arup_journal_2-2008

8/9/2019 Arup_Journal_2-2008

1/60

2/2

The Arup Journal

8/9/2019 Arup_Journal_2-2008

2/60

2 The Arup Journal 2/2008

1.

When the world’s largestobservation wheel, the SingaporeFlyer , was opened on 15 April 2008by Singapore’s Prime Minister,Lee Hsien Loong, it was very mucha national celebration betting thisiconic structure. Prime MinisterLee struck a symbolic beat ofthe ceremonial drum, and initiateda spectacular light show andreworks display (Fig 1).He proudly stated: “I am veryhappy with the project; it is ontime and on schedule. I thinkit’s achieved what we hoped.”

Contents

2 The Singapore Flyer Andrew Allsop Pat Dallard Heng Kok Hui

André Lovatt Brendon McNiven

15 The Virtual Building Peter Bailey Daniel Brodkin John Hainsworth

Erin Morrow Andrew SedgwickMartin Simpson Alvise Simondetti

26 Designer’s Toolkit 2020 : A vision for the design practice Alvise Simondetti

34 Terminal 5, London Heathrow:The new control tower

Jeremy Edwards Richard MatthewsSean McGinn

40 CCTV Headquarters, Beijing, China:Building the structureChris Carroll Craig Gibbons Goman HoMichael Kwok Richard Lawson Alexis LeeRonald Li Andrew Luong Rory McGowanChas Pope

52 The European Extremely Large Telescopeenclosure designDavar Abi-Zadeh Philip Bogan Jac Cross

John Lyle Pieter Moerland Hugo MulderRoland Trim

8/9/2019 Arup_Journal_2-2008

3/60

The Arup Journal 2/2

Inception and government support

The Singapore Government plans to positionSingapore as a leading tourism hub for Asia.It has set ambitious targets for the tourism industry– to triple receipts to S$30bn, double visitor arrivalsto 17M, and create 100 000 additional tourism-related jobs by the year 2015. It aims to transformthe tourism landscape to realise this vision.

The Singapore Flyer exemplies what is to come This giant observation wheel (GOW) occupies aprime site in the Marina Bay area and is one ofthe “necklace of attractions” planned to alterthe future landscape of downtown Singapore.

The Flyer was conceived as the key element ina development-led project by Melchers ProjectManagement Pte Ltd (MPM), a subsidiary ofC Melchers GmbH & Co, an internationallogistics and engineering services company.

The

Singapore Flyer Andrew Allsop Pat DallardHeng Kok Hui André LovattBrendon McNiven

Set amidst a “necklace” ofattractions, the Singapore Flyer has become the latest additionto the city skyline. Arup builton knowledge gained duringthe design of the London Eye to develop a “next generation”rim structure. The resulting two-dimensional “ladder truss” rim isboth larger in diameter and lighterthan that of its predecessor.

8/9/2019 Arup_Journal_2-2008

4/60


SuntecConvention

Centre

ConventionCentrestation

ConradHotel

PanPacificHotel

FutureMilleniastation

East CoastParkway

Raffles Avenue

RafflesBoulevard

Open parkingfor 40 tour

buses

Terminalbuilding

GOW

Concertamphitheatre

45m spanlink bridge

East CSinga

Stra

CBD,Marina Bay

Ritz-CarltonHotel

MandarinOriental

Hotel

MarinaMandarinHotel

TheEsplanade

Two-storeycar parkbuilding

(280 cars)

3. Singapore Flyer location plan.

The proposal to develop the Singapore Flyer as amust-see, must-do tourist attraction in Asia wasagreed in 2003. The huge wheel was to be an iconiclandmark and a compelling draw for foreign visitorsto the garden city. The Singapore Tourism Boardsupported the project by purchasing the land for thedevelopment and leasing it back to Singapore FlyerPte Ltd, initially for 30 years but with an option for afurther 15 years. The land was rent-free up to therst day of operation.

Overview

The Flyer is located on the peninsula of land thatseparates Marina Bay from the Kallang Basin, and isoriented to overlook the new downtown aroundMarina Bay in one direction and to provide aspectacular view of the East Coast and SingaporeStraits in the other (Figs 2-4). As the project formspart of the government’s tourism blueprint to developMarina Bay’s new waterfront, its prime location issited close to the future Millenia Mass Rapid Transit

(MRT) train station. This iconic visitor attraction offers passengers

a spectacular sightseeing experience. The 28 fullyair-conditioned capsules, each accommodating28 people, are attached to the outer rim of the150m-diameter wheel, which at the top of itsrevolution reveals a 45km panorama of Singapore,Malaysia, and Indonesia.

Visitors board the capsules via access gantriesand loading platforms on the third storey of theterminal building at the base of the wheel.

The building not only houses the passenger owinfrastructure required, but also includes 15 000m 2 ofretail shopping space. A tropical rainforest attractionreplete with water features is incorporated in thecourtyard space immediately below the wheel toadd to the visitor experience.

A 280-lot car park space located across Rafes Avenue is linked to the terminal building by apedestrian bridge. This fully-covered access allowsvisitors to appreciate the environs while makingtheir way to the main building. The surrounding areaalso accommodates a concert amphitheatre forperformances and other artistic pursuits.

The Japanese architect Kisho Kurokawa preparedthe design concept for the building works (Fig 5),whilst DP Architects in Singapore (Fig 4) carried out

all the nal documentation, acting post-concept asthe local architect of record.

At its highest point, the Singapore Flyer stands atotal of 165m tall, making it the world’s largest GOW.It surpasses the well-known London Eye by 30m,and thanks to a more efcient and innovative design,is not only larger but also lighter and slimmer than itspredecessor.

2. The Flyer stands above the three-storey terminal building.

8/9/2019 Arup_Journal_2-2008

5/60


The detailed design was then followed through by theGOW’s contractor, Mitsubishi Heavy Industries, in adesign/build form of contract, with Arup acting in anovated role as the engineer of record, signing andsubmitting to the local authorities.

Giant observation wheels: a short history

GOWs form one lineage in a family of visitorattractions known as iconic viewing platforms (IVPs).Gustave Eiffel’s Tower, the centrepiece of the 1889Paris Exposition, was perhaps the rst purpose-builtIVP of modern times, and remains one of the world’smost successful with more than 200M visitors sinceit was opened.

The founder of the GOW lineage was GeorgeFerris’s Wheel, designed and built as the principalengineering attraction of the 1893 Chicago World Fair(Fig 6), and with the intention of creating anengineering marvel to rival the Eiffel Tower’sspectacular success. This original Ferris Wheel was 76m in diameter and had 35 cabins, each of

which was able to accommodate up to 60 people.It was demolished in 1906.

Two years after the Ferris Wheel began operation,an 86m diameter, 40-car rival was built by theGigantic Wheel and Recreation Towers Company Ltdfor the Empire of India Exhibition, Earl’s Court,London. Several more GOWs were subsequentlycommissioned and built around the world,characterised by their size and advancedengineering. Notable amongst them are:

Riesenrad , 61m diameter (Fig 7).Burnt down in 1944 and rebuilt the following year,albeit with only 15 cabins of 12-person capacityrather than the original 30 cabins, it wasimmortalised in the 1949 movie The Third Man as the location of the famous speech by thecharacter Harry Lime (Orson Welles) toHolly Martins (Joseph Cotten).

La Grande Roue , Paris ExpositionUniverselle (Fig 8), approximately 80-100mdiameter, 36 cabins with 8-10-person capacity.It was demolished in 1937.

London Eye (“Millennium Wheel”)(Fig 9), 135m diameter, 32 capsules with25-person capacity.

Foundations

The geology of the Marina Bay area is typicallyrecent marine and uvial sediments of the KallangFormation, varying from unconsolidated tonormally consolidated. These materials overlaythe Old Alluvium present at 15-30m depth.

About 30 years ago the site was reclaimed usingll over the existing strata.

Arup’s role

The design of the Flyer itself was very muchengineering-led. Arup built upon knowledge gainedduring the design of the London Eye to develop athinner, lighter, “next generation” rim structure with amore efcient structural geometry and cablearrangement. The Flyer ’s two-dimensional “ladder

truss” rim structure gives it less bulk than the LondonEye ’s three-dimensional triangular rim, as well asreducing the wind loads.

At the outset of the project, Arup worked closelywith MPM in a nancial risk/reward partnershiparrangement that involved reduced initial fees, thensupplemented by success payments upon theproving of project feasibility. Arup took the designevolution from the initial conceptual ideas throughscheme development up to tender stage.

4. Perspective looking north-west.

6. The original FerrisWheel , ChicagoWorld Fair, 1893.

7. Vienna Riesenrad (“Giant wheel”), 189

8. La Grande Roue ,Paris Exposition,1900.

9. 2000: London Eye(“Millennium Whee

5. Architect’s initial concept sketch for the terminal building.

3G design

Arup developed a “third generation” rim design, usingstate-of-the-art technology that makes the most ofthe strength and arrangement of the cables to reducethe size and visual appearance of the rim to a two-dimensional truss (which from a distance seems todisappear in relation to the size of the wheel).

