ARUP Structural Engineering Design

The Arup Journal 2/2005 3

‘Who says that structure should not be re i n v e n t e d ?. . . Who says that reinventing structure cannot bec re a t i v e ? ’

I n t ro d u c t i o n

G rowth in China is happening at an historically unparalleled rate. China CentralTelevision (CCTV), the principal state-run bro a d c a s t e r, currently has 13 channels,but by 2008 it plans to be operating over 200 channels and competing successfullywith CNN, NBC, Sky, and the BBC in the global market. To enable this expansion,and to place CCTV firmly on the global map, a new headquarters facility wasneeded, with the entire television-making process housed in one location withinB e i j i n g ’s newly-designated Central Business District (Fig 1).

In early 2002, CCTV organized an international design competition, whichattracted some of the biggest names in arc h i t e c t u re. After much effort it was won inAugust 2002 by Rem Koolhaas’s practice Office for Metropolitan Arc h i t e c t u re(OMA), based in Rotterdam, working with Arup. To secure the project, OMA formedan alliance with the East China Arc h i t e c t u re and Design Institute (ECADI), whichwould act as the local design institute (LDI) of re c o rd for both arc h i t e c t u re andengineering. Working with an LDI is a statutory stipulation for all projects in China,and the LDI’s local knowledge and contacts can make the re l a t i o n s h i pvery beneficial.

CCTV Headquarters, Beijing, China:S t ructural engineering design and appro v a l s

Chris Carro l l Paul Cro s s Xiaonian Duan Craig GibbonsGoman Ho Michael Kwok Richard LawsonAlexis Lee Andrew Luong Rory McGowan Chas Pope

A rchitectural concept

The client stipulated in the competition brief that thefacility should all be housed on one site, but notnecessarily constrained to one building. In hisa rchitectural response, however, OMA decided thatby doing just this, it should be possible to bre a kdown the ‘ghettoes’ that tend to form in a complexand compartmentalized process like the making ofTV programmes and create a building whose layoutin three dimensions would force all those involved –the creative people, the producers, the technicians,the administrators – to mix and produce a bettere n d - p roduct more economically and eff i c i e n t l y.

The winning design thus combines administrationand offices, news and broadcasting, pro g r a m m ep roduction and services – the entire process – in asingle loop of interconnected activities. Thespecifics of the structure evolved in tandem with thespecifics of the building as they in turn evolved, anotable example being the placement of double-height studios within the Towers and Base, whichsignificantly influenced the structural form.

The public facilities included in the project arelocated in a second building, the Television CulturalC e n t re (TVCC); both buildings are serviced from asingle support building which houses major plant aswell as the site security.

When the SD (scheme design) started, thep roject was divided between Arup offices inHong Kong and London. The core team,including six staff seconded from Hong Kongand one from Beijing, was located in London towork closely with OMA in Rotterdam. Given themany engineering disciplines involved, and theneed for dedicated project teams for each of thet h ree buildings on the site, Arup had a near-permanent presence in OMA’s offices. Four ofE C A D I ’s engineers joined the Arup team inLondon for most of the EPD (extendedp reliminary design) phase, while theira rchitectural colleagues worked alongside OMAin Rotterd a m .

Another Arup team in Hong Kong pro v i d e dinformation and guidance on Chinese design andp ro c e d u res, maintained contact with local

authorities and the client, and off e red specialistinput such as wind and fire engineering. As thedesign pro g ressed, additional input was re c e i v e df rom Arup offices in Beijing and Shenzhen.

This close co-operation proved invaluable indelivering a scheme design within four monthsand the EPD and EPR (expert panel re v i e w )a p p rovals within a further six.

Arup provided engineering and consultancy inputfor structural, building services, geotechnical,f i re, communications, and security design,leading the engineering design through SD, EPDincluding the associated approvals processes tot e n d e r, working with ECADI engineers. ECADIc u r rently leads the production of the finalconstruction information and is to provide siteassistance with support from Arup.

The design team

Rem Koolhaas, from a discussion at Tsinghua University, 5 August 2003.

1 . CCTV in the new Beijing CBD.

