arcpress structural design for architecture

STRUCTURAL DESIGNFOR ARCHITECTURE

ANGUS J MACDONALD

Structural Design for ArchitectureAngus J. Macdonald

Architectural Press

Architectural Press225 Wildwood Avenue, Woburn, MA 01801-204An imprint of Butterworth-HeinemannLinacre House, Jordan Hill, Oxford 0X2 8DPA division of Reed Educational and Professional Publishing Ltd

A member of the Reed Elsevier plc group

OXFORD BOSTON JOHANNESBURG

MELBOURNE NEW DELHI SINGAPORE

First published 1997Reprinted 1998

© Reed Educational and Professional Publishing Ltd 1997

All rights reserved. No part of this publicationmay be reproduced in any material form (includingphotocopying or storing in any medium by electronicmeans and whether or not transiently or incidentallyto some other use of this publication) without thewritten permission of the copyright holder exceptin accordance with the provisions of the Copyright,Designs and Patents Act 1988 or under the terms of alicence issued by the Copyright Licensing Agency Ltd,90 Tottenham Court Road, London, England W1P 9HE.Applications for the copyright holder's written permissionto reproduce any part of this publication should be addressedto the publishers.

British Library Cataloguing in Publication DataMacdonald, Angus J.

Structural design for architecture1. Architectural design 2. Structural design1.Title721

ISBN 0 7506 3090 6

Library of Congress Cataloguing in Publication DataMacdonald, Angus, 1945-

Structural design for architecture/Angus J. Macdonald.p. cm.

Includes bibliographical references and index.ISBN 0 7506 3090 61. Buildings. 2. Structural Design. 3. Architectural design.1. Title.TH846.M33 97-27237624. 1'771-dc21 CIP

Composition by Scribe Design, Gillingham, KentPrinted and bound in Great Britain

Contents

Foreword vii

Preface ix

Acknowledgements xi

1 Structure and architecture 11.1 The role of structure in architecture 11.2 Structural requirements 41.3 Structure types 51.4 Structural materials 111.5 Structural design 17

2 Structural design for architecture 222.1 Introduction 222.2 The relationship between structural

design and architectural design 242.3 Selection of the generic type of

structure 342.4 Selection of the structural material 402.5 Determination of the form of the

structure 412.6 Conclusion 47

3 Steel structures 493.1 Introduction 493.2 The architecture of steel - the factors

which affect the decision to select steelas a structural material 49

3.3 The properties and composition ofsteel 61

3.4 Structural steel products 633.5 Performance of steel in fire 723.6 Structural forms 73

4 Reinforced concrete structures 994.1 Introduction 994.2 The architecture of reinforced concrete -

the factors which affect the decision toselect reinforced concrete as astructural material 100

4.3 A brief introduction to concretetechnology 118

4.4 Structural forms for reinforced concrete130

5 Masonry structures 1475.1 Introduction 1475.2 The architecture of masonry - factors

which affect the decision to usemasonry as a structural material 147

5.3 The basic forms of masonry structures164

6 Timber structures 1796.1 Introduction 1796.2 Timber and architecture 1806.3 The material, its properties and

characteristics 1906.4 Properties of timber 1926.5 Grading of timber 1966.6 Timber components 1986.7 Structural forms for timber 215

Selected bibliography 233

Appendix 1: The relationship betweenstructural form and structuralefficiency 235

Appendix 2: Approximate methods forallocating sizes to structural elements 239

A2.1 Introduction 239A2.2 Structural analysis 239A2.3 Element-sizing calculations 249A2.4 Steel structures 258A2.5 Reinforced concrete structures 262A2.6 Masonry structures 263A2.7 Timber structures 263

Index 265 V

Angus Macdonald states that this book is symbolised and high tech (i.e. celebrated orprimarily for architects. In my view it is also an expressionist) is apt but contentious and couldextremely good reference book on architectural result in some lively discussion betweenstructures for students and practising architect and engineer,structural engineers. The book then divides into sections on the

He stresses that buildings are designed as a major structural materials - steel, concrete,collaborative task between architects and masonry and timber. Each of these sectionsengineers and that the earlier in the design follows a similar pattern and includesprocess this happens, the better the result. properties, advantages and disadvantages,Current teaching ideas in many universities common structural forms, etc.are, at last, acknowledging the benefits of joint Structural Design for Architecture is astudent working and it has certainly been my comprehensive and up-to-date work on theexperience that close working produces the relationship of structure to architecture and willbest product. form an extremely useful reference work for both

The early part of the book covers the history, students and practitioners of architecture andtechnology and structural philosophy of engineering. I highly recommend it and looknumerous buildings and building types and forward to having a copy in our office library,has a very comprehensive review of structuralsystems with excellent examples of seminalbuildings and their structures. It also covers Professor Tony Huntthe history of structural'material development. Chairman

The section on structure in relation to Anthony Hunt Associatesarchitecture: structure ignored, accepted, June 1997

Previous pageis blank

vii

Foreword

The architect who considers him or herself tobe an artist, dealing through the medium ofbuilt form with the philosophical preoccupa-tions of the age in which he or she lives, issurely engaged in a titanic struggle. Oneaspect of that struggle is the need to deter-mine building forms which are structurallyviable. All artists must acquire mastery of thetechnology of their chosen medium but fewface difficulties which are as formidable asthose who choose buildings as their means ofexpression. The sculptor has to contend withsimilar structural problems but his or her diffi-culties are trivial by comparison with those ofthe architect. The difference is one of scale -the size of a building, compared to that of awork of sculpture, means that the technicalhurdle which must be surmounted by thearchitect is of a different order of magnitude tothose which are faced by most other artists.

The structure of a building is the armaturewhich preserves its integrity in response toload. It is a bulky object which is difficult toconceal and which must somehow be incorp-orated into the aesthetic programme. It musttherefore be given a form, by the building'sdesigner, which is compatible with otheraspects of the building's design. Several funda-mental issues connected with the appearanceof a building including its overall form, thepattern of its fenestration, the general articula-tion of solid and void within it and even, pos-sibly, the range and juxtaposition of thetextures of its visible surfaces are affected bythe nature of its structure. The structure canalso influence programmatic aspects of a build-ing's design because the capability of the struc-ture determines the pattern of internal spaceswhich is possible. Its span potential will deter-

mine the maximum sizes of the internal spacesand its type affects the extent to which thesizes and shapes of the spaces can be variedboth within an individual storey and betweenstoreys.

The relationship between structure andarchitecture is therefore a fundamental aspectof the art of building. It sets up conflictsbetween the technical and aesthetic agendaswhich the architect must resolve. The mannerin which the resolution is carried out is one ofthe most testing criteria of the success of awork of architecture.

This book is concerned with structuraldesign for architecture. It complements myprevious volume, Structure and Architecture, anddiscusses the selection of structure type, theselection of structural material and the deter-mination of structural form. It deals primarilywith the development of the idea of the struc-ture for a building - that first stage in thestructural design process which is concernedwith the determination of the elementary formand arrangement of the structure, before anystructural design calculations are made. It isintended primarily for architects and it ishoped that it will enable students andmembers of the profession to gain a betterunderstanding of the relationship betweenstructural design and architectural design. Thebasic structural layouts and approximateelement sizes which are given in Chapters 3 to6 should, however, also allow building design-ers to use the book as an aid to the basicplanning of structural forms.

Angus MacdonaldEdinburgh

July 1997 ixPrevious pageis blank

Preface

1 would like to thank the many people who are due to all those who supplied illustrationshave assisted me in the making of this book. and especially to the Ove Arup Partnership,These are too numerous for all to be George Balcombe, Sir Norman Foster andmentioned individually, but special thanks are Partners, Paul H. Gleye, Pat Hunt, Tony Hunt,due to the following: Stephen Gibson for his the late Alastair Hunter, Jill Hunter, Denysexcellent line drawings, the staff of Lasdun Peter Softley and Associates, Ewan andArchitectural Press for their hard work in Fiona McLachlan, Dr Rowland J. Mainstone andproducing the book, particularly Neil Warnock- the Maritime Trust. I am also grateful to theSmith, Zoë Youd and Sarah Leatherbarrow. I British Standards Institution for permission towould also like to thank the staff and students reproduce tables,of the Department of Architecture at the Finally, I should like to thank my wifeUniversity of Edinburgh for the many helpful Patricia Macdonald for her encouragement anddiscussions which I have had with them on the support and for her valuable contributions totopics covered in this book. the preparation of the manuscript and

Illustrations, other than those illustrations,commissioned especially for the book, areindividually credited in their captions. Thanks Angus Macdonald


xi

Acknowledgements

Chapter 1

Structure and architecture

1.1 The role of structure inarchitecture

The final form which is adopted for a work ofarchitecture is influenced by many factorsranging from the ideological to the severelypractical. This book is concerned principallywith the building as a physical object and, in

particular, with the question of the structuralsupport which must be provided for a buildingin order that it can maintain its shape andintegrity in the physical world. The role of thebuilding as an aesthetic object, often imbuedwith symbolic meaning, is, however, alsocentral to the argument of the book; onestrand of this argument considers that the

Fig. 1.1 Offices, Dufour'sPlace, London, England,1984. Erith and Terry, archi-tects. As well as having aspace-enclosing function theexternal walls of this build-ing are the loadbearingelements which carry theweights of the floors androof. [Photo: E. & F.McLachlan]

Structural Design for Architecture

Fig. 1.2 Crown Hall, 1IT, Chicago, USA, 1952-56. LudwigMies van der Rohe, architect. This building has a steel-frame structure. The glass walls are entirely non-structural.

contribution of the structure to the achieve-ment of higher architectural objectives isalways crucial. Technical issues are accordinglyconsidered here within a wider agenda whichencompasses considerations other than thoseof practicality.

The relationship between the structural andthe non-structural parts of a building may varywidely. In some buildings the space-enclosingelements - the walls, floors and roof - are alsostructural elements, capable of resisting andconducting load (Fig. 1.1). In others, such asbuildings with large areas of glazing on theexterior walls, the structure can be entirelyseparate from the space-enclosing elements(Fig. 1.2). In all cases the structure forms the

basic carcass of the building - the armature towhich all non-structural elements are attached.

The visual treatment of structure can besubject to much variation. The structuralsystem of a building can be given great prom-inence and be made to form an important partof the architectural vocabulary (Fig. 1.3). At theother extreme, its presence can be visuallyplayed down with the structural elementscontributing little to the appearance of thebuilding (Fig. 1.4). Between these extremes liesan infinite variety of possibilities (see Section2.2). In all cases, however, the structure, byvirtue of the significant volume which itoccupies in a building, affects its visual charac-ter to some extent and it does so even if it isnot directly visible. No matter how the struc-ture is treated visually, however, the need fortechnical requirements to be satisfied mustalways be acknowledged. Structural constraints


Fig. 1.3 HongkongBankHeadquarters, Hong Kong,1979-84. Foster Associates,architects. The structure ofthis building is expressedprominently both on theexterior and in the interior.It contributes directly aswell as indirectly to theappearance of the building.[Photo: Ian Lambot.Copyright: Foster &Partners ]

Fig. 1.4 Staatsgalerie,Stuttgart, Germany,1980-83, lames Stirling,architect. This building hasa reinforced concrete struc-ture and non-structuralcladding. Although thestructure plays a vital rolein the creation of thecomplex overall form it isnot a significant element inthe visual vocabulary.[Photo: P. Macdonald]


1.2 Structural requirements

therefore exert a significant influence, overt orhidden, on the final planning of buildings.

This book is concerned with the program-matic aspects of the relationship betweenarchitecture and structure. Chapter 2, in par-ticular, deals with the process by which theform and general arrangement of structures forbuildings are determined - with the design ofarchitectural structures, in other words.Information on basic forms of structure - therange of structural possibilities - is essentialto the success of this process; this is providedin subsequent chapters which deal separatelywith the four principal structural materials ofsteel, reinforced concrete, masonry and timber.More general aspects of the topic are reviewedbriefly here.

Fig. 1.5 The first of the frameworks here is capable ofachieving equilibrium under the loading shown but isunstable. The insertion of the diagonal element in thesecond framework renders it capable of achieving stableequilibrium.

state of static equilibrium but is not stable andwill collapse if subjected to a small lateraldisplacement. The insertion of a diagonalbracing element in the second frameworkprevents this and renders the system stable.Most structural arrangements require bracingfor stability and the devising of bracingsystems is an important aspect of structuraldesign.

As the simple diagrammatic structure in Fig.1.6 illustrates, the structural elements of abuilding provide the link between the appliedloads and the foundation reactions in orderthat equilibrium can be achieved. To be effec-tive the elements must be of adequatestrength. The strength of an element dependson the strength of the constituent material andthe area and shape of its cross-section. Thestronger the material and the larger the cross-section the stronger will be the element. It ispossible to produce a strong element eventhough the constituent material is weak byspecifying a very large cross-section.

In the case of a particular structure, once therequirements for stability and equilibrium havebeen met, the provision of elements withadequate strength is a matter firstly of deter-mining the magnitudes of the internal forceswhich will occur in the elements when the peakload is applied to the structure. Secondly, astructural material of known strength must beselected and thirdly, the sizes and shapes ofcross-sections must be chosen such that eachelement can safely carry the internal forcewhich the load will generate. Calculations are

The principal forms of loading to which build-ings are subjected are gravitational loads, windpressure loads and inertial loads caused byseismic activity. Gravitational loads, which arecaused by the weight of the building itself andof its contents, act vertically downwards; windand seismic loads have significant horizontalcomponents but can also act vertically. Toperform satisfactorily a structure must becapable of achieving a stable state of staticequilibrium in response to all of these loads -to load from any direction, in other words. Thisis the primary requirement; the form andgeneral arrangement of a structure must besuch as to make this possible.