“First generation” : laced compression spoke wheelstypical of fairground park attractions

“Second generation” : 3-D box or triangular truss withtension cables as in the Ferris Wheel or the London Eye

“Third generation” : 2-D “ladder” truss rim of theSingapore F lyer.

8/9/2019 Arup_Journal_2-2008

6/60


It was decided at the outset not to incorporatebasements in the project, to avoid unnecessary costand impact to the programme. The foundations forthe buildings and wheel are bored piles between600mm and 1500mm in diameter, and penetratingup to 52m in depth, socketed into the Old Alluvium(Fig 11). The piles were fully cased through the exof the reclaimed ll and soft marine clays.

Support ing structure

The wheel is supported by two 2.85m diametercolumns, founded in the courtyard of the terminalbuilding below and stabilised at the spindle level bfour cable stays. Each cable stay comprises six100mm diameter locked coil cables prestressedto 17MN (Fig 13).

The lateral components of the stay pre-tensionsare resolved through the spindle structure at the higlevel, and through the terminal building’s ground structure (acting as a compression annulus), at thelow level. The result is a relatively stiff closed

structural system that distributes and balances thelateral components of the permanent pre-tensionforces in the structure. The piles in essence arethen only required to resist the vertical uplift anddownwards reactions, and the net lateral forcearising from wind loading, etc (Figs 12, 14).

Term inalbu ild ing p iling

37-52m

Term inalbu ild ing p iling

37-52m

Temporary steel cas ingto base of soft clays

8m recla imed (f ill)

5-18m mar ine clay

Old Alluv ium

Car parkp iling

42-47m

11 . S ite geology .

13 . Anchorages in the courtyard for the cable stays .

10 . The wheel is supported by two 2 . 85m d iameter columns .

12 . Ground oor annulus n ite element model .

kN/m

1122-242 . 5-1067-2927-4337-5702-7066-8431

8/9/2019 Arup_Journal_2-2008

7/60


Rim and spoke design

The rim and spokes are the components thatdifferentiate wheels from all other types of structure,and which pose some of the principal engineeringchallenges in designing GOWs.

Three external load cases generate signicantforces in the rim and spokes. These forces aredescribed assuming that the spokes can resistcompression. Firstly, gravity causes tension in thelower spokes and compression in the upper ones,along with compression in the lower half of the rimand tension in the upper half. Secondly, wind causes

tension in spokes attached to the windward side ofthe hub and compression in those attached to theleeward side. Thirdly, temperature differentialsbetween the rim and spokes cause spoke tensionand rim compression, or vice versa.

The Singapore Flyer uses cable spokes thatneed to be prestressed to resist compression.

The prestress is set such that under factored loadsnone of the cables go slack, so they remain effectivein controlling the displacement of the rim. While theprestress is necessary, the compression it inducesin the rim dominates the rim design. Achieving anefcient design for the rim requires the prestress tobe minimised (Fig 16).

The 2-D ladder truss helps reduce the wind loadon the Flyer rim. This is important as, even thoughthe wind load is only about a 10th of the weight,it generates approximately the same prestressrequirement because the cable angles areunfavourable for resisting lateral load. To minimise theprestress required against wind, the width of the Flyer hub was maximised and cables were selectivelycrossed to the opposite side of the rim to increasetheir efciency.

Ground floor plate provides“annulus” stiffness, balancinginternal forces (prestressing,dead load, live load,temperature, etc).

Whole of building +GOW superstructuremoves as one piece,floating on top of piles.

14. Design sketch showing force resolution for thesupporting structure.

15. The Flyer rim is a “third generation” two-dimensional lattice truss.

16. Spoke cable components of force.

+ + =T TT

T

C

CCC

CT

T

C

T T

TT

T

T

T

T

C C

C

C

C

C

C

Dead + l ive loads W ind loads Prestress Total: no cablescompress ion (slack

Pr imary load cases

8/9/2019 Arup_Journal_2-2008

8/60

8/9/2019 Arup_Journal_2-2008

9/60


that the variance in dead load that could be expectedin the structure was low when compared to that of atypical building. Here, the weights of the capsulesand other applied dead loads were known muchmore accurately than typical building dead loads.

As a result, dead load factors more akin to thoseused in bridge design were adopted for the designof the rim structure.

Dynamics

Passenger comfort is a key design consideration forGOWs. Comfort in terms of vibration dependsparticularly on the wind response of modes involvinmovement out of the plane of the wheel. The teamstudied how the properties of these modes wereaffected by changes in lateral restraint at the bottomof the wheel, and changes in the stiffness of thesupport structure cable stays. The optimum level ofdamping to be added was also examined.

The studies showed that comfort benets couldbe gained by increasing the size of the supportstructure cable stays over that required for strength,so as to enhance their stiffness. They also concludedthat damping should be introduced at the base ofthe wheel, so this was incorporated into thepassenger deck structures along with the drivetrain mechanisms (Fig 21).

Lower rimcompression

Reduceddead load

Lower spokeprestress

L

20. Virtuous dead load cycle.

21. Wind response modes: (a) in plan rotation (0.2Hz); (b) Later al displacement at top of rim (0.42Hz); (c) Longitudinal displace ment of support structure (0.60Hz);(d) Four-lobe displacement of rim (0.65Hz); (e) Six-lobe displacement of rim (1.1Hz).

19. Passengers mount the Flyer from the boarding deck at the third storey of the terminal building.

a) b) c) d) e)

8/9/2019 Arup_Journal_2-2008

10/60


Wind tunnel testing

In view of the importance of the wind loads in helping to determine minimum prestrelimits, a segment of the rim and a capsule was tested in MHI’s Nagasaki wind tunnelfacilities to verify the assumptions made on wind drag (Fig 22). Only a segment of thwheel was tested in this large and high-speed tunnel, since the model needs to be ata scale where Reynolds’ number effects can be managed. Measurements were takenfor a variety of wind approach angles and rim inclinations to enable accurate

application of the results in the design model. There was some doubt about the extra drag that would result from the cylindricalshape of the Flyer capsules, compared to the better aerodynamic shape of those onthe London Eye . It was also necessary to model accurately the service bus-bars anddrive plates, etc, which signicantly increase wind drag compared to the bare tubesof the rim structure itself.

Due to programme constraints, the foundations were designed using moreconservative assumptions on overall drag, prior to the wind tunnel results beingavailable. The more rened wind tunnel test results were incorporated into thesuperstructure design.

Aeroelastic stability

Questions were also raised about the risk of large amplitude vibrations due to effectssuch as “galloping”, “vortex shedding” and “utter”. The porous nature of the rim an

the low sustained wind speeds in Singapore both pointed away from problems withresponse of the whole wheel. Local vibrations of long slender tubular elements andcables were also considered. The main elements of the rim were found to be stable,but the possible need for cable dampers was kept on the risk register.

The main strut columns were found to fall within the range of potential vortexshedding. Tuned mass dampers were installed at mid-height in each of the columnsafter site measurements of the natural inherent structural damping were found to bebelow the values required to mitigate response.

Some vibration in the cable spokes was also observed on site during constructionand ascertained to be due to wind/rain-induced responses. Rivulets of water runningdown the spokes alter the geometric form and result in a dynamic response.Stockbridge dampers tuned to the third and forth natural frequencies of the cables(those frequencies at which resonance was observed), were provided subsequentto operations commencing.

Wind loading

Climate and design wind speeds

Singapore experiences unique wind conditions:“Sumatra” squalls blow in from the Straits, and in thismixed wind climate monsoons and thunderstormsare also commonplace.

While in general, wind speeds are low, the peakgusts in Singapore, resulting from thunderstorms,can arise very quickly and with limited warning.It is therefore difcult to reliably manage evacuationsof the wheel in advance of strong winds as can bedone on the London Eye . Fortunately such stormsnormally consist of only a few strong wind gusts.

The variation of wind speed with height inconvective events (such as thunderstorms) is knownto be quite different from the standard code proles,and often the strongest winds occur below 100mheight. Unfortunately there are currently no

procedures that can be considered reliable formodelling this kind of behaviour, so in accordancewith current design practice a standard wind modelwas assumed to t the predicted 50-year gust speedat 10m height. This model is likely to overestimate thewind gust speeds as the top of the wheel but mayunderestimate the dynamic response factor -a rational compromise, given the unknowns!

An allowance for the provision of dampers on the rimwas made in the design should they have proven tobe required under actual wind conditions.

During normal operation, a wind speed limit of13m/sec average at 10m height was used, togetherwith gust and dynamic response calculations basedon the ESDU (Engineering Sciences Data Unit) windmodel, which is compatible with British Standardcode design. Given the unpredictable nature ofsquall/thunderstorm conditions in Singapore,however, a design acceleration limit (comparable tothat experienced on the MRT trains) was imposedunder the full design wind condition. Damping wasalso provided to ensure movement dies out quicklyand any passenger alarm quickly alleviated.

