Steel Details
Publication No. 41/05
Apart from any fair dealing for the purposes of research or private study or criticism or review, as permitted under the Copyright Design and Patents Act 1988, this publication may not be reproduced, stored or transmitted in any form by any means without the prior permission of the publishers or in the case of reprographic reproduction only in accordance with the terms of the licences issued by the UK Copyright Licensing Agency, or in accordance with the terms of licences issued by the appropriate Reproduction Rights Organisation outside the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publishers, The British Constructional Steelwork Association Ltd at the address given below. Although care has been taken to ensure, to the best of our knowledge, that all data and information contained herein are accurate to the extent that they relate to either matters of fact or accepted practice or matters of opinion at the time of publication, The British Constructional Steelwork Association Ltd, the authors and the co-ordinator assume no responsibility for any errors in or misinterpretation of such data and/or information or any loss or damage arising from or related to their use.
The British Constructional Steelwork Association Limited (BCSA) is the national organisation for the steel construction industry: its Member companies undertake the design, fabrication and erection of steelwork for all forms of construction in building and civil engineering. Associate Members are those principal companies involved in the purchase, design or supply of components, materials, services related to the industry. Corporate Members are clients, professional offices, educational establishments which support the development of national specifications, quality, fabrication and erection techniques, overall industry efficiency and good practice. The principal objectives of the Association are to promote the use of structural steelwork; to assist specifiers and clients; to ensure that the capabilities and activities of the industry are widely understood and to provide members with professional services in technical, commercial, contractual and quality assurance matters. The Association’s aim is to influence the trading environment in which member companies operate in order to improve their profitability. A current list of members and a list of current publications and further membership details can be obtained from: The British Constructional Steelwork Association Ltd 4 Whitehall Court, Westminster, London SW1A 2ES Telephone: +44 (0) 20 7839 8566 Fax: +44 (0) 20 7976 1634 Email:
[email protected] BCSA’s website, www.SteelConstruction.org, can be used both to find information about steel construction companies and suppliers, and also to search for advice and information about steel construction related topics, view publications, etc.
Publication Number First Edition ISBN
41/05 November 2005 0 85073 049 X
British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library © The British Constructional Steelwork Association Ltd Designed and Printed by Box of Tricks www.bot.uk.com
FOREWORD Whether it is true that some deity is in the details is uncertain. What is true is that steelwork costs are as much determined by connection details as they are by raw material costs or site methods and conditions. Steelwork Contractors will confirm that their businesses depend on economic detailing and for this reason it is important that the BCSA should ensure that sound advice is available concerning steel details. In terms of engineering connections, the "Green Books" promoted by BCSA and SCI have proved a valuable source of standard details. But there is more to detailing than is covered by the Green Books. This publication is intended to develop an understanding of the wider issues in two ways. Firstly, it provides a forum for steelwork experts to explain their views on what engineering issues affect efficient detailing. Secondly, it takes architectural details from actual structures and allows both engineers and architects to interrogate these. Whilst actual details evidently differ from standard ones, the fundamentals of good detailing advocated by our authors still apply. It is common knowledge that during the last 25 years, there has been a marked improvement in steel’s UK market share for both buildings and bridges. However, what is less widely appreciated is that the nature of steel construction has changed. During that period the mix has moved from being predominantly industrial to being predominantly commercial. Factories with lattice cantilever columns supporting heavy overhead travelling cranes have been replaced by offices with atria featuring complex glass-supporting space structures. At the same time as steel has been developing the shed structures that are now the common feature of our industrial landscape, it has been improving its competitive position for commercial and residential multi-storey frames, and most notably has established itself as the modern material without equal for highly-visible prestige structures. Perhaps it is no coincidence that during the same period many of the UK’s leading architects have established their credentials as household names on the world stage. We trust that the mix of architectural details that we have selected to feature in this publication is a fitting testimony to the pioneering work of those leaders and the structural engineers who have supported them. The contributory authors are experts in their particular fields and sincere thanks are given to them. Our gratitude is also due to Susan Dawson who drafted and compiled the architectural details and to the co-ordinator of the book, Dr Roger Pope.
Dr Derek Tordoff, Director General The British Constructional Steelwork Association Limited.
CONTENTS ENGINEERING CONNECTIONS Chapter 1
Structural details are engineering - John Rushton Introduction ............................................................................................................................ 1 Design-analysis-detail design: a virtuous circle .................................................................... 1 Implementation of the design on site .................................................................................... 3 Challenges and opportunities ................................................................................................ 4
Chapter 2
Connection design in relation to frame design - Alastair Hughes Introduction ............................................................................................................................ 5 ‘Simple’ versus moment-resisting frames ............................................................................. 5 Methods of global analysis .................................................................................................... 5 Simple construction ............................................................................................................... 6 Continuous Elastic design ..................................................................................................... 6 Continuous Plastic design ..................................................................................................... 6 Semi-continuous Plastic design ............................................................................................ 7 Semi-continuous Elastic design ............................................................................................ 7 Composite construction......................................................................................................... 8 Tubular construction .............................................................................................................. 8 The Green Books and the Connections Group ..................................................................... 8 Conclusion ............................................................................................................................. 9
Chapter 3
Connection design programmes - Alan Rathbone Introduction .......................................................................................................................... 10 Underlying design models ................................................................................................... 10 Range of software ................................................................................................................ 12 Some particular issues ........................................................................................................ 13 Advances foreseen .............................................................................................................. 15
Chapter 4
From laboratory to laptop - David Nethercot Getting the knowledge base ................................................................................................ 16 Use of numerical analysis .................................................................................................... 16 Future needs ........................................................................................................................ 16
Chapter 5
Connection rotational characteristics - Graham Couchman Introduction .......................................................................................................................... 18 Connection types ................................................................................................................. 18 Conclusion ........................................................................................................................... 18
Chapter 6
Simple connections and basic fabrication - Dave Chapman Introduction .......................................................................................................................... 19 Basic fabrication machinery ................................................................................................ 19 Column bases and HD bolts ................................................................................................ 21 Beam connections ............................................................................................................... 21 Bracing connections ............................................................................................................ 22 Column splices .................................................................................................................... 22 Cantilevers ........................................................................................................................... 23 Truss details ......................................................................................................................... 23 Weld details .......................................................................................................................... 24 Fabrication details ................................................................................................................ 24 Erection details .................................................................................................................... 24 Conclusion ........................................................................................................................... 25
Chapter 7
Connection costs - Kim Dando Introduction .......................................................................................................................... 26 Keep it simple! ..................................................................................................................... 26 Connection groups .............................................................................................................. 26 Comparative costs ............................................................................................................... 27 General details ..................................................................................................................... 31 Conclusion ........................................................................................................................... 32
Chapter 8
Structural fasteners - Thomas Cosgrove Introduction .......................................................................................................................... 33 Background .......................................................................................................................... 33 Current UK practice ............................................................................................................. 33 Existing European preloaded bolts ...................................................................................... 34 Preloaded bolts to EN 14399............................................................................................... 34 Suitability test....................................................................................................................... 35 Installation ............................................................................................................................ 35 Non-preloaded bolts to EN 15048 ....................................................................................... 36 CE Marking .......................................................................................................................... 36 Conclusion ........................................................................................................................... 37
Chapter 9
Cladding connections - Richard B Barrett Introduction .......................................................................................................................... 38 Loading and design issues .................................................................................................. 38 Air leakage and robust details ............................................................................................. 39 Thermal bridges ................................................................................................................... 40 Cladding systems ................................................................................................................ 40 Roofs ....
41
Wall claddings ...................................................................................................................... 41 Gutter and downpipe supports............................................................................................ 41 Construction issues ............................................................................................................. 42 Health & Safety .................................................................................................................... 43 Conclusion ........................................................................................................................... 43
Chapter 10
Acoustic details - Andrew Way Background on acoustics and sound .................................................................................. 44 Acoustic regulatory requirements ........................................................................................ 45 Principles of acoustic detailing ............................................................................................ 46 Floor and ceiling treatments ................................................................................................ 49 Integration of columns and services .................................................................................... 49 Conclusion ........................................................................................................................... 49
Chapter 11
Composite connections - Mike Banfi Introduction .......................................................................................................................... 51 Principles.............................................................................................................................. 51 Limitations for plastic design ............................................................................................... 51 Design considerations ......................................................................................................... 52 Serviceability ........................................................................................................................ 53 Conclusion ........................................................................................................................... 53
Chapter 12
Joints under fire conditions - Ian Burgess, Roger Plank Introduction .......................................................................................................................... 54 Response of steel composite frames to fire ........................................................................ 54 Design considerations for joints .......................................................................................... 54 Behaviour of joints in fire ..................................................................................................... 55 Connection temperatures .................................................................................................... 55 Observed behaviour in Cardington fire tests ....................................................................... 55 Observed behaviour at World Trade Center ........................................................................ 56 Suggested variants of simple joints ..................................................................................... 57 Eurocode requirements ........................................................................................................ 57 Moment-rotation at high temperatures ................................................................................ 57 Component-based approach............................................................................................... 58 Conclusion ........................................................................................................................... 58
Chapter 13
Vibrations - Stephen Hicks Introduction .......................................................................................................................... 59 Acceptability of floors for walking vibrations ....................................................................... 59 Simplified approach ............................................................................................................. 61 Conclusion ........................................................................................................................... 64
Chapter 14
Hollow section joints - Eddie Hole Introduction .......................................................................................................................... 65 Welding
65
In-line jointing ....................................................................................................................... 65 Joints in lattice construction ................................................................................................ 67 Beam-column joints ............................................................................................................. 68 Mechanical fixings ............................................................................................................... 69 Bolted connection systems ................................................................................................. 70 Castings ............................................................................................................................... 70 Conclusion ........................................................................................................................... 70
Chapter 15
Bridgework connections - Richard Thomas Introduction .......................................................................................................................... 71 Plate girders ......................................................................................................................... 71 Box girders ........................................................................................................................... 74 Footbridges .......................................................................................................................... 76 Conclusion ........................................................................................................................... 76
Chapter 16
Tension connections - Roger Pope Introduction .......................................................................................................................... 77 Applications ......................................................................................................................... 77 Key issues ............................................................................................................................ 77 Connection types ................................................................................................................. 81 Design issues ....................................................................................................................... 84 Conclusion ........................................................................................................................... 84
Chapter 17
Lattice towers and masts - Brian Smith Leg-to-leg joints ................................................................................................................... 85 Stay linkages ........................................................................................................................ 86 Bracing connections ............................................................................................................ 86 Conclusion ........................................................................................................................... 87
ARCHITECTURAL DETAILS Dublin's Design Pinnacle ..................................................................................................... 92 Fête of Twist ......................................................................................................................... 96 Revolutionary Motion ......................................................................................................... 102 Barajas Airport, Madrid, Spain........................................................................................... 106 Office Building, St Mary Axe, London ............................................................................... 110 A Stand of Two Halves....................................................................................................... 116 Racing Times ..................................................................................................................... 120 Making a Splash ................................................................................................................ 124 Pump Up the Volume ......................................................................................................... 128 School Campus, Bromley, Kent ......................................................................................... 132 Roof in a Landscape .......................................................................................................... 136 Watertight Design............................................................................................................... 140 Stretching a Steel Hyperbole ............................................................................................. 144 Footbridge, East Ham, London ......................................................................................... 148 A Bridge for Shanks's Pony ............................................................................................... 152 Tyne and Cheer .................................................................................................................. 156 Steps to Better Study ........................................................................................................ 160 Office, Devonshire Square, City of London ....................................................................... 164 Right Royal Restoration ..................................................................................................... 168 Factory for Rolls-Royce, Goodwood, West Sussex .......................................................... 172 Theatre, Plymouth .............................................................................................................. 176 Space Odyssey .................................................................................................................. 180
References
185
ENGINEERING CONNECTIONS An introduction by Roger Pope
To the Steelwork Contractor the core of the organisation has always been the Drawing Office. It is there that the client’s aspirations, the architect’s whims and the engineer’s scheme are converted from ideas into nuts and bolts. Until the shop fabrication details are finished not a piece of steel can be cut, and without cut steel there is no fabrication and there is no erection. The pressure is always there, such that the works are always pleading for drawings to be released for fabrication – otherwise overhead recovery will be lost and contractual delay, disruption and damages will ensue. Not an easy atmosphere in which to collaborate creatively with others in the design team; yet this creativity is what happens on every new project each with its unique challenges, as John Rushton’s article emphasises. In particular, Alastair Hughes describes how the engineer designing the frame needs to interact with the connection designer. A further pressure arises because the engineering connections shown on the detailed drawings determine the extent of nearly every cost within the Steelwork Contractor’s control. Decisions on the number of bolts, the thickness of welds, or the location of splices are committed once the connection is drawn. This emphasis on cost implications runs through the articles of both Kim Dando and Dave Chapman. Other than buckling (member and frame stability), connection behaviour is the area of steel construction that has been researched most extensively. Articles by David Nethercot, Graham Couchman, Ian Burgess and Roger Plank provide an insight into these ongoing developments. The “Green Books” represent the distillation of this research into practical everyday designs, and Alan Rathbone explains how much the current generation of connection design software relies on these standard references.
The details shown in the Green Books serve to meet the needs of a wide section of the steel construction market, but there are sectors with specialist requirements. To illustrate these needs, there are articles by Richard Thomas on bridgework, Eddie Hole on hollow section joints, Roger Pope on tension connections, and Brian Smith on lattice towers and masts. Thomas Cosgrove brings readers up to date on developments with structural fasteners – the nuts and bolts basics again. As is so richly illustrated in the Architectural Details section of this publication, steel details are much more than structural connections. Even in engineering terms, the interfaces with cladding and other building elements raise thermal, acoustic and vibration issues that need to be solved by the connection. The articles by Richard Barrett, Andrew Way and Stephen Hicks describe what is involved. It can be seen that connection design and detailing is not an easy job, but the steel construction industry has thrived through the efforts of generations of steel draughtsmen with the personal skills and nous to sort out such problems. I like to think that this publication is a testimony to all their efforts. It is surely no coincidence that the largest proportion of senior managers in the industry started their careers in the Drawing Office.
CHAPTER 1 Structural details are engineering John Rushton, Peter Brett Associates
Introduction
Design - analysis - detail design: a virtuous circle
Structural detailing is in danger of becoming neglected by structural designers. Perhaps this arises from possible misleading connotations of the word detail; the Web gives us definitions such as: ‘something considered trivial enough to ignore’, and ‘the smaller and less significant features of a subject’.
Design Whatever the structural material, whether it is formed on site or manufactured and assembled on site, at the initial stage there is the idea, the concept. It may use bespoke and/or proprietary components but whatever concept is used, ‘joining’ details are an intrinsic part of the concept.
Modern construction practice tends to compartmentalise suppliers and their detailing input to facilitate manufacture. The importance of details can be lost as designers bypass design and go straight to analysis. Practical experience and buildability advice are often not readily accessible when design decisions are being developed further exacerbating the situation.
In addition to structural details making the frame work as a whole there are ‘key’ details that drive concept decisions.
Structural details are an intrinsic facet of structural designs in any material. When details go wrong failure often results.
Shear head failure
Soffit cover failure after striking due to congested bars and lack of compaction
Structural ‘joining details’ visible in the final construction
Countersunk splice on 9m continuous spans used to achieve thin floor depths
Upstand box beam used to provide clear zone below slab edge adjacent to service riser
1
In summary, details are major parameters affecting engineering decisions and, in the case of any structure made off site there are issues of manufacture, transport, erection and connecting on site, a whole new set of parameters to be considered!
Analysis The engineering decisions outlined above are often driven by detail, transport, and joining concepts in the design phase. In industry there are constant complaints from contractors about the construction difficulties leading on from designers' ‘impractical’ designs which can often be linked back to a focus on computer analysis often with little or no attention paid to detailed construction aspects. Overly complex codes which obscure an understanding of the mechanical principles in action pose a further risk. In conceiving a design in either steel or concrete the engineer needs to have made conscious decisions on the buildability aspects of his design. These decisions are made in the design stage described above. They precede the analysis stage and will involve key architectural and structural details which are reflected in the way the model is constructed and analysed. Perhaps a better way to describe the complex transition of concept through to construction is in terms of a ‘virtuous circle’ of design, analysis and detail design.
Twin 152 UC integrated with external walling for flush internal walls
Internal exposed soffits
Detail design
Countersunk column splice to allow use of compact column casings
The third stage occurs after analysis and involves detail design. This often leads to redefining the details initially assumed following mathematical validation. Change to the concept design parameters can often be negotiated if under the control of others or adjusted within the structural concept for example by changing structural principles, the assumed material characteristics or construction method. The process then iterates in a virtuous circle of design, analysis and detail design.
Flat slabs to meet M&E engineer’s wish for clear soffits
Manufacture, transport, erecting and connecting As noted above, these are design constraints and major subjects in their own right. They exist in parallel with the design, analysis, design processes and sometimes dominate in a structure where offsite manufacturing methods are employed. Steelwork with its light weight and high strength characteristics tends to have handling and transport length issues to deal with. Precast concrete tends to have weight as the dominant factor. Both need to have simple and safe connecting techniques. In hybrid construction a mix of materials can be used in both the members and connection details.
Summary Structure free floor depth adjacent to main service riser on the right, to suit large ducts emerging into ceiling space
Details are part of the design brief and an essential consideration in modelling for analysis. They validate the conversion of structural design principles into an engineered solution.
2
Insitu concrete Steel construction is undertaken alongside other forms of construction – timber, masonry, precast concrete, insitu concrete and composite, mixed or hybrid combinations of these. It is important to review and capitalise on the significant developments taking place in these other fields. In the UK designs in insitu concrete were traditionally done by the designer who provided all the information needed for construction on site. The design of details in a monolithic concrete construction has tended to be closely linked with member and building component designs. Significant advances in flat slab construction cost and construction speeds have been made on the basis of proprietary formwork suppliers influencing designs to suit their products. As with steel, the most efficient structural form and details have developed as much from process improvement as from enhanced material properties.
21+m long roof beams being transported to site, lengths and joints designed to be within economical transport limits
In situ concrete frames owe their efficiency to the principles of structural continuity at moment transferring connections between vertical and horizontal members. Where concentration of shear has posed a challenge to the use of shallow/thin floors engineering ingenuity in the development of special shear reinforcement configurations has been applied. Use of preformed steel meshes/rolls, shear head assemblies and post tensioned slabs further promoted change. As with steel, the detail design challenge is transmitting forces and moments through the congested connection zone between horizontal and vertical building elements.
Splices in top and bottom chord introduced to limit lift weight in the factory for this heavy suspension transfer truss
This is a good example of how a design approach has been integrated with the practical realisation of the structural concept. In architecture separation of concept from practical details is not uncommon. Engineering is the practical application of scientific principles and consideration of detail needs to be applied at each stage as does the equally important concept of overall structural stability. Details make the structure work and there is a continued need for a single point of responsibility for ensuring compatibility of details and design in the overall structure.
Changes in designer working practices have also occurred. Reinforcement drawings and schedules by specialist teams of detailers and more contractor designed components in the structure have created new, conceptual and practical interfaces. An understanding of both the concept and construction of details has remained as important to the overall designer as it was before.
In the design process as the concept gets nearer to being validated as acceptable in all aspects of design, iteration occurs in the virtuous circle of design – analysis – detail design described above.
Implementation of the design on site Introduction Construction procurement is not considered here except to accept and acknowledge the risks implicit when interfaces and overlaps exist in design concept and detailing for construction. It is worth noting that new health & safety regulations are being introduced soon which will govern the actions of designers tending to formalise design for construction responsibilities. The responsibility for good engineering practice has always existed in the past but has sometimes been obscured by working practices. Buildability considerations are becoming a more formal process in design and if the approaches described above have been effectively carried out, success will result.
Proprietary shear stud rails
3
aids, prefabricated reinforcement assemblies and advances in offsite production of components. There is a convergence evolving between the two industries where the designer/consultant integrates the work of specialists in concrete structures in the same way as has evolved in steel structures. Just like steelwork, success will depend on concrete details being developed that satisfy the parameters of design and practical construction.
Heavy joints for offset precast cladding support
Buildings have become more complex and use of structural materials more varied as designers respond to the challenge of integrating structure with building services and architectural challenges of exposed structures and natural ventilation. Demands for rapid construction on site and continued pressure on price and cost of labour combine to act as further impetus to use off site manufacturing methods.
Bracket extended through inner wall skin to laterally support precast cladding
Structural steelwork Common practice in UK steel construction is for the engineer/ overall designer to conceive the design and the steelwork contractor to detail and manufacture the structural ‘joining’ details, i.e. the main structural connections. There are advantages and disadvantages in this traditional decoupling of frame analysis from connection design. The two have come together in some building forms, notably single storey buildings in the UK. Here there is generally a steelwork design and construct contractor and the single leading designer, in this case the contractor, will be responsible for overall stability as well as the detail design. Outside the UK specialist detailers, independent of the steelwork contractors manufacturing the building, are sometimes used.
Hybrid construction Hybrid construction involves combining site and off-site manufactured steel and/or concrete members. This form of construction seems likely to make an ever increasing impact on the UK. Hybrids exploit the most favourable structural characteristics in the materials. In buildings which incorporate different structural materials, there are key architectural and structural details that drive the structural design. A mix of architectural and building service parameters will tend to dictate structural positions and shapes. The structural ‘joining’ details will generally incorporate high strength material to facilitate factory manufacture, delivery and erection. There is an enormous body of research and knowledge in the design and detailing of conventional steelwork and in the UK the steelwork industry remains at the forefront in the off-site production and site assembly of structures.
Whatever developments occur in procurement processes, details will remain at the heart of steelwork design.
Summary Neither detail nor concept can exist without the other when applied to construction engineering. Designers, contractors and detailers have different primary and secondary roles. Whatever the construction material, all participants need to have sufficient competence in detailing for them to carry out their respective roles.
Steel details have a major role to play where transfer of loads needs to be done in small structural zones allowing the maximum volume of material to be in members manufactured off site. Challenges of interface management and by implication design of details will become crucial for UK based designers and contractors if they are to compete with well-developed continental expertise in hybrid construction (see Goodchild 1995 in References).
Challenges and opportunities Introduction So how do we draw on the design and detailing skills of the designer, and the contractor making the structure? The steelwork industry has a tradition founded in craft-based workshop skills. The rigour imposed by modern methods of factory production has, by necessity led to relatively close integration of design and construction techniques.
Here again we see the importance of details to structural designers and a big opportunity for the structural steelwork industry to apply its manufacturing and logistical skills. It is well positioned to exploit, both in areas of research and construction, expertise in mechanical joint behaviour, off-site manufacture and site assembly in new hybrid structural forms brought about by the agenda for sustainable design.
Integrated steel design and manufacture is comprehensively reviewed in the ‘Computer integrated manufacture of steel’ (CIMSTEEL) program (SCI P-178). This initiative has not had the publicity it deserves and the reader is strongly recommended to consult this reference. The concrete industry has made significant advances over the last few years in terms of rationalising site construction, design
4
CHAPTER 2 Connection design in relation to frame design
bracing. In a braced frame, it offers reduced beam size in return for modest extra connection cost. In an unbraced frame the benefit comes in the form of less elaborate, hence less costly, connections.
Alastair Hughes, Tube Lines
Methods of global analysis
Introduction
For any frame which is not statically determinate the designer must exercise another important choice – to analyse it elastically or plastically. This election has a profound influence on the properties the connections require. For instance, an understrength connection would obviously invalidate a plastic analysis in which members are assumed continuous. In an elastic analysis, the same condition does not apply; the connection only needs to be as strong (moment-resistant) as the design bending moment envelope locally demands. However there is a different precondition. In order for the elastic bending moment distribution to be valid, the connection must be Rigid. In other words the angle between the connected members must stay the same, as our computer assumed.
Frame design and connection design are inseparable. It is in the nature of steel construction that continuity cannot be taken for granted. Essentially we connect together prefabricated lengths, by bolting or welding. Both fastening methods have their merits, but for practical site connections bolting is usually the first choice. Welding is favoured in the workshop, but portability and erection convenience limit the size of a welded unit. It is therefore common to find connections located at junctions between beams and columns. One member, usually the column, can continue through the joint, so that the beams come as individual span lengths with a connection at each end. The most common, and certainly the most challenging, connection in steel building design is the beam-to-column connection. What makes it challenging is the tendency of the bending moment diagram to peak right where we want to make the connection.
Real connections are not perfectly rigid. The codes recognise this, and set a rotational stiffness (relative to that of the member connected) which is high enough for the flexibility of the connection to be neglected in the analysis. This is a condition for the validity of conventional elastic analysis, not a condition for structural adequacy of the frame. It is not relevant to plastic analysis.
Mostly, we are not prepared to weld in the field and must accept the constraints which come with bolting. Bolts are versatile and convenient to use, but the capacity of a bolt is limited – we standardise on M20 8.8 and are reluctant, in buildings at least, to push the size beyond M24 or the strength beyond 10.9. Available space also limits the number of bolts which can be made to act. The problem is much more severe where bending moment, not just shear force, is to be transmitted. Indeed it is not too far from the truth to regard the 'ideal' bolted connection – one which matches the bending and shear resistance of the connected member – as an unrealistic concept. For all but the smallest beam sections it can only be achieved if the depth is locally increased by a haunch. Haunches do have their place, in portal frames, but they are too space-consuming and workmanship-intensive to feature strongly elsewhere.
A semi-continuous frame is, by definition, one in which the relevant condition for continuity – either to be full strength or to be Rigid – is not satisfied. The frame is moment-resisting, i e not simple construction, but the global analysis needs to take account of the connections. This does dictate that the connections and the frame are designed together, not one after the other. At this stage it should be acknowledged that the advent of Eurocode 3 (EC3) has been helpful in imposing a clear distinction between the plastic and elastic approaches to semi-continuous design, historically having been somewhat obscured in BS 5950. The term 'semi-rigid' was applied to both. Admittedly, this term was familiar before plastic theory was invented, but the confusion is regrettable. The 'partial strength' connection in a semi-continuous plastically analysed frame needs very different attributes from the 'semi-rigid' connection in a semi-continuous elastically analysed frame. Although any given connection could be both semi-rigid and partial strength, it could equally be one and not the other.
'Simple' versus moment-resisting frames The customary response to this problem in UK engineering practice is to adopt what is known as 'simple construction' in which the connections are not asked to transmit moment. The disadvantage is that the beams have to be designed as simply supported. Also, bracing is compulsory, for lateral resistance. Nevertheless the method is economic and enduringly popular. Part of its appeal has always been that beam design and column design are decoupled by the nominally pinned connection. In the slide rule era, hyperstatic frame analysis was much more involved than it is today!
To summarise the principles:
There will always be circumstances in which simple construction is not an option. If bracing is unacceptable the frame must be moment-resisting. Pressure on structural depth, or the pursuit of lateral stiffness in a high-rise structure, may dictate a continuous frame. There is also the appealing middle course of semi-continuous construction, applicable with or without
CONTINUOUS
implies connections are FULL STRENGTH or RIGID
SEMI-CONTINUOUS
implies they are PARTIAL STRENGTH or SEMI-RIGID
SIMPLE
implies they are NOMINALLY PINNED
For elastic analysis, blank out the term including the word 'strength'. For plastic analysis ditto for 'rigid'.
5
Actually, it isn't the connection but the frame (or the design) that is semi-continuous. It only takes one connection that falls short of full strength or the Rigid assumption (whichever is relevant) for the frame analysis to need to take account of it. Nominally pinned connections, on the other hand, will not invalidate the analysis if they are modelled as such (and are not present in such numbers as to create a mechanism!).
It is at this point that it must be admitted that a gap exists between theory and practice. Code rules vary between the vague and open to interpretation (Traditional British) and the unambiguous but maybe unachievable (Modern European). Curiously, EC3 is over three times as demanding for frames that are unbraced, requiring a rotational stiffness of 25EI/L. That is something more than six times the rotational stiffness of the member end. Even for a braced frame, 8EI/L is about twice what BS 5950 might have been construed to require. The validity of elastic analysis of the unbraced frame is arguably less, not more, sensitive to connection flexibility than if the frame is braced, so more than one motive may be in play here. It does seem logical to relate the required rotational stiffness to that of the member, but EC3's multipliers are perplexing.
To make this conspectus complete, elasto-plastic analysis (taking account of every property the connection has!) is mentioned, fleetingly, and it is acknowledged that connections also appear in statically determinate situations where none of this subtlety applies. The bending moment diagram is immune to rotation at a cantilever connection or a midspan splice in a simply supported beam (though serviceability is not).
It's all rather academic, really, in the absence of a satisfactory way to quantify the connection's rotational stiffness for comparison. In practice designers exercise judgement, and some cling to the idea that if a connection is designed for strength it will be stiff enough to qualify as Rigid. Others are more particular; some of them feel better if the column web is stiffened. Judgement is not consistently applied, and this is as good an example as any of the hazards which lie in the path of those who would computerise the design process from start to finish. The BCSA/SCI Connections Group attempted to fill the guidance vacuum in the Green Book on Moment Connections (P-207) but this cannot be expected to remain the last word on a very problematic subject, especially if EC3 influence becomes more assertive.
In contrast to their predecessors, who could classify connections almost like airships (types A:rigid and B: limp), today's designers face an extended range of choice. The options are reviewed below.
Simple construction The nominally pinned connection is only required to transmit force: usually shear force, sometimes axial force such as code tie force. The characteristics it needs are rather negative ones. To realise the design assumption its rotation capacity, measured as the change in the angle between the connected members, might typically need to be something between 1 and 2 degrees. Throughout this rotation the connection must offer negligible resistance to moment. Of course the word 'negligible' is open to interpretation, and most practical connections fall some way short of perfect limpness. Does this matter? Yes, at least sometimes. In the world of elastic analysis there is always the possibility that if the stiffness of one part of the frame is wrongly judged, or wrongly ignored, the bending moment distribution will be invalidated, to the disbenefit of a member or connection somewhere in the frame. For example, the unrestrained bottom flange of a beam might find itself under compression, or columns might receive significant moments not designed against. So in practice efforts are made to ensure that 'simple' connections are reasonably flexible, while maintaining the degree of mutual restraint that is assumed in the design of the individual members and makes for safe erection. Standard approaches presented in the Green Book on Simple Connections (P-212) can be relied upon for most situations.
Continuous Plastic design Not least among the advantages of plastic design is that all this is irrelevant. All that matters for the analysis is that the connections are full strength – moment resistance not less than the member – and that can be calculated with reasonable confidence. There is, however, a concern that allowance ought to be made for overstrength members. Steel sections sometimes exceed their minimum yield strength (of 275 MPa or whatever) by a substantial margin. The danger foreseen is that a plastic hinge, supposed to form harmlessly in the member, would transfer itself to an adjacent connection not capable of acting as such. Arguably, this is a rather theoretical concern, because the stronger the members are the less likely that any plastic hinges will need to form and rotate, but it is a concern that EC3 latches onto with its provision that overdesigning the connection by 20% will do nicely. In view of the difficulty of designing a '100%' bolted connection, still less a '120%' one, this could be said to expose EC3's lack of practitioner influence. A more reasonable opinion is that only the brittle components of the connection – essentially, the fasteners – are at risk and so long as design rules for these remain relatively conservative '100%' will suffice.
Continuous Elastic design Continuous construction is the opposite of 'discontinuous' simple construction, and is the other traditional option. Elastic analysis is the traditional method of global analysis and when these come together the connections are required to be Rigid. Not rigid in the sense of resistant to rotation (not a pin), not rigid in the sense of perfectly stiff (the theoretical ideal) but Rigid in the sense of rigid enough to qualify according to code rules. Let's give this special meaning of the word a capital R.
There's no gainsaying the fact, however, that even the 100% connection tends to be an elaborate and expensive fabrication involving haunching and stiffening. With the significant exception of portal frame sheds, continuous plastic design has had very limited success in penetrating the market.
6
In unbraced frames, semi-continuous plastic design should be thought of as the new name for the very traditional wind-moment method, which in its original form was regarded with affection by some and with suspicion by others. The connections were supposed to possess 'unusual powers of discrimination' between wind- and gravity-induced moments. In reality, they just need to be ductile, so that the frame can adapt plastically. Provided that the proper connections are used, the method is now thoroughly respectable and recommendable. A point to watch is that strength design is only part of the process; the issues of serviceability and stability must also be addressed. Up to date guidance is found in SCI's P-263 and P-264. Standard ductile connections for wind-moment frames are tabulated in the Moment Connections Green Book; they differ only in having the same pattern of bolts top and bottom (because wind moments can reverse). BCSA HQ: Simple construction
Another use for standard ductile connections has been observed. Safe erection is an industry preoccupation, and there are situations in simple construction where a connection designed to be limp is far from ideal in the incomplete frame. If a full depth end plate connection is desired to cater for this condition, the ductile connection can provide it – together with the rotation capacity that allows it to be harmlessly substituted for the nominally pinned connection of the original design.
Semi-continuous Elastic design For completeness, semi-continuous elastic design has to be mentioned. This is semi-rigid design in the strict sense of the word. It is not recommended, because reliable formulae for connection stiffness (to be fed into the analysis as rotational springs at member ends) are not available. Even if they were, questions would remain about the sensitivity of the resulting bending moment distribution to the accuracy of the input. Sharp-pencilled designers would be tempted to contrive a SCI HQ: Wind-moment frame
Semi-continuous Plastic design The answer may lie in semi-continuous plastic design. Design with partial strength connections, usually at beam ends, virtually guarantees that plastic hinges will form at the connections, and if they are ductile they are qualified to perform as such, rather as a class 1 'plastic' cross section is. Ductility implies rotation capacity without loss of strength. Ductile connections need to be carefully designed, but can in practice be selected from the standard range of predesigned connections, to be found in the yellow pages of Design of Semi-continuous Braced Frames (P-183), published by SCI. They are quite ordinary looking end plate connections in which plate thickness is controlled (relative to bolt tension) to ensure safe deformation. Their performance has been verified by testing. In terms of elaboration and therefore cost they have more in common with 'simple' connections than with those for continuous construction. This approach promises to be the next evolutionary step in multistorey braced frame design, and it is worth emphasising that the benefit which comes in the form of reduced beam depth will often outweigh the direct cost saving. Semi-continuous joints were used on the Arup Campus
7
bending moment diagram with equal end and span moments, so as to size the beam just as if it had been plastically analysed but without the ductile connections. There is no guarantee that a semi-rigid connection will qualify as ductile.
the responsible frame designer must, at least, take steps to establish that a practical connection can be made to transmit the design forces and moments. Of course this holds true for many non-tubular connections as well; indeed it can be argued that the traditional designer-contractor handover of responsibility for connection design is only really valid for simple connections.
However, semi-rigid analysis does have a valuable role, and that is for serviceability checks, which are done elastically irrespective of how the analysis for strength design was done. The more refined the frame design, the more likely that serviceability will control. Connection stiffness contributes immeasurably to serviceability (true to some extent even of nominally pinned connections). More reliable formulae for connection stiffness are worth pursuing for serviceability checking purposes, not least for plastically designed frames. EC3's formula, although an improvement on its predecessor, is not well respected. It sets out to encompass a wide population of connection styles. Restricting the scope to a single family, such as the standard ductile connections, proves to be a more successful approach (see Brown et al, 2001 in References).
The Green Books and the Connections Group A shared understanding of connection design principles and best practice is vital to the smooth running of our industry and our projects, and helps to minimise the issues which arise at the interface between design and construction. Standardisation, of design approach as well as detail configuration, can deliver process efficiencies for all styles of connection. In the case of ductile connections for semi-continuous construction, it does more than that; it is the key which unlocks obstacles and makes the method a practical proposition. For UK steel construction, the vehicle for connection design guidance is the series of Green Books prepared by the BCSA/ SCI Connections Group, with industry and taxpayer support, and jointly published by BCSA and SCI. These not-so-slim volumes are sometimes described as Industry Standards; their authority aspires to the level of British or European Standards but ultimately derives from the esteem of the practical people who use them. They will never be all-embracing, but over the years since the Group was established in 1987 most areas of connection design practice have received attention.
Composite construction The popularity of composite construction in the multi-storey market is such that it may seem remiss not to have mentioned it so far. In fact, for simple construction, there is really no difference in connection design. Moment connections in composite construction are another matter. Compression in the bottom flange of the steel beam is coupled with tension in the concrete slab, or to be more precise the rebar, whose elongation capacity becomes an important consideration. Another is the wish to avoid spending more on welding compression stiffeners at bottom flange level than is being saved elsewhere. Designers of this type of connection are operating not far behind the research front and are advised to consult the Composite Connections Green Book (P-213) and be aware. An encouraging development in the rebar supply chain is the promised availability of a new, more ductile, 'Class C' bar.
Currently the Connections Group has around 25 members, representing steelwork contractors, design firms, researchers, steelmakers, bolt manufacturers and the increasingly important stakeholder group of software developers. Under the chairmanship of SCI Director Graham Owens, the Group's next challenge is to adapt to the new 'correctness' coming our way from Europe. There are positive as well as negative aspects to the inevitable wholesale revision exercise, one of which is the prospect that a wider steelwork community may benefit from the Group's efforts and investment in this field.
Tubular construction A section in a general survey cannot do justice to tubular connections, which are a specialist subject – and a fast moving one, with the development of innovative 'blind' fastening techniques for bolting direct to the face of a hollow section. These may increase interest in hybrid multi-storey frames using square hollow columns partnered with conventional beam sections. The latest edition of the Simple Connections Green Book (P-212) has been expanded to include these systems. Connections between two (or more) hollow sections, commonly encountered in trusswork and the like, are covered in the familiar CIDECT design guidance which is published in several forms, now including EC3. As all its users are aware, joints are rarely capable of transmitting the full capacity of the intersecting member, and the geometry-dependent reduction factors can be quite low. This has implications for the division of labour between the connection designer and the frame designer. The way our industry operates means they are liable to be different people working in different organisations, but
Illustration of ductile connection under test at Dundee University
8
Conclusion To conclude, we do well to remind ourselves that nobody told the frame how it was analysed. The distinctions drawn between elastic and plastic analysis, and the important differences identified between needed connection properties under the different design regimes, are a function of the means as much as the end. The state of stress in the real frame is influenced by our design, but it is also influenced by such wild cards as initial lack of fit, differential settlement, unforeseen composite action and load history. It is a comfort that steel frames, however designed, have a forgiving nature as evidenced by past performance over a century and more.
9
CHAPTER 3 Connection design programmes
Most simple connections are driven by good detailing practice i.e. if it is a ‘good’ detail then it will most likely be adequate structurally. However, variations from the ‘standard’ do occur but using the principles laid down in the design models many of these can be accommodated within the scope of the software. Anything that is unusual or very variable e.g. gusseted brace and truss connections, tend not to be blessed with definitive design models. Consequently, you may find these beyond the scope of your current software. Without well researched design models the software suppliers have some difficulty providing robust and safe calculation routines. Usually it is not commercially viable to carry out the research themselves be it desk study and/or physical testing.
Alan Rathbone, CSC (UK) Limited
Introduction For some time now, the majority of steelwork design has been carried out by software. This is particularly true of moment connections. For connections in ‘Simple Construction’ e.g. flexible endplates, fin plates and double angle cleats, the size of the connection is usually governed by good detailing. It is no surprise then that the prevalent 3D steelwork modelling/detailing software incorporate modules for simple connection design. Similarly, moment connections are, by their very nature, not simple to design. The latest and most efficient design models for moment connections have been written assuming that nobody would attempt such a design by hand. Hence, nearly all of these types of connection are designed using software.
For those simple connections that are within the scope of the software (or can be modelled within the given scope), the program will check the ultimate resistance of the various components based on the distribution of forces within the connection arising from applied vertical shear. Where structural integrity requirements also need to be met (more often since the revised version of Part A of the Building Regulations was published in December 2004), the program should carry out the appropriate strength checks. Thus for simple connections the design considers only Ultimate Limit State checks for applied shear and tension due to tying forces. There is no consideration in the design model of applied moment – the software input should reflect this or, if integrated within a wider design or detailing model (see below), the data passed to the design engine in the program should be limited to this.
In both cases (simple and moment connections) not all of the connections that ‘appear’ on site can be dealt with clearly in the software. Hence, there is often a need to either ‘model’ the connection within the scope of the particular program or resort to hand calculations. Whether the connection you are designing is clearly within scope or you are manipulating the software to give some reasonable answers, it should always be remembered that the software is only a tool. It is important that the designer understands connection design per se and then uses the software to assist with the (boring!) calculations.
The Ultimate Limit State loading as described above (i.e. no moment) is consistent with including pins at the member ends in the analysis model. However, this assumption of pinned ends infers that the beam end and hence the connection will rotate relative to the supporting member. This requires both ‘ductility’ and ‘rotation capacity’ from the ‘simple’ connection. It is believed that most current software packages currently do not include any calculation to check these requirements. Indeed the underlying design models in the Green Book provide no method for doing so. Ensuring ductility and rotation capacity is handled in the design models by applying detailing rules which research has shown will provide an adequate measure of both. This must be borne in mind if you are tempted to stray from the standard details (usually the default settings in the software) or you are using the program to model an unusual configuration.
Underlying design models Simple connections The publication of the ‘Green Book’ on simple connections in 1991(P-205) provided a definitive set of design models for the most common types of simple connections – flexible end plate, fin plate and double angle cleat. (Note that at the time of writing the first edition of the Green Book, fin plates were widely used in some parts of the world but not in the UK. The Green Book was ground breaking in not only providing design models and standardised connections but also in promoting a new type of simple connection.) These models have become the de facto standard for such connections. The range of connections covered includes splices and base plates. This publication was later enhanced by the separate publication of a set of worked examples and safe load tables (P-206). Most recently, these two publications have been combined into one volume (P-212) with additional material and in compliance with BS 5950-1: 2000.
Moment connections The design of moment connections is the subject matter of another publication in the ‘Green Book’ series which was published in 1995 (P-207). This provided a definitive set of design models for the most common types of moment resisting connections,
The design model component was written in a form that was clear and concise but more importantly supported the logic associated with implementing the design models into software. The safe load tables and worked examples ensured that quick hand designs could still be carried out by those who do not have access to the requisite software. The safe load tables, of course, also provide the function of allowing quick design estimates.
• bolted end plate – flush or extended with or without haunches; • direct welded; • splices; • column bases. These models have become the de facto standard for such connections.
10
However, prior to the introduction of this Green Book there were a number of design models for bolted end plate and other types of moment connection (see BCSA’s Manual 1982 and Horne & Morris 1981 in References) that were based on an essentially linear distribution of bolt forces. Using elementary structural mechanics these bolt forces were distributed into and tracked through the connection components. Each component was then checked for the force that it ‘saw’. If the component had insufficient strength then its size or strength had to be increased or stiffened in some way.
of stiffness is dependent upon the type of connection e.g. haunched connections tend to be very stiff, and the situation in which it is to be used. You are advised to read Section 2.5 of the Green Book on moment connections (P-207). A new but logical design model for column bases is provided in the Green Book. It draws on the principles embodied in but is different to the design model for simple bases. It is most appropriate for bases with significant applied moment and some axial load (as opposed to an axially loaded base with some moment).
The approach in these earlier models differed from that contained in the Green Book. The latter is a so called ‘capacity model’. That is, the resistance of each component is calculated and the minimum resistance of each component ‘summed’ to give the overall moment of resistance. This is then compared with applied moment. Under most circumstances this allows plastic redistribution to occur which can lead to more efficient designs than the earlier models. It also has an effect on the approach adopted in software (see below).
Composite connections There is a publication P-213 in the Green Book series that deals with composite connections. The design model takes a ‘reinforced concrete approach’. The publication also deals with the effects on frame design. Such connections are thought to be little used and the author is unaware of any software that deals with this type of connection. Further discussion is, therefore, beyond the scope here and the reader is recommended to refer directly to the Green Book.
Current and past design models have not been updated to comply with BS 5950-1: 2000. However the effects are believed not to be overly significant. The author is not aware of any ‘official’ information on the effects of these changes on the earlier design models. Responsible software vendors are likely to have amended their programs to suit these changed requirements whilst continuing to comply with the principles of the current or earlier design models as appropriate.
Complex connections In all the above design models, the approach is based primarily on the application of relatively simple structural mechanics. Resistances are generally based on code requirements, normal engineering criteria e.g. yield or semi-empirical requirements based on either physical tests or experience. This approach suits the vast majority of connection types that you will come across in every day design. Specific software usually exists for most of these (see below).
The design model component of the Green Book was written in a form that was clear and concise but more importantly supported the logic associated with implementing the design models into software. Indeed they have been written assuming that nobody would attempt such a design by hand! The safe load tables provide the function of allowing quick design estimates for those who do not have access to the requisite software. It should be noted that those provided for portal haunch connections are rather conservative. Worked examples ensure that the design process can be understood and hence assist in the confirmation of expected results from the software.
However, there may be occasions when connections are sufficiently large, complex or critical and different that a more fundamental approach is required. In this case there may be no given design models and the performance of the connection can be ascertained using full finite element modelling. This will provide an analysis of the connection only and not the design. To model these sufficiently accurately will necessitate the use of ‘high end’ analysis software. You need to know what you are doing and be confident to consider such items as:
Ultimate Limit State loading includes moments (of course), shear and axial load – the last is usually glossed over or ignored in the earlier models. This is consistent with including ‘fixed ends’ to members in the analysis model. In a 3D model this is equivalent to providing rotational restraint about the local major axis. Most bolted end plate connections have little moment of resistance about their vertical axis and hence in the analysis model the rotational restraint about the local minor axis should be free (i.e. pinned out of plane). It should be noted that the assumption of fixed ends infers that the beam end and hence the connection will remain at the same angle to the supporting column until its reaches its maximum moment of resistance. This requires an assessment of the stiffness of the connection. It is believed that most software packages currently do not include any calculation to check these requirements. Indeed, the underlying design models in the Green Book provide no method for doing so. Ensuring adequate stiffness is usually a question of experience. This must be borne in mind if you are tempted to stray from the usual details or you are using the program to model an unusual configuration. The importance
• complex finite element types e.g. 3D brick elements; • complex material behaviour e.g. post yielding behaviour; • complex interconnectivity e.g. allowing for lack of fit and tolerances; • complex stress states e.g. built-in stresses due to welding or residual stresses due to rolling. Such detail is obviously beyond the scope here.
