Available online at www.sciencedirect.com

The Twel

Tin Sh

Abstract

This paper constructionHong Kongfoundation da major chakarst area. Fthe collapse and how thestages, especcavities withdiameter boparameters hallowable shproject.

© 2011 Pub

Keywords: Roc

(Northwest New

a Correspondinb Presenter: leu

A

lfth East As

Rock-SGeophy

hui Wai

C. T. W

Arch

will present n of a public lg. An appraisdesign and to allenge to engFormation of a

of cavities. Te sinkhole hazcially the exphin the karst aored piles havhave been pubhaft friction b

lished by Else

ck-Socketted Larg

w Territories)

g author: wongctungmk@archsd.go

Procedia Engi

Available onlin

sia-Pacific C

Socketedysical SuPublic L

WONGa, K

hitectural Serv

the geologiclibrary cum ansal was carrieassess the sui

gineers when a sinkhole usu

This paper wilzard was inveerience in usiarea. Moreoveve widely beeblished for thebetween marb

evier Ltd.

ge Diameter Bore

t@archsd.gov.hkov.hk

ineering 00 (201

ne at www.sci

Conference

d Large Durvey inLibrary C

. P. YIM, M

vices Departm

cal and enginn indoor recreed out to dettability of difffounding a st

ually leads to l describe how

estigated and/oing cross-holeer, in Hong Ken employed e piles installeble and concr

ed Piles, Caverno

1) 000–000

encedirect.co

on Structur

Diametern CavernCum Ind

M. K. LEU

ment, HKSAR

neering consieation centre itermine the e

fferent types otructure and/oa sudden dep

w the most apor dealt with e seismic tomoKong, althoug

as foundatioed in cavernourete pile at th

ous Karst, Cross-H

www

m

ral Engineer

r Bored nous Kardoor Rec

UNGb, S. C

Government,

iderations in in cavernous keffect of geof foundation sor carrying ouression on the

ppropriate founat the planninography to deh rock-socket

on, limited lous karst area.he socket sec

Hole Seismic Tom

ProceEngine

w.elsevier.com/lo

ring and Co

Pile andst Area: creation

C. FUNG

China

the planningkarst area at Tlogical constr

systems. Sinkhut excavation e ground surfandation system

ng, design andlineate the spatted steel H-pical experiencThis paper wition being ad

mography, Sched

ediaeering

ocate/procedia

onstruction

d

Centre

g, design andTin Shui Wai,traints on thehole hazard isin cavernous

ace caused bym was chosend constructionatial extent ofiles and large

ce and designill discuss the

dopted in this

duled Area No. 2

d,essynnfenes

1877–7058 © 2011 Published by Elsevier Ltd.doi:10.1016/j.proeng.2011.07.219

Procedia Engineering 14 (2011) 1744–1751

C.T. WONG et al. / Procedia Engineering 14 (2011) 1744–1751 17452

1. Propos

The site (the junctionstructure wibasement tothe geologic

Figure 1(a) Geo

Figure 1(b) Se

sed developm

(trapezoidal shn of Tin Fuk ith steel truss

o house the plcal map and ge

ological Map (So

ection across the S

Aut

ment

hape on plan wRoad and Pin

ses for the inant rooms andeological sect

ource: Adapted fr

Site

hor name / Proce

with an area ong Ha Road.

ndoor recreatiod to serve as ion of the site

om GEO 2007: 1

edia Engineering

of 5,450m2) isThe propose

on centre andthe car-park.

e.