The distinction between the requirementsfor stability and equilibrium is an importantone and the basic principles are illustrated inFig. 1.5. Equilibrium occurs when the reactionsat the foundations of a structure exactlybalance and counteract the applied load; if itwere not in equilibrium the structure wouldchange its position in response to the load.Stability is concerned with the ability of astructural arrangement which is in equilibriumto accommodate small disturbances withoutsuffering a major change of shape. The first ofthe beam/column frameworks in Fig. 1.5 is in a4

Fig. 1.6 Force system in abuilding's structure. The gravi-tational load on the roof isconducted, via the roof trussand the walls, to the founda-tions where it is balanced byreactions from the substrata.The same is true of loadsimposed on the floors whichare transmitted by the floorstructural elements and wallsto the foundations. The rooftruss, wall and floor elementsmust be strong enough tocarry the internal forces gener-ated by the load.


an essential aspect of this process and arerequired both to determine the magnitudes ofthe forces in the individual elements - anactivity known as structural analysis - and thento calculate the required sizes of the elementcross-sections.

A fourth property which a structure mustpossess, in addition to the requirements ofequilibrium, stability, and strength, isadequate rigidity. All structural materialsdeform in response to load and it is necessarythat the overall deflection of a structure shouldnot be excessive. As with strength, the rigidityof the structure depends on the properties ofthe material and the sizes of the cross-sections, which must be large enough toensure that excessive deflection does notoccur. Like strength, rigidity is checked andcontrolled through the medium of calculations.

To summarise, the basic requirements of thestructure (the firmness element of the archi-

tectural shopping list of 'firmness', 'commod-ity' and 'delight') are the ability to achieveequilibrium under all possible load conditions,geometric stability, adequate strength andadequate rigidity. Equilibrium requires that thestructural elements be properly configured,stability is ensured by the provision of abracing system; and adequate strength andrigidity are provided by the specification ofstructural elements which are of sufficient size,given the strengths of the constituentmaterials.

1.3 Structure types

1.3.1 Post-and-beam structuresMost architectural structures are of the post-and-beam type and consist of horizontalspanning elements supported on verticalcolumns or walls. A characteristic of this type 5

Non-loadbearingpartition wall


of structure is that the horizontal elements aresubjected to bending-type internal forcesunder the action of gravitational load(normally the primary load on an architecturalstructure). This has two consequences. Firstly,it requires that the structural material becapable of resisting both tension and compres-sion (e.g. steel, reinforced concrete, timber).Secondly, it is an inefficient type of structure(larger volumes of material are required tosupport a given load than are necessary withother types of structure).1 The post-and-beamstructure has the great advantage that it issimple and therefore cheap to construct. Thisgroup of structures can be subdivided into thetwo categories of 'skeleton-frame' structuresand 'panel' structures. The latter are alsoloadbearing-wall structures.

Skeleton-frame structures consist of anetwork of beams and columns which supportfloor slabs and roof cladding and to which wall

cladding is attached (Fig. 1.7). The configur-ation of the beam-and-column grid which isadopted in a particular case depends on theoverall form of the building concerned, on theinternal planning requirements and on theproperties of the particular structural materialwhich has been chosen - see Chapters 3 to 6.

In this type of arrangement the structureoccupies a relatively small volume and this isin fact one of its principal advantages.Considerable freedom is available to the build-ing designer in the matter of internal planningbecause both the internal partition walls andthe exterior walls are non-loadbearing. Largewall-free spaces can therefore be created in theinteriors and different plan-forms adopted atdifferent levels in multi-storey buildings. Thechoice of external treatment is wide. Relativelyfragile materials such as glass can be used andlittle restriction is placed on the locations ofdoors and windows (Fig. 1.8).

A consequence of the small structuralvolume of the skeleton frame is that structuralloads are concentrated into slender columnsand beams and these elements must therefore1 See Appendix 1 for an explanation of this.6

Fig. 1.7 Steel skeletonframework. In this arrange-ment, which is typical of amulti-storey steel-framestructure, concrete floorslabs are supported by agrid of steel beams whichis in turn supported byslender steel columns.These elements form thestructural carcass of thebuilding. External wallsand internal partitions arenon-structural and can bearranged to suit planningand aesthetic require-ments.


Fig. 1.8 NenfeldwegHousing, Graz, Austria,1984-88. Gunther Domenig,architect. The structure ofthis building is a reinforcedconcrete framework. Itsadoption has allowedgreater freedom to beexercised in the internalplanning and external treat-ment than would have beenpossible with a loadbear-ing-wall structure. [Photo:E. & F. McLachlan]

be constructed in strong materials such assteel or reinforced concrete.

Panel structures are arrangements of struc-tural walls and horizontal panels (Fig. 1.9). Thewalls may be of masonry, concrete or timber -the last of these being composed of closelyspaced vertical elements - and the floors androof of reinforced concrete or timber - againthe configuration with timber is one of closelyspaced elements, in this case floor joists ortrussed rafters (Fig. 6.39).

Many different combinations of elementsand materials are possible. In all cases thevolume of the structure is large in relation toskeleton-frame equivalents with the result thatthe structural elements are subjected to lower

levels of internal force. Structural materials oflow or moderate strength, such as masonry ortimber, are therefore particularly suited to thisform of construction (Fig. 1.10).

Panel structures impose greater constraintson planning freedom than skeleton-frameequivalents because structural considerationsas well as space-planning requirements mustbe taken into account when the locations ofwalls are determined. The creation of largeinterior spaces is problematic as is the vari-ation of plans between different levels inmulti-storey arrangements. The advantage ofthe panel form of structure is that it is simplerto construct than most skeleton frames andconsiderably less expensive. 7


Fig. 1.9 Multi-storeyloadbearing-wall structure.In this arrangement thefloors and roof of the build-ing are carried by the walls.The structural sections givean indication of the action ofthe walls in response togravitational and windloading. The plan arrange-ment of the walls must besuch as to provide goodstructural performance andthe plan-form must bemaintained through alllevels. These are constraintswhich must be acceptedduring the internal planningof this type of building.

Fig. 1.10 Housing,Gogarloch Syke, Edinburgh,Scotland, 1996. E. & F.McLachlan, architects. Thesehouses are good examples ofpanel structures in loadbear-ing masonry. | Photo: KeithHunter; copyright: E. & F.McLachlan |

8


1.3.2 Vaults and domesVaults and domes are structure types in whichthe dominant feature is an upwards curvaturetowards the dominant, downward-acting gravi-tational load (Fig. 1.11). They belong to a classof structure in which the internal forces arepredominantly axial rather than of the bendingtype,2 and, in the case of vaults and domes,this internal force is compressive. They aretherefore normally constructed in materialswhich perform well in compression, such asmasonry or concrete.

The axial-compressive-stress-only conditionwhich is associated with vaults and domes hastwo important consequences. First, it allows

2 See Macdonald, Structure and Architecture, Chapter 4 for aclassification of structure types. The principles aresummarised here in Appendix 1.

Fig. 1.11 TGV Station, Lyon-Satolas, France 1989-94,Santiago Calatrava, architect/engineer. A vaulted structurein reinforced concrete is used here to achieve a relativelylong span. [Photo: E. & F. McLachlan]

large horizontal spans to be achieved withmaterials, such as masonry or unreinforcedconcrete, which have little tensile strength (Fig.5.4): large-span interiors can be created inmasonry only by the use of domed or vaultedstructures. This was the principal reason forthe use of this type of arrangement prior to theinvention of modern materials such as steeland reinforced concrete which allow largespans to be achieved with post-and-beamforms due to their ability to resist bendingeffectively.

Secondly, and perhaps more importantly forthe buildings of today, it allows loads to beresisted with much greater structural efficiencythan is possible where bending is the principal 9


Fig. 1.12 Building for IBM Europe travelling exhibition.Renzo Piano, architect/engineer, Ove Arup and Partners,structural engineers. A complex, highly efficient vaultedstructure like this would not normally be used for a short-span enclosure. It was justified in this case due to arequirement for a lightweight structure for a portablebuilding. The choice of lightweight materials - timber andplastic-was also sensible. [Photo: Ove Arup & Partners]

result of the application of load.3 In modernpractice, vaults or domes are normally used toachieve high levels of structural efficiency,either because a very long span is required orbecause a special requirement must be satis-fied such as the need for a very lightweightstructure (Fig. 1.12).

1.3.3 Tents and cable networksTents and cable networks are tensile equiva-lents of domes and vaults (Fig. 1.13). The in-ternal forces which occur in these structuresare those of axial tension and they are there-

fore, like their axial compressive equivalents,potentially highly efficient in resisting load.4 Aswith domes and vaults they are used in situa-tions in which high structural efficiency isdesirable, such as for long spans or where alightweight structure is required.

1.3.4 Combined-action structuresA fourth category of structure is one in whichthe load is resisted by a combination ofbending and axial internal forces. The ubiqui-tous portal frame (Fig. 1.14) is perhaps thebest-known example of this but any structurewhich is neither purely 'form-active' nor purely'non-form-active' will carry load through thecombined effect of axial and bending action.These structures have properties which areintermediate between those of the post-and-beam arrangement, which is inefficient butsimple to construct, and the arch, vault orcable network, which are highly efficient but

10 3 See Macdonald, Structure and Architecture, Chapter 4. 4 See Appendix 1.


complicated to construct. Combined-actionstructures are therefore used in situations inwhich intermediate levels of efficiency arerequired, for example in the medium-spanrange. They are most often found in the formof skeleton-frame arrangements.

Fig. 1.13 Olympic Stadium, Munich, Germany, 1968-72.Behnisch & Partner, architects, with Frei Otto. The struc-ture of this canopy consists of a network of steel wires (thevery fine square mesh) supported on a system of mastsand cables. The pattern of heavy rectangular lines resultsfrom the flexible joints between the cladding panels.Highly efficient structure types such as this are requiredwhere long spans are involved. [Photo: A. Macdonald]

1.4 Structural materials

The form and general arrangement of architec-tural structures are greatly influenced by theproperties of the materials from which they areconstructed. For this reason the basic structuretypes appropriate to the four principal materi-als of steel, reinforced concrete, masonry andtimber are described in separate chapters.

Each material has its own individual charac-teristics in terms of physical properties and

manufacture which contribute to determiningthe structural forms for which it is mostsuitable. These issues are considered in detailin the chapters on individual materials. Onlythe most general aspects are reviewed here.

The properties of materials which affect theload-carrying performance of a structure arestrength, elasticity and, to a lesser extent,specific gravity (which determines the self-weight of structural elements). Other signifi-cant physical properties are durability (i.e. 11


12

Fig. 1.14 Palmerston Special School, Liverpool, England,1973-76 (demolished 1989). Foster Associates, architects;Anthony Hunt Associates, structural engineers. Semi-form-active portal frames of steel hollow-section are used hereas the primary structural elements in a multi-bay arrange-ment with relatively short spans. The moderately highefficiency of this type of structure has permitted veryslender elements to be adopted. |Photo: lohn Donat]

susceptibility to both physical and chemicaldeterioration) and performance in fire. Non-physical, but interrelated, properties which arerelevant are cost, availability and environmen-tal impact. The last of these is concerned withthe environmental issues (depletion of mater-ial resources and energy sources, pollution, thehealth of workers, etc.) which will arise fromthe manufacture, installation and use of struc-tural elements of a particular material.

Of the purely physical properties, perhapsthe most important so far as structural per-formance is concerned is strength, althoughthe ratios of strength to weight and strength to

elasticity are also significant because thesedetermine the efficiency with which a materialcan be used. Of the four principal structuralmaterials, steel and reinforced concrete maybe thought of as high-strength materials andtimber and masonry as low-strength materials.Each of the four has a unique combination ofproperties which makes it perform best inparticular types of structural arrangement.

Another set of factors which influences thetypes of structure for which a material issuitable are the conditions of its manufactureand finishing. These determine the type ofproduct in which the material becomes avail-able to the builder. Steel, for example, is avail-able in the form of manufactured elementswhich are straight and of constant cross-section. The construction of a steel structure istherefore a process of assembly of prefabri-cated components. Concrete, on the otherhand, normally arrives on a building site inliquid form and the building is literally formed


Fig. 1.15 The high strength of steel allows the creation ofstructures with very slender elements. In buildings whichare supported by steel frameworks the volume occupied bythe structure is low in relation to the total volume of thebuilding. (Photo: A. Macdonald).

Fig. 1.16 Multi-storey steel frameworks are typically acombination of l-section beams and H-section columns.(Photo: A. Macdonald).

on the site by pouring the concrete intomoulds.

The strongest of the structural materials issteel, which is therefore used for the tallestbuildings and the longest spans. It is a highlyversatile material, however, and is also usedover a very wide variety of building types andspan sizes. Because it is very strong the struc-tural elements are slender and of low volume,so that steel is used almost exclusively in theform of skeleton-frame structures (Fig. 1.15).The majority of these are assembled fromstandard rolled sections; these are elementswith I- and H-shaped cross-sections (Fig. 1.16)and longitudinal profiles which are straight 13


14

and parallel-sided, and which lend themselvesto use in straight-sided frameworks. Most steelstructures therefore have a regular, rectilineargeometry. The range of possible forms hasbeen extended in recent years by the develop-ment of techniques for bending rolled sectionsinto curved shapes and by the increased use ofcasting to produce structural steel elements.However, the fact that all steel structures areprefabricated tends to require that regular andrepetitive structural geometries be adoptedeven though the individual components are ofirregular or curvilinear shape.