9247_Arup.qxd 16/9/05 4:53 pm Page 3

The Arup Journal 2/20054

P ro g ress to construction

Given the challenging and unprecedented nature of the winning design, thecompetition was followed by a further period of justification and persuasion, duringwhich Arup applied considerable effort at substantial risk. During the next fourmonths, feasibility studies were made and two key proof-of-concept meetings heldbetween the client’s technical advisors and key members of the Arup team. Theseprimarily addressed the safety, buildability, and overall cost of the scheme, andconcluded that though there was no precedent for such a building, it could beachieved. Once the client was convinced of this, contracts were signed and theexperienced international design team re q u i red to deliver the project was mobilized.The official gro u n d - b reaking ceremony took place on 22 September 2004, butalthough construction started relatively re c e n t l y, it is felt that rather than wait morethan three years until completion, the fantastic story of the realization of CCTVshould start to be told now. This first article is principally a description of thestructural design, analysis, and approvals process for the main CCTV building.Subsequent editions of The Arup Journal will contain episodes on the buildingservices and security engineering for the main building, on the TVCC and supportbuildings, and on the construction process to completion and opening.

The new CCTV headquarters development

The entire CCTV development has a site area of 187 000m2 and will provide a totalof 550 000m2 g ross floor area. The estimated construction cost is around 5bn RMB(or $US600M), and the project as a whole (Fig 2) includes:• the China Central Television headquarters building (CCTV building)• the Television Cultural Centre (TVCC)• a service and security building• a landscaped media park with external feature s .

The 450 000m2, 234m tall (Fig 3), CCTV building consists of a nine-storey ‘Base’and thre e - s t o rey basement, two leaning Towers that slope at 6° in each dire c t i o n ,and a nine to 13-storey ‘Overhang’, suspended 36 storeys in the air, all combiningto form a ‘continuous tube’. Viewed in other terms,the total building form can beseen as four distinct volumes, each approximately the size of One Canada Squarein London’s Canary Wharf, two of them leaning towards each other from oppositec o rners of the site, and joined at the top and bottom by the other two, bothhorizontal and with opposite 90˚ angles in their middles.

2. The site layout, showing programme distribution. 3. The functions and layout within the CCTV building.

Public space and circ u l a t i o n

Studio and bro a d c a s t

S t a ff and VIP facilities

Sky lobby

To w e rl o b b y

To w e rl o b b y

E x p re s se l e v a t o r s

E x e c u t i v ee l e v a t o r s

VIP elevators

Sky studio

P u b l i cl o b b y

C a n t e e n

P re s i d e n ts u i t e

S t u d i o s

S t u d i o s

S t u d i o s

O p e ns t u d i o s

E x e c u t i v ef l o o r

B ro a d c a s t i n g

V I Pl o u n g e

V I Pl o u n g e

Cafe andm e e t i n g

VIP lounge

V I Pl o b b y

V I Pl o b b y

Central kitchen

S t a ff canteen

H e a l t hc e n t re

S p o rts hall

M a r k e t i n gd e p a rt m e n t

G y m

C a n t e e n

9247_Arup.qxd 16/9/05 4:53 pm Page 4


Structural form

S u p e r s t ru c t u re – the ‘continuous tube’

Early on, the team determined that the only way todeliver the desired architectural form of the CCTVbuilding was to engage the entire façade structure ,c reating in essence an external continuous tubesystem. Adopting this approach gave pro p o r t i o n sthat could resist the huge forces generated by thecranked and leaning form, as well as extre m eseismic and wind events.

This ‘tube’ is formed by fully bracing all sides ofthe façade (Fig 4). The planes of bracing arecontinuous through the building volume in order tore i n f o rce and stiffen the corners. The continuoustube system is ideally suited to deal with the natureand intensity of permanent and temporary loadingon the building, and is a versatile, efficient structurewhich can bridge in bending and torsion betweenthe Towers, provide enough strength and stiffness inthe Towers to deliver loads to the Base, and stiff e nup the Base to re i n f o rce the lower Tower levels anddeliver loads to the foundations in the mostfavourable possible distribution, given the geometry.

Vertical cores housing lifts, stairs, and risers areoriented and stepped so that they always sit withinthe footprint of the sloping Towers. Sloping core s ,to allow consistency of floor plate layout, werec o n s i d e red but ruled out due to constraints on thep ro c u rement of the lift systems.

