International Conference And Exhibition on Trenchless Technology and Tunnelling, 7 9 March 2006, Hotel Sheraton Subang, Subang Jaya,

, Malaysia

GEOTECHNICAL ASPECTS OF THE SMART TUNNEL

Siow Meng Tan SSP Geotechnics Sdn Bhd (sspg@sspsb.com.my)

ABSTRACT: The SMART tunnel spans across the eastern side of Kuala Lumpur. It is 9.7km long and is located in Kuala Lumpur Limestone well known for its highly erratic karstic features. Both ends of the tunnel alignment, adding up to more than half of the total tunnel length, are in ex-tin mining lands. This paper presents the various underground features encountered along the tunnel alignment particularly the limestone karstic features and the engineering properties of the limestone. 1. INTRODUCTION The SMART tunnel spans across the eastern side of Kuala Lumpur in a north east-south west direction, starting near the confluence point of Sg. Ampang river and Sg. Klang river in the north and ends at the lake at Desa Water Theme Park. The total tunnel length is 9.7km with a bore diameter of 13.26m. The cover thickness above the tunnel is about 1 to 1.5 tunnel diameter. There are six shafts: One at each end of the tunnel for TBM retrieval; the TBM launch shaft in the mid-alignment is the largest, measuring 140m long, 20m wide and 30m deep; two junction boxes and a stand alone ventilation shaft. The tunnel is located in Kuala Lumpur Limestone which is well known for its highly erratic karstic features. The areas at both ends of the tunnel alignment, adding up to more than half of the total alignment length, have been subjected to tin mining in the past. This paper presents the various subsurface features encountered along the tunnel alignment particularly the limestone karsts and some of the engineering properties of the Kuala Lumpur Limestone, mainly based on the information collected from the site investigation during the design stage. 2. GEOLOGICAL SETTING The alignment of the SMART Tunnel is superimposed on the map extracted from GSM 1995 as shown in Figure 1. The tunnel is located in Kuala Lumpur Limestone. The areas at both ends of the tunnel alignment, , have been subjected to tin mining in the past as shown in Figure1 and Figure 2. Kuala Lumpur Limestone belongs to Upper Silurian marble. It is finely crystalline grey to cream, thickly bedded, variably dolomitic rock. Banded marble, saccharoidal dolomite, and pure calcitic limestone also occur as described by Gobbett & Hutchison 1973. The properties of the limestone are given in Section 5. Kuala Lumpur Limestone is well known
1

for its highly erratic karstic features (Tan 2005, Chng 1984, Chan & Hong 1986, Ting 1986, Yeap 1986) 3. TIN MINING Tin mining activities in Kuala Lumpur started in 1857 when the first mine was operated in Ampang. Tin mining was rampant in the past and concentrated in the limestone area of Kuala Lumpur as shown in Figure 2. Note that most information concerning the tin mining industry of Selangor before the Second World War was lost or destroyed during the war (Yin 1986), and as a result, it is not possible to have a complete Figure 1 Geological map: Ex-mining area is dotted; Solid lines are and accurate record of all the fault lines, dashed lines are inferred fault lines (GSM, 1995) mining areas. Most tin mine tenures expired in the early 1980s. The common mining method was open cast and gravel pump. This method involved excavation by big machines such as bucket wheels and navies. At confined places, such as potholes and pinnacles, the sediments were first broken by water jet and washed down to a pool which was then pumped to flow down along a sluice built on a tall wooden framework called palong (Figure 3), thus concentrating the heavy minerals including the tin ore cassiterite (Ayob 1965). The mining activities left behind numerous ponds and remnants mainly consisting of sand and clay slime, forming a highly heterogeneous sequence of overburden materials over the limestone as illustrated in Figure 4. 4. KARSTS OF KUALA LUMPUR LIMESTONE 4.1 Development of Karsts Karst topography in limestone is formed by a chemical dissolution process when groundwater circulates through the limestone as illustrated in Figure 5. Carbon dioxide from the atmosphere is fixed or converted in the soil in an aqueous state and combined with rainwater to form carbonic acid, which readily dissolves carbonate rocks. Karstic features develop from a self-accelerating process of water flow along well-defined pathways such as bedding planes, joints and faults. As the water percolates downward under the force of gravity, it dissolves and enlarges the pathways. Enlargement of a pathway allows more water flow which increases the dissolution rate. As the enlarged pathway transmits more water, it pirates drainage from the surrounding rock mass. Over time, this process results in very jagged appearance, sometimes dissecting vertically and deeply into the rock terrain.