22. Capsule and rim segment during wind tunnel testing.

23. Wind tunnel tests conrmed the drag on the cylindrical capsules.

8/9/2019 Arup_Journal_2-2008

11/60


After the support columns and spindle were in place, the wheel itself was erected in a“pie slice” fashion. Rim segments were delivered to site and laid to level on atemporary stage. Cables were then installed in a slack condition. Temporarycompression struts were provided between the hub and the rim enabling eachsegment of the rim to be stable in its own right.

Upon completion, each segment was rotated to clear the way for installing the nextsegment and so on (Figs 24, 25). Additional strengthening was provided to the rim inthe form of lightweight chords forming bowstring trusses and maximising the size ofthe segments able to be built. Once the full wheel was in place, the cables werestressed in two stages, and the capsules installed.

Laser technology measuring microtremors in the cables was used to ascertain theforce in each of the cable spokes at different stages in stressing. A full set of cabletensions were measured over a three-night period from survey stations at groundlevel, and checked against analytical predictions.

Erection method

Possible erection methods (Fig 26, overleaf) werestudied in detail with both client and contractor, so asto satisfy several constraints:

tolerances

The horizontal lifting method used on the LondonEye , whereby after assembly on platforms on the river

Thames the entire wheel was raised by strand jacksto the nal vertical position, was not favoured for theFlyer . This was primarily due to space constraints onsite, but also because of geometric clashes with theterminal building and the support legs during lif ting.Instead, a vertical erection method was used.

First, the main support structure columns wereerected in segments using bolted splices, and thenthe hub and spindle arrangement (180 tonnes) was

lifted by strand jacks off a temporary gantry spanningbetween the tops of the main columns.

Initially the rim segments were intended to bebarged to site using access from the adjacent MarinaBay and Singapore Straits. In the end, however, thisproved impossible due to the barrage (a projectconverting the entire Marina Bay and Kallang Basinwater bodies into a freshwater reservoir), sealing offaccess to the Straits. Instead the steelwork was sizedand detailed to allow transportation by road.

24. Stages of erection: The wheel was erected in a “pie-slice” fashion (a). Each segment was rotated (b) until all segments had been installed (c). Once the wheel was erected,the spoke cables were stressed in two stages (d). The temporary struts were then removed, leaving only the rim attached to the central hub by cables (e), followed by theinstallation of the cabins (f).

25. View looking north-west during erection.

a)

d)

b)

e) f)

c)

8/9/2019 Arup_Journal_2-2008

12/60


MEP engineering

The terminal building is unusual in that it is a buildviewed mostly from above. With this in mind, thearchitects were keen to minimise the amount ofrooftop clutter and services required. Most of themain services are therefore sited across the road in compact drum area appended to the car park

building and connected to the terminal building viathe link bridge across Rafes Avenue. The terminal building was designed to maximis

non-air conditioned open public spaces through theprovision of circulation areas that provide shade anpromote natural air ow. The building’s “doughnutshape also allows a high degree of connectivity to central rainforest attraction, resulting in a comfortatropical feel to the building environment. One of thearly buildings to be so assessed, the terminalbuilding achieved a Green Mark award under thelocal environmental accreditation scheme.

Trafc

Arup carried out the original trafc impactassessment required for the project. The impact onsurrounding intersections during the opening year, future case of the year 2015 (under trafc forecastsprovided by the Land Transport Authority), as wellthe suitability of the level of car parking to beprovided, were all studied.

Passenger boarding platforms/bridges

The passenger boarding bridges are the interface where all the requirements of theGOW operation come together at one point. These requirements often conict.

They include lateral structural support and damping to the base of the wheel, cateringfor the forces imposed by the drive motors and braking requirements, delivery ofelectrical power, provision for operating equipment and operations staff, and nally thenecessity for a column-free slot to allow passengers to board and disembarkunhindered. These all had to be considered in arriving at the nal architectural andstructural form.

A curved composite steel/concrete drive deck was nally adopted. Capable ofsupporting the various drive motors and dampers, etc, it also affords some acousticprotection to the passengers and operators immediately below. The deck is in turnsupported off a large CHS triangular truss capable of resisting the torsions generatedfrom the eccentricities of the deck and cantilever passenger platforms. The wholearrangement was supported three storeys off the ground by steel towers acting ascantilevers to resist lateral and longitudinal loading (Fig 27).

Passenger boarding bridges span the gap between the platforms and the terminalbuilding, and movement joints at the building interfaces ensure the whole arrangementacts independently.

Fire engineering

Arup’s re engineers used a performance-based re strategy for the terminal building, This enabled the stairs to be reduced, yielding considerable nancial benets for the

client. Compared with prescriptive methods this approach saved 6m of requiredegress width, the equivalent of approximately 400m 2 of oor area (Fig 28). This wasclassed as part of the developable oor area permitted by Singapore’s UrbanRedevelopment Authority, and freeing it up as lettable area improved the building’snet-to-gross ratio as well as providing for a more considered re safety strategy.

27. Steelwork isometric of passenger boarding deck bridges.

28. Advanced computerised pedestrian sof tware (STEPS) wasused to simulate the escape patterns and evacuation timerequired to aid the re engineering design.

26. Alternative erection methodologies.

Method A Method B Method C

8/9/2019 Arup_Journal_2-2008

13/60


The immense scale of development slated for theMarina Bay area means considerable increases in thetrafc currently being serviced by these intersections.

The land title setbacks governing the placement ofthe building works allow for the duplication of Rafes

Avenue at some time in the future.Subsequent to these studies a new Formula One

night race event was introduced to Singapore.With the pit lanes sited immediately adjacent to theFlyer and with the project being ring-fenced by thetrack, the trafc environment will be drastically alteredat least once a year!

Comparisons with the London Eye

The London Eye was an architecturally-led projectformulated to mark the turning of the Millennium forthe city of London. The Singapore Flyer was acommercially-led development supported by theSingapore government, to inject investment into thecountry’s tourism economy. In both instances, theimportance of creating a world-class attraction of

exceptional quality and appearance was recognisedas essential to success.

Arup developed the London Eye design to tenderstage, when the design was for a 150m diameterwheel with 36 capsules. The design was takenforward by others at a slightly reduced size, leadingto the 135m, 32-capsule Eye that exists today.

30. Passenger boarding bridge.

Project awards

At a glance

outside of capsules)

building)

two soccer elds

station

Indicative weights

Other notable measurements and statistics

180 tonnes 100m long, 100mm diameter, 8 tonnes

Each main stay cable can carry over 6000tonnes of load. The main stay cables are sizedto limit wheel movements and much strongerthan they need to be to resist wind loads.

Each spoke cable is capable of carrying

diameter

Each of the 28 capsules is approximately

2

), and can carry up to30 passengers.

The structure is supported by 38 foundationpiles up to 1.5m in diameter, and bored up to

main strut columns, and ve under each ofthe four support cable stays.

Key facts about the Singapore Flyer

29.

8/9/2019 Arup_Journal_2-2008

14/60


The design of the London Eye was strongly inuenced by architectural requirements.From the outset, the architects envisaged it as being supported from one side only,and that the rim would be a triangular truss. There was a strong preference for limiting

the number of spokes and for them to connect to the central inner chord of the truss.While the advantages of a wide hub were recognised, the Eye hub width was

limited by the distance that even a very thick walled spindle could be made tocantilever. The idea of connecting cables to the edge of the rim to increase itstorsional stiffness was accepted, but they were limited to eight pairs, the minimumnumber that would effectively inhibit the four-lobed buckling mode.

The design of the Singapore Flyer was engineer-led. It was felt appropriate tosupport the spindle on both sides, which made it easier to achieve good supportstiffness, as well as allowing a much wider spindle to be used and consequentlyimproving the angle and efciency of the spoke cables. This increased efciency,together with a spoke arrangement developed to resist both lateral and radial forcesand provide torsional restraint to the rim, meant that the Flyer rim structure could bereduced to a bare minimum.

The two differing erection methods were both effective. The horizontal liftingapproach employed on the Eye made use of the River Thames as additionalconstruction site area, and was well suited to the one-sided support framing.

The vertical method adopted on the Flyer was ideally suited to the two-sided supportarrangement. It also minimised the plan area required on site for erection, allowing thesurrounding retail construction to proceed unhindered.

Conclusion

The Singapore Flyer is a private development investing in the Singapore tourismeconomy. Arup worked closely with developers in the rst instance and subsequentlyas part of the consultant/contractor team to add value where the rm was best placedto contribute.

The design was an engineering-led process that recognised the importance ofseveral geometric constraints on the structure’s efciency, and built upon knowledge

gained during the design of the London Eye . Differing site constraints from those ofthe Eye , as well as alignment with the development driver of reducing cost, resulted ina more efcient structure being developed. The two-dimensional truss form of theSingapore Flyer is both taller and lighter than the London Eye , and brings a newlightweight elegance to the design of GOWs.