In addition to the cores, the floor plates of theTowers take support from many vertical columns.Given the nature of the sloping Towers it is notpossible to continue vertical column lines from topto bottom, so a two-storey deep system of transfertrusses is used at approximately mid-height. Thefloor plates of the Overhang are also supportedby vertical columns that are transferred to thee x t e rnal tube structure via a two-storey deeptransfer deck (Fig 5).

The continuous tube structure has behind itsfinal irregular arrangement a regular base pattern ofperimeter steel or steel-re i n f o rced concrete (SRC)columns, perimeter beams, and diagonal steelbraces set out on a typically two-storey module.The regular base pattern was tuned or optimized byadding or removing diagonals and changing braceplate thickness to match the strength and stiff n e s sre q u i rements of the design.

The two-storey base pattern was chosen tocoincide with the location of several two-storey highstudios within the Towers. A stiff floor platediaphragm can only be relied on every two store y s ,hence lateral loads from intermediate levels aret r a n s f e r red back to the principal diaphragm levelsvia the internal core and the columns.

The braced tube structure gives the leaning Towers ample stiffness duringconstruction, allowing them to be built safely within tight tolerances before they areconnected and propped off each other. The tube system also suits the constructionof the Overhang, as its two halves will cantilever temporarily from the To w e r s .

R o b u s t n e s s

The continuous tube has a high degree of inherent robustness and re d u n d a n c y, ando ffers the potential for adopting alternative load paths in the unlikely event that keyelements are removed. This was studied in detail and provides the building with afurther level of safety.

S u b s t ru c t u re and foundations

The main Towers stand on piled raft foundations. The piles are typically 1.2m ind i a m e t e r, and about 52m long. Given the magnitude and distribution of the forc e sto be transferred to the ground, the raft is up to 7.5m thick in places and extendsbeyond the footprint of the Towers to act as a toe, distributing forces morefavourably into the ground. The foundation system is arranged so that the centre ofthe raft is close to the centre of load at the bottom of each To w e r, and nopermanent tension is allowed in the piles. Limited tensions in some piles are onlypermitted in major seismic events.

For the Base plus thre e - s t o rey basement, a traditional raft foundation is used,with tension piles between column locations to resist uplift from water pre s s u reacting on the deep basement. 15-20m long, 600mm diameter tension piles will bearranged under the raft with additional 1.2m diameter piles under secondary core sand columns supporting large transfer trusses from the studio areas (Fig 6).

5 .(a) Internal columns start i n gf rom pilecap level.

(b) Internal columns support e don transfer stru c t u re s .

6 .The foundation system.

4 .Principles of the tube stru c t u re: regular grid of columns and edge beams+ patterned diagonal bracing = braced tube system.

9247_Arup.qxd 16/9/05 4:53 pm Page 5


The diagonal braces within thecontinuous tube structurevisually express the pattern off o rces within the structure anda re an important aestheticaspect of the cladding system.The bracing pattern wasdetermined through an intenseiterative analysis and in closecollaboration with the arc h i t e c t .

The principal structure of thebuilding was modelled in OasysGSA and re p re s e n t a t i v eloading applied, including astatic equivalent load for a level1 earthquake. Initially, a uniformbracing pattern (Fig 7) wasadopted, and the SRC andsteel columns sized

a p p ro p r i a t e l y. The distributionof forces within the braces wasthen investigated, and theresults categorized into thre eg roups with an appro p r i a t eaction applied to the braceswithin a particular gro u p :

• densify the mesh by addingbraces; ‘doubling’

• keep the same

• rarefy the mesh by re m o v i n gbraces: ‘halving’.

The structure is highlyindeterminate, so changing thebracing pattern resulted in anew distribution of forc e swithin the columns, braces,and edge-beams.

Changes in stiffness alsochanged the dynamicbehaviour of the structure andhence the seismic forces thata re attracted. Pattern i n gbecame an extremely involvediterative pro c e s s .