Water will continue to percolate downward until it reaches the water table, below which all pore space is occupied by water. The water table fluctuates as a result of seasonal change and creates a zone of preferential dissolution along the zone of fluctuation. Over time, this process creates solution channels. The dissolution of limestone is a very slow process compared to the human life span. The dissolution rate is expressed in ka, 1000 years (Kaufmann 2004).

SMART TUNNEL L

4.2 Limestone Profile The limestone profile along the tunnel alignment obtained from some 40 boreholes are depicted in Figure 6(c). The limestone profile varies with depth varying from a few metres to more than 30m at both ends of the tunnel, with Figure 2 Ex-tin mining areas in Kuala Lumpur (Mining Dept., isolated depressions in unpublished) between. The drop in limestone profile at the northern end could be attributed to the fault line (see Figure 1). The bedrock profile is expected to be a lot more erratic if more boreholes had been sunk. Such erratic features are exposed in the excavation of some shafts. Some extreme cases are shown in Figure 6. Steep and deep depressions on a limestone plateau are shown in Figure 6(d). The maximum depth of a depression is 34m as determined by construction of the contiguous bored pile wall. In another shaft, potholes as shown in Figure 3(b1&b2) were encountered. The largest measured 11m in diameter and 8m deep. Bigger potholes have been observed by Ayob (1965) and Yeap (1986). Deep depressions and potholes were suspected during the site investigation stage but it was very difficult if not impossible to delineate the shape of such features with reasonable accuracy without a huge number of boreholes. This was not practical and feasible.

Figure 3 Palong in an opencast tin mine in Segambut (Gobbett, 1973)

Figure 4 Formation of tin mine tailings using gravel pump method (Chan & Hong 1986), Above: Mining remnants are being deposed from the palong after tin ore extraction. The fine sliming and clayey materials settle much slower than the course sandy materials, thus is floating on top and separated from the sand. Below: As a result of the deposit mechanism above, the mining remnants form lenses of material of heterogeneous properties.

Figure 5 Process of limestone dissolution (UCGS 2000)

(a) (b1) (c) (d) (b2)

TUNNEL CROWN

RC WALL BENDED

RC RETAINING WALL LIMESTONE PLATEAU

CONTIGUOUS BORED PILE WALLS & ANCHOR TIE BACK

TUNNEL, 13.2M O.D. DEPRESSION SHOWN IN (a)

Figure 6 Limestone exposures from the shaft excavations for the SMART tunnel: (a) A depression of 20m deep maximum as measured by a bored pile. (b1 & b2) A huge pothole and layout of the potholes at the North Junction Box near the Kg Pandan roundabout. (c) Limestone profile along the SMART Tunnel. (d) A 3-D image of the limestone topography at the North Ventilation Shaft and the TBM launch shaft near Jalan Cheras. 5

Geophysical methods that are common locally include seismic refraction survey, seismic reflection survey, resistivity and ground penetration radar. These methods have achieved limited success in the past in detecting erratic limestone profile and existence of cavities. The applications of these methods would also be hindered by encumbrances at the site and interference of ambient noise particularly traffic noise, underground utilities such as metal pipes, electrical and telecommunication cables. Micro-gravity method was used for a stretch of 2.7km during the design stage and was relatively successful in identifying the locations of large karstic features in the limestone but the results are indicative. Further trials during the construction stage shows that 2D-resistivity tomography was promising and was carried out extensively along the alignment in advance of tunnel boring to forewarn the existences of unfavourable karstic features and allowed time for implementation of mitigation measures. The design of the retaining walls for the shaft excavations had to cater for various bedrock depths. Reinforce concrete and gabion walls were adopted for shallow bedrock of a few metres. Contiguous bored pile (CBP) and secant bored pile walls were used where bedrock is deeper. Diaphragm walls were not considered suitable due to highly erratic bedrock profile. The retaining wall design for the shafts was designed to be fully flexible to cope with the expected highly erratic rockhead. As a first stage, the soil at the retaining wall location was excavated down to bedrock and the excavation inspected by a geologist. Where competent rock at depths less than 6m was encountered a RC cantilevered retaining wall was constructed. In areas of deeper rockhead bored piles were constructed. Once the depth of the bored piles was known, the numbers, spacing and loads for the ground anchors could be determined based on predetermined designs for various heights of wall. In some circumstances alternative designs were adopted by the Contractor for cost, programme and practicality reasons. These alternative designs included the introduction of cornere struts (Figure 6a), as opposed to anchor tie backs, jet grouting at the rear of the CBP walls or realigning the rc wall, to minimise the number of bored piles required. In one case the CBP wall needed to be Figure 7 A cavity underneath a bored pile wall. realigned to prevent intrusion of the reinforced piles into the tunnel eye in poor ground. Where exceptionally deep or erratic rockhead was encountered a double row of bored piles was required in order to provide additional support and to ensure that pile toes were adequately socketed into the limestone In another case, the RC retaining wall was realigned to get around the pothole as shown in Figure 6(b2); Strengthening of pile toes or filling up cavities underneath the wall, such as the one in Figure 7.