105)

g 00 (2011) 000–0

s located in Tid developmend swimming Figures 1(a)

000

in Shui Wai, Hnt is a reinforpool. There iand (b) show

Hong Kong atrced concreteis a 7m deep

w respectively

tepy

1746 C.T. WONG et al. / Procedia Engineering 14 (2011) 1744–1751 Author name / Procedia Engineering 00 (2011) 000–000 3

2. Geology of the site

According to Environment, Transport and Works Bureau Technical Circular (Works) No. 4/2004, the site is located within the Scheduled Area No. 2 (Northwest New Territories), which is an elongate curved valley (Figure 1(a)) with granitic and volcanic rocks forming the high relief topography on either side and weathered meta-sedimentary rocks forming low hills (Frost 1989). The valley is covered with superficial deposits. A two-staged approach was adopted to determine the geology of the site, with the first stage to obtain general information of the geology and the subsequent stage to obtain more detailed conditions of the karst features as revealed in the first stage (Lee and Ng 2004). A geological assessment was also conducted to develop the geological and hydrogeological models of the site. A total of 50 drillholes had been sunk, which revealed that the site is underlain by fill materials of 5m to 7m thick overlying alluvial soils of about 10m thick followed by completely weathered meta-siltstone and then the bedrock. The bedrock is of 3 types: pure marble, meta-siltstone and granite. Three lines of fault zones run across the site. Although the geological map (Figure 1(a)) shows that a large portion of Tin Shui Wai area is underlain by volcaniclastic breccias, the site in this particular project is distinct, because the underlying rock is pure white marble belonging to Ma Tin Member of the Yuen Long Formation. The pure marble of Ma Tin Member comprises virtually pure calcium carbonate, which renders it susceptible to dissolution (Frost 1989). Cavities of maximum “height” of 24.5m have been recorded (Lai 2004). For this site, extensive cavities with maximum height of 11m (infilled with clayey silt) are found in the pure marble at north-eastern corner of the site at about 50m below ground. Ground water levels vary from 2.4m to 5.6m below ground.

Chan (1989a, 1989b) classifies sites in karst area into “easy”, “fair” and “very difficult” using his developed Marble Quality Designation (MQD) values. According to his classification, the site can be classified overall as an “easy site”. However, if the site is divided into zones, different classification may be required. There are extensive large cavities as well as sheared zones at the north-eastern side. The western portion and north-eastern portion of the site should therefore be classified as “fair site” and “very difficult site” respectively, instead of an “easy site” for the remaining portion. Therefore, a borehole rating using MQD values in classifying site with cavernous karst should be used with caution and supplemented by professional judgment and adequate geological modelling.

3. Cross-hole tomographic survey

One of the concerned geological features of the site is the karstic formation with extensive cavities. To locate the spatial extent of cavities in a karst site, conventional method by sinking drillholes would be very time consuming and expensive (Morris et al 2005). In particular, sinking drillholes can only provide information on the “vertical height” of the cavities but not the “spatial extent” of the cavities. Geophysical methods are possible technique for the three-dimensional exploration of subsurface conditions in detecting the presence and spatial extent of cavernous karstic formation. The number of drillholes can be reduced by 50% with appropriate use of geophysical methods (Hoover 2003). Cross-hole seismic and ground penetration radar are among the various geophysical methods ranked by BS5930:1999 as “excellent potential but not fully developed” in cavity detection (there being no geophysical method was ranked as “excellent with the technique well developed” by BS5930:1999). In Hong Kong, ground penetration radar method in the form of cross-hole radar was used in several cases in the karstic formation (Lee et al 2000). Therefore, it was determined to employ cross-hole seismic method on this site following the pre-drilling works. 13 cased drillholes with depths between approximately 25m and 43m and eight sections were selected to conduct the tomographic survey. A vertical spacing of the sparker source (which serves as the emitters) at 1m c/c was chosen, and a 24-channel hydrophone spaced at 1m c/c was used at

C.T. WONG et al. / Procedia Engineering 14 (2011) 1744–1751 17474

the receiverHowever, thdistances an1.5m) or thinwhen emplohave to be tomograms s

4. Choice

In construdriven H-pialong slopinthe spatial ebuilding waof the site aconstructing