A typical steel-frame building thus has arelatively simple overall form and an interiorwhich is open and unencumbered by structuralwalls. Great freedom is therefore available tothe designer so far as the internal planning ofsuch buildings is concerned: the interiorvolumes may be left large or they may besubdivided by non-structural partition walls;different arrangements of rooms may beadopted at different levels and a free choice isavailable in the treatment both of the externalwalls and of the internal partitions.

An advantage of prefabrication is that steelstructural elements are manufactured and pre-assembled in conditions of very high qualitycontrol. Great precision is possible and this,together with the slenderness which resultsfrom high strength, means that structures ofgreat elegance can be produced. Steel is there-fore frequently selected as much for itsaesthetic qualities and for the stylistic treat-ment which it makes possible as for its struc-tural performance.

Reinforced concrete, the other 'strong'material, is of lower strength than steel withthe result that equivalent structural elementsare more bulky. It too is used principally inskeleton-frame structures of regular geometryand therefore offers similar advantages to steelin respect of internal planning and exteriortreatment.

Concrete structures are normally manufac-tured on the building site by the pouring ofliquid concrete into temporary formwork struc-tures of timber or steel which are speciallymade to receive it. This allows a wide choice of

Fig. 1.17 Goetheanum, Eurhythmeum, 'Glashaus' studio,Dornach, Switzerland, 1924-28. Rudolf Steiner, architect.The complexity of form which is possible with in situreinforced concrete is well illustrated here. [Photo: E. & F.McLachlan]

element shape to be available. Continuitybetween elements is also easily achieved andthe resulting statical indeterminacy5 facilitatesthe production of structures of complicatedform. Irregular geometries in both plan andcross-section, with cantilevering floor slabs,tapering elements and curvilinear forms mayall be produced more easily in reinforced

5 See Macdonald, Structure and Architecture, Appendix 3, foran explanation of the phenomenon of statical indeter-minacy and its relevance to the determination of struc-tural form.


concrete than in steel (Figs 1.17 and 4.20). Theshapes of the elements are usually, however,more crude at a detailed level.

Masonry is the term for a range of materialswhich have the common characteristic thatthey consist of solid elements (bricks, stones,concrete blocks, clay tiles) which are bedded inmortar to form piers and walls. A range ofother materials with similar physical propertiesto masonry, such as various forms of dried orbaked earth, are suitable for the same types ofstructural configuration.

The principal physical properties of thesematerials are moderate compressive strength,relatively good physical and chemical durabil-ity and good performance in fire. Very signifi-cant properties are brittleness and low tensilestrength. The last of these, in particular, has aprofound effect on the structural forms forwhich masonry is suitable. Lack of tensilestrength means that bending-type load ofsignificant magnitude cannot be resisted sothat masonry structural elements must besubjected principally to axial compression

Fig. 1.18 Casa Pfaffli, Lugano, Switzerland, 1980-81.Mario Botta, architect. The structure of this buildingconsists of loadbearing masonry walls supportingreinforced concrete horizontal structural elements. [Photo:E. &F. McLachlan]

only. They can be used as walls, piers, arches,vaults and domes but not as slab-type horizon-tally spanning elements. When used as walls,they must be supported laterally at regularintervals due to their inability to withstandout-of-plane bending loads such as mightoccur due to the effect of wind pressure.

Masonry is therefore used in the loadbear-ing-wall form of structure (Fig. 1.18) to producemulti-cellular buildings in which the principalwalls are continuous through all levels, givingsimilar arrangements of spaces on everystorey. The horizontal elements in such build-ings are normally of timber or reinforcedconcrete but may be of steel. Structural con-tinuity between these elements and thesupporting masonry walls is difficult to achieveand the internal forces in the structural 15


16

elements must be maintained at modestlevels. Spans are therefore normally kept smalland in modern practice loadbearing masonrybuildings are usually fairly small in scale. (Thisis in contrast to the very large-scale masonrystructures of previous ages which wereachieved by the use of masonry vaults anddomes as the horizontal spanning elements -see Section 5.2. and Fig. 5.1.)

Although modest in scale, modern loadbear-ing masonry structures exhibit very goodcombinations of properties and produce build-ings which are durable and fireproof and whichhave walls which perform extremely well inrespect of thermal and acoustic insulation.They are therefore ideal for all kinds of livingaccommodation.

Timber is a structural material which hassimilar properties to steel and reinforcedconcrete in the sense that it can carry bothtension and compression with almost equalfacility. It is therefore capable of resistingbending-type load and may be used for alltypes of structural element. It is significantlyless strong than either steel or reinforcedconcrete, however, with the result that largercross-sections are required to carry equivalentamounts of load. In practice, large cross-sections are rarely practicable and timberelements must therefore normally be used insituations where the internal forces in thestructural elements are low, that is in buildingsof small size, and, in particular, short spans.

A significant advantage which timber hasover other structural materials is that it is verylight, due to its fibrous internal structure andthe low atomic weights of its constituentchemical elements. This results in a high ratioof strength to weight. Other advantageousproperties are good durability and, despitebeing combustible, relatively good perform-ance in fire.

Timber is commonly used for the horizontalfloor and roof elements in loadbearingmasonry structures (Fig. 6.39). Loadbearing-wall 'panel' construction, in which the struc-ture of a building is made entirely of timber, isanother common configuration (Fig. 1.19). Wallpanels consist of closely spaced timber posts

Fig. 1.19 Timber loadbearing-wall structure. Everythinghere is structural. The wall and floor structures consist ofclosely spaced timber elements. Temporary bracingelements, which provide stability until non-loadbearingcross-walls are inserted, are also visible. [Photo: A.Macdonald]

tied together by horizontal timber elements atthe base and top, and the panels are arrangedin plan configurations which are similar tothose used in masonry construction. Thegrouping together of timber posts into panelsensures that the load which each carries isrelatively small. Even so, such buildings rarelyconsist of more than two storeys. Timberloadbearing-wall structures are simple toconstruct, using components which are light inweight and easily worked. The use of timber inskeleton-frame configurations for multi-storeystructures is relatively rare and normallyrequires that columns be closely spaced to


Fig. 1.20 Forth Railway Bridge,Scotland, 1882-90, Henry Fowlerand Benjamin Baker, engineers.Two types of structure are visiblehere. Short-span viaducts,consisting of parallel-chordtriangulated girders, are used ateach end where the groundconditions allow closely-spacedfoundations to be provided. Thetwo deep river channelsseparated by the island arespanned by an arrangement ofthree pairs of balancedcantilevers. The configuration ofthe bridge was determined by acombination of site conditions,structural requirements andfunction. [Photo: P. & A.Macdonald]

keep beam spans short. Very large single-storey structures have, however, beenconstructed (see Section 6.2).

1.5 Structural design

As with any other type of design, the evolutionof the form of a structure is a creative actwhich involves the making of a whole networkof interrelated decisions. It may be thought of

as consisting of two broad categories of activ-ity: first, the invention of the overall form andgeneral arrangement of the structure and,secondly, the detailed specification of theprecise geometry and dimensions of all of theindividual components of the structure and ofthe junctions between them.

In the case of an architectural structure bothof these activities, but especially the first, areclosely related to the broader set of decisionsconnected with the design of the building. The 17


18

overall form of the structure must obviously becompatible with, if not identical to, that of thebuilding which it supports. The preliminarystage of the structural design is therefore virtu-ally inseparable from that of the building, takenas a whole. It is at this stage that architecturaland structural design are most closely relatedand that the architect and engineer, be theydifferent persons or different facets of the sameindividual, must work most closely together.The second stage of the design of the structure,which is principally concerned with the sizing ofthe elements and the finalising of details, suchas the configuration of the joints, is principallythe concern of the structural engineer.

The different aspects of the structural designactivity are most easily seen in relation topurely engineering types of structure, such asbridges. It will be instructive here, beforelooking at the process as it takes place in thecase of a building, to consider the design of aprominent example of this type. The ForthRailway Bridge in Scotland (Fig. 1.20), despitebeing now over 100 years old, provides a goodillustration of the various stages in the evolu-tion of a structural design. The issues involvedare broadly similar to those which occur in anyengineering design project, including those ofthe present day. The same issues will beconsidered again in Chapter 2, where they arediscussed in relation to architectural design.This preliminary review, in the context ofengineering, serves to identify the essentialaspects of the structural design process.

As is normal in bridge design, the mostsignificant sets of factors which influenced thedesign of the Forth Bridge were thoseconnected with its function and with itslocation. The ground level at each side of theestuary of the River Forth, where the bridge issituated, slopes steeply up from the shore andthe railway therefore approaches from a level ofapproximately 50 m above water level at eachend. This, together with the requirement thatthe busy shipping channels which the bridgecrosses should not be blocked, dictated thatthe railtrack level should also be relatively high(50 m above sea level). At one shore a broadstrip of low lying ground occurs close to the

Fig. 1.21 Forth Railway Bridge, Scotland, 1882-90, HenryFowler and Benjamin Baker, engineers. The main part ofthe structure consists of three pairs of balancedcantilevers. In this shot the central tower of one pair ofcantilevers has been completed and the first elements ofthe cantilevers themselves have been added. The arrange-ment was adopted so that the uncompleted structurecould be self-supporting throughout the entire period ofconstruction. [Photo: E. Carey; copyright: British RailBoard Record Office]

edge of the water. At the other, the groundrises steeply from the water's edge but a broadstrip of shallow water occurs close to the shore.Between these two flanking strips of relativelylevel ground the estuary consists of two verydeep channels separated by a rocky island. Thebridge was therefore broken down into threeparts and made to consist of two long approachviaducts, each with a sequence of girdersspanning relatively short distances betweenregularly spaced piers, and a massive centralstructure spanning the two deep channels.


The structure which spans the centralchannels consists of three sets of balancedcantilevers - a configuration which wasadopted because it could be built without theneed for temporary supporting structures. Thearrangement was essential due to the nearimpossibility of providing supporting struc-tures in the deep channels and due to theneed to maintain the shipping lanes free ofobstruction during the construction process.6

The method of construction which wasadopted was to build three towers first andthen extend pairs of cantilevers simultaneouslyon each side of these (Fig. 1.21). The structurewas therefore self-supporting during the entireprocess of construction.

The basic form of the structure is showndiagrammatically in Fig. 1.22 together with thebending moment diagram7 which results fromthe peak load condition (a distributed load

Fig. 1.22 Forth Railway Bridge, Scotland. 1882-90,Henry Fowler and Benjamin Baker, engineers.(a) A diagrammatic representation (not to scale) of themain elements of the bridge. The two river channels arecrossed by three sets of balanced cantilevers. Viaductsconsisting of short-span girders supported by closelyspaced piers provide the necessary link with the highground on the approaches to the bridge.

(b) The bending moment diagram which results from theaction of a uniformly distributed load across the entirestructure gives an indication of the variation which occursin the magnitudes of the internal forces across the span.(c) The insertion of two extra hinges creates a cantilever-and-suspended-span arrangement which alters thebending moment diagram (d) and reduces the magnitudeof the maximum bending moment.(e) The external profile which was finally adopted is closelyrelated to the bending moment diagram. The triangulationof the internal geometry was necessary to achieve highstructural efficiency.

6 The device of constructing the main elements of thebridge on the shore and floating these into the finalposition, which had been used earlier in thenineteenth century at Saltash (I. K. Brunei) and theMenai Straights (Robert Stephenson and WilliamFairbairn) was impractical due to the very long spansinvolved.

7 The bending moment is the internal force produced bythe load on the structure in the type of arrangementshown in Fig. 1.22. The bending moment diagram is agraph which shows how the intensity of this internalforce varies across the span. For an explanation ofbending moment see Macdonald, Structure and

Architecture, Chapter 2. 19

(b)

(c)

(d)

(e)

(a)


Fig. 1.23 Forth RailwayBridge, Scotland, 1882-90,Henry Fowler and BenjaminBaker, engineers. The railwaytrack is carried on an internalviaduct which is supported atthe junctions of the triangu-lated main structure. [Photo:A. Macdonald]

20

across the entire span). This shows that theintensity of internal force is at its highest atthe locations of the support towers and falls tozero at the mid-span points, where theadjacent cantilevers are joined. The distribu-tion of internal force was modified by theinsertion of two hinge-type connectionsbetween each set of cantilevers (Fig. 1.22c)which had the effect of reducing the magnitudeof the maximum internal force at each supporttower.