11
Range of software
• design may be carried out by the building designer, the steelwork contractor or the design-and-build contractor usually with most effort expended at the detailed design stage;
Be aware that the all the various connection types – simple, moment, connections, splices, bases – may be packaged differently e.g. they may be one product or several products.
• software will be available as ‘stand-alone’ programs as well as integrated into the building design or detailing software. Whilst the Green Book has not been updated to the latest version of BS 5950-1, you should ascertain whether the necessary changes have been made to the software. If not then you will need to make a judgement as to whether for your connections in your structure the impact of the changes is significant.
Simple connections The nature of the steelwork industry in the UK is such that simple connections are usually ‘detailed’ rather than ‘designed’ and the work is carried out by the steelwork contractor and not the building designer. The consequence is that, for simple connections,
There is also a wide range of software for moment base design.
• connection design software is usually built into the steelwork detailing or 3D modelling software; • a facility for design in building modelling software might be useful; • consequently, programs for simple connections are not often run in ‘stand-alone mode’. Clearly, for ‘design and build’ the situation is slightly different and designers have the choice to check their simple connections either at the building model design stage or at the steelwork detailing stage. This type of software will usually deal with the standard details for simple connections e.g. beam to beam and beam to column using flexible end plates, fin plates or angle cleats. Simple bases and splices are normally included. A range of non-standard but reasonably common connections may also be included e.g. skewed or offset connections.
Eaves connection (courtesy CSC Fastrak)
Gusset connections e.g. as used in trusses between double angle bottom chords and single angle internals or as used between braces and the main framing steelwork can vary widely in their configuration. Consequently software applies principles rather than discrete design models. It should be remembered that, in trusses, the final connection design influences the member design. For members in tension, the number of bolts and their layout will determine the ‘effective net area’ required by BS 5950-1: 2000. For members in compression, the orientation of the members and number of bolts in each connected leg determines the effective length used to establish the buckling resistance.
Apex connection (courtesy CSC Fastrak)
Moment connections
Other connections
There is a wide range of software for the design of moment connections. The most common forms of moment connection that you are likely to encounter are the haunched beam to column bolted end plate connection and the beam to beam bolted end plate connection found at the eaves and apex of portal frames (respectively) as illustrated. The software may be based on the Green Book approach and/or the earlier design models described above. Unlike simple connections, these are major connections requiring significant design effort. This can have a number of effects,
Software is also available for, • hollow section connections typically used in tubular trusses; • Vierendeel connections; • hybrid connections e.g. gusset plate to CHS or CHS/SHS internals to I- or H-section chords. It should be noted that software is not available for 100% of the connections you may encounter.
• the complex nature of the calculations effectively means that software is essential;
12
Some particular issues
due to the asymmetrical nature of a single sided weld are usually ignored owing to the comparatively stiff nature of the surrounding detail.
Welds Designers may have noticed an increase in the required size of certain welds over recent years. This has been brought about by two effects,
Secondly, shear stiffeners are often required in the web panel of portal frame eaves connections. There are various configurations including the ‘Morris stiffener’. Due to the usual shape of the web panel, these stiffeners are often steep and abut the column flange at a very acute angle. This angle is often so shallow that,
• In BS 5950-1: 1990, there was a rule for symmetrically placed fillet welds that allowed their combined resistance to be taken as equal to the capacity of the joined plate when the sum of the throat thickness was equal to or greater than the thickness of the joined plate – the so-called "sum of throat thickness" rule. It is suggested that this rule was based on original research using S275 steel and that later work showed that this could not be assumed for S355 steel. Hence for higher grade steels this rule was not safe and consequently the rule was removed from the 2000 version of BS 5950-1.
• the welds fall well outside the ±30º rule in BS 5950-1: 2000; • there is insufficient access for the weld apparatus such that placing the weld with any degree of consistency, accuracy and consequent quality is very difficult. This can be exacerbated by the juxtaposition of a compression stiffener at the same location (see illustration). Justifiable but conservative design assumptions are made until such time that the steelwork industry can establish a reliable calculation model for this area.
• The modern, efficient design models encapsulated in the Green Book for moment connections (and in EN 19931-8 – the part of EC3 dealing with connection design) requires plastic redistribution of bolt forces. Also, to ensure adequate ductility in a connection, failure of the brittle components (welds) should be avoided. These criteria lead to the necessary provision of ‘full strength’ welds in certain locations. As an example of the combination of these two effects, consider the weld between the web of a beam and an end plate in a moment connection. This is required to be ‘full strength’ i.e. be able to generate full yield in the web. Taking a web of 8.5mm thickness and applying the ‘sum of throat thickness’ rule, a fillet weld with a 6mm leg length applied each side would suffice for both S275 and S355 steel. Since this rule no longer appears in BS 5950-1: 2000 these welds must be ‘designed’. Assuming matching electrodes, the required leg lengths of the weld are 6.1 and 6.9mm for S275 and S355 steel respectively. This results in an increased weld size for the higher grade material (assuming that you agree that 6.1mm is within an engineering tolerance of 6.0mm!). The effect becomes more complicated if you are welding an S355 beam to an S275 plate and/or you are using non-matching electrodes!
Detail of cap plate
Although not part of the same effect there are a couple of other issues that have been addressed in rethinking the weld requirements contained in BS 5950-1: 2000. Firstly, a ‘cap plate’ is sometimes provided to the top of a stanchion to assist with connecting parapet posts for instance. When present, a cap plate can also be considered as a stiffener and indeed may be provided for that sole purpose. They are normally cut short of the flanges to avoid any interference with other components of the connection or frame (as illustrated). There is a temptation to provide a weld between the cap plate and the edge of the flange but there is a fear that this could become ineffective due to lamellar tearing in the stanchion flange. Hence, typically only the full profile weld to the underside of the cap plate is provided and relied upon in design. This can cause difficulties in achieving a sensible weld size since all of the design force is resisted by a single sided weld. Any bending effects
Detail of junction of stiffeners
13
Moment connections
where members are ‘design’ members and not ‘analysis’ members then the end fixity can be set in a more appropriate manner by default e.g. simple beams have pinned ends. This terminology refers to an element between two nodes in analysis, whereas this would be modelled as an actual beam or column in design.
The difference in approach for moment connections between earlier models and that contained in the Green book was described above. Essentially the earlier models used elementary structural mechanics to distribute and track the bolt forces through the connection components. Each component was then checked for the force that it ‘saw’. If the component had insufficient strength then its size or strength had to be increased or stiffened in some way. This allowed logic to be built into software to decide how to deal with any component that failed a design check.
Hollow section connections Hollow section connections are most typically used in tubular trusses. Despite the fact that the internals are fully welded to the chords, this type of connection behaves as if it were pinned at the final stage of loading i.e. at Ultimate Limit State. This is due to the relatively thin walls of hollow sections and the large deformations that such connections can sustain. They can, of course, also be designed to resist significant moments so that Vierendeel action can be generated.
The approach contained in the Green Book is based on a so called ‘capacity model’. That is, the resistance of each component is calculated and the minimum resistance of each component ‘summed’ to give the overall moment of resistance. This is then compared with applied moment. None of the components fail as such and thus it is only the overall moment of resistance that is adequate or not. Since no individual item fails it is much more difficult to develop a logical strategy that can be built into software to automate the process of achieving a satisfactory configuration i.e. a PASS.
If the truss is designed first, then due to the efficient way in which hollow sections carry the applied forces (mainly axial), relatively small, very thin walled sections can be found to be adequate. Later, it may be found difficult to justify the connections between such minimum weight members and stiffening may be required. Since most of the work and hence cost is in the preparation and welding of the member ends then attempting to minimise the section size is counterproductive. When using software it may be best to be prepared for some iteration in the analysis/design process i.e. guess the member sizes for the truss analysis, being generous on the sizes selected, check the connections, update the analysis/design model and then check the member sizes.
The Green Book model can be applied to connections used in the Wind Moment Method and to semi-rigid design for the calculation of resistance at the Ultimate Limit State. However, current UK design guidance does not include methods for the calculation of stiffness and this has a number of impacts, • the stiffness of the connection may be insufficient to hold the connecting members at the same angle all the way through the load response history i.e. beams cannot be assumed to have fixed ends. This can have a significant effect on the deformations of the structure under Ultimate Limit State loading and significantly increase any second-order effects;
Bases
• if the connections are ‘partial strength’ e.g. as assumed in the Wind Moment Method then they must have adequate rotation capacity. The wind moment connections detailed in the Green Book do possess this attribute;
The design model for simple bases i.e. those subject to axial load and no moment is consistent between the Green Book and BS 5950-1: 2000. For moment resisting bases there is no definitive design model in BS 5950-1: 2000 but is well covered in the Green Book. This model (as mentioned above) works well for bases with significant moment and some axial load. However, the design model for moment bases with significant axial load and very little moment does not coincide with the simple base model. For example a simple base with 1000 kN axial load will not give the same result as a moment base with the same axial load and 1 kNm moment. You should bear this in mind and decide whether it is possible to ignore very small moments and design the base as simple. Since the design models are discrete then it is likely that the software will also be discrete, if not at the interface level then certainly in the calculation routines ‘behind the screen’.
• as for the Ultimate Limit State in the first bullet point, the lack of stiffness can markedly increase the deflections under the Serviceability Limit State loading. All of these effects are taken into account in the Wind Moment Method but over a limited range of building arrangements. Whilst the software will calculate the strength of semi-rigid connections for use in the general case, the other effects described are not taken into account and these have an effect on the frame design. In a similar vein, the design models for both strength and stiffness of connections into the web of I- or H-sections and into hollow sections are as yet undeveloped. Such connections are still often called for by the designer. With no design models, that means no software and difficult fabrication, all leading to increased costs. Software for general analysis was often blamed for these occurrences since in a general analysis package the assumption has to be that all member ends are fixed. This is so that the model will analyze without any instability in the solution and relies on the designer making a positive effort to change the end fixity to something more realistic. However, with modern building model software
14
Advances foreseen
Software
Design models
Any changes to design models as described in the previous sub-section will obviously affect the software. Any safety issues that arise would of course be addressed by the software vendors whilst, given a commercial case, new connection types are likely to be included as the design models become available. A good example of the latter would be the design of moment connections to webs – once the research is complete and a robust design model exists it is only necessary for customers to perceive a benefit in having this type of connection for the software vendors to react and include them.
The terms stiffness, ductility and rotation capacity have been mentioned above in the context of both simple and moment connections. Whilst the assessment of the adequacy of the behavioural characteristics is currently based on experience, methods for their calculation are becoming available with the introduction of the Eurocodes and accompanying documents. This may be a double-edged sword in that it will now allow us to calculate such things but in some cases we may not like the answers! The approach to strength design in the Eurocodes will see some relatively minor changes to the way we carry out some of the checks, but the main topics that are likely to be influenced by the Eurocodes either directly or indirectly are:
One of the main areas of change for software for connection design is in ‘integration’ that is allowing the connection design from within a larger software package. This has a number of advantages including keeping all the data together, facilitating change control, no replication of data, no re-input of data (and hence fewer mistakes) etc. This has already been present in detailing software particularly for simple connections for a number of years. It is becoming increasingly popular for building modelling software.
• the checks for structural integrity in simple connection may improve with work that is going on in parallel with the introduction of the Eurocodes under the auspices of Technical Committee 10 of the European Convention on Constructional Steelwork (ECCS, the umbrella body in Europe for all the trade associations such as our own BCSA). Design forces for structural integrity checks will be contained in EN 1991-1-7, Actions – Accidental.
Whatever type of connection you are designing, it should always be remembered that the software is only a tool. It is important that you understand connection design per se and then use the software to assist with the calculations.
• the design model for moment connections contained in the Green Book is based on the ENV version of EC3 – the socalled Annex J method. Whilst this model is no longer in an annex the requirements remain essentially unchanged in the forthcoming part of EC3 that deals with connections, EN 1993-1-8. This should ease the adoption of the Eurocode rules into the UK for those types of connection. • EN 1993-1-8 also provides a design model for bases. This draws heavily on the principles developed for moment connections and as such it is very different from current design models. This an area of potential concern since the design model is so new and untried that there may be idiosyncrasies and flaws that might produce designs that are quite different to those currently in use. • occasionally moment connections into one side of the web of I- or H-sections columns are required. Such connections are normally avoided due to the difficulty of making and proving an adequate detail. However, recent research by some members of the TC10 committee of ECCS has shown that with little stiffening such connections can sustain substantial moments. The research is incomplete and has not yet addressed the thorny issue of stiffness. However, in time full design models for such connections are likely to become available.
15
CHAPTER 4 From laboratory to laptop
contradictory behaviour. In one case the ratio of moment to shear applied to the joint appeared to have no effect; for the other it appeared to influence the ability of the connection to develop its full moment capacity. We postulated various explanations but did not really have sufficient evidence from the different measures of response (deflections, strains, rotations etc) taken in both series against which to check these. It was not until we developed a fully validated FE model, that we were able to appreciate subtle changes in the exact mode of failure between the two series. Subsequent studies, making small changes to key parameters, clearly showed that both sets of tests were correct but that without the supplementary numerical studies we simply did not have sufficient test results to understand all aspects of the problem. My assessment is, therefore, that there will always be a need for experiments that both:
David Nethercot, Imperial College
Generating the knowledge base Research necessary to provide the improved understanding of the behaviour of connections, on which to base better design methods started in the early part of the last century. Pre World War 2 such work was largely confined to laboratory testing. After the war this became more scientific, with better conceived tests, more comprehensive monitoring of performance and more serious attempts to convert this knowledge into behavioural models for various connection types. The past few decades have seen the emergence and growth of numerical work, during which time the role of laboratory testing has, to some extent, changed. The table below contrasts the main features of the experimental and the numerical approach.
• Permit the key behavioural features to be identified by observation. • Provide confidence in the numerical approach by careful validation.
Use of numerical analysis The key question is "can numerical analysis replace physical testing and if so when?" The answer is, of course, "it all depends on what features you are considering". To take an example from a different field: when did you last read of an elastic test on a perspex model of a bridge deck designed to produce a set of influence coefficients? Forty years ago this was normal practice. Nowadays numerous commercially available FE packages do the job far more quickly and cheaply. For steelwork connections, if we were interested in the elastic stresses in an all welded beam to column joint using UB and UC sections, then an appropriately configured FE run would be much more convenient than a Laboratory test. However, obtaining the full load – deformation response for a bolted fin plate connection represents a substantial challenge – even for the most expert research team. For the present, the correct response to the question is that both approaches are required and that the most influential gains in understanding are likely to come from an intelligent combination of the two.
Future needs Currently, the following features of connection behaviour remain difficult to model: • Friction and slip of bolts loaded in shear positioned in clearance holes; • Bolt pretension for bolts subsequently loaded in tension and bending; • Residual stresses due to welding, e.g. in tubular joints; • Lack of fit effects; • Shear connector/metal decking/rebar interaction in composite joints.
No doubt members of the research community in many parts of the world are currently working on these and other challenging issues but for the foreseeable future the laboratory will remain one important component; the contrast with the situation a century ago is that it now has a powerful ally in numerical analysis.
To take one illustration from my own experience, some 15 years ago we were attempting to understand composite action in connections between steel columns and composite floors. In two separate series of tests we had observed seemingly
Experiment
Numerical
Exact replica of practice
Need to model the interaction of several complex features and phenomena, some of which are still "too difficult"
Full range behaviour
Post peak loading response often difficult to follow
Expense (time and cost)
Once validated, software is cheap to run
Needs purpose built facilities and skilled manpower
May be run on any suitable platform; technician and experimental skills replaced by "the intelligent user"
Sharp rise in costs and complexity with increase in physical size
Physical size of the connection studied has negligible effects
Impossible to fully cover the range of all important parameters
Now feasible to conduct many runs with systematic variation of all key parameters
Possible experimental malfunction e.g. premature failure due to inappropriate specimen/rig design
Possible numerical problems and/or false results
16
Illustration of lab test (above) Reproduced courtesy of LUSAS. For more details of the worked example by Jim Butterworth see: http://www.lusas.com/pdf/CS502_FEA_of_Structural_Steelwork_Beam_to_Column_Bolted_Connections.PDF Illustration of FE graphic (below)
17
CHAPTER 5 Connection rotational characteristics
example is connections that, to be compatible with the frame member designs, need to have a certain strength (partial), certain stiffness (semi-rigid) and high ductility. These are most commonly used in what are termed wind-moment frames, but also in semi-continuous braced frames (SCI publication P-183). The SCI publications covering wind-moment design (P-263, P-264) include standard connections that have been shown by testing to perform appropriately. Personal experience suggests that designers who do not follow this guidance but detail non-standard connections, normally by considering only the strength requirements, do so at their peril. Whilst these connections are likely to be sufficiently stiff, they are unlikely to have sufficient ductility (choosing connection components that deform to give ductility is often in conflict with a need to provide strength and stiffness, so a fine balance must be achieved).
Graham Couchman, BRE
Introduction Prompted by the development of Eurocode 3 in recent years (decades), specialists in the steelwork sector have been giving considerable thought to the rotational characteristics of steel, and indeed steel-concrete composite, connections. These are: • Strength (moment resistance); • Stiffness (or rigidity);
Building on this theme of connection characteristics, composite connections present a whole range of issues particular to themselves. At face value their use seems very interesting; high levels of strength and stiffness can be achieved without the need for complex (costly) steel detailing, simply by adding local slab reinforcement. This can provide significant benefits in terms of reducing beam deflections, but care is needed to ensure the connections ‘work properly’. However, whilst a book of standard composite connections has been produced by BCSA/SCI as part of the Green Book series (P-213), more guidance is needed before they will find more widespread adoption.
• Ductility (rotation capacity). Somewhat in contrast to this, the traditional approach to steelwork connection design/detailing in the UK has been to consider two types of connection, namely ‘pinned’ (or ‘simple’) and ‘rigid’. A third type, ‘semi-rigid’, has occasionally been used, for example in wind-moment frames (even if the connections were not always recognised as such). As considered below, even the names of these three types are potentially misleading.
Connection types Rigid connections are designed using one of various models, or perhaps tabulated data for standard connections from the BCSA/SCI Green Book on Moment Connections (P207), to calculate their strength. No explicit consideration is normally given to either the stiffness or ductility of such connections. The former characteristic is just as necessary as the connection strength, but is normally achieved inherently because the detailing needed to achieve strength (‘thick’ plates, lots of ‘big’ bolts etc) is compatible with that needed to achieve strength. Rigid connections do not generally need to be particularly ductile, so the inability of ‘thick’ plates and ‘big’ bolts to deform, allowing rotation, is not a problem.
Conclusion The rotational behaviour of composite connections is complex. Substantial amounts of local reinforcement are needed to achieve a moment resistance that is not too low compared with the sagging moment resistance of the composite beam (which of course is much higher than that of the steel section alone). Using substantial reinforcement means that the need for connection rotation, to redistribute moments into the span, can be limited. The rebar also helps to ensure the connection has reasonable ductility and stiffness. However, using large amounts of reinforcement in a confined area poses practical, including anchorage, problems. A particular problem with composite connections is their use with unpropped beams, which are clearly the most common. Different components (the connection rebars, the lower fibres in the beam) are then strained as part of different regimes under dead and imposed loads. Analysis becomes rather complicated!
Pinned connections are designed primarily based on past experience (and previous test data), perhaps by using tabulated data for standard connections in the Green Book on Simple Connections (P-212). The shear resistance of such connections is explicitly considered, but all the requisite rotational characteristics (low strength, low stiffness, high ductility) are assumed to be implicitly achieved based on past experience. Whilst these characteristics are very important (excessive strength could lead to columns being overloaded, lack of ductility could lead to connection failure at low load), there is nothing to suggest this reliance on past experience is inadequate. However, it is precisely because of the importance of all the connection’s characteristics that designers should seek to understand the basics of how they perform. Whilst good practice detailing will normally ensure that ‘compatible’ characteristics are achieved (as noted above), there is a danger that this will not always be the case. This is a particular problem when more complex performance is needed. An
18
CHAPTER 6 Simple connections and basic fabrication Dave Chapman, Corus Construction and Industrial
Introduction Simple details and simplifying awkward details are the key to ensuring a successful project. It would be simplicity itself for the steelwork contractor to deliver plain beams to site just cut to length and leave the erectors to site weld the elements together! This happens on sites in countries where works facilities are rare and the weather is fine, but for UK construction the objective of a design engineer is to apportion the tasks in a way that all personnel in the chain here work efficiently. Another factor to be considered is the machinery available to a particular fabrication shop and therefore choosing details to suit. For example bolted cleats may be favoured if an automated drill line and angle line are available, and fully welded trusses should be avoided if a limited lay down area is present.
Circular cold saw
There are good publications on the design of connections, however there are many practical aspects beyond calculation, which an engineer will learn through experience or adversity. The items below are written with an engineer in mind who has recently started designing steelwork.
Basic fabrication machinery Available equipment
Drill line
The tools and machines owned by the various steelwork contractors in this country vary widely from the largest with CNC fully automated everything to the smallest just with a saw, welder and portable magnetic drill. The medium sized steelwork contractors (50-300 tonnes/week) complete the majority of structures and the following equipment is likely to be available to them.
are available, but here it is efficient if a standard 22mm diameter hole for an M20 bolt is used throughout to avoid downtime to change the drill bits. The drills only operate at a certain point on the drill bed and therefore offer maximum economy if the holes in the flanges and web are in line, allowing all three drills to work at the same time. If drills are liquid or air-cooled this allows an increase of three times the spindle speed. These two items combined in a fully automated machine comprise the ubiquitous "saw and drill line".
Processing main members Circular saws rotate slowly cutting away sizeable chips of steel rather than the small particles obtained from a hacksaw. This blade descends on the member, which is clamped in the hydraulic jaws. The saw is accurate to a fraction of a millimetre for length, and within 0.2% of the depth of the cut for square. The blade speed adjusts itself automatically on its way through the work piece dependent on the thickness of material. An option is a swivelling saw bed that allows bevel cuts of up to 45º. Rotating the saw to form bevel cuts takes time and the detail should take into account whether it is easier to form the angle required with the fitting or incoming member.
The members still may require coping or notching. Notching is often still done by the traditional method of marking out then cutting with the "gas axe" (oxy-propane or oxy-acetylene torch). Occasionally a CNC controlled coping machine capable of cutting in three dimensions will be available.
Band saws are used and although more suited to cutting smaller members and tubes are becoming economic for midsized profiles. Three axis drills are CNC controlled and include a measuring head. The piece is clamped between rollers that move the piece along the conveyors. Multi-head and automatic drill changers
Manual notching
19
Automatic coping
Fittings
Angle line
Processing fittings
The angle line is a very versatile and swift tool being able to punch, crop and mark angles and flats up to 150mm x 150mm x 15mm. The illustration shows a component that has been punched and marked prior to cropping. Punching holes in steelwork is much faster, and therefore less costly, than drilling; its use, however, is generally limited to predominantly statically loaded structures with limited thickness, or to secondary members. The maximum thickness where punching is applicable depends on the material grade and quality. Standard end plates, angle cleats, fin plates, stiffeners and flat bracing plus many other items can be formed using the angle line.
Alongside the processing of the main members, the fittings will be made to the draughtsman’s details. Fittings can be end plates, base plates, truss gussets, purlin cleats, bracing tongues and caps. These are formed from plate, flat or short lengths of section. Wide plate is quickly cropped on a guillotine to make floor plate, large gussets, or a required width of flat. The plate is marked, clamped and sheared. However at a thickness of 25mm some distortion of the cut edge may become evident. A flat bar processor drills, marks and burns large plates – typically up to 600mm wide x 50mm thick with 40mm diameter holes. This machine is ideal for baseplates, splice plates and large end plates. As well as rectangular plates, complex plate profiles can be produced.
Assembly The fittings and members are brought together for shop bolting occasionally or more usually "MIG welding" (strictly speaking this is MAG welding). This is still conducted manually except for the most advanced production of plate girders. Welding is a complex process of using an electric arc to generate heat to melt the parent material in the joint. A separate filler material supplied as a consumable wire electrode also melts and combines with the parent material to form a weld pool. The weld pool is susceptible to atmospheric contamination and is shielded by gas during the critical liquid to solid freezing state. The wire electrode is fed from a spool, but stick welding is used for awkward locations and on site.
Plate guillotine
Flat bar processor
Manual welding
20
Column bases and holding down bolts In an office building frame where the front, side and internal columns may be at different orientations and formed from UBs or UCs the temptation is to use the minimum size base plates resulting in rectangular plates. A frequent call from site is that bolts have been cast at 90º to the direction required, resulting in more work for the design engineer. To avoid this potential risk all bases and bolt setting should be square and use easy dimensions to aid the groundworker. If a portal or gable column forms part of the office area a similar bolt layout can be employed to avoid differences or eccentric centre lines. This is an example of the draughtsman considering "buildability" by making life easier on site.
Square base plate
Base plate eccentric to common holding down bolts
End plated beam
Holes in beam web
Fin plate on beam
Bolting access
Beams to flange and web
Rectangular bracing gusset
To help ensure the correct holding down (HD) bolts end up securing the correct column only one diameter should be used for a certain grade and length. For example, do not specify M24 grade 8.8 x 600mm long and M24 grade 4.6 x 450mm long on the same site, as once installed they look the same. There is a wide range of bolts and anchorages existing, the common choices are shown in this table.
Beam connections There is a choice of two common forms of beam-to-beam and beam-to-column connections; end plates and fin plates. An end plate is welded to the end of a beam and then bolted to a beam or column. A similar beam is drilled and bolted to a fin plate welded to the supporting member. A fin plate is preferred in many locations due to speed and ease of erection. If a primary beam has incoming fin plated secondary beams supported by it, all of the welding is on the primary beam, which allows the secondary beams to flow straight to the paint shop and thus reduce handling costs. Access for bolting beams to small columns, especially if they are moment joints, needs to be carefully considered. A subsequently erected beam must avoid the bolts already present, and the first erected beam must not prevent fixing of the next beam. In some cases a larger or different shape column will provide a better overall solution. The problem of bolt and tool access is not restricted to columns as it also applies to the top of steep apex joints and trusses/splices where large air powered socket drivers are used to tighten the bolts. As seen in the illustration, the endplate does not need to be the full width of the flange and this is also true for moment connections.
Preferred sizes of holding down bolts and anchor plates Diameter HD bolts
Anchor plates (mm)
M20 Length (mm)
M24
M30
300 375
375
450
450
450
600
600
Size (mm)
100 x 100
120 x 120
150 x 150
Thickness (for 4.6 bolts)
12
15
20
Thickness (for 8.8 bolts)
15
20
25
Standard lengths of HD bolts for each diameter are shown in bold type
21
Vertical braced column
Column splice
Flat bracing
After some design practice on similar joints it will become clear that for a given load the same group of bolts, weld and plate will give a balanced answer where each element is working equally hard.
Cantilever eaves
dictate the use of 2 or 3 storey shafts. There are two types of splice. The first is "direct bearing" where the upper shaft stands on the top of the lower one. This style is ideally suited to very heavily loaded columns because when fitted on site any small gap between the two pieces due to lack of fit would require filling. This may appear the simplest and most logical method but the real risk of gaps encourages the use of the second type "non-bearing" or "bolt bearing".
Bracing connections "K" bracing resists compression and tension forces and is frequently formed by using circular or square hollow sections.
In non-bearing splices a distinct gap between columns is detailed and the loads are transferred from one stack to the other through the bolts and splice plates only. This form is by far the most common until the number of bolts and size of splice plates becomes very large when bearing splices have to be employed.
Computer detailed gusset plates are most simply drawn using a rectangular plate. It is possible that, although all the member centroids node (i.e. meet at a single point), the line of action passes through thin air and is not coincident with the centre of the weld group. This eccentricity produces a moment in the weld group that may result in the originally designed 6mm fillet weld needing to be 8 or 10mm.
The illustration shows a splice using bolts that have been countersunk into the splice plates to avoid the protruding bolt heads. This is a bearing splice using a thick division plate allowing the spread of load from the smaller section above to the resisting lower flanges. The bolts present may be resisting a moment or providing a tensile resistance from forces to avoid progressive collapse. For splice design forces, a critical condition is the temporary condition of a high level column under wind load before any beams are attached.
Although a continuation of a similar rectangular plate to the tongue of the bracing is the best use of material, a vertical side does assist accurate placement of the part. In braced bays it is prudent to use end plates as these typically have greater vertical capacity than fin plates and the shear in the beam due to wind can be greater than the gravity loads. Crossed flat bracing works in tension and is weak horizontally and can be moved by hand on site leading to allegations of lack of fit. This is not the case as a 15mm movement in a 6m flat is equivalent to less than 0.04mm change in length. However, to ensure a taut bracing system the flats are drawn a few millimetres short and have to be fought into place, or a simple tensioning system can be added to one end. Flats can be full length but reducing their length with a central plate aids erection if they become unwieldy.
In a bolt-bearing splice a further complication would be to run the splice plates inside the flanges and save the thickness of the plate from the overall width. If this flange were then countersunk the column would require checking for the net section. Countersinking bolts is expensive and time consuming. If we consider a similar modest arrangement to that in the photo where 32 regular M24 grade 8.8 bolts are required, the bearing capacity in a 12mm thick flange is 132kN and for a countersunk bolt 66kN. After a column passes through the drill line countersinking is often completed manually, taking about 10 minutes per hole including handling, at a workshop rate of say £40/hour (still cheaper than many car main dealerships). Is the extra cost of a few hundred pounds worth it to save 32mm (2x16)?
Column splices Column splices are required in buildings over 4 or 5 storeys to keep the lengths transportable and to break the structure into manageable phases. Also stability during erection may
Bolt diameter (mm)
Bolt head depth (mm)
Typical comparative cost of bolts per 100 Grade 8.8
Countersunk
M20
14
£24
£72
M24
16
£48
£144
M30
20
£160
£400
22
Cantilever floor
Truss noded at column
Truss noded at end plate
Cantilevers
is already short it may not be able to be sawn automatically due to roller infeed spacing and other reasons. It could be cut manually by a saw or more likely with the oxy-gas burner during fit up. The fit up of the internals will be awkward as every internal fixes to its adjacent members, leaving no tolerance for adjustment. This operation is made more awkward if the booms are precambered. After fit up, the welding is made harder because of the joint on the overlap.
Shallow bolted connections for cantilevers inevitably rotate a small amount; therefore as a minimum the tip should be preset upwards by the deflection under dead load. For cantilevers and members subject to torsion, stiffness and depth are much more important properties than strength as these members will exhibit significant deflection and movement. One easy solution is to bury the cantilever beam within the floor or roof structure running it over the support.
If however between the internals a small gap of about 2540mm dependent on weld size is developed, these problems can be eliminated. This change must be conducted with the approval of the member designers and the induced local effects checked by them or the joint designer. Even if a boom size has to change or the internals made narrower as viewed, the changes are still very likely to outweigh the cost of partial overlap joints. BS 5950-1 states "eccentricity should be introduced" for hollow sections to give a gap or allow adequate overlap to transfer the forces involved.
Truss Details If an economic member design of a truss has been completed independently it is a fair certainty that the ensuing connection design will take longer, produce more calculations and may require strengthening of local joint zones in the main members. We are taught member centroids should coincide to ensure the forces resolve to zero at one point, unfortunately this avoidance of eccentricity can cause more problems than it solves. The illustrations show two arrangements of a truss top boom and the last diagonal fixing to a column. The fully noded example will experience horizontal shear in the region indicated along with local tension at the top and bearing/buckling below. This may not be a problem for a portal column but for a 203 or 254UC in simple construction it will. The example noding on the end plate has the vertical shear resisted at the same place, which will also be applied to determine a nominal moment based upon an eccentricity of 100mm from the column flange according to BS 5950-1 cl.4.7.7. The big advantage is the horizontal forces are resolved within the truss. If insufficient contact is present between the boom and diagonal elements a triangular infill plate can be attached to transfer the horizontal shear. This also applies to tied portal frames and bracing structures.
If the position of the boom splices can be determined, a pin can be introduced into the computer model to make the splice design more economic by removing any bi-axial moments. Similarly it is reasonable to pin the internal joints as they are often welded to flanges of UCs or side walls of hollow sections which are flexible. The analysis will change a minute amount from the fully fixed case but it may be a better model and will certainly aid economy by reducing the amount of connection design and welding. A bottom boom made of a UC on its side may be a suitable section especially when subject to wind suction on the roof when the member may be in compression. If it is in compression, the truss restrains its weak axis vertically and the horizontal strong axis may only need one or two bottom boom restraints to be stable over the full span. Hollow section internals are efficient sections and may work well. While this is true, a real problem may occur when directly welding internal hollow sections to UC booms. Stiffened gussets may have to be welded to the boom first and then the tubes slotted and welded to the gussets. This becomes less economic than turning the UC boom upright and adding a few extra bottom boom restraints.
Some re-noding can significantly increase the fabrication speed of trusses. One of the fully noded internals on the bottom boom requires a second cut at both ends, because it
Truss boom gap
Truss boom noded
Truss boom welding
Lintel welding
23
Fabrication details Following the calculation design rules a perfectly good fitting can be drawn and made, but plates should either be fully symmetrical plates or obviously asymmetrical to help the plater.
Stitch welding
Individual tubular bracing members in areas such as a roof bracing system may all appear the same, but small offsets and varying situations can mean the lengths vary by just a slight amount. To allow batch sawing of members and easy location on site, the draughtsman can adjust gussets and setting out points to produce as many identical items as possible.
Asymmetric plates
Adequate restraint is often present for truss booms, thus they are often selected to work very near their maximum axial capacity, which is fine for the member design. When a series of holes is taken out of a boom at a splice location there may not be enough steel left to take the load. Consider that in a splice 6 M24s (2 in each flange and 2 in the web) represent a length of 6x26=156mm which is a significant portion of any member. Thankfully BS 5950-1 cl. 3.4.3 allows an effective increase in net area of 20% for S275 and 10% for S355.
Within a project where S355 material is specified, using S275 fittings material is perfectly acceptable. Mitres at corners should be avoided to simplify the fabrication. Where the tips of two members are unsupported, the stiffer member should be run through as a cantilever to ensure the other has simple pinned ends.
Erection details
If the allowed increase in effective area is not successful, then cl. 3.4.4.3 gives guidance on staggered holes but this leads to longer plates and increased complication for drawing, manufacture and erection. In conclusion, if the splice were anticipated to be in an area of high stress, a check that a bolted splice is suitable would be prudent by the member designer.
The designer can aid the safe erection of elements by thoughtful detailing. Where two beams join to a web using common bolts, additional erection bolts should be added to one beam to avoid the need for a triple-ply fit up. Where the erector is working out of an access platform, spare holes can be positioned local to joints to allow a harness clip to be easily secured. If the final structure requires a column to be supported by a beam, the temporary case is often overlooked. The supporting beam needs to be wide flanged and it helps considerably if a further beam frames at right angles into its web to provide a stable base for erection. Although this may mean additional steel the design avoids risks involved in installing, using and removing of temporary guy lines.
Weld details Welding produces a massive amount of heat to produce molten metal, and when it cools the weld area contracts and can distort the member. This deformation is usually too small to notice, but can be considerable if the size and length of welding is over-specified or conducted without using proper procedures. For example, consider the lintel illustrated, with a plate to the bottom that supports the outer leaf of masonry. The short plate may be cheaper but the section will tend to contract to the left and exhibit a bow to the right and thus may fail to provide adequate bearing for the straight outer leaf. A wider plate allows balanced symmetrical welding to avoid horizontal deformation, however the member will try to deform upwards by a small amount. Although not calculable this can be considered a beneficial pre-cambering for the applied load from above. For economy and reduction of deformation, intermittent or stitch welding is usefully employed where gaps are left in long welds. Stitch welding is used to secure compound sections, portal frame haunches, shelf angles on beams plus many others (see BS 5950-1 cl. 6.7.2.5). As discussed above over-specification of welds can be a problem and very costly. Full strength full penetration butt welds are not required in the majority of structural cases. Partial penetration butt welds, either with or without overlaid fillet welds, are often more suitable. If a full 45º 8mm fillet weld is required to an end plate upstand above the beam flange, the upstand must clearly be more than 8mm to allow for tolerance. Also, with a bare 8mm upstand, the natural tendency is for some welders to deposit more weld on the larger surface and thus create a non-symmetrical and slightly weaker fillet weld.
Web erection bolts
24
Single bolt fixings may have adequate strength but with two bolt holes there is a real erection advantage. One hole is used, along with a podger spanner to line up the other for bolt insertion. Single bolt fixings can be used for small sections such as purlin and rail stays but for larger members an additional podger hole is better. The keenest purlin design may require two anti-sag bars to stabilise the member under wind reversal, but one (or fewer) anti-sag bars and a larger purlin may be prudent to reduce the erection time and risk when fixing sag bars at height.
Conclusion It is possible to design every connection individually for a one off specific economy, but consistency of style throughout the project for varying member sizes used in a certain manner is important to allow the draughtsman to follow the logic of the calculations, as well as ensuring an easier design check. Nowadays with 3D solid modelling anything can be drawn and plasma plate profilers can produce any shape, however it is still good practice to keep bolt layouts, plate shapes and flat widths simple. All steelwork contractors will have their own company standard fittings, dimensions, cross-centres, set-outs, welding and plating preferences which should be accommodated, as far as possible, to maintain their common overall work flow across other projects through the drawing office, workshop and site.
25
CHAPTER 7 Connection costs
Keep it simple! There are a great many connection forms, but it does help to rationalise these into a few groups in order to understand where costs are being expended. This could be seen as being a gross over simplification, but we have witnessed many times needlessly over-complicated connections, simply because there has been no clear understanding of bracing or stiffness requirements, unworkable geometry, aesthetic constraints which have little or no consideration for the forces involved, impractical (usually too small) member sizes, and no appreciation or understanding of fabrication practises. A golden rule must be – keep it simple! In this context simple does not mean ‘pinned’ and complicated ‘fixed or moment’, but correctly proportioned, un-notched/stiffened, and based upon recognised standards like the BCSA/SCI Green Books, using standard achievable bolt sizes and spacings etc. It really cannot be stressed enough; the strategy with connections must be simplicity, symmetry, and sensible rationalisation (acknowledging haunched moment connections will invariably not be symmetrical).
Kim Dando, Frank H Dale Limited
Introduction Many years ago when considering how to incorporate a then, ‘state of the art’ 3D modelling package into a steelwork contractor's drawing office, it became apparent that structural steelwork fabrication is only about connections. What do we do? We take plain lengths of steel and add connections to them. Of course we paint and erect structural steelwork as well, but those are not our core activities and in turn can be heavily influenced by the manner in which the connections are designed and detailed. The UK fabrication industry is a world leader, and that is largely about economy and efficiency. The design and fabrication of connections is therefore a fundamental part in establishing and maintaining this position [Previous studies have suggested connections can influence between 40 and 60% of the total frame cost, a similar proportion to the cost of steel itself]. Recent significant raw material increases will have had an impact on the consideration of efficient use of steel balanced with minimum workshop effort. In other words minimising the use of expensive steel is commendable, but not at the expense of increasing labour and workshop costs – a fine balance. Taking a controversial line for a moment, it could also be suggested that insufficient consideration is generally given to connection design at the main project design stage, which at times appears to be completely divorced from the main member design process. Admittedly, for most architecturally sensitive projects and the highly engineered super projects like T5 and Swiss Re (see page 110), the connection design is clearly an integral part of even the concept design.
CAD drawing of simple connection
This criticism is aimed at the much more straightforward, less high profile projects which form the bulk of our steel fabrication production – single storey, low/medium rise and even some high rise structures. Advancement in computer technology has driven us forwards and backwards at the same time, 3D modelling has become commonplace in the fabrication world, and has been for some years, but now it is making serious inroads into the design process, and so it should. However, how many times does the design process stop at member end fixity definition and member design, with the steelwork contractor taking over often unnecessarily complicated member connections, making no allowance for beam eccentricities and actual member sizes? The days of the poor steelwork contractor being stuck with connections which are impossible to develop or massively expensive to fabricate at his expense are largely gone. We are older and wiser and qualify our bids according to the perceived complexity, or lack of information available. This, of course, protects our interests but does nothing to advance the quality of steel framed projects or reduce costs, and the problem could so easily be addressed!
Connection groups It is customary to group connections in terms of their structural performance. Two broad classifications as mentioned earlier will be ‘Pinned’ and ‘Moment’ connections. We hesitate to use the word ‘fixed’ from ‘fixed ended’ because this suggests a totally rigid connection with no rotation – rarely, if ever achieved. This classification can be applied to both beam end connections and column bases. It must be understood however that as well as not normally achieving full fixity, in most structures true ‘pinned’ connections are also not achieved, hence the concept of flexible end plates. Complicating the situation further we have conditions like ‘nominally pinned’ and ‘semi rigid’ all introducing different degrees of stiffness and rotational limitations on the connection form. Generally speaking a designer should define either from experience, or a particular structural requirement at an early stage in the design process, which ‘fixity’ classification will be adopted for the connection. This is an important choice, and if not given adequate consideration will at best give an unnecessarily expensive solution, and at worst a seriously flawed design. It is very important not to apply fixity to all joints in the analysis model on the premise that members will have minimum size and cost. This will not be the case, as I’m sure you will have heard before, ‘minimum weight does not mean minimum cost’.
26
Comparative costs
Having made this point, we must assume that these choices are undertaken correctly and concentrate more on considering connections in terms of their cost of manufacture and ease of erection. [Design for Manufacture Guidelines - SCI publication P-150 has some useful information giving relative load capacities of major connection types, shear and moment, as well as information on basic workshop practises and general cost centres].
Introduction The table below shows comparative costs for the 19 connection types described further below. The costs are based on comparable capacities across types 1 to 12, but types 13 to 16 would have significantly different capacities to types 1 to 12.
Fin plates (types 1, 5) This is a good general-purpose connection particularly suited to beam-to-beam connections where secondary beams can be detailed with no welding i.e. bolt holes (and notches) only. Primary beams can also adopt fin plates but as they are usually connected to columns flexible end plates may be a more appropriate condition, especially if it is required to introduce some rotational stiffness to aid erection, or some additional shear capacity. CAD drawing of moment connection
Fin plate
Type
Description
Material
Fabrication
Erection
Total
Beam to beam connections 1
Fin plate notched
2
2
1
5
2
Angle cleats notched
2
2
2
6
3
Flexible end plate notched
1
1
2
4
4
Blind fixed
3
2
3
8
Beam to column connections 5
Fin plate
2
1
1
4
6
Angle cleats
2
1
2
5
7
Flexible end plate
1
1
1
3
8
Blind fixed
3
1
3
7
Tube bracing connections 9
Tee
2
2
2
6
10
Slotted tube
2
3
3
8
Column splices 11
Bearing
3
2
4
9
12
Non-bearing
4
2
3
9
Moment connections (beam to column) 13
Unstiffened
3
3
3
9
14
With stiffeners
4
5
5
14
15
Haunch unstiffened
4
4
3
11
16
Haunch stiffened
5
6
4
15
Base plates 17
Pinned
2
2
2
6
18
Moment
5
4
3
12
19
Fire moment
4
3
2
9
27
Fin plates are very inexpensive to manufacture, with current automated machinery producing these in minutes. Assembly and welding is normally very accessible, and shaping or fitting of plates is not normally required. Beam-ends will require notching in certain configurations, the only real downside of this connection type.
Use a limited range of standard flats as shown in the Green Book, again not plate. Modern plate punching and cropping machinery also handles flats and can produce these fittings very quickly, so intelligent grouping of connections is worthwhile, especially when handling time changing different bars vastly exceeds the time taken to produce one or two components.
The use of ‘rolled Tees’ as bolted outstands should be avoided. Steel and manufacturing costs are high for these sections. If a bolted Tee section is to be used, fabricate it from standard flats (not plate).
There are two detailing points to consider. Firstly, always make the end plates symmetrical top to bottom so that they cannot be welded on upside down. Secondly, align the top edge of the plate with the top of the beam noting the current design guidance recommends not fully welding the plate to the flange – short tacks, however, should go unchallenged. This should potentially allow two components to be correctly assembled without actually ‘studying’ the detail drawing. In the case of notched beams, align the plate with the top of the notch.
Most erectors welcome the use of fin plate connections as the beams are normally short and therefore easy to position and also partly self-locating, they feed along and can rest on the outstanding fin plate and only a few connection bolts are required in most instances before unhooking the crane. One detail point to watch, to negate erection nightmares, is to ensure the beams are always positioned on the ‘leading’ side of the joint, especially when there are beam-to-beam and notched conditions. Also stiffeners in columns hinder all forms of beam erection, and can cause significant difficulties in multi storey situations where beams are expected to ‘drop down’ columns.
Angle cleats (types 2, 6) In many ways double or single angle web cleats offer similar qualities to fin plate connections, but may have lost some popularity these days mainly because of higher steel costs and time spent on or off site assembling the connection prior to erection.
Whilst it is not conventional to design for any end fixity, fin plates can sustain (very) small moments which may be used to good effect when tweaking mid-span moments to justify or utilise a particular beam section. In addition they are quite efficient and easy to analyse when designing for tension forces (disproportionate collapse).
Seating cleats or shear plates used in conjunction with web cleats can be very efficient in certain circumstances and provide fast ‘off hook’ capabilities when lifting beams into position, and can be used with most connection types.
Particularly with fin plate connections, especially double row bolted connections, check the design span of the beam. Can it be shortened because the beam is spanning between column flanges, not centre lines? Where is the actual point of support? Is there any advantage in moving it from the end of the supported beam to a point within the connection? Normally this would not be recommended because of torsion and other connection design complications, but once in a while it does pay off, particularly with large runs of identical members.
Flexible end plates (types 3, 7) Welded angle cleats
This is another good general-purpose connection, probably being one of the most widely used (and abused) forms. Unfortunately there is a tendency with increased rationalisation to overlook the fact that the end plates must remain flexible, so standardising on 12mm flat is not the way forward.