Figure 2 Tomo

Commonfoundation; diameter bopad or raft f

r side. Abouthe method ofnd the geometrn rock coverin

oying seismicverified and superimposed

e of most app

ucting an annles were dam

ng bedrock (Szextent and dists re-planned sas far as pos

g the foundatio

ographic Image o

nly adopted ffloating pile

ored piles or rfooting was ru

Aut

t 16,000 travf tomographic rical density ong may not ha tomographicsupplemented

d with the resu

propriate foun

nex to TWG Hmaged from hze 2006). Wittribution of thsuch that the psible, and henon works.

f the Site

foundation syd-foundation

rock-socketteduled out, as fo

hor name / Proce

vel times wersurvey is lim

of ray paths. Cave clear iden method is th

d by the resulults of drillhol

ndation solut

Hospitals Kwoard-driving thh a clear geol

he cavities undproposed strucnce avoiding

stems in cav(e.g. Pakt-in

d steel H-pilesounding the bu

edia Engineering

re determinedmited in spatiaCavities with ntification. Thuhat the measults of drillholles obtained in

tion

ok Yat Wai Chrough the kalogical and hyderneath the pcture was shifpotential dif

vernous karst n-Place (PIP) s. For this prouilding on eith

g 00 (2011) 000–0

d and used fal resolution smaller dimenus, one salien

urements of sees. Figure 2

n this site.

College adjacearstic layer ofydrogeologicalproposed devefted away fromfficulties and

areas in Honpiles); driven

oject, shallowher the fill or

000

for tomographdepending onnsions (e.g. lent point to be beismic tomogrshows one o

ent to this sitef marble and l model showielopment, the m the north-eareducing cost

ng Kong incln steel H-pile

w foundation inalluvium laye

hic inversion.n the drillholeess than 1m orborne in mindraphic survey

of the seismic

e, many of thepile slippage

ing especiallye layout of theastern portiont and time in

lude: shallowes; and largen the form ofers will result

.erdyc

eeyenn

weft

1748 C.T. WONG et al. / Procedia Engineering 14 (2011) 1744–1751 Author name / Procedia Engineering 00 (2011) 000–000 5

in excessive settlement in the building. For a floating piled-foundation using PIP piles, GEO Technical Guidance Note No. 26 limits the increase of vertical effective stress (at most 10% due to overburden soil) at the karst surface, so as to prevent the collapse of any cavities in marble. This limit is rather conservative. Waltham (2005) and Sowers (1996) suggested that if the rock thickness is greater than the width of cavity, it is likely to be safe. Indeed, GEO Technical Guidance Note No. 26, besides specifying the 10% rule, also permits the use a rational design approach to determine the increase of the vertical stress in the cavity. In this project, the founding level of the pile tips is just at a few metres above the karst surface at some locations, and the thin layer of intact rock above the cavities may cause the roof over cavities to collapse. Wong (2003) summarized the criteria for the successful employment of floating piled-foundation in cavernous karst areas as: the karst surface lying at deep level so that the stress induced at the pile tips will not overstress the roof of cavities, and the building being of low to moderately loaded and thus can be supported by the relatively low pile capacity of PIP piles.