4.3 Sinkholes A sinkhole refers to a depression on the ground surface caused by dissolution of the limestone near the surface or the collapse of an underground cave. There were a number of sinkhole incidents in Kuala Lumpur and the surrounding areas in the past as summarised by Tan (2005). Almost all sinkholes are triggered by construction activities. The main triggering factors are lowering of groundwater table thus loss of fines through groundwater seepage.

(a)

(b)

(c)

Figure 8 (a) A solution channel which is originally covered and stable. (b) Water drains into the solution channel when there is a dewatering activity. There will be extra groundwater flow after rains, expediting migration of fines into the channel and causes upward erosion. (c) As the upward erosion eventually reaches the ground surface, the soils collapse, creating a sinkhole. Sinkholes can also be pre-existent and filled up until construction activities come along to trigger off new collapses (after Zhou et al 2002).

An obvious case of ground subsidence in limestone area related to groundwater extraction was reported at an industrial park near Subang Jaya, about 13km from Kuala Lumpur (SSPG 1998). The ground subsided significantly within a period of two months during the illegal pumping of groundwater at an adjacent vacant land. When the pumping was stopped, the rate of subsidence reduced significantly. Other sinkhole triggering factors include imposing of additional loads and vibrations. In a few occasions, it is due to direct punching of cavity cover by borehole or piling activities. The mechanism of sinkhole formation is illustrated in Figure 8. Locations where overburden are thin are more susceptible to occurrences of sinkholes due to lack of buffer and bridging effect. The construction of the 150m long TBM launch shaft at Jalan Cheras road demonstrated the wide spread effects of groundwater table drawdown due to the extensive and interconnected solution system within the limestone. Figure 9 shows the existence of

Figure 9 A solution channel in sound rock masses

Figure 10 A sinkhole being filled up by concrete

solution channels in sound rock masses. Not until extensive grouting work was undertaken to seal the shaft were the incidences of surface subsidence and sinkholes stabilised. The fissured limestone and overburden soils harbour a high groundwater table. Slurry Mixshield TBMs were used to prevent groundwater drawdown to avoid ground subsidence and triggering of surface sinkholes. There were a few sinkholes incidents related to the shaft excavations for the SMART tunnel. The sinkholes occurred at places surrounding the shafts where the overburden soils are a few metres thick. Where overburden thickness was about 10m as observed in one incident, there were ground depressions but open sinkholes as the one shown in Figure 10 did not form. It has been observed that the ground water flow via solution channel was not constant, sometimes it was almost dry but the flow increased during raining period. After a certain time interval, a big flow would occur. The big flow normally was accompanied by sand particles and muddy water. It is believed that as the groundwater was substantially discharged, the flow reduced and the soil in-fills in the solution channel started to build up and blocked the flow further. Groundwater accumulated after the blockage. As the groundwater reached a certain weight, a sudden flush was triggered. This process was repeated until a sinkhole finally appears on the surface unless mitigation measures are carried out on time. The mechanism of sinkholes occurrence along the tunnel alignment during tunneling is different from the above mechanism in Figure 8. They occurred due to soil feature penetrates below the tunnel crown level in general. 5. SOME ENGINEERING PROPERTIES OF THE LIMESTONE 5.1 Mineral Contents The mineral contents of the limestone as determined by X-ray diffraction analysis are summarised in Table 1. The main mineral is calcite. Eight out of the ten specimens analysed consisted of more than 77% calcite, some as high as 100%.
TABLE 1 MINERAL CONTENT BY THIN SECTION AND X-RAY DEFFRACTION ANALYSIS Calcite Other Major Accessory* No of Sample 77%-100% 0-23% 8 1 50% Dolomite 30% 20% 1 34% Fine grained ground mass 67% *consists of microline, iron oxide, grossularite