Figure 1.22c represents, in diagrammaticform, the basic configuration which was finallyadopted for the bridge. It was modified to giveimproved load-carrying efficiency by matchingthe longitudinal profile of the structure to thepattern of internal forces so that the structuralmaterial was concentrated at the locations ofhighest internal force. It was further improvedby the adoption of a triangulated internalgeometry. Yet another decision taken by thedesigners was to carry the railtrack on a

viaduct (similar to the approach viaducts) size of each cross-section, thickness of metal,located within the primary structure and etc.). At this stage structural calculations weresupported by it at regular intervals (Fig. 1.23). employed to determine the magnitudes of theThe primary structure therefore provided internal forces which each element would carryregularly spaced supports for this internal and to check that sufficient thicknesses ofviaduct rather in the manner of the equally metal were specified to maintain the stressesspaced piers which carry the approach within acceptable limits. The detailedviaducts. determination of the configuration of the joints

Figure 1.22e shows diagrammatically the between the elements, which could only befinal form of the bridge. It represents the done once the sizes of the elements wereculmination of the first stage of the design known, was then carried out.process in which the form and general arrange- The description given above illustrates, inment of the entire structure were determined. much abbreviated form, the typical sequenceThe sequence of decisions which led to this of decisions which is involved in the design ofproposal illustrates, in a much abbreviated any major civil engineering structure andform,8 the key stages of the preliminary design allows the two key stages in this - the initialprocess and allows the nature of the process determination of form and the final realisa-which concerns us here to be appreciated. In tion of the form - to be appreciated. Aparticular, it shows that the basic form of this sequence of decision-making which is broadlypurely engineering structure was determined similar to this is involved in the design of allby the designers from a consideration of the structures whether civil engineering or archi-function of the bridge, from the constraints of tectural.the site, from a knowledge of structural behav- In the case of an architectural structure, theiour and from an awareness of the vocabulary form and general arrangement must also beof structural possibilities which was available entirely compatible with that of the buildingto them. The design was an imaginative which it will support. The preliminary stage inresponse to these various influences. the design of an architectural structure is

The second stage in the design process, the therefore inseparable from the design processrealisation of the structure, led to the detailed of the building as a whole. Once the overallspecification of the various structural elements form of a building has been determined, thefrom which the bridge was constructed. This choice of structural arrangement will normallyinvolved decisions on the precise shape of the be fairly narrow. The act of architectural designindividual elements (in longitudinal profile and is therefore also an act of structural design incross-section) and on the precise amount of which the most fundamental decisions relatingmaterial which would be specified (the overall to the structure are taken.


21

8 Several alternative arrangements were in fact consid-ered by the designers before the chosen form wasadopted.

Chapter 2

Structural design for architecture

22

2.1 Introduction

The purely technical aspects of the design of astructure were reviewed at the end of the previ-ous chapter. The complex relationship betweenstructural design and architectural design willnow be considered.

As was seen in Section 1.5 the process ofstructural design may be subdivided into twoparts: there is a preliminary design stage, whenthe form and general arrangement of the struc-ture are devised, and a second stage in whichstructural calculations are performed and thedimensions of the various structural elementsare determined. In the case of an architecturalstructure many of the decisions associatedwith the preliminary stage of the design of astructure are taken, consciously or uncon-sciously, when the form of the building isdetermined. The general arrangement chosenfor a building will normally determine the typeof structure which will have to be adopted tosupport it and will probably also dictate theselection of structural material.

In the case of the Willis, Faber and Dumasbuilding (Fig. 4.17), for example, where therewas a requirement for a large wall-free interiorand glass external walls, there was no alterna-tive to the adoption of a frame-type structure.The requirements for a curvilinear plan-formand for columns which were set back from theperimeter, dictated that reinforced concreterather than steel be employed as the structuralmaterial. The outcome in this case was a build-ing in which architectural and structuralrequirements were satisfied in equal measureand the building stands up well to both archi-tectural and technical criticism.

Fig. 2.1 Rooftop Remodelling in Vienna, Austria, 1988.Coop Himmelblau, architects. This glazed, irregular formrequired that a skeleton framework structure be adopted.[Photo: Gerald Zugmann]

The complex arrangements of the RooftopOffice in Vienna by the Coop Himmelblaugroup, to take another example (Fig. 2.1),would have been unrealisable with any othertype of structure than a skeleton framework,which had to be of steel to ensure that theelements were sufficiently slender. The Hysolarbuilding of Behnisch (Fig. 2.2) is a similar typeof building and could only have been realisedwith a skeleton framework of structural steel.The buildings of Richard Meier (Fig. 4.18) andFrank Gehry (Fig. 4.20), on the other hand,required that reinforced concrete structures beadopted.

Thus, although some aspects of the designof structures, such as the precise geometry of a

Structural design for achitecture

Fig. 2.2 Solar ResearchInstitute (Hysolar),Stuttgart, Germany,1988-89. Gunther Behnisch& Partner, architects. Theirregular geometry andlarge areas of glazingrequired that a skeletonframework structure beadopted to support thisbuilding. The choice ofsteel produced a particularaesthetic quality - themost notable aspects ofthis are the slenderness ofthe structural elementsand the refined appearanceof the exposed steelwork.[Photo: E. &F. McLachlan]

beam or column grid or the dimensions of theelements, may remain undecided until a laterstage in the design, many important structuralchoices will be closed once the overall form ofa building has been determined. The initialconcept for a building, which determines itsoverall form and the disposition of solid and

void within it, therefore exerts a dominatinginfluence on its subsequent structural make-up. The architect is thus, consciously or uncon-sciously, a structural designer.

The design of the structure which willsupport a building is an identifiable anddiscrete part of the overall design process in 23


which four broad categories of decision-makingmay be identified. These are: first, the decisionon the kind of relationship which will existbetween the architectural design and the struc-tural design; secondly, the selection of thegeneric type of structure for the building;thirdly, the selection of the structural material;and lastly, the determination of the detailedform and layout of the structure. In a particularcase these sets of decisions may not be takenin an ordered sequence and some - forexample on the relationship which is to beadopted between structural and architecturaldesign - may not even be taken consciously.They are nevertheless decisions which aretaken by the designer of every building.

In the design of the majority of buildings, inwhich the relationship between architecturaldesign and structural design is not one of'structure ignored' (see Section 2.2 for anexplanation of this term), the preliminarystructural decisions are taken consciously,usually by a team of designers which includesarchitects and structural engineers. The natureof the decision-making sequence is explored inthis chapter.

It is quite possible to project whole formsin the mind without any recourse tomaterial, ...'

More recently, Wolf Prix of CoopHimmelblau expressed the view that:

... we want to keep the design moment freefrom all material constraints ...2

It is possible to say therefore that at leastsome European architects have, since theRenaissance, considered it feasible to ignorestructural considerations when they invent theform of a building, believing that a preoccupa-tion with technical matters inhibits the processof creative design. The second of the quota-tions above is a fairly explicit comment from acontemporary architect whose work manywould regard as controversial but it neverthe-less describes the reality of a very significantproportion of present-day architectural designactivity, including much of what might bethought of as belonging to the mainstreamrather than the fringe. Many architects, in fact,pay little attention to structural issues whendetermining the form of a building. Few,however, find it appropriate to make state-ments such as that quoted above.

Just as it is possible virtually to ignore struc-tural issues when carrying out the preliminarydesign of a building, it is also possible toallocate the highest priority to the satisfactionof structural requirements. In such cases theaesthetic and programmatic aspects of thedesign are accorded a lower priority than thoseconnected with the structure and are notallowed to compromise the quality of thestructural design. This approach can occurthrough necessity, when the limits of what isfeasible structurally are approached due, forexample, to the extreme height of a building or

2.2 The relationship betweenstructural design and architecturaldesign

2.2.1 IntroductionThe detailed design of a structure is normallycarried out by a structural engineer (moreprobably a team of engineers) but, as wasnoted above, the overall form of an architec-tural structure is determined by that of thebuilding which it supports and therefore princi-pally by the architect (or architectural team).This raises the issue of the extent to which thearchitect should be preoccupied by structuralconsiderations when determining the form andgeneral arrangement of a building.

There are several possible approaches tothis. Leon Baptista Alberti, who in his treatiseon architecture written in the fifteenth centurymore-or-less outlined the job description ofthe modern architectural profession, stated:24

1 L. B. Alberti, On the Art of Building in Ten Books, 1486,

Trans. Rykwert, Leach and Tavernor, London, 1988.2 Quotation from 'On the Edge', the contribution of Wolf

Prix of Coop Himmelblau to Architecture in Transition:Between Reconstruction and New Modernism, N o e v e r (ed . ) ,1991.


to the need for a very long span (e.g. Fig. 3.10).It can also occur through choice.

Some architects hold the view that one ofthe marks of a well-designed building is thatall potential conflicts between the architecturalprogramme and its structural consequenceshave been resolved without either aspectdominating the other. This represents yetanother relationship between structure andarchitecture and it is one with which mostmainstream architects would openly, ifperhaps somewhat hypocritically, concur. Inthis scenario well-designed structure isregarded as a necessary precondition for goodarchitecture. Such an approach requires thatthe structural make-up of a building be evolvedin conjunction with all other aspects of itsdesign and that structural issues be consideredfrom an early stage in the design process andallowed to play as significant a part in thedetermination of the final form of a building asthe aesthetic and space-planning programmes.In this approach the objective of design is todetermine the form in which all requirementsare satisfied equally.

It is possible therefore for structure andarchitecture to be related in several differentways and one of the first decisions which hasto be taken by a design team which has a fullawareness of the activity in which it isengaged, is concerned with the nature of thisrelationship. Often the issue is allowed toremain unclear and some architects even denythat there can be more than one properrelationship between structure and architec-ture. It has been suggested however3 that therelationship can take a number of differentforms and that the totality of possible relation-ships between structure and architecture maybe summarised in the four categories of struc-ture ignored, structure accepted, structure symbolisedand structural 'high tech'. The implications ofthese categories of relationship are nowreviewed.

2.2.2 Structure ignoredIt is possible, principally due to the existenceof structural materials such as steel andreinforced concrete, to invent architecturalform without considering the structural impli-cations of that form. The wide variety of formsinto which steel is fashioned in the manufac-ture of motor cars, ships, marine oil produc-tion platforms and consumer durables, forexample (Fig. 2.3), all of which have shapeswhich are determined principally from criteriaother than those connected with structure,illustrate the almost limitless possibilitieswhich are present in the matter of form when

3 See Macdonald, Structure and Architecture, Chapter 7.

Fig. 2.3 Steel can be fabricated into almost any shape asthe structures which are built for the marine environmentdemonstrate. The construction of a ship requirestechniques of fabrication which are significantly morecomplex and expensive than those which are normallyused in the construction of buildings. There is, neverthe-less, no technical reason why such shapes should not beemployed in architecture. [Photo: P. Macdonald] 25


Fig. 2.4 Sailing Yacht'Gipsy Moth IV. In theconstruction of ships andyachts timber has beenused to produce a widevariety of complex shapes.The material itself placeslittle restriction on formbut the production ofgeometries like that illus-trated is highly labourintensive and thereforeexpensive. [Photo: TheMaritime Trust]

26

the objects concerned are made from steel. Asimilar statement could be made of reinforcedconcrete but in this case the exemplars wouldperhaps be grain silos or water towers. Timbertoo is capable of being used in a very widerange of shapes although, because timber isless strong than steel or reinforced concrete,these are of a smaller scale. Again, ships andyachts serve as examples of the level ofcomplexity of form which is possible (Fig. 2.4).In all of these examples a criterion other thanstructure, such as hydrodynamic performanceor adherence to fashion, was the dominatinginfluence on the form which was chosen.

The structural factor which is common to thethree materials considered above (steel,reinforced concrete and timber) and whichallows them to be shaped into almost anyform, is that they can all resist tension,compression and bending.4 The high strength

They can therefore be used to make any type of struc-tural element: form-active, semi-form-active or non-form-active. The same cannot be said of masonry,which has very limited tensile strength and thereforealso limited bending strength. As is shown in Chapter5, the need to prevent significant tensile stress fromdeveloping imposes constraints on the structural formsof masonry.

of steel and reinforced concrete, together withthe fact that very effective structural joints canbe made between components of these mater-ials, are additional factors which make possi-ble the creation of practically any form.

It may seem surprising therefore that, in theperiod since these materials became widelyavailable to the builder in the late nineteenthcentury, architects have availed themselves solittle of the potential for the free invention ofform which they made possible. Instead, archi-tects have, for the most part, generally contin-ued to produce buildings with plane verticalwalls and horizontal or pitched roofs which arenot significantly different in overall shape fromthe traditional architectural forms which wereevolved from the much more restrictive struc-tural technology of masonry. There have beenseveral reasons for this, most of which werenot technical.

Perhaps the most significant reason for theapparent lack of basic (as opposed to superfi-cial) complexity in the architectural forms ofthe twentieth century has been cultural. In theperiod in which the freedom to experimentwith form has been available it has not gener-ally been fashionable for buildings to be givenirregular or curvilinear forms. For ideologicalreasons, modernity preferred to accept thevocabulary of orthogonality, which symbolised


Fig. 2.5 East Pavilion,Groninger Museum,Groningen, Netherlands,1990-94, Coop Himmelblau,architects. Parts of this build-ing were manufactured in ashipyard using shipbuildingtechniques. It is a relativelyrare example in architectureof the employment of alarge-scale industrial processto create a complex form.[Photo: R. Talbot]

rationality-, of regularity and repetition, whichsymbolised the production line; and of straightlines and sharp edges, which symbolisedmanufacture by machine rather than by handcrafting (e.g. Figs 1.2 and 4.6). Thus, at the verytime when developments in structuralengineering could have released architectsfrom the tyranny of the straight line and theright angle they voluntarily restrictedthemselves to the use of little else. In recentyears the mood has changed, however, andarchitects of the late modern Deconstructionschool, in particular, are now availingthemselves of the freedom which their prede-cessors failed to exploit.

This may be seen in buildings such asthose illustrated in Figs 2.1, 2.2 and 4.20. It isinteresting to note that in the East Pavilion ofthe Groninger Museum (Fig. 2.5), majorelements of the steelwork were manufacturedin a local shipyard using shipbuildingtechniques. This allowed the use of shapesand configurations which, in the context ofbuilding, were exciting and new. It is a matterof conjecture whether this can be takenseriously as a method by which buildingsshould be constructed but it does show thatarchitects are finally making use of the fullpotential of the materials which industry hasplaced at their disposal.