An unfolded stress patternview of the structure wasdeveloped (Fig 8) to clearlydisplay the results of the wholebuilding in one view, ora l t e rnatively folded up into a3-D model (Fig 9). This enabledthe design team to visuallyp rocess the results quickly,while keeping the arc h i t e c tup-to-date and involved in thedevelopment of the pattern .

A performance-based design appro a c h

Chinese approvals pro c e s s

The legal framework in China governing buildingdesign practice is similar to those of Japan andsome continental European countries where thedesign codes are legal documents published ande n f o rced by the state government. Designengineers must comply with the codes whendesigning buildings and structures covered by theirscope, but equally the codes provide legalp rotection to the design engineers who are re l i e v e dof any legal responsibilities by virtue of compliance.The Chinese code for seismic design of buildings(GB50011 – 2001), sets out its own scope ofa p p l i c a b i l i t y, limiting the height of various systemsand the degree of plan and vertical irre g u l a r i t i e s .Design of buildings exceeding the code must got h rough a project-specific seismic design expertpanel review (EPR) and approval process as set outby the Ministry of Construction.

Although the 234m height of the CCTV buildingis within the code’s height limit of 260m for steeltubular structural systems (framed-tube, tube-in-tube, truss-tube, etc) in Beijing, its geometry is non-compliant. The Seismic Administration Office of theBeijing Municipal Government appointed an expertpanel of 12 eminent Chinese engineers andacademics to closely examine the structural design,focusing on its seismic resistance, seismic structuraldamage control, and life safety aspects.

Arup realised the importance of engaging theexpert panel early in the design process, and thre einformal meetings were held to solicit feedback andgain trust before the final formal presentation. Thepanel was under enormous public and statescrutiny to closely examine Arup’s design, ther i g o rous nature of which - aided by successfulcollaboration with ECADI - was of vital importance.F rom day one the building received wide publicity inChina, and concerns were voiced by both thegeneral public and the government over the safetyof the adventurous design - indeed it came to bere f e r red to by the public and the media as ‘We iFang’ (the dangerous building): an indicator of thescrutiny the project and hence Arup were under.

Seismic re q u i re m e n t s

As the seismic design lay outside the scope of thep rescriptive Chinese codes of practice, Arupp roposed a performance-based design appro a c hf rom the outset, adopting first principles and state-of-the-art methods and guidelines to achieve setperformance targets at diff e rent levels of seismicevent. Explicit and quantitative design checks usinga p p ropriate linear and non-linear seismic analysisw e re made to verify the performance for all thre elevels of design earthquake.

Developing and optimizing the bracing pattern

7 .Brace stresses for au n i f o rm grid.

9 .Models illustrating thedevelopment of thefaçade pattern .

8 .Unfolded view of the stru c t u re, showinga reas to densify or rarefy the mesh.

9247_Arup.qxd 16/9/05 4:53 pm Page 6


The criteria for this performance-based design arebeyond those usually applied to such buildings inChina, and were set by the Arup design team inconsultation with the expert panel to reflect theimportance of the building both to the client and tothe Chinese Government. The basic qualitativeperformance objectives were :• no structural damage when subjected to a level

1 earthquake with an average re t u rn period of50 years (63% probability of exceedance in50 years)

• repairable structural damage when subjected toa level 2 earthquake with an average re t u rnperiod of 475 years (10% probability ofexceedance in 50 years)

• s e v e re structural damage permitted but collapsep revented when subjected to a level 3earthquake with an average re t u rn period of2500 years (2% probability of exceedance in50 years)

For the CCTV development site, the peak horizontalg round acceleration values associated with thet h ree levels of design earthquake are 7%, 20% and40% of gravity re s p e c t i v e l y.

Elastic superstru c t u re design

With the bracing pattern determined from the initialconcept work, a full set of linear elastic verificationanalyses were performed, covering all loadingcombinations including level 1 seismic loading, forwhich modal response spectrum analyses wereused. All individual elements were extensivelychecked and the building’s global performanceverified. Selected elements were also initiallyassessed under a level 2 earthquake by elasticanalysis, thus ensuring key elements such ascolumns remained elastic.