Bracing connections (types 9, 10) These largely fall into two camps, tubular brace connections and all others. The intention is not to dwell too much on the ‘others’, which can take virtually any form from simple angle and flat tension bracings, right up to very specialist highly loaded complex three-dimensional assemblies. One point to mention with all bracing joints is to check load paths carefully especially, with highly loaded members. Designers can be quite relaxed about spreading joints to get members to fit, or quite pedantic about members intersecting on their neutral axis. Both are correct of course provided eccentricities are accounted for in the main member designs, and an experienced designer will appreciate the real difficulties of actually nesting these members. One detail, which crops up time and time again in the ‘shed’ market particularly, is
Flexible end plate
28
Tee connector
stopping vertical bracings above finished floor level. This can be anything up to a metre away from base level (even more in dock leveller conditions) – a fair eccentricity with a highly loaded vertical brace onto the minor axis of a column. This is a correct and proper detail from a construction viewpoint, but often not considered. Returning to tubular bracings there are many ways to connect tubes to sections, the two main forms being ‘T’ connectors and slotted/bird-mouthed ends. ‘T’ connectors are the most straightforward to fabricate, but have limitations on their load carrying capacity. This can be improved by not using rolled Tees, but fabricated Tees from flat. Even then the design of the joint can be quite intricate, especially when considering deformation of the flat under load welded to the end of the tube, and resulting weld stresses concentrated around the stalk. This is largely why the slotted/inset version is utilised, the design process being much more straightforward, but the fabrication costs are significantly greater. One easy trap to fall into is to forget the tube has to cope with compression, which means the ‘T’ stalk is acting as a strut, and needs to be designed as such. Reinforcing ‘gluts’ on alternate sides of the joint may be required to reduce the flat thickness to acceptable dimensions.
Column splice
automated machinery, the final choice of bearing or non-bearing types tends to be company specific rather than project specific, although splicing together dissimilar sections will probably sway more towards bearing splices, if only to cut down on time in manufacturing and erecting long and heavy packing plates. It may also be reasonable to assume that an erector would prefer nonbearing splices because, although there are more bolts to place, the connection is initially more ‘flexible’, allowing bolts to fit better. In addition column alignment (or even misalignment, pulling columns apart in the upper lifts to allow for easier high level beam placement), has its advantages. From a design point of view, the increasing importance of disproportionate collapse introducing tension into the splice, also probably promotes the greater use of non-bearing splices these days.
Beam splices tend to be at positions of higher moment. This can fundamentally change the detail away from flange and web plates towards bolted end plates, which may or may not be preset to account for end plate misalignment, connection slippage and beam deflection. As always, a balance needs to be struck between locating the joint in the centre of the beam (high forces/expensive connection), but probably more economical steel purchasing and no long lengths to worry about, against an asymmetric arrangement. On paper a one third/two thirds arrangement will always score better, but in practice things may be different.
Fully welded tubular assemblies are required on certain projects. These are best suited to specialist steelwork contractors in terms of both design and fabrication capability. Needless to say they can carry ‘mind blowingly’ high fabrication costs as a result of complex setting out, machining, jigging, assembling, welding, inspecting and testing, handling and painting, and inefficient transportation. This is then followed by different but equally problematic (which means expensive) site issues. Finally the welded assembly will have built-in two sets of tolerances, one from fabrication and the other from erection. The fit-up process will have to cope with these.
With ‘cranked’ members, rather than butt weld two sections together, consider introducing welded or bolted division plates.
Moment connections (types 13, 14, 15, 16) Moment connections are significantly different from all other connection types particularly with regard to design and fabrication. Firstly, it is very difficult to ‘guess’ a moment connection. In this context guess is typically to call upon years of design experience and try and correctly proportion the connection. With other connection types bolt shear and tension capacities can be broadly remembered, weld lengths and leg lengths have known capacities, web shear problems can usually be spotted, and end plate thicknesses can be recalled. With little consideration, a workable detail can usually be put together without too much calculation, to be verified as necessary at a later date. The same is not true of a highly loaded moment connection. Also, in the portal frame ‘shed’ market the connection design becomes an integral part of the main frame analysis.
Splice connections (types 11, 12) These are normally associated with joints in columns, but of course can apply when joining any member within its length, usually limited by transportation or site access difficulties, or even steel availability issues. Considering simple column splices initially, these are primarily designed for axial load, but may also be subject to shear and moment forces which in the worst cases are bi-axial. Design guides are available for both bearing and non-bearing conditions. Assuming manufacturing of components is by
29
In wind moment designs great care has to be taken to demonstrate rotation capacity, which in turn depends on deformation capacity. In other words, we are designing the connections to deform. How does this affect manufacturing costs? Standards are quite difficult to adopt. The Green Book virtually stands alone in prescribing a set of moment connections. The principles are excellent (as far as they go), and certainly if the specific design condition is not tabulated, the methods adopted will likely allow a particular condition to be engineered, maybe with the exception of moment connections into column webs, arguably a significant omission.
Plates can be sheared, flame cut, plasma cut, punched, drilled or any combination of these operations, depending upon required quantities, material availability, and workshop machinery. It is also possible to ‘buy in’ pre-manufactured plates from a specialist, generally thicker and profiled. Reinforcing base plates, usually associated with moment bases, is very labour intensive as weld sizes are usually also large, so as a (very) general principle simple bases are preferred in the first instance. Fire bases (a form of moment base designed to cope with the perimeter boundary requirement in fire) are also generally not reinforced. Thankfully, it is very unusual these days to machine (for flatness) column base plates and column shafts, but this may be required in special circumstances. Pre-heating before welding may also be required with thicker material (in practice not usually below about 50mm thickness, but do check!). Putting aside moment bases, weld sizes should be kept to a practical minimum. Usually axial compression is transferred by direct bearing, shear is not normally a major factor, so uplift, especially on braced columns is the normal design criterion. Moment bases like all moment connections require careful design. Juggle the geometry to pare away at steel and weld sizes, and avoid any form of butt-welding if possible.
Haunched rafter
Unfortunately most moment connections are bespoke in some form or another, and this will require a specific hand or computer calculation. Work hard on the design and remove all stiffeners possible! Fitted stiffeners are hugely expensive, cap plates are not. Weld sizes are important, single passes are possible with leg lengths of 6 and 8mm, whereas legs of 10mm and upwards require multi-pass welds which treble or quadruple or more the welding times. Minimise haunch lengths where possible. Consider every weld, especially the many metres of weld between the haunch web and beam flange. Intermittent/staggered or one sided welding is possible, but may cause issues with painting and long term corrosion protection (exposed edges are notoriously difficult to paint adequately).
Blind fixing
Miscellaneous (special) connections These are usually driven by member constraints (connections to closed sections), unusual geometry, sound and thermal issues, differing materials, and demountability requirements.
Base plates (types 17, 18, 19) Base plates will generally be manufactured from ‘plate’, although there is no reason why narrower plates cannot be manufactured from standard flats. The amount of time taken to manufacture and assemble base plates, and attach them to column shafts is a function of the plate size and thickness, the amount of stiffening if any, and weld sizes.
The possibilities are virtually endless, but three or four relatively common situations have spawned solutions which are becoming widely adopted. • Blind fixings for closed sections eg tubes. These usually take the form of ‘drop’ bolts or ‘expanding’ anchors. The connected member is usually standard, but accuracy of the drilled hole diameters can be critical especially for expanding (hollow) bolts, and capacities are usually not up to grade 8.8 black bolt values. The fixings themselves can also be very expensive, so are to be used sparingly. • Another mechanism for connecting members to tubes is the ‘flowdrill’ method. This requires special ‘drilling’ equipment, so not all steelwork contractors will be geared up for this, but it does enable standard bolts to be used, and should not impact greatly on fabrication or erection periods.
Fire moment base
30
• Quicon®, a method of connecting components without using bolts but special studs and key-hole slotted connectors. Whilst not exclusively for demountable situations, it is obviously ideal in this condition. Fabrication costs will probably be higher than normal but the system scores on significantly reduced erection periods (costs).
bolted temporary joint as well – and that will generally need to be removed and 100% inspection and full treatment will the generally be needed.
• Finally it is worth remembering that steel is sometimes connected to other materials, usually concrete. In some respects the ideal solution is to supply and cast in (attach to formwork) a pre-engineered steel faceplate. This enables a steel-to-steel connection of some form to be adopted. Alternatively there are many cast-in Halfen® type fixings and ‘T’ bolts which allow adjustment during fixing. These can be significantly more expensive than conventional fixings, but again can be successful in reducing site durations in the right circumstances.
Do not specify or design connections using non-standard bolt diameters. These bolts are more expensive and on longer delivery lead times. Some years ago the industry standardised on a limited range of fully threaded 20mm diameter bolts, grade 8.8. This has worked well, but other standard diameters are available including 12, 16, 24 and 30mm diameters for the ‘rare cases’ when 20mm diameter bolts will not do! Similar conditions apply with holding down bolts and, if in doubt, a reputable bolt supplier will quickly advise which ‘standard’ lengths and diameters are stocked. Whilst on the subject of holding down bolts, it is worth pointing out that sometimes alternatives to cast-in bolts have to be used. Most alternatives (mechanical expanding fixings or resin type anchors) have significantly reduced load-carrying capacity in comparison with their conventional holding-down bolt counterpart; so direct substitution is rarely possible. Resin anchors in particular should be load tested after installation, prior to erection commencing, and to a given value. Moisture, dust and other factors can sometimes mean that pullout capacities are dangerously lower than expected, which can manifest itself during the erection of the steelwork with potentially serious consequences.
Bolting
General details Welding As mentioned earlier, adopting a MIG welding process on the shop floor usually means that 6mm and 8mm fillet welds can be executed quickly and efficiently. It also makes possible prefabrication blast cleaning, as slag and weld spatter is minimal. Welders get used to these sizes, and can produce them all day long with minimal problems and consistent sizing. Welds with longer leg lengths, however, slow the process down, requiring multi passes, greater operator skill, and enhanced inspection and testing. Very large welds, often associated with thicker steel, may require pre-heating as part of the process. This again will have time and cost implications.
HSFG bolts have particular performance-enhancing features, mainly in connections where any form of ‘slip’ or movement would be unacceptable. However, their application is limited because of additional fabrication controls (masking connections from paint etc), additional erection procedures (torque control and checking), and higher unit price.
Butt welding also requires greater operator skills, and usually the steelwork contractor will step up the inspection routine which may sometimes include 100% NDT testing of these welds, again incurring time and cost. Butt welds often do not need to be full strength, and partial penetration welds, especially on a repeating detail, may be cost effective.
Tension Control Bolts are in many ways similar to HSFG bolts, but offer improved site performance capabilities (see SCI P334). [Remember to use some form of locking device on nuts where either movement or cyclic loading of a joint is expected, or where members are suspended].
Site welding to most steelwork contractors is like ‘holy water to a vampire’. If operatives are taken away from their normal workplace and environment, then obviously efficiencies will drop. There are additional hazards with site welding, both for site operatives and protection of the more combustible parts of the site. Weather will also play a part. A guaranteed quality is also more difficult to achieve, usually resulting in enhanced inspection procedures. There are certain conditions however when site welding is the right way forwards, and if undertaken by specialist operatives or contractors, can be surprisingly quick and efficient to execute. Consider each condition carefully, and if it can be avoided then do so. If it cannot, then recognise this fact and implement a configuration and process which can be undertaken as efficiently as possible, on site.
Punching The use of punched holes is a long-established means of forming holes in flats and angles. Modern punches are only limited by design constraints imposed on connections, moment connections in particular. The manufacturing goal posts have
Complex geometrical welded arrangements carry a further problem, both in the shop and on site, in that it is often impossible to rotate the assembly to give the welder the best position, meaning less than ideal ‘positional’ welds are required, requiring more expertise, inspection and therefore cost. Furthermore, accurate and secure location of two pieces to be joined by site welding will often involve quite a significant
Punching machine
31
moved, and it is recognised that punching 26mm diameter holes in 25mm thick steel is readily achievable. Interestingly, beam-punching equipment is not so commonplace in the UK unlike America. Punching standard slots is also efficient, but limits the length of the slot. Multi-punching to form a long slot is to be avoided if possible, but often crops up when fixing brickwork restraint members which are adjustable for brick placement and coursing.
cheap, fabrication and erection are not, and the correction of a poorly engineered or incorrect detail on site will be a designer's nightmare. Connection design as suggested earlier can be seriously under-considered in the project design process. On certain types of work this design can be more complicated and time consuming than the main structure design. When undertaken correctly it brings into consideration all aspects of the steel construction process, which carries its own rewards when undertaken correctly, and increases costs at an alarming rate when not.
Steel Grades It should be pointed out that plate and fittings material generally will not be in higher-grade steel for any connection type unless there are specific project constraints (irrespective of the main section grade). Stocking and availability aspects far outweigh any benefits usually derived from using higher-grade fittings. Connection designers must obviously bear this point in mind. The welding of varying grade steel sections at any one time usually means that welding wire or electrodes are normally chosen to suit the higher grade material, rather than continually changing to suit higher and lower grade steels. This will arguably introduce a small extra measure of safety with the weld design (not to be taken into consideration), accepting that the weld does become parent metal at some point. This is in addition to the fact that the code safety factor for weld sizing etc is greater than that for main material as well.
Geometry Consider the assembly of the joint on site. Can the bolts be positioned from either direction? Is there enough room to turn a spanner? Can a socket be fitted over the bolt head (surprisingly large for a 30mm diameter bolt)? Will welds from stiffeners encroach upon the head or washer positions? Do not use different diameter bolts in the same connection. Avoid connections with bolts in two planes eg vertical and horizontal – invariably one set will not fit. Avoid single bolt connections (except knee stays) – the point of a ‘podger’ spanner is essential to locate the joint. Do not mix grades of the same diameter bolts. These points although seemingly minor can ‘make or break’ a project on site. The use of 3D modelling systems has enabled engineers and draughtsman to more ably appreciate the physical problems associated with site assembly. The use of inbuilt clash detection mechanisms and also the careful setting up of the connection macros, together with the facility to spin and view the joint from any perspective should in the hands of an experienced CAD operator reduce, and hopefully eliminate the instances of connections not being able to be assembled.
Conclusion It is hoped what is described here conveys a little of the mystery of what happens behind the ‘beam basher’s’ doors, and why connections cost money. It’s not a black art but, like so many engineering problems it’s about compromise, juggling the design and drawing, fabrication, and erection requirements, usually not satisfying any of them in an ideal manner, but recognising that design and drawing is relatively
32
CHAPTER 8 Structural fasteners
may coexist. This option will remain with the introduction of Eurocodes for steelwork. Progress was made when it was agreed to develop European product standards for all the major bolt solutions that already exist across Europe and not impose the bolts from one country on another. Therefore the preloaded bolt standard is likely to have at least 10 and perhaps even 11 or 12 parts when complete. At first glance some of these bolts will appear strange. A German engineer will find strange the European version of our common load-indicating washer (the BS 7644 Direct Tension Indicator) from the UK, whereas a British engineer will find strange the European version of the preloaded fit bolt with its actual diameter larger than its nominal diameter, although it is often used in German bridges.
Thomas Cosgrove, SCI
Introduction The launch of Eurocodes for steelwork is associated with the introduction of new European product standards for structural bolting. The purpose of this introduction is to provide a brief description to the background and major issues that are likely to be encountered in connection with structural bolting over the next few years in the UK. The bolt descriptions; ‘preloaded bolt’ for high strength friction grip bolts and ‘non preloaded bolt’ for ordinary or black bolts are used throughout. At the time of writing the European standards dealt with in this introduction are at various stages of development. Some have passed formal votes and are awaiting notification in the Official Journal before becoming BS ENs by the addition of national forewords where required. Others are at the prestandard or prEN stage and have been issued for public comment and enquiry; while others still are at the committee document stage. For the purpose of simplicity in this general introduction all European standards are referred to as EN documents irrespective of their official statues. Furthermore, at this time, many of these European standards are out for public comment and enquiry and thus they may change. As the purpose of this introduction is to give a general overview, the standards should be consulted when finally completed and issued for a detailed understanding of their provisions. In time, more detailed guidance will be available through the SCI and the BCSA.
In the long term the wider European market may choose to focus on one or other of these bolts – let the market decide – but in the immediate future it is likely that each country will start to use the European bolts or product standards that have been developed from their own national standards. This is to be encouraged for the foreseeable future until the Eurocodes have bedded down and experience has been gained in their use as well as with their supporting product and execution standards.
Current UK practice Since approximately 1970, the bolt standards used in the UK have remained largely unchanged although the steelwork design standards in that time have evolved from an allowable stress to an ultimate limit state approach. The main British Standards for bolts (BS 3692 and BS 4190 for non preloaded bolts, and BS 4395 for preloaded bolts installed to BS 4604) are tried and trusted. Around 1990 attempts were made to introduce other international standards in the UK however the market continued to specify non preloaded bolts to BS 3692 and in particular to BS 4190. That generation of BS EN standards was listed in the 4th edition of the National Structural Steelwork Specification and for grade 8.8 fasteners included non-preloaded bolts to BS EN ISO 4014 and 4017, and nuts to BS EN ISO 4032. The use of such BS EN ISO fasteners has been very rare as they have been more expensive than the equivalent fasteners to BS 4190.
Background The task of agreeing common European product standards for structural bolts has been very difficult and protracted. Recently however, much progress has been made and the first product standards (actually Parts of Standards) have been completed and will be available shortly. These standards will allow bolts to be specified that are suitable for steelwork designs to the Eurocodes. The relevant technical sub-committee is currently working on the outstanding standards which should be published over the next year or two. The major reason for the difficulties is that several of the larger European countries have their own successful fastener industries, dating back to the 19th century in most cases. Not surprisingly these industries, in conjunction with their steelwork sectors, have developed unique and successful but different solutions for structural fasteners, particularly for preloaded bolts. These solutions are based on assumptions of site control and supervision which vary from country to country. Preloaded (i.e. “pre-loadable”) bolts have a greater market share in certain European countries than in the UK because they are commonly used in a preloaded or non preloaded manner depending on the tightening specified for their installation. Although permitted by design and execution standards, this is contrary to long-standing British practice because it requires greater site supervision and control on projects where different connection types
Typical tension control bolt assembly (TCB)
33
The emergence of the tension control bolt (TCB) as a preloaded bolt in the last 10 years has been the one major development in the UK market since 1970. A typical TC bolt assembly is illustrated. Although the original Torshear bolt was developed in the UK many years ago, the modern TCB has been introduced into Britain from improved versions currently used in the USA and Japan. At least three variations are presently available in the UK. Although a British Standard has not been developed, the SCI has produced an industry standard for one of the types on the UK market (P-324).
Therefore in developing the European product standard (EN 14399) for preloaded bolts it was agreed to develop two general parallel systems, HR (British/French) and HV (German), which reflect the above two philosophies. Both systems have under head radius and thus require chamfered washers to be placed under the heads of the bolts. The HR system will have two grades, 8.8 and 10.9 while the HV system will only have one grade, 10.9. It is important to avoid mixing up the components of both systems. Therefore bolts and nuts for both systems are standardized in separate parts of the product standard EN 14399 and the marking of the same system is uniform. Bolts and nuts from the same system will be stamped with their system designation, HR or HV, in order to avoid confusion. In addition bolts and nuts will be stamped with their property class (i.e. grade), 8.8 or 10.9 for bolts and 8 or 10 for nuts as appropriate. For the HR system the following possibilities exist:
Typically the TC bolt head is round but may be hexagonal. Preloading is normally carried out by an electrical shear wrench at the threaded end of the bolt (i.e. the nut or spline end), and preloading is complete when the spline shears off. A further variation relies on paint inserts in the washer which squirts out bright coloured paint due to compression when the preload has been achieved. The spline does not shear off with this variation. Owing to their splines, TC Bolts are quite distinctive in appearance.
• Bolts to class HR 8.8 with nuts to class HR 8; or • Bolts to class HR 10.9 with nuts to class HR 10.
Existing European preloaded bolts
The HR 8.8 bolt is very similar to the Part 1 general grade HSFG to BS 4395 and likewise, the HR 10.9 bolt is very similar to the Part 2 higher grade HSFG bolt to BS 4395. Design standards will permit preloaded bolts to class 10.9 or 8.8 to carry applied tension.
Across Europe two different philosophies exist to achieve the necessary ductility in preloaded bolt/nut/washer assemblies. Both systems have long histories and are well proved. The British/French approach, BS 4395 for example, has been to use deep nuts and longer thread lengths in the bolt assemblies and thus obtain ductility predominantly by plastic elongation of the bolt. The longer threaded length is necessary to ensure that the induced strain is not too localised. In load tests these bolts tend to fail in a ductile manner by rupture in the shank of the bolt at the root of the thread – i.e. the bolt breaks in two. These bolts are generally insensitive to over tightening during preloading and thus require less site control. Furthermore, if severely overtightened during preloading, the ductile failure mode of the bolt assembly offers an indication of pending failure. BS 4395 bolts only require plain hardened washers. The fact that the thread may be locally subject to plastic strain during tightening means that such fasteners may not be re-used if removed.
Preloaded bolts to EN 14399 EN 14399 has the general title: ‘High-strength structural bolting assemblies for preloading’ and has been divided into the following parts: • EN 14399-1: General requirements • EN 14399-2: Suitability test for preloading • EN 14399-3: System HR – Hexagon bolt and nut assemblies • EN 14399-4: System HV – Hexagon bolt and nut assemblies
The German approach, using DIN 6914 bolts, has been to use shallower nuts and shorter thread lengths in the bolt assemblies and thus obtain ductility by plastic deformation of the engaged threads rather than in the threaded shank zone outside of the engaged length. In load tests these bolts tend to fail in a brittle manner by thread stripping – i.e. the nut flies off. Although the failure mode of the bolt assembly is brittle, it is argued that it is a more ductile failure mode for the steelwork connection because the bolt shank remains in place after failure and may act as a ‘peg’. These bolts are generally more sensitive to overtightening during preloading and thus require more site control. If severely overtightened during preloading the brittle failure mode of the bolt assembly offers little indication of pending failure. DIN 6914 bolts have an under head radius and therefore require a washer with an internal chamfer (and for this to be installed in the correct orientation if only chamfered one side). It is quite common to use DIN 6914 bolts in non preloaded situations.
• EN 14399-5: Plain washers • EN 14399-6: Plain chamfered washers • EN 14399-7: System HR – Countersunk head bolt and nut assemblies • EN 14399-8: System HV – Hexagon fit bolt and nut assemblies • EN 14399-9: Load indicating washers • EN 14399-10: System HRC – Bolt and nut assemblies with calibrated preload EN 14399-10 will cover TC bolts and it still has to be decided whether HV countersunk head bolts and HR fit bolts will be added to the series. The standard constantly emphasises that preloaded bolted assemblies are very sensitive to differences in manufacture and lubrication. It is therefore important that the assembly
34
is supplied by one manufacturer who is always responsible for the function of the assembly. For the same reason it is important that hot dip galvanizing is under the control of one manufacture. This is reinforced by the arrangements for CE marking as this relies on testing of complete fastener assemblies to EN 14399-2. Bolts and nuts cannot be CE marked separately.
Suitability test In view of the differences in backgrounds and philosophies between the HR and HV systems it was decided to develop a suitability test (EN 14399-2) as part of the European product standard for preloaded bolts. EN 14399-2 states that the purpose of this test is to check the behaviour of the fastener assembly so as to ensure that the required preload can be reliably obtained by the tightening methods specified in the execution standard with sufficient margins against overtightening and against failure. The tests shall be carried out on test assemblies in the condition of delivery without alteration of the lubrication of the various components. The manufacture has to test 5 bolt assemblies from a single assembly lot or extended assembly lot (see EN 14399-1 regarding assembly lots). If the bolts delivered to site have had their surface condition altered for whatever reason, they should be retested.
Idealised torque/bolt force curve
The various parts of EN 14399 specify minimum angles ∆_2 for the HR and HV systems depending on the length of the bolt. This is the angle of rotation between the nut and bolt after the specified preload has been achieved and before the bolt force falls back to the specified preload again. The test may be stopped if the angle ∆_2 has been realised and the bolt force is still greater than the specified preload. In addition EN 14399 requires that the maximum bolt force recorded in the suitability test must be greater than 0.9 fubAs. The angle ∆_2 and the limitation on the maximum bolt force ensure sufficient margin against failure from overtightening when the tightening methods permitted in the execution standard are employed. The torque/bolt force curve or an analysis based on the data necessary to produce it will only be required if torque is used in the tightening method employed on site.
Section 5 in EN 14399-2 states that the principle of the test is to tighten the bolt assembly and to measure, during tightening, the following parameters: • the bolt force; • the relative rotation between the nut and the bolt; • the torque, if required; • the bolt elongation, if required. From the recorded data several graphs may be plotted for each test but the two most likely to be used are the rotation test curve and the torque test curve (see illustrations).
Installation EN 14399-1 requires that at the time of ordering the manufacture shall obtain the specified k-class of the bolt assembly. According to EN 14399 a bolt assembly may have one of three k-classes; K0, K1 and K2, and if no k-class is specified, k-class K0 applies. The k-class is related to how the coefficient of friction or k-factor, ki, of the bolt assembly is calculated from five specimens subjected to the torque suitability test. The three k-classes are as follows: • K0 – No requirement for k-factor; • K1 – Individual test value ki; • K2 – Mean test value km and coefficient of variation of kfactor Vk.
Idealized rotation/bolt force curve
The k-class of the bolt assembly is integral to the tightening method employed on site and the data required to calculate the torque, if required for the tightening method, needs to be obtained from k-classes K1 or K2. If torque is not part of the tightening method employed, as with load indicating washers
35
Non-preloaded bolts to EN 15048
for example and probably TC bolts, then k-class K0 may be specified. If k-class K0 is specified then the bolt manufacture need only carry out the rotation suitability.
There are many more types or variations of non-preloaded bolts across Europe than preloaded bolts although their technical differences are not as large as that of preloaded bolts. Therefore a much simpler product standard is being developed for non-preloaded bolts than for preloaded. EN 15048 has the general title ‘Non-preloaded structural bolting assemblies’ and is divided into two parts as follows:
EN 1090-2 permits the selection of a tightening method from the three methods outlined below unless otherwise specified:
Torque control method
• EN 15048-1: General requirements;
The torque is to be applied in two steps. The first step, after bedding of the joint, is to apply a torque of up to 75% of the required torque value to all the bolts. The second step is to apply an additional torque to each bolt such that the total applied to a bolt is up to 110% of the required nominal torque value. The extra 10% is to offset the subsequent preloading force decrease. The required torque value is determined using the data from k-class K2.
• EN 15048-2: Suitability test. All the usual strength grades from 4.6 to 10.9 are included. However, to allow a variety of manufacturers’ products to comply, the standard does not specify bolts directly being more of a performance standard than a prescriptive one. EN 15048-1 sets out mechanical and other minimum requirements and EN 15048-2 specifies suitability tests for non-preloaded assemblies. If a bolt to any national standard can satisfy the general requirements and pass the suitability test it may be placed on the market as a non-preloaded bolt to EN 15048. In the UK bolts to BS 4190, with a slightly deeper nut which is required to pass the suitability test, should be able to conform to EN 15048.
Combined method This method is a combination of torque control and the traditional “part-turn” method. After the joint is bedded the preloading takes place in two steps. The first step is to apply a torque of up to 75% of the required torque value to all bolts (as with the torque control method). The second step is to apply to each bolt a predetermined rotation or “part-turn” to a specified angle, depending on the bolt length. As an alternative to using the specified angles, the angle of rotation may alternatively be determined by test. The required value for the pre-torque is determined using data from either k-class K2 or K1. It should be noted that the simple part-turn method used in the UK for many years will no longer be permitted under EN 1090-2 due to the risk of overtightening bolts, which is the same reason that the simple part-turn method has not been allowed for Part 2 higher grade HSFG bolts to BS 4395.
CE Marking The driver for common product standards across Europe is the imperative to open the market and gain comparable efficiencies of scale that suppliers enjoy in the US with its common standards. Since the Construction Products Directive (CPD) was promulgated in 1989 this initiative has been thwarted by an absence of harmonised standards for fasteners. Two factors have delayed this development: • Abortive attempts to widen the common approach to develop ISO standards for structural fasteners which stumbled due to conflicting commercial interests in Europe, the US and elsewhere.
Direct tension indicator method This method is the most popular in the UK and relies on protrusions on special washers often called load indicating washers. These protrusions create a gap prior to preloading in the installed assembly. After the joint is bedded down, on tightening the bolt the gap reduces as the protrusions depress under load. When the actual gap is less than the maximum specified gap the bolt force will not be less than the specified preload. The k-class may be K0 when the direct tension indicator method is used to control the preload in the bolt. It is expected that K0 will also be used in conjunction with TC bolts.
• The two systems, HR and HV, by which fastener performance is linked to design values and the associated differing methods for executing and controlling installation (as described above). Finally, the necessity for CE marking has resulted in Europe being on the threshold of getting common product standards. However, as noted above with respect to EN 15048 in particular, the common standards are driven by performance requirements and the associated suitability tests. At its simplest, the CPD asks manufacturers to demonstrate that their products meet such test standards before placing them on the market, and for this reason it is whole assemblies that need testing.
Details of the procedure test for determining the slip factor in connection with preloaded bolts are found in Annex G of EN 1090-2. It is somewhat more complex than the traditional British approach in that it requires five tests instead of three and slip is measured at a higher value of 0.15mm instead of 0.10mm. However, the slip load is averaged over the five tests instead of taking the lowest of the three test results as in the BS 4604 approach.
36
Conclusion It remains to be seen what effect the new standards for structural fasteners will have on the European market. It is possible that the wider market will be competitive, but it is also possible that the requirement to re-tool or invest in additional testing will mean that the “new” products will be more expensive – at least initially. For that reason, it is likely that BS 4190 bolts will remain in the market for a few years yet. However, there is likely to be a greater need to change in public sector works than in the private sector owing to the additional directives that cover public purchasing. Hence, for bridges in particular, one may see preloaded bolts to BS EN 14399 replacing HSFG bolts to BS 4395 sooner rather than later.
37
CHAPTER 9 Cladding connections
The general cladding issues described here are those that need to be taken into account when designing and erecting structural steelwork for “sheds” such as warehouses, retail stores, factories and leisure centres, and focuses in particular on the specific actions needed to comply with Part L. Also considered are the other cladding related issues that main contractors, steelwork designers and steelwork contractors need to consider in order to successfully build quality sheds.
Richard B Barrett, Barrett Steel Buildings Limited
Introduction One of the primary functions of all buildings is to provide a controlled internal environment that is protected from the varying external environment. The separation of these two environments is provided by the building’s envelope.
Loading and design issues
The UK government is trying to reduce energy consumption in an effort to reduce levels of greenhouse gas emissions. Operationally, buildings account for approximately half of all energy consumption, and consequently the thermal performance of new buildings has become a key focus of the government’s CO2 strategy. Part L of the Building Regulations, introduced in 2002, sets tough standards and these will get even tougher with an amendment due out at the beginning of 2006. The building envelope is a crucial factor in the energy performance of the building, and is therefore under ever increasing levels of scrutiny.
The roof and wall claddings are an important structural element of the building. They must not only carry the external loads such as wind and snow, but are also required to provide lateral and torsional restraint to the purlins and side rails, which in turn provide stability to the main frame itself. It is important that consistent design loads are used for the design of both the steelwork and the cladding. For example, if the steelwork has been designed for wind loads to BS 63992 using the hybrid method in the SCI guide “Recommended Application of BS 6399-2 (Electronic publication SCI ED001)”, this will have resulted in purlins and rails distributed to suit these loadings. The cladding designer needs to use the same load calculation method for the design of the cladding system, otherwise there is a risk that certain cladding sheets will be overloaded, or insufficient fixings provided.
A vital factor in the performance of the building’s envelope is the supporting steelwork. Crucial to that performance are the design, detailing, tolerances and accuracy of the steelwork, particular of secondary items such as purlins and rails.
Debenhams Distribution Warehouse, Peterborough
38
Manufacturers of purlins and rails normally assume that their products receive lateral and sometimes torsional restraint from the cladding. Designers need to satisfy themselves that the sheeting and its fixings can provide this restraint. Most purlin manufacturers give guidance on which types of cladding systems adequately restrain their purlins, and this guidance should be followed carefully.
riding issue is the ability of the cladding sub-contractor to fix the cladding to purlins and rails without compromising air-tightness. Difficulties can arise due to three main causes; poorly designed supports, deformed secondary steelwork, and out of position steelwork. Looking firstly at the design elements, the key requirement is to provide a robust and continuous fixing face for the cladding joints and gutter supports. This is straightforward for the main runs of purlins and rails, but care is needed at interfaces such as hips, corners, parapets, and openings. For example, at the hip ridge, two slopes come together at a complex raking angle. The cladding sheets need a robust continuous support to both slopes. One effective solution is a purpose made dihedral angle supported directly from the hip rake, as shown in the photograph. Too often details like this are left to be developed on site with pieces of light gauge pressed metal angles. The result is often a poorly built hip, with the risk of serious air leakage.
Purlins and rails are often required to provide lateral restraints to rafters and columns, particularly in portal frame buildings. For the outer flange this will be provided directly by the purlin connection. In cases where the inner flange is in compression, diagonal struts or rafter stays will need to be provided. It is important that the purlins or rails are large enough to have sufficient stiffness to act as restraints. As a rule of thumb, a purlin should be at least 25% of the depth of the member being restrained.
Air leakage and robust details
Deformed purlins and rails are another potential cause of air leakage. Cold rolled sections are easily damaged on site. If damage is severe it will prevent the cladding from being properly fixed to form an air tight envelope. The longer a cold rolled element is on site before being fixed, the more likely it is to incur damage. Therefore on a well planned project it is normal for the delivery of cold rolled components to be phased to suit precisely the erection sequence. Items that need to
Part L requires that new buildings are air tested to check that the building meets demanding targets to minimise air leakage – draughty buildings lose lots of heat. A failure to achieve the required standard would have potentially serious consequences for the design and construction team, as it is very difficult retrospectively to remedy air leakage problems in a satisfactory way. For the steelwork element, the over-
Hip: robust support using a purpose made dihedral angle
39
be left off in the first fix – such as rails across access bays or purlins for canopies – should be kept off site and delivered later. If, for some reason, cold rolled steelwork has to be stored on site, it should be placed in a safe, traffic free area. The main contractor needs to be made aware, and asked to keep vehicles out of the way.
insulation line, a neoprene thermal isolator packing should be used to reduce heat loss. This is particularly effective on minor brackets and connections such as parapet posts. However, a thermal packing will not be possible on a substantial moment connection, such as a canopy cantilevered from the main columns through the insulation line. Such architectural details should be avoided, if possible, as they will create thermal bridges, the effects of which will need to be accounted for in calculations of the building’s thermal performance.
Finally, crucial to the performance of the cladding system is the accuracy of the installation of purlins and rails. There is little guidance on specific erection tolerances for secondary steelwork in the UK. Whilst some tolerance may be deduced from BS 5950-7 (now withdrawn in favour of DD ENV 1090-2) and from the National Structural Steelwork Specification, there is a concentration on primary frame elements such as beams. A new publication on metal cladding is expected from the SCI in late 2005 (“SCI Cladding Guide”) and this may address specifically a number of cladding/steel related issues, including secondary steelwork tolerances.
Cladding systems There are many types of cladding, but here the focus is on the two most commonly used; modern twin skin built-up systems, and composite panels, made from profiled steel sheets. These systems dominate the “shed” market due to their economy, performance and installation speed.
In the meantime, it is probably best to approach the issue from a practical point of view. The requirement is for the cladding to be fixed without compromising air tightness. Some cladding contractors report problems in this respect due to secondary steelwork being out of position. This may be due to poor fabrication or bad site practice. Poor fabrication may result in purlin cleats being out of position on the back of the rafter or being fixed to the rafter out of square. This will result in the purlin being offset from the joint in the roof cladding, and therefore exceedingly difficult for the roofer to fix properly. Alternatively the angle of the rafter pitches may vary slightly, causing the roof ridge to vary up and down the building. Finally, poor site fixing could lead to the purlins festooning down the slope. This is probably the most serious tolerance problem, as it may result in the cladding fixings “missing” the purlin altogether, so that the sheet is not actually attached to the purlin at all. These are all practical problems, which can be avoided by selecting high quality experienced steelwork contractors, who will ensure that fabrication is carried out to good standards, and site fixing is undertaken by competent erectors. One way to help ensure that steelwork contractors have these competences is to select them from the Register of Qualified Steelwork Contractors, details of which are available from the BCSA.
Built-up systems consist of the following site assembled elements: • Liner sheet generally made from steel with a shallow trapezoidal profile. The liner tray restrains the purlin or rail, supports the insulation and forms a weathertight layer prior to the fixing of the outer sheet. • Insulation comprising a mineral fibre quilt (e.g. rock wool). • Spacer system to support the outer sheet at the required distance from the liner. The spacer is always positioned directly over the purlin and fixed securely to it. • Top sheet or weather sheet which has the primary role of keeping out the weather, but is also the structural element that supports the external loads through to the spacer and then to the structural frame. These sheets are usually zinc coated steel painted to provide the required appearance. Composite panels comprise of a rigid layer of insulation bonded between two metal skins, which results in a strong, stiff panel. Most commonly panels are factory made units, although they can be assembled on site. Unlike built-up systems, there is no need for a spacer system, as the insulation is strong enough to support the outer sheet. The liner and weather sheets are usually similar to those used with built-up systems. The whole panel acts as a restraint to the purlin, with the fixing screw passing right through from the outside of the weather sheet to the purlin top flange.
Thermal bridges Part L also considers the impact of thermal bridges on the thermal performance of the building. Steel is an excellent conductor, and therefore any penetration of steelwork through the building envelope will create a thermal bridge. This will result in heat loss from the building and also create a risk of condensation.
For roofs, an alternative to the trapezoidal weather sheet, used in both built-up systems and composite panels, is a “standing seam” outer sheet. This form of cladding is attached using a clip system, and hence has no penetrations through the weather sheet. Consequently it can be used down to lower pitches. However the clip system may not give the level of restraint to the purlin that is required, and the purlin manufacturer should be consulted for guidance on purlin restraint for these types of system.
Linear thermal bridges, in which a line of steelwork, such as a capping channel or eaves beam, penetrates the envelope, are not acceptable. The steelwork needs to be detailed to ensure that the whole of the linear member is either inside or outside the envelope. Point thermal bridges, such as brackets penetrating the envelope, are also not desirable and should be avoided wherever possible. If the steelwork has to penetrate the
40
Roofs
When the cladding spans horizontally, there are a variety of ways to form a supporting system, all of which are more complex and hence more expensive than for vertically fixed cladding.
With regard to air tightness, built-up roof systems are more tolerant of minor problems with the steel frame than composite panels. This is because the provision of a lap between liner sheets allows moderate deviations to be accommodated. The new SCI Cladding Guide suggests that a variance in dimensions between adjacent purlins of up to 25mm is acceptable on built-up systems.
For built-up systems, a relatively simple solution is to fix the liner panels to span vertically, and then fix the spacer bars at 90 degrees to the rails, that is vertically between the rails. The outer sheet then spans horizontally between the spacers. In this way the horizontal outer sheet can be fixed without the need for a complex framing system for the rails.
Composite panels require tighter tolerances in the positioning of the purlins because there is no adjustment possible in the liner panel laps, and there are strict end bearing requirements for this type of panel. The SCI Cladding Guide suggests just 10mm tolerance is acceptable as the variance in dimension between adjacent purlins, so great care needs to be taken in fabricating the steelwork, and when fixing and aligning purlins for this type of cladding.
Alternatively the rails need to be designed to support horizontally spanning cladding sheets. This can be done in a number of ways – for example horizontal rails can be fitted with a secondary vertical grid between them forming the cladding supports, or the rails can span vertically onto a number of substantial steel members spanning between the columns. Rail manufacturers show various options in their literature. It is important with all these systems that care is taken in design and fixing to ensure the system is adequately supported to stop the rails sagging. Another problem that can arise is with air tightness; vertical spanning rails need to be correctly aligned so that their outside flanges line up across the supporting members, creating a smooth fixing face for the cladding.
Wall claddings Built-up systems and composite panels can be specified to span either vertically or horizontally on side walls, depending upon architectural preference.
One particular issue with horizontal panel cladding is the tolerance required by the cladding manufacturer. Whilst in many cases it is not a particular problem, sometimes ludicrously tight tolerances are specified, often for no apparent reason. These can be almost impossible to achieve, and will certainly need some form of secondary adjustment mechanism, such as packing or slots. This will be costly, and will also have implications for the building programme; such fine tolerances can only be achieved with extensive surveying and numerous adjustments. Therefore, when comparing cladding systems it is prudent that manufacture’s recommended tolerances are carefully checked, so that any extra steelwork costs and extended programme times can be taken into account before a selection is made.
Gutter and downpipe supports Gutters used to be relatively light weight items, made from galvanised pressed steel. Supporting them was relatively easy, usually through the careful positioning of an appropriate purlin or rail. However, to achieve modern insulation standards, gutters are now usually made with insulation bonded to them, similar to a composite panel. These are quite heavy, weighing up to half tonne over a 7 metre bay width and they need a more substantial steel member to support them. The thickness of the gutter is also considerable, and needs to be known when positioning steelwork supports.
Kingspan Multibeam rails with anti-sag system
When the cladding spans vertically, the side rail system operates in a similar way to the purlin system on the roof, with side rails running horizontally, spanning from column to column. It is important to get the rails into good alignment, in order that the cladding sheets can be properly fixed. Normally side rails are prevented from sagging with a manufacturer’s anti-sag system. These consist of a set of struts or rods between the purlin to set the correct horizontal spacing, together with diagonal wires or rods to take the vertical dead loads back to the supporting columns. These should be fixed in the correct sequence, as specified by the manufacturer, in order to achieve a good level line to the rails.
When positioning steelwork under gutters, it is also necessary to consider the gutter drainage method. Space may be needed for sump outlets, and if syphonic drainage is specified, the pipe runs need to be considered to ensure there are no clashes with the steelwork.
41
Erecting sheeting rails from mobile elevating working platforms
Construction issues
Finally, issues may arise relating to allowing safe access for the cladding operatives to get to particular areas. For example, it is safer and quicker to fix vertical cladding above a canopy or lean-to area before the steelwork is erected in that area. Therefore the project programme should show this as a later fix item, and the relevant steelwork should be planned for later delivery than the main building.
When developing the project programme, the sequencing of operations between the steelwork erection and cladding fixing needs to be taken into account. Firstly, to load the cladding materials onto the roof, the cladding contractor will probably require some purlins to be left out, so that a hoist can be used to load the roof. In some cases in may be necessary to alter the sequence of roof steelwork, to allow a crane to lift the cladding onto the roof. This could affect positioning of bracing etc., and so needs to be determined at the earliest possible stage. It is also important that the unfixed roofing materials are placed appropriately on the roof. Until the roof sheets are fixed, the purlins are completely unrestrained, and their load carrying capacity is much reduced. Consequently, packs of cladding should be kept relatively small, and positioned at rafter positions, and not in the mid-span of purlins. If in doubt, the cladding contractor should consult the purlin designer before loading the roof. As noted above, the gutters are now generally heavy items. On some contracts it may be more appropriate for the steelwork contractor to lift the gutter sections into place, for final fixing by the cladding contractor. For example, this may avoid the cost of a substantial crane to reach over the finished steelwork to a distant valley gutter location.
The Big One, Prologis Developments Ltd., Northampton
42
Health & Safety There are certain safety issues for the installation of the roof cladding that may have implications for the steelwork and its erection. To fix the roof cladding safely, it is now common practice in the UK for the roof area to be netted, so that cladding fixers do not fall to the ground if they fall off the leading edge of the roof. The nets are tied to various points on the steel by qualified netting installers. However, as a consequence of the roof being netted, it is not possible to access the roof steelwork from below. Therefore it is essential that all this steelwork is completely finished before netting begins. Minor items of snagging, such as a missing rafter stay or bolt, need to be attended to before netting takes place, or otherwise it may have to wait for a number of weeks to be rectified. Around the perimeter of the roof, the cladding contractor will need to fix a safety handrail to protect operatives from falling. This handrail will be attached to the steelwork. Often this is done using scaffolding, the posts of the handrail being formed by clamping scaffold poles to the top two or three lines of sheeting rails. This will only provide a satisfactory solution if the rails are adequate and in the appropriate position. It also suffers a couple of drawbacks; the clamps can potentially damage the rails, and the handrailing system prevents wall cladding operations until the roof is completed and the handrail scaffolding removed. A better solution is to attach the handrailing to sacrificial brackets built into the steelwork, preventing any clamping damage and allowing wall cladding to start before the roof is finished. To achieve this level of coordination will require a well managed supply chain, with early appointment of all the key contractors.
Conclusion Most single storey buildings are straightforward, but the increasing sophistication of cladding systems has increased the number of issues that need to be considered when designing and erecting these buildings. In particular, the impact of Part L of the Building Regulations has imposed a number of new demands on the steelwork, and these will increase as the Regulations become more onerous. However, if care is taken with the issues covered here, and competent contractors are selected, then the steel-framed single storey building offers a high-quality fast-build solution for today’s clients.
43
CHAPTER 10 Acoustic details
In buildings, sound insulation methods can be divided into two types: • Airborne sound insulation;
Andrew Way, SCI
• Impact sound insulation.
Background on acoustics and sound Introduction Sound is caused when objects vibrate in air. The movement in turn causes air particles to vibrate giving rise to rapid pressure fluctuations which are detected by the ear. The manner in which humans perceive sound governs the way it is measured and described. Two important characteristics of sound which humans can detect are:
The frequency range of building acoustics (Hz)
• the level – how loud a sound is, and • the frequency – whether it is high or low pitched.
Airborne sound insulation
Sound levels and sound insulation values are expressed in decibels (dB), whilst pitch or frequency is expressed in Hertz (Hz). In the case of sound levels, the dB rating is a representation of the volume of the sound whilst in the case of sound insulation values it is a measure of the amount by which sound transmitted from one room to another is reduced by the separating construction. Some typical sound levels and sound insulation values are illustreated. Sound Level
Airborne sound insulation between rooms can be measured by generating a steady sound of a particular frequency content in one room (the source room) and comparing it with sound received in a second adjacent room (the receiving room). These measurements are made at a number of different frequencies. The difference between the two levels is referred to as the level difference D. This level difference is influenced by the amount of acoustic absorption around the receiving room. The absorption can be estimated by measuring the reverberation time T - the time taken for the reverberant noise to decay by 60 dB. In order that measurements in different buildings may be compared, the level differences can be adjusted to a standard reverberation time of 0.5 seconds. This gives the standardised level difference DnT.