As a testament to its innovative design, the Singapore Flyer was awarded theStructural Steel Design Award 2007 by the Singapore Structural Steel Society, for the“distinguished use of structural steel for its creativity, value and innovation”.

Andrew Allsop is a Director of Arup in the Advanced Technology and Research group in London, and thecompany’s leading wind engineering specialist. He wasthe wind engineer responsible for the wind tunnel tests

and other related wind aspects for the project.Pat Dallard is an Arup Fellow and a Director of Arup inthe Building London group. He specialises in advancedstructural design and analysis. The buckling designapproach that he originated was instrumental in thedesign of the London Eye and in this project, where hewas responsible for the original scheme designs.

Heng Kok Hui is a senior engineer in Arup’s Singaporeofce. He was the geotechnical engineer for thefoundation design for this project.

André Lovatt is a Principal of Arup and the ofce leaderfor Arup in Singapore. André provided the re safetyconsultancy for this project.

Brendon McNiven is a Principal of Arup, and was theProject Director for the Singapore Flyer project. Heleads the buildings team for Arup in Singapore. Brendonspecialises in architectural building structures and h isexpertise is in lightweight structures.

Credits

Client: Singapore Flyer Pte Ltd/Melchers ProjectManagement Design architect: Kisho Kurokawa

Architect of record: DP Architects SMEP and reengineer, and transportation planner: Arup - Andrew

Allsop, Easy Arisarwindha, John Brazier, Henry Chia,Mak Swee Chiang, Ho Chong Leong, Pat Dallard,

Andrea De Donno, Alex Edwards, Gary Goh, Jean Goh, Andrew Henry, Liew Kim Hoe, Heng Kok Hui, Lui VuiLee, Andre Lovatt, Peter MacDonald, Dexter Manalo,Brendon McNiven, Wong Siew Moh, Jane Nixon,Christopher Pynn, Sigrid Sanderson, Margaret Sie,Jonathan Sze, Jeffrey Willis Associate structuralengineer: MHI Associate building services engineer:

Alpha Engineering Landscape architect: ICN DesignInternational Contractor (building works): TakenakaCorporation Contractor (GOW): Mitsubishi HeavyIndustries Illustrations: 1, 2, 15, 29, 31 Singapore FlyerPte Ltd; 3, 11, 16, 17, 20, 26 Nigel Whale;4 DP Architects; 5 Kisho Kurokawa & Associates;6 Illinois Institute of Technology; 7 Edi Mitterlechner/ Dreamstime; 8 centerblog.net; 9 Mark Arkinstall;10, 12-14, 18, 21, 22, 24, 25, 27, 28, 30 Arup; 19 SoonWee Meng/Dreamstime; 23 Benglim/Dreamstime.

31. The East Coast, Singapore Straits vista from the top of the Flyer.

8/9/2019 Arup_Journal_2-2008

15/60

The Arup Journal 2

Introduction

For at least the near future, the intuition and know-how of experienced designers and builders willremain fundamental to successful building projects.However, much more can be done in the virtualworld both now and in the future to help designers,builders, and owners avoid some of the time-consuming and costly trial-and-error approachescurrently accepted within the industry.

The next decade will see the emergence andapplication of a holistic, technology-driven approachto the building process - a revolution in the making.

Thanks to the new virtual technologies, thepotential exists to rely more on hard facts ratherthan just design intuition. The concept of the “virtualbuilding” will eventually enable designers to develop

The VirtualBuildingPeter Bailey Daniel Brodkin

John Hainsworth Erin Morrow Andrew Sedgwick Martin Simpson Alvise Simondetti

a fully-tested building solution with condence not just in the building’s constrbut also in its long-term operational performance. The emerging virtual procesbecoming fundamental to design innovation, producing results that could not hbeen predicted before the advent of these technologies. This process will includsupplement current cutting-edge use of 3-D computer-aided design/drafting (Cand building information modelling (BIM).

What is the “virtual building”?

Answer: a concept in which all design, construction, environmental performanoperational problems are visualised, solved, and optimised using integrated comsimulation. The virtual building is intended to support stakeholders throughoutproject’s lifetime in the following areas:

Exploration: a constantly evolving tool for exploring new directions in designconstructionCommunication: enabling project teams to quickly and accurately communic

design forms, functions, and behaviours to other team members and the brocollection of stakeholders

Integration: providing an environment where design and facility team membshare and co-ordinate project information quickly and efciently

Optimisation: facilitating analysis tools that are capable of optimising performsustainability, and costs to meet both short-term and long-term goals.

Tools and techniques used in the virtual building are constantly evolving. This focuses on the possibilities for virtual design in the building industry now , what is nand cutting-edge, and what can be expected to come next that will change the wwe design buildings in future.

Emerging technology is moving us closer to thedream of the “virtual building”: a fully dened,integrated and operationally tested virtual prototypeof the nished building.

1. The structure of the Beijing Aquatics Centre (“Water Cube”): projects like this are now beyond conventional two-dimensional design and documentation methods.

8/9/2019 Arup_Journal_2-2008

16/60


Now2-D drafting vs 3-D modelling

Drawings in two dimensions are still the construction industry’s main form ofcontract documentation. They are also one of the main causes of conict, with poordocumentation estimated to cost billions of dollars each year. The problems with 2-Ddocumentation usually relate to poor co-ordination and poor detailing, due to the

limitations of designers in fully representing a physical object, ie a building describedonly in two dimensions on documents produced by separate disciplines.

3-D modelling, on the other hand, is the building block of the virtual building,offering signicant improvements over conventional drawing production (Fig 2). A 3-D model of a building created early in the process forces the designer/drafter tothink and resolve the proposed solutions in all three dimensions and in all parts ofthe building. In essence, 3-D modelling pulls the activity of co-ordination forward intothe process of design, creating a vehicle for true design integration. Once the spatialarrangement and detailing are resolved, then 2-D drawings can be extracted directlyfrom the 3-D model.

As the drawings are a “by-product” of the model, almost limitless permutationsof sections, plans, elevations, and isometric views can be produced in any direction.More importantly, as the drawings reect the model, they are fully co-ordinated withone another and will only present consistent information. Through 3-D representation, the building can be far more easily understood not onlyby the design disciplines, but by clients and builders as well. As a communicationtool, the 3-D modelling approach is thus far superior to 2-D and is already showingresults in producing better products with less rework. Once a basic 3-D model is setup, the possibilities of how this information can be developed, utilised, interrogated,and supplemented are endless.

New Virtual construction

As the density of systems increases, spacemanagement becomes increasingly important producing an efcient and well-integrated builBy combining 3-D models from the various de

consultants, the architectural and engineeringdesign can be co-ordinated by overlay andvisual comparison. This process can be aided clash detection software, but is most effectivelimplemented at virtual construction workshopBy producing a virtual model of building systecomponents, it is possible to effectively visuaand manage design co-ordination, thereby impcondence in the design and reducing the chanof late changes and clashes between buildingsystems on site.

This process is best enacted if all consultanthe same software. If this is impossible, data cexchanged using Industry Foundation Classesinteroperability standards1. Alternatively, softwsuch as NavisWorks 2 can be used to import andmodels from different software platforms and virtual design workshops. During the review pwe can rotate and zoom in on issues, isolate thredline, add appropriate comments, and then a

2. 3-D model of the Sydney Opera House.

8/9/2019 Arup_Journal_2-2008

17/60

The Arup Journal 2

actions, resulting in a Word document annotatedwith 3-D views from the model. Closer collaborativeworking practices should develop, using these tools.

One benet may be to avoid duplication of effort.For example, Arup is currently working with industryleading architects to integrate the structural andarchitectural models, leading to signicant time andcost savings for architects through not having tocontinually redigitise structural frame information.

During construction, subcontractors’ models canbe added to the process to provide further assuranceon t. In cases where subcontractors do not yethave 3-D modelling tools, information can be takenfrom their 2-D drawings and developed in 3-D by amodelling team. In this way, full 3-D co-ordinationby clash detection, or “virtual construction”, can becarried out before physical construction commences. This can be considered a virtual dress rehearsal forthe construction process, saving potentially costlyremedial works on site, and estimated to reduceconstruction costs by between 2-10%.

A combination of the architectural, MEP, façade,and structural designer and subcontractor modelswithin a single interactive, free-to-view model offersa very powerful design review tool. The ability tocombine 3-D models over one another in the virtualbuilding environment (Fig 3) may promote a “rightrst time” approach to the design, procurement, andconstruction process.

Common models

The next step beyond virtual construction is tointroduce a common model approach from theoutset of the project - this is where a 3-D model isshared centrally with all members of the design team. A shared central model requires agreed protocolsregarding who can alter what and how, and when itmay be updated. The model will need to be hostedon a central server located at the ofce either of theclient or any member of the design team, or by aspecialist modelling rm appointed to the project.