5.2 Physical Properties Majority of the core samples tested has density of 26.5 kN/m3 to 27 kN/m3 as shown in Figure 11(a). Poissons ratio is determined by attaching strain gauges on the rock core specimens in uniaxial compression tests. The Poissons ratio ranges from 0.16 to 0.35 for majority of the samples as presented in Figure 11(b), with an average value of 0.27. 5.3 Strength Properties The results of the uniaxial compression tests are presented in Figure 11(c). The average uniaxial compression strength, UCS, is 54MPa. This value falls within the range of values obtained from other sites in Kuala Lumpur as tabulated in Table 2.
8

300 99.% 100.% 100.%

120.%

30 26 92.% 22 96.% 100.%

120.%

250 89.% 200

FREQUENCY

240

95.%

100.%

25

25

100.%

80.%
FREQUENCY

20 70.% 15 14 43.% 10

80.%

150

60.%

60.%

100 14.% 18 5728.% 24

40.%

40.%

50

25 6.%

7.% 3 2501-2550

9.% 8 2551-2600

20.% 19 1 1 >2900 .%

5 2 2.% 0 <0.11 0.11-0.15 2 4.%

18.%

20.%

2601-2650

2651-2700

2701-2750

2751-2800

2801-2850

2851-2900

<2501

.% 0.16-0.20 0.21-0.25 0.26-0.30 0.31-0.35 0.36-0.40 >0.40

POISSON'S RATIO, v

DENSITY, (Kg/m3)

(a) Density
30 120.00%

(b) Poissons Ratio

20 18
100.00%

19 18 17
95.% 97.% 99.% 100.%

120.%

25

25

26 95.96% 91.92% 22

100.00%

16
80.00%

100.%

20 69.70%

14
FREQUENCY

14
77.%

80.%

FREQUENCY

12 10 8 6 4 2 1 1.% 4
17.% 10.% 5.%

Frequency 15 Cumulative % 14 43.43% 10 40.00% 60.00%

10
44.%

63.%

60.%

7 5
27.%

40.%

5 2 2.02% 0 2 4.04%

18.18%

20.00%

2 1

20.%

.00%

0 <11 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 101110
YOUNG MODULUS, E(GPa)

.% >110

<0.11

0.11-0.15

0.16-0.20

0.21-0.25

0.26-0.30

0.31-0.35

0.36-0.40

>0.40

POISSON'S RATIO, v

(c) UCS
30 99.% 100.% 120.%
20 18

(d) Youngs Modulus

19 100.% 15 13 61.% 51.% 45.% 8 5 4 3 1 3.% 1.% 2.60-3.00 <2.60 1 6.% 4.% 3.01-3.50 3.51-4.00 9.% 32.% 26.% 22.% 18.% 14.% 6 6 6 6 60.% 8 7 6 5 4 20.% 40.% 93.% 15 89.% 85.% 95.% 80.% 100.% 120.%

25

23

24 85.% 77.% 64.% 16

92.% 94.%

100.%

16 14

20 FREQUENCY

80.% FREQUENCY

12 10 8 6 10

70.%

80.%

15 12 10 6 5 2 2.% 0 <0.51 0.51-1.00 1.01-1.50 1.51-2.00 2.01-2.50 2.51-3.00 3.01-3.50 3.51-4.00 4.01-4.50 4.51-5.00 5.01-5.50 5.51-6.00 6.01-6.50 >6.50 3 4.% 2 6.% 10.% 9 27.% 17.% 2 1 45.% 10 9 7