Sound insulation 120
Pneumatic drill 100 Inside underground train
Specialist broadcast studio wall
80
Normal conversation 60 Living room (suburban) Bedroom at night Threshhold of hearing
Individual building elements such as partitions, doors or windows can be tested in acoustic laboratories. These laboratories comprise two massively constructed adjacent rooms which are isolated against flanking transmission (see below) and connected by an aperture containing a test panel of the building element. The level difference is measured between the two rooms and the result adjusted to be independent of both the area of the panel and the acoustic absorption of the room. The resulting value is the sound reduction index R.
Solid brick wall (225mm)
40
Acoustic double glazing
20
Single sheet of steel No wall
0 (dB)
Typical sound levels and sound insulation values (dB)
Impact sound insulation
The sound insulation properties of walls or floors vary with frequency and, as most sounds are a mixture of several different frequencies, certain frequencies within a sound are likely to be attenuated more effectively than others by a given construction (low pitched sounds are normally attenuated less than high pitched sounds). In view of this, the sound reduction characteristics of walls and floors are usually measured at a number of different frequencies across the hearing range. The normal frequency range of measurements is illustrated.
Impact insulation tends only to be relevant to floors. A standard impact sound source (a tapping machine) is used, comprising a row of hammers which strike the floor repeatedly at a standard rate. The resulting sound in the receiving (downstairs) room is measured and this value is termed the impact sound pressure level L. Measurements in buildings can be standardised to a reverberation time of 0.5 seconds. This gives the standardised impact sound pressure level L’nT (a field measurement). Tests in laboratories, normalised for area and absorption, give the normalised impact sound pressure level Ln. Thus, the better the impact sound insulation, the lower the value of L’nT or Ln.
Sound insulation can be described in a variety of ways. This can initially be confusing for the architect or designer trying to interpret specifications or manufacturers’ literature. The following attempts to explain some of the main terms. More comprehensive descriptions are given in BS EN ISO 140-1.
44
Single figure rating values
Acoustic regulatory requirements
Sound insulation is normally measured at a number of different frequencies - usually 16 one-third octave bands from 100 Hz to 3150 Hz. However, for many purposes, including the requirements for dwellings given in Building Regulations, a single figure rating is required. There are several methods of reducing the sound insulation values at the 16 individual frequencies to a single figure value. An obvious method is to take the arithmetic mean, but very high levels of sound insulation at some frequencies can offset poor performance at others. The most common method of overcoming this is to compare the measured results with a set of 16 reference results i.e. a reference curve. The rating is made by considering only those sound insulation values which fall short of the reference curve. In this way, one or two very good results have much less effect on the single figure value. The method used for airborne sound is given in the figure. A similar method is used for impact sound.
Introduction Acoustic standards are a product of both physical needs i.e. the need to sleep or to hold a conversation, and the general expectancies of building users. It is generally acceptable, for instance, for a certain amount of sound to cross between a typical domestic kitchen and living room, however it is not acceptable in modern homes for conversations to be audible between dwellings. It is apparent that the standards of acoustic insulation that are required in different parts of a building will vary, and that the performance required of individual building elements will reflect this variation. For example, walls between offices and a workshop are likely to require greater acoustic insulation properties than those between a trade counter and the same workshop.
Regulations The acoustic requirements for dwellings and rooms for residential purposes are specified in Approved Document E of the Building Regulations for England and Wales. The equivalent document in Scotland is Section 5 of the Domestic Technical Handbook, in Northern Ireland it is Technical Booklet G. For hospitals, Health Technical Memorandum 2045 “Acoustic design considerations” produced by NHS Estates, specifies the requirements. For schools, Building Bulletin 93 “The acoustic design of schools” produced by the Department for Education and Skills, should be adopted.
Calculation of single figure value DnT,w
• Weighted sound reduction Rw when generated from R;
The acoustic requirements detailed in the documents stated above are expressed using different terms and methods as appropriate to the different building types. Therefore, a direct comparison of requirements is not straightforward. However, the principles of good acoustic detailing are consistent.
• Standardised weighted impact sound pressure level L’nT,w when generated from L’nT;
The acoustic requirements from Approved Document E for separating walls and floors are given in the table.
The single figure values are called: • Standardised weighted level difference DnT,w when generated from DnT;
• Normalised weighted impact sound pressure level Ln,w when generated from Ln.
Building type
Walls Airborne DnT,w+Ctr
Airborne DnT,w+Ctr
Impact L’nT,w
Purpose built dwellings
≥45 dB
≥45 dB
≤62 dB
Dwellings formed by material change of use
≥43 dB
≥43 dB
≤64 dB
Purpose built rooms for residential purposes
≥43 dB
≥45 dB
≤62 dB
Rooms for residential purposes formed by material change of use
≥43 dB
≥43 dB
≤64 dB
45
Floors
Note: The Ctr term is a spectrum adaptation term, which is generally negative and adjusts the airborne performance index to take additional account of the low frequency sounds that often cause problems in residential buildings. Ctr is generally in the region of -5 dB to -8 dB for steel framed type construction.
that the sound insulation of a solid element will increase by approximately 5 dB per doubling of mass. The mass law does not however apply to lightweight framed constructions which achieve far better standards of sound insulation than the law would suggest owing to the presence of a cavity, and the degree of isolation that is achieved between the various layers of the construction. It has been demonstrated that the sound insulations of individual elements within a double skin partition tend to combine together in a simple cumulative linear relationship. The overall performance of a double skin partition can therefore generally be determined by simply adding together the sound insulation ratings of its constituent elements. In this way, two comparatively lightweight partitions of 25 to 30 dB sound reduction can be combined to give an acoustically enhanced partition with a 50 to 60 dB sound reduction, whereas the mass law alone would have suggested only a 5 dB improvement. This is the basis of many lightweight partition systems, and is further illustrated here.
Approved Document E explains that there are two methods of demonstrating compliance: • Carry out on-site tests to measure the acoustic performance of separating walls and floors, to confirm that the performance standards in Approved Document E are met. • Use Robust Details (RDs), as published in the Robust Details Handbook, throughout the building. Before construction the developer must also register the site with Robust Details Limited, who administers the RD scheme.
Principles of acoustic detailing Direct and flanking transmission When a room is separated from another room, airborne sound can travel by two routes: directly through the separating structure - direct transmission, and around the separating structure through adjacent building elements - flanking transmission. These routes are indicated in the illustration.
Sound insulation by layers
It is important to provide adequate sealing around floors and partitions since even a small gap can lead to a marked deterioration in acoustic performance. Usually walls are sealed by the plaster finish; however, where walls abut profiled metal decks, or similar elements, sealants may be required. Where there are movement joints at the edges of walls, special details are likely to be necessary and advice can be sought from manufacturers.
Transmission of sound
Sound insulation for both routes is controlled by the following three characteristics:
Ideally, wall linings should not be penetrated by services. This is particularly important for separating walls between dwellings. Where service penetrations do occur in sensitive locations particular attention should be given to the way in which these are sealed, see following.
• Mass; • Isolation; • Sealing. Direct transmission depends upon the properties of the separating wall or floor and can be estimated from laboratory measurements. Flanking transmission is more difficult to predict since it is influenced by the way in which the building elements are configured and detailed. It is notable that, in certain circumstances, such as where separating walls have a high standard of acoustic insulation but side walls are constructed to lower standards and are continuous between rooms, flanking transmission can account for the passage of more sound than direct transmission.
Separating floor and wall junctions The following details are examples of steel framed construction that will meet the acoustic requirements for domestic buildings. A wider range of details is given in SCI publication P-336, “Acoustic detailing for multi-storey residential buildings”. Alternative, proprietary details are available from some product manufacturers. A combination of the acoustic principles of mass, isolation and sealing can be observed in the details shown in the following figures.
Sound transmission across a solid wall or a single skin partition will obey what is known as the mass law. This law may be expressed in a variety of ways. In principle the law suggests
46
Notes
External cavity wall with light steel internal leaf and shallow deck composite floor (with downstand beam)
This detail is a Robust Detail (RD) when it is used in conjunction with an approved RD floating floor treatment, A ≥ 80mm, B ≥ 130mm, C ≥ 300mm, the concrete density is at least 2200 kg/m3 and the light steel frame inner leaf has insulation between the studs. Dimension D depends on the ceiling treatment used. Decking may be trapezoidal or re-entrant and may span in either direction. Where decking profiles are at a right angle to the walls, voids (above the beam) are filled with profiled mineral wool inserts and caulked with acoustic or flexible sealant. Ceiling board should not be in direct contact with any steel beams or columns. Wall outer leaf may be masonry or precast panels. Wall inner leaf must not be continuous between storeys. Notes
External cavity wall with light steel internal leaf and deep deck composite floor (with RHS or ASB edge beam)
This detail is a Robust Detail when it is used with an ASB edge beam, in conjunction with an RD floating floor treatment, A ≥ 80mm, C ≥ 300mm, the concrete density is at least 2200 kg/m3 and the light steel frame inner leaf has insulation between the studs. Dimension D depends on the ceiling treatment used. The edge beam may be an RHS with welded plate or an ASB. However, if an RHS is used it does not have Robust Detail status and on site testing is required. Decking may span in either direction. Ceiling board should not be in direct contact with any steel beams or columns. Wall outer leaf may be masonry or precast panels. Wall inner leaf must not be continuous between storeys. Notes
Performance levels similar to those of a Robust Detail could be expected with A ≥ 80mm, B ≥ 130mm, E ≥ 200mm and F ≥ 50mm. Dimension D depends on the ceiling treatment used. Concrete density should be at least 2200 kg/m3. Decking may span in either direction and may be trapezoidal or re-entrant. Where deck profiles are at a right angle to the walls, voids (above the beam) are filled with profiled mineral wool inserts and caulked with acoustic or flexible sealant. Wall and ceiling boards should not be in direct contact with any steel beams or columns. Floor treatment should not be continuous under separating wall. Internal light steel separating wall and shallow composite deck floor (with downstand beam)
47
Notes
Performance levels similar to those of a Robust Detail could be expected with A ≥ 80mm, E ≥ 200mm and F ≥ 50mm. Dimension D depends on the ceiling treatment used. Concrete density should be at least 2200 kg/m3. Decking may span in either direction. Floor treatment should not be continuous under separating wall. Wall or ceiling boards should not be in direct contact with any steel beams or columns.
Internal light steel separating wall and deep deck composite floor (with ASB) Notes
Performance levels similar to those of a Robust Detail could be expected with A ≥ 40mm, B ≥ 150mm, E ≥ 200mm and F ≥ 50mm. Dimension D depends on the ceiling treatment used. Precast unit density should be at least 300 kg/m2. Screed density should be at least 80 kg/m2. Precast units must butt tightly together and all voids between units must be grouted. Wall or ceiling boards should not be in direct contact with any steel beams or columns. Floor treatment should not be continuous under separating wall. Voids between the wall and floor must be filled with acoustic or flexible sealant.
Internal light steel separating wall and precast floor (with downstand beam)
48
Floor and ceiling treatments
Deep batten floor
Floor treatments The six illustrated floor treatments may be used with floor slabs constructed from insitu concrete with deep or shallow profile metal decking or with floor slabs constructed from precast units and screed topping. The details are given in descending order of estimated acoustic performance.
Ceiling treatments
Cradle and battern floor
All separating floors should have a ceiling treatment of at least one layer of nominal 8 kg/m2 of gypsum-based board. Ceiling boards may be supported by: • Propriety metal frame system with a void dimension D of at least 75mm; • Timber batterns and counter batterns with a void dimension D of at least 100mm; • Resilient bars with a void dimension D of at least 15mm.
Standard batten floor
Exact minimum void dimension D depends on the mass of ceiling board being used and the floor slab construction. See SCI publication P-336 for further guidance.
Integration of columns and services Any element, whether it be services or structural, that penetrates or is built into a separating wall or floor must be detailed appropriately to ensure that the acoustic performance of the separating wall or floor is not impaired.
Platform floor
The usual solution for services that penetrate separating floors is to box in the service with two layers of gypsum-based board. The usual solutions for services within a wall are to stagger services on either side of the wall and provide additional layers of gypsum-based board where the wall board is penetrated or to construct a special void within the wall to house the services. Possible detailing solutions for structural columns built in to a separating wall are provided in the two figures. Advisory desk note AD 287 “Acoustic detailing of steel columns within masonry separating walls” deals with acoustic detailing of steel columns within masonry separating walls and provides further detailing solutions.
Shallow Platform floor
Conclusion Detailing of buildings to diminish the effects of airborne and impact sound is an important factor in providing a pleasant environment for occupiers. Modern styles of living probably create greater volumes of noise, but regulations are now in place to define suitable acceptance criteria. Scientific evaluation of sound insulation performance enables designers to select suitable building details, and the steel construction industry has been active in building a portfolio of pre-qualified Robust Details. This portfolio will grow as more and more buildings are completed that satisfy the current requirements.
Screed floor
49
Integration of column in a masonry separating wall
Integration of column in a light steel separating wall
50
CHAPTER 11 Composite connections
Principles The basic principle of composite connections is to use reinforcement in the slab as an additional bolt line in an end plate moment connection. A significant tensile force at an increased lever arm can make a massive difference to the potential moment capacity of a connection. For example, for a 457x191x74kg/m UB a flush end plate connection could have a capacity of about 200kNm if connected to a suitable column. This compares to about 450kNm if eight T16 reinforcement bars are added in the slab. Although this may seem a significant moment it is about half of the composite moment capacity at midspan.
Mike Banfi, Arup
Introduction One of the major steel successes of the past 25 years has been the use of composite construction for multi-storey construction. It took into account lessons learnt from North America and has developed to give steel a 70% share of the UK market in this field. A typical feature of this construction is that the beams are designed as simply supported despite the fact that in many cases the slab and steel beam are continuous either side of the support. Composite connections (designed to transmit moment) are rare; the reasons for this are a mixture of technical, managerial and cultural. An explanation of these reasons will help people identify where they may be appropriate and implement them successfully. The composite connections considered here fall into two basic types: beam-to-column and beam-to-beam as illustrated.
Forces in composite connection to an internal column (from P-213)
BS 5950-3.1 and Eurocode 4 give advice on the use of elastic global analysis for continuous composite beams. Redistribution is allowed even for semi compact sections. However even with redistribution the resulting support moments are likely to be of the same order of magnitude or greater than the midspan moments. This will not be an efficient use of the strength of the beams. A plastic analysis will allow the capacities to be used but requires knowledge of the connection capacity and sufficient rotation at the connection. This is the process proposed in P-213. Composite connections can be used in cantilevers and of course in that case there can be no redistribution. In these cases the capacity of the connection during construction may govern the design.
Limitations for plastic design The plastic rotation of the connection means that components of the connection must deform. The rotation capacity of the connection is from elongation of the reinforcement, slip of the shear connections and compression of the bottom flange. The most important factor is usually the reinforcement. Because of the lack of ductility of welded mesh this cannot be taken into account in the design and bar reinforcement must be used to provide the required area. Normal bar reinforcement has an elongation at maximum load of 5% and in some cases more ductile reinforcement may be required to ensure the required rotation. To make sure that the reinforcement is strained over a significant length it is recommended that the first stud is located away from the connection i.e. at least 100mm from the face of the column or 200mm from the centre of the connection where there is no column.
Typical composite connections (from P-213)
The connections have many common characteristics but significant differences. The connections are also covered in Eurocode 4. The BCSA/SCI have published a guide on composite connections as part of the Joints in Steel Construction series (P-213). This should be read by anyone using, or considering, this form of construction.
51
Design considerations
The moment capacity of the connection, the beam and the span influence the amount of rotation required at the connection. Obviously the stronger the connection the less plastic rotation it will need to undergo. Similarly if the moment at midspan is reduced there will be less curvature and less rotation at the end connection. Longer spans will also give more rotation. The SCI guide suggests limits for these three factors:
Moment connections at the ends of beams can be used for two main purposes. They can improve the performance of the beam at both Serviceability and Ultimate limit states and/or they can be used to frame into columns and provide stability to a building. As composite connections are likely to be partial strength their use in frames can be quite complicated but simple methods are available in certain situations. Another SCI publication (P-264) gives guidance on Wind-moment Design of Unbraced Composite Frames. The method is limited to frames in one direction only, between 2 and 4 storeys high and between 2 and 4 bays wide. There are other limits on dimensions and loading. The connections proposed are those in P-213 and similar limits and restrictions apply; in particular in many cases high ductility reinforcement is required.
• the connection capacity should be at least 30% of the sagging moment capacity of the beam; • the applied moment at midspan should be limited to 85% of the beam capacity; • the ratio of the span to the total depth of beam and slab should be limited to: - 25 for beams subject to UDL, multiple point loads or a central point load;
The use of composite connections in moment frames means the connections will include a column. Where they are used to help the beam they may be beam to beam connections as well as beam to column. Beam to column connections are often end plate connections anyway so an upgrade to a composite connection does not mean a major change to the steelwork detailing. On the other hand beam to beam connections in simple design are likely to be fin plate or even angle cleats and the change to an end plate connection can mean a change in normal practice. Beam to beam connections are also very unlikely to have significant moment capacity at the construction stage and hence there will be higher stresses due to construction dead load and possible increased rotation of the connection as described above. In beam to column connections the tensile load in the reinforcement is transferred to the column by being anchored into the concrete behind the column and the concrete bearing on the column flange. The reinforcement must therefore be placed close to the column and the concrete bearing capacity can limit the force that can be transferred. At the edge of the building there cannot obviously be a beam to beam connection and for beam to column connections the anchorage of the reinforcement behind the column can be difficult.
- 20 for beam subject to point loads at third points. For unpropped construction the stress in the steel due to the construction dead load must also be considered. In this case the weight of the concrete slab is not applied to the composite connection and the connection can undergo slightly more rotation. However this advantage can be more than offset by the additional rotation requirement. Because of the locked in stresses in the steel section the steel at midpsan will yield at a lower applied load and as the load is increased further the stresses in the beam section at midspan must redistribute. This leads to increased curvature in the middle of the beam and more rotation at the support. The following table (see Byfield 2005 in References) shows limits for the stress due to construction dead load. They are based on the support moment being greater than 30% of the span moment and the maximum design sagging moment being 85% of the capacity. The limits are given in terms of the maximum allowable ratio of the maximum stress in the steel prior to hardening of the slab concrete (σdl) to the nominal yield stress (σy). (Values assume support to span moment ratio (Msup/Mspan) is not less than 0.3 and the sagging moment is not greater than 0.85 Mp.) The depth to be used in evaluating the span to depth ratio is the total depth of the composite beam including the slab. The three columns for each ratio relate to different loading patterns, 2 Point Loads, a UDL or 3 Point Loads.
It is worth remembering that composite connections rely on there being continuous lines of beams. Beam to column connections also need the beam to be on the centreline of the column. Irregular layouts and offsets are not going to be suitable for this form of construction. One advantage of simple construction is that alterations are relatively easy. They will need a lot more consideration with composite connections. Removing a beam to create a major opening will require assessment and possible strengthening of adjacent beams and columns. Another advantage of composite construction
Limiting ratio Steel grade
Beam location
S355
Internal
S275
Span/depth = 15
Span/depth = 20
Span/depth = 25
2PL
UDL
3PL
2PL
UDL
3PL
2PL
UDL
3PL
0.89
1.00
1.00
0.44
0.53
0.73
0.21
0.29
0.44
External
1.00
1.00
1.00
0.54
0.51
0.67
0.30
0.26
0.39
Internal
0.68
0.80
1.00
0.31
0.4
0.58
0.13
0.18
0.31
External
0.81
0.78
1.00
0.42
0.53
0.53
0.21
0.16
0.28
52
is the ability to form small service holes in the slab. A popular area for these is adjacent to columns i.e. the area where the main reinforcement for the connection is located. Reinforcement will be bars most likely 16 or 20mm diameter. They should be located with 20mm cover to the deck. In many cases there is only 60 to 70mm of concrete above the deck. Locating one set of bars in the right location will need some care. Locating bars in two directions will be quite difficult and with normal slab thicknesses composite connections will be limited to one direction.
Serviceability It can be seen that for efficient strength design with composite connections the stress in the beam under dead load may need to be controlled and it may be necessary to specify non-standard reinforcement. Where serviceability governs the restrictions may be less and the benefit of continuity in reducing deflections will be significant. It should be noted that composite connections will not really improve the vibration performance of the floor because normal connections have enough stiffness to be considered as continuous under these conditions. When the deflection under imposed load needs to be controlled to less than the normal span/depth ratios composite connections can be useful. On the edges of buildings the cladding often requires strict limits on deflections. Composite connections may be appropriate for these cases but often it is difficult to incorporate the reinforcement adjacent to the columns. Another possible use due to the increasing floor spans where the normal span/depth ratios imply quite significant deflections i.e. span/360 for an 18m span is 50mm. These deflections are not usually a problem in the main part of the floor but where, for example such a beam spans adjacent to a core this would be a significant differential deflection and may need to be controlled. Composite connections have been used for this reason on a recent project.
Conclusion The idea of using composite connections as end connections to improve the design of composite frames may appear a good idea. Unfortunately the strength of typical connections is such that normal elastic designs are not efficient. Plastic design must therefore be used and the resultant rotation of the composite connections can require nonstandard reinforcement and/or limits on stresses during construction. Where deflections govern it is more likely that composite connections can make a useful contribution. In the countries where composite connections are very common steel does not have a high market share. The simple construction method used in the UK is adapted to the particular behaviour of steelwork and has been very successful. However in particular situations composite connections can have a benefit and they should not be disregarded.
53
CHAPTER 12 Joints under fire conditions
temperature increases), and secondly by expansion. These lead to changes in structural behaviour which interact with other parts of the building so that the net structural response can be very different from that at ambient temperature. This depends on the continued integrity of the joints, which are subject to marked changes in force during a fire, especially in the case of composite floors. As temperatures increase the exposed steel beams initially heat rapidly and expand, with little reduction in strength, whilst the concrete slab heats much more slowly. The resulting temperature differential causes thermal bowing towards the fire, inducing very high permanent compressive strains in the weakening steel beams. These are further increased as a result of restraint to thermal expansion from the cool structure surrounding the fire compartment. As the temperatures increase further, very large flexural deformations can develop, which are acceptable provided that the fire is contained within the compartment of origin. Under these conditions tensile action can develop, particularly within bare steel beams, and the dominant action in the joints is therefore very different from that at ambient temperature, when the moment-rotation characteristics are most important.
Ian Burgess and Roger Plank, Sheffield University
Introduction All structural elements (slabs, beams, joints and columns) must maintain their load-bearing function for the required fire resistance time. Those which separate different compartments must also retain sufficient integrity (there should be no cracks or openings which could allow penetration by hot gases or flames) and provide sufficient insulation (the temperatures on the non-exposed surface of separating elements must not exceed ignition temperatures) to prevent fire spread. These elements have traditionally been designed for fire by considering them in isolation. In real buildings structural elements form part of a continuous assembly, and building fires often remain localised, with the fire-affected structure receiving significant restraint from cooler areas surrounding it. The real behaviour of these connected structural elements can therefore be very different from that indicated by standard furnace tests, especially for composite construction. Practical beam-column and beam-beam joints, which may be treated as perfect hinges in the ambient-temperature design process, usually have a residual stiffness which becomes significant when the connected beams have lost much of their own strength and stiffness at high temperature. In composite construction, continuous slab reinforcement can create a much stiffer overall joint, further magnifying this effect. Hence, restraint to thermally-induced movements may exist in respect of both translations and rotations at the ends of beams and the edges of slabs. There is therefore considerable scope for load-sharing and for both advantageous and disadvantageous effects of restraint to rotation and horizontal movement.
Once a real fire starts to decay temperatures reduce and the process reverses. The heated beams regain strength and stiffness and also try to contract. However, they have effectively become shorter during the fire, partly as a result of the permanent compressive strains developed during heating, and partly because of the inevitable bending deformation which they have suffered. As the beams try to shorten, they are restrained by their connection to the surrounding structure and exert increasingly high tensions on the corresponding joints.
Design considerations for joints The fire resistance of joints must be at least the same as for the connected members. This means that beam-to-column joints should be able to transmit the internal forces during the whole fire resistance time. When passive fire protection is used on the members this requirement is generally considered to be fulfilled if the same thickness of fire protection is applied to the joints. Because of the concentration of material, the temperature of the joint tends to be lower than that of the connected members, and therefore they retain a greater proportion of their strength. However, because of the changes in joint forces described above, this is can be a considerable oversimplification.
It is implicitly assumed in fire engineering design approaches that joints retain their structural integrity, yet evidence from the collapse of the World Trade Center buildings and full-scale tests at Cardington indicates that joints may be particularly vulnerable during both heating and cooling. If joint failure occurs, the assumed response of the structure will not be able to develop fully, thereby compromising safety levels. Furthermore it is important that the fire should be contained within its compartment of origin, and the physical integrity of floors needs to be maintained, even at very high distortions. If joints fail, deformations locally are likely to be increased dramatically. Whilst there may be sufficient redundancy within the structure to sustain this, the concrete floor slab may have very limited ductility and may not be able to accommodate such deformations without significant cracking, causing loss of compartmentation. Ultimately, joint failure can precipitate failure of the structural floor system, which may in turn either overload lower floors causing progressive failure or may allow the supporting columns to buckle, leading to a much more extensive structural collapse.
For simple braced frames the joints typically possess very low rotational stiffness and are assumed to transfer only vertical shear forces into the columns. In fire conditions such joints can be exposed to much more onerous conditions than at ambient temperature. In addition to the development of forces described above, joint rotation is very much higher, initially because of differential expansion of the steel and concrete, and subsequently because of the large beam deformations at high temperature. This, and the fact that connection temperatures are often lower than the connected members, means that significant moments can be generated, even in joints which are designed as ‘simple’ for ambient temperature.
Response of steel composite frames to fire The structural behaviour of steelwork during a building fire is strongly influenced firstly by softening of the material (the progressive degradation of the stress-strain curves as its
54
Vertical Shear Catenary Tension Hogging Moment Local buckling and potential force concentration in top bolts
This can contribute significantly to reducing deflections of the connected beams, although simplified fire engineering design calculations, using a residual end moment for the beam, must be used with caution because of the possibility of local buckling as observed in a number of tests.
simplified temperature calculation methods for joint zones. Three degrees of simplification are allowed. The simplest is to treat the connection temperature as uniform and to calculate using the maximum value of the section factors A/V of the connected steel members. Alternative approaches involve incremental calculations and allow the temperature to be estimated for the different connection components separately, or for the connection as a whole but with a linear or bi-linear temperature gradient through the depth of the connection.
Little research has so far been done on the behaviour of semirigid and rigid joints in fire, and it is doubtful whether they can maintain their rigidity in fire, when local buckling is likely to occur at the beam-end adjacent to the connection.
Observed behaviour in Cardington fire tests
Behaviour of joints in fire
Partial-depth end plates were used extensively in the Cardington fire tests, and in many cases local buckling was seen in the bottom flange of the beams, in combination with shear buckling of the webs. However, given that the beams had been designed as simply supported, this local buckling did not appear to prejudice the overall structural performance.
The principal structural effects which would normally be considered as “failure” at joints are fracture due to tension and shear, and local buckling due to compression and shear. The latter is most likely to occur in parts of the structure which are restrained against thermal expansion. Local buckling of the lower beam flange adjacent to the joint does not in itself constitute a failure in the fire situation, but is known to trigger shear buckling in the web as illustrated. The diagonal tension field action caused by this shear buckling has the potential to concentrate the shear and tying forces at the top part of the connection, especially when the beam is at high deflection and catenary tension is developing. This could trigger a progressive fracture of the joint from the top downwards, which is a genuine structural integrity failure. Depending on the design details, this could involve failure of the bolts, bolt-holes, welds, beam web or end-plate. Even if no fracture occurs during the heating phase, the same progressive fracture can take place during cooling. Under unfavourable loading conditions, e.g. explosion induced damage and/or fire induced weakening of materials, this joint fracture may lead to progressive structural collapse, and when designing the building for robustness this is a key consideration.
During cooling a partial failure of these connections was often observed, in which the end plate fractured on one side of the beam web adjacent to the welds, as illustrated. In no case did this happen on both sides of the end plate. This suggests that
Connection temperatures The temperature distribution in a joint zone is usually considerably lower than that of the members it connects, especially if some of these members are unprotected, because of the local concentration of material and relative lack of exposed surfaces. Eurocode 3 Part 1.2 provides
Partial fracture on an end plate connection after cooling
55
the increased flexibility produced by the fracture allowed the tensile forces developing in the cooling beam to be relaxed through deformation of the joint – i.e. the remaining connection performed in a ductile fashion and was sufficient to transmit the structural actions. Fin plates were used at Cardington to connect secondary beams to their supporting primary beams. In several cases it was observed that the bolts had fractured in shear at the interface between the fin plate and the beam web, as illustrated. This again is thought to have occurred as the secondary beam contracted during cooling, but in other cases might happen as it expands during the heating phase. Fin plates rely on steel in direct tension and shear, and so will always behave in a less ductile fashion than a bending element such as a partial-depth end plate. There is also no mechanism of partial failure which would allow the development of ductile behaviour as was the case with the partial depth end plates.
Shear failure of bolts in fin plate connections in cooling
Observation of behaviour at World Trade Center The final report of the World Trade Center Building Code Task Force identified a number of issues relevant to the safety of all buildings, including the need to understand connection performance under impact loads and during fire, and to quantify this for design. It identified joints as critical components in structural frames, and referred to the importance of assessing performance of structural members and connections as part of a structural system in building fires. At a more detailed level the report provided evidence that bolted shear splices can be vulnerable to catenary forces in fire if they are not designed for an appropriate tying capacity. Building 5 of the World Trade Center, which was subject to a severe fire following the collapse of the Twin Towers, was constructed using pre-fabricated “column trees” supporting suspended beam spans. The joint illustrated was fire protected but failed as a result of the effects of fire alone. The upper storeys of the building were not damaged and remained intact while several floors below collapsed during the fire. It is probable that the failure initiated at one level and the resulting impact on successive levels below led to progressive collapse. The connection utilised a simple shear plate, designed to transfer the vertical shear force along the member. The eccentricity between the bolt rows generates a small moment, but otherwise the bolts and bolt hole positions are designed for vertical forces only, which explains the “nominal” edge distance between the bolt holes and the end of the steel plate. Under the catenary forces generated at high deflection, combined with vertical shear, these edge distances are clearly inadequate and block shear failure occurs.
Column tree detail from WTC5 (from FEMA Report)
Failed tab plate and column tree edge from WTC5 (from FEMA Report)
56
Alternative simple beam-column joints: seating/locating cleats, sliding bearing, seating bracket/web cleat, slotted fin plate
As temperature increases, the strength of bolts and welds is assumed to degrade according to strength reduction factors similar to those for normal structural steel. In the case of friction grip bolts, it is assumed that the heating effectively relaxes the contact pressure, so that they are assumed to have slipped. Under fire conditions, such connections are therefore treated in the same way as for ordinary bearing bolts.
Suggested variants of simple joints For optimal behaviour in fire a simple joint ideal needs the following characteristics: • High rotation capacity to cope with the large beam deflections in fire; • Relatively flexible behaviour in horizontal tension, so that catenary tension is reduced by allowing the effective shortening caused by large deflection;
Moment-rotation at high temperatures
• Sufficient strength when distorted to resist the catenary tension; and
As structural fire engineering becomes increasingly based on the behaviour of whole structures, there is an increasing need to understand how joints behave in fire. Research is currently progressing in this area but detailed design procedures have yet to be developed. Adopting a similar approach to ambient temperature studies on real connections, the work initially centred on moment-rotation characteristics and the effect of these on limiting beam deflections. Data has been gathered from full-scale furnace tests, typically on cruciform arrangements, and finite element analyses. These ignored any effect of axial thrusts. Based on these studies, semi-empirical rules have been developed, allowing the progressive degradation of strength and rotational stiffness
• Sufficient vertical shear resistance to carry the vertical load component appropriate to the fire limit state at the final steel temperature. Some suggested details for beam-to-column joints are illustrated. The details shown are variants which may be useful in providing some of the above requirements in particular circumstances. The slotted fin plate shown and the seated web cleat can be designed to give the flexibility in rotation and axial movement which is required to minimise catenary forces, together with the final strength to resist these forces. The sliding bearing can accommodate high axial movement and rotation, but has no tying capacity other than that provided by the slab reinforcement.
Eurocode requirements Eurocode 3 Part 1.2 has relatively little to say about joints, in contrast to the highly advanced treatment which is possible for joints at ambient temperature under Eurocode 3 Part 1.8. Annex D is “informative” and deals only with simplified connection temperature calculation, and the reduced strength of bolts and welds at elevated temperatures. It does not allow any of the load-deflection behaviour to be predicted.
Moment-rotation curves for steel-to-steel and composite beam-column joints (see Al-Jabri 2002 in References)
57
to be represented as illustrated. Composite connections have also been considered, but with less consistent patterns of behaviour than for the bare steel case.
Each of the components is represented as a non-linear spring, with its own strength and stiffness in tension, compression or shear, which degrade as the temperature rises. Thus, combinations of moment and thrust are simply different combinations of the horizontal forces in each of these nonlinear springs, as illustrated.
The results have typically been presented as a family of curves for specific connection details, with individual curves representing a particular temperature. Whilst this work has provided a very sound basis for understanding some of the ways in which fire influences joint behaviour, the Cardington full-scale fire tests have demonstrated th at design procedures based on isolated member behaviour have limited validity. For joints this means that rotational behaviour has to take account of the accompanying axial forces caused by restrained expansion, potentially a very significant effect. The development of moment-rotation-thrust surfaces at different temperatures would require prohibitive numbers of expensive tests.
Simplified analytical models have been developed for the characteristics of some of the main components of flush and extended end plates at elevated temperatures, and these have been validated against furnace tests and against detailed finite element simulations. Components which have so far been evaluated in this way are: • the tension zone comprising the end plate, top bolts and column flange, • the compression zone in the column web, in the absence of high column axial force. These are sufficient to represent the moment-rotation characteristics without axial thrust in the beam or column, and compare very well with the earlier furnace tests on cruciform arrangements as illustrated. Further components are currently being investigated and these will enable a more generalised approach to be developed for a range of different connection types under various combinations of axial load and bending moment.
Component-based approach More recently the focus has been to consider the joint as an arrangement of different zones, each contributing to the overall behaviour, allowing simpler treatment of the large number of variables and a clearer assessment of robustness. This is a more practical alternative and is a logical extension of the principles of the “Component Method” of joint analysis used at ambient conditions to design at elevated temperatures. In principle any joint can be considered as an assembly of individual zones, each including several components, as illustrated. A steel joint under the action of a member endmoment alone can be divided into the three principal zones shown: the tension, compression and shear zones.
Comparison of component-based joint model against momentrotation characteristics taken from cruciform tests The three zones and components in an end plate joint under moment
Conclusion The principles of the Component Method can be used directly in either simplified or finite element modelling, without attempting to predict the overall joint behaviour in fire, to enable semi-rigid behaviour to be taken into account in the analytical fire engineering design of steel-framed and composite buildings. This approach is also being used to examine the ultimate strength of joints and this will be of major importance in considering the robustness of buildings.
Illustration of the action of component springs under moment and thrust
58
CHAPTER 13 Vibrations
Acceptability of floors for walking vibrations The basic design procedure for assessing the vibration performance of floors subjected to human activities is as follows:
Stephen Hicks, SCI
Introduction
Frequency of the floor structure
Specifiers are more frequently facing the need to consider the vibration response of floors in detail. The key issues are summarised here with a consideration of the implications for steelwork connection details.
The frequency determines how the structure will behave when subject to human activities. For structures that possess a ‘low frequency’ of between approximately 3 to 10 Hz, harmonic components of the walking force can coincide with the floor frequency causing resonant excitation; in this case, the level of damping is important as it defines the amount that the response will be magnified (the lower the damping ratio, the higher the response).
Any structure will vibrate if subjected to cyclic or sudden loading. In most cases, the vibrations are imperceptible, and so can be neglected in building design. However, in some circumstances, the response of the structure to a common cyclic load (e.g. walking activities etc.) is sufficient to produce a response that is perceptible to the occupants of the building. The most common example of such small, but perceptible, vibrations occurs in floor structures. This is not a new phenomenon, but may be more noticeable within the working environment of modern offices and residential dwellings as spans have increased.
For structures that possess a sufficiently ‘high’ frequency (greater than approximately 10 Hz), such that it is effectively ‘tuned’ out of the main harmonic components of walking force, the response is dominated by impulsive excitation. In this case, the response is relatively insensitive to the level of damping.
Modal mass of the floor
A simple model that illustrates the vibration behaviour of a floor structure is shown in the figure. The bending stiffness is modelled as a spring of stiffness k, and the floor mass is modelled by a point of mass m. All practical structures will have some damping, conveniently modelled as a viscous (or oil-pot) damper. Damping refers to the loss in mechanical energy within a mechanical system. Practical floor structures possess a low level of natural damping (normally in the order of 1%).
For both resonant and impulsive excitation, the amount of mass that is mobilised is needed to calculate the response (in a similar way to the damping ratio, the lower the modal mass, the higher the response).
Floor response The response is simply calculated using Newton’s second law of motion in that acceleration is proportional to the force (from the walker) divided by the vibrating mass of the floor. • For ‘low’ frequency floors (frequency fº between 3 to 10 Hz), the root-mean-square (rms) acceleration is given by:
arms =
αn P0 1 √2M 2ζ
(1)
where αn is the Fourier coefficient of the nth harmonic (which can be taken as 0.4 for floors with f0 < 3.55 Hz or 0.1 for other cases), P0 is the static force exerted by an ‘average person’ (normally taken as 76 kg × 9.81 = 745.6N), M is the modal mass and ζ is the damping ratio.
Idealised single-degree-of-freedom system (a) basic components; and (b) undamped free vibration response
If the spring-mass is set into motion in some way and it is assumed that there is no damping, the resulting vibration response is shown in the figure. As can be seen the period T defines the time taken for one complete cycle of oscillation to occur. In design, the inverse of the period is normally considered, which is defined as the frequency; this parameter provides a measurement of the number of cycles per second (or Hertz) the structure will vibrate. However, the most important parameter of vibration response is the amplitude of the motion, as this is used to assess the acceptability of the structure. Although there are many possible ways in which the magnitude of the vibration response can be measured, it is often convenient to describe the amplitude of the motion in terms of acceleration.
• For ‘high frequency’ floors (frequency f0 > 10 Hz), the rms acceleration is given by:
arms = 2πf0
I M
(2)
where I is the impulsive force (in Ns) from the walker (which can be taken as I = 190 / f01.3)
59
Frequency-weighted floor response The human perception of the floor response is affected by both the direction and frequency of the vibrations. This is accounted for in the current vibration standard BS 6841 (which is similar to ISO 2631-1) by adopting the ‘basicentric’ coordinate system shown in the figure below (note that the z-axis corresponds to the direction of the human spine), and an appropriate frequency-weighting curve such as that also shown in the figure. Since weighting factors are ≤ 1.0, a factor of 1.0 may conservatively be applied for preliminary design. However, should a more systematic check be made, the rms acceleration should be multiplied by the following factors for z-axis vibrations: Weighted
arms = arms × 0.5√f0
BS 6841 (a) directions of basicentric coordinate systems for vibrations influencing humans (b) frequency weighting characteristics for z-axis vibration
for 3 Hz < f0 < 4 Hz
or Weighted
arms = arms
or Weighted
arms = arms
×
for 4 Hz
8 f0
For x- and y-axis vibrations, Weighted for f0 > 3Hz.
≤ f0 ≤
accepted multiplying factors cited in BS 6472 for different environments corresponding to a ‘low probability of adverse comment’ are provided in the table below (note that calculated multiplying factors greater than those shown correspond to an unacceptable human reaction if the vibrations are applied continuously).
8 Hz (3)
for f0 > 8 Hz
arms = arms
×
f0 2
Vibration dose values The multiplying factors presented in the table above are based on continuous vibrations, and are therefore appropriate for floors that are very heavily trafficked with walkers continually present. For less heavily trafficked floors, walking activities will produce intermittent vibrations, and a cumulative measure of the floor response may be made through the use of vibration dose values (VDVs). In these circumstances it can sometimes be shown that the floor would be acceptable, even when the calculated multiplying factor is greater than the values given above.
Human perception limits The current standards define the level of vibration at the threshold of human perception by a ‘base value’ of rms acceleration. According to BS 6472, the base value for z-axis vibrations corresponds to 5 × 10-3 m/s2, while for x- and y-axis vibrations the base value reduces to 3.57 × 10-3 m/s2. Satisfactory vibrations for different environments are defined by the ratio of the calculated weighted rms acceleration to the appropriate base value. These ratios are known as ‘multiplying factors’ or ‘response factors’. Some commonly Place
Time
Operating theatre, precision laboratories Residential, hospital wards
Multiplying factor for continuous vibration 1
Day
2 to 4
Night
1.4
Offices, general laboratories
4
Workshops
8
Place
Time
Vibration dose value (m/s1.75) or VDV z-axis
x- & y-axis
Day
0.2 to 0.4
0.14 to 0.28
Night
0.13
0.09
Offices, general laboratories
0.4
0.28
Workshops
0.8
0.56
Residential, hospital wards
Note: It is not considered appropriate to use a dose value assessment on especially sensitive floors, such as within an operating theatre
60
Simplified approach
A recent study of steel-framed floors (see Ellis 2001 in References) showed that VDVs can be estimated from the following equation:
VDV = 0.68 × a(rms) × t 0.25
The following simplified approach can be adopted to assess the vibration response of typical steel-framed floors.
(4)
Floor frequency The fundamental frequency of the complete floor f0 may be calculated by considering an appropriate mode of vibration, summing the deflection calculated from the beam and slab components, and placing this value within the equation below.
where a(rms) is the weighted root-mean-square (rms) acceleration and t is the total duration of the vibration exposure (in seconds). Should the designer wish to undertake a VDV assessment, the calculated dose values should be less than or equal to the values presented in the table below, which correspond to a ‘low probability of adverse comment.
f0 =
18
(5)
√δ
Total deflection of a floor panel for a variety of framing arrangements Framing arrangement
Secondary beam mode of vibration
4
δ=
Condition when mode shape is governed by the motion of the primary beams
Primary beam mode of vibration
-
-
16b3 Ip ≤ 3 Ib L
3 3 b ωb 64b L L + + δ= Ib Islab Ip 384E
3
ωb 5L b + 384E Ib Islab
As above
Ip ≤
As above
4
3
92b I L3 b
4
δ=
Notes: ω is the load per unit area, b is the spacing of the secondary beams, L is the span of the secondary beams, Ib and Ip are the dynamic second gross second moments of area of the composite secondary beam and primary beam respectively (which may conservatively taken as that used in the static design and increased by 10%) and Islab is the second moment of area of the composite slab (which may be determined from the table below)
Dynamic second moment of area for composite slabs with different deck types Profile type
Deck height, hp (mm)
Dynamic second moment of area per metre width, Islab NWC
LWC
3.7
0.65 h3.5
Re-entrant deck
51
0.37 h
Trapezoidal deck
60
0.23 h3.7
0.40 h3.5
Trapezoidal deck
80
0.19 h3.7
0.37 h3.5
Trapezoidal deck
225
0.05 h3.7
0.12 h3.5
Notes: NWC normal weight concrete, LWC lightweight concrete and h is the overall slab depth
61
3
3 b ωb 368b L L + + Ib Islab Ip 384E
where σ is the total deflection (in millimetres) based on the gross second moment of area of the composite beam and slab (for cases when the floor grid is regular, the value of σ may be determined from the table below), with a load corresponding to the self weight, and other permanent loads, plus a proportion of the imposed load that can be considered as permanent (10% of the imposed load may be considered as an upper limit in modern offices).
in collaboration with Richard Lees Steel Decking, which can effectively double the damping values given above for modern composite floor construction. This product comprises a layer of high-damping visco-elastic material of approximately 3mm thickness, which is sandwiched between two thin steel plates; this is then introduced at the interface between the top of the steel beam and the underside of the composite slab towards the ends of the beam. In the remainder of the span, through deck welded stud connectors are introduced to provide a partially composite beam.
Damping Damping is often the most difficult property to predict for dynamic analysis. Higher damping depends on the energy dissipation through non-structural components such as partitions, which are largely dependent on frictional forces. For design purposes, it is suggested that the following damping ratios (ζ) should be used for estimating the response of ‘low frequency’ composite floor systems:
Modal mass The effective vibrating mass M in equations (1) and (2) may be taken as equal to mSLeff / 4, where m is mass per unit area (kg/m_) of the floor plus any loading that is considered to be permanent. The values of S and Leff should be taken from the table below (reproduced from the SCI guide P-076), where:
• 1.1% for completely bare floors or floors where only a small amount of furnishings are present; • 3.0% for normal, open-plan, well-furnished floors; • 4.5% for a floor where the designer is confident that partitions will be appropriately located to interrupt the relevant mode(s) of vibration (i.e. the partition lines are perpendicular to the main vibrating elements of the critical mode shape).
RFp
is the relative flexibility of the primary beam
S*
is the effective width of the floor participating in the vibration, calculated from the effective slab stiffness, given by:
*
S = 4.5
Damping can be dramatically enhanced by dampers based on: • Inertial devices;
EIslab mf02
1/4
(m)
where EIslab is the dynamic flexural rigidity of the slab in Nm2 per metre width, e.g. for a 140mm NWC slab with a 60mm trapezoidal deck
• Discrete damping elements connecting between two points; • Incorporation of high-damping materials within the form of construction.
EIslab = 0.205
Tuned mass dampers (TMD’s) are common inertial devices. A TMD consists of a mass mounted on a structure via a spring system and a viscous damper, preferably in a location where the floor structure’s deflections are the greatest. The spring and mass are ‘tuned’ so as to have a natural frequency close to that of the floor structure. However, it should be noted that TMD’s have a narrow frequency range of effectiveness and, for floors that may possess a number of modes of vibration, multiple devices may be required to damp an entire floor structure.
L*
× 0.23 × 1403.7 = 4112969 Nm2
is the effective span of the secondary beam participating in the vibration, calculated from effective composite beam stiffness, given by:
*
L = 3.8
Viscous dampers are an example of discrete damping elements. These devices function in a similar way to shock absorbers and dissipate the energy by the movement of a piston passing back and forth through a fluid. Although successfully used on the Millennium Bridge in London, discrete dampers are difficult to incorporate within floors since they need to connect two points that are moving relative to each other (along the axis of the damper) in the vibration mode.