For driven steel H-pile with or without pre-boring and allowance on pile redundancy for uncertainties, appropriate pile hammer should be used during the driving operation, which should be light enough not to damage the pile during hard driving; but heavy enough to achieve the required set (Sze 2006; Wightman and Lai 2006; Sowers 1996). A dynamic pile-driving analyzer should be used to monitor the stress during the whole driving process, and the piling specification of the Architectural Services Department of the Hong Kong SAR Government limits the maximum driving stress to 80% of the yield strength of the steel H-pile. However, in this project, driving steel H piles without pre-boring was not technically feasible as founding H-piles on top of karst surface might again possibility cause the collapse of the roof over cavities within the karst. Driving steel H-piles with pre-boring was also undesirable because of the environmental sensitivity of the site, with three secondary schools, a West Rail Station and a Light Rail adjacent to the site. Large diameter bored pile socketted into rock was therefore adopted, especially due to relatively shallow rockhead (about 20m to 25m below ground) at over three-quarters of the site. The base of the socket was located at a depth where there are no cavities existing within the zone below the pile base to a depth equal to the diameter of the pile base. Karst area also contains steeply inclined bedrock surfaces inducing additional stress from one pile onto the adjacent piles and/or the adjacent cavities. In designing the length of rock socket of pile, the founding level of rock socket was required to pass through the adverse joints, and to follow the commonly adopted rule of thumb in Hong Kong with an “angle of stress dispersion” between 30o and 45o to the vertical (Wai 1991). Similar to large diameter bored pile socketted into rock, rock-socketted steel H-piles could also be a solution. However, for cavernous karst area, such foundation system may have difficulties in pre-boring and forming rock sockets, as it is difficult for the casing shoe to bring the casing through the cavities and to form rock socket on the sound bedrock, except that a secondary permanent casing is left for pile installation; but this would lead to a high construction cost of foundation. Kwong et al (1997) stated that unknown and uncontrolled quantities of soil might be removed when air-flushing is not well controlled during pile installation which may lead to excessive ground movements and collapse of surrounding soil and ground. Indeed, there is no published case on the successful employment of rock-socketted steel H-piles in cavernous karst area in Hong Kong. Hence, a compromise scheme was to adopt rock-socketted steel H-piles only at areas with cavity-free marble and/or granitic bearing rock underneath, whereas large diameter bored piles are used at areas with cavernous marble bedrock.

The final adopted foundation design involved a total of 50 nos. of large diameter bored piles with either 1.5m or 2m diameter, and 99 nos. of rock-socketted steel H-piles. The piling layout plan is shown as Figure 3. With the adoption of two different piling systems, the design of the pile caps and superstructure was also required to cater for the possibility of differential settlement due to the difference in the elastic shortening of reinforced concrete bored piles and structural steel rock-socketted steel H-piles.

C.T. WONG et al. / Procedia Engineering 14 (2011) 1744–1751 17496

Two parathe allowaballowable en(Meigh 199Department but no corresitu measurevaries over codes/designmini-piles, ain this projethe allowablhave been inDepartment level surveyindustry is n

Figure 3 Piling

5. Constr

The presamount of methods to with lean cowas also pconstruction

ameters are reble end bearinnd bearing pr1). For the shaonly gives th

esponding figuements (e.g. Sa large range

n guidelines cand noted thatect, a conservale bearing prenstrumented wof Civil Engi

y, strain measuneeded.

Layout Plan of t

ruction work

ence of karstoverbreak leaform rock-so

oncrete, or inspaid to avoidn due to in t

Aut

equired to calng pressure, aressure for Caft friction, th

he values for ture for marblSerrano and Oe, and there icontaining the t the average aative value ofessure and shawith vibrating ineering of theurement and e

the Site

ks on site

t cavities mayading to conc

ocketted large stalling a permd sinkhole suthe increase o

hor name / Proce

culate the loaand the shaftlass I or II m

he Code of Prahe allowable e. Overseas, tOlalla 2006) ss yet any conallowable sha

allowable shaff 0.3MPa was aft friction, fowires strain g

e University oxtensometer h

y impose concrete loss frodiameter bor

manent casingubsidence byof the effecti

edia Engineering

ading capacityt friction betw

marble in Honactice for Foushaft friction theoretical stushow that thensensus. Kutsaft friction beft friction rangadopted. To v

our nos. of rogauges and exof Hong Konghas still been i

nstruction diffom pile and red piles throug in the cavitiey controlling ive stress ove

g 00 (2011) 000–0

y of rock-sockween rock anng Kong is tyundations (200for socket in

udies (e.g. Tane allowable shchke et al (20tween grout ages from 0.42validate the adock-socketted tensometers w. The instrumein progress on

ficulties, suchhence affectinugh cavities ies before conc

the dewaterer the cavitie

000

ketted bored pnd the concreypically taken04) issued by granitic and vn and Chow 2haft friction in004) quoted t