This process has been trialled on very few projectsaround the world. One example in which Arup wasinvolved is the One Island East project for SwireProperties in Hong Kong (Fig 4), which was entirelydesigned and procured using the Digital Projectplatform. The client bought hardware and softwarefor the entire team to use to ensure a consistent

approach. A central 3-D co-ordinator was appointedto oversee and supervise the central model allthrough the design and construction process. The client sees this as a way of rationalising hisapproach to all the projects in his portfolio, withbenets owing into how he manages his assets.

4. One Island East, Hong Kong, designed using a central model.

3. Princeton University Chemistry Laboratory: overlay of all engineering disciplines.

8/9/2019 Arup_Journal_2-2008

18/60


Construction scheduling (4-D)

Planning a construction process is notoriouslydifcult. Industry reports suggest that resourcesare only used at 40-60% efciency. 4-D modellinis a powerful new tool that provides an interactivability to visualise, inform, and rehearse construsequences, dr iving more efciency into the

construction process . “4-D” is an acronym that has developedinthe industry to represent the addition of the timedimension to a 3-D model. In simple terms, the model contains “ob jects” controlled and driven bGantt chart3 timeline. The application of the “foudimension” allows the sequence of ob jects to bemanipulated with almost limitless permutations. If we wish to amend the staging process, we amthe Gannt chart, not the“3-D images” (which arsimply a by-product of the process).

In the early stages of a time-critical pro ject itcan be useful to produce simple visualisation/ AVI presentations of the construction and sitemanagement sequenc ing. Sequential stills andmovies of the process can be produced to helpdisseminate the information clearly.

Later in the pro ject, as more deta iled programare required, the model can be used to describethe complex sequence of building without theneed to read and understand pages of charts. Tkey aim is to optimise overall construction timeby highlighting bottlenecks and site constraints istaging the works. Site management is ass istedillustrating the true scope of works and the stagnecessary to solve key constructability issues . Ita highly effective planning communication tool f

disseminating construction impacts to stakeholdor to overlapping and multiple subcontractors. This approach has already been used with

great success by Arup on many pro jects, includdemolition scheduling on the Leadenhall Streetpro ject in London (Fig 6), and ma jor works stagifor K ings Cross and St Pancras stat ions.

5-D scheduling

When we combine the automated extraction ofquantities over a timelined 4-D model we add adimension, commonly known as“5-D”. The pow5-D scheduling allow us to exploit the relationshbetween the ob jects’ timeline within the 4-Denvironment, and then report on their subsequenquantity or cost at particular points in time.

In simple terms, the consequence of taskoccurrences (or not), and their relationships toone another, allows us toinvestigate limitlesspermutations of quantum at any point in time. Sexamples of this would be to extract cubic metrconcrete to be poured in the following week ontdayworks schedule, or a $ value of work compin a monthly cost plan forecast. In a recent shop

Simpler versions of the central model, such as centralised database modelling, arealready being used. For example, the architect’s extruded shape geometry can befused with the engineer’s analytical centreline geometry with scripted links for softwareinteroperability, facilitating the comprehensive inclusion of design changes on a s ingleparametric platform.

In practice, the central model approach is not yet perfect and the pro ject teamcan expect numerous procedural problems. But though the approach may not savedes ign and documentation time, it can be expected to cons iderably reduce effortand save money during the site phase . In order to maximise the benets, centrallycontrolled models will require a transformation in the way pro ject teams work, with“master modellers” expected to assume control of all design information on pro jectsin the near future.

Bu ild ing Informat ion Modell ing (BIM)

BIMis a tool for adding information other than geometry to a 3-D model,its mainpurposes including:

Right now, BIMis proving useful (as stated by Autodesk ) “in providing continuous andimmediate availability of pro ject des ign scope, schedule, and cost information that ishigh quality, reliable, integrated, and fully co-ordinated”. The ability to attach this typeof information already exists within the common 3-D software packages, but we arestill developing an understanding of how to select and organise the data . BIM offersthe potential to verticallyintegrate the entire construction supply chain, as well ashorizontallyintegrate the design team (Fig 5).

Quantities and costs

It is already becoming common practice to extract the precise measurement ofmaterials or components from 3-D models we produce. All the geometric informationneeded has already been used to create the model, so it is a s imple extension to

extract that information in summary form once complete. The benet of this is thatthe manual take-off of quantities - often prone to human and scaling error - can beveried, or indeed may become superseded .

Once the quantities are extracted in a usable format,it becomes a s impleextension to add unit costs to the quantities measured to extract a representative costplan. One of the great benets of this is that rapid assessment and reassessment ofcosts is now possible once the 3-D modelis set up. Any changes to the model andits impact on cost can be qu ickly (and automatically) assessed.

FEASIBILITY DESIGN CONSTRUCTION OPERATION

Integrated documentation/virtual construction

Quantities/costs

Environmental/performance simulation

Optimisation/parametrics

Construction planning (4D/5-D)

Supply chain management

Asset managementBIM

5 . Virtual bu ild ing processes cover the full cycle of a bu ild ing’s l ife .

8/9/2019 Arup_Journal_2-2008

19/60

The Arup Journal 2

During the early stages of a project, designers teto use generic components to represent the buildsystems. Such components can be used to produaccurate tender information, but eventually willreplaced by specic components that the generacontractor and subcontractors intend to use forconstruction. The object-oriented nature of the vbuilding model means that components at varyilevels of detail can be easily inserted or exchanat any stage of the process.

The virtual building process thus enablesalternative layouts and building system strategibe modelled quickly and accurately, including clash detection and installation procedures. Thedigital model can also be linked to order informallowing components to be tracked from producto delivery, storage on site, and nal installation

Asset or fac i l i t i es management

The virtual building is not only useful during thdesign and construction process, but will soon ban effective tool for facility management througthe building’s lifetime. By linking components virtual building to a facility management databathe building manager could operate and run the using a visual interface. The virtual building dacan be designed to hold drawings, specicationand maintenance history for the components withe model. Hence an asset manager could simpl“click on a room” to nd relevant information f Alternatively, the manager could move directly the database to the location in the model to iden

centre project, moving the bars on the Gantt chart ripples over the 4-D model andonto the 5-D documentation, presenting the number, location, and availability of carpark spaces available at any point in time during the refurbishment. Such methods areideal for optioneering and assessing the client’s risk and nancial implications.

The clear downstream benets of 4-D and 5-D during the construction phase of aproject means that selection of design consultants with the requisite modelling skills isnow more important than ever.

D i rect manufacture

The virtual building process enables advanced manufacturing technologies which

extract fabrication data directly from 3-D models using computer numericallycontrolled (CNC) technology, eliminating the need and risk associated withinterpreting 2-D drawings.

Digital fabrication can be used for routine assemblies, but can also enable morecomplex shapes and assemblies that would not be possible using conventionalmethods. This technology is used extensively in the steel industry, but can be adaptedfor precast concrete construction as well. A recent example is “The Travellers”sculptures in Melbourne4 for the 2006 Commonwealth Games, where no drawingswere produced. All components were fabricated direct from the 3-D design modeland associated spreadsheets.

The potential to save money and time by eliminating the design drawing and/orworkshop drawing process is self-evident – a pointer to the potential for a “drawing-free” future, and a key step towards the “virtual building”.

Supply cha i n management Having guided a collaborative design and planning effort, the virtual buildingmodel can be manipulated and interrogated to further effect during construction.Interactive project review meetings with builders and subcontractors can be hosted,and discussions documented with views from the model. This promotes cross-tradeco-ordination through the trial construction, and helps maximise the benets of thecollective specialisms offered by the subcontractors. Interactive and free-to-viewmodels can be distributed to all, offering quick and effective project visualisation;this helps subcontractors immediately understand what is required of them andreduces much of the risk aspect of their pricing.

6. 122 Leadenhall Street, London, project: 4-D construction modelling.

8/9/2019 Arup_Journal_2-2008

20/60


a component in question, or the model could be setup to warn of faults or scheduled maintenance, ormonitor energy usage.

The process of reordering components orscheduling maintenance becomes greatly simplied,as the manager only need point to the element inquestion in the model for all relevant specicationsto be brought up from the database. This couldbe particularly powerful for façade elementswhere breakages are common and geometric andperformance data must be precisely adhered towhen reordering.

Parametric and generative modelling

Parametric modeling is a process using associativemodelling software which, according to BentleySystems, “captures and exploits the criticalrelationships between design intent and geometry”via scripts, algorithms and rules. By capturing thedening parameters of a building, ie geometricconstraints, environmental issues, or material

limitations, and their relationship to the buildingform, the design process can be automated anddesign iterations accelerated. Designers are thusempowered to explore limitless expressions in formthat are not arbitrary, but instead responsive to thecritical needs of the project.

The impact on building design is liberating.For example, current trends in architecture forcurving, non-orthogonal building forms are beingdriven by this new-found power in parametricmodelling. Parametric software facilitates the designand setting-out of complex non-orthogonal buildingforms in two respects. Firstly, it allows users togenerate the rst form, which is often too complex toderive using simple computer programs or scripts. Then, since the form is generated from a system ofrules applied to a few key variables, the shape canbe changed rapidly by adjusting the variables, andtested for efciency, aesthetics and performance.