EIb mf02
1/4
(m)
where EIb is the dynamic flexural rigidity of the composite secondary beam (Nm_) and b is the secondary beam spacing (m)
Constrained layer damping can, in principle, be concealed within the structure of the floor. The advantage of this form of damping is that it can simultaneously damp several modes of vibration, thereby reducing the overall floor response. In the UK a system known as ‘Resotec’ has been developed by Arup
62
W
is the width of the floor plate under consideration (m)
Lm
is the span of the primary beam (m)
Lmax
is the total length of the secondary beam when considered to act continuously (m)
Values for dimensions Leff and S used in determining the effective mass of the floor Qualifying conditions
Leff (m)
S (m)
RFp < 0.2
L
S* but ≤ W
RFp > 0.2
L
Greater of S* or
Mode shape governed by motion of secondary beams
Indicative floor layout
Lm but ≤ W
l=L
2L
0.8 L < l < L
2L
As for Case (1) above
Mode shape governed by motion of primary beams
1.7 L l < 0.8 L
L
RFp < 0.6
2L W
RFp > 0.6
L* but ≤ Lmax
W2 = W1
W2 > 0.8 W1
W2 < 0.8 W1
2 W1
As for Case (3) above
1.7 W1
W1
Vibration dose values For cases when the designer wishes to take advantage of the fact that floor vibrations occasioned by walking activities are intermittent, the graphs illustrated may be used. In the illustration the time taken for a walker to walk from one end of a corridor to another is maximized by considering the slowest
pace frequency that may be practically achieved (1.5 Hz). Based on multiplying factors that exceed the requirements for continuous vibrations from BS 6472 quoted above, the number of crossings per hour have been calculated which would provide VDV’s consistent with a low probability of adverse comment also in accordance with the earlier table.
Maximum number of walking crossings per hour for various multiplying factors and corridor lengths for: z-axis vibrations in office, residential and general laboratory environments during a 16-hour day (VDV = 0.4 m/s1.75); and x- and y-axis vibrations in residential and hospital ward environments during an 8-hour night (VDV = 0.09 m/s1.75)
63
Conclusion The engineering connections and architectural details chosen can have an influence on the dynamic response of the structure to vibrations. Factors to be considered that will affect the assessments outlined above include: • The framing arrangements, especially whether beams are designed as continuous or simply supported as this affects the mass mobilised; • Whether beams frame onto columns as primary beams or onto other beams as secondaries; • The stiffness of the beam-to-beam or beam-to-column connections, and whether there is potential for hysteretic damping within the connections; • The choice of normal weight or lightweight concrete for floor slabs on metal deck and whether composite action is required; • The choice of prestressed or simply reinforced slabs when precast units are used; • Whether beams act compositely with the slabs, and whether special hysteretic materials are used to increase damping; • The positions and the connection points necessary for installing and maintaining active dampers.
64
CHAPTER 14 Hollow section joints
Open ends of hollow sections can be simply sealed by use of welded cap plates. Although simple and often non-load bearing, care and attention is required to weld preparation and weld quality to ensure sound non-porous welds are deposited.
Eddie Hole, Corus Tubes
Introduction Joints with hollow sections are often said to be complicated and expensive but in reality they can be simple and cost effective. Any joint can be approached from two directions, first as a basic mechanical joint and second as an architectural joint.
Welding Welding represents a major method by which structural hollow sections are connected and, with the exception of bolted connections, is involved in all forms of hollow section joints. Welding can be considered in a number of stages: design, preparation and detailing, welding procedures, welder approval and qualification and weld testing. The principal reference source for general rules for execution of welded structures and supplementary rules for hollow sections are given in DD ENV 1090-1 and DD ENV 1090-4 respectively. Industry guidance can be found in Corus Tubes publication CT15 SHS Welding with specification guidance given in the National Structural Steelwork Specification.
Basic mechanical joints serve to connect two or more components in the simplest and therefore usually cheapest way. Plates are cut by cropping or profile burning and, whilst sharp edges and burrs are removed, no additional shaping is made. Joints are made in the most accessible way for shop fabrication or site erection and will normally be left exposed. Given freedom of choice and considering that, in today’s economic climate, it’s the lowest bid that usually wins, steelwork contractors will, unless requested otherwise, plan for basic mechanical joints. In many cases hollow sections are being used for their aesthetic appearance and in these cases care is required in the choice of joints that must still develop the mechanical strength required but, dependent on their locality and proximity to public attention, may need further attention either by simple means of shaping plates or by totally enclosing the connection. Whilst it is true to say that very few things are impossible, there is a financial price to be paid for the increased workmanship required in architectural joints.
In-line jointing Introduction As most structures are higher or wider in dimension than the length of an individual hollow section component, they will need to be jointed. A common connection will be an in-line joint of two similar sized or possibly different sized hollow sections.
Many architectural joints are often approached with the aim of being “unseen” and, as a result, a number of the simple but exposed types are not used. In reality, when the actual joint is seen in the context of the whole structure, the more basic joints are often more than acceptable and can even be used to create an architectural expression. In essence good architectural detailing can, provided a sense of proportion or scale is maintained, be achieved without incurring substantial extra costs.
Welded butt joints In-line joints of equal size members are easily made by full penetration butt joints. Full use should be made of the fact that for a given size of section the thickness change is internal and not external, thus a constant external dimension can be achieved even when differing thickness are being joined. Use of these techniques, especially when used at natural joint positions (i.e. at ends of individual mill length members), can achieve clean lines with economy.
The choice of joints and the means of connecting individual members can have an influence on the choice and the performance of the corrosion protection system used. Cleaner lines, simpler details and absence of areas that entrap water or debris contribute to a longer life for the chosen protection system. This is essentially true for external structures but also will have an effect for internal steelwork.
up to
Also of consideration for external structures is the effect of using of jointing systems or details that penetrate the hollow section wall or leave open-ended sections. Many details will automatically seal a hollow section and in this situation no internal protection is required. Where however the jointing system penetrates the tube wall without sealing or leaves the section open, consideration must be given to the question of water ingress and internal corrosion protection. In general it is often not internal surface corrosion that is the issue but rather the effects of entrapped water. External structures can experience lower temperatures than internal steelwork and entrapped water, which can freeze, can cause deformations or lead to splitting of members.
up to
up to
over In-line welded joints with backing and weld preparations
65
In-line joints need careful alignment and jigging to ensure correct fit-up. This is not a problem in the workshop where equipment is readily available and components can be jig assembled. For site work provision is needed for temporary cleats to provide alignment and restraint or scaffolding for temporary support. Site work will also be affected by weather variances and consideration must be given to welding and weld procedures.
Flanges can be of the ‘blank’ or ‘ring’ type, the latter being useful for external elements such as tubular towers that require galvanizing both internally and externally. Ring flanges are thicker than blank flanges but have the advantage of permitting welding from both sides, thus reducing the effects of distortion caused by single sided welding needed for use of blank flanges. Ring flanges also permit internal galvanizing but when bolted together cannot be guaranteed to stop ingress or water. Attention must therefore be paid to water drainage in external structures. Some typical examples used for lattice towers and masts are shown later in this publication.
Flange joints
Assuming the same grade of material is used for the flange as the main hollow section, a minimum tube thickness of three times wall thickness is required for ring flanges and twice wall thickness for blank flanges. Thickness is based on strength and may need increasing to control distortion.
Flanges are particularly suited to site joints as the final connection is by bolting. In-line bolted flanges can be of any shape or size but those most generally used are rectangular, square, round or triangular. They should be kept as small as possible but have sufficient thickness to give adequate joint strength and prevent distortion during welding.
Similar joints can be used for RHS and CHS in lattice construction. Whilst in close up these can look bulky, in overall perspective their effect is diminished. Flange joints can equally be used in lattice or multi storey construction with either the same or different size members. For multi storey construction the flange joint can normally be accommodated within the floor construction depth. Flanges can be blank or, where composite concrete filled columns are used, provided with a central hole or be ring type to permit filling. At heads of columns ring flanges can provide hand access to permit bolting of connecting floor/roof beams.
Flange bases Column bases are effectively flange joints welded to the column, either thick enough to withstand any direct load and bending or provided with stiffeners to reduce the bending and plate thickness used. Rather than using welded stiffeners, it is generally more economic to use thicker plates for flange rings or blanks, except where the plates become so thick that preheat and other special welding procedures are required.
In-line flange joint
Splice joints As an alternative to flanges, in-line joints can be made by bolted splice joints. Plates, angle sections or stub column sections welded to closing plates on the main member are connected by use of bolted splice plates. The joint can either be left exposed or, by use of connection cover plates, have a smooth external appearance.
Secondary joints Secondary joints frequently referred to as fittings, cleats or brackets often form a part of many joints whether for main components or secondary items. They can vary from simple flat plates or off-cuts of rolled sections to complicated brackets. As with all joints simplicity is the key and simple joints are often the most economical and can be visually acceptable. Also, as a further bonus, a simple joint usually creates a direct load path between the elements being connected.
Flanged CHS member connections
66
End forks with single or multiple bolts or pins can be made from standard rolled sections or from plates. Double forks reduce plate thickness and put bolts or pins into double shear. Single forks can be either on or off centreline. End plates or sections also seal the SHS member and negate internal corrosion, noting that sealed SHS members must be vented if galvanized.
Whilst overlapping of bracings is usually not preferred, the joint strength can be increased by the bracings overlapping. This will however increase the fabrication work due to the double shaping required unless a full overlap is made. Full overlaps can reduce the double shaping required but will mean the cutting of members and the sequence of fitting members into the truss during fabrication has to be carefully controlled.
As an alternative to adding plates, ends of CHS members may be fully flattened and drilled or punched to make bolted connections. Cold flattening may be used, but repair by welding may be required on the weld and outer edges of the tube due to the amount of cold working carried out.
When two or more bracings overlap using a central division plate can reduce the shaping required. This is especially useful for overlapped CHS bracings or where an additional vertical bracing is required when a division cap plate can be added to form a T connection.
Joints in lattice construction Joint strength
Introduction
As with all structural joints, the capacity of the hollow section joints must be adequate to resist the loads being imposed through it. The strength capacity of such joints is dependent on a number of co-existing issues being: the relative size of the bracing and chord members, the angles of the bracings, whether the bracings have a gap between or overlap each other, the thickness of the chord and the order of the chord load if in compression.
Structural hollow sections are highly efficient in compression as well as tension and the added fact of a clean, slender appearance means they are regularly used in lattice construction. A warren bracing system is generally preferred as, being highly efficient in compression the hollow section strut member need not be constrained to the shortest possible length as with the N brace systems. In addition, it can reduce the number of bracings to be fitted when compared to an N brace system. Small vertical members can be used at purlin positions to remove the top chord bending but these may increase joint complexity.
As previously noted, the member selection is usually made on the basis of centreline noding which will, dependent on member sizes and angle, produce a gap or overlap condition. The resultant condition will directly affect the joint capacity as will the issue of chord face deformation, which is affected by the ratio of the bracings to the chord. All of these factors are determined by the selection of the relative main member sizes and therefore, unlike joints in open profiles, the designer pre-determines the capacity of the joint (perhaps without realisation) when selecting members whilst carrying out structural frame calculations and initial member load capacity checks during the scheme design.
Lattice bracing joints are basically in two types, gap joints or overlap joints. The gap joint is the easiest for fabrication as it confines the work to single bevel cutting or end profiling of the bracing members. A minimum bracing angle of 30° is recommended to ensure adequate access for welding. The design procedure usually involves initial sizing of members on the basis of centreline noding (i.e. intersection at a single setting out point). For a given combination of bracing angle and member sizes this may mean that bracing members are overlapped at the chord face connection. Creating an eccentricity and producing a gap joint can often eliminate this overlapping. Such eccentricities need to be included in the chord design check, but the design rules allow this check as well as provision to take eccentricities within the depth of the chord into account.
This factor has too often caused a problem with hollow section construction. The problem being that to remedy insufficient joint capacity in the main member it may be necessary to increase chord thickness, change bracing sizes or joint geometry or indeed a combination of all. Such changes are relatively simple (and economic) to effect at the design stage
Warren braced truss configuration
Central plate reduces fabrication complications
67
but, due to the nature of the hollow section profile, are not easy to affect an increase in capacity once members have been selected and purchased. It is therefore essential that the designer making the structure calculations undertakes initial checks of the overall joint capacity at the time of structural design. Local connection design of the welds between components within the joint can then be confidently passed to the fabrication detailer. Design rules have been established for most of the common range of joints in lattice construction. Using the joint geometry, member sizes and co-existing member loads the capacity of the joints can be established. Formulae and joint parameter limits are given in: Eurocode 3 which is based on CIDECT design guide publications or the Corus Tubes publication CT16 Design of Welded Joints.
Short length CHS for multiple connections clearance
This can reduce the bracing complexity and also allows the girders to be split for fabrication and transport. Where multiple bracings occur the bracing connection can be moved back from the intersection point until a simple joint is formed. This can be achieved by introducing a short length of CHS or by employing hollow spheres. Spheres give the advantage of a square cut bracing end regardless of the intersection angle but are limited in their source.
Profiling For joints involving CHS members end profiling needs to be considered, although not all bracings need to be profiled or saddled. When bracings are between 1/3rd and 2/3rds of the chord diameter partial profiling is needed and when over 2/3rds full end profiling is needed to achieve the weld fit up requirement. For bracings smaller than 1/3rd of the chord diameter a single end cut will normally achieve the required fit up which is then filled with the attachment weld.
Beam-column joints A wide range of varieties of end plates, cap plates and shoes may be used with structural hollow sections to make joints between columns and rolled beam sections or fabricate trusses/lattice girders. In part these alternatives will vary dependent on the form of the column profile, i.e. if the column is an open UC or UB profile or hollow section of either a circular or rectangular form. There is scope for the designer to exercise ingenuity to produce aesthetically pleasing connections, which will keep fabrication to minimum, but which are adequate to perform the task for which they are required.
Square cut, partial or full profile cutting ranges
With CHS bracings the technique of partial flattening can be used to increase welding access and reduce/avoid profiling or double shaping. This method will however increase the bracing width and the resultant fit-up should be considered. Joint complexity can increase with multi-planar joints such as those found in triangulated girders. In these cases necessary weld clearance can be obtained by partial flattening of the ends of the bracings, double shaping the bracings or alternatively the main member diameter may be increased to ensure necessary weld clearance is obtained. When RHS members are used in triangulated girders, the selection of member sizes for the bracings and chords must be done by considering their relative sizes and fit up. Failure to do this can cause extra fabrication and complicate details. Rotating the chord member and avoiding expensive ‘bird mouthing’ can often eliminate additional work. It should always be remembered that chords could be made of twin members.
T end and angle cleat connections
68
Design rules for joints in simple construction have been developed for joints between beams and hollow section columns. These are published by SCI and BCSA in the Green Book series which include design procedures and examples. In all cases it is the joint area that requires the fabrication work. In many cases pre assembled joint locations can be made and added to the overall column – hence minimising the need to handle the total column length. Beams can be joined to SHS columns through end plates or angle cleats in conjunction with proprietary mechanical methods such as Hollo-bolt or bolted systems such as Flowdrill (see Flowdrill and Hollo-Bolt jointing below). Where joints are located at the end of the section, hand access for installation and tightening can be obtained and a single side fixing made. Cover plates can be used to restore the visual appearance of clean lines. For joints at other locations a joint box technique (using a heavier wall insert and a hand access hole) can be adopted. Where four-way beams connect the final face will require captive nuts, a threaded column wall or a one sided mechanical fixing to be used.
End plate connection to welded stub
When using UC or UB profiles, connections can use the flange to transmit the loads and it becomes, effectively, a simple plate flange. Such connections are equivalent to flexible end plates or angle cleats and are probably the cheapest type of joint. Care must be taken during fabrication to avoid cumulative tolerances, which may give rise to difficulties during erection and necessitate an excessive amount of ‘shimming’. It should also be borne in mind that the connecting bolts transmit all the internal loads in the truss or girder.
Mechanical fixings A range of mechanical fasteners or fixing systems is available from specialist suppliers. Such mechanical fixings may be used with CHS but the flat surfaces of RHS are better suited to their use. Generally fixings can be put into two categories, those that require special tools and those that can be used in normal clearance holes without special tools. Details are shown below. Through bolting can be used as an alternative to single side fixings, the section is drilled and spacer tubes welded in position. This method prevents distortion of the RHS and seals against water ingress. Alternatively, where thicker sections are used, the walls of the sections may be drilled and tapped, or, if the section thickness is insufficient, threaded pads or nuts can be welded on. Studs can be welded onto the section face. Some methods leave a collar at the root and as a result holes in connection plates must be chamfered or recessed to clear the collar or clearance washers fitted.
Double angle cleat connections using blind bolting
When RHS columns are used it is possible to bolt through the column wall as with open sections but from a single side. A range of single-sided connections is available for such joints and further information is given below. Alternatively, face connections can be made by combination of plates or section off cuts. Combinations of simple connections, often exposed, can with ingenuity create aesthetic and highly visual structures. As with open profile structures, beams can be twin channels (parallel beam system) that provide access for vertical services and can be connected to the column through a simple plate welded to the column and bolted to the beams.
Cladding may be fixed direct to hollow sections by using selfdrilling self-tapping fasteners. It is important that these use twin seal washers to avoid water penetration. A wide range of other fixings including J and U bolts is available.
For joints in simple construction the fin plate is the generally chosen method. In such cases flexibility of the connection parts must be developed with sufficient robustness. In the case of RHS and CHS columns checks are required on the local connection face for tearing etc.
Flowdrill showing thermally drilled holes before and after threading
69
Bolted connection systems Flowdrill and Hollo-Bolt give a choice of two methods to produce single sided bolted joints in hollow sections. Both systems offer the following benefits: • They produce bolted joints of structural capacity in hot finished structural hollow sections (HFRHS). • They minimise the change in the fabrication process by using connection details which are standard in the construction industry.
Hollo-Bolt
Castings can range from those with simple form to those having complex highly intricate forms. The final shape can accommodate blending of section profiles which reduce stress concentrations, a particular advantage when considering fatigue sensitive structural joints.
• They reduce fabrication by removing the need to weld plates or other fittings onto the outside surface of the RHS. • They simplify erection by using fully threaded bolts – an increasing practice in the construction industry.
Metal can be added at required points and to high dimensional accuracies, giving thickness variations where required, without excessive weight penalties. A wide range of surface finishes can be achieved, depending on the casting process employed. Advice and guidance on castings can be found in SCI publication P-172.
• They maintain aesthetics by producing a flush face on the RHS after fabrication. The design guidance for Flowdrill and Hollo-Bolt systems with grade 8.8 bolts in conjunction with Corus Tubes hot finished structural hollow sections is given in the BCSA/SCI Green Books. The guidance results from Corus Tubes initial research work undertaken on connections with CIDECT and it has been consolidated into the published information. Procedural checks are given for bearing, shear and local bolt pull out of the RHS wall and for the combined effect of the column axial load and the structural integrity tensile load of BS 5950-1. The combined check for the column axial load and the structural integrity tensile load recognises that the flexibility of the RHS face caused by the tensile load can, in the presence of the column axial load, reduce the overall joint capacity. Both Flowdrill and Hollo-Bolt use fully threaded bolts that allow standardisation of bolt lengths throughout the construction. Where beams are connected to adjacent faces of an RHS column a check must be made with the chosen bolt length to ensure that assembly is possible. Both Flowdrill and Hollo-Bolt are suitable for use with the standard grade of Corus Tubes structural hollow sections to BS EN 10210-1 of S355J2H. At present, application of the Flowdrill process is limited to RHS thicknesses up to and including 12.5mm. For thicknesses of 16mm and over, conventional drill and tap methods are recommended, although due to the RHS material strength being lower than that of the grade 8.8 bolts, pull out strengths may be below the bolt tension capacity.
Cast joint in CHS roof structure
Conclusion Hollow sections are widely used because they combine clean aesthetic lines with structural efficiency, particularly in welded truss configurations. However, achievement of clean lines relies on not having to provide extensive stiffening in the joint zone where main members interconnect. These zones are precisely where the material is working hardest as it copes with global forces as well as the local forces arising from eccentricities, punching etc. created by the detailed joint configuration itself. As it is generally not possible to decouple these detail design factors from frame design with hollow section joints, the scheme designer needs to pay early attention to the implications of connection details by basing main member thicknesses on typical connections.
Castings Castings can be effectively used in tubular jointing and are particularly useful for multiple member joints where joint complexity is high. They can be designed to locate the hollow section ready for welding and allow graceful joints to be made. Economy depends on complexity of joints and numbers of castings required. As an alternative for small numbers, it may be economic to machine connection fittings of similar size and strength from solid steel blanks. Care is needed with both machined and cast fittings to ensure that all material properties (including weldability and toughness) are suitable for fabrication and use, particularly with thicker items whose metallurgy is constrained by the production route.
70
CHAPTER 15 Bridgework connections
a 1 in 4 taper in the thicker material but the designer should be aware that there are increased costs associated with this, particularly where there are significant differences in plate thickness at the joint.
Richard Thomas, Rowecord Engineering Limited
Transverse web and flange butts are welded in this way prior to girder assembly, usually using a mechanised submerged arc process for speed and integrity, although there is no reason not to use an alternative arc welding process provided the appropriate procedure approval is in place and inspection and testing is conducted in accordance with the specified requirements. Non-destructive testing is carried out using ultrasonic testing for sub-surface examination or magnetic particle inspection for surface examination. For bridgeworks, destructive testing of production run on/off extension pieces is undertaken to confirm the mechanical properties of the welded joint.
Introduction The nature of bridgework manufacturing and site installation substantially dictates the types of connection and joints required for assembly of component parts. Welding is normally the key workshop process to manufacture all types of bridge structure; but at site there are design and economic issues to consider before choosing a welded or bolted connection. The design and construction process has to achieve a balance of economics, productivity and integrity to ensure the best possible solution. Most major steelwork contractors have sophisticated numerically-controlled machinery to cut, saw, drill and weld sections and plates with a high degree of precision and repeatability. Simplification of connection design enables the steelwork contractor to maximise the use of machinery and therefore produce quality components at budgeted cost.
Shop web to flange joints Fillet welds are usually specified for these joints. Butt welds are very expensive in comparison not only in terms of weld metal volume but also in the additional cutting operations to produce the preparation and the increased inspection and testing required to demonstrate weld integrity.
Leaner construction programmes and cost conscious customers demand nothing less; but it is important to involve the steelwork contractor at an early stage of the work to ensure that production performance achieves design and functional requirements.
Many steelwork contractors use purpose-built mechanised equipment to assemble and weld plate girders into the Iconfiguration. These machines are usually capable of producing Z-girders or “off-centre” flanges and can also accommodate a degree of camber inherent in the design of the girder. Plate thickness changes are an added difficulty but experienced steelwork contractors devise methods to overcome the problem, generally preferring to maintain overall girder depth constant and shaping the web plate to accommodate changes in flange thickness. Web plate thickness changes are generally shared either side of a common centreline. Most machine welds are deposited using the submerged arc process and the steelwork contractor then takes advantage of the deep penetration characteristics of the process to achieve the design throat thickness using a smaller weld size.
The purpose here is to summarise the significant types of connection/joint required on various bridge structures and to highlight the advantages or disadvantages to the designer and steelwork contractor. Major steel bridges in the UK are predominantly plate girders rather than box girders but both are important. Typical details for tubular truss footbridges are also considered.
Plate girders Shop in-line joints The design of plate girders often necessitates making up available material lengths to suit the span or size of the structure. The position of these joints is usually agreed by the interested parties and to suit material availability. Transport and site constraints may also influence this decision. In workshop conditions, the steelwork contractor prefers to weld flange and web full penetration joints, either in a single or double vee configuration depending on the plate thickness. Typical weld preparations are illustrated.
If there are no machine resources or perhaps the parameters of the girder are beyond the capability of the machine in terms of dimensions and/or weight, steelwork contractors resort to traditional methods of manual assembly.
Stiffener configurations
Thicker joints necessitate in-process turning to minimise distortion and to maintain plate flatness. For bridgeworks changes in material thickness are accommodated by machining
Web stiffeners are provided to improve the web capacity as well as to provide an attachment for bracing systems. Bearing stiffeners are provided at support locations and these are normally “fitted” to the bottom flange to distribute the bearing loads. Flat plate, angles, and tees are used to form the web stiffeners. steelwork contractors usually prefer to manufacture stiffeners from plate material as this is cheaper to purchase and modern process machinery can produce component parts quickly and accurately. In addition, web stiffeners produced from plate are frequently shaped to provide adequate area for connecting bracing systems.
Typical weld preparations
71
Typical stiffener details
The corners of the stiffeners need to be shaped in order to miss the flange to web weld. Either cope holes or sniped/ chamfered corners are used to achieve this. In the case of the radiused cope hole it is difficult to prepare the surface and apply adequate protective treatment to the internal face of the hole. Large clearance snipes, whilst being easy to cut, are not suitable for welding and testing and are even less likely to receive adequate protection. The preferred method is thus to cut a small snipe to just clear the web to flange weld and to completely seal by welding continuously through the corner. This detail is much easier to weld and inspect and enables effective corrosion protection.
Typical bracing arrangements
Cross girder connections
Bracing systems
The configuration of cross girder connections is to a large extent a function of loads to be transferred. The arrangements can become quite complex with significantly skewed structures which have to accommodate a camber. Simplicity is again the key for steelwork contractors to manufacture and erect cost effectively and bolted connections are preferred.
Bracing systems are usually provided at the supports as well as intermediate positions within the span. Bracing members at supports provide a load path for the transverse loads to the bearings/substructure. Bracings at intermediate positions are provided to stabilise the girders in the span during the erection of the structure prior to casting the composite deck as well as to provide a load path for the transverse loads. It is good practice in fabrication and erection at site to keep the bracing systems as simple as possible and generally these are bolted connections with possibly a welded sub-assembled frame detail. Maintaining square ends of sections, minimising the numbers of holes and providing adequate access for fastener installation and tightening operations are the principal rules for simple solutions to bracing design. Some typical arrangements are illustrated
Connecting the cross girder to one side of the web stiffener is the most economical option. A variation is to connect to both sides using splice plates in which case packing plates may be required to overcome any differences in web or flange thickness. The number of bolts in the connection may increase but the overall cost is likely to be less than that for an endplate connection where a tee section web stiffener is necessary in the plate girder cross section. Typical arrangements are illustrated.
Typical cross girder arrangements
72
Site welded joints
benefits is necessary to offset the increased equipment costs. Gas shielded processes are more susceptible to adverse weather conditions and therefore shelters need to be more efficient in preventing draughts.
The design of site welded joints is dependent on several factors including the location of welding, for example, whether the joints are welded at ground level for subsequent lifting as long components or whether the joints are welded in situ with the girders erected. Temporary works considerations include the structural stability in the unwelded condition, shelter and protection from the elements and method and means to accurately align the joint prior to welding to ensure that construction tolerances are achieved.
In all cases inspection, non-destructive testing and the provision of samples for destructive tests are required as for shop welded joints. It is desirable to leave welds in the as-welded condition with neat capping runs, however if welds are required to be smooth, careful flushing by grinding is necessary to maintain the plate thickness at the joint.
It is good practice for the joint configuration to be staggered as illustrated and this also provides a landing support, although it is sometimes necessary to provide additional support where flanges are thin and liable to bend. It is also necessary to provide a draw cleat arrangement to provide basic stability and initial positioning. In addition, cope holes are required in the web at the flange centreline to enable the flange weld to be continuous. The designer instructs whether these are filled after main component welding.
Site bolted joints Main girder connections are usually formed by drilling the ends of the flanges and webs and connecting by splice plates fastened either side of the part effectively clamping them together. A typical connection is illustrated. Major fabricators drill the component parts using numerically controlled machines utilising data downloaded electronically from drawing information. Precision manufacture ensures that site connections can be assembled quickly and accurately.
Staggered joint configuration
The size and configuration of bridgework girders means that much of the welding is carried out in the vertical (web) and overhead (underside of flanges) position. There are considerations of time and cost, together with enhanced skill level requirements for welders when welding in these positions. Every effort should be made to weld as much of the work in the flat position where productivity is higher. Therefore it is normal to weld the flanges with a single vee preparation up to say 40 mm thick and only to change to a double vee for thicker materials. Web joints up to 20 mm are again single vee changing to a double vee for thicker materials. Experienced fabricators have welding procedures to address this type of joint configuration. It is important to emphasise that correct sequencing is followed to limit welding distortion and maintain the girder alignment in all directions.
Main girder connection
The design normally requires a preloaded bolt in the connection which, when tightened, applies a tension in the bolt sufficient to impart frictional resistance to the interfaces thus preventing slip. This is achieved, either using high strength friction grip bolts or perhaps nowadays more commonly, using tension control bolts (often termed TCBs or TC bolts). The advantage of using a tension control bolt is that consistent tension is applied clamping the plies together so that appropriate friction between the interfaces is achieved. More information on developments in structural fastening using bolts is given elsewhere in this publication.
The predominant site welding process remains manual metallic arc welding because of equipment portability and versatility. Major projects with extensive site welding are often set up to weld with flux-cored wire or the submerged arc welding processes using mechanisation for increased deposition rates. There are significant productivity gains in using these processes and a suitable balance of the economic
73
High strength friction grip bolts are tightened in accordance with standard procedures. There are three principal procedures: • Torque control method – the tension in the bolt is established by applying a controlled torque to the bolt. It is necessary to calibrate the equipment and method to ensure that the appropriate level of tension is achieved. • Combined (or part-turn) method – this is a two stage procedure requiring an initial bedding torque to bring the plies into contact, followed by a measured turn to apply the preload.
Diaphragm at box girder cross-section
Box corner joints
• Direct tension indicator method – tension or load indicating washers are included in the bolt assembly. These are devices with raised protrusions designed to compress under load. During the tightening procedure the residual gap is monitored to ensure that the appropriate load is generated. The use of this type of fastener assembly may cause difficulties in successfully applying a protective treatment system owing to the residual gaps above the indicator washer.
Internal access is frequently the deciding factor in selection of joint detail. Fillet welds on both webs are adequate for most applications. For small section girders, the internal weld is deposited in an open box situation, i.e. without one flange in place. The closing weld is then a single bevel partial penetration or full penetration weld. The full penetration weld is normally deposited on a backing strip with sufficient root gap to provide good access for welding, as illustrated.
With all large bolted connections it is necessary to sequence the tightening method to ensure proper seating of the plies and this generally involves working outwards from the centre of a bolt group. The connection designer needs to be aware of some potential difficulties with bolted joints. Particularly, the size and weight of plates likely to have to be man-handled to final position and the need for packing plates where there are changes in thickness of flanges or webs. In addition for composite construction, the uppermost top flange splice plate is likely to require stud shear connectors in between bolt positions. These have to be carefully positioned to avoid causing an obstruction during bolt tightening operations.
Box girders Shop in-line joints In principle, pre-assembly joints are welded in exactly the same way as for plate girder joints using similar preparations and processes. Inspection and testing requirements are common to both types of girder.
Weld details
The difference between the two forms of construction occurs when lengths of box girder are shop joined into a long length for delivery. If the section is too small for safe internal access it becomes necessary to weld the transverse web to flange joints from the outside only. The usual method of doing this involves the provision of internal backing, either as a strip or as a diaphragm. A single vee joint preparation with a root gap of 6 to 10mm provides good access to complete the joint. Correct sequencing is necessary to minimise distortion effects and to distribute the residual welding stresses in the joint. The provision of a diaphragm at the joint assists in maintaining good cross-sectional shape. A typical configuration is illustrated. In this situation designers need to be aware of the difficulty in providing continuity of any longitudinal flange or web plate stiffening.
External long welded joints are usually made using mechanised equipment, but care is necessary to balance the weld sequence to avoid distortion effects inducing twist and out of straight flanges.
Internal stiffening and diaphragms The box girder is very effective in resisting bending due to the wide bottom flange. In addition, the closed shape of the box offers substantial rigidity in resisting torsion. Depending upon the size of the box and the design loads the web and flange plates may require longitudinal stiffening formed from welded flats, rolled sections or formed trough sections.
74
Stiffening arrangements
Preferred details
75
Internal diaphragms are required to reduce cross-sectional deformations and to maintain shape, particularly at support positions. These are usually profiled plates or a sub-assembled framework of rolled sections fillet welded into the box girder. Frequently, the fabricator will require additional internal stiffening/ frames to maintain the shape of the box during fabrication. Typical internal stiffening arrangements are illustrated.
Overlapped joints where diagonal members meet are used only where the intersection points cannot be spaced sufficiently due to the loading or the geometry of the structure. For short to medium spans the common construction is a halfthrough truss where the compression chord is restrained by means of U-frame action. To transfer the moments generated efficiently, it is usual to ensure the transverse member forming the horizontal arm of the U-frame has the same depth as the chord section. Where this is not the case, there may be a need to have stiffeners slotted in the bottom chord section. This is to be avoided as it is a very expensive detail. Some simple preferred details suitable for footbridges are illustrated.
In open-top boxes, top lateral bracing is required mainly to stiffen the box girder during transport and erection, although it is sensible to utilise the bracing arrangement in considering overall structural performance.
Site in-line joints Conclusion
Access and the consequent health & safety issues remain dominant with respect to selection of joint preparations. Unless the box section is large with good access and egress it is likely that most site welded joints are welded externally using a backing arrangement described in the section on shop welded joints. Temporary works are necessary to provide structural stability and are often designed to set the joint alignment quickly and accurately.
Designing suitable connection details for bridgework poses a different challenge in that the structure being built is generally made of larger components. Thus the type and positioning of joints is often dictated by the chosen delivery and erection process. In particular, bridgework is likely to require site welded connections and friction grip connections made using high strength preloaded bolts (neither of which is generally common in building frames). Readers requiring a fuller understanding of the issues involved are recommended to refer to the Guidance Notes on Best Practice in Steel Bridge Construction (SCI P-185) which contains ten data sheets on steel bridge detailing. In terms of the detailed construction engineering decisions that underpin the delivery and erection methods, readers are referred to the BCSA Guide to the Erection of Steel Bridges.
Welding, inspection and testing requirements are the same as for plate girders.
Bolted in-line joints This type of connection is only possible when there is internal access and the connection configuration is the same as for plate girders. The designer should be aware of the even greater difficulties of man-handling splice plates inside a closed section.
Footbridges Designers and architects generally prefer to use hollow sections in the construction of footbridges. The use of appropriate connection details is essential for fabrication of these structures. The current bridge standard, BS 5400 does not cover hollow section connection design; however publications are available from CIDECT to assist with the connection details. The CIDECT design rules for joint design have largely been incorporated into Eurocode 3 Part 1.8. Additional advice on hollow section details is given elsewhere in this publication. The final selection of the joint type is a function of the design and sometimes it is possible to achieve an economic connection by simply increasing the wall thickness of the hollow section forming the structural member. Wherever possible the width of the section forming the web member should be less than that of the chord section. This arrangement facilitates joining the sections using fillet rather than butt welds. The ideal width ratio is 90%. Substantially reducing the width of the vertical (or inclined) web section may result in having to reinforce the chord section against punching/tearing and this is clearly undesirable.
76
CHAPTER 16 Tension Connections
Key issues
Roger Pope, BCSA Technical Consultant
In the zone where the connection is made, the spread of the load into the main structure will be determined by relative strains in the 3D force field. This may require finite element methods to model sufficiently accurately if hot spots are to be eliminated.
Concentrated forces
Acknowledgement is due to Ian Firth, Flint & Neill Partnership, who provided all the information upon which this article is based, to Wilkinson Eyre (architects for Lockmeadow, Halgavor and Swansea bridges) and to Rambøll (for the Malmö bridge).
If the tension connection does not properly align with the tension member it secures, then significant secondary bending may be imposed on the tension member that could critically jeopardise its performance. Furthermore, tensile structures can be relatively lively such that they often deflect significantly as load patterns vary with time. The designer needs, therefore, to review the behaviour of tension connections over a 3D envelope of predicted local geometries for the load paths.
Introduction Structural members subjected to tension are often made from relatively high strength steels (e.g. wire strands of strength grade 1570 MPa) that accommodate relatively high and highly concentrated loads. The detail designer’s primary task is how to engineer a tension connection to distribute these concentrated localised forces back into the main structure.
Applications The traditional application for tension members has been guyed masts and towers and of course bridgework – with suspension bridges now being largely superseded by cable-stayed bridges except on the very longest spans. Industrial cableways, cable cars and ski-lifts are also common applications where the terrain is remote or mountainous. Structurally, the need for long column-free spans also occurs in stadia and some very large industrial or commercial buildings. If deep trusses are not used, the alternative would be to support shallower roof members from masts using tension members. Large lightweight roofs and canopies for stadia and other open spaces have been provided by cable net structures. These require tension connections for their steel edge cables and mast stays. The attractively light appearance of structures that make extensive use of tension members has now led to them being favoured for footbridges particularly those in scenic or heritage areas. Their architectural attractiveness has also led to their use in the public circulation areas of prestige buildings such as office atria and station concourses.
Varying geometrical set out in 3D on Swansea Sail Bridge
Conventional structural assemblies are relatively much stiffer than those using primary tension members, and even if designed as nominal pins, conventional connections generally offer enough strength/stiffness to maintain the local geometry. Tension connections, however, are often formed using actual pin joints and the structures can be close to becoming mechanisms. High localised force concentrations combined with potentially large deflections may also give rise to P-Δ effects that cause lateral stability problems to arise in the overall structural assembly. Thus the designer needs to review how movements might influence the global behaviour and perhaps modify the layout to “triangulate” the forces into a stable configuration under large potential deflections.
Halgavor footbridge – steel masts, steel cables, grp deck, timber planking, concrete foundations
77
Economy and aesthetics As tension connections are prominently visible, they often need to look aesthetically clean. To some extent the conceptual design of a tension connection may emerge directly from the load paths in the way that form follows function. However, taking forces from 1570 MPa wire strand into a connection made from 275 MPa carbon steel would dictate that the connection material has at least six times the area and this can look relatively out of proportion. The higher safety factor needed for the connection materials plays a part here also. Additionally, the connection zone may become congested with duplicated tension members required for reliability or replacement reasons, and further attachment points for jacking equipment and access equipment used during installation or removal. If pins are used for all these attachment points, the holes will require quite generous edge distances to meet code requirements for avoiding net section failure. Hence, keeping a tension connection neat and tidy can be quite a challenge.
Lockmeadow Bridge showing splayed stays
Environmental exposure
Neat and tidy group of connections visible at mast head
In bridges and many other applications, the completed connection is externally exposed to the weather, experiencing low temperatures and wet conditions from rain and condensation. The low temperatures mean that the connection designer needs to be careful about the fracture toughness of the materials specified as this will require steel with improved Charpy properties where thick materials are necessary.
Neat and tidy connections on Malmö swing bridge
Whilst toughness requirements can be tackled through the specification, coping with moisture in the connection zone
Roped access used for installation and maintenance
78
requires considerably more design ingenuity. Firstly, water is notoriously good at penetrating seemingly sealed spaces owing to capillary action and relative pressure differences induced by temperature and humidity changes. Secondly, unless the water is standing, the continuing flow of water (with air) will result in continuing potential corrosion. Thirdly, under load or temperature effects local movements can cause surface-applied corrosion prevention products to crack.
Lightning protection straps may need to bridge the connection zone, and an additional precaution relevant to high strength and zinc coated materials in particular is the need to detail in a way that eliminates any direct contact with cementitious materials.
Fatigue Tension members exposed to the wind will vibrate and this vibration can induce cyclic stress variations in the connection zone. Over time a pattern of high enough stresses can lead to fatigue failure as the number of cycles mounts. Again, the designer has more than one weapon to use. The vibration can be reduced by mounting dampers on the tension member itself, or the fluid dynamic behaviour and mode of vibration of the whole member might be modified in other more sophisticated ways. In the connection zone, local details can be used that mitigate fatigue effects. These include smooth transitions that avoid sharp edges and sharp changes of shape as presented to the load path. For threaded couplers etc., they also include specifying rolled threads that have a higher resistance to the propagation of fatigue cracks.
The designer has four weapons to use in coping with this problem: stainless or “rust-free” steels, water-resistant and water-repellent products (e.g. paints, greases and pastes), impermeable covers (e.g. sealants and wrappings) and planned maintenance. A typical problem is posed by a threaded coupler or turnbuckle used to join or adjust a bar tension member. The threads cannot be fully treated with paint etc. until after installation. Hence, access to that position to complete the corrosion protection system will be necessary during installation, and this access can be used later if, say, grease needs replacing. One detail that can assist is the provision of a suitable circumferential crevice that assists in retaining the grease or sealant in place.
Manufacture and installation The key concern during execution is that cumulative tolerances do not impair performance or aesthetics. As tension members form generally shallow catenaries, tolerances affect their as-installed shape. This is because their effective stiffness arises from elastic stretch, lay-up of the wire and catenary sag. The sag is directly linked to the relative length of the tension member compared with the direct distance between its connection points. For example, less than 5mm difference in installed length between a pair of nominally parallel ties will be readily visible in terms of differing sags. Not what the architect expected! If the local geometry around the connection zone is complex, it may be necessary to ensure that reference points for settingout are clearly marked in accessible locations with known offsets from, say, true intersection points.
Sleeves used to prevent ingress of water running down tie bars
Cable sealed with zinc paste and cover plate sealing connection zone from ingress of rainwater running down cable
Complex local geometries
79
Local bending
In some cases careful control of specified lengths may be sufficient when combined with progressive checks during trial and final assembly. In other cases, e.g. where wire ropes are being used, pre-stretching of the assembly will usually be necessary to remove any inelastic stretch, to control its length and to check its effective elastic tensile stiffness. In many cases there will need to be a facility for adjustment of length after installation, possibly using a turnbuckle in a tie bar.
Another installation precaution concerns the use of “pin” connections, as these might seem to permit local relative rotation between members but only do so during the installation process. As load is increased, the ease of rotation reduces such that at some point it “freezes” into a fixed alignment (unless the connection is furnished with a properly engineered lubricated bearing sleeve). Above this load it may be assumed that the connection is no longer pinned but stiff and it will remain so in service. The installation procedure needs to ensure that the tension connection at the end of the tie itself is “presented” at the correct tangent set angle through the pin to its attaching plate etc. Otherwise possibly significant secondary moments will be induced in the tie.
Adjustments are also needed to correct local alignment, possibly using pairs of tapered packings, or cam adjustment to the coupler attachment pin at the end of a cable tie. Detailers should keep remote connection points simple and incorporate the more complex adjustment mechanisms at the more readily accessible end.
A pair of tapered packings used for fine adjustment
Local alignment yet to be finalised during erection
Connection at lower end facilitating adjustment for length or load in tie
Collars fitted to limit local bending
80
Similarly, high local bending can arise from relative movement in service, even if the initial installation is correctly aligned. An example is the “short” hanger on a suspension structure. Although “pinned”, the end assembly will be much stiffer than the bending stiffness of the hanger tie such that movements of the whole structure will cycle the hanger through a bending regime that can cause fatigue conditions to arise locally where the hanger enters its end coupler. Collars can be detailed to spread the bending zone and limit the local effects.
Eccentricity Reflecting on the highly concentrated nature of the forces in the tension member, it can be seen that even quite small eccentricities can induce unintended high local bending in the connection itself. This may arise from overall misalignment in the design where lines of action do not reach a common intersection point. Alternatively it can arise from local effects such as slip in a bolted connection group or slop in a pin connection. Thus, the local force distribution around a pin assembly can be critically affected by even quite tight manufacturing tolerances. For instance, to ensure that the pin is symmetrically embedded in the joint, the detail should use “keeper” plates to retain the pin in its correct position.
Foundation end connections for splayed cables
Jacking shoe with threaded permanent attachment used for adjustment at end connection
Keeper plate securing pin in connection to foundation
Connection types End connections These are generally made with tie end couplers as spades or forks attached using pins to gusset plates. At nodes where several ties join together, plates may be shaped as “deltas” etc. In simpler structures, traditional rigging tackle is often used, e.g. “D” shackles. Special cases arise where the end connection attaches to a foundation. In the case of very large span structures the cable anchorages become major structures in which the layout and securing of the fan of cable anchorages requires careful thought.
Neat and well-aligned connection group on Malmö swing bridge
It may be necessary to use the end connection to make adjustments to load and/or length during installation.
81
Saddles and end connections on the Wye Bridge
Hanger clamped to suspension cable on Halgavor footbridge
Mid connections In bar tensile members these are made using threaded couplers. All threads may be right-handed unless adjustment is needed. Otherwise, longer “turnbuckles” would be used that require left-handed threads at one end. Some means of turning the turnbuckle when under load may be needed. As the bars and the couplers might be sourced from different suppliers, specifiers may need to be careful about embedded thread lengths, thread tolerances and fit to ensure the assembly does not fail prematurely by thread stripping.
Non-aligned connections
Proprietary “Stylite” end connection from Bridon
These are either clamps or saddles. A clamp would be used where a hanger attaches to a suspension cable, a saddle where the suspension cable passes over a pier or other structural support. In both cases care has to be taken to ensure that the attachment does not chafe the primary cable by using, for example, a soft metal zinc liner inside a clamp.
The manufacture of the connection components varies with type. Smaller proprietary components are usually cast or forged. Larger and bespoke components are usually machined from steel bars or plates. Although plates are available in structural steels to BS EN 10025 (e.g. S355J2 up to 400mm thick), the larger bars are generally only available as wrought steels to BS 970, and to facilitate machining some higher strength grades require strength improvement by heat treatment after machining. If bespoke castings are used, then it is prudent to specify over-production to ensure that suitable as-produced specimens are available for testing as destructive methods may be the only ones possible.
Proprietary types Bespoke design is often necessary, but for general use there are manufacturers who offer proprietary systems. Generally their designs have been proved by full-scale testing combined with code verification such that there should be a high degree of confidence backed by the manufacturer’s warranty if the products are used in accordance with the manufacture’s recommendations (see Bridon and Macalls in References).
82
Hanger replacement in progress on the Severn Bridge
Welding is needed for irregularly shaped items that are too large to be economically machined from a single solid billet. As noted above, specifiers need to ensure that thicker materials do reach the required toughness in terms of the Charpy value, and the same applies to procedures for heavy welds. If bespoke items are designed to accommodate significant outof-plane forces, it may also be necessary to specify throughthickness properties in the parent material. BS EN 1993-1-10 provides recommendations on both these issues. A family of bespoke machined spade end connections
The mast of the Swansea Sail Bridge – engineering or sculpture?