and various typ5MPa to 0.62dopted paramlarge diamete

with the assistaentation in tern site. More re

as a potentiang the pile iinclude: fillingcreting. Particring during tes. Sheet pile

piles, namely:ete/grout. Then to be 5MPathe Buildings

volcanic rock;2006) and in-n rock socketthree nationalpes of rock in

20MPa. Hencemeters for both

er bored pilesance from therms of preciseesearch by the

al for a largeintegrity. Theg the cavities

cular attentionthe basementes around the

:eas;-tlne,hseee

eesnte

1750 C.T. WONG et al. / Procedia Engineering 14 (2011) 1744–1751 Author name / Procedia Engineering 00 (2011) 000–000 7

perimeter of the basement were driven to a maximum depth of 13m below the existing ground level providing an effective cut-off. The excavation was divided into two zones, and each zone was carried out in two stages, so that when dewatering commenced on site, the increase in the effective stress would have partially been compensated by the removal of overburden soil. Settlement markers, tilting check points, and standpipe piezometers were also installed to monitor the effects of the works on adjacent structures and grounds. The piling works was completed in March 2009, and the superstructure works commenced in April 2009. No sinkhole subsidence was noted throughout the construction period.

6. Conclusion

(a) Ground investigation is crucial for the success of designing building layout, and design and construction of foundation. Wherever necessary and affordable, geophysical survey method using cross-hole seismic tomography can provide more information on the sub-surface ground condition and the spatial extent of cavities and hence help to remove uncertainties or risk in foundation design and construction.

(b) With more comprehensive information on the complicated geology underneath the site, the building layout can be relocated to minimize costs and construction time. Similarly, alternative foundation systems are available. Floating piles or shallow foundation in general are most economical if the loadings are not high. If, however, cavities are present at high level near ground surface, rock-socketted large diameter bored piles passing through cavities are recommended.

(c) Roofs over cavities within the karst may collapse due to the increase in the stress and construction activities, and it is vital for engineer to study carefully the increase in the stress induced by the foundation and the dewatering process. GEO Technical Guidance Note No. 26 has limited the additional vertical pressure on karst surface and this should be reviewed in due course.

Acknowledgments

The authors would like to record their thanks to the Director of Architectural Services for her kind permission of publishing the paper, and to the staff in the Architectural Services Department, Hong Kong SAR Government for their help in preparing the manuscript.

References

[1] BSI, BS5930:1999 - Code of Practice for Site Investigations (London: BSI); 1999.

[2] Chan, Y C, GEO Report No. 29: Classification and Zoning of Marble Sites (Hong Kong: Geotechnical Engineering Office);

1989a.

[3] Chan, Y C, GEO Report No. 32: Classification and Zoning of Marble Sites (Hong Kong: Geotechnical Engineering Office);

1989b.

[4] Frost, D V, “Palaeokarst of Yuen Long, North West New Territories, Hong Kong”, Proceedings of the Third

Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, St Petersberg, Florida,

2-4 October 1989; 1989, pp 239-46.

[5] GEO, GEO Technical Guidance Note No. 26: Supplementary Guidelines for Foundation Design in Areas Underlain by

Marble and Marble-bearing Rocks (Hong Kong: Geotechnical Engineering Office); 2005.

[6] GEO, GEO Publication No. 1/2007: Engineering Geological Practice in Hong Kong (Hong Kong: Geotechnical

Engineering Office); 2007.

[7] Hoover, R A, “Geophysical Choices for Karst and Mine Investigations”, Presented at the 3rd International Conference on

Applied Geophysics, Hotel Royal Plaza, Orlando, Florida, 8-12 December 2003; 2003, pp. 529-38.