60.%

40.%

20.%

4 2

.%

9.51-10.00

10.01-10.50

10.51-11.00

11.01-11.50

4.01-4.50

4.51-5.50

5.01-5.50

5.51-6.00

6.01-6.50

6.51-7.00

7.01-7.51

7.51-8.00

8.01-8.50

8.51-9.00

9.01-9.50

POINT LOAD STRENGTH (MPa)

BRAZILIAN TENSILE STRENGTH (MPa)

(e) Point Load Strength

(f) Brazilian Tensile Strength

FIGURE 11 STATISTICS PLOTS OF THE TEST RESULT

>11.50

.%

TABLE 2 LIMESTONE STRENGTHS OF VARIOUS SITES IN KUALA LUMPUR Locations Bore-hole Data no. UCS, MPa No. Range Avg. 47 223 10 115 54 SMART Tunnel 29 46 25 85 48 Sentul, Site 1 Sentul, Site 2 62 113 8 117 46 Ampang 15 19 12 92 40 Komoo (1989) 28 120 Tan & Chng (1987) 108 10 120 -

The Youngs modulus, E calculated from the stress-strain curves of the uniaxial compression tests are presented in Figure 11(d). The average E of the rock core is 61GPa. The results of point load tests and Brazilian tensile tests are presented in Figure 11(e) and Figure 11(f). The above test results are summarised in Table 3. The relationships between Youngs modulus, point load strength, Brazilian tensile strength and UCS are illustrated in Figure 12. Point load strength, Is, has been widely used to estimate UCS of rocks. The correlation is found to be: UCS = 16 Is loaded diametrically on rock cores (1a) UCS = 22.5 Is loaded axially on rock cores (1b) The point load indexes established above are lower than the value of 24 recommended by IAEG (1981). There is hardly any relationship between Brazilian tensile strength and UCS. The correlation between the Youngs modulus and UCS is: E = 1250 UCS (2) The ratio of 1250 is rated as high modulus ratio according to Deere & Millers chart in Figure 13. The strength of the limestone ranges from Medium Strong (MS) to Strong (S).
TABLE 3 SUMMARY OF TEST RESULTS Range Property 26.3-29.5 Density, kN/m3 Poissons Ratio 0.08-0.61 Youngs Modulus, GPa 9-117 Uniaxial Compression Strength, MPa 10-115 Brazilian Tensile Strength, MPa 1.1-13.4 Mean Value 26.7 0.271 61 54 7.8 Sample No. 396 99 100 223 148

5.3 Aggregate Properties The aggregate properties are summarised in Table 4 and plotted against the UCS in Figure 14. No obvious relationship can be observed between UCS and the flakiness index, aggregate crushing value and Los Angelas abrasion value. Nevertheless the reduction in water absorption value is obvious as the limestone is getting stronger.
TABLE 4 SUMMARY OF AGGREGATE PROPERTIES
Test Standard Flakiness Index Aggregate Crushing Value(%) BS812:1988 & ASTM C131-98 Water Absorption(%) Chloride Content(%) Los Angelas Abrasion (%) ASTM-C535 & C131 Min 13 22 0.1 <0.01 25 Max 42 38 1.1 <0.01 57 Average 25 31 0.5 <0.01 41

10

P o in t L o a d Te st, P L S (M P a )
12 0 12 0

Y o u n g M o d u lu s, E (G P a )

B ra zilia n Te n s ile S tre n g th , TS (M P a )

12 0

11 0

11 0

11 0

10 0

10 0

10 0

U n co n fin e d C o m p re ssive Stre n g th , U C S (M Pa )

U C S = 2 2 .5 Is
90 90

E =1250UC S
90

80

80

80

70

U C S = 1 6 Is

70

70

60

60

60

50

50

50

40

40

40

30

30

30

20

20

20

10

L E G E ND P oint Load S t reng th ( Diam et ral) P oint L oad S tr eng th ( A xial)

10

10

0 0 2

0 10 0 20 40 60 80 10 0 12 0

0 0.0 0 4.0 0 8.0 0 1 2.0 0 1 6.0 0

R e la tio n sh ip b e tw e e n U C S w ith P L S (D ia ) a n d P L S (A xia l)

R e la tio n sh ip b e tw e e n U C S w ith E

R e la tio n sh ip b e tw e e n U C S w ith T S

(A)

(B)

(C)

Figure 12 UCS versus (a) point load strength, (b) Youngs modulus and (c) Brazilian Tensile strength

Figure 13 UCS versus Youngs modulus plotted on Deere & Millers Chart, superimposed on data of Hong Kong limestone from GCO 1990.