Programming and scripting have, it is true,been used in various forms for many years,such as generating geometry and analysis models,or for specic uses such as venue sightline analysis.In the past, however, scripting was only accessibleto those with computer programming skills, but nowsimpler scripting languages, and more compatibilitybetween languages and new programs that use the

same scripting principals but present the user with agraphical user interface, have made parametric andgenerative modelling more accessible.

Proprietary parametric software includeDigitalProject by Gehry Technologies5 and BentleySystems’ Generative Components 6 (Fig 7).

7. A sculptural arts centre and a twisted building created using Generat i ve Components.

8/9/2019 Arup_Journal_2-2008

21/60

The Arup Journal 2

have condence in their performance. Small stehave already been taken towards assessing theacoustic performance of spaces dened by 3-Dmodels. Simplied models can now be extractefrom a detailed central model and tested and reas Arup has done in modelling the upgrade tothe Sydney Opera House Opera Theatre. Furthedevelopment is needed on the direct interrogatioof central models.

Similar testing levels are possible for smokemodelling as part of an overall performance-basre engineering approach. Smoke modeling canuse geometry directly from the design 3-D modproviding a more precise assessment of evacuattimes and smoke control performance (Fig 9).

As an example, the proposed roof of the new Olympic Park stadium in Melbournewas studied parametrically to nd the optimum shape, performance, and cost byvarying the height of the leading edge of the roof and thus causing an automaticupdate of the key geometry of the rest of the roof. Structural and façade elementvariation could thus be studied to nd the optimum set from a cost and visual pointof view (Fig 8).

It is not difcult to imagine how multiple variations of buildings could be designedfrom standard components. A predened façade suite could be programmed to

populate the building face automatically, knowing its geometric and environmentallimitations, as the geometry changes. Other components could also respond to theirinputs. The designer would then select the preferred combination depending onclient, site, environmental requirements, and individual preference. This has enormouspossibilities in reproducible or adaptable buildings such as schools and apartmentbuildings, especially when combined with direct manufacture.

Environmental performance modelling

The principles of virtual building lend themselves to exploring project improvementsthrough quick assessment and comparison of alternative environmental performanceoptions. Pioneering methods are emerging that will assist in planning optimal space,material and energy utilisation, allowing teams to assess the optimum sustainabledesign outcome. These design options can be maintained throughout the designperiod, with the rapid ability to schedule, analyse, and compare options concurrentlyas they develop. For instance, a 3-D model now offers a central database from whichcompliance reports for environmental rating systems such as LEED7 in the US andGreen Star8 in Australia can be automatically created.

Sustainable design assessments can focus at a micro-level - for instance,embodied energy in the concrete - or at a macro-level, to determine, for example,urban amenity, over-shadowing, or street acoustics in whole precincts. In either case,changes and improvements can be readily interpreted using visual and aural models.

There will be no more important development in this regard than the integrationof thermal/energy, air quality, and daylight modeling into a central virtual buildingmodel. Using these tools we can hope to achieve more sustainable buildings and

9. Smoke modeling in the Sydney Opera House model.

8. Parametric modelling of Melbourne rectangular pitch stadium roof, including roof panels and structural forms.

8/9/2019 Arup_Journal_2-2008

22/60


10. City model of Ancoats Village, Manchester.

The optimisation routines used will depend onproblem to be solved. Routines are often set tooptimise a single parameter (eg steel tonnage)but it is now more common to try to optimisemultiple or competing parameters.

In these cases, one process is based on “antcolony” optimisation. Ants nd the optimum rthrough unknown terrain by emitting pheromosimilarly, sets of solutions are developed thatbest meet the design team’s objectives. Once acomputational solution set has been built, altedesigns can be explored by varying the param

Design parameters can be incorporated intocomplex algorithms that will nd the best set solutions to meet the objectives set by the desiteam. Once a computational solution set has bbuilt, alternate designs can be explored by varthe parameters.

This approach has been widely used in theaerospace and automotive industries, and is onnow beginning to take hold in the building ind

Optimisation’s appeal for architects is that is pan objective basis for design, but is in no way replacement for design itself.

The design team and client must control thesubjective process of selecting and weighing tparameters. The strength of this approach is thproject solutions can be assessed without anypresupposition about form, and condence incof nding the best solution.

City modelling

Whole cities can now be modelled to demonstrate client and community-widebenets - a “virtual city” of virtual buildings. The existing city is modelled by gatheringgeographic spatial information, either from existing information or aerial or terrestrialsampling, and storing it in a manageable format. The virtual building model forthe new development is then inserted into the city model (Fig 10), where it can beaccessed for such uses as integrating and assessing new developmentsfor planning purposes, accessibility assessments, and visual and otherenvironmental assessments.

NextReal-time analysis

Currently, design is a time-consuming iterative process whereby design teams meet,conceive options, and then go away to investigate and test those options. A weekor two later the team meets again and the process repeats. Tools are now beingdeveloped to enable design to be optimised quickly in “real time” in the design studiowith the whole design team. Computational uid dynamics (CFD) is used to assessthe environmental performance of a space, but to date has been very time-consumingto set up and run, often taking days or weeks. But computer power and memory aredeveloping rapidly, and hence the ability to run these routines on the spot and helpthe design team work through options more rapidly.

Optimisation

This process uses computational routines to assess and sort options to nd anoptimal set of solutions, providing a support to design intuition rather than replacingit. Any number of parameters in a design can be varied, including for example, views,daylight levels, thermal efciency, and costs (Fig 11).

8/9/2019 Arup_Journal_2-2008

23/60

8/9/2019 Arup_Journal_2-2008

24/60


It is now possible to provide an accurate aural footprint of a space using acousticsimulation rooms such as Arup’sSoundLab . InSoundLab , the acoustic performanceof a space can be demonstrated at any position inside the space using surroundspeakers, with visual clues provided by a 3-D model on a screen. It is thus possible todemonstrate the view and sound at any given seat in specic performance spaces.

Engaging the visual senses is also being explored using 3-D projections or virtualreality goggles, which provide some ability to immerse yourself within a spacemodelled in 3-D. There are shortcomings, however, as current screen and projectiontechnologies are unable to closely replicate the visual bandwidth perceived by thehuman eye, and hence form a barrier to true “reality”, particularly when in varyingshades of light and dark. These tools are still under development and far frommainstream. As for air temperature and movement, attempts have been made toprovide a visual representation so that we can see how a space is behaving.CFD is the current tool; experiments to present the results in 3-D have not so farproven successful.

The goal is a room that can simulate the appearance, sound, air movement,and temperature performance of a space, providing a true immersive experience. This might be formed by creating a box in which the building model is projectedonto the inside walls to simulate standing or walking in the room in question,while surround speakers, fans, heaters, and air-conditioners simulate the plannedenvironmental conditions direct from the virtual model (Fig 13).

Populating virtual buildings

Software now exists that permits the virtual building space to be inhabited byagents, preprogrammed with human behavioural patterns to see how they will reactto different physical environments. One such program is Arup’sMassMotion , aninternally-funded research and development initiative that staff in the rm’s Torontoand New York ofces developed in response to the needs of the Fulton Street Transit Center (FSTC) project in New York City. Since then, further development hastaken place in Toronto with technical input from staff in the New York, Melbourne,Westborough, and San Francisco ofces.

Developed relatively economically in comparison with other comparable programs,MassMotion is a completely new suite of tools, though the developers leveragedcommercially available 3-D software fromSoftimage to streamline development andrapidly build out functionality.MassMotion is also very cost-effective.

MassMotion produces highly instructive animations of pedestrian ow, and itshould be stressed that these are not merely animations, but the results of analysingthe cumulative effect of the decisions of the individual agents. In addition to theanimations,MassMotion produces ow and occupancy counts, queue sizes, anddensity maps; all of which inform the design.

The process involves the creation or adaptation of a 3-D model with all theprimary physical and spatial features that one would nd in the nal built form. Then the agents can be programmed to behave in ways that mimic human behaviour,for instance pausing at a café for a cup of coffee or stopping at a travel informationboard, passing through a turnstile or going up an escalator, based upon destinationpreferences. The FSTC model agents were given attributes from the eld surveys,ie male/female ratios (as women on average walk at a slightly shorter step and pace),and whether they were commuters (know where they are going) or tourists (not sure

where they are going). The agents are then left free to populate the model, enabling the users to

observe and assess how the space performs. The result is the potential for a realisticassessment, as true pedestrian systems are more random and chaotic than previousmodeling tools allowed. The performance of the space can then be assessed againstlevel of service metrics and to identify bottlenecks, as well as egress assessment. Trafc simulation can also provide further opportunities.

13. Immersion in a virtual reality room modelling sight, soundand comfort.

The breakthrough with this technology is that up endless possibilities for testing any sort of interaction. For example, the likely success oflayouts could be proven.