C.T. WONG et al. / Procedia Engineering 14 (2011) 1744–1751 17518 Author name / Procedia Engineering 00 (2011) 000–000

[8] Kutschke, W G, Tarquinio, S and Kartofilis, D, “An Alternative Foundation Solution State Route 22, Section A02-

Lewistown Bypass”, Proceedings of Geo-Trans 2004, 27-31 July 2004, Los Angeles, California, ; 2004, pp. 1327-36.

[9] Kwong, S M, Lee, M K and Tse, S H, “Foundation Design and Construction Aspects in Marble (Ma On Shan) – an

Overview of Geotechnical Control”, Proceedings of the Nineteenth Annual Seminar, Geotechnical Division, The Hong Kong

Institution of Engineers, Hong Kong; 1997, pp. 231-38.

[10] Lai, K W, “The Influence Factors of Buried Karst upon the Foundation Design of Hong Kong”, Proceedings of the

Conference on Foundation Practice in Hong Kong, Hong Kong, 4 September 2004; 2004, pp. I1-10.

[11] Lee, D M and Ng, M, “Design Strategies for Deep Foundation in Areas of Marble Formation in Hong Kong”, Proceedings

of the Conference on Foundation Practice in Hong Kong, Hong Kong, 4 September 2004; 2004, pp. C1-17.

[12] Lee, D M, Pun, W K, So, K O and Wai, C C, “Foundation Design and Construction in Hong Kong – Present and Beyond?”,

Proceedings of the Seminar on the State-of-the-Practice of Geotechnical Engineering in Taiwan and Hong Kong, Hong

Kong, 20 January 2006; 2006, pp. 153-80.

[13] Meigh, A C, Technical Note TN 3/91: General Report on Foundation in Area Underlain by Marble and Associated Rock

Volume 1 (Hong Kong: Geotechnical Engineering Office); 1991.

[14] Morris, V, Gehrig, D and Bryant, T, “Detection of Three-Dimensional Voids in Karstic Ground”, Proceedings of the Tenth

Multidisciplinary Conference, 24-28 September 2005; 2005, San Antonio, Texas, pp. 562-71.

[15] Serrano, A and Olalla, C, “Shaft Resistance of Piles in Rock: Comparison between In-Situ Test Data and Theory Using the

Hook and Brown Failure Criterion”, International Journal of Rock Mechanics and Mining Sciences, 43(5); 2006, pp. 826-30.

[16] Sowers, G F, Building on Sinkholes: Design and Construction of Foundations in Karst Terrain (New York: ASCE Press);

1996

[17] Sze, W C. “Case Study for a Driven Steel H-Pile Foundation in the Tin Shui Wai Marble Area”, Proceedings of the Seminar

on Geotechnical Works in Karst in South-East Asia, Hong Kong, 26 August 2006; 2006, pp. 155-78.

[18] Tan, Y C and Chow, C M, “Foundation Design and Construction Practice in Limestone Areas in Malaysia”, Proceedings of

Seminar on Geotechnical Works in Karst in South-East Asia, Hong Kong, 26 August 2006; 2006, pp. 21-43.

[19] Wai, M, Technical Note TN 2/91: Checking Procedures and General Requirements for Foundations in Areas Underlain by

Marble (Hong Kong: Geotechnical Control Office); 1991.

[20] Waltham, T, Bell, F and Culshaw, M, Sinkholes and Subsidence: Karst and Cavernous Rocks in Engineering and

Construction (Berlin: Springer/Praxis); 2005.

[21] Wightman, N R and Lai, A, “Investigation and Foundation Design in Marble/Karst Designated Areas of Tung Chung and

Ma On Shan”, Proceedings of Seminar on Geotechnical Works in Karst in South-East Asia, Hong Kong, 26 August 2006;

2006, pp. 109-40.

[22] Wong, H Y, “Design and Construction of Soil Friction Bored Piles in Hong Kong, with Particular Reference to Marble

Areas”, Proceedings of the 23rd Annual Seminar, Geotechnical Division, The Hong Kong Institution of Engineers, 9 May

2003; 2003, pp. 265-82.