11

0 0

F la k i n e s s In d e x (% ) 20 40 60 80

100 0

A g g re g a te C r u s h i n g V a l u e (% ) 20 40 60 80 100
0

W a te r A b s o r p tio n ( % ) 0 .5 1 1 .5

2 0

L o s A n g e la s A b ra s io n V a lu e (% ) 0 20 40 60 80 100

10

10

10

10

20

20

20

20

Average Unconfined Compressive Strength, Average UCS (MPa)

30

30

30

30

40

40

40

40

50

50

50

50

60

60

60

60

70

70

70

70

80

80

80

80

90

90

90

90

100

100

100

100

(a) (b) (c) (d) Figure 14 UCS versus (a) flakiness index, (b) aggregate crushing value, (c) water absorption, and (d) Los Angelas abrasion value.

110

110

110

110

Figure 15 Pressumeter modulus versus RQD.

12 10 LUGEON 8 6 4 2 0 0 20 40 RQD, % 60 80 100

Figure 16 Lugeon versus RQD 12

5.4 Rock Mass Properties Pressuremeter Modulus Pressuremeter tests were conducted within the rock mass. Menard type pressuremeter was used and the maximum test pressure was 60bar, apparently not adequate for stronger rock mass. The pressuremeter modulus data are shown in Figure 15. The values vary from 0.5GPa to 3GPa. The average value is 1.5GPa, about 40 times lower than E obtained from intact rock cores. Although there is a trend of higher pressuremeter modulus sound rock mass as reflected in the higher RQD value, the data is scattered widely. Permeability The permeability of the rock masses was determined by means of the water pressure test known as packer or Lugeon test conducted in the boreholes. Permeability was measured by the flow of water pressed into isolated sections of a borehole. The permeability is expressed in Lugeon. According to BS5930, 1 Lugeon unit (LU) is defined as under a head above groundwater level of 100m (10 bar), a 1m length of borehore section accepts 1 litre per minute of water. The test results are presented in Figure 16. Most of the Lugeon values fall below 3 while a few tests give higher values of 8 to 10. No meaningful relationship can be established between Lugeon and RQD. The permeability of the rock masses is considered manageable. The main concern is drainage of groundwater table through limestone solution features. Rock Mass Quality Rating, Q The Q rating system was developed by Barton et al of NGI in 1974 (Bieniawski 1989). It is widely used in tunnel engineering. It is defined as: Q=(RQD/Jn)(Jr/Ja)(Jw/SRF) (3) Where RQD is the rock quality designation (0-100), Jn is the joint set number coefficient (0.5-20), Jr is joint roughness number (0.5-4), Ja is joint alteration number (0.75-20), Jw is joint water reduction number (0.05-1) and SRF is stress reduction factor (0.5-20). Q values were calculated based on rock cores obtained from the boreholes. The last component (Jw/SRF) in Eq. (3) is assumed unity for convenience. The distributions of the Q values are tabulate in Table 5.
TABLE 5 DISTRIBUTIONS OF Q VALUES MEASURED FROM ROCK CORES BY ASSUMING JW/SRF=1 0-5 5-10 10-20 20-30 30-50 50-75 75-100 >100 Q

Worst Case, % Best Estimate,% Average Frequency, %

84.3 5.6 45.0

9.0 5.6 7.3

1.1 31.5 16.3

3.4 5.6 4.5

2.2 21.7 12.0

0 5.3 2.7

0 12.4 6.2

0 12.3 6.1

As the rocks were exposed during construction, most of the exposed was dry. Therefore Jw is 1. It was observed that sheared zone, weak zone with clay band were common and SRF of 2.5 was adopted in most cases. This gives Jw/SRF of 0.4. The Q values expected from rock excavation is summarised in Table 6.