One of the reasons for the continuation, intothe twentieth century, of the relatively simplestructural geometries of architectural traditionwas the elementary one of convenience.Arrangements of rectangular rooms withhorizontal floors and ceilings are moresuitable, for most human purposes, thanenclosures made from curvilinear surfaces orwall and roof planes which intersect at acute oroblique angles. There was no technical reason,however, why buildings with complex forms,with non-vertical walls, inclined roof planesand curvilinear enclosures could not have beenconstructed once materials such as steel andreinforced concrete became available to thedesigners of buildings in the late nineteenthcentury.

Yet another factor, which is not strictlytechnical but which may have inhibited the useof irregular forms, is obviously that of cost.Complicated forms are difficult and thereforeexpensive to construct. This is why artefactssuch as motor cars or aeroplanes are relativelymore expensive than buildings, despite theeconomies of mass production which areassociated with their manufacture. The costper tonne of steel used in a motor car is anorder of magnitude greater than that of a struc-tural steel framework for a building. The sameis true of timber used for yacht construction 27


compared to structural timber in buildings andof the aluminium alloys used in aircraftconstruction. The high relative costs of themotor car, the sailing yacht and the aeroplaneare not due to the basic costs of the materialsbut to the costs of fashioning these materialsinto complex shapes and configurations. Bycomparison with a vehicle, of whatever type,most buildings are very simple objects, beingcomposed of basic components assembledinto relatively simple geometries.

The virtually unlimited range of forms whichis available to the designer of the motor car orthe ship or the aeroplane is therefore at thedisposal of the architect. The factors whichhave mitigated against the use of complexforms have for the most part, as stated above,not been technical and therefore not real, inthe sense of ultimate physical, constraints.

There is, however, one technical factorwhich, in the context of buildings, does inhibitthe freedom to invent form. This is scale, ormore particularly the size of the span involved.If this is small, then the architect can indeedassume virtually unlimited freedom in thematter of form. The larger the span the morerestrictive are the structural constraints onform.

The scale of a building is therefore of criticalimportance if its form is to be determined freefrom any consideration of structural perform-ance. This is because the building will have tocontain a structure with sufficient strength toresist the loads applied to it. If the form hasbeen determined 'freely' in the visual sense(i.e. without consideration of technical implica-tions) it is likely that the structure will besubjected to bending-type internal forceswhich will result in an inefficient use of struc-tural material.5 It is probable also that themagnitudes of these internal forces, for a givenapplication of load, will be high, which willfurther compromise efficiency. If the internal

forces are so high that, even with strongmaterials such as steel or reinforced concrete,the sizes of the cross-sections required tocontain stresses within acceptable limits areexcessive, the structure will not be feasible. Asis stated above the critical factor here is scale,and more particularly span.

This situation may be appreciated by con-sidering the effect of size on any physicalobject. If the object is small - say less than onemetre across (a model of a building forexample) - it would be possible to make it froma weak material such as cardboard or papiermâché. If the span is 10 m (small buildingscale) a stronger material (reinforced concrete,steel, timber, plastic) would be required. Thelimits of what was possible would depend onthe material and on the nature of the form.With reinforced concrete, for example, the limitcould be reached at a span as low as 20 m butmight not be reached until the span was 200 m,depending on the form. If steel were used thelimits would be higher.

Scale is a significant consideration with anytype of structure but it is particularly importantif the relationship between structure and archi-tecture is that of 'structure ignored'. The largerthe span the greater is the possibility that theform will not be feasible. With the materialscurrently available, virtually any shape is possi-ble up to a span of around 30 m. Most shapesare possible up to a span of around 60 m. Forspans greater than 60 m there is an increasingpossibility that a shape which has been chosenwithout regard to structural considerations willnot be feasible. Built forms which have beendetermined without regard to structuralconsiderations are likely to be significantlymore costly to produce, however, than thosewhich have.

To summarise, the architect who wishes todisregard structural considerations when deter-mining the form of a building must be mindfulof the considerations of cost and scale. If anysignificant departure is made from the basicforms of structure outlined in Chapters 3 to 6here, then a cost penalty is likely to beinvolved. If the span is large there is also apossibility that the form chosen will not be

5 It will, in other words, most likely be either a semi-form-active or a non-form-active structure, neither ofwhich is structurally efficient. See Appendix I for theexplanation of this.28


feasible. It would be imprudent, therefore, foran architect to adopt the 'structure ignored'methodology in the context of large enclos-ures, long spans and/or a parsimonious client.

2.2.3 Structure acceptedWhen this type of relationship between struc-ture and architecture is preferred the objectiveis to produce a building in which more-or-lessequal importance is attached to all aspects ofthe design. The aesthetic and programmaticissues must, as always, be brought to asuccessful conclusion but the structure mustalso be considered to be satisfactory whenjudged by purely technical criteria. It must not,in other words, be regarded simply as aprovider of support which occupies the hidden,nether regions of a building and which doesnot, of itself, have to be well designed.

Buildings in which this type of relationshipbetween structure and architecture exists havebeen constructed throughout the entire historyof western architecture and include some ofthe most notable buildings in that tradition(Figs 5.4 and 5.7). The temples of Greek an-tiquity, the massive basilicas and bath housesof the Roman period and the Gothic cathedralsof medieval Europe are prominent examples.These have many equivalents in the modernworld: the buildings by Le Corbusier which arebased on reinforced concrete frameworks, forexample (e.g. the Villa Savoye (Fig. 4.6), theUnité d'Habitation and the monastery of LaTourette), whose rectilinear geometries areideally suited to economical forms of construc-tion in that medium, are the twentieth-centuryequivalents of the Greek temples or the Gothiccathedrals, in terms of the relationship whichwas achieved between structure and architec-ture. The Willis, Faber and Dumas building byNorman Foster (Fig. 4.17) is a more recentexample. The majority of buildings constructedin the present day, though not necessarily themost famous, also fall into this category.

In all of these buildings the structural andaesthetic programmes co-exist in harmony.The structures have not been expressed in anovert way but their properties, requirementsand limitations have been accepted and visual

vocabularies have been adopted within whichthey have been easily accommodated. Theaesthetic and technical issues have, in otherwords, been reconciled.

To achieve this type of integrated designarchitects and structural engineers have towork together from an early stage in the designprocess. For a given size of building, the rangeof structural options which are sensible may berelatively small and an outcome in which allaspects of the design are deemed to have beensatisfactorily resolved will be obtained only if aflexible strategy is adopted with regard to thebuilding's appearance. The aesthetic strategymust, in other words, be compatible with therange of structural options which is mostpractical, given the spans involved.

The criteria by which the quality of a struc-ture can be judged have been discussedelsewhere,6 where it was argued that theprimary technical objective of structural designwas the satisfaction of load-carrying require-ments with maximum economy of means.Well-designed structures were shown to be acompromise between complexity, which isrequired to economise on material, andsimplicity, which allows economy to beachieved in the activities of design andconstruction. The principal factors which deter-mine the optimum compromise betweencomplexity and simplicity were shown to bespan and the intensity of applied load. Thelarger the span, the greater was the level ofcomplexity which was justified, because highlevels of structural efficiency are required toachieve long spans. The greater the appliedload, the lower was the level of structuralefficiency which could be tolerated and thesimpler therefore the structural form which wasappropriate.

Thus, it was seen that simple structuralforms composed of basic elements (timberbeams or reinforced concrete beams and slabswith simply shaped cross-sections) assembledinto basic post-and-beam arrangements areappropriate for short-span structures. These

6 See Macdonald, Structure and Architecture. Chapter 6. 29


represent the best compromise betweencomplexity and simplicity for buildings which,though not necessarily of small scale, have norequirement for large interior spaces andtherefore for long spans. At the other extreme- the very long span - highly complex systemssuch as steel-cable networks or doubly-curvedreinforced concrete shells were seen to providethe best structural solutions. For intermediatespans, intermediate levels of complexity areappropriate.

A happy consequence of regarding theachievement of maximum economy of meansas the principal criterion of good structuraldesign is that it is also the condition which islikely to result in the lowest-cost structure.Although monetary cost is, in important ways,an artificial yardstick, it is in fact related to thetotality of the resources, of all kinds, whichmust be committed to a structure.

A factor which affects the precise level ofcomplexity which is appropriate for a particularstructure is the economic climate in which itwill be built. If, for example, the cost of labouris low in relation to that of materials, which isthe normal situation in a non-industrialisedeconomy, a more complex (and therefore moreefficient) structural form is justified than if thereverse is true. The structure type which ismost appropriate for a particular span in anindustrialised economy, in which labour costsare high, might therefore be simpler than themost appropriate type of structure for thesame span in a 'developing' country.

As was noted above, good examples of well-resolved architecture of the 'structureaccepted' type are found in the early modernperiod. This was due to the happy compatibil-ity of the favoured orthogonal aesthetic withthe post-and-beam structure, which is themost appropriate structure type for therelatively short spans which occur in themajority of buildings.7

To summarise, the building designer whowishes to achieve technical as well as aestheticand programmatic excellence must select atype of structure which is truly appropriate forthe span and load conditions involved. This islikely to have a fairly profound effect on theoverall form of the building and in such a casethe technical and aesthetic agendas must bemade compatible. Because the range of struc-tural options may be limited, particularly bythe span involved, the approach to theaesthetics must be flexible. Where thismethodology is adopted the structural solutionwill inevitably be an adaptation of one of thebasic forms of structure outlined in the follow-ing chapters.

2.2.4 Structure symbolisedOne of the features of buildings designed inaccordance with the 'structure accepted'approach discussed above is that the structureitself rarely constitutes a prominent part of thevisual vocabulary and may even be entirelyhidden from view. The objective of the 'struc-ture accepted' approach is not overly toexpress the structure visually but to ensurethat it will be well designed and wellintegrated with all other aspects of the build-ing. Where the relationship between structureand architecture is of the 'structure symbol-ised' type, the structure is emphasised visuallyand constitutes an essential element of thearchitectural vocabulary.

In the 'structure symbolised' approach thestructure is treated as a set of visual motifsand decisions concerning the size, shape andarrangement of the structural elements areinfluenced as much by visual as by technicalcriteria. The technical performance of thestructure is secondary to its aesthetic role andthe technical quality of the structural design isfrequently compromised as a result. This is aninevitable consequence of this method ofworking and must be accepted if it is used.

In recent architecture the 'structure symbol-ised' approach has been employed almostexclusively as a means of expressing the idea oftechnical progress and has therefore beenassociated principally with the Modern

7 In the context of architectural structures thepost-and-beam arrangement is non-form-active. Forshort spans, this is, in most cases, the best structuralsolution. See Macdonald, Structure and Architecture,Chapter 6 for a discussion of this.30


Fig. 2.6 RenaultWarehouse andDistribution Centre,Swindon, England, 1983.Foster Associates, archi-tects, Ove Arup &Partners, structuralengineers. Many of thefeatures which areassociated with highstructural efficiency arevisible here, such as thesemi-trussing of themain elements, the useof tapered profiles, the1-shaped cross-sectionsand the circular lighten-ing holes. They wereadopted for visualreasons, however, as fewof these complexities canbe justified on technicalgrounds due therelatively short spaninvolved. [Photo: AlastairHunter]

Movement. In the early Modern period, in theworks of architects such as Mies van der Rohe,the use of technical imagery was accomplishedwith restraint and refinement, and confined tothe repetitive use of elements such as parallel-sided I-section beams (see Fig. 1.2). Laterpractitioners of the so-called 'high-tech' genreused a much richer palette of structural images,many of which were 'borrowed' from the fieldsof vehicle and aeronautical engineering ratherthan from structural engineering. The purposewas to create an architectural style which wascelebrative of the idea of technical progress andof the condition of modernity.

In the high-tech style the 'structure symbol-ised' approach resulted, with a few notableexceptions, in an architecture which was basedon steel framework structures, principally

because steel was the only structural materialwhich possessed the necessary high-techimage and which also provided a large enoughnumber of different element and componenttypes to give the architectural language areasonably large vocabulary.8

The use of steel frameworks in this waymeant that the constraints which are associ-ated with steel had to be accepted and the

It is significant that one of the very few buildings of thistype to have a reinforced concrete structure - theLloyd's Headquarters Building in London -wasintended originally to have a steel frame and wasdetailed as though the structural material were indeedsteel. This provides a good illustration of the fact that,contrary to the claims which are often made for it, thisis not a structurally 'honest' type of architecture. 31


building forms which resulted therefore had tohave rectanguloid geometries with regularlayouts of elements. The shapes of the indi-vidual structural elements were more complex,however, than would have been specified forequivalent structures under the 'structureaccepted' philosophy and the various deviceswhich are associated with high structuralefficiency were freely employed to add visualinterest. These included complex shapes incross-section and longitudinal profile, thetapering of elements, the liberal use of thecircular lightening hole (used in aircraft struc-tures to reduce the weights of components)and the exaggeration of joints and fasteningcomponents (Fig. 2.6). Much of this vocabularywas 'borrowed' from aerospace or motortechnology, where its use was, of course, justi-fied on purely technical grounds. In the field ofarchitecture, its employment was purely stylis-tic but was necessary to convey the impressionof 'state-of-the-art' technology.

The need to expose the structure to view,which is an inevitable consequence of its useas a set of visual motifs, carried with it arequirement to provide adequate corrosionprotection schemes, which added to the struc-tural costs, especially if, as was often the case,the steelwork was exposed on the exterior aswell as in the interior of the building. Wherethe building had more than one storey, it wasalso necessary to make adequate provision forfire protection. This posed quite severeproblems which were often difficult to resolve9

and which frequently required the adoption ofone of the more complex forms of fire protec-tion (see Section 3.5).