Since its application for FSTC,MassMotion hbeen developed further. It can now simulate a range of pedestrian activities including emergevacuation, navigation by familiarity or by sigbehaviour in access-controlled areas such as fagates, and dynamic response to scheduled eve

A wide range of project types, including trastations, bus stations, and airports, as well as sand ofce towers, have now been designed wihelp ofMassMotion .

Conclusion

Full virtual prototyping of buildings is no longdream for the distant future. Powerful tools arimplemented in the virtual building environmeallow us to partially simulate the performancebuilding before it is constructed. As the technodevelops, the potential exists for the creation ocomplete virtual building in which all its aspeinternal relationships can be tested and undersin an automated fashion.

The challenge for the property and construcindustries today is to embrace and accept the3-D-enabled technology now on offer, to proda more streamlined, right-rst-time approach tbuilding design, construction, and operation.

Forward-thinking clients already expect 3-D

design. As technology advances these are the who will expect the model’s object content to packed with all conceivable aspects of data to them nancial or operational certainty. The resvirtual building models will open far-reachingopportunities within the future management anbusiness operations related to the building indand Arup will contribute a key role in this pro

8/9/2019 Arup_Journal_2-2008

25/60

The Arup Journal 2

Peter Bailey is a Principal of Arup, Buildings sector leader in the Sydney ofce, and a memberof the rm’s global Building Sector Board.Daniel Brodkin is a Principal of Arup and the Edison, New Jersey, ofce leader. He is also theBIM leader for the Americas Region.John Hainsworth is an Associate of Arup in the Sydney ofce and BIM leader for the

Australasian region.Erin Morrow is a senior consultant in computer modeling with Arup in the San Francisco ofce.He led the development of the MassMotion simulation program.

Andrew Sedgwick is a Director of Arup in the Buildings London 4 group, design and technicalleader in the Buildings sector, and global leader of the arts and culture business.Martin Simpson is an Associate Director of Arup in the Manchester ofce.

Alvise Simondetti is an Associate of Arup in the Foresight, Innovation and Incubation team,London. He is also the global manager of Arup’s virtual design skills network.

Credits

Nat i onal Aquat i c Centre, Be i j i ngClient: Beijing State-owned Assets Management Co L

Architect: PTW (Australia) & CSCEC & DesignOne Island East, Hong Kong Client: Swire Properties Ltd

Architect: Wong & Ouyang (Hong Kong) Ltd122 Leadenhall Street, London Client: British Land Co plc

Architect: Richard Rogers Partnership LtdMelbourne Olymp i c Park rectangular p i tch stad i umClient: Melbourne & Olympic Park Trust

Architect: Cox Architects & PlannersSydney Opera House Opera Theatre refurb i shmentClient: Sydney Opera House Trust

Architect: Utzon Architects/Johnson Pilton WalkerPr i nceton Un i vers i ty Chem i stry Laboratory Client: Princeton UniversityDesign architect: Hopkins Architects LtdExecutive architect: Payette Associates IncMar i na Bay Sands Integrated Resort, S i ngaporeClient: Marina Bay Sands Pte LtdDesign: Architect: Moshe Safdewith Aedas Al Raha t ower, Abu Dhab i

Client: Aldar Properties Pjsc Architect: Asymptote ArchitectureFulton Street Trans i t Center, New YorkClient: Metropolitan Transit Authority CapitalConstruction New York

Architect: Grimshaw ArchitectsIllustrations: 1 Ben McMillan; 2 Stuart Bull; 3 VincenFiorenza; 4 Swire Properties; 5 Nigel Whale; 6 SimonKerr, 7 Matt Clark, John Legge-Wilkinson and Stuart (© Arup + Marina Bay Sands Pte Ltd); 8 John Legge-Wilkinson 9, 10 Simon Mabey; 11 Gianni Botsford

Architects; 12 Alvise Simondetti; 13 Tristan Simmond14 Robert Stava.

References

(1) http://cic.nist.gov/vrml/cis2.html#IFC(2) http://www.navisworks.com(3) http://en.wikipedia.org/wiki/Gantt_chart(4) BAHORIC, J,et al . “The Travellers”.The Arup Journal, 42pp32-37, 1/2007.(5) http://www.gehrytechnologies.com(6) http://tinyurl.com/4ag6am(7) http://www.usgbc.org(8) http://www.gbca.org.au

Acknowledgements

The authors would like to acknowledge the key contributithis article from the following Arup experts:

Americas: Matt Clark, Anthony Cortez, David Farnsworth Vincent Fiorenza, Ken Goldup, Zak Kostura, Murat KuraRobert Stava, Ben Urick

Australasia: Peter Bowtell, Stuart Bull, John Legge-WilkiChris PynnEast Asia: Maverick Chan, Kelvin LamEurope: Francesco Ànselmo, Gavin Davies, Alexej Goehr

Anne-Marie Gribnau, Alejandro Gutierrez, Peter Head, AJenkins, Scott Kerr, Vahndi Minah, Aston Wisdom, BrauliMorera, Tristan Simmonds, Steve Walker, Neill Woodger, Woolf, Russell Yell.

14. Fulton Street Transit Center, New York: MassMot i on modelling.

8/9/2019 Arup_Journal_2-2008

26/60


This study acknowledges changes in tools andfabrication processes for the built environment.

Also, and in contrast to most designers in industryand academia alike, it considers these changingprocesses to be fundamental to design innovation.

Based on observation of the current position,the paper proposes, as a way forward, that acommon vision should be shared by practice,industry, and academia - as one way to accelerate amuch-needed transformation of design practice.

Designer’s Toolkit 2020 adopts the frameworkproposed in the USA National Research Council(NRC) study “Beyond productivity: IT and thecreative practice” 1, with four levels of research anddevelopment investment risk and return (Fig 2):

predicted.

Unusually for a review in this eld, all four levels aretaken into consideration.

Introduction

Most research that focuses on exploring the ever-shifting design requirementsinstigated by dramatic changes in society remains based on the assumption ofunchanging tools and fabrication processes. The present study, by contrast, focuseson changing design tools, on making tools, and on the effect of this on design.

In 2006, the author conducted a review study, Designer’s Toolkit 2020 , to explorethe drivers for, and what might plausibly be, the designer’s desktop scenario around15 years in the future. He interviewed 22 thought leaders* - PhD candidates toindustry board members - from across the design world, with contributions fromdesigners outside the built environment professions. Where possible, the interviewswere conducted face-to-face; if not, by via video-conference or telephone.

Designer’s Toolkit 2020 : A vision for the design practice

Alvise Simondetti

* Professor Mark Burry, RMIT, Melbourne; Reed Kram, Kram Design, Stockholm; Charles Walker, Zaha Hadid Architects; Jeffrey Yim, Swire Properties, Hong Kong; Axel Kilian, MIT,Boston; Jose Pinto Duarte, Technical University, Lisbon; Joe Burns, Thornton Tomasetti, Chicago; Mark Sich, Ford Motor Company, Michigan; Phil Bernstein, Autodesk, Boston;Lars Hesselgren, KPF, London; Bernard Franken, Franken Architekten, Frankfurt; Martin Fischer, Stanford University, San Francisco; Dr. Kristina Shea, Technical University, Munich;Prof. Chuck Eastman, Georgia Tech, Atlanta; Professor Donald E. Grierson, University of Waterloo, Ontario, Canada; Mikkel Kragh, Mike Glover, Duncan Wilkinson, Arup, London;Colin Stewart, Arup, Birmingham; Peter Bowtell, Arup, Melbourne; Tristram Carfrae, Arup, Sydney.

The designer’s toolkit is rapidlychanging. Design practices needa shared vision for the short,medium, and long terms.

Designer’s Toolkit 2020 focused on design researchprojects and individuals working within a project-based research methodology. As explained byMartin Fischer 2 among others, this contrasts withlaboratory-based research methodologies.Project-based research methods involve identifyinga non-trivial challenge in a specic practical context,and solving that specic challenge within theproject’s deadline.

Researchers often use bespoke tools andprotocols, and in this their methods are not differentfrom standard project practice. However, there arefurther steps: revisiting the challenge; focusing onwhat is novel in the solution; generalising it fromthe specic project; rigorously testing the solution’svalidity; confronting the ndings within the researchcommunity; and nally contributing to knowledge

through publication of the results. This project-based methodology inherently guarantees thepractical signicance of the solution, something oftenquestioned in design research.

1.

8/9/2019 Arup_Journal_2-2008

27/60


The following expands on the four big ideas. All quotations in italics have beenselected from the personal interviews.

(1) Process transfer, not technology transfer

Transferring technologies from other ofces and/or indust ries has provided greatbenet, allowing the design and construction of projects that couldn’t otherwise havebeen built 4. However, those working with new technologies, including parametricrelational modelling and building information modelling (BIM) point out the limitationsof this approach 5 and the necessity for a whole new one.