83
Design issues Components of tension connections need to be designed using common rules for other connections. Thus, depending on the code, whilst a simple steel tension member may be designed using a factor of safety of 1.0, the connection should be sized using a partial safety factor of 1.25. However, the main tie member might be increased in size or doubled up for reasons of added safety or redundancy. There could also be significant over-strength in the tie material that exceeds the likely over-strength in the connection components. For these reasons a decision will be needed as to what connection components should be upgraded to retain the principle that the tie should fail before its connection. When considering cable replacement etc., it is practical to base the validation on the serviceability limit state when using traditional UK codes or in some cases the accidental limit state when using the Eurocodes. Even where proprietary systems are available, it may be necessary to interrogate the manufacturer’s data carefully to ensure that the designs are validated on assumptions that are consistent with the particular design philosophy the specifier is using (e.g. whether fatigue has been considered). Where fatigue is a factor, it is sound practice to keep the cyclic component as a low proportion of the total stress, and for welds to keep the nominal imposed stresses well below yield.
Conclusion As with most steel structures, there are many examples of actual connections readily visible around us and in technical publications. Given the applications favoured for the use of tension connections, they are some of the most readily visible. Furthermore, their nature is such that it is often possible to develop a model of their structural behaviour “by inspection”. However, there are several aspects of the detailed design that require special attention and that are not so readily appreciated at first glance. The most important structural issue is the need to avoid local effects arising from very high forces at unintended eccentricities. For externally exposed connections, it is also important to give careful consideration to the avoidance of corrosion.
84
CHAPTER 17 Lattice towers and masts
When type (a) is used the ends of the legs need to be ground to provide even end bearing and the flanges need to be designed for prying action. Bolts need to be preloaded and care is needed in designing the welds, particularly if fatigue is an issue.
Brian Smith, Flint & Neill Partnership
Important details
Type (b) relies on compression and tension being carried through the ring flange and in this case the load should be able to be carried by the outer welds as the inner welds cannot be inspected. Fatigue again becomes an important design aspect. Again prying action of the flange needs to be taken into account.
For lattice towers and masts, the most important details are those described here which are associated with leg-to-leg joints, the stay linkages and bracing connections.
Leg-to-leg joints
Type (c) is the simplest but is not generally a good fatigue detail and when used with tubular legs, drainage (or sealing) to prevent corrosion needs to be considered.
When angle sections are used for leg members joints are normally through conventional cover splice plates designed to transmit axial loads (in tension and compression) through bolt shear to the plates. However when tubular or solid round legs are used several alternative details can be used. The most common is to use ring or disc flanges and direct compression is either taken directly through bearing of the leg (only when solid round legs are used) or through the flanges. Tension is carried by bolt tension through the flanges. Typical details are illustrated.
Unacceptable fabrication control on flange plate details can occur, and the illustration shows a situation where the edge distance of the bolt holes is inadequate. The use of wing splice plates as in (d) is rare although this detail was adopted for the very tall Clyde Crossing transmission towers where large solid round legs were used and a flange plate detail became impracticable.
Typical leg-to-leg connections
85
To minimise the risk of cable vibration causing excessive local bending in the anchorage steelwork, freedom for the guy to rotate in both the vertical and lateral planes is usually provided. The important vertical freedom is achieved by horizontal pin connections at the anchor block, and through the tensioning system. Lateral freedom has been achieved by using a pin plate at the guy connection which allows the guy itself to be free to rotate in the lateral plane. Here ‘spherical’ holes have been used in plates, although more sophisticated details have been necessary elsewhere.
Bracing connections Lattice towers, in particular, have been designed over many years using strut curves developed from full-scale testing to destruction of transmission towers. Full-scale testing was practical as such structures were used in situations where many towers of an identical type were needed on a given transmission line.
Unacceptable leg flange joint
Stay linkages In guyed masts it is necessary to provide adjustment at the guy anchorages, both to adjust the tensions in the guys and, during the life of the structure, to facilitate replacement of the guys themselves. For tall guyed masts (of say 300m height) the installation and tensioning of the guys, and the subsequent re-tensioning due to any loss of tension due to ‘bedding in’ of the guys (which may occur even if they have been prestretched prior to installation), needs an extensive system as may be seen in the illustration. The long threaded rods provide the adjustment and facility for inserting a tensioning head to tension the guys.
Economy thus became paramount and as a result the strut curves tended to provide higher strengths for a given configuration than use of standard building code curves. In addition better account was taken of the specific end details in terms of the number of bolts used and the associated eccentricities. The resulting details thus need to be exactly maintained if the enhanced strut curves are to be relied upon. Close control of back-marks, of the end and edge distances of bolts and tolerances on bolt hole diameter all need to be specified and maintained.
Stay connection details (courtesy of National Grid Wireless)
86
As opposed to strut or strut-tie bracings, the use of a tension bracing system can provide lighter and more pleasing structures, but possibly at the penalty of increased costs and the need to pretension the bracing system. A good example is illustrated. The final illustration here shows unacceptable bracing (and leg) connection details: poor welding, lack of continuity of the shear-bracing system, and inaccuracies in the lengths of adjacent leg members can all be seen. The leg joint here is made through bolted spigots but this introduces problems of misalignment that can also be seen in the illustration.
Conclusion Lattice towers and masts are a specialist field of steelwork in which, owing to the repetitive nature of transmission lines etc., it is often practical to use full-scale testing to validate design details. Designers can take advantage of this catalogue of successful past practice provided that the details are repeated faithfully as departure might bring in unforeseen factors like local bending or fatigue sensitivity. The author is currently preparing a publication for Thomas Telford Limited that gives much wider advice on design issues related to towers and masts.
Tension bracing system (courtesy of Rambøll)
Unacceptable fabrication of guyed mast
87
88
Architects’ Working Details Architects’ Working Details were first published in 1953 and ran to a series of 15 classic black-bound volumes. After a long lull they were revived in 1988; the most recent volume, the tenth in the new series, is a selection of details published in the Architects’ Journal (AJ) in 2001 and 2002. The Working Details are intended to collate ideas about the detailing and construction of buildings; some details demonstrate new and innovative techniques, others refine tried and tested methods of construction. They serve an additional purpose - to enable architects to exchange information on contemporary problems in design. Too many architects, faced with an unfamiliar detailing problem, start afresh - with all the expense of time and energy that that entails - when they could be building on the experience of their fellow professionals. Although the working details are grouped under subject headings, the volumes are not intended to be a comprehensive examination of all aspects of construction. They originate from the building studies which are published each week in the AJ. Each detail is placed in context by reference to an AJ building study or feature, and is accompanied by photographs and an explanatory commentary. For consistency and legibility, the architects’ drawings were re-drawn and scaled to fit the pages of the AJ, with the sizes of individual components annotated rather than drawn to a particular scale. The BCSA acknowleges the assistance of the AJ in compiling this selection of Working Details.
89
90
ARCHITECTURAL DETAILS AN INTRODUCTION by Susan Dawson Designing the details of a building involves important practical decisions; clearly the building must be structurally stable, fulfil Building Regulation requirements and achieve a weatherproof enclosure and so on. But the design of a building or structure is also a unique, imaginative response to its brief, the site and its requirements; the details are the means by which this unique response is achieved. Like architecture itself, good details are a fusion of imaginative design and sound construction.
Jean Prouvé’s refinement of industrial detailing and use of lightweight sheet metals have been explored by the recent generation of High-Tech architects such as Richard Rogers and Nicholas Grimshaw. Cor-Ten weathered steel can be used, not only for a bridge structure but also as a rainscreen façade to a theatre. I have also included details which demonstrate what Italo Calvino has called “the second industrial revolution, which does not present us with such crushing images as rolling mills and molten steel, but with ‘bits’ in a flow of information travelling along circuits in the form of electronic impulses. The iron machines still exist, but they obey the orders of weightless bits.”
The following 23 Architects’ Working Details were originally published in The Architects’ Journal between 2000 and 2005 and are examples of how, over the last few years, architects and engineers have explored and developed details in steel. This selection is not intended to give a comprehensive exposition of standard steel details; rather it shows how imagination and creativity are an integral part of the design and detail of steel elements.
Today the processes of design and fabrication are being dramatically altered by these weightless bits. The structure and fabric of the Spiral Café, based on the Fibonacci series, were developed with a parametric 3D model and the steel plate ribs were cut on a computer controlled plasma-cutting machine. The façade of Foster and Partner’s St. Mary Axe building, a complex curtain wall of diamondshaped panels which are faceted at their edges in both plan and section, was also developed with a 3D modelling tool. New programs can handle complexity more efficiently; this offers great future potential for detailing in steel.
Many ideas about steel detailing have their roots in the Modern Movement - as a material it has had a radical influence on architecture. In the 19th century steel skeletal structural frames liberated buildings from the inhibitions of the loadbearing wall and related construction, making it possible to build upwards; the heroic scale of American cities was determined by steelframed skyscrapers. Nowadays a deeper understanding of the structural behaviour of steel has allowed longer spans to be achieved with minimally-sized steel elements. The sports and school buildings which follow are examples of how steel can provide lightweight almost ‘hands-free’ enclosures; the four bridges, each utterly different, show steel structures of striking delicacy and grace, or, in contrast, a structure of sturdy and rugged Cor-Ten, inspired by the sculptures of Richard Serra. The development of stainless steel at the beginning of the 20th century provided an environmentally stable metal that could sustain a polished lustrous appearance. It has been used with great elegance in Ian Ritchie’s Spire of Dublin, a tapering stainless steel spire 120 metres high. After the war, the transfer of technology from military and aeronautical industries generated new metal forms. Mies van der Rohe’s reductionist structures, of steel channels, angles and universal beams, were tautly elegant expressions of a new minimalist aesthetic. A classic UK example of this, Peter Foggo and David Thomas’s Space House in East Grinstead, was built in 1965 and has been recently updated. In the past one of the key problems of designing in steel was that the desire to express the steel structure conflicted with the need to protect it from fire. Bennetts Associates has detailed an office building in Devonshire Square to solve that problem.
91
DUBLIN’S DESIGN PINNACLE The Spire of Dublin’s simple elegance belies the complex design and engineering process behind its construction BY SUSAN DAWSON Dublin has a new monument – a slender spire at the heart of the city which soars 120m above the rooftops. By day, the stainless steel spire reflects the city’s changing light and shadow; by night, the base glows gently while the illuminated tip shines out like a beacon over the city.
The stainless steel surface was given a shot-peened finish, a process which uses stainless steel beads to create a lightly textured surface that enhances corrosion resistance and diffuses light. The effects of light on a peened surface are much softer than with a conventional brushed finish. The top 12m of the spire is perforated with approximately 12,000 holes to allow light to escape and to illuminate the upper surface of each hole.
The concept and design is by Ian Ritchie Architects, the winner of an international competition for a monument to celebrate the third millennium and to be the centrepiece of wide-scale urban improvements.
The spire rises from foundations below ground and passes through a 7m diameter bronze base plate set at pavement level. Bronze was chosen for its historical associations with the development of art in Ireland and for the fact that it weathers well. An elegant logarithmic spiral design was machined into the surface of the 16 bronze plates to produce an anti-slip surface.
Its position – on O’Connell Street at the junction with Henry Street and North Earl Street – is significant in Dublin’s history. This was the site of Nelson’s Pillar, a memorial built in 1808 and blown up in 1966 by the Irish Republican Army. It is also next door to the General Post Office, the headquarters for the Easter Rising in 1916 and a place of pilgrimage for Republicans everywhere.
In concept and basic structure, the spire is extremely simple – a cone which tapers from 3m to 150m at the pinnacle in a simple cantilever. But the research undertaken into the materials, finishes and structural implications belies the simple elegance of the final result.
The spire is a stainless steel cone, tapering from 3m in diameter at the base to 150mm at the tip. Ritchie’s inspiration for the spire’s form evolved from a study of standing stones and obelisks and from an ambition to create a monument which would be as slender and vertical as possible as it escaped the roofline of Georgian Dublin – and, crucially, would capture the light of Ireland’s skies. ‘The spire reintroduces a vertical counterpoint to the prevailing horizontal nature of O’Connell Street’s buildings, without visually interrupting the streetscape,’ says Ritchie. It was decided that the maximum desirable base diameter should be 3m and this led to a proportion of base width to height of 1 in 40 – an elegant but technically challenging specification. The spire was fabricated in eight sections from stainless steel grade 316L (L denoting low carbon content) manufactured and supplied in accordance with EN 10088. This grade is extremely resistant to corrosion, a necessary factor in an environment where traffic pollution and wind-borne marine salts are both present.
CREDITS ARCHITECT Ian Ritchie Architects: Ian Ritchie, Robin Cross (project architect), Gordon Talbot, Phil Coffey CLIENT Dublin City Council STRUCTURAL AND SERVICES ENGINEER Arup
QUANTITY SURVEYOR Davis Langdon & Everest LIGHTING Ian Ritchie Architects/La Conch MAIN CONTRACTOR SIAC/Radley Engineering joint venture SUPPLIERS Steel forming Barnshaw Steel Bending
92
Structure of the spire
mass is connected to four viscous damping units and tuned to a specific frequency by adjustment of the length of the suspension cables. This technology has been around for some time, but the spire marks the first use of tuned mass dampers in Ireland . The mass damping system had to be coordinated to fit into a very tight space without hitting the internal access ladders or the electrical cables supplying the lighting system.
The Spire of Dublin is a true collaboration between architecture and engineering. Engineer Arup worked with Ian Ritchie Architects at the competition stage and undertook detailed research into the design of the structure and its geotechnic, mechanical, electrical and public health aspects. The company also provided wind engineering services and supervised onsite construction.
Many other aspects of the design – the selection of materials, wind analysis and wind profiling, pile design, fatigue analysis – involved fundamental research into the physics of specific problems. The complexities were further compounded by a specified design life of more than 120 years.
The project was full of engineering challenges, but the really satisfying side to it was the holistic approach to design and the use of processes from outside the mainstream construction industry, such as polishing and peening. The spire is supported on stainless steel reinforced concrete walls of an underground access chamber, which houses electrical and drainage systems. This in turn is supported on eight steel-reinforced 900mm diameter concrete piles anchored into the bedrock. Once the piles were in place, the pile cap was monolithically connected to the bottom of the concrete chamber.
But the most demanding issue was the height of the spire and the effects of the wind’s many moods; these factors determined plate thickness, joint design criteria, weld fatigue criteria, damper design and tolerances.
To counter dangers of wind-induced vortex shedding, a tuned mass damping system was built into the fifth of the Spire’s eight frustums (sections). Designed by Arup and Motioneering, of Guelph, Ontario, this is a passive system consisting of two stainless steel masses suspended on cables. Each
93
A stainless steel conical spire The spire is 3m in diameter at its base; it tapers to 150mm in diameter at the tip, which is 120m above ground level. It is fabricated from Grade 316L rolled stainless steel plate with a shot-peened finish. The plate is 20mm thick, except for the lowest 4m of the spire, where it is 35mm thick, and the topmost section, which is 10mm thick.
level 124.800m
stainless steel spire at 224mm radius 150mm zone to raise and lower specialist lighting
PLAN AT 112.286m
level 112.286m
threaded connection stainless steel spire at 540mm radius 150mm zone to raise and lower specialist lighting
The spire was fabricated in eight sections (frustums) of up to 20m in length. Each frustum was formed from a series of short sections, fabricated as follows. A stainless steel plate was cut to a curved template, rolled to form a half-section and welded with vertical joints to form a complete short section. A series of these were then welded together with circumferential joints to create a frustum.
PLAN AT 85.606m
The thicker frustums at the base of the spire were curved using a modern asymmetric plate-forming machine; the upper frustums, which were too small for conventional rolling, were formed using a 1,000tonne press brake.
PLAN AT 40.946m
The lowest seven frustums are connected by bolted flanges. The uppermost frustums have a threaded connection.
apex of Grade 1.4404 stainless steel spire at 75mm radius
PLAN AT 124.800m
bolted flange connection
stainless steel spire at 741mm radius maintenance platf platform hatch with removeable panel
level 85.606m
intermediate aviation warning lights 150mm zone to raise and lower specialist lighting
PLAN AT 69.467m
stainless steel spire at 1071mm radius
level 69.467m
150mm zone to raise and lower specialist lighting bolted flange connection
stainless steel spire hatch with removeable panel maintenance platf platform ss winch
PLAN AT 22.266m
level 40.946m carriageway bollards Grade 1.4404 stainless steel spire
The tip of the spire is fitted with lighting and with aviation warning lights. These are maintained by means of a lift and lower mechanism – a pulley is housed at the very tip of the spire – and by ladder access with integral fall-arrest system.
150mm zone to raise and lower specialist lighting
level 22.266m
granite cover to perimeter drainage channel 12mm gap between spire and bronze base granite paving
PLAN AT 4.800m PA PAVEMENT LEVEL
94
N
The ladder is set in the central void of the spire and leads to an open grille platform at about 70m above ground level. The lighting unit is composed of lightemitting diodes and is 10m in overall length. The ladder access also allows the mass damper system to be adjusted.
bronze base plate with spiral surface profile
level 4.800m
KEY ELEVATION ELEVATION VA
top plate
insert threaded into top plate to receive aviation warning light
folding footrest to give access to damping devices
476mm
sheave housing
threaded plugs to match spire surface
level lev el 73.285m level el 73.150m
circumferential rings circumf
circumferential rings circumf
support and lighting fins support and lighting fins
bolted flange connection
flange
150mm clear zone in which to raise and lower aviation warning light
conical stem
internal access ladder with integral fall-arrest system
threaded connection to apex and spire
intermediate aviation warning light
level el 71. 566m
el 70.817m
perforated rforated sur rf to stainless steel spir zone for electrical cables galv steel open grille platform floor platf removeable panel in opening hatch sized to allow passage of aviation warning light
lighting – shown during raise and lower operation
level 69.476m angle UB trimmer 700 x 790mm hatch in platf platform floor
DETAIL SECTION THROUGH PLA TFORM DET threaded extension welded to flange
zone ffor electrical cables UB trimmer internal access ladder with integral fall-arrest system
threaded hole ffor turning rods, fitted with flush threaded plugs to match spire
removeable panel in opening hatch sized to allow passage of aviation warning light
150mm clear zone in which to raise and lower apex aviation warning light and specialist lighting
galv steel open grille platform floor platf 150mm clear zone in which to raise and lower aviation warning light
gutter
700 x 790mm hatch in platf platform floor
vice
DETAIL SECTION AT APEX DET
ELE VATION AT APEX
PLAN AT PLATFORM AT LEVEL 69.467m
95
FÊTE OF TWIST Marks Barfield’s Spiral Café is both a resting place and a sculpture in Birmingham’s regenerated Bullring centre BY SUSAN DAWSON. PHOTOGRAPHS BY ALASTAIR CAREW-COX AND PETER DURANT Amid the retail frenzy of Birmingham’s Bullring Shopping Centre, it is refreshing to come across the elegant, shell-like spiral of copper that is the Spiral Café. Designed by Marks Barfield Architects, it is only 60m2, small enough to be viewed as a piece of sculpture in its own right, yet large enough to allow shoppers to sit comfortably and enjoy the view. The café flanks the curved walkway that runs along the western side of St Martin’s Square, a new space that links the new shops and central boulevard, including the Selfridges store by Future Systems, to Edgbaston Street, stepping down in a series of wide terraces to St Martin’s Church.
Although the café floor spans across the shell, the integrity of the structure is maintained by making the ribs continuous; they curve under the floor and emerge in the main interior of the café as a low curved wall that acts as the servery. The floor spans across the ribs, partly in the form of laminated glass so that the continuity of the structure is visible. This continuity maintains the structural integrity of the shape and works well in practice, as the terrace had to be reinforced substantially below ground level anyway to take the weight of the café. A small store to the café is slotted against the rear side of the ribs, with a curved wall to match their profile. Its side walls are made of steel plate and brace the main structure.
If you find the shape of the café particularly pleasing, it’s because it is based on the Fibonacci sequence – a mathematical sequence related to the Golden Section and identified by Leonardo Fibonacci in the 13th century, in which each term is the sum of the previous two, thus: 0, 1, 1, 2, 3, 5, 8, 13, etc. The sequence has been found to have an almost mystical relationship with the natural elements, from the shape of shells, the pattern of seed heads and pine cones, to complex fractals in galaxies.
The curved ribs are clad on the outside with post-patinated copper sheets and on the inside with smooth bronze plates finished with clear lacquer. ‘We saw it as something organic,’ explains Parker. ‘The outside is weathered, but you look inside to a smooth interior – like opening an oyster shell.’ To maintain the logic, the front face of the servery, where it emerges on the inside, is also clad in copper. The Bullring regeneration project has commissioned art and sculpture to enhance its public spaces. The shell-like form of the new café could well be taken for one of these pieces, but one that also serves a good cup of cappuccino.
‘We looked for a natural mathematical form as an appropriate concept to bridge the space between the commercial and the spiritual,’ explains project designer Ralph Parker. The concept is simple: take the Fibonacci number sequence and lay it out as a series of squares; draw quarter circles in each square to form a spiral. Based on the spiral, make a series of steel ribs from curved steel plate; set them 9° apart on plan and tilt each one higher than the one before. Cover them with a copper roof and fill the end openings with frameless glass. This elegant yet dramatic statement is the essence of the new café. The spiral tilts upwards towards the church, its large end opening filled with a Planar bolted-glass wall, which frames the church perfectly.
96
97
98
Forming the structure The form of the cafe is derived from sweeping a Fibonacci spiral to create a shell-like canopy, writes Tim Lucas. The structure is contained between the inner and outer curved surfaces of the canopy; it both supports the cantilevering roof and provides accurate formwork from which the rest of the construction can take its shape. The eight structural ribs are arranged radially in plan and each tilts up relative to its neighbour to create the shelllike form. A series of CHSs are set diagonally between the ribs; together they act as a cantilevering shell structure. The ribs are supported at points under the floor of the servery and at the roof level of the rear annexe. Further stiffness is generated by the CHS braces between the ribs, which together act as a cantilevered truss supported at the outer tips of the first three ribs. To simplify fabrication as much as possible, the structure was made from mild-steel plate ribs cut on a computer-controlled plasma cutting machine. This meant that the form of the building could be manufactured easily, with a number of curved ribs defining its shape. Obviously three-dimensional modelling was critical to the design. A product design programme was used to develop a parametric three-dimensional model. The programme was used to model all structural elements and many architectural elements, such as cladding profiles and glazing interfaces. As part of the design process, the architect and the structural engineer worked together on the same three-dimensional model, taking sections and profiles from it to develop further details that were not modelled three-dimensionally. Drawings generated from the three-dimensional model included a set of true plans of each component. This allowed the fabricator to take the profiles and add additional information, such as bolt holes and splice locations, before cutting the metal. A very high degree of fabrication accuracy was achieved in this way. The level of accuracy in the steelwork allowed us to use the frame as a building-wide template for production of other information later in the programme. Elements such as capping pieces could be designed on the three-dimensional model with full confidence that they would fit to what was on site.
CREDITS CLIENT The Birmingham Alliance ARCHITECT Marks Barfield Architects: Ralph Parker STRUCTURAL ENGINEER Price & Myers 3D Engineering: Tim Lucas
Tim Lucas is a structural engineer with Price & Myers 3D Engineering.
PROJECT MANAGER Gardiner & Theobald MAIN CONTRACTOR Thomas Vale City and Interiors SUBCONTRACTORS AND SUPPLIERS Steel fabricator Sheetfabs; glass Pilkington (Planar); glazing contractor Ide Contracting; copper and bronze supplier KME; copper patination Capisco
WEBLINKS THE BIRMINGHAM ALLIANCE www.birminghamalliance.co.uk MARKS BARFIELD ARCHITECTS www.marksbarfield.com PRICE & MYERS 3D ENGINEERING www.3dengineering.co.uk GARDINER & THEOBALD www.gardiner.com THOMAS VALE CITY AND INTERIORS www.thomasvale.com
99
A structure of copper-clad curved steel ribs The café structure takes the form of a series of eight steel ‘ribs’ that curve – each tilted in relation to its neighbour – to create the enclosure. The curve continues below the floor and emerges as a low wall that acts as the servery; the continuity maintains the structural integrity of the shape. The construction involved two individual curved ribs of 15mm steel plate, laser-cut and welded between a series of diagonal CHS struts. The resulting frames were taken to site and the ribs of adjacent curved frames were bolted together into pairs, some reinforced with an additional 15mm plate rib. On the outside, the rib structure is clad with 0.7mm-thick postpatinated copper sheet with standing seams, laid on to a bitumen ‘sandwich’ waterproof membrane and a ply deck, with 20mm high-performance foillattice insulation. The internal finish – panels of lacquered bronze sheet on a vapour-control layer – is also laid on a ply deck; both ply decks are curved to follow the rib profile. The ribs of the structure project beyond the bronze panels and taper at the upper edge to a delicate bull-nose; this incorporates a notch in which the glass panels are restrained. The 12mm toughened low-iron glass panels are fixed with Planar stainless steel bolts to toughened glass fins. They are supported from below and only restrained at roof level; that is, the roof is not supported on the glass. The existing reinforced-concrete terrace structure was adapted to support the weight of the new building and to provide some extra terrace space around it. A new pit incorporating a concrete beam spans between points of support to provide the main foundation.
100
101
REVOLUTIONARY MOTION The world’s first rotating boat lift, the Falkirk wheel, is the impressive link between two of central Scotland’s canals BY SUSAN DAWSON. MAIN PHOTOGRAPHS BY NICK HUFTON/VIEW As a piece of steel engineering, the Falkirk wheel is revolutionary – in every sense. It is the first-ever rotating boat lift, which lifts boats 35m from one canal to the other and vice versa. Set in the countryside near Falkirk in the Scottish Lowlands, the boat lift is the most exciting and dramatic part of the Millennium Link – a £78 million investment to link the east and west coasts of Scotland with an inland waterway.
By making two arms with circular openings and using hooked leading ends to add a sense of direction, the wheel was born. Its advantages over the original Ferris wheel proposal are that only two caissons are needed, the structure is simplified and it is made more dramatic. The wheel is built on a massive scale that is hard to visualise. The 35m lift is equivalent to a nine-storey building, and the weight of four boats plus water is about 400 tonnes. Manufacturing took place in the workshops of Butterley Engineering in Derby, a firm that specialises in heavy engineering for shipyards.
The project also includes an extension of the Grand Union Canal, a tunnel underneath the Antonine Wall, a new aqueduct and a lock. Together with the wheel, they connect two longderelict canals: the Grand Union Canal and the Forth and Clyde Canal, set 25m lower. Formerly, a staircase of 11 locks, dismantled in 1933, connected the canals.
The rotation of the elements created serious problems in design, as individual elements face repeated 100 per cent stress reversals as they turn, alternating from compression to tension. Specialist equipment was used to carve the huge pieces of steel, based on very accurate steel templates. The curved boxbeam arms are made from 12 pieces. Steel thicknesses range from 10mm to 50mm, depending on the stresses involved. Around 15,000 bolts were required, with 45,000 bolt-holes drilled into the steel sections and flange plates.
Boats approaching from the higher Grand Union Canal now sail along a new extension before entering a tunnel that takes them underneath the ancient Antonine Wall. When they emerge, they travel along an 80m-long semi-circular concrete aqueduct that passes like a thread through the eyes of a series of five giant concrete needles – the support piers. At the end, a caisson awaits them to take them a half-revolution on the wheel to the basin below.
The canal now has a new purpose: to serve Scotland’s leisure and tourist industry. As well as being practical, the wheel is a major tourist attraction, and even boasts its own visitor centre, also designed by RMJM.
The wheel consists of two massive steel arms that revolve around a central hub. Each arm contains a caisson – a steel container – that can hold two boats. The caissons have a set of lock gates at both ends to allow boats to enter from one side and leave from the other. When the wheel is stationary the gates open, allowing boats to enter. The gates then shut and the wheel rotates. As it moves, stability gears maintain the horizontality of the caissons. The half-revolution takes 15 minutes, after which the boats leave the caissons and the cycle begins again. British Waterways conceived the idea for a link to reconnect the two central Scottish canals some years ago but the scheme was delayed by funding problems and a subsequent redesign. After joint-venture contractor Morrison-Bachy Soletanche won the contract in 1999, the company, together with British Waterways, decided to take the opportunity to create a new tourist attraction. RMJM was invited to help with ideas, and architect Tony Kettle produced a concept for a rotating bridge, using a Lego model to demonstrate his plan.
CREDITS ARCHITECT RMJM CIVIL ENGINEERING CONSULTANT Arup STRUCTURAL AND MECHANICAL ENGINEERING SUB-CONSULTANT MG Bennet and Associates
‘Our starting point,’ says Kettle, ‘was to reinvent the brief, which had produced the original proposals. Previous schemes had been developed on the simple idea of the wheel as a circle, as in a Ferris wheel. RMJM’s concept grew from the design for the connecting aqueduct. Its organic form – the semi-circular canal resting in circular openings – was inspired by the design of a fish skeleton.’
DETAILED STRUCTURAL ENGINEERING SUB-CONSULTANT Tony Gee & Partners MAIN CONTRACTOR Morrison-Bachy Soletanche WHEEL SUPPLIER Butterley Engineering
102
103
A steel boat lift in the form of a wheel The wheel is at the head of a semicircular reinforced concrete aqueduct standing on piers at 25m centres. The joint between the two is made by a section of steel aqueduct that accommodates movement while taking account of the tolerances required to provide a waterproof seal. The wheel consists of two massive hooked steel arms, each with a central circular aperture which supports a water-filled steel caisson. The arms, carrying up to two canal boats in each caisson, rotate through 180 degrees on a 3.5m diameter axle. A series of five stability gears is used to transmit the rotation of the central axle to the caissons. The central stability gear is fixed in a static position around the central axle. Two smaller gears revolve around this as the wheel turns, maintaining the horizontality of the final pair of outer stability gears, one around each circular opening in the arms, which are attached to the ends of the caissons. In this way the caissons cannot rock backwards and forwards but are positively located at all times. (The gears do not drive the arms of the wheel – they are ‘followers’.) At the base of the wheel is a dry well that allows the caissons to move without coming into contact with the water in the basin. The dry well includes two deep slots, which allow the hooked arms to rotate freely. Because the arms are counterweighted and geared, only 95kW of power are used in the operation. The system is driven by one hydraulic powerpack, which pumps oil through 10 hydraulic motors, each the size of a telephone. The motors use synchronised gears to turn the wheel at a controlled pace of one revolution in eight minutes.
104
105
BARAJAS AIRPORT, MADRID, SPAIN
The new airport terminal building at Barajas Airport - together with the associated satellite, rail and road connections, and car parks for 9,000 - is probably Europe’s largest current construction project. The terminal and associated satellite totals 700,000m2. It is designed to handle 35 million extra passengers annually. Including the car parks and rail station, over a million square metres of space is under construction. This is a landmark project for Richard Rogers Partnership (RRP), a marriage of architecture and engineering that evokes the spirit of Beaubourg, the building that launched Richard Rogers’ global career a quarter of a century ago.
The new airport metro link and the ambition to connect Barajas to the European TGV rail network reflect Madrid’s increasingly expansive mood. Anthony Hunt Associates, a firm that first worked with Rogers in Team 4 days, collaborated with the architects on the design of the terminal and satellite structures. In terminals on the scale of Barajas the single-level solution is likely to lead to passengers having to walk daunting distances. The competition scheme depicted a series of parallel ‘bars’ across the building, each with its own distinct function, through which passengers progressed from drop-off areas/car-parks/rail station through check in, passport and customs control to lounges and finally to the waiting aircraft. A very long pier - a prerequisite of the competition brief and 1.2km long as built - allowed for the installation of up to 40 boarding stands attached to the main terminal. And a satellite provided extra capacity, with a second satellite as a long term move.
Estudio Lamela, a large Madrid practice, invited RRP to become part of the joint venture consortium being formed to enter the 1996 competition. After winning the competition - four teams had been shortlisted - the RRP/Lamela team worked initially from the Rogers office in London, moving to Madrid when work started on site in summer 2000. During this period Luis Vidal was responsible for assembling the project team, including, for example, acoustics, landscape and lighting consultants.
The main terminal is covered by an undulating roof, equally legible as a series of wings, supported in each bar on central ‘trees’ and propped at the edges, where it oversailed to provide sun shading. The roof, free of all services, is punctuated by rooflights providing controlled daylight on the upper (departures) level of the building. But given the multi-level section, a strategy was needed to bring natural light to the lower level. The solution was the introduction of ‘canyons’ separating
The new Madrid terminal are designed as international and intercontinental hubs, with large numbers of transfer passengers. Spain sees Madrid as the key link between Europe and Latin America. For Madrid, which a decade ago lagged behind Barcelona in terms of its global image, the project is part of a reassertion of the capital city’s supremacy.
106
the linear blocks of the terminal - full-height spaces, spanned by bridges, in which arriving and departing passengers, though segregated, could share a really imposing space with a sense of place far removed from the blank corridors which greet arriving passengers at most airports. The canyons are also directional locators within the terminal, underlining the clear sense of direction fundamental to the scheme. Baggage handling (the largest system ever built) and other servicing are concentrated at basement level.
The essence of the Barajas project, in constructional terms, is its use of modular components on a standard grid. It may look crafted, but it was designed for rapid realisation and maximum flexibility. The luminous steel and glass pavilion of the terminal is classic Rogers, but it sits on a massive base of concrete, like one building sitting on another. Services are designed to fit around the modular structure and located at the edge of the floor slabs. Full air conditioning is provided only in enclosed spaces. Elsewhere a low velocity displacement ventilation system is used. It is in the skilful use of natural light that the project most clearly reflects an energy-saving agenda.
The roof is intended to float over the building, emphatically propped rather than supported at the perimeter, so that the impact of the main façade was deliberately minimised. Internally, the aim was to achieve a smooth, curvaceous form for the roof, requiring a layer of cladding covering the steelwork and the thick layer of acoustic and thermal insulation. Laminated strips of Chinese bamboo, a renewable material, are used to form the terminal ceilings. Much thought went into their setting out to achieve a seamless and flowing look, and into addressing demanding Spanish fire regulations. Natural stone is used as a flooring material throughout the terminal, a highly practical material, unlike the carpets ubiquitous in British airports. It adds to the sense of texture and integrity in the building. (The floor is curved up where it meets the façade to avoid damage from trolleys - a simple and elegant device.)
CREDITS ARCHITECT Richard Rogers Partnership STRUCTURAL ENGINEER Anthony Hunt Associates STEELWORK aluminium roof covering and deck – Corus PHOTOGRAPHS Simon Smithson
107
A glazed facade with ‘kipper’ trusses The three-storey terminal building is covered with a sinuous curved and vaulted roof, supported internally on pairs of canted columns. It extends over the glazed north and south facades to shade them, propped on external canted Y-shaped props at 18m centres. The glazed facade and its ‘kipper’ truss support structure were designed to be minimal and delicate, so as not to break the flow of the roof, inside to out. The trusses are tensioned to resist wind loading and formed of five horizontal cast stainless steel arms connected by pairs of stainless steel rods. The disposition and varying lengths of the arms create the fish-like shape (hence the name) and reflect the line of structural forces. The kipper trusses run at 9m centres, aligning with the main roof beams. They are pinned to the bottom flange of the beams with forked connectors and anchored at their bases to a posttensioned concrete edge beam. Each arm of the truss supports a transom, a set of paired 76mm diameter CHSs, bolted together with welded flanges and clad with profiled aluminium. Each transom, in turn, supports three 2.34 x 3m long doubleglazed panels. Two stainless steel drop rods at 3m centres between the kippers take out vertical deflection from the transoms – 75mm gaps were left between the glass panels to accommodate rotational movement. They were sealed with a concertina-like extruded rubber gasket. The trusses were factory-assembled, delivered to site and loosely pinned and bolted. A set of ties and jacks were set up behind them and tensioned to draw together the roof and floor beam to a maximum loading of 60 tonnes. The trusses were then attached, tested and loaded to 40 tonnes. The maximum horizontal movement of the facade will be about 15mm. SUSAN DAWSON
108
109
OFFICE BUILDING, ST MARY AXE, LONDON The real driving force behind tall office buildings, the critic Martin Pawley has argued, ‘has been the production of an enclosed interior landscape, a process whose external drama has been drained by repetition for nearly a century.’ In its reinstatement of ‘external drama’ as a key element in the design of the tall building, Foster and Partners’ 40-storey 30 St Mary Axe tower (formerly Swiss Re), has captured the public imagination and was the popular, as much as the critical, choice for the 2004 Stirling Prize. The award was also a welcome reassertion that the design of workplaces matters as much as that of the public, cultural and infrastructural projects that have virtually monopolised the Stirling since it was established.
of tall buildings in the eastern quarter of the City was a strong argument in its favour and helped to secure the blessing of English Heritage. The site for 30 St Mary Axe was created by the 1992 IRA bomb that wrecked the Baltic Exchange, a lavish but, in truth, not especially distinguished structure of the 1900s. Swiss Reinsurance (SwissRe), which occupied a number of buildings in this quarter of the City, took an interest in the site late in 1997. The Foster scheme was developed by SwissRe in association with Skanska. Norman Foster sees St Mary Axe as ‘a considered attempt to break down barriers and improve the quality of life in the workplace’. If the sculptural, curvilinear form of the building has precedents in earlier Foster projects such as the Bilbao metro and Canary Wharf station, the issue of rethinking office design is one that has preoccupied Foster since Team 4 days. Willis Faber, the Hongkong Bank and the Frankfurt Commerzbank (credibly claimed as the first ‘green skyscraper) are landmarks in the quest, which continued in the Foster scheme for the World Trade Center site.
Tall buildings have long been a contentious issue in London. It is nearly 70 years, long before the advent of modern highrise buildings here, since the City promulgated the St Paul’s Heights policy to protect distant views of Wren’s cathedral. At 590 ft (180m) St Mary Axe easily oversails its sixties neighbour, the Commercial Union tower, though, probably deliberately, it allows the NatWest tower (completed in 1981 and now renamed Tower 42) to retain its pre-eminence as the City’s tallest building. The location of the Foster tower in a cluster
110
CREDITS ARCHITECT Foster and Partners STRUCTURAL ENGINEER Arup STEELWORK steel frame – Victor Buyck-Hollandia metal deck -Richard Lees Steel Decking curtain wall –Schmidlin steel frame and glazing to apex – Waagner-Biro architectural metalwork - Glazzard PHOTOGRAPHS Nigel Young
The profile of the tower went through a series of revisions – at one stage it was a portly egg - reaching its final form during 1998 after protracted negotiations with City planners and with English Heritage (EH), which was prepared, in effect, to sacrifice the remains of the Baltic Exchange for an outstanding new building. While EH was eventually convinced by the basic rationale of the Foster proposal, with its spiraling internal atria corkscrewing through the fingers of office space, and taken with its aerodynamic shape, it wanted a slimmer profile even if that meant an increase in overall height from the 130m proposed at concept stage. The final version at 180m is more bullet-like than gherkin-like. A planning application was lodged in June 1999 and the project began on site in 2001. This provided for a building containing around 46,000m2 net office space, considerably more than that on offer from the groundscraper solution, though a reduction on the potential area of the earlier egg scheme.
spiraling light wells clad in grey glazing, and the use of paint to emphasise the vertical four-storey diamonds in the diagrid cladding, with horizontal elements played down to reflect the structural agenda, emphasise the verticality of the tower.) While much attention had focused on the long-distance impact of the building, Foster was equally concerned with the way it would be experienced at close hand. ‘The City within a city has always been full of surprises,’ he noted. The most impressive views of St Paul’s, for example, are the incidental glimpses seen through gaps in the building line. The great strength of this project is the degree to which every aspect of the programme – social, structural, environmental, spatial – is fundamental to the whole. In its entirety, it represents the realisation of aspirations that have been present in Foster’s work for many years. It is claimed, for example, that, largely thanks to the double-skin ventilated façade, energy consumption is potentially half that of a prestige airconditioned office building of similar size (though fine-tuning by the building’s users will be critical in this respect). Foster and Partners claim that this is ‘a radical building… a paradigm of the responsible environmental practice that is a quest for both architect and client.’ For Norman Foster, it is the practical embodiment of Buckminster Fuller’s vision of ‘a micro-climate within an energy-conscious envelope’.
The site of 0.57ha was an irregular rectangle, bounded by two narrow passages to north and south and with the modest thoroughfares of St Mary Axe and Bury Street (with Berlage’s famous Holland House) to the west and east. The Baltic Exchange had filled it completely. The churches of St Helen, Bishopsgate, and St Andrew Undershaft, both precious survivors of the Great Fire, are near neighbours. (The
111
A glazed dome with a glass lens at the apex In shape, the 40-storey building is a prolate spheroid; the perimeter superstructure, apart from the top three storeys, is a diagrid formed of rings of two-storey A-shaped steel frames bolted together and clad with a glazed curtain-wall system of faceted triangular panels. The upper three storeys – in effect a dome, 30.5m in diameter at the base and 22.5m high – contain private dining rooms, a restaurant and bars for the users of the building. The dome is clad with triangular doubleglazed panels – 10mm toughened (heat-soaked) grey body-tinted outer with interpane coating to the inner face, 16mm argon-filled cavity and 12.76mm laminated inner. Its structure is self-supporting and rests on the final ring of the A-shaped superstructure. It consists of fabricated mullions and transoms welded to follow the complex 10º faceted geometry of the cladding. The mullions and transoms were welded at the factory into a series of ‘ladder’ members and ‘loose’ members. The size of each ladder member was determined by the maximum size for transport. Two ladder members were placed in a 30º jib on site and welded together with the loose member between, an arrangement determined by the lift capacity of the tower crane. They were lifted to the top of the building and welded together to form the structure. At the top, the dome terminates in a convex glass lens; it rests in a circular ‘spider’ structure, which was welded to the tops of the ladder members. The lens is a double-glazed unit, 2420mm in diameter. It is flanked by a ring of insulated aluminium panels connected to air-extract and smoke fans. The fans are concealed by curved aluminium panels, which also house recessed light fittings. SUSAN DAWSON
112
113
A diagrid perimeter structure of steel columns and nodes The 40-storey building is a prolate spheroid in shape. It has a central steel core from which floor beams radiate to a perimeter superstructure, creating column-free interiors. The superstructure is a diagrid, formed from a series of two-storey A-shaped frames bolted together. Each frame consists of a pair of columns bolted to a fabricated steel node. The columns have circular end plates which are bolted to matching circular connection plates, welded to the nodes at the precise angle needed to create the diagrid. Both sets of plates are machined smooth and flat in the factory and drilled to achieve an accuracy of within 1mm when bolted face-to face. By prefabricating components to precise dimensions before assembly, the complex angular relationships between components are resolved before erection. The sequence of construction is as follows: as soon as two floors of core structure are erected, a perimeter row of A-shaped frames is craned into position. Tie plates at both sides of each node are bolted to hoop-ties by means of tapered flanges, connecting the nodes horizontally. A series of floor beams, radiating from the core, are bolted to the backs of the nodes. They support a composite concrete structural slab, the metal deck of which is laid in trapezoidal sections. The diagrid steelwork is fire-protected with foil-backed mineral wool blanket and clad with V-shaped panels of white polyester powder-coated aluminium. Cladding brackets bolted to the node and to the hoop-ties support a curtain wall of diamond-shaped panels which are faceted at their edges in both plan and section. SUSAN DAWSON
114
115
A STAND OF TWO HALVES Careful planning of Chelsea FC’s West Stand, now in its second design stage, has enabled AFL to create an extra 750 seats BY SUSAN DAWSON. PHOTOGRAPHS BY RICHARD BRYANT/ARCAID
In 1905, when the first fans passed through the gates of Chelsea Football Club’s Stamford Bridge ground, the only shelter was a single covered stand on the east side of the pitch. The other three sides were terraces, open to the weather and formed of banked earth – conveniently provided by excavation of the Underground’s District Line. Stands were built and improved over the years, but an attempt in the 1970s to develop the ground into a 50,000 all-seater stadium ended in disaster – rising debts forced the club to sell the ground to developers. After some anguished years, it was finally bought back in 1992.
more than 750 extra seats in without compromising comfort or increasing the plan or overall building height.’ (Like many other traditional football grounds, Stamford Bridge is hemmed in by adjacent buildings, which severely restricted the size and height of the new stand). Whitby Bird & Partners was the structural engineer for both phases. Stand design is determined by spectators’ need for a good view of the whole pitch over the head of the person immediately in front. The sectional shape of the second phase was determined by the rake of the upper tier, with 23 rows of seats, and the need to slot in three levels of executive seating between the upper and lower tiers.
The Stamford Bridge site on Fulham Road has now been transformed into the largest league ground in London, with covered stands on all four sides of the pitch, seating 42,450 fans. Three stands have been rebuilt in the past five years, with the last of these, the West Stand, now completed.
The tiers rest on a cast in situ frame six levels high, which accommodates concourses, WCs, lifts, staircases and circulation areas. A row of executive boxes with glazed fronts and outdoor seating – eight people per box – runs at the back of the lowest tier with two lounges at the rear. Projecting above it, in an 11.5m cantilever, is the prime viewing area, eight-seats deep and projecting where club directors sit. The chairman’s lounge and two VIP suites are at the rear, together with two lounges for 300 people on each side.
The West Stand was built in two phases: the concept and first phase – the lowest tier and ground-floor concourse – was designed by KSS Architects; detailed design of the second phase – three upper tiers, executive and conference suites and the roof – was won by Atherden Fuller Leng (AFL) in competitive interview. ‘We won the job,’ explains architect John Roberts, ‘because we found that we could, with careful planning, get
116
A third level, two-seats deep and projecting in an 8.2m cantilever over the directors’ seating, is linked to the Millennium Suite, a series of 24-person self-contained suites. They are available for daily use as individual offices and are designed to ensure that the club maintains an income whether or not matches are being played. The suites, which have fully glazed openings facing the pitch, contain conference and dining facilities including kitchens and WCs, and are backed up by additional clubrooms at the rear. The upper tier is the largest and most dominant element when seen from the pitch – it rises at the rear to just below the back edge of the roof, a gently arched steel canopy of steel cellular beams supported by a pyramid mast structure. The new stand gives fans comfortable seats, shelter and a clear view – something to celebrate in addition to scoring goals. CREDITS ARCHITECT Atherden Fuller Leng (AFL) Project architect John Roberts STRUCTURAL ENGINEER Whitby Bird & Partners Ben Rowe, Des Mairs PROJECT MANAGEMENT MPM Capita
MAIN CONTRACTOR Multiplex Constructions SUPPLIERS Steelwork Westbury Tubular Structures Roof covering Lexan polycarbonate, PMF profiled steel sheet
117
A steel canopy roof supported by a pyramid mast structure The roof is a gently arched canopy of steel cellular beams supported by a pyramid mast structure that spans 104m to shelter the stand. It is covered with profiled steel sheet at the rear, and with profiled polycarbonate sheet at the front to ensure that high levels of light are transmitted through it to promote healthy grass growth on the pitch. The sheeting is supported by rectangular steel purlins 2.8m apart. These are supported in turn by 1,014mm-deep cellular beams at 6m centres. The beams are propped at the rear of the stand, and span over an inclined truss to cantilever 15m beyond it to the edge. The inclined truss runs the length of the stand, with a central span of 72m and cantilevers of 16m at each side. The top chord of the truss follows the curved profile of the roof while the bottom chord is straight, so that the truss depth varies across the length of the roof. A walkway gantry for lighting and speakers cantilevers from the bottom chord of the main truss. Vertical truss members are rigidly connected to the cellular beams to carry the gantry and the end cantilevers. The truss is supported at two pinned suspension points, 72m apart, by ‘pyramid’ structures, each an asymmetric ‘A-frame’ of three 610mmdiameter CHS supports. Two of the supports transmit large compression forces, while the third, at the rear corner of the roof, transmits a large tension force. Each pyramid structure is supported by an 864mm-diameter CHS mast at the gable; the mast has a splayed base for connection to a 1,500 x 300mm steel column in the reinforced-concrete gable wall.
118
119
RACING TIMES Foster and Partners and Whitby Bird & Partners have created a grandstand design that avoids the obvious cantilever roof BY SUSAN�DAWSON. PHOTOGRAPHS�BY�NIGEL�YOUNG The new Tattersalls Grandstand at Newbury Racecourse, Berkshire, demonstrates that a simple cantilever roof is not the only way of sheltering large numbers of viewers – whether of racehorses or football. The building is formed of six huge steel X-frames, 12m apart. The top legs support the roof – rather like the legs of an ironing board – while the lower legs enclose a large betting hall on the ground floor and follow the line of the ‘steppings’ – the stepped terraces on which spectators stand to watch the races. They have open corridors between them that give access to the betting hall.
The steppings rise to a bar, with fully glazed walls that run the length of the grandstand. The floor above houses a 600seater restaurant, also fully glazed for racecourse views, with projecting balconies set between the upper legs of the Xframes. The rear of the grandstand, a conventional steel-frame structure set within the X-frame legs, is four-storeys high and accommodates staircases, WCs, plant and services. The new grandstand is designed primarily for racing but, like other new sports buildings, the restaurant and bar have been designed so that they are suitable for receptions, banquets, exhibitions and
120
CREDITS
The X-frame provides structural stability that can cope with both the static loads and the dynamic loads when racegoers move about. It was modelled using Xsteel software
ARCHITECT Foster and Partners: Ken Shuttleworth, Charles Rich, Dominic Skinner, Matthew White, Michele McSharry, Viktorie Smejkalova STRUCTURAL ENGINEER Whitby Bird & Partners Charlie Benson, Eloise Plunkett MAIN CONTRACTOR Heery International STEELWORK CONTRACTOR Watson Steel
The structure was modelled using XSteel software, a package that forms a three-dimensional model of the steel structural members from the earliest stages of a project. The model was passed to Watson Steel, which won the contract to fabricate and erect the steelwork, and was used to provide all workshop drawings. An electronic format of the model was also issued with the tender documents, so that an exact schedule of components could be extracted from it. This removed the need for a detailed Bill of Quantities and shortened the tender period.
conferences, and are available to generate revenue throughout the year, not just on the days of racing fixtures. Steel was the obvious material for the structure. The simple logic of the X-frame structure ensured value for money – the whole project cost only £9 million. Steel’s ability to be prefabricated ensured fast construction – the client had stipulated that the old stand be demolished and the new one be in place between the dates of the annual November Hennessy Gold Cup race meetings. With less than a year allowed for construction of the grandstand, the design made the most of off-site fabrication so that it could be delivered to site as a ‘kit-of-parts’, ready to be assembled.
The front and side walls of the grandstand, giving important views of the racecourse, are largely glazed. The four-storey structure at the rear is clad with grey profiled metal sheeting – the ‘crinkly tin’ familiar on industrial estates, but here enhanced by an overlay of stainless steel mesh. The two materials combine to produce a material of classic quality with a matt finish so that horses are not dazzled by reflections as they race towards the winning post.
Charlie Benson of engineer Whitby Bird & Partners explains the advantages of the design: ‘Structural stability is inherent in the plane of the X-frames, although the layout created significant “out-of-balance” forces which had to be stabilised’. These forces, which include the load created by the large balconies cantilevering from the front edge of the restaurant floor, would tend to cause the upper sections of the X-frames to pivot forward at the node crossing. To prevent this, an exposed vertical tie was fixed between the rear upper and lower tip of each X-frame leg, and connected to an adjacent column at the rear of the restaurant support beam. Unlike football fans, racegoers move around between races. This imposes unusual dynamic loads on what is essentially a lightweight structure. Due to the angle of the raking legs, the Xframes do not behave like simple columns but combine vertical deformation with sway. A finite-element analysis showed that, even with a restaurant full of spectators jumping up and down in unison, there was no risk of excessive movement.
121
A steel X-frame structure The new grandstand has a structure of six exposed X-frames set 12m apart. The legs are formed of CHS members, chosen for their capacity to accommodate high compressive loads. The node crossings were prefabricated in the factory and brought to site, where the legs were welded to them. The lower legs are formed of 660mm diameter CHSs; the upper legs are formed of 508mm diameter CHSs from the node crossing to the balconies, and of 275mm diameter CHSs from the balconies to the tips. At foundation level, the ends of opposing legs span 18m; they are set in precast shoes and tied together with a 356 x 406mm UC which is cast into the ground-floor slab so that only vertical and lateral wind loads are transferred to the piled foundations. The six lower members of the X-frame enclose the main betting hall. The floor of the bar above rests on a central steel spine beam fixed just below the node crossings of the X-frames. The stepped beams that support the steppings are fixed to the spine beam. The restaurant floor rests on fabricated steel beams fixed between the upper legs of the X-frames. The beams are connected to them by saddle-plates; each end is also bolted to a vertical column that prevents the tendency of the upper section of the X-frame to pivot at the node crossing. Bar and restaurant floors are constructed of precast concrete planks with structural topping, which reduced the need for secondary steelwork and shortened the construction programme. The diagonal framework of the roof structure acts as a diaphragm and ties the top ends of the X-frames together to prevent sway. It supports a profiled metal deck covered with a single-ply membrane.
122
123
MAKING A SPLASH The Manchester Aquatics Centre, designed by FaulknerBrowns, hosted the Commonwealth Games in 2002. BY SUSAN DAWSON PHOTOGRAPHS BY IAN LAWSON
As the name implies, this new building is more than just a swimming pool; it offers flume rides and a toddlers’ paddling pool but also caters for every aspect of competitive swimming including solo diving, synchronised diving, synchronised swimming and water polo. The complex will be one of the major venues of the Commonwealth Games to be held in Manchester. Commissioned by Manchester City Council in collaboration with the three universities in the city, it can be used by students and locals.
Supporting accommodation is arranged in a three-storey concrete frame structure along the south side of the pool hall. The changing village and health suite are set on each side of the double-height reception, with a café overlooking the pool. Upper levels contain a fitness suite and a dance studio, and give access to the main spectator gallery of 1000 seats which will be used for the Commonwealth Games. For such occasions extra temporary seating can be provided, in part, over the leisure area to give up to 2500 spectator spaces.
The new pool building fronts the busy Oxford Road and forms part of the UMIST campus; student halls of residence are behind and the building is adjacent to the new School of Management, designed by ORMS. The gable end on the west faces the main road; the 100m long south side - a series of stone-clad service ‘chimneys’ and glazed walls – forms one side of a new pedestrian public space and contains the main entrance. Visitors enter a wide double-height reception which leads to the café and pool hall, an ‘heroic’ 32 x 100m space.
The mechanical and water treatment plant is contained in the basement. This also contains a unique facility, an additional 50 metre training pool whose purpose is to provide dedicated elite swimming training. The pool has a traversible boom to adjust its length to 25m, and a retractable floor to allow water of varying depth for teaching; it has its own storage and changing rooms, a sports science room, a sports medicine room and a land training room containing isokinetic fitness equipment.
The central pool, the 50 metre competition pool, has underwater observation windows and a constant 2m depth which can be made shallower by two movable floors; the pool can also be divided by submersible booms.
The structure The 32 x 100m pool hall is a column-free space formed by a steel roof structure which arches asymmetrically over the pools reaching almost 20m at its apex and spanning 37m over the spectator seating. Along the north elevation and facing the spectator seating and entrance runs a series of plate girders 7.7metres apart, which spring from stone-clad thrust blocks. They rise, increasing in depth and curving to meet a flattened ridge, a double ridge truss flanked by rooflights; at this point the plate webs change to trusses which slope down to the south elevation to bear on cast-in-situ concrete columns and core
At the east end a 5 metre deep diving pool incorporates underwater observation windows and a retractable floor. The cast insitu concrete diving platforms have 3m springboards together with 3, 5, 7 and 10 metre fixed platforms. At the west end of the hall is a free-form shallow water lagoon for children; a play area, flume rides and bubble pools are landscaped with artificial palm trees, emphasising the leisure, as opposed to serious swimming, function.
124
walls beyond. They are partly screened by a waved acoustic ceiling which helps to provide the acoustic environment needed for the voice alarm evacuation system.
solar shading, articulate the façade, and give access to the curtain wall for maintenance.
Protecting the steel structure
Eight of the plate girders are deeper than the others and are arranged in four pairs, with horizontal and diagonal CHS purlins and bracing members between them. The pairs are located at the gables and at mid-point, at each side of the main staircase. The gable girder pairs are braced with additional CHS props which are pinned from the bottom boom of the outer girder to the top boom of the inner girder to stabilise the gable and transfer wind loads from the gable to the roof plane. These pairs of frames support the ridge trusses which in turn support shallower intermediate frames.
Steel was the natural choice of structure for the pool, for its ability to span large volumes economically. The technology of corrosion prevention has been highly developed for areas with much greater risk, such as oil exploration platforms. The fabricated steel components were shot-blasted and treated with 50 micron epoxy zinc phosphate primer, a barrier coat of 125 micron micaceous iron oxide followed by a 50 micron decorative top coat.
The top and bottom flanges of the main girders are made up from 254 x 254mm universal columns with 12mm thick steel plate webs welded between them. A 20mm thick 254mm wide steel plate replaces the universal column for the top flanges of the intermediate frames. The girders are set at 7.7m centres and are braced by a series of horizontal CHS purlins welded to the webs, providing lateral stiffness and creating a simple and elegant structural rhythm along the north wall. The webs of the girders were welded in sections; the welded junctions are designed to occur where the CHS purlins are welded to the webs. The junction is concealed by stiffener plates which, together with the site-welded purlins, give lateral stability to the structure. The difference between the paired girders, acting as box trusses, and the smaller girders which run between them is emphasised by a subtle change in soffit treatment. The smaller girders are lined with a perforated acoustic structural ‘Kal-tray’; the large paired girders are lined with structural ‘Kal-deck’, a more articulated trapezoidal profiled liner.
CREDITS
The west gable of the pool flanks the pavement and busy highway of Oxford Road. The strongly curved girder form is reflected in the verge which is clad with silver-coated Luxalon sheet, cut and flashed to replicate the girder shape behind it. The curtain walling system below it comprises a polyester powder coated aluminium frame with bands of double-glazed units of clear glass inset with adjustable micro-blinds, and opaque ceramic-backed glazing units backed with insulated aluminium trays. The façade is divided into horizontal bands by painted steel walkways with GRP grid floors; they give
ARCHITECT WATER TREATMENT FaulknerBrowns ENGINEERING Nick Deeming, Jean Paul Colback FaulknerBrowns Engineering Services STRUCTURAL ENGINEER
125
Ove Arup & Partners, Newcastle on Tyne Gordon Mungall, Sarah Clemmetsen, John Gregory
QUANTITY SURVEYORS Tozer Capita
MECHANICAL & ELECTRICAL ENGINEER Ove Arup & Partners, Manchester
STEELWORK CONTRACTOR Watson Steel
MAIN CONTRACTOR John Laing Construction
Steel mullions supporting a glazed gable wall As the roof structure runs for 100metres without movement joints, it was necessary to design the gable wall as an element of structure which allowed the roof to move above it. The gable wall structure consists of a series of fabricated steel mullions at 3.6m centres. They support the gable wall and the walkways with a top fixing that accommodates movement of the roof structure in three planes. In section each mullion is an asymmetric oval, tapering at one end. It is formed of an inner 406 x 178mm UB lined with 12mm steel plate. The curved ends are formed by CHS tubes cut in half; a half–tube of 244mm diameter is welded to one flange; the other flange is shortened and welded to a half-tube of 114mm diameter. Each mullion has a welded lug at the top which slots inside a pair of plates welded to the bottom flange of the outer gable girder. The slot allows the girder to move up to 10 – 15mm vertically and horizontally. An insulated slotted movement joint is incorporated in the cladding on the same plane. The gable mullions are restrained by a series of CHS props which retain the mullions within the slots. The lower props are horizontal; the topmost prop is curved to follow the line of the girder. The walkways are supported on brackets which cantilever from the main mullions and are additionally propped by 25mm diameter rods suspended from the verge. The rods are held by tapered T-sections connected to the bottom boom of the gable frame. A 50mm structural thermal break penetrates the façade at all connections.
126
127
PUMP UP THE VOLUME The new Wilmersdorf pumping station is a delicate glass and steel structure, featuring an interior gallery that overlooks the pump areas and gives visitors an insight into its workings BY SUSAN DAWSON
The city of Berlin squats on a flat plain with a very high water table. This has meant that for nearly a hundred years drainage and sewage have been distributed by mechanical pumping. The city has 140 pumping stations, some approaching 100years-old, standing as wonderfully robust monuments to its industrial past. The old red-brick station at Wilmersdorf, on the west of Berlin in the heart of a densely developed urban district, was built in 1903-6. Although its structure is sound, the equipment has reached the end of its life.
A visit to the gallery gives a true picture of the building. Running along each side of a huge well, the gallery measures 8m wide and nearly 40m long. Look over the railings and a Piranesian underworld is revealed, containing pumps, ducts, generators, sewage tanks and plant. This is where the real technical business is done. Staircases and gantries in the central well give access to the four levels of plant. The vast basement is flanked by concrete walls measuring more than a metre thick. The whole basement block was cast above ground as a caisson and then lowered 18m to its final position by flushing out the sandy subsoil beneath it. Although this may seem a relatively unusual construction method, it is frequently used in Berlin because of the high water table.
Following a proposal to build a new pumping station next to the old one, which is being converted into a technical museum, the Berlin water authority organised a competition. This was won by the Munich architect Ackermann und Partner. In a deliberate contrast to the cathedral-like proportions of the old station, the new building announces itself as a delicate steel and glass structure, two-storeys high. The interior – a gallery resting on a mezzanine – is visible through a glass facade that minimises the impact of the bulky 20 x 40m building. The gallery is open to the public and while most of the visitors are expected to be schoolchildren, it can also be used for exhibitions, book launches or small parties.
Above ground, the mezzanine – a solid enclosure formed of concrete slabs supported by sheer walls and composite steel columns – runs at the sides of the central well. The first-floor gallery, with its steel structure, sits on top of the mezzanine. Clearly visible through the glass skin, it consists of four 20m triangulated truss girders, which bridge the central well. These are supported at each end on trestle-like props with canted legs
128
(see Working Detail). The connection between the trusses and the props – the most visible part of the structure – is made by an articulated cast-steel node, giving the structure a machinelike quality; a reference to the great pumping machines below. The composite roof is fitted with a series of large rooflights above the well. These are fitted with prismatic glass, which directs daylight into even the deepest parts of the basement. Because it is closely surrounded by houses and flats, the huge glass facade of the pumping station was designed to provide sound insulation and save energy by exploiting solar gains. It consists of two separate glass layers. There is an external layer of 15mm toughened glass sheet connected by stainless steel structural bolts, an intermediate space 700mm wide that allows for maintenance and gives a means of escape in a fire, and an internal skin of steel-framed double-glazed units. The facade support structure is independent of the main building. It uses a series of 200 x 30mm steel columns at 3.3m centres on the outer layer of the facade, braced at the corners with tension rods. Similar columns at 1.65m centres support the inner layer. Although there are no permanent workplaces in the pumping station, temperature extremes had to be avoided. Even under exceptional weather conditions, the double-layer steel and glass facade is designed to maintain an internal temperature of about 14°C. No additional energy is used to keep this temperature constant. The six electric and diesel pumps emit 199kW of waste heat, with a total capacity of 20,000 m3/hour.
CREDITS ARCHITECT Ackermann und Partner
Above: the steel structure of the gallery is clearly visible through the glass facade.
PROJECT ARCHITECTS Christof Simon, Eoin Bowler
Right: the truss girders are supported by trestle-like props with canted legs, connected by articulated cast-steel nodes. Below: the pump station plan, showing the welllike structure
STRUCTURAL ENGINEER Buro Happold, Bath and Happold Ingenieurbüro, Berlin: Terry Ealey, Michael Vitzthum
129
A trussed girder structure with canted supports A framework of tubular trusses with canted supports rests on a solid mezzanine floor. Both are visible through the two-storey glazed facade of the 20 x 40m pumping station. The structure was designed as a delicate filigree of steel tubes, which would clearly indicate the line of forces. It consists of four triangulated truss girders, each measuring 1.3m deep and spanning 20m, which support a steel/concrete composite roof. The paired top booms and bottom boom (168.3mm diameter CHSs) are connected by canted 114.3mm diameter CHS props. The load from the roof is transferred to the trusses by 168.3mm diameter CHS compression members. These are coupled to the top and bottom chords of the trusses by 40mm diameter steel tension rods, creating a cantilever action on both sides of the truss. Each truss is supported at its ends by two pairs of canted 193.7mm diameter CHS legs, which rest on floorplates alongside the glass facade. The plates also act as fixing points for a pair of vertical tension rods, which transfer loads from the top chords of the truss down to the concrete floor. The plates are positioned on concrete pockets to distribute the additional horizontal force component from the diagonal legs, through the concrete table and sheer walls, down to the concrete slab below ground. As some of the sheer walls did not match the position of the legs, loads were modelled on a computer program. The roof consists of 180mm deep insitu concrete cast on a profiled steel deck. At the perimeter, the roof is supported by a 300mm deep universal beam, which is bolted to the top booms of the trusses. To reduce the level of noise from the pumps, the roof and the facade are separated from the supporting structure by elastomeric anti-vibration bearings.
130
131
SCHOOL CAMPUS, BROMLEY, KENT Hayes School in Bromley comprised an existing Victorian house, part of a motley collection of typical post-war school buildings (mostly from the ‘50’s and ‘60’s) and temporary huts. The competition brief set out a requirement for several general classrooms, music and IT spaces, a new library, multipurpose hall and additional changing rooms. The architect has taken the opportunity to give the school a new image, using new buildings to join the disparate collection of buildings together with new circulation routes, thereby giving the campus a new sense of identity.
residence – and the visitor is presented with the bold concave facade of the new library. This fragment of a circle is rooted to the radial geometry of the paving and balanced by the Gothic geometry of the ‘lodge’ (which appropriately contains the reception and administration). The second of the project’s main features is the school ‘street’. This new circulation route fills in the gaps between the library and the multi-purpose hall, around the main school hall. It then forms a wide, glazed colonnade along the south side of one teaching block and links to the end of a second. What were dull, awkward and left over outside spaces have been transformed into the unifying spine of the school’s circulation. But it is much more than a corridor, and as it emerges from
The first new feature, and the one which sets the image for the whole school, is the creation of a new ‘square’ – a round one – at the entrance. The school is approached from the road through mature trees – these were the grounds of a substantial
132
CREDITS ARCHITECT PCKO Architects STRUCTURAL ENGINEER Price & Myers PHOTOGRAPH Grant Smith
between buildings into the daylight it is a genuine promenade and a usable space it its own right.
For the two particularly visible elements of structure – the columns which support the welcoming roof overhang to the library, and the row of seemingly random columns in the school street – the architect has devised, with the engineer, a slender tree-like column, branching at the top into four or six struts.
The third part of the development is an L-shaped range of classrooms for music, media and general teaching in the angle between hall and library that helps to define a separate, subsidiary, courtyard.
There are a number of nice spatial touches. The corridor around the new music rooms is in the middle of a deep space and, to bring light in, there are three very long and narrow curved lightwells in the first floor slab. One of these lightwells is cut out on both sides of the wall separating the corridor from the library, leaving the glazed corridor intriguingly freestanding. Where the school street meets the existing refectory, an internal terrace with sinuously curving front wall makes a dramatic transition between old and new.
Two particular methods have been adopted to create a sense of unity – and an individuality that lifts the campus above its previous ordinariness. Generally speaking, the new buildings are constructed using simple, everyday materials and techniques: steel frame, rendered blockwork, aluminium curtain walling the profiled metal roofing are all used, without reinventing details.
133
A glazed ‘street’ supported by branched columns The space between two ‘60s two- and three-storey teaching blocks has been enclosed with an 8m high roof and a glazed wall to form a ‘street’, 7m wide and 60m long. It links the old buildings with new ones and creates an informal space for children to meet as well as a more structured space for exhibitions or displays. The monopitch roof rises 8m to overlap the eaves of a three-storey block; a row of aluminium louvres set below the ridge ventilates the street by stack effect. The glazed roof is supported by five branched steel column ‘trees’, each with a 323.9mm diameter x 25mm CHS ‘trunk’ with a splayed oval top plate. The branches, 114.3mm diameter CHS struts, are fixed to the trunk with pin joint connections at irregular heights and inclinations; their design refers to the informal arrangement of large mature trees on the site. The branches are triangulated with the CHS roof members and lateral support is provided by fixing two branches back to the second-floor slab of the adjacent classroom block. The roof is composed of single-glazed 6.4mm laminated glass panels on 134 x 51mm aluminium glazing bars at 600mm centres. Glazing bar cleats are fixed to angle brackets welded to the CHS structure. Four panels of glass form the roof slope, jointed together with silicone. To reduce solar heat gain, the two upper panels are laminated with a white translucent interlayer. Rainwater drains into an aluminium channel fixed to the ends of the glazing bars. The curtain wall is supported by a row of 139.7mm diameter CHS columns braced by horizontal CHS members.
134
135
ROOF IN A LANDSCAPE Hampshire County Council’s architects department has designed three buildings in a visitor centre, united by an oversailing roof BY SUSAN DAWSON PHOTOGRAPHS BY JUSTYN WILLSMORE Harold Hilliers (1905-1985 ) had a passion for plants. He owned a large and successful nursery but his dream was to create a significant plant collection. In 1952 he bought a modest country house set in 32 hectares to the east of Romsey in Hampshire to realise this ambition. Hilliers exchanged plants from every corner of the world; this allowed him to rescue and propagate plants which had been on the verge of extinction – not only from the wildest places in the world but also from neglected gardens nearer home.
better access and circulation and to create a logical site for a new visitor centre at the main public entrance. For the visitor walking up from the main car park, it is the roof of the centre which catches the eye, extending out beyond the building enclosures to shelter pedestrian routes and forming an effortlessly daring cantilevered canopy at the south-west corner which acts as at the main entrance. As you approach the canopy the layout of the visitor centre becomes clear; it consists of three simple yet separate pavilions arranged on three sides to form a courtyard around a landscaped pool. The roof extends over the three sides of the courtyard in a U-shaped plan; it links the pavilions together and seems to hover effortlessly over the walls, extending beyond them and sloping gently upwards to terminate in an elegant tapered eaves.
Today the Sir Harold Hilliers Gardens and Arboretum contain one of the greatest collections of hardy trees and shrubs in the world and are an important centre for education and conservation – 80,000 visitors a year come to visit. Hilliers secured the future of the gardens by forming a charitable trust, of which Hampshire County Council was the sole trustee. Since then the gardens have expanded: they now cover 72 hectares with 42,000 plants of around 12,000 taxa (types) and house the largest number of National Plant collections to be found on any one site.
‘The design’ explains Georgina Hall, project architect, ‘aims to create a new threshold for the visitor to experience the landscape. We felt the need to embed the building in its surroundings. The roofline has been kept as low as possible but it has a very delicate edge which cants upwards’.
By 2000 the popularity of the garden was such that access, circulation and visitor facilities had to be improved to cope with the numbers. Hampshire County Council asked its architects to design a new visitor centre. It was developed in collaboration with landscape architect Colvin & Moggridge;, some years previously they had drawn up a detailed master plan which sought to grasp the whole area and gently re-shape it to give
The choice of roof design allowed the three pavilions, with their individual requirements for enclosure and open space, to be unified. The entrance canopy, through which visitors pass to enter the courtyard, is flanked by the main foyer space on the left with its ticket desk and, beyond, the restaurant. To the right is a self-contained shop and WC block; the block at the west
136
side of the courtyard contains offices, a small laboratory and classrooms which can be adapted to act as a lecture theatre. The restaurant is glazed on three sides to give views northwards to the oak fields and eastwards to the winter garden. Here the roof extends beyond the glazed north wall to give shelter to a terrace – an extension to the restaurant for use in warm weather. On the northern side of the courtyard a pool creates a subtle natural boundary to prevent visitors from entering the garden without a ticket. The structure comprises a series of inverted steel trusses (see Working Detail) which rests on a central row of main columns at 6.6m centres, with secondary small posts buried in the walls on the high side of the pavilions The steelwork contract was won by structural steelwork fabricators Allslade, who used XSteel software to create a 3-dimensional model of the steel structural members. The software files were then transferred to Allslade’s cutting machines which cut and labelled each steel element in the sequence in which it was going to be erected. The trusses and their tubular spigots were welded in the factory; they were then taken to site, lowered onto the columns and connected by purlins. CREDITS
The roof is covered with terne-coated stainless steel and the soffitts are clad with The roof is covered with terne-coated stainless steel and the soffits are clad with cedar boarding. To make the most of the views, and natural light the walls are glazed on all sides of the courtyard . Around the outside edges of the pavilions local Michelmersh bricks are used, with clerestorey windows above.
ARCHITECTS Hampshire County Council Richard Gooden Georgina Hall Colin Henry, Paul Bulkeley Philippa Dickson LANDSCAPE ARCHITECT Colvin & Moggridge QUANTITY SURVEYOR HCC QS Department
The visitor centre opened to the public in June 2003, and over the next few years the new tree planting, and extensive landscape works will mature, to further integrate the building into its surroundings.
STRUCTURAL ENGINEER Price and Myers Paul Batty, Simon Jewell
137
MAIN CONTRACTOR Brazier Construction Division of Kier Regional STEELWORK SUBCONTRACTOR Allslade SUPPLIERS Roof - terne-coated stainless steel
A roof structure of steel trusses connected with spigots to circular columns The visitor centre consists of three single-storey pavilions grouped to create a courtyard with a landscaped water garden. The roof structure, a series of inverted steel trusses, forms a continuous shelter, U-shaped in plan, over the three pavilions. The trusses are supported on pairs of columns at 6.6m centres; a 244mm diameter CHS main column and a small 150 x 100mm RHS post propping the high point of the truss. Each truss is over 10m long; sloping downwards to cantilever over 3.3m on the inner, courtyard side. Cantilevered beams with via thermally broken connections project from the ends of the trusses to support the roof edges. The trusses are formed of fully welded 152 x 152mm UC members and their connection to the circular columns needed to have some moment capacity to provide portal action. Instead of the more conventional approach – which would rely on a bolted connection, the structural engineer used a spigot detail. Over each column, the beams of the truss are welded to a 323.9mm diameter CHS sleeve which forms the spigot. After the columns were erected, each truss was individually craned over a column and the spigot was lowered onto it and after lining and levelling the spaces between them were then filled with grout. This elegant connection requires only a few locating bolts where the truss sits on the column. The trusses were subsequently bolted together with 152 x 152mm UC purlins. The cantilever at the corners required a slightly different steel structure – a diagrid. The same detail was used but with additional side plates to allow the diagrid members to be slotted together. Flitch plates are set in to the external walls of the pavilions; they act as continuations of the trusses while allowing thermal breaks to be incorporated.
138
139
WATERTIGHT DESIGN For its new operations centre, Wessex Water wanted a green design to reflect its commitment to sustainability. Bennetts Associates delivered an unintrusive, eco-friendly solution BY SUSAN DAWSON. PHOTOGRAPHS BY PETER COOK It is hard to imagine a more beautiful site. The new operations centre of Wessex Water faces south on the edge of the Limpley Stoke Valley, an area of outstanding natural beauty on the outskirts of Bath, and an appropriate place for a very green building. A BREEAM assessment has shown the Wessex Centre to be the greenest commercial office in the UK, consuming less than a third of the energy required to power a standard headquarters office building.
east-west axis to catch the prevailing wind, enhancing natural ventilation and giving everyone the chance of views over the wooded valley. Communal spaces, such as the restaurant and meeting rooms, are set at the west side of the ‘street’ and act as an environmental buffer zone, protecting the offices from the afternoon sun. The layout has produced a building with a great diversity of space and a wonderful quality of light. You arrive (served by a shuttle bus to reduce car use) at the upper level. Through the glazed entrance is the hall, which is roofed with a series of vaulted coffers, supported by a delicate steel framework, and floored with natural stone. The hall flows into the width of the ‘street’, revealing glimpses of the open-plan office spaces on the east side. It steps down, with pauses for a waterfall sculpture, to continue at a lower level, where it opens out to meeting rooms and a glass-walled restaurant.
Bennetts Associates has snuggled a substantial 10,000m2 headquarters for 580 staff into the landscape. The building follows the contours of its sloping site in three steps, and nowhere is it higher than two storeys. The plan is E-shaped, with the long spine of the E – the ‘street’ or main circulation route – running north-south and stepping down the slope with two staircases. The three horizontal legs are open-plan office spaces. These spaces are set on an
140
Detailed design of the exposed steel structure Studies were made to establish the size of a secondary beam capable of spanning 9m unrestrained, but with a top flange narrow enough to allow the precast coffer units to be lowered onto the bottom flange. As it would be exposed, a slender bottom flange was considered desirable. A simple 457 x 191mm universal beam offered the advantages of minimal fabrication and a slender bottom flange. The shape of the coffers was designed to fit this section within defined tolerances. It was considered safer to install the coffers by keeping them horizontal when lowering them, but the top flange interfered with this process. This problem was solved by lowering each coffer at 45° to the beam until it slotted between the flanges, and then rotating it onto the bottom flange. The top flanges were cut back at the perimeter, to allow the last two coffers to be lowered without rotation. CREDITS ARCHITECT Bennetts Associates STRUCTURAL AND SERVICES ENGINEER Buro Happold
SUPPLIERS steelwork and composite precast Wescol Glosford Structures; cladding, structural, solar shading Merlin Specialist Products
LANDSCAPE ARCHITECT Bernard Ede/Grant Associates CONSTRUCTION MANAGER Mace
The layout of the new operations centre is designed to accentuate light and space
The three wings of office space open off the east side of the ‘street’. Each has an identical layout, with its own business centre, kitchenette, meeting rooms and WCs. To optimise natural ventilation and daylight, the floors are only 15m wide. On the south elevations the facade is protected from solar gain by internal blinds and a two-storey external brise-soleil.
On its previous low-energy office structures, Bennetts used insitu and precast concrete coffer slabs to achieve the required thermal mass to cool the building. At Wessex Water the aim was to achieve similar thermal performance but using less concrete, to reduce embodied energy and consumption of resources. In conjunction with Buro Happold, the structural and building-service engineers, a system of precast exposed concrete coffer units was designed. The coffers and structural topping screed use 50 per cent less concrete than earlier designs, giving savings in the cost of foundations and allowing the use of a lightweight steel structure.
The structure of a typical office wing consists of a box-section spine beam of paired channels, supported on a central row of steel columns. Secondary beams run at right angles to the spine beam at 3m centres, and extend to the north and south facades to rest on delicate 120 x 120mm SHS columns. Precast concrete coffer slabs run between the secondary beams. Their solidity contrasts with the delicacy of the steel structure. Wessex Water has a commitment to sustainability and wanted its headquarters to be ‘an exemplar of environmentally sensitive architecture’. It commissioned Bennetts Associates because the firm has long recognised the link between economic prosperity, low-energy issues and the quality of the workplace. Bennetts Associates’ approach to sustainable architecture has been continually refined since it designed Powergen, its first low-energy headquarters building (AJ 2.3.95), followed by the John Menzies headquarters (AJ 30.11.95) and the BT headquarters (AJ 7.10.99) in Edinburgh.
141
A steel structure with precast coffers The structure of a typical 15m office wing consists of a box-section spine beam, formed by paired 430 x 100mm channels set toe-to-toe and supported on 180 x 180mm steel columns. These divide the office space into a 9m and a 6m bay. A series of secondary 457 x 191mm universal beams runs at right angles to the spine beam. These beams are set 3m apart and extend to the north and south facades, where they rest on delicate 120 x 120mm SHS columns. Rows of curved precast concrete coffers, each 750mm wide, span between the secondary beams and rest on the bottom flanges. The design is reminiscent of nineteenth-century ‘jackarch’ construction, where brick arches span between wrought-iron beams. An in-situ concrete topping ties the coffer units together, which appear as a series of smooth white-painted vaults with dove-grey steel strips between – these being the visible bottom flanges of the supporting beams. The vaults act as the path along which natural cross-ventilation from the highlevel windows is channelled (allowing heat exchange between the air and the concrete for overnight cooling). To maintain this path, there are no downstand edge beams – the spacing of perimeter columns allows the last coffer unit to act purely as a tie. This arrangement also enables the lightness of the structure to be expressed through the double-glazed facade, as the shape of each coffer frames one opening and one fixed window. The construction of the spine beam from toe-to-toe channels allows services to be carried in the void. The beam is perforated with 200 x 200mm holes with radiused corners, to continue the path of natural cross-ventilation. External solar shading to south and west elevations modulates and controls direct solar heat gains within the building.
142
143
STRETCHING A STEEL HYPERBOLE A new footbridge – a hyperbolic paraboloid of steel – celebrates Manchester’s restored commercial centre BY SUSAN DAWSON. PHOTOGRAPHS BY PETER COOK/VIEW
On 15 June 1996 a huge bomb exploded in the centre of Manchester, injuring 220 people, causing immense physical damage to buildings and disrupting the social and economic fabric of the city centre. One of the most vivid images of that day is the shattered footbridge which ran across Corporation Street, connecting two shopping centres. As part of the restoration programme a competition for a new footbridge was held; it was won by Hodder Associates and structural engineer Ove Arup & Partners with what may well be the first hyperbolic paraboloid bridge. Both sides of Corporation Street have now been restored; to the west is a new Marks & Spencer, to the east is the new Arndale Shopping Centre. The completion of the bridge between them is a symbol of Manchester’s recovery.
structure, rather than the walkway deck, is the dominant element. It is in the form of a cylindrical hyperbolic paraboloid - circular in section, but with a slender waist at the centre which increases in diameter towards each end. The curved shape enhances its transparency, giving uninterrupted views to those walking through the bridge while shoppers below in Corporation Street can see through the tructure. The symmetry of Hodder’s design resolves a number of problems in changes in levels and sizes of openings – a 1.2m drop between the Arndale opening and the lower, narrower Marks & Spencer opening – with public access and height regulations. As Hodder describes it, ‘the downward slope of the walkway at 2.8m regulation width was restricted by the lower circumference; raising the walkway increased the width but reduced available head height’. The problems were resolved by 3D modelling with CAD.
Corporation Street, a canyon-like route running north/south through the city, terminates in the civic space of Albert Square. Stephen Hodder conceived the bridge as a transparent link stretched 19m across the street, a delicate membrane of steel tubes glazed to enclose the walkway inside it. The steel
A curved hyperbolic paraboloid structure, though fiendish to draw in conventional plan and section, is relatively simple in principle. Imagine two rings with steel wires stretched between
144
castings bolted to it. Each casting consists of an inner and an outer disc with a ‘pickled’ finish which clamp six triangular glass corners together. The panels are faceted on all three edges to accommodate the three-dimensional curvature, and the joints are silicone-sealed. Above walkway deck level each laminated glass triangle comprises an outer pane of 12mm toughened glass, a PET (polyethylene terethalate) interlayer laminated with poured resin and an inner pane of 8mm heat-strenghened glass. The space below the walkway deck level acts as a plenum for the heating system and is fitted with lighting. Here the glass had additional performance requirements: it had to support maintenance workers walking on it, and in fact had to be a safe support even when broken. These glass triangles have a double PET interlayer laminated with three poured resin layers. The Ove Arup team studied ways to prevent a broken glass panel falling out of the casting due to the small bearing surface at the tip of the glass. The solution is ingenious; the interlayer is extended beyond the glass corner as a projecting tab (as shown in the connection at A, overleaf), riveted to an aluminium flat and wrapped round it to help build friction; it is then inserted into a slot in the outer plate of the casting to which the inner plate is then bolted. The two plates are sealed with site-applied silicone.
them; hold one ring still and just twist the other a half-turn to create the paraboloid. The bridge structure consists of 18 straight 114mm diameter steel tubes alternating with 18 straight 30mm diameter solid steel rods, fixed in a ring at each end-collar and canted to create the paraboloid. The rods are post-tensioned to add stiffness to the structure. At each end the tubes and rods are anchored to a 300 x 200mm RHS double collar, braced with a circular lattice truss, which is bolted to the structure of the adjacent building. The collars are clad with perforated steel sheet.
CREDITS ARCHITECT Hodder Associates: Stephen Hodder, Peter Williams, Stewart Jones
SERVICES ENGINEER Ove Arup & Partners: Andy Sedgwick
A series of circular hoops within the network of rods and tubes restrains the structure against buckling and supports universal beams on which the walkway deck rests. The hoops are formed of 75 x 75mm solid steel sections with machined grooves top and bottom, forming an H-shaped profile which provides connection points for the steel structure, the cladding and the delicate emergency-light fittings.
STRUCTURAL ENGINEER Ove Arup & Partners: Richard Houghton, Richard Summers, Andy Foster, Colin Jackson, Paolo Micucci, Stuart Bull;
SUPPLIERS steelwork Watson Steel, stainless-steel casting Dane Architectural Systems, glass panels Design-a-Glass, flooring and balu-strades J W Tayler
The bridge is clad with triangular laminated-glass panels which lie just inside the structure and are fixed to stainless-steel
Arup Facade Engineering: Stuart Clarke, Graham Dodd, Simon Webster
145
CONSTRUCTION MANAGER Bovis
A tubular steel footbridge clad with triangular glass panels The bridge, spanning 19m at first-floor level across Corporation Street, is a curved hyperbolic paraboloid. The structure and shape is created by a network of straight 114mm diameter steel tubes and 30mm diameter solid steel rods. A series of circular steel hoops runs within the network of rods and tubes to restrain the structure and support the bridge walkway deck. The hoops are formed of paired 75 x 75mm H-sections. The bridge is clad with triangular laminated glass panels which lie just inside the steel structure, faceted at all three edges to form the curve. They are fixed to a series of stainless steel castings which are bolted to the structure. Each casting consists of an inner and an outer disc which each have a ‘pickled’ finish; these clamp six triangular glass corners together. Detail A shows a casting bolted to the main CHS structure at the lower part of the bridge, below the deck walkway. Panels with double-laminated PET interlayers were used here for safety; they can be walked on for maintenance and the interlayers will maintain the integrity of the panels should the glass break. To prevent the panels dropping out of the clamp when damaged, the PET interlayer was extended with projecting tabs which were riveted to aluminium flats, wrapped around them and inserted into slots in the casting. Detail B shows an alternative version of the casting bolted to a 75 x 75mm H-profile steel hoop near the top of the bridge. It is in two sections to straddle the hoop; each half-casting accommodates the ends of three glass panels. SUSAN DAWSON
146
147
148
FOOTBRIDGE, EAST HAM, LONDON The bridge in Plashet Road, East Ham, is, indeed, anything but a shrinking presence in the relatively humdrum landscape of this corner of the borough of Newham. As you turn the corner from the Tube station it hits you in the face, a snaking white tube, carried on angular blue-painted steel legs, striding across the busy road.
The bridge is much more than a functional connection. Firstly, it reflected a concern for the site, its increasingly sinuous form allowing it to bypass a fine mature tree which the previous proposal would have felled. Secondly, what would have been a grim experience, like marching along an Underground corridor, was turned into a potentially pleasant one.
The practical brief was straightforward: Plashet School for Girls (1,300 pupils) inhabits two buildings on either side of Plashet Road. One is the former girls’ grammar school, completed in 1932 in depressingly institutional municipal Neo-Georgian style; the other a surprising 1960s work which housed the local secondary modern in a striking eight-storey tower, whose generosity of scale, vision and precise detailing puts its clumsy pre-war neighbour to shame.
Locals have reportedly compared the bridge to a wagon train. By cladding the structure in white Teflon fabric (with a life span of 25 years), held taut on a series of T-section galvanised steel hoops – ‘like a corset’, Russum comments – the architects have created a light and cheerful interior. The curving plan means that entering the pace is something of an adventure. At the centre, a steel-clad section contains two recessed bays, a break-out point, with windows looking out, a place where girls can pause briefly between lessons and swap tales from school.
For many years, the pedestrian crossing on Plashet Road was the vital link between the two sites for pupils and teachers. The obvious solution was a bridge, although the disposition of the two buildings – indeed, their lack of any obvious relationship, physical or stylistic – made it an awkward proposition. Unabashed, Newham’s borough engineers weighed in with proposals in 1998 for a heavyweight straight steel corridor slung across the road. The local planners were unimpressed and in 1999, threw out the plans as a potential blot on the borough’s landscape. Enter BPR, working with engineer Matthew Wells.
CREDITS ARCHITECT Birds Portchmouth Russum STRUCTURAL ENGINEER Techniker PHOTOGRAPH Nick Kane
149
A curved steel bridge with a fabric canopy The bridge takes the form of an S-shape, spanning 65m over a busy road to link two school buildings. The U-shaped composite structure – a pair of universal beams with a box girder deck welded to the bottom flanges – is supported on steel piers and covered with a lightweight PTFE glass fibre fabric stretched over a series of galvanised steel hoops. The steel hoops are assymetrical in shape and alternate along the length of the bridge to give the fabric rigidity. The 913 x 305mm universal beams are clamped to the two central piers by wedges driven through large hollow pins. In contrast the hoops, hoppers and gargoyles are of lightweight steel and are bolted to the beams. A 193.7mm diameter CHS welded to the top of each beam acts both as a handrail and a conduit for roof drainage, lighting cables and telecom cables. Rainwater collection is, in the architect’s words, ‘expressively detailed to provide delight on rainy days’. Rainwater collects in an upstanding fabric gutter on each side of the canopy and is directed through galvanised steel hoppers into a pipe set inside each CHS handrail. Steel gargoyles with chains at their ends spout the rainwater down the inside face of the piers to ground level gulleys. The top flange of each hoop has a pair of 16mm diameter steel bars welded along the outer edges. A 12mm diameter bar runs between them, with welded plates which are bolted to the flange. The edges of the PFTEcoated glass fibre fabric canopy, sleeved with 10mm diameter stainless steel cables, are stretched over the outer bars and fixed to the central bar with steel bands.
150
151
152
A BRIDGE FOR SHANKS’S PONY A Cor-Ten bridge across the River Nene has won plaudits for its design and sits comfortably in the landscape of the Fens BY SUSAN DAWSON. PHOTOGRAPHS BY RICHARD BRYANT Shanks Millennium Bridge is a dynamic curve of Cor-Ten steel that carries walkers, cyclists and ponies across the River Nene. It acts as the final link in a network of footpaths, cycleways and bridleways that surrounds the town of Peterborough in Cambridgeshire.
forms such as cable-stays or arches, which in any case were not appropriate in the flat landscape.’ The box girder is formed of steel plate which is stiffened internally with smaller plates at right angles where the stresses are highest. This is an efficient use of steel, optimising weight and cost. Cross diaphragms at 1.8m centres help to support the deck plates and prevent the deck distorting as it twists. The depth of the box girder below the deck has been minimised to reduce the length of the approaches to the bridge (every additional 100mm deck-depth meant a 2m longer ramp at a slope of 1:20). The box was therefore stiffened by a canted upstand on the outside of the curve that acts as the base of the balustrade.
Known as the Green Wheel project, the network runs into the city with a series of ‘spokes’ to connect urban cycleways with rural communities and even neighbouring towns and cities. (The Fenland countryside, flat to its horizons, is perfect cycling country). The project has a cultural element – cyclists, riders and walkers on the paths will pass sculptures, new engineering structures and interpretation boards which will tell them something of Peterborough’s 3,000-year history. The new bridge is a steel box-girder supported by four tapering Cor-Ten piers and flanked by concrete abutments. It crosses the river to the east of the city on the site of an ancient ford used by the Romans. The original Roman track approached the river at right angles to it, before crossing and turning sharply to run parallel to the bank. The curve of the bridge reflects the geometry of the ancient track, allowing the footpath, cycleway and bridleway to be seamlessly connected. Boats, mostly pleasure cruisers, pass under the bridge – it rises very gently at the centre to allow them to pass below.
The piers are tapered structural boxes fabricated from CorTen steel plate with the stiffening ribs exposed on the outside. The ribs accentuate the slender form of the piers and express the load distribution. The deck is fixed structurally to the large north abutment, so that there is no visible joint between the ground and the bridge deck at this point. All other supports at the piers and the south abutment have sliding bearings which allow the deck to move to accommodate variations in temperature. The bridge has already received accolades: it won the award for bridge design in the Royal Fine Art Commission/BSkyB Building of the Year 2001, and a commendation in the Structural Steel Design Awards 2001.
The bridge is a sturdy and rugged structure – suitable for the passage of horses and appropriate to the Fenland landscape. The use of Cor-Ten was inspired by the sculptures of Richard Serra. Cor-Ten is a partially recycled material that will weather gracefully in the rural environment and will respond to the curve by weathering in gradations of colour according to its orientation. It also reflected the council’s environmental policy – the metal does not need to be painted and maintenance will be minimal. The top surface of the box girder forms a 2.5m-wide bridleway. The pedestrian walkway/cycleway runs alongside, cantilevered on L-shaped Cor-Ten brackets. To give solidity, the bridleway is surfaced with concrete and protected by a canted timber screen that shields riders from wind and prevents their silhouettes from disturbing rare river birds. The bridleway and walkway are separated by a guardrail that acts as a ‘leaning seat’ for walkers to stop and look out at the riverscape.
CREDITS ENGINEER Whitby Bird & Partners
To accommodate the torsion caused by its curve, a closed boxgirder construction was used for the bridge deck. ‘Because the span – just over 30m – was relatively short’, explains Des Mairs, the engineer in charge, ‘we wanted to make the box girder carry the loads without additional help from structural
MAIN CONTRACTOR May Gurney STEELWORK CONTRACTOR Fairfield-Mabey
153
A curved Cor-Ten box girder with a bridleway and pedestrian walkway The curved Cor-Ten box girder which forms the bridleway deck is of 10mm and 12mm plate with internal stiffeners; it was fabricated in 12 sections which were butt-welded together on site. The girder has a tapering canted upstand along its outer edge to provide stiffness and to act as a base for the balustrade. The deck is surfaced with 90mm cast-in-situ concrete to give a durable and solid surface suitable for horses, while helping to dampen potential live movement within the structure. The 1.8m high balustrade is canted to give horses and their riders as much space as possible and to deter horses, if startled, from climbing the parapet. It consists of a series of Cor-Ten frames of 80 x 10mm flats; panels of horizontal 40 x 30mm cumaru hardwood slats, with 10mm gaps between them, are fixed to the frames. The 1.6m wide pedestrian walkway is raised above the bridleway deck to give the structure a lighter, less massive feel. It is supported by a series of paired L-shaped Cor-Ten arms which cantilever from the edge of the box girder; they run at 1800mm centres and are bolted to the sides of CorTen outriggers welded to the deck. A Cor-Ten box-frame bolted between each pair of arms supports a deck of 100 x 31mm cumaru boards with slipresistant inserts. The balustrade consists of rows of 45 x 8mm Cor-Ten flats welded 100mm apart to canted 32 x 10mm lugs, which are bolted to the raised L-shaped arms. The arms project above the balustrade to support a 60mm diameter polished stainless-steel handrail.
154
155
TYNE AND CHEER Wilkinson Eyre’s striking steel bridge is a worthy addition to the proud list of structures connecting Newcastle and Gateshead BY SUSAN DAWSON. PHOTOGRAPHS BY GRAEME PEACOCK
The Gateshead Millennium Bridge is a structure of breathtaking elegance; a pair of soaring parabolic arches which span 105 metres across the River Tyne.
bridge in the world. Further upstream is Robert Stephenson’s high-level railway bridge of 1845-49. The new bridge continues the North East’s proud tradition of engineering and is worthy of comparison with the other great engineering structures on the Tyne.
One arch is a delicate crossing deck for pedestrians and cyclists; in everyday mode it forms a gentle curve across the river with the other arch acting as its support above it. Seen from the quayside in this position, the slender profile of the main arch frames its neighbour, the famous Tyne Bridge.
The design was the winning entry in a 1997 competition promoted by Gateshead Metropolitan Borough Council, Newcastle’s less well-known neighbour across the Tyne on the south bank of the river.
To allow ships to pass below, both arches pivot from their springing points; as they do so the composition metamorphoses into a new parabolic shape, a ‘grand arch’ of great width and space, with an action reminiscent of a closed eye slowly opening.
The bridge links Newcastle’s newly developed quayside district with ambitious new developments on the Gateshead side of the river – including the new visual arts centre at the Baltic Flour Mills by architect Ellis Williams and the Northern Regional Music Centre by Foster and Partners.
This is a striking new example of an innovative steel structure, in a tradition that has been realised in many of the bridges that have linked Newcastle and Gateshead since Roman times. A medieval bridge crossed the river close by, at the point where the swing bridge, built in 1878, now stands; that opened to allow tall-masted ships to travel further upstream. When it was built in 1928 the Tyne Bridge, the best known of the six bridges which now exist on the river, was the longest single-span
The design for the Gateshead Millennium Bridge had to reconcile two seemingly opposed requirements: to retain a clear channel for shipping – a navigable width of 30m and a height of 25m – and to provide a low-level crossing for pedestrians and cyclists. The solution is deceptively simple: a pair of arches, one forming the crossing deck, the other supporting it. Both pivot
156
The stunningly elegant Gateshead Millennium Bridge was put in place by an enormous floating crane. The structure has a unique opening mechanism which tilts to resemble a blinking eye, and is a valuable addition to the varied bridgescape of the Tyne
at their springing points to allow ships to pass beneath them. The crossing deck and the support arch are connected by a series of cables. The motion is efficient and rational, yet full of drama and solves an old problem in a unique way. The springing points of the bridge are on two new concrete islands which have been created parallel to the quaysides on each side of the river and which house the operating gear. Glazed halls on each side offer dramatic views of town, bridge and riverscape. The supporting arch is a gently tapered box beam that forms a parabola with a radius of 46m at its crown; in section it is kite-shaped. The crossing deck is also a tapered box beam which forms a similar parabola. The main deck forms the footway; the cycleway runs outside it, supported on cantilevered steel brackets. The box beams were formed of steel plate that varied in thickness between 15 and 25mm, reinforced with transverse and longitudinal stiffeners. The beams were fabricated into 13 huge tapered segments at Watson Steel in Lostock, Bolton, Lancashire and assembled into a single element at Hadrian’s Yard, Wallsend, Newcastle. The footway and cycleway are separated by a screen – the ‘hedge’ that accommodates the change in level between the two and gives some wind protection to walkers and cyclists. Gaps in the hedge are fitted with steps and tubular stainless steel handrails. At intervals of about 30m along the deck, the hedge is reconfigured to form benches.
CREDITS ARCHITECT Wilkinson Eyre Architects: Keith Brownlie, Martin Knight
The Gateshead Millennium bridge opened to the public for the first time on 17 September with celebrations including a runpast by competitors in the Great North Run.
STRUCTURAL AND SERVICES ENGINEER Gifford & Partners: Peter Curran
The River Tyne has one of the world’s finest collections of innovative and historic bridges. This, its first bridge of the new millennium, is a worthy addition.
LIGHTING CONSULTANT Jonathan Speirs & Associates
157
MAIN CONTRACTOR Harbour & General STEELWORK SUBCONTRACTOR Watson Steel ARCHITECTURAL METALWORK D&B Darke
A box beam arch forming a footway and cycleway with a ‘hedge’ screen The hedge and benches are made of perforated stainless-steel sheet panels on a stainless-steel support frame so as to be maintenance-free and resistant to vandalism. The lozengeshaped perforations are grouped in rectangular bays; they make the steel panels appear lighter and at night they disseminate the lighting from cold cathode-ray tubes that are fixed to the steel support frame inside. The top edge of the hedge, 1,100mm high, is canted; it can be used for walkers to lean on and enjoy the view without becoming a shelf for litter. All along the cycleway the base of the hedge is fitted with louvred panels backed with 3mm opalescent acrylic lenses; the panels are hinged with stainless-steel piano hinges to give access to the lighting for maintenance. The opalescent lenses direct light onto the cycleway surface. The cycleway is supported by a series of I-shaped steel sections with 160mm flanges and 10mm thick tapered webs which cantilever from the box beam at 3m centres and taper at their ends to a steel nosing. The cycleway surface is an aluminium Neatdeck open-grid deck system of which the upper surface is sandblasted to give grip to tyres. The material is strong and resilient yet when the bridge deck is raised it appears partially transparent. It rests on 152 x 102mm steel T-shaped sections. The two balustrades are canted inwards to discourage people from climbing on them. They are formed of 12mm paired steel balusters bolted to the box beam on the pedestrian side and to the I-sections on the cycleway side. The balustrade is made of steel flats and is topped with a 48mm diameter stainless-steel handrail on projecting lugs.
158
159
STEPS TO BETTER STUDY A curved helical ramp provides London’s LSE with much improved library circulation BY SUSAN DAWSON. PHOTOGRAPHS BY PAUL RATIGAN/VIEW For more than 100 years, the London School of Economics & Political Sciences (LSE) has had its home in Aldwych, in the heart of London, but the great convenience of a citycentre location has been offset by the difficulties of finding room to grow. As part of plans to refurbish and expand, the library, housing what is considered to be the largest and most important archive of the social sciences in the world, has been renovated and enlarged.
triangles squashed around a lightwell. The resultant deep floor plates had perimeter glazing which provided poor light and ventilation. The listing of two facades precluded conventional improvement, and the structure was in a poor state of repair. ‘We wanted to inject a social form into the building,’ explains Robin Partington of Foster and Partners. ‘At the same time we had to find a way of getting daylight into the centre, especially the basement, and of solving the acoustic problem which meant separating students who are meeting and talking from those who are studying’.
Since 1973 the library has been in the Lionel Robbins building on the north side of the campus, the former headquarters and warehouse of WH Smith built in 1916. Comprising four storeys and a basement, it is an awkward shape – think of a pair of
The solution is dynamic. The lightwell at the centre has been transformed into a light and airy atrium, a cylinder of space open on all sides. At the top a dome with a canted glazed
160
rooflight pours daylight into the heart of the building, while eliminating glare and solar gain. A stepped helical ramp curves around the atrium from basement to third floor providing a social and circulation hub – a spiral of constant movement. A pair of glass lifts provides an alternative vertical route. The students walk from the ramp along passages defined by bookstacks which act as superb acoustic attenuators between the social hum of voices in the atrium and quiet study areas at the perimeter of the building, facing the windows. Part of the ground-floor slab has been removed to create a double-height study area in the basement. The fourth floor and a new fifth floor, which house a research centre, have a separate entrance and lift. The stepped form of the ramp has its own logic – a conventional sloped helical ramp from floor to floor would have made too steep a gradient. And it is not a continuous helix, rather a series of helical segments connected to projecting landings at every floor. The sinuous steel structure makes a dramatic statement in the atrium, curving from floor to floor, supported by a ring of six slender steel tubular columns and framed with a delicate curved steel balustrade. To accommodate torsional forces, the ramp is constructed of a steel rectangular box-like section with two canted steel plates welded to the underside to form a tapered, hull-like profile. A series of tubular props and splayed plates shaped like outstretched hands extends from the columns to give support to the underside of the ramp. Pinned steel brackets above the props and fixed to the outer string act as restraints.
The stepped helical ramp at the London School of Economics creates an open cylinder of space at its heart
CREDITS
The ramp was assembled in sections; both ends have welded connection plates that align the position of each tapered baluster. The sections were lifted in by a tower crane through the rooflight opening, bolted together and strip-welded through specially designed holes in the strings. The ramp and the atrium form part of the natural-ventilation strategy. Air is drawn perimeter windows, now double-glazed and building management system and extracted the dome.
ARCHITECT Foster and Partners: Norman Foster, Ken Shuttleworth, Robin Partington, Andy Purvis, Lulie Fisher, STRUCTURAL ENGINEER Adams Kara Taylor
lighting and the in through the operated by the through vents in
SERVICES ENGINEER Oscar Faber QUANTITY SURVEYOR Davis Langdon & Everest
161
CONSTRUCTION MANAGER Bovis SUPPLIERS ramp AFS Steel Fabricators, dome Cowley Structural Timberwork
A stepped helical ramp Originally a book warehouse designed in 1916, the new library is five storeys high plus a basement. The deep plan was originally lit by a large lightwell which has been transformed into the main vertical circulation hub; an 18 x 18m tear-shaped void has been created through which rise a pair of glass lifts and a stepped helical ramp. The space is covered with a dome 14m in diameter and 8m high, with a canted inset rooflight. The ramp takes the form of a broken helix with each segment bolted at its ends to the floor slabs. It is supported by six 323.9mm diameter CHS columns that run at the perimeter of the void.
fifth floor
In section the ramp is a rectangular steel box with sides of 275mm deep curved string plates and a hull-like base of two canted steel plates, designed to give a less bulky appearance. Structurally, each section of the ramp is a torsional box restrained at the landing positions. The box sections were fabricated from steel plate cut to radial curves, twisted to create the incline, and then welded together.
fourth floor
third floor
second floor
The ramp is supported by arms and restrained by brackets, welded one above the other, to each column. Each arm takes the form of a plate, in the shape of an outstretched hand, supported by a 76.1mm diameter CHS prop pinned at the end with forked connectors welded to the column.
first floor
ground floor
Services run through voids in the box. The ramp treads are topped with screed to reduce vibration. The balustrade is of 20mm solid rods welded to tapered balusters of 20mm thick steel flats and topped by a 48mm diameter CHS handrail. The balusters align with the ramp steps and the connection plates.
basement
KEY PERSPECTIVE DIAGRAM OF RAMP
162
323.9mm dia. x 16mm 48mm dia. x 4mm SS CHS handrail baluster of 20mm plate tapering from 80mm at base to 50mm at top
welded brackets pinned with forked connector 275 x 10mm string plate line of step 75mm structural screed 10mm thick thread plate
1100mm
handrail with welded spigot bolted to baluster
20mm dia. solid bar balustrade
275 x 10mm string plate 210 x 10mm secondary string plate
75 x 6mm curved plate 10mm soffit plate 100mm dia.hole as service duct M24 bolts in 26mm dia. holes
75mm structural screed
80mm splice plate CHS pinned to splice plate with forked connector
76.1mm dia. CHS
45mm thick profiled plate welded to splice plate 76.1mm dia. x 5mm CHS pinned to profiled plate with forked connector
850mm
CHS pinned to column brackets with forked connector
DETAIL SECTION A-A
DETAIL SECTION B-B
323.9mm dia. CHS column 323.9mm dia. CHS column ramp bolted to original floor structure
275mm string plate tapered baluster
275mm string plate structural screed
210mm secondary string plate
ramp sections bolted together 275mm string plate
210mm secondary string plate tapered baluster 48mm dia. x 4mm SS CHS handrail 20mm dia. solid bar balustrade 75mm screed as deck 10mm soffit plate service duct
tapered baluster
275mm string plate
76mm dia. CHS as ramp support bracket pinned with forked connector
DETAIL PLAN OF RAMP
AXONOMETRIC SKETCH
163
164
OFFICE, DEVONSHIRE SQUARE, CITY OF LONDON Houndsditch, even today, lacks cachet – hence the rather curious address (2.5 Devonshire Square) for the new office development that occupies a site between the busy street and the still-tranquil square tucked away behind, that retains a few Georgian houses. The natural point of entry to the building is on the north-west corner of the site. From here, it is a short hop to Liverpool Street station, via a pedestrian alley used by thousands of commuters daily. There is a steep fall in pavement level from north to south, hence the decision to locate the main reception area at first-floor level – areas of the ground floor are occupied by the service yard and by a bar accessed from Houndsditch. The building includes nine office floors and three basement levels, which are used for parking, services and storage. The office floors are conventional in terms of their scale and appearance, providing the requisite open-plan flexible spaces – the perforated beams which gave the interiors a strongly industrial character at pre fit-out stage have now been covered by ceiling panels that conceal a plethora of services running through them. A stepped series of atria, rising above level 3, provides increased natural light, but these remain standard City floors, where the comparison with a Victorian factory is not inappropriate. Service cores are located at three points, on the western perimeter, where the main bank of lifts is adjacent to the reception area, and on the south-east and north-east corners of the building. The strategy has obvious advantages in terms of delivering unencumbered office floors but also reflects Bennetts’ desire to animate the edges of the floors, to allow the building’s occupants views out and visual contact with their surroundings. The fine materials and careful detailing seen in the staircases (framed by stone-clad ‘bookends’), with their granite treads and landings, reflects this idea, as well as the overall air of quality in the completed scheme. Seen from the surrounding streets, for example, up St Mary Axe, the building is a sober but far from oppressive presence. On Devonshire Square it provides a welcome contrast to the bulk of an adjacent sub-station, clad in crude Post Modern brick which does nothing to reduce its impact.
CREDITS ARCHITECT Bennetts Associates STRUCTURAL ENGINEER Whitby Bird & Partners/Waterman Partnership
Though the building is air-conditioned, there is an attempt to minimise energy use by means of sensible low technology – in-cavity blinds, for example, and sunshade louvres on the southern elevation.
STEELWORK CONTRACTOR Wescol PHOTOGRAPH Peter Cook/VIEW
165
A facade with an exposed structural steel frame In sympathy with nearby Victorian warehouses, the office building expresses its structure with a rugged, load-bearing steel and glass facade. The exposed steel frame, which is constructed to half National Structural Steelwork Specification toler-ances, supports six stories of composite concrete slabs on cellular steel beams. The steel frame consists of a series of 356 x 406mm universal columns and 686 x 254mm universal beams. Fire protection is achieved by a combination of strategies; the members are oversized in relation to their loading, and they are backed with a series of 60-minute integrity and insulation fireresistant composite panels, which extend beyond them at the sides, shielding them from heat. The gap between structure and panel is drained and ventilated to avoid condensation. Thermal bridging is minimised by a thermal break between internal and external structural elements. Shear bolts with ferrule spacers separate column and beam connection plates, and the gap formed by the spacers is filled with rigid insulation. The cladding is a grid and panel system supported by thermally isolated brackets. The polyester powder coated aluminium frames are structural silicone-glazed. The double-glazed units have a toughened outer pane and a laminated inner pane with a Low-E coating to the outer face. The lower glazed panels open to vent the office floor for smoke evacuation. On the south facade, the glazing is screened with solar shades which are pivot-hung to allow the glazing to be cleaned. The steel frame is painted to a high specification using a zinc-rich primer and M10 barrier coat with a decorative finish to match the cladding.
166
167
RIGHT ROYAL RESTORATION The Royal Court Theatre in London’s Sloane Square was kept alive by the architectural equivalent of open-heart surgery BY SUSAN DAWSON. PHOTOGRAPHS BY ANDY CHOPPING On 8 May 1956 the Royal Court Theatre in Sloane Square, London, staged John Osborne’s Look Back in Anger. The play was to change British drama forever, yet its showcase was an utterly conventional Victorian theatre in what was – and still is – a swanky shopping area. But all that has now changed; the theatre, lately refurbished and extended by the architect Haworth -Tompkins, now reflects the radical and contemporary approach of the stage company under its roof.
entrance gives tantalising glimpses of a curved vermilion wall beyond. This, the original rear wall of the auditorium that rises through the building and is visible on all levels, was painted by artist Antoni Malinowski. The heart of the scheme is the original auditorium, which has been stripped down to its essentials. The frilly plasterwork – a post-war pastiche – has been removed to reveal the tough, bolted iron plates of the original structure which supports the upper circle.
In 1995, the theatre faced closure. The structure was unsound, and front-of-house and backstage conditions were damp and overcrowded. Haworth Tompkins was asked to double the floor area of the building, adding new facilities while respecting the theatre’s history.
The original theatre needed refurbishment, but it was also seriously short of ancillary space: backstage facilities were cramped, there was no access for scenery and the stage itself lacked any modern technological systems. But the site was restricted, flanked by Sloane Square underground station and the residue of the River Cranbourne, which is culverted in a sewage pipe at basement level. The only two directions to expand were along the side of the theatre and beneath Sloane Square.
‘It was a fantastically delicate operation’ explains Steve Tompkins. ‘Instead of a complete reconstruction behind the original facade, which was the easy option, we undertook a kind of open-heart surgery’.
New backstage and administration facilities are housed in a fourstorey extension at the side of the theatre, while a cavernous
The Victorian Sloane Square facade has been retained – it was listed – but is now more open and ‘permeable’, and the inviting
168
The four-storey extension features sliding panels to allow for shade and privacy, while the theatre’s listed facade on Sloane Square remains intact
new bar/restaurant was created by burrowing beneath the road and the square. Matching the stripped-down aesthetic of the refurbished interior and auditorium, the new spaces are clad with simple materials that will weather and age naturally. The bar is fitted with reclaimed timber and dark leather seating, and its walls are of polished cast in situ concrete. The four-storey extension houses dressing rooms and offices on the upper floors. These are clad with an open rainscreen of flat weathering steel (Cor-Ten) panels, -perforated to allow actors and office workers to control any problems of overlooking, sound attenuation, solar glare or gain (the side extension runs alongside an alley which is only 3m wide in some places). Sliding doors and hinged screens of Cor-Ten are fixed in front of the oak-framed windows to give further control of view. They are operated by a pulley-and-wheel system. The ground floor is clad with solid profiled weathering steel Cor-Ten panels, painted to take additional wear and tear. ‘Cor-Ten is a very attractive material,’ says Harry Montrésor, who acted as cladding consultant to the project, ‘but the detailing is critical. All joints and profiles must be designed so that rainwater can run off – there must be no retention of water. And all contact with other materials – fixings, electrical conduit, lighting – must be avoided.’ In practice this meant that the Cor-Ten facade panels and their stainless-steel fixings had to be clearly separated (see Working Details, overleaf) with nylon spacers, washers, bushes and sleeves. An additional problem is the rust-coloured rainwater run-off from the panels which occurs during the first few years after installation. A wide stainless steel gutter has been positioned at first floor level to collect this. Run-off from the painted Cor-Ten panels below is directed into a stainlesssteel box gutter inset into the floor slab and covered with a perforated stainless-steel plate.
CREDITS ARCHITECT Haworth Tompkins STRUCTURAL ENGINEER Price & Myers SERVICES ENGINEER Max Fordham
Cor-Ten was chosen as a cladding material for its inherent richness and durability. Since it has been installed the cladding has weathered from light orange to a subtle dark purple-brown. Like the theatre, it should grow old gracefully.
CLADDING CONSULTANT Montrésor Partnership FABRICATOR Custom Metal Fabrications
169
Rainscreen of weathering steel panels The new four-storey extension to the theatre, containing offices and dressing rooms, is clad with a rainscreen of solid and perforated weathering steel (CorTen) panels. They are sized to reflect storey heights and spandrel areas and are fixed, except in front of window openings where they are formed into sliding and hinged screens. The 150mm thick cast insitu concrete structure was waterproofed with a liquid bituminous membrane and fitted with a series of stainless steel support and restraint brackets, which were lined and levelled. Mineral-fibre insulation bats, 75mm thick, were then pinned to the structure; they are lined on the outside with black glass-fibre tissue to assist in shedding water and to give a neutral background to the perforated screens. The oak windows were sealed to the waterproofing and a second series of stainless steel support and restraint brackets was fixed to the first, with a thermal resistor between them. Each 5mm thick perforated Cor-Ten panel is fixed through the unperforated perimeter zone to a bracket with a stainless steel buttonhead socket screw. To ensure that the stainless steel and Cor-Ten are totally separate, the screw is sleeved in nylon spacers and washers; the spacer is thick enough to avoid rainwater bridging from Cor-Ten to stainless steel and vice versa by surface tension. Stainless steel captive nut clips were fitted over the top of each bracket to enable the button-head socket screws to be tightened from the outside. Cor-Ten earthing connector tabs are used to achieve electrical continuity between panels. At first-floor spandrel level, a stainless steel gutter is set below the rainscreen to collect the run-off. The ground floor is clad with red-painted solid profiled panels of Cor-Ten, fixed in the same way as the upper panels.
170
171
FACTORY FOR ROLLS-ROYCE, GOODWOOD, WEST SUSSEX
The Rolls-Royce site on the Goodwood Estate, a few miles from Chichester and close to the famous racecourse and car-racing circuit was earmarked for gravel extraction, a process that could have extended over some decades. It is now occupied by a building that is a model of discretion and sensitivity in terms of its landscape impact and a rational development of the new workplace model, derived from the USA and pioneered by Grimshaw, Rogers and Foster in the 1960s and 70s.
BMW resolved to re-create the Rolls-Royce car from scratch and its new Phantom model (basic cost: £250,000, 500 bhp engine, 0-60 mph in 5.7 seconds) has been acclaimed as a reinvention of the marque. The new factory was commissioned specifically to build the new Rolls. The choice of the Goodwood site reflects the demands of a luxury market. Many cars are bespoke models with a future owner visiting the factory to select everything from the body colour to the details of the hand-sewn leather upholstery. Each car takes 260 hours to build, starting with body shells made in Germany - the plant turns out just five cars a day. The availability of craft skills in the area (there is an old-established boatbuilding tradition, for example) was another factor that attracted the company to Sussex.
The Rolls-Royce car celebrates its 100th birthday this year. It was on 1st April 1904 that engineer Henry Royce’s first model emerged from the factory in Cooke Street, Manchester. Shortly afterwards, Royce met Charles Rolls for dinner in the city’s Midland Hotel and a famous partnership was launched. Car production moved to Derby, then, in 1947, to Crewe – Rolls-Royce meanwhile became a major manufacturer of aero engines. After a period in which the marque seemed to lose its way in design terms, the right to produce Rolls-Royce cars was sold to BMW in 1998.
The basic function of the 55,000m2 factory is to build cars, though sales, design and managerial facilities are also provided. There were two other major imperatives. Firstly, though there was general local support for the project, with its promise of hundreds of jobs, yet the site is in an Area of Outstanding
172
Natural Beauty. The impact of the building on the landscape had to be benign. Second, the BMW group is a progressive employer that places stress on good amenities for staff and on quality design generally.
While the heavy-duty, steel-framed structure of the complex on a 20m square grid is essentially straightforward, considerable attention was given to the fine-tuning of the facades to ensure optimum environmental conditions (and also, of course, to temper the impact of the plant on its setting). Beyond the primary steel and glazed cladding, a second cladding layer forms an environmental screen, tailored to the specific needs of external elevations (and those facing internal courts where simple profiled aluminium is used). The main elevation of the production area, facing west, and that of the reception area and offices in the pavilion, facing south, feature bespoke sunscreening systems. Elsewhere, the secondary cladding is far more economical; panels of western red cedar in removable frames that allow for the alteration and possible extension of the buildings in the future. The timber is already starting to weather to a satisfying silver grey hue.
‘The height of the building was a key issue’, says project architect Paul McGill. ‘A clear 8m inside was a basic client requirement, while locals wanted the factory kept low – given the nature of the landscape, views of the building from above were also important.’ The excavation of the valuable gravel from the site prior to construction work starting late in 2000 allowed the complex to be partly sunk into the ground. Material left over from excavation was used to create earth mounds. The planting of 400,000 trees and shrubs will also help to blur its impact. The 50,000m2 sedum-planted roof sets a new record for the UK. The Rolls-Royce factory and headquarters is a carefully composed family of buildings. A more formal notion of architecture has displaced the machine aesthetic of the recent past. The production area is contained within the largest of these buildings, which extends on a north/south axis, with staff parking areas at the south end and the paint shop at the north end – bodies are sprayed here as the first step in the production process. Visitors to the site, including potential buyers, arrive from the west in a large paved courtyard which is flanked, to the south, by the final production area and, to the north, by the two-storey pavilion that houses VIP lounges, showroom, boardroom and offices. The restaurant, used by all staff, is contained in a pod, raised on piloti, that forms a link between the pavilion and the production building.
CREDITS ARCHITECT Nicholas Grimshaw and Partners STRUCTURAL ENGINEER WSP South Cameron Taylor Bedford STEELWORK steelwork – Stahlbau Plauen architectural metalwork McGraths, Littlehampton Welding louvres – MBC Precision Castings steelwork fixings -Hilti
The significance of the pavilion is emphasized by the use of areas of stone cladding. Any notion of management sitting in luxury with the workers consigned to a basic shed is immediately dispelled, however, by views into the production areas from the courtyards. Production staff enjoy ample natural light and views out to the courtyard and the country beyond.
PHOTOGRAPHS Edmund Sumner/View
173
A glazed facade with cedar louvres The west facade of the manufacturing building faces the main entrance courtyard and is single-glazed to give views of the main production line. Workers are protected from glare by a series of vertical louvres which are motorised to respond to sunlight. A fixed brise-soleil at the eaves cuts out glare from high-level summer sunlight. The 10m high facade is supported by elliptical steel columns which run behind it at 10m centres and are connected to the main roof structure with torsion brackets. The louvres are supported at top and bottom by a series of splayed, cast aluminium outriggers. A pair of steel channels, set one behind the other and connected by insulated steel studs, supports the outriggers. The outer channel runs along the facade and the outriggers are bolted to it with M12 bolts. The inner channel is bolted to the columns and is braced with a profiled 15mm steel flat and 75mm dia CHS stiffener between each column. Each louvre is formed of two 25mm western red cedar boards framed with leading and trailing edges of aluminium and set between a vertical aluminium centre post. The centre posts of the louvres rest on an aluminium extrusion which houses the motors. The aluminium trailing edge is perforated to allow a small amount of light through, avoiding sharp contrasts when the louvres are almost closed. Below the louvres the facade is of 12mm toughened glass. Behind them it is 10mm annealed glass, set in aluminium frames and supported by 120 x 60 x 6mm RHS vierendeel frames. The clerestory glazing is of 12mm annealed glass. Single glazing could be used since the frames are fully supported and the louvres reduce thermal shock. SUSAN DAWSON
174
175
THEATRE, PLYMOUTH Ian Ritchie Architects’ production centre for Plymouth’s Theatre Royal is a finely crafted, unashamedly industrial building, that combines traditional function and beauty BY SUSAN DAWSON. PHOTOGRAPHS BY PAUL RATIGAN/VIEW
176
Peter Moro’s replacement for the old Theatre Royal, one of many depressing losses inflicted on Plymouth by wartime bombing, was completed in 1982. Moro’s building has worn well and is popular with audiences and performers. As the only major theatre west of Bristol, it hosts visits by major national companies including Birmingham Royal Ballet and Welsh National Opera, and is a leading producing house, generating shows that transfer to London and beyond.
where a director and cast can work out ideas – the largest allows sets for the main Theatre Royal stage to be assembled for rehearsals. The remainder of the building - the central circulation/office spine and the big assembly studio/paint-shop block to the west - is clad in a mixture of zinc sheet and zinc-galvanised steel, Glazing is either of a standard double-glazed variety or insulated U-profile sheets of low iron glass which provide insulation and solar protection while giving the building a satisfying glow by night and calm, well-lit interiors by day.
Moro’s theatre stands on a tightly constrained site, surrounded by roads. As a consequence, Ian Ritchie Architects’ new production centre for the Theatre Royal, christened TR2, is located a mile or so from the city centre in a recently inaugurated business park. In some respects this is unfortunate, since TR2 is Plymouth’s most distinguished new building in decades yet contributes little to the public face of a city eager to transform itself.
The client’s desire to create a building that was ‘architecture, not an industrial box’ generated a passionate response from the architects. Externally, TR2 is rooted to its site by a landscape of broken rocks, an industrial Zen garden, plus the gabion walling that Ritchie pioneered but which has a particular resonance on this site. The generosity of the building is apparent even from the double-height reception area extending right across the building and well-equipped to handle a busload of school students. Beyond this is the communal route which extends along the entire eastern elevation, fully-glazed on both sides with views to the river on the one side and into workshop spaces on the other. A parallel route is provided along the centre line of the building at mezzanine level, with views into working spaces. Together these routes, forming the main circulation artery, can be used as a useful introduction for visitors to the activities going on within TR2, while creating minimum disturbance to staff. The central circulation zone reads as a great slit through the building from north to south, punctuated by staircases and connecting bridges.
Unfortunately the brief for the Moro building was cut even as the scheme was developed. The opportunity to house rehearsal spaces and workshops for scenery, props and costumes on the main site was lost for ever. For 20 years the Theatre Royal has devolved these activities to church halls, warehouses and other ad hoc spaces around the city, an arrangement that has been uneconomic and inconvenient. TR2 is, however, much more than a technical support facility. Like other publicly-funded arts organisations, the Theatre Royal is expected to be increasingly accessible, to attract wider audiences and greatly develop its educational role. Also, the Theatre Royal has a number of community based performance groups, some 500 strong, which use the new building – its education department has one of the most ambitious programmes of any theatre in the provinces and it finally has the spaces it needs to expand its work.
TR2 is a big-boned, tough building in an unashamedly industrial mould. The steel-framed structure, with floors of pre-cast concrete slabs (useful for acoustic insulation and providing the thermal mass which is part of a low energy services strategy), is designed with flexibility in mind. The architects envisage that spaces will be freely reconfigured as the practical needs of the Theatre Royal develop. ‘The building may become unrecognisable, but the architectural intent will endure’, says Ian Ritchie.
The TR2 site at Neptune Park is close to the estuary of the river Plym, east of the city centre, on land owned by the Duchy of Cornwall. The decaying terrain of boat shops and storage sheds has been transformed by a process of reclamation, with rock quarried nearby used to create infill land from mud flats. Ian Ritchie Architects was appointed in June 1997. With funding from the Lottery, EU and elsewhere, the £5.8 million building was constructed in just under two years between January 2001-November 2002. The client brief was for four rehearsal spaces, workshops for assembly and of painting sets, for props and costumes plus a dedicated education base with offices and classrooms. The building was designed for controlled public access – workshop areas, for example, had to be secure, though they are included in regular guided tours for educational groups. Each element in the building has its own functional and engineering requirements and this is reflected in the plan – TR2 can be seen as a ‘family’ of buildings. The rehearsal spaces are contained within perimeter ‘pods’ – three have been built and the fourth waits for funding - separated from the central spine of the building. One advantage of this arrangement is that the spaces are acoustically self-contained. The pods are wrapped in a woven wire cladding of phosphor bronze, designed to weather and evolve in the sea air. The cladding is fixed with stainless steel strips on cedar battens, infilled with geo-textile matting. ‘This is the first soft metal-clad building’, says Ian Ritchie. ‘It’s tactile in every way’. Ritchie sees the pods as beached structures, driftwood – the stains left by seagulls and the wind and rain are part of their character. They are private spaces set apart,
CREDITS ARCHITECT Ian Ritchie Architects STRUCTURAL AND SERVICES ENGINEER Arup STEELWORK steelwork - Bison Structures mesh panels - Lockerwire Weavers zinc cladding - Boss Metals architectural metalwork - Taunton Fabrications PHOTOGRAPHS Jocelyne Van den Bossche, Ian Ritchie, Toby Smith
177
A rainscreeen cladding of phosphor bronze alloy mesh To isolate them from noisy workshops, the theatre’s three rehearsal spaces are designed as ‘pods’, linked by acoustic lobbies. Their steel portalframe structures and precast concrete plank roofs are covered with an insulated single-layer roof membrane. The walls are ‘wrapped’ with phosphor bronze woven wire mesh cladding with galvanised steel-framed glazing below. Parapets are clad with natural zinc to match other elevations. The bronze mesh is a ‘soft and tactile’ pillowed rainscreen to an insulated and weatherproof metal stud wall. To accommodate thermal expansion and to avoid ‘flapping’ and work hardening, the mesh is tensioned against three layers of compressive geotextile mat, giving it form. The mesh is a ‘Dutch’ weave with a 0.4mm diameter warp wire set at 10 to the inch and a 0.3mm diameter weft wire set at 95 to the inch; this provides the required longevity, visual opacity and limits stretch within the mesh. Phosphor bronze is a 95/5 copper alloy needing no surface treatment and weathering to form a patina of black, green and gold. The mesh is fixed in 1m-wide panels, up to 7.5m high and separated by a 10mm joint. The vertical edges of each panel are wrapped around stainless steel strips and perforated with eyelets, providing sufficient pull-out strength for the stainless steel security fixings. All edges are fixed through a cedar batten covered with bronze mesh and sealed to the stud wall by a butyl strip. All materials used in the cladding system are non-ferrous to avoid bi-metallic corrosion with the bronze. A stainless steel drip sheds the bronze run-off water away from the glazing below. SUSAN DAWSON
178
179
SPACE ODYSSEY Peter Foggo and David Thomas’ pioneering 1960s Space House has been refurbished, gently reorganised and, dare one say it, improved by Lee/Fitzgerald Architects BY BARRIE EVANS. PHOTOGRAPHS BY RICHARD BRYANT/ARCAID
Space House in East Grinstead was a speculative prototype, built to a budget, which became one of the icons of its time in 1965. It well deserved listing but luckily this was one that got away, leaving its present owners free to update and ‘improve’ on the design of Peter Foggo and David Thomas. Space House is a raised, single-storey pavilion in three zones - living, serving and sleeping - with four rectangular steel trusses braced with internal and external tie rods (prefiguring Reliance Controls). It is glazed wall-to-ceiling at the front and back. Inside, the flow of light was muted by the use of timber boarding for floors, walls and ceilings, finishes that have darkened over the years. Along with other changes, notably black-coated aluminium external doors in place of glazed cedar, Space House had grown darker and heavier as it aged.
Lee/Fitzgerald Architects won the job of bringing it into the new century through the recommendation of Foggo Associates, with which the practice had worked in refurbishing the earlier Foggo/Thomas timber-framed deck house, Sorrel House, in Bosham Hoe, West Sussex. The owners, Andrew Spurgeon and Ann Kelly, were keen to retain the spirit of the place and lived on site, managing the project themselves. The result has been to increase the feeling of space and the experience of natural light inside and out. Painting the originally black frame in white makes a startling difference, lifting the building ‘closer to the cloud. The extensive single glazing is retained in its original neoprene gaskets, which are
180
CREDITS ARCHITECT Lee/Fitzgerald Architects
in excellent condition. External cedar cladding, which lurked under a heavy build-up of varnish, has been sanded and re-coated. Yet more light has been let into the house by replacing some cedar walls and yellowed British Colombian pine ceilings with white painted plaster. The architect favoured more cedar reinstatement than the client, who preferred the lightness of painted plaster; only the truss walls are now timbered, which does have architectural clarity.
back of thehouse to improve the connection between the (refitted) kitchen and the garden. Wall insulation has not been upgraded and the single glazing remains, hence the powerful warm-air heating system. Natural ventilation comes from sets of vertical louvres while doubleskin pleated blinds help reduce solar gain. The two somewhat cramped core bathrooms have been reorganised to produce a large bathroom and a shower room.
The deep through living-dining room now feels more connected to the outdoors and the removal of corridor doors round the service core increases the flow of space and light. Full-height central glazed double doors have been added to the rear set
Outside a large oval pool has been removed, allowing the house to breathe a bit more though there will always be the sense that it needed a larger site.
181
A new pair of double-glazed doors for a 1960s house The detached single-storey house, designed in 1964 by Peter Foggo and David Thomas, takes the form of a flatroofed pavilion raised off the ground on a steel-framed structure. It is clad with large panes of glass or panels of vertical cedar boards. Foggo and Thomas produced a series of original and economical details, drawn to full-size scale, some of which are illustrated here (far right). The exposed steel structure frames and supports the glass and cedar panels. The fascia is a 203 x 76mm RSC with an upstand welded to the top flange and a downstand welded to the bottom. The upstand trims the roof covering, the flanges face inward to house the joists and the downstand is connected to the (original 1/4” plate) glass panes by a neoprene ‘zipper’ gasket. At the sill a similar upstand, welded to a 203 x 203mm UB, is zipper-gasketed to the bottom edge of the glass. The glazing bays are subdivided by Tshaped steel mullions, of which the ends are directly zipper-gasketed to the glass. Originally the kitchen opened out on to the rear terrace through two doors set at each side of a glazed wall. They have been replaced by a pair of centrally placed outward-opening glazed doors. Following the original front door detail, the new doors have a top frame with a rebated ledge, which fits directly behind the fascia downstand and is screwed to it. The upstand at floor level is slotted into a groove in the door sill to act as a water bar. Each jamb is supported by a 65 x 65mm RSA; the other legs of the RSAs are zipper-gasketed to the adjacent glass panes. SUSAN DAWSON
182
183
184
References
Corus Tubes: SHS Welding, CT15 Design of SHS Welded Joints, CT16
Industry Publications
CIDECT Design Guides: Circular hollow section (CHS) joints under predominantly static loading, No. 1 Rectangular hollow sections (RHS) joints under predominantly static loading, No. 3 Fabrication, assembly and erection of hollow section structures, No. 7
Websites: BCSA www.steelconstruction.org SCI www.steelbiz.org Corus Construction Centre www.corusconstruction.com Construction and Industrial www.corusconstructionandindustrial.com Tubes www.corustubes.com
Bridon steel ropes for structural applications www.bridonltd.com Macalls bar and cable for tension structures www.macalloy.com
Green Books: Joints in Steel Construction: Moment Connections, P-207 Joints in Steel Construction: Simple Connections, P-212 Joints in Steel Construction: Composite Connections, P-213
Resotec damping system www.rlsd.co.uk
Standards Publications BS 970 Specification for wrought steels for mechanical and allied engineering purposes
[Note: P-212 superseded two previous publications Joints in Simple Construction: Volume 1: Design Methods, P-205; and Volume 2: Practical Applications, P-206]
BS 3692 ISO metric precision hexagon bolts, screws and nuts (superseded)
Black Book: National Structural Steelwork Specification for Building Construction, 4th Edition, P-203
BS 4190 ISO metric black hexagon bolts, screws and nuts Specification
Grey Book: Commentary on the National Structural Steelwork Specification for Building Construction, 4th Edition, P-209
BS 4395 Specification for high strength friction grip bolts and associated nuts and washers for structural engineering
Brown Book: Structural Fasteners and their Application, Publication 4/78
BS 4604 Specification for the use of high strength friction grip bolts in structural steelwork. Metric series
Manual on Connections for Beam and Column Construction, Publication 9/82
BS 5400-3 Steel, concrete and composite bridges. Code of practice for design of steel bridges
Guide to the Erection of Steel Bridges, Publication 38/05
BS 5950 Structural use of steelwork in building BS 5950-1 Code of practice for design. Rolled and welded sections BS 5950-3.1 Design in composite construction. Code of practice for design of simple and continuous composite beams
Design Guide on the Vibration of Floors, P-076 Castings in Construction, P-172 Design for Construction, P-178 Design of Semi-continuous Braced Frames, P-183 Guidance Notes on Best Practice in Steel Bridge Construction, P-185
BS 6472 Guide to evaluation of human exposure to vibration in buildings (1 Hz to 80 Hz)
Wind-moment Design of Low Rise Frames, P-263
BS 6841 Guide to measurement and evaluation of human exposure to whole-body mechanical vibration and repeated shock
Wind-moment Design of Unbraced Composite Frames, P-264 Tension Control Bolts, Grade S10T, in Friction Grip Connections, P-324
BS 7644 Direct tension indicators
Design of Multi-Storey Braced Frames, P-334
BS EN 1090-2 Execution of steel structures and aluminium structures. Technical requirements for the execution of steel structures (provisional title)
Acoustic Detailing for Multi-Storey Residential Buildings, P-336 Interfaces: Curtain wall connections to steel frames, P-101 Connections between steel and other materials, P-102 Electric lift installations in steel framed buildings, P-103 Design of steel framed buildings for service integration, P-166 Steel supported glazing systems, P-193 Recommended Application of BS 6399-2: ED001
185
Government Publications
BS EN 1991 Eurocode 1: Actions on structures BS EN 1991-1-2 General actions – Actions on structures exposed to fire BS EN 1991-1-7 General actions – Accidental actions from impact and explosions
Building Regulations: Approved Document E, Resistance to the passage of sound, HMSO Building Regulations: Approved Document L2, Conservation of fuel and power in buildings other than dwellings, HMSO
BS EN 1993 Eurocode 3: Design of steel structures BS EN 1993-1-1 General rules and rules for buildings BS EN 1993-1-2 General rules – Structural fire design BS EN 1993-1-8 Design of joints BS EN 1993-1-10 Material toughness and throughthickness properties
Building Standards (Scotland) Regulations: Domestic Technical Handbook, Section 5 Noise, HMSO Northern Ireland Building Regulations: Technical Booklet G Sound, HMSO Health Technical Memorandum 2045: Acoustics – Design considerations, HMSO
BS EN 1994 Eurocode 4: Design of composite steel and concrete structures BS EN 1994-1-1 General rules and rules for buildings BS EN 1994-1-2 General rules – Structural fire design
Building Bulletin 93: Acoustic design of schools, HMSO Robust Details Handbook www.robustdetails.com World Trade Center Building Performance Study: Data Collection, Preliminary Observations, and Recommendations, FEMA 403, May 2002
BS EN 10025 Hot rolled products of structural steels BS EN 10210 Hot finished structural hollow sections of nonalloy and fine grain steels BS EN 14399 High-strength structural bolting assemblies for preloading (provisional title)
Other Publications Al-Jabri, The behaviour of steel and composite beam-tocolumn connections in fire”, Ph.D. Thesis, University of Sheffield, 1999
BS EN 15048 Non-preloaded structural bolting assemblies (provisional title)
Brown, Hughes & Anderson, Prediction of the initial stiffness of ductile end-plate steel connections, ICE Structures & Buildings 146, February 2001
BS EN ISO 140 Acoustics. Measurement of sound insulation in buildings and of building elements BS EN ISO 4014 Hexagon head bolts. Product grades A and B
Byfield, Couchman, Dhanalakshmi & Gwyder, Limitations in the use of composite connections with unpropped construction, Eurosteel 2005
BS EN ISO 4017 Hexagon head screws. Product grades A and B
Ellis, Serviceability evaluation of floor vibration induced by walking loads, The Structural Engineer 79, 2001
BS EN ISO 4032 Hexagon nuts. Style 1. Product grades A and B
Griffin. Handbook of human vibration, Academic Press Ltd, London, 1996
DD ENV 1090 Execution of steel structures DD ENV 1090-1 General rules and rules for buildings DD ENV 1090-2 Rules for cold formed thin gauge members and sheeting DD ENV 1090-4 Supplementary rules for hollow section structures
Goodchild, Hybrid Concrete Construction, British Concrete Association, 1995 Horne & Morris, Plastic design of low-rise frames, Granada, 1981
ISO 2631-1 Mechanical vibration and shock. Evaluation of human exposure to whole-body vibration. Vibration in buildings (1 Hz to 80 Hz) DIN 6914 Sechskantschrauben mit großen Schlüsselweiten für Stahlkonstruktionen (HV-Verbindungen) (Ausgabe 10.89)
186