13

TABLE 6 EXPECTED QUALITY OF ROCK IN EXCAVATION BASED ON AVERAGE Q DISTRIBUTION FROM TABLE 5

Rating Q Frequency, %

V. Poor to worst 0<1 22

Poor 1-4 30

Fair

Good

4-10 21

10-40 21

V. Good to Best >40 6

6. CONCLUSIONS The SMART tunnel is a good showcase of the highly erratic karstic features of the Kuala Lumpur Limestone. Such features posed challenges in the excavation of the deep shafts. It is impractical to rely on boreholes to delineate the limestone rock profile and solution features as a great number of boreholes will be needed. Probing of rock profile prior to construction was cost effective but is limited to probing depth of around 10m and cannot detect any feature below the rock head. Through experience learnt from the SMART, it is hoped that the accuracy of the geophysical survey using resistivity survey in detecting limestone features has been improved to be more reliable for future projects. The solution system in the Kuala Lumpur Limestone is well connected and spreads far. Sealing by means of grouting in a strategic manner should be undertaken before a deep excavation is carried out to minimise ground subsidence or sinkholes due to excessive loss of groundwater via the solution system. ACKNOWLEDGEMENT The author wishes to express his appreciation to Mr. C. L. Lee of Sepakat Setia Perunding Sdn Bhd for providing impressive photographs and valuable information on sinkhole and rock mass quality, to Mr David Parks of Mott Macdonald on information related to the construction. Appreciation is also extended to the authors colleagues, Mr C. S. Lim, Mr. Soh L P and Mr T. W. Chang for their kind assistance, Ms Hazel Hooi and Mr. F. K. Sek for proof reading the manuscript. REFERENCES Ayob, M., 1965, Study in bedrock geology and sedimentology of Quaternary sediments at sungai besi tin mines, Selangor, BSc.(Hons.) Thesis, Geology Department, Univ. Malaya Bieniawski, Z. T. 1989, Engineering rock mass classifications, John Wiley & Sons, pp. 73-82. Chan S. F. & Hong, L. E., 1986, Pile foundation in limestone areas of Malaysia, Foundation Problems in Limestone Areas of Peninsular Malaysia, Geot. Tech. Div., IEM China Press, 29-04-2004 Chng S. C., 1984, Geologi Kejuruteraan Batukapur Kuala Lumpur, Malaysia, BSc.(Hons) Thesis, Geology Department, UKM, Bangi, Selangor, Year 1983/84 Geological Survey Malaysia, 1995, Geological map of Kuala Lumpur and surrounding areas, Wilayah Persekutuan Series L8010, Part of Sheet 94a, 94b, 94d, 94e & 94f, Digital process 1995 Gobbett, D.J. & Hutchison, C.S. 1973, Geology of the Malay Peninsula, New York: Wiley-Interscience.

14

GCO 1990, Foundation properties of marble and other rocks in the Yuen Long-Tuen Mun area, Geotechnical Control Office Pub. No. 2/90, Civil Engineering Services Department, Hong Kong Kaufmann G., 2004, Karst system modelling, Course lecture, Inst. of Geophysics, Univ. of Goettingen, Gemany, http://www.uni-geophys.gwdg.de/~gkaufman/work/karst/ index.html, 16-09-2004 Mining Department, 1980-1982, Ex-mining Land Map in Kuala Lumpur and Adjacent Area, unpublished. Sin Chew Daily 3-04-2004 SSP Geotechnics Sdn Bhd (SSPG), 1998, Geotechnical Investigation Report on Cracks and Settlement of Factory Lots at Subang Hi-tech Park (Subang Square), Selangor, Nov. 1998, Job 31259. Tan, S. M., 2005, Karstic Features of Kuala Lumpur, Cover story, Jurutera, IEM Bulletin, Jun 2005, pp. 6 11. Ting, W. H., 1986, Foundation problems in limestone areas, Foundation Problems in Limestone Areas of Peninsular Malaysia, Geot. Tech. Div., IEM. U.S.Geological Survey (USGS), 2002, Coastal and Marine Geology Program web site, Jan 18 2002, http://coastal.er.usgs.gov/publications/ofr/ 00180/intro/karst.html Yeap E.B., 1986, Irregular Topography of The Subsurface Carbonate Bedrock in The Kuala Lumpur Area, Foundation Problems in Limestone Areas of Peninsular Malaysia, Geot. Tech. Div., IEM. Yin E.H., 1986, Geology and Mineral Resources of Kuala Lumpur-Klang Valley (Draft), Geological Survey Malaysia District Memoir.

15