The elastic analysis and design was principallyperformed using SAP2000 and a custom-writtenChinese steelwork code post-pro c e s s o r, whichautomatically took the individual load cases appliedto the building and combined them for the limitstate design. Capacity ratios were then visuallydisplayed, allowing detailed inspection of the criticalcases for each member. Due to the vast number ofelements in the model – 10 060 elementsre p resenting nearly 90 000m of steel and SRCsections – and the multitude of load cases, fourp o s t - p rocessors were run in parallel, one for steelcolumns, one for SRC columns, one for braces, andanother for the edge beams that together form thecontinuous tube. The SRC columns used a modifiedp o s t - p rocessor to account for the diff e re n c e sbetween the steel and SRC codes; sectionp roperties of these columns were determined usingX Tract, which also computed the properties for thesubsequent non-linear analyses.

The post-processor provided a revised element list which was imported back intoSAP2000, and the analysis and post-processing repeated until all the design criteriaw e re met. As the structure is highly indeterminate and the load paths are heavilyinfluenced by stiffness, each small change in element property moves load aro u n dl o c a l l y. Optimizing the elements only for capacity would result in the entire loadgradually being attracted to the inside corner columns, making them pro h i b i t i v e l ylarge, so careful control had to be made of when an element’s section size could bereduced and when there was a minimum size re q u i red to maintain the stiffness ofthe tube at the back face.

To further validate the multi-directional modal response spectrum analyses,level 1 time-history checks were also made using real and artificially-generatedseismic re c o rd s .

Non-linear superstru c t u re seismic design and perf o rmance verification

For the performance-based design, a set of project-specific ‘design rules’ werep roposed by the design team and reviewed and approved by the expert panel,c reating a ‘road map’ to achieve the stated seismic performance objectives.A p p ropriate linear and non-linear seismic response simulation methods wereselected to verify the performance of the building under all three levels of designearthquake. Seismic force and deformation demands were compared with theacceptance limits established earlier to rigorously demonstrate that all thre equalitative performance objectives were achieved.

Inelastic deformation acceptance limits for the key structural brace members inthe continuous tube were determined by non-linear numerical simulation of thepost-buckling behaviour. LS-DYNA, commonly used to simulate car crashb e h a v i o u r, was used for this work. The braces are critical to both the lateral as wellas the gravity systems of the building and are also the primary sources of ductilityand seismic energy dissipation. Non-linear numerical simulation of the braces wasneeded to establish the post-buckling axial force/axial deformation degradationrelationship to be used in the global 3-D non-linear simulation model. It was alsoused to determine the inelastic deformation (axial shortening) acceptance limit inrelation to the stated performance criteria. Post-buckling inelastic degradationrelationship curves illustrate the strength degradation as the axial shorteningi n c reases under cyclic axial displacement time history loading. The acceptableinelastic deformation was then determined from the strength degradation‘backbone’ curve to ensure that there was sufficient residual strength to supportthe gravity loads after a severe earthquake event.

Many software packagesw e re used to deliver theCCTV structural design, somedeveloped in house by Arup:

CSI SAP2000: limited non-linear structural analysis anddesign with static, dynamicand push-over capability

Oasys LS-DYNA Enviro n m e n t :non-linear explicit time historya n a l y s i s

Oasys GSA: linear structuralanalysis with static andresponse spectrum analysisdynamic capability

Oasys Vdisp: plug-in for GSAto analyze non-linear soils t i ff n e s s

Oasys GSRaft: iterative non-linear soil-structure interactiona n a l y s i s

Oasys Compos: compositesteel beam and concrete slabanalysis and design

Oasys Adsec: composite(SRC) cross-section analysis

Oasys ADC: re i n f o rc e dc o n c rete design package

Xtract: non-linear large straincomposite (SRC) cro s s -section analysis

MSC/Nastran: heavy dutyfinite element analysisp a c k a g e

Xsteel: 3-D CAD package.

1 0 . Non-linear finite element simulation model showing local buckling of a typical steel brace.