One of the ironies of the high-techmovement was that when structural form wasmanipulated according to the dictates ofaesthetic rather than technical criteria, theresult was often technically flawed and it is afact that most of the high-tech buildings were

not well engineered if judged purely on theirtechnical performance.10 That which symbol-ised technical excellence was not, in otherwords, itself technically excellent. This,however, was an inevitable consequence of thesymbolic use of structure.

It should be noted, however, that the 'struc-ture symbolised' approach is capable of beingapplied to architectural agendas which aredifferent from the Modernist preoccupation withthe celebration of technology. If, for example,the intention was to celebrate the idea of asustainable architecture the use of the 'structuresymbolised' method would yield a quite differ-ent architecture in which attention was drawn tothe measures which are required to minimisethe input of energy and resources, of all kinds,which are involved in the construction, main-tenance and running of a building. Such anapproach could lead to the development of acompletely new architectural symbolic vocabu-lary of masonry and timber. This would, ofcourse, also be equally 'dishonest' and probablyas basically inefficient, in structural terms, aswas 'high tech'. This development is thereforeperhaps less likely to occur than was 'high tech'because the potential clients would, due to itslack of engagement with physical reality, be lesssympathetic to the idea of adopting it and morelikely to favour the 'structure accepted' or truestructural high-tech' approach.

Where the 'structure symbolised' approach isused it is essential that the designer has a clearawareness of the type of architectural statementwhich the building is intended to make. Theform of the structure must then be distortedand exaggerated so as to make the nature ofthat statement unequivocally clear otherwisethe result will be architecturally feeble.

2.2.5 True structural high techIn this, the fourth type of relationship betweenstructure and architecture, the design of thestructure is accorded the highest priority and

9 It was in fact due to the impossibility of meeting thefire regulations that the change of structural materialfrom steel to reinforced concrete was made in the caseof the Lloyd's building.

10 See Charles Jencks, The battle of high tech: great build-ings with great faults', Architectural Design, 58, (11/12) pp.

19-39, 1988.32


allowed to determine completely both theoverall form of a building and the nature of thearchitectural vocabulary which is adopted. It isa method which is normally used for reasonsof necessity, when the limits of what is pos-sible structurally are being approached.Obvious examples of this are the very tallbuilding and the very long-span enclosure. Inthe case of the former, the principal structuralproblem is the resistance of lateral load, inwhich case the systems required to accommo-date this become prominent features of thedesign." The achievement of a long spanrequires that a highly efficient type of structurebe adopted, such as a steel cable network or areinforced concrete shell. These form-active12

structure types have distinctive geometrieswhich therefore dictate the overall form of thebuilding.

Another example of a building type in whichthe highest priority must be given to structuralmatters is the portable building. Here there isa requirement for demountability and also acritical need to save weight and therefore toadopt an efficient form of structure. Form-active structures are therefore frequently speci-fied for this type of building. The tent, in all itsmanifestations, provides an example of this.Demountable buildings for travelling exhib-itions (Fig. 1.12) and any other type of tem-porary accommodation are further examples.With this type of building very few concessionsare made to style, and none which conflict withefficiency.

The 'true structural high-tech' approach canbe applied to the design of any building.Architects rarely favour it, however, and willrarely use it unless technical necessity makes itunavoidable, because it leaves them with littleto do other than administrate the constructionof the building. It leaves, in other words, little

scope for artistic statement. The methodologyrequires that no concessions be made withregard to the design of the structure and theresulting architecture might be described as atype of vernacular. For the majority of build-ings, in which spans and loads are modest andthere are no special requirements which favourthe adoption of a highly efficient form of struc-ture, the best structural solution is likely toinvolve the use of a very simple type of struc-ture such as a post-and-beam form composedof basic elements such as masonry walls orreinforced concrete slabs and columns.

The 'true structural high-tech' approach isthe most straightforward of the possiblerelationships between structure and architec-ture. The preliminary design of the buildingbecomes simply the design of a structuralarrangement which is appropriate for the spanand load involved. In most cases this willfavour the selection and adaptation of one ofthe basic forms of structure outlined in thefollowing chapters. Aesthetic, space-planningand other considerations are given a secondaryrole and are not allowed to compromise theintegrity of the structural solution.

33

2.2.6 Conclusion

The various possibilities concerning therelationship between structural design andarchitectural design have been reviewed in thissection. The distinctions considered have onlybeen possible in the twentieth century follow-ing the development of the structuraltechnologies of steel and reinforced concrete -the strong materials which released architec-ture from the constraints imposed by thetechnology of structural masonry. It was notuntil this had occurred that the methodologiesof 'structure ignored' and 'structure symbol-ised', which have a tendency to generate struc-tural geometries which are far from ideal,became possible.

The distinctions between the fourapproaches outlined above to the relationshipbetween structure and architecture arefrequently misunderstood, by both architects

11 In all of the very tall buildings the structure which isrequired to resist wind loading is concentrated on theexterior of the building and therefore affects its appear-ance.

12 See Appendix 1 for an explanation of the term'form-active'.


and critics, especially in connection with thosestrands of Modern architecture in which theidea of drawing attention to the tectonicaspects of a building has been fashionable.The confusion which arises concerns theallocation of priorities between technical andnon-technical issues. If technical issues have,in reality, been allocated the highest prioritythe building will fall into the category of 'truestructural high tech'. If aesthetic consider-ations have been given a higher importancethen the building will be an example of 'struc-ture symbolised' or 'structure ignored'.

The distinction is brought into focus by aconsideration of the consequences of allowinga high design priority to be given to technicalconsiderations ('structure accepted' or 'struc-tural high tech'). If the structural problem isspectacular, such as a very long span, theresulting structure, and therefore building, willalso be visually striking. If the structuralproblem is modest - a building of small ormedium span - the best structural solutionwill almost certainly also be modest and of thepost-and-beam kind. In the early years of archi-tectural Modernism - the 1920s and 1930s -such forms were compatible with the prevailingaesthetic theories and, as a consequence,many early modern buildings are goodexamples of the 'structure accepted' approach.In the present day, post-and-beam forms arefrequently considered to be visually dull and inthis situation the temptation arises tomanipulate the structure for visual or symbolicreasons ('structure symbolised') or to ignore itsrequirements entirely ('structure ignored').

As is discussed above, the architecturalsymbolists of the so-called 'high-tech' school,who have often claimed that the structures oftheir buildings are examples of genuine techni-cal excellence,13 provide a good example of thetype of unclear thinking which has surroundedthis topic. The confusion has led, in manycases, to the creation of buildings which have

an unresolved quality, because the full poten-tial offered by the purely symbolic use of struc-ture has not been exploited. A betterarchitecture would probably have resulted ifthe true nature of the relationship betweenstructure and architecture had been more fullyappreciated and acknowledged.

This last statement is generally true: thefinal outcome of an architectural designprocess is more likely to be satisfactory if thearchitect is fully aware of the nature of therelationship between technical and aestheticissues.

2.3 Selection of the generic type ofstructure

2.3.1 IntroductionStructures for buildings may be placed into thethree broad categories of 'form-active', 'semi-form-active' and 'non-form-active'.14 In thecontext of gravitational loading, which is theprincipal form of load on most architecturalstructures, post-and-beam structures are non-form-active and can be further subdivided intothe two categories of loadbearing-wall struc-tures and skeleton-frame structures. It is fromthis limited range of possibilities that thestructure for a building must be selected.Within each category an almost infinite varietyof structural possibilities exists, however,depending on the types of element which arespecified and the manner in which these areconnected together.

It is important to recognise that the processof structural design is not so much one ofinvention as one of selection and adaptation. Mostnew structures are in fact versions of one of arange of basic structural forms which haveevolved in practice as the best arrangementsfor the particular material concerned. Thesebasic forms of structure are reviewed inChapters 3 to 6 and they represent the vocabu-lary of structural form from which the designer

13 The late modern deconstruction architects have notbeen troubled by this particular item of theoreticalbaggage.

14 See Appendix 1 and Macdonald, Structure andArchitecture, Chapters 4 and 5.34

must begin the process of selection andadaptation. An insight into the working of thisprocess is given in Section 2.5.

An important factor in the selection of astructure type for a building is the nature ofthe relationship which has been adoptedbetween structure and architecture. If this is inthe 'structure ignored' category, then theoverall form of the building, and therefore ofthe structure, will have been determinedwithout consideration of structural require-ments and the structure will most probablyhave to be of the semi-form-active type. This isbecause, for a given load condition, form-active and non-form-active structures haveunique geometries. A structure whose geom-etry has been determined without regard tostructural considerations is therefore likely tohave a semi-form-active shape. In this situ-ation the choices available to the structuraldesigner may be very limited.

The structural material will almost certainlyhave to be either reinforced concrete or steelbecause a material with the ability to resistsignificant bending will be required. The choicebetween these two materials will be influencedby the nature of the building. Concrete resultsin a more massive structure but allows the easycreation of curvilinear forms. Steel is strongerand is therefore capable of producing a lighterstructure (both literally and in appearance) butmust normally be used in the form of a skele-ton framework. The distinction between theirrespective suitabilities can be seen by compar-ing the Vitra Design Museum building by FrankGehry (Fig. 4.20) with the rooftop office inVienna by Coop Himmelblau (Fig. 2.1).

In the Vitra Design Museum building thepresence of strongly curvilinear forms, ofcantilevered volumes of complex shape and ofthe limited number of openings in the build-ing's envelope favoured the use of reinforcedconcrete. The mouldability of the material andthe level of structural continuity which itallows are particularly useful properties. In theRooftop Office in Vienna the extensive areas offaceted glazing favoured the adoption of askeletal structure of slender steel elements.The minimal amount of curvature in the roof

structure was accommodated by steel and thegeneral arrangement of the structure, despiteits apparent novelty, conforms to the fairlystandard primary/secondary beam format.

If the relationship between structure andarchitecture is other than 'structure ignored' itwill be necessary to select a structure typewhich is compatible with the aesthetic andprogrammatic aspects of the building andwhich is also sensible from a structural pointof view. The choice of structure type should,however, be larger because there should be areadiness to adjust the aesthetic programmeto accommodate structural requirements andthus produce a building with a structure whichis satisfactory technically. The factors on whichthis depends are now reviewed.

2.3.2 The effect of scaleThe span of an architectural structure, which isdetermined by the required sizes of the spaceswhich are enclosed by it, has a very significanteffect on both the generic type of structurewhich should be used and on the selection ofthe types of structural element of which itshould be composed. The underlying principlewhich governs the relationship between spanand structure type is that the ratio of self-weight to load carried should be satisfactory.For a given type of structure the strength-to-weight ratio tends to become less favourableas the span increases.15 More efficient types ofstructure must therefore be specified as spansare increased to maintain the ratio at anacceptable level.

Table 2.1 shows the ranges of span for whichbasic types of structure are most suitable andtherefore normally specified. The figures mustbe regarded as approximate but serve to givean indication of the applications for whicheach is most appropriate. Separate figures areprovided for floor and roof structures to allowfor the variations which are caused by thesignificantly different levels of gravitationalload to which they are subjected.

15 See Macdonald, Structure and Architecture, Chapter 6. 35


Table 2.1 Normal span ranges for commonly used structural systems

A Timber structures

Structural system Normal span range (m) Span/depth ratio

softwood planks floor 0.6-0.8 25-35roof 2-6 45-60

plywood board floor 0.3-0.9 30-40roof 0.3-1.2 50-70

softwood timber joist floor 3.5-6 12-20roof 2-6 20-25

stressed skin timber panel floor 3-6 20-30roof 3-9 30-35

laminated timber beam floor 5-12 14-18roof 4-30 15-20

plyweb beam floor 5-18 8-10roof 6-20 10-15

trussed rafter roof 5-11 4-6parallel chord truss roof 12-25 8-10gluelam portal frame roof 12-35 30-50gluelam arch roof 15-100 30-50lattice dome roof 15-200 40-50

B Steel structures


profiled decking floor 2-3 35-40roof 2-6 40-70

profiled decking with composite concrete topping floor 2-6 25-30beam (hot-rolled section) floor 6-25 15-20

roof 6-60 18-26cold-formed open-web joist roof 4-30 15-25hot-rolled triangulated parallel-chord truss floor 12-45 4-12

roof 12-75 10-18pitched, triangulated truss roof 25-65 5-10space deck roof 10-150 15-30braced barrel vault roof 20-100 55-60portal framework roof 9-60 35-40cable-stayed roof roof 60-150 5-10hanging cable roof roof 50-180 8-15

C Reinforced concrete structures


one-way span solid slab reinforced 2-7 22-32prestressed 5-9 38-45

two-way span solid slab reinforced 4.5-6 30-35one-way span ribbed slab reinforced 4-11 18-26

prestressed 10-18 30-38two-way span coffered slab reinforced 6-15 18-25

prestressed 10-22 25-32beam/column frame with beam/slab floor slab 3.5-6 30-36

beam 6-12 15-20portal framework 12-24 22-30arch 15-60 28-40


36


All structure types have a potentialmaximum span and the less efficient thestructure, the lower is this practicalmaximum span. An indication of this is givenin Table 2.1 in which it will be observed thatthe maximum span given for each type ofstructure is related to its potential efficiency.Thus, elements with simple, solid rectangularcross-sections, such as sawn timber joists orrectangular cross-section reinforced concreteslabs, have relatively low maximum spans.Simple elements with 'improvements', suchas l-section steel beams or triangulatedtrusses of timber or steel, have highermaximum spans. The highest spans areachieved by highly efficient vaulted shells orcable networks.

