant PPV was 5.2 mm/s. Figure 11 shows the damage to the rock and
DSM toe following the incident.

Figure 11. DSM toe and damage to rock below DSM following a close proximity blasting

224 DFI-PFSF 2017 Conference, Melbourne


On CBP walls, a source of concern was over-excavation of the proposed rock benching. This was as a
result of hasty excavation works and highly jointed and weathered rock profiles. In one case, where the
remaining benching was insignificant, additional level of ground anchors was quickly added to pin the
CBPs and prevent potential pile toe instability (refer Figure 12).

Figure 12. Over-excavation of rock benching (left), and remedial measures taken by adding a third
row of ground anchors (right)

4. COMPARISON OF RETAINING WALL PERFORMANCE: EXPECTED VS. ACTUAL

The basement excavation and response of the retaining walls were monitored throughout the excavation
and basement construction process. Wall lateral deformations along the four walls were monitored using
eight inclinometers. About 19 numbers of ground and wall settlement markers were also installed to
monitor the effect of excavation on surrounding areas, roads and buildings. Four standpipes were also
installed around the walls to monitoring groundwater variation throughout the excavation period. Two
of these had the dual purpose of ground water level monitoring and serving as recharge wells, if
necessary. Layout plan of the different instruments is as shown in Figure 13.

Figure 13. Instrumentation layout plan

Alert and alarm values, together with action plans were set up at the beginning of the excavation based
on local construction practice. The maximum lateral movement of the wall was limited to 0.5% of the

225 DFI-PFSF 2017 Conference, Melbourne


excavation height. During the excavation works, monitoring of the installed instruments (inclinometers,
ground settlement markers and standpipes) was carried out at a frequency of twice per week. This was
gradually reduced to a weekly and fortnightly basis during basement construction and anchor
destressing/removal.

From the monitoring data, it was noted that the DSM wall barely moved, showing a maximum lateral
movement of about 1.0 mm up to the end of excavation period. The predicted maximum deformation
from FEM modelling was about 8.0 mm. It is to be noted that in FEM modelling for simplicity the rock-
soil/DSM interface is modelled as a straight plane (horizontal or otherwise). However, owing to the
karstic profile of the rock, the actual rock–DSM interface is jagged. This allows more contact area and
mechanical roughness, thereby providing higher basal resistance against sliding and resulting in a better
performance of the block retaining wall.

In contrast, the CBP wall performance showed some variation. During each stage of excavation, the
wall deformations were checked. The maximum lateral deformations ranged from a minimum of 3.0
mm to a maximum of about 20.0 mm. Generally, the actual performance of the CBP walls was better
than the predicted deformation, as shown in Figure 14. Particularly at the South wall, the performance
of the CBP wall was much better than the predicted due to better soil conditions than expected (Figure
14, bottom left). The dipping bedrock profile at this wall may have resulted in lower lateral pressures
on the wall. The exerted surcharge load on this wall was also less than the design, since there was about
8.0 m clear distance to the road and the space was mostly vacant.

Figure 14. Comparison of predicted and actual wall deformations; West Wall (top left), North Wall
(top right), South Wall (bottom left), and East Wall (bottom right)
The highest recorded lateral movement was at the North wall (INC-4), where highly-fractured limestone
was encountered and the rock benching was over-excavated (Figure 14, top right). In another case at
the East wall, where the rate of wall deformation was larger than predicted as a result of deeper rock
head (INC-9 in Figure 14, bottom right), a corrective measure in the form of installing additional ground
anchors at that level was carried out.

226 DFI-PFSF 2017 Conference, Melbourne


Plots of normalized ground settlement and distance of monitoring points from the excavation face for
the different walls is shown in Figure 15. As shown in the monitoring results plotted in Figure 15, on
one side of the site there was localized settlement of the ground. This was on a pedestrian walk way on
the south west corner of the site. The reason behind the ground settlement was found out to be mainly
due to loss of support to kerb wall. About 1.0 m high earth was removed behind the CBP walls upon
which the kerb wall was leaning to make way for site offices and storages.

Further, the monitoring data recorded for this project will be used to refine designs for future basement
excavation projects in similar ground conditions.

Zone I: Sand and soft to hard clay (average


workmanship on support construction);

Zone II: Very soft to soft clay – Condition 1


(difficulties in construction);

Zone III: very soft to soft clay – Significant depth


below excavation bottom

Figure 15. Normalized plots of ground settlement vs. excavation depth and distance from excavation
face. Plots show actual, recorded settlements compared to Peck’s (1969) plot for excavations

5. SUMMARY AND CONCLUSIONS

This project shows the effectiveness of mixed solutions for basement retention particularly in karst
ground. The provided solution for excavation retention proved successful. Performance of the retention
system was satisfactory and within the predicted movement envelop. Excavation was carried out
efficiently; on account of no obstructions due to struts, no problems with ground water inflow, and
faster excavation was possible.

Pre-execution site investigation works, such as rock probing, were vital to arrive at an optimized design.
Precise mapping of the rock head profile was used to determine the wall types to be implemented. Using
a mixed solution for the excavation retention enabled tailor-made solutions for each potential risk, while
also allowing cost and time saving to the Client.

In this project, it was evident that the construction aspect of basement excavations is important. In karst
ground, as the variables that affect the design are numerous, it was vital to follow the construction
closely. The construction process of the retention system was examined closely to ensure the design is

227 DFI-PFSF 2017 Conference, Melbourne


being realized and the required quality is being achieved. Monitoring was carried out consistently and
design was revised progressively when the need arose.

REFERENCES

Gobbett, D.J., Hutchison, C.S. and Burton, C.K., 1973. Geology of the Malay Peninsula.

Hoek, E., Carranza-Torres, C. and Corkum, B., 2002. Hoek-Brown failure criterion-2002 edition.
Proceedings of NARMS-Tac, 1, pp.267-273.

Horvath, R.G. and Kenney, T.C., 1979. Shaft resistance of rock-socketed drilled piers. Deep
Foundations, pp.182-214.

Liew S.S., Khoo C.M., Tan S.T., and Loh Y.E., 2010, May. Review of Load Test Performance of Bored
Pile Foundation in Weathered Meta-Sedimentary Formation and Kuala Lumpur Limestone. The 17th
Southeast Asian Geotechnical Conference Taipei, Taiwan.

Marinos, P. and Hoek, E., 2000, November. GSI: a geologically friendly tool for rock mass strength
estimation. In ISRM International Symposium. International Society for Rock Mechanics.

Marinos, P., Marinos, V. and Hoek, E., 2007. Geological Strength Index (GSI). A characterization tool
for assessing engineering properties for rock masses. Underground works under special conditions.
Taylor and Francis, Lisbon, pp.13-21.

Peck, R.B., 1969. Advantages and limitations of the observational method in applied soil mechanics.
Geotechnique, 19(2), pp.171-187.

Tan, B.K. 2006. Urban geology of Kuala Lumpur and Ipoh, Malaysia. Engineering geology of
tomorrow’s cities. Geological Society, London, Eng Geol Spec Publ, 22.

Tan, Y.C. and Chow, C.M., 2006, August. Foundation design and construction practice in limestone
area in Malaysia. In HKIE-IEM Seminar on Geotechnical Works in Marble, Hong Kong.

Topolnicki, M., 2004. In situ soil mixing. Ground Improvement, 2, pp.331-428.

Yee, Y.W. and Chua, C.G., 2008. Case Studies: Deep Soil Mixing in Excavation. In Seminar on
Excavation and Retaining Wall.

228 DFI-PFSF 2017 Conference, Melbourne

View publication stats

You might also like