S o f t w a re

9247_Arup.qxd 16/9/05 4:53 pm Page 7


Having established the inelastic global structure andlocal member deformation acceptance limits, thenext step was to carry out non-linear numericalseismic response simulation of the entire 3-Dbuilding subjected to level 2 and level 3 designearthquakes. Both the non-linear static pushoveranalysis method and the non-linear dynamic timehistory analysis method were used to determine theseismic deformation demands in terms of themaximum inelastic inter- s t o rey drifts and themaximum inelastic member deformation. Thesedeformation demands were compared against thes t r u c t u re ’s deformation capacities store y - b y - s t o re yand member-by-member to verify the seismicperformance of the entire building. All global andlocal seismic deformation demands were shown tobe within their respective acceptance limits,demonstrating that the building achieves thequantitative and hence qualitative performanceobjectives when subjected to level 2 or level 3e a r t h q u a k e s .

Foundation design

The design of the foundations re q u i red that the applied superstructure loads beredistributed across the pilecap (raft) so as to engage enough piles to pro v i d eadequate strength and stiffness. To validate the load spread to the pile group, acomplex iterative analysis process was used adopting a non-linear soil model.

The superstructure loads were applied to a discrete model of the piled raftsystem. Several hundred directional load case combinations were automated in as p readsheet controlling the GSRaft soil-structure interaction solver (Fig 11).This pro c e d u re iteratively changed the input data in response to the analysis re s u l t sto model the redistribution of load between piles when their safe working load wasreached. The analysis was then repeated until the results converged and all pilesw e re within the allowable capacities. The envelope of these several-hundre danalyses was then used to design the re i n f o rcement in the raft itself.

C o n n e c t i o n s

The force from the braces and edge-beams must be transferred through and intothe column sections with minimal disruption to the stresses already present in thecolumn. The connection is formed by replacing the flanges of the steel column withlarge ‘butterfly’ plates, which pass through the face of the column and thenconnect with the braces and the edge-beams. No connection is made to the webof the column to simplify the detailing and construction of the concrete around thesteel section.

The joints are re q u i red to behave with the braces, beams, and columns as‘ s t rong joint/weak component’. The connections must resist the maximum pro b a b l eload delivered to them from the braces with minimal yielding and a relatively lowd e g ree of stress concentration. High stress concentrations could lead to brittlef r a c t u re at the welds under cyclic seismic loading, a common cause of failure inconnections observed after the 1994 Northridge earthquake in Los Angeles. Tw oconnections, re p resenting the typical and the largest cases, were modelled from theoriginal AutoCAD drawings using MSC/NASTRAN. The models were analyzed,subjected to the full range of forces that can be developed before the bracesbuckle or yield - assuming the maximum probable material properties - to evaluatethe stress magnitude and degree of stress concentration in the joints. The shape ofthe butterfly plate was then adapted by smoothing out corners and notches untilpotential regions of yielding were minimized and the degree of stress concentrationreduced to levels typically permitted in civil and mechanical engineering practice.CAD files of the resulting geometry of the joints were exported from the finiteelement models and used for further drawing production (Fig 12).

Transfer stru c t u re s

Whilst the external tube structure slopes to give the unique geometry, the intern a lsteel columns and cores are kept straight for functional layout and to house lift andservices shafts. This resulted in a diff e rent configuration for every floor - the spansf rom core to façade, and internal column to façade, change on each. Consequently,i n t e rnal columns can be removed where the floor span decreases sufficiently onone side of the core. Similarly, additional columns are needed up the building heightw h e re the floor spans increase significantly on the other side of the core. Tr a n s f e rtrusses support these additional columns, spanning between the internal core andthe external tube structure. They are typically two storeys deep and located in plantfloors so as to be hidden from view and to minimize the impact on floor planning.

The sizes of the transfer trusses mean that they could potentially act asoutriggers linking the external tube to the internal steel cores - undesirable as thiswould introduce seismic forces into the relatively slender internal cores. The designp re f e rence is to keep all the seismic forces in the more robust diagrid framing of thecontinuous tube. The transfer trusses are thus connected to the internal cores andthe external columns at singular ‘pin-joint’ locations only. Detailed analyses weremade to ascertain that no outrigger effects result from the transfer truss geometry.

1 2 .The von Mises stress distribution of a large connection plateunder the most unfavourable loading combination.

11. GSRaft model of the piled foundation.

Loads extracted from 3-Ds u p e r s t ru c t u re model

5 x 5m rafts p r i n g s

5 x 5m piles p r i n g s

51.5m long,1.2m diameterpiles. Loadt r a n s f e rred tosoil at 46.5mbelow baseof raft

0

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

6 0 0

M P a

9247_Arup.qxd 16/9/05 4:53 pm Page 8


C re d i t sC l i e n t : China Central Television A rc h i t e c t : O M AStedebouw BV Geotechnical, structural, MEP, fire ,and security consultant: Arup - Abdel Ahmed, Chuan-Zhi Bai, Cecil Balmond, Carolina Bartram, Chris Carro l l ,Wayne Chan, Mark Choi, Dean Clabrough, Paul Cro s s ,Roy Denoon, Omar Diallo, Mimmy Dino, Xiaonian Duan,Gary Ge, Craig Gibbons, Sam Hatch, Guo-Bo He,Xue-Mei He, Colin Ho, Goman Ho, Yi Jin, JonathanK e r r y, Michael Kwok, Francis Lam, Peter Lam, RichardLawson, Alexis Lee, Jing-Yu Li, Zhao-Fan Li, Ge-QingLiu, Peng Liu, Quan Liu, Pierre Lui, Man-Kit Luk,A n d rew Luong, John McArthur, Rory McGowan,Hamish Nevile, Gordon Ng, Xiao-Chao Pang, JackPappin, Steve Peet, Bill Peng, Dan Pook, Chas Pope,Qun Shi, Andrew Smith, Stuart Smith, Derek So,G Y Sun, George Thimont, Alex To, Fei To n g ,Paul Tonkin, Ben Urick, Bai-Qian Wan, Yang Wa n g ,Will Whitby, Robin Wilkinson, Michael Wi l l f o rd, MichelleWong, Stella Wong, Eric Wu, Jian-Feng Yao, AngelaYeung, Kenneth Yeung, Te rence Yip, George Zhao,Julian Zheng (analysis, geotechnical, structural)I l l u s t r a t i o n s : 1, 9 ©OMA; 2 Nigel Whale; 3 -14 Arup

The next article in this series, discussing the buildingservices design of the CCTV development, will appearin The Arup Journal 3 / 2 0 0 5 .

14. Site work begins, September 2004.

The floors cantilevered from the two leaning Towers to form the Overhang areenclosed by the continuous tube structure on the outside. This supports a two-waytransfer deck in the bottom two storeys of the Overhang, carrying the columns forthe floors above.

The Base also contains major transfer trusses, spanning over the principalstudios to support the columns and floors above.

Construction issues

The building’s unique form necessitated careful consideration of the constructionmethod throughout the design process. Both the method and sequencing of theworks (Fig 13) will affect the permanent distribution of dead load through thecontinuous tube.

To allow the contractor some flexibility in method and programme, upper andlower bound analyses were performed, using staged construction and loading tobuild up the final dead load incre m e n t a l l y. The lowest bound of loading when thetwo Towers are connected puts the highest stresses into the Overhang structuresince it acts as a prop between the Towers, while the upper bound puts the largests t resses into the Towers since they carry more of the load in bending as ac a n t i l e v e r. Between these two extremes there is scope for the contractor to choosehis programme, and to propose alternative erection pro c e d u re s .

The timing of the connection between the Towers is also important, so as tominimize the relative movement between them from thermal and wind effects. It isalso important to minimize future thermal movements between the Towers thatcould put large stresses into the Overhang where it is restrained by the To w e r s .

C o n c l u s i o n

The structural design of CCTV posed manytechnical challenges to the large international team.They were successfully overcome, within a very tightp rogramme. Arup’s unique global depth ofexperience and knowledge made this possible,enabling the right people to be involved at the rightstages of the project. The Arup team delivered thedesign through a seamless global collaboration,transcending time zones, physical distance,c u l t u res, cost centres, and even the SARSo u t b reak. Foremost in the team’s collective mindwas the need to deliver the complex design on timefor the client and in so doing win the approval ofthe Chinese Ministry of Construction expert panel.

1 3 . One of several construction sequence loading arrangements considere d .

9247_Arup.qxd 16/9/05 4:53 pm Page 9

Documents

ARUP Structural Engineering Design