The effect of the variation in the maximumspan potential of different types of element isthat the choice of element types which is avail-able to the structural designer diminishes asthe span increases. If the span to be achievedis small (say 5 m) virtually all structure typesare available. At this scale the designer couldtherefore choose to use any type of structurefrom simple, solid beams or slabs to sophisti-cated forms such as the arch or the vault. Inthe context of contemporary architecture, itwould probably be regarded as technicallyinappropriate to use a complex form for ashort span, unless some special requirementfor high efficiency existed, because muchsimpler post-and-beam forms would performadequately. It would nevertheless be a choicewhich the architect could make. As the spanincreases the number of different types ofstructure which would be viable decreasesuntil, at the very long span (say 200 m), onlythe most efficient form-active types, such assteel cable networks or thin concrete shells,are feasible. In summary, from the point ofview of the designer, the choice of structuretype is large if the span is small and becomesprogressively more limited as the spanincreases.

The most basic post-and-beam types ofstructure (loadbearing-wall arrangements inmasonry or timber with simple horizontalelements such as timber beams or reinforced

concrete slabs) are suitable for short-spanstructures in the 5 m to 10 m range. The spanrange can be extended by the use of moreefficient types of horizontal structure such astriangulated trusses in either timber or steel.The use of walls with 'improved' cross-sections(fin wall or diaphragm wall) allows the verybasic structural system to be used for largerenclosures with high external walls, such assports halls, where spans of up to 30 m havebeen achieved.

The post-and-beam frame (in eitherreinforced concrete or steel) is a more sophis-ticated and therefore more flexible system thanthe loadbearing wall. In the multi-storeyversion the span range is slightly more exten-sive than that for loadbearing-wall structures(5 m to 20 m). The most basic types of elementare used at the low end of the range (solidreinforced concrete slabs, rolled-steel sections)and at the upper end more efficient types suchas coffered, reinforced concrete slabs andhollow-web steel beams. Spans greater than20 m are unusual in multi-storey buildings butwhere these occur, efficient types of structuralelements must be specified (e.g. triangulatedgirders used to achieve a span of around 25 mat the Centre Pompidou in Paris).

In single-storey structures the change fromthe most basic forms to more efficient struc-ture types (e.g. space frame horizontal struc-tures or semi-form-active structures) normallyoccurs when the span is in the range 20 m to30 m, with spans greater than 30 m usuallyrequiring the use of a semi-form-active struc-ture such as a portal framework. As with short-span structures, the type of element which isused can vary and the tendency is always forthe level of efficiency (and therefore ofcomplexity) to increase as the span increases.Thus, short-span portal frameworks (20 m)would normally be accomplished with rolled-steel sections such as the universal beamwhile a triangulated-truss longitudinal profilemight be specified for a longer-span version(say 50 m).

The transition from the semi-form-active tothe highly efficient fully form-active structureoccurs in the range 40 m to 60 m. 37


2.3.3 The effect of cost example by columns. There have, of course,Another factor which influences the choice of been exceptions to this, the 'prairie houses' ofstructure type for a particular span is cost. Frank Lloyd Wright being good examples ofStructural efficiency is achieved by structural interiors with free flowing internal spacescomplexity which is costly to produce. The being accommodated within loadbearing-wallbuilder of the long-span structure has no structures. These buildings are of relativelychoice but to use an expensive, complex type small scale, however, and the free-flowingof structure. The builder of the short-span spaces exist within patterns of walls which arestructure does have a choice: both complex fairly regular (Fig. 2.7).and simple types will be feasible, but the The degree of regularity of a plan affects thecomplex structure will be more expensive. The preferred type of structure. As is mentioneduse of the latter, in the cost-conscious environ- above, loadbearing-wall structures mustment of the late twentieth century, would normally be given a regular arrangement (seetherefore normally have to be justified on Chapter 5) and the same is true of frames,some other grounds such as a particular which perform best if based on a regularrequirement for a lightweight structure. The column grid. If the pattern of vertical supportreason for choosing a complex form might, of is not regular, this is best accommodated bycourse, be simply a desire for the appearance the adoption of a system for the horizontalof complexity or extravagance. Depending on structure which has the capability to span inthe priorities of the client, therefore, cost more than one direction. For floor structuresmight or might not be a significant factor in the simplest element of this type is the two-determining the structural form which is way-spanning slab of in-silu reinforcedadopted. concrete. For roof structures it is the fully

triangulated space deck.2.3.4 Internal planning Although both of these structural types workThe nature of the plan which is intended for a best when provided with a regular pattern ofbuilding will normally influence the choice of vertical support they can, due to their staticalstructure. The principal factors which affect indeterminacy,16 be supported on irregularthis are the degree to which the interior is arrangements of columns or loadbearing walls,subdivided, the degree of regularity required in The greater the degree of irregularity thethe internal planning and the degree of en- stronger, and therefore deeper, must theseclosure which is envisaged. elements be.

The extent to which a building will be If the irregularity of the design of a buildingcompartmentalised normally has a consider- extends to a variation in plan between differentable influence on the choice of structure type. floor levels, this is best accommodated by theMulti-cellular buildings lend themselves to adoption of a frame-type structure. The wallsloadbearing-wall-type structures, subject to the which subdivide the interior are then non-provisos that the individual compartments are loadbearing partitions which are simply carriednot too large, that a reasonably regular pattern by the floors and whose location in the planof walls can be adopted and, in multi-storey can be varied between levels. If the variation inbuildings, that the plan is more or less the plan between levels is such that the continuitysame at all levels. If any of these conditions of columns through different levels is not

16 See Macdonald, Structure and Architecture, Appendix 1 fora discussion of statical indeterminacy in relation tostructural design.

cannot be met, then a frame structure willnormally be required even if the plan iscompartmentalised.

If the interior of a building is to consist oflarge areas of open space a frame structure willnormally be required even if the space can beinterrupted by some vertical structure - for38


Fig. 2.7 Plan of the Winslow House, River Forest, near Chicago, USA, 1893. Frank LloydWright, architect. The open plan of the Winslow House is accommodated within aconventional arrangement of parallel loadbearing walls (see Fig 5.23) with lintel beamsspanning the large openings between the principal spaces. The small scale of the build-ing together with the relatively large areas of vertical structure which are provided haveallowed the open interior to be achieved without the use of an expensive frameworkstructure. 39


possible, this must be achieved by the use ofsubstantial horizontal structure at the levels atwhich discontinuities in vertical structureoccur. A deep structure will be required,however, and appropriate provision for thismust be made.

If the form of a building is highly irregular inplan, elevation or section an unconventionalform of structure will be required. This willnormally have to be based on either reinforcedconcrete or steel - see above.

2.3.5 Exterior treatmentThe treatment which is envisaged for theexterior of a building can affect the choice ofstructure type, especially if large areas of thewalls will either be glazed or covered with acladding material which is incapable of actingas part of the support structure. Full glazing ofthe exterior walls, in particular, is normallyassociated with a skeleton-frame structure. It isworth noting, however, that this type of treat-ment of the exterior need not, by itself, neces-sarily dictate that a frame structure be adoptedas it is possible to arrange the plan of aloadbearing-wall building in such a way thatthe exterior walls are non-loadbearing (e.g.with a cross-wall-type plan). Full glazing of theexterior of a building is, however, normallyassociated with a relatively open plan forwhich a frame structure would in any case beappropriate.

2.3.6 ConclusionThe selection of the generic type of structure istherefore based on a number of factors, themost important of which are: scale, whichdetermines the spans involved; internalplanning which dictates the nature of the inter-nal spaces required; and external treatment.The structural choices available are form-active, semi-form-active and non-form-activestructures, in all their manifestations. The finalselection will be determined by the issues ofcost and technical feasibility, given therequirements of scale, space planning andexternal treatment.

17 Not because the material is inherently stronger butbecause the volume of the structure itself, for a giventhickness of wall, is greater. The timber wall will consistof a series of closely spaced vertical elements alternat-ing with voids. The masonry wall is continuous andtherefore contains more structural material fora giventhickness.

18 Traditional half-timbered buildings were frequentlyhigher (six or seven storeys), especially in the towns ofnorthern Europe. These were based on very largetimber sections, however. This type of structure is veryrare in the present day.

2.4 Selection of the structuralmaterial

The choice of the structural material is anotherfundamental decision in the planning of astructure. It is an aesthetic as well as a tech-nical decision. Each of the four principalstructural materials (steel, reinforced concrete,masonry and timber) produces a building witha distinctive visual quality.

The preferred generic type of structure willalso affect the choice of structural material andindeed these two issues are frequentlyresolved together. If, for example, a loadbear-ing-wall structure is adopted this will favourthe use of either masonry or timber.Reinforced concrete and steel would be thenormal choices for skeleton-frame structures.

If a loadbearing-wall structure is being usedfor a domestic-scale building there may belittle to choose between masonry or timberfrom the point of view of structural action. Amasonry wall will normally have greaterstrength in relation to compressive load,17

which means that the use of masonry is essen-tial if the building has more than threestoreys.18 It also favours the use of masonry insituations where internal spaces are large(greater than 8 m across) and in other situ-ations, such as the adoption of an irregularpattern of loadbearing walls, in which theloads applied to individual walls are likely tobe high. Other considerations, such as localavailability of materials, speed of erection orany requirement for the prefabrication of the

40

structure are, however, more likely to influencethe choice of structural material than thestrength required.

Skeleton-frame structures can beconstructed in steel, reinforced concrete ortimber. The use of timber is rare for multi-storey buildings, however, and the timberskeleton frame is most usually associated withsingle-storey enclosures in which the timberstructure is exposed to view. Often the struc-tural elements are of spectacular appearance(large triangulated trusses) and the structuretherefore contributes significantly to theappearance of the building. The relationshipbetween structure and architecture mustnormally be of the 'structure accepted' or 'truestructural high-tech' type rather than the'structure symbolised' type, when timber isused for skeleton frames. This is because thereis normally insufficient reserve of strength toallow the structural performance to be com-promised for purely visual effect.

Reinforced concrete is rarely used for single-storey buildings because the high self-weightof the structural elements renders themunsuitable for situations in which the imposedloads are small. For the same reason,reinforced concrete is rarely used for roofstructures in multi-storey buildings except forflat roofs, in the form of terraces or roofgardens, to which access is permitted andwhich are therefore required to carry greaterthan normal roof loadings. Steel is suitable forboth multi-storey and single-storey buildings.

The normal choices available to the designerof a skeleton-frame structure are thereforebetween steel and timber for single-storeyframes and between steel and reinforcedconcrete for multi-storey frames. The relativeadvantages and disadvantages of the materi-als, from a structural point of view, areoutlined in Chapters 3, 4 and 6.

The normal reasons for the selection of steelare its high strength, its appearance (especiallyin the context of the 'structure symbolised'approach) or the speed of construction whichit allows. The first of these is obviously impor-tant if long spans are involved or if there is aneed to produce a structure which is of low

volume and which has slender elements.Appearance is frequently also the principalreason for the selection of timber for a skele-ton framework although the lightness of theelements can also be a factor. Timber may alsobe selected due to its durability, for examplewhere hostile internal environments, such asoccur in swimming pool buildings, wouldcause corrosion of a steel structure.

Reinforced concrete can be selected for anumber of reasons. Often, it will provide thelowest cost structure due to both the lowinherent cost of the material and the elimina-tion of the need for additional finishing materi-als or fire proofing. The good durability ofconcrete is one of its considerable advantages.Concrete may also be selected due to theopportunities which it provides in relation toform. The mouldability of the material and theease with which it allows structural continuityto be achieved favour its use in situations inwhich complex structural geometries are to beproduced, especially if these involve curvilinearforms.

All materials impose constraints as well asprovide opportunities. Thus, unless the 'struc-ture ignored' method is being practised, thelimitations of the material must normally berecognised once it has been chosen. Thenature of these constraints is something whichis taken into account during the selection ofthe material to ensure that it is compatiblewith the architectural intentions.

2.5 Determination of the form of thestructure

The final form and general arrangement of astructure is normally an adaptation of one ofthe basic forms described in Chapters 3, 4, 5and 6. This is now discussed in relation toreinforced concrete structures to illustrate howthe process of adaptation is carried out toproduce buildings in which structural andarchitectural requirements are reconciled. Themethodology described is applicable to allstructural materials. 41



Fig. 2.8 Plan of the Willis. Faberand Dumas office, Ipswich,England, 1974, Foster Associates,architects, Anthony HuntAssociates, structural engineers.The floors of this building arereinforced concrete coffered slabswhich are supported on a conven-tional square column grid to giveoptimum two-way-spanningcapability. Small-diametercolumns at close spacings areprovided around the perimeter tosupport those parts of the planwhich project beyond the maincolumn grid. The complex curvi-linear plan-form has beenachieved by a relatively minoradaptation of the basic arrange-ment for two-way-spanning slabs(see Fig 4.47).

The Willis, Faber and Dumas building (Fig. ideally suited to deep-plan office buildings.4.17) is a good example of a two-way-spanning The flat-slab structure depends for strength onflat slab structure of reinforced concrete. The structural continuity and performs best inbuilding has three storeys and a very deep plan situations in which several structural bays areconsisting mostly of open-plan office accom- provided in two orthogonal directions. It is amodation. The requirement for the latter ruled form of structure which is suitable for relativelyout the possibility of using a loadbearing-wall high levels of uniformly distributed imposedstructure. The curvilinear plan of the building load as in the standard pattern of officefavoured the use of reinforced concrete rather loading. The absence of deep beams in flat-than steel as the structural material. slab structures facilitates the creation, by use

The floor slabs are supported on a square of a false ceiling, of a continuous services zonecolumn grid which conforms to the standard under the floor structure and the straightfor-plan for this type of two-way-spanning struc- ward layout of reinforcement and relativeture (Fig. 4.47). It is a configuration which is simplicity of the structural form makes the42


Fig. 2.9 Florey Building,Oxford, England, 1967-71.lames Stirling, architect,Felix Samuely & Partners,structural engineers. Thisbuilding has a fairly complexform which was realised by arelatively straightforwardadaptation of a standardreinforced concrete framingsystem. [Photo: PMacdonald]

construction process simple and thereforeinexpensive.

In the Willis, Faber and Dumas building thecurvilinear plan-form conflicted with thesquare column grid and the problem wasresolved by the use of a series of small-diameter perimeter columns (Fig. 2.8). Thistechnique by which the incomplete squares ofslab in the main structural grid were providedwith the support which they required wasmade possible by the two-way-spanningcapability of the floor slab. The relatively largespan involved (16 m) justifies the use of thecoffered version of the flat slab. The distinctiveform of the Willis, Faber and Dumas building

was therefore made possible by a skilfuladaptation of one of the basic forms ofreinforced concrete structure.

The Florey Building in Oxford, England (Figs2.9 and 2.10) has a reinforced concrete framestructure with a one-way-spanning floor deck.This is a university hall of residence with amulti-cellular interior and it is also a relativelysmall building. These factors would normallyfavour the use of a loadbearing-masonry struc-ture but the distinctive cross-section of thebuilding, together with the requirement for anopen plan on the ground floor, resulted in theselection of a skeleton framework in this case.The relatively complex form in both plan and 43


44

Fig. 2.10 Structural plan of the Florey Building, Oxford,England, 1971. lames Stirling, architect, F.J.Samuely &Partners, structural engineers. The structural plan of thisbuilding shows the layout of the primary elements, whichare beam/column frameworks of in situ reinforced concrete.These carry a one-way-spanning floor system. As in thecase of the Willis, Faber and Dumas building a complexform has been achieved by the straightforward adaptationof a basic structural arrangement (see Fig 4.52).

cross-section favoured the use of reinforcedconcrete rather than steel.

The choice of a beam/column frameworkrather than a flat-slab structure was sensible fortwo reasons. Firstly, the depth of the building isthat of a single structural bay. The structuralbenefit of continuity over several structuralbays, which is desirable for the efficient workingof a flat-slab structure, was therefore not avail-able. The crescent-shaped plan was alsoproblematic in this respect because it reducedthe level of structural continuity which could beachieved in the along-building direction.Secondly, the cross-section of the building, andin particular the outward cant of the rear wall,favoured the use of rigid frames consisting ofwide columns and deep beams. A beam/columnframework in reinforced concrete was thereforethe logical choice of structure for this building.

The crescent-shaped plan was produced bysimply distorting the basic plan-form of theone-way-spanning frame (Fig. 4.52) to create a

Fig. 2.11 New City Library, Munster, Germany, 1993.Architekturburo Bolles-Wilson, architects, Buro Thomas,Buro Menke & Kohler, structural engineers. The photo-graph shows the edge of one of the two parts of this build-ing. The main part of the building is to the right and has areinforced concrete frame structure. One of the lines ofvertical structure is seen and it will be noted that this is, ineffect, a loadbearing wall which has been punctured inplaces by large openings to leave pillars of varying cross-sectional shape. The non-structural cladding wall seen onthe left of the photograph is supported by a lean-toarrangement of cranked laminated timber elements.[Photo: E. & F. McLachlan]

structural plan which was in fact an alternatingseries of triangles and rectangles. The arrange-ment, and in particular the triangular parts ofthe plan, were more easily constructed inreinforced concrete than would have beenpossible with steel. The mouldability ofconcrete was also exploited to create thedistinctive cross-section of the building. Itsdurability allowed the elimination of additionalfinishing materials over most of the structuralsurfaces.

45

Fig. 2.12 New City Library, Munster, Germany, 1993.Architekturburo Bolles-Wilson, architects, Buro Thomas,Buro Menke & Kohler, structural engineers. The plans ofthe two parts of this building are shown here. Althoughthese have a relatively complicated form the structuralarrangement is conventional. In each case there are fourlines of vertical structure which run longitudinallythrough the building, one in each perimeter wall and twoin the interior. At the perimeter these are literally wallswith holes punched in them for doors and windows. Inthe interior they are lines of columns. The floors spanbetween these lines of vertical structure.



Fig. 2.13 Exhibition andAssembly Building, Ulm,Germany, 1993. RichardMeier, architect. Therelatively complex form ofthis building was accommo-dated by a structure whichwas a relatively straightfor-ward adaptation of one ofthe standard reinforcedconcrete arrangements.[Photo: E. & F. McLachlan]

46

The distinctive form of this building wastherefore made possible by an imaginativeexploitation of the structural possibilities ofreinforced concrete which involved a relativelysimple modification of one of the standardplan-forms for multi-storey reinforced concretestructures. The structure stands up well topurely technical criticism and the building fallsinto the category of 'structure accepted'.

A similar relationship between structure andarchitecture is seen in the Public Library

Building at Munster in Germany by Boles andWilson (Fig. 2.11). The building is in two linkedparts which are separated by a pedestrianstreet. Each part has a relatively straightfor-ward section but is more complex in plan (Fig.2.12). The juxtaposition of solid and void onthe exterior walls is also fairly complicated.Scrutiny of the plans reveals, however, that thestructural make-up of the building is straight-forward. In each part a reinforced concreteframe has been used. In the case of the part

Fig. 2.14 Plan at level threeof the Exhibition andAssembly Building, Ulm,Germany, 1993. Richard Meier,architect. The structural plan ofthis building is based on a 9 mgrid. Columns are placed onthe grid points and at thelocations where the grid linesintersect the perimeter. Beamsrun between these exceptwhere this would carry themacross a void. The structuralarrangement is simply anadaptation of the basic two-way-spanning skeleton framelayout (Fig 4.51b).

with the curved exterior wall, for example, fourlines of vertical structure are used to supportthe floors, one in each perimeter wall and tworunning the length of the interior of the build-ing. These lines of vertical structure form aseries of rigid frameworks between which thefloors span. The structural make-up of theother half of the building is similar. Thesepost-and-beam frameworks are thereforeadaptations of the standard one-way-spanningframe plan and constitute a sensible structuralsolution in view of the spans and loadingsinvolved. The relationship between structureand architecture is one of 'structure accepted'.

The Exhibition and Assembly Building atUlm in Germany by Richard Meier (Fig. 2.13) isyet another example of a complex form whichis based on a reinforced concrete framework.As in the case of the Munster Library the build-ing is in two linked parts. The more compli-cated of these has a plan which is based onconcentric circles. The arrangement is furthercomplicated by the juxtaposition of solid andvoid which occurs in both the external wallsand in the floors of the building.

The structure is based on a 9 m span grid(Fig. 2.14). Columns are placed on the gridpoints and at the locations where the gridintersects the concentric circles which definethe building's perimeter. Beams run betweenthe grid points, except where they would crossvoids, and carry the floor slabs. This produces apattern of support which is sufficient to carry

all areas of the floors. The structural continuityof reinforced concrete is required, however, toaccommodate the complex plan-form of thefloors. The square column grid would normallybe used in conjunction with a flat-slab floorstructure. The use of deep beams in this casewas due to the lack of structural continuitycaused by the large areas of void in the floorstructures. The degree of adaptation of thestandard form was therefore greater in thisbuilding than in the examples described above.

2.6 Conclusion

The buildings described above give an insightinto the process by which the design of anarchitectural structure is carried out. Thevarious aspects of the decision-making processby which the form of structure for a building isdetermined may be summarised as follows.First, there are several significant global factorswhich influence the choice of structure type.Perhaps the most important group of these arethe architectural objectives being set by thearchitect(s). The building form which resultsfrom the deconstructionist approach followedby, for example, Coop Himmelblau, is likely tobe very different from those whose approach iscloser to the modernist mainstream, such asMeier or Boles and Wilson. Another form islikely to result from high-tech modernists such 47



as Rogers or Grimshaw. The relationship which which are required. All of the above factorsexists between the structural design and the influence the choice of structural material andarchitectural design, as discussed in Section the basic structure type. These choices are2.2, is yet another global factor which is likely made from the vocabulary of basic structuralto influence the final form of the structure. forms which is outlined in the following

Factors particular to the building being chapters. Once the basic type of structure hasdesigned will obviously exert a strong influ- been selected a process of adaptation andence on its structural make-up. These include modification must occur, as described in thethe conditions at the site, the scale of the examples given above, so that it may satisfybuilding and the character of its internal layout the individual requirements of the particularas determined by the types of accommodation building.

48

Chapter 3

Steel structures

3.1 Introduction

Steel is a material which has excellent struc-tural properties. It has high strength in tensionand compression and is therefore able to resistbending and axial loads with equal facility; it isthe strongest of the commonly used structuralmaterials, being approximately twenty timesstronger than timber and ten times strongerthan concrete. It is therefore used to make thetallest buildings and the enclosures with thelongest spans. Its most common application,however, is for building frames of moderatespan, in a wide variety of configurations.

3.2 The architecture of steel - thefactors which affect the decision toselect steel as a structural material

3.2.1 The aesthetics of steelThe visual vocabulary which is associated withsteel structures contains some of the mostpowerful images of modern architecture. Theglass-clad framework (Fig. 3.6), the use ofslender, precisely crafted structural com-ponents as visual elements (Figs 3.1 and 3.9)and the celebration of structural virtuosity inthe form of either breathtakingly long spans(Fig. 3.10) or very tall buildings (Fig. 3.11) areall different aspects of the expressive, andimpressive, possibilities of steel. Theseaesthetic devices have been used from thebeginning of the modern period and are stillpart of the twentieth-century architecturalpalette. They are often the primary reason forthe selection of steel as the structural materialfor a building.

Steel became available as a buildingmaterial in the second half of the nineteenth

century, following the development of econom-ical processes for its manufacture, and,although its expressive possibilities were notused initially, it was quickly absorbed into thewell-established tradition of metal-framebuilding which had arisen in connection withthe use of cast and wrought iron (Fig. 3.2). Iron

Fig. 3.1 Channel 4 Headquarters, London, England,1994. Richard Rogers & Partners, architects. Concepts suchas neatness, precision and high quality control are readilyconveyed by the use of an exposed steel structure. [Photo:E. & F. McLachlan] 49


Fig. 3.2 Cross-section and details of Bage's Mill, Shrewsbury, England, 1796. This is a very early example of an iron-frame industrial building. The floors consisted of brick jack-arches, topped by a non-structural filling, supported on askeleton framework of cast iron beams and columns. No walls were required in the interior. Columns were not providedin the perimeter walls which were therefore of loadbearing masonry. [Illustration. Mitchell's History of Building]

50

Iron roof frames

Fireproof floors

Normalcolumn

Window

Specialcolumn

machineryshaft

forIron

columns

Specialcolumn

formachinery

shaft

Smallestsection

near topand bottomof column

Maximum atground floor

150mm/6 inches

Ironcolumn

Irontie-rod

Ironbeam

Swellingat middleof column

Brickjack-arches

Iron-column 2.6m

(8.5ft) high

Columnswells outat middle

Floor,

Detail oftypical column

beam andjack-arches

Typical column100mm (4 inches)

across attop and bottom

Floor.

Column topTie-rod

Socketfor column

Skewback

Skewback Tie-rod Socketfor column

•Wall

SIDE VIEW

won

TOP VIEW

•Tapering flange

.Skewback

100 mm

5 inches

Brickjack-arch

Floorsurface

10 metres

Ironcolumn

Socketfor column

Tie-rod

CUT-AWAY VIEWshowing iron frameand fireproof floor

Taperingflange

Skewback

Fireproofceiling

Brickwall

SECTION

BEAM SECTION

Ironbeam

5

5

15 20 25 30 feet100 5

1 0

0

0

Undersideof ironbeam

COLUMNSECTIONSOF SOLID

CAST IRON

frameworks had been developed in the earlynineteenth century principally for industrialbuildings such as factories, warehouses andrailway stations but were little used in thetypes of building which, at that time, wereconsidered worthy of the descriptive termArchitecture. As the century progressed,however, metal frameworks were graduallyabsorbed into the world of architecture. Theywere used mainly for 'new' types of building,such as department stores and multi-storeyoffices, where they allowed uncluttered inter-iors to be created within buildings which wereof conventional external appearance. InEurope, this was the age of revivals and thestreets of the commercial districts of latenineteenth-century European cities becamefilled with imitation Greek and Roman templesand Italian Renaissance palaces which were infact steel frameworks clad in masonry andwhich, in their scale and internal arrange-ments, bore little resemblance to their historicpredecessors. In North America, the techno-logy of the steel frame and the existence of arapidly expanding economy generated a newtype of building, that of the skyscraper, whichfound architectural expression in the work ofLouis Sullivan.

The structural technology on which thesebuildings were based was that of the skeletonframework, a characteristic of which was thatloads were channelled into slender elements oflow volume in which stress levels were high.Large-scale interior spaces were created withminimal interference from vertical structure,and plan arrangements were varied betweenlevels in multi-storey buildings, within theconstraints of a regular column grid, becausewalls were simply non-loadbearing partitions.The frameworks themselves were normallyhidden. Internal columns were cased in fire-proofing materials and finished frequently asone of the classical orders of architecturewithin a conventional scheme of interiordecoration (Fig. 3.3). External walls were ofloadbearing masonry and were fashioned intohistoric stylisations. Eventually, the frame wasto take over the structural function of theexternal walls, although, in the maioritv of

cases, the walls would continue to be ofmasonry