“Our edge comes from us and the way we think, not just our tools.” Transferring new technologies is insufcient if one doesn’t also expend the

energy needed to understand their methods and how to use them. Methods, unliketools, always need to be understood and adapted to our industry, and cannot bedirectly translated. The same tool might be used in a dramatically different way whentransferred from, say, the automotive industry to architecture - as in the case of rapidprototyping, originally developed to produce prototypes overnight and speed up thedesign development but, when adopted in architecture, used to produce uniquedesigns accurately.

“Our children in their bedroom are using more sophisticated technology to make decisions within games than we’re using in the planning environment.”

In computer games, users make decisions based on quantitative and real-

time feedback from their actions. Process transfer is the ability to learn from otherofces and/or industries how they go about producing their designs and makingdecisions, how they think with their tools, what their protocols of interaction are, whothey interact with, and who has control. For example, Toyota’s lean manufacturingmethods are based on accurate real-time information travelling up and down thesupply chain.

“We used to have computer programmers and designers, now we havedesigners who can program. The ability to program what you want, when you want

it, has already brought larger gains for the project, for the client, and our challenge is to turn them into designer’s gains.”

Traditionally, tools and methods were selected by a master designer, based onyears of experience. However, tools and methods have now become disjointed, withdigital tools selected by apprentices and applied to the master designer’s traditionalmethods. Methods must also be selected with new tools in mind.

“New graduates have no fear of programming, no use for primitives.” It is a challenge for those who haven’t learned how to write a computer program,

even a simple one, to understand the power, and the risk, and the limitations of thework of their junior staff. How can the design director sign off the latest deliverablein the shape of a BIM automatically generated by a script where no two-dimensionalsection is similar to any other?

“When model managers are third parties, they take control. Project management is the ideal place to nd the lateral thinking and specicunderstanding necessary to be a custodian, or master modeller, or model manager.This might be a temporary role.”

Traditionally, the architect took control of the design. In other industries, however,model managers hold all the design information and have eroded that kind of control.

The master modeller’s role includes acting as the gatekeeper who gives information

privileges, makes sense of information coming in, and knows what informationgoes out to different teams at the time they need it. Possibly “just-in-time” designinformation could bring similar quakes to design and construction as “just-in-time”manufacturing did to its industry. In a similar way to manufacturers seeing warehousesfull of components disappear, designers might experience servers full of unusable andredundant design data disappearing too.

Designers, however, must ensure that, in the long term, control will return tothem when interoperability, access control, and versioning - the current challengesin the industry - are overcome. The nancial industry has automated access controlmethods already.

Findings

Four “big ideas” emerged from the interviews:(1) Transferring technologies from other industries

has provided great benets, but has generated theneed to transfer processes as well - the processesby which other industries produce their designs andmake decisions.

(2) Despite most of industry’s and academia’sfocus on development of the designer’s toolkit toincrease efciency, the main dr ivers for change arethe new ways of making . Naturally the toolkit hasdeveloped faster and further in supporting changesat the end of the construction supply chain; however,

tools for the early stages of design are creatinggreater gains for designers.

(3) The gains from the interaction and interplayof discipline-specic algorithms are greater than fromincreasing the individual sophistication of single-discipline algorithms.

(4) Designers are getting used to “just-in-time”information being available anywhere - fast, recentand relevant - and are now expecting this to applyalso to design information.

2. The NRC’s four levels of risk against return.

3. Building theory in practice, as visualised by the Center f orIntegrated Facility Engineering (CIFE) 3.

8/9/2019 Arup_Journal_2-2008

28/60


Steelwork fabrication quickly adopted component-based modelling to improve its processes.

This in turn is now rapidly t ransforming the designer’stoolkit from lines, points, and layers (inherited fromthe designer’s hand drawings that were developedto communicate with 19th century craftsman) tocomponents and assemblies.

Virtual prototyping of the build environment,ie BIM8 or BEM (built environment modelling) 10 ,is reducing construction risk and waste. In thepast, designers kept separate from construction- a business with a different risk prole. However,reducing the risk has seen the proliferation of “garagecontractors” who thrive on their green credentialsbecause of the reduced waste and reliable delivery.

“There will be something like a pre-emptive modelling of the building process that will knowexactly what’s going to happen with the building.Today, if you go to have your appendix out, youdon’t hope you’re going to come out alive; it’s a

near mathematical certainty that today you’ll survive

an appendix operation.” Conversely, the current limitation of virtual

prototyping is that it is unregulated. (This is tobe expected and is common to all new formsof representation.) Practitioners are left with thechallenge of selecting the appropriate level of detailand, most importantly, of communicating it to theteam so that everyone knows what the prototyperepresents and what it doesn’t.

“We should enhance the front end of the design process that’s going on in all design ofces. I think many design ofces miss out on a major possibilityof increased productivity or an improved design -the decisions made in the initial design stage affect80% of what happens thereafter.”

Designers should focus on developing tools thatwill support the conceptual stage of the designprocess, this rst stage being arguably the mostdifcult of all. It is highly unstructured, and has noreal algorithmic bases, at least not ones that can bereadily perceived.

(3) Develop algorithms for integration,not specialised knowledge

A cycle seems to be happening: we have had 20years of developing algorithms, including niteelement analysis modelling, that have made explicit

our industry’s specialised knowledge and greatlyenhanced the development of performance-baseddesign in engineering. However, it was pointed outthat few academic papers in this area have beensubmitted in recent years. The current research focusis in enabling integration. Similarly in design practice,larger gains seem to accrue from optimising howdisciplines interact than from how they do theirtasks individually.

“We will see a proliferation of experts, as the rst rule of modelling, ’junk in, junkout’, is still valid.”

Master modellers aren’t the only emerging specialists. Construction industrydesigners might take notice of the role of the mathematical modeller in the automotivedesign industry. The electronic math modeller, also referred to as the “digital

sculptor”, is the individual who takes a free form and then matches a mathematicallyrepresentable form to it to create the computational representation.

“I’m now involved with people in economics, in applied mathematics, who have nothing to do with engineering, but who have little expertises that I don’t have.”

Computation is shifting the boundaries between disciplines, with the result thatmodels from other disciplines are becoming of interest to designers. This is not new.What is new, however, is that these are explicit computational models that require setprocedures to translate.

(2) Design for new ways of making , not for design efciency

One of the greatest changes occurring in our industry is in how we make (or build)things, specically our increasing ability to produce unique and complex mass-customised designs 6, 7 at the same or even improved speed, cost, and quality asrepetitive and simple mass-produced ones.

“We focus on novel design, not only measurable improvements.” Traditionally, designers have tailored their abstract representations (scaled

plans, sections, and elevations) to communicate their ideas and solutions tovarious audiences including, crucially, fabricators and contractors. Now that designinformation feeds automatically into computer numerically controlled (CNC) machinery,novel representations are needed in the form of spreadsheets of machine commandsor databases - assembly instructions, as well as interactive visualisations, that helpconvince the fabricator that the script as well as the machine is doing the right thing.

Traditional representations of plan, section, and elevation are becoming redundantfor the fabricator and the contractor. This could have profound implications fordesigners who have used these representations as “tools to think with”.

“Plan, section, and elevation will disappear as we know them today; however 2-D schemes will grow.”

There will be implications for other disciplines that have used designers’ drawingsto, for example, extract quantities, provide planning advice, bring evidence in court,and calculate fees. It is possible that these disciplines may adapt to the novelrepresentations now used to communicate between designer and fabricator. In oneexample, a court used an accurate representation of the 3-D design geometry tosupport the case of a fatal accident on a building site. In another, a High Speed 1(Channel Tunnel Rail Link) contractor used earthworks machinery driven by on-boarddigital terrain models - which in turn is helping transform the rail design industry fromvector to meshed representation.

4. How the master modeller role might evolve.

8/9/2019 Arup_Journal_2-2008

29/60


“A holistic approach to sustainability drives multi-physics simulation? Absolutely, and with that will come a legal framework that will force you to do it. It’s happening already in projects in Switzerland, also in Singapore and Finland.”

“It’s a multi-phase analysis; you need to do it at the conceptual design stage andat various stages all the way through. How to develop good evaluation technologiesand requirements at each of these phases? It’s a challenge to do that well and to beable to cross-link across phases.”

Horizontal bidirectional links between the analysis and geometrical models, alsoreferred to as “round-tripping”, enables faster design cycles 11 and allows for manualdesign optimisation. In some projects, including stadium design, the geometry that isnally built might be as much as the 27th design version. Bidirectional links betweenanalysis and design also allow for computational design optimisation (CDO) 12 .For example, in the design of spaceframes for long-span steel roofs, CDO is beingused to reduce steel member sizes.

“The survivor will be the one who understands the need to connect.” The ultimate goal would be to take advantage of the interaction or interplay

between discrete analysis as it occurs, for example in re/structural analysis 13 . Theintegration of the different discrete sub-models allows the designer to identify areas ofoverlap and interaction, and feedback loops.

“The next drivers are going to be [from] biology and I think it is biological modelling that is going to drive the next 10 years.”

Good design is holistic, and the science of biology has developed tools andmethods to understand

arup_journal_2-2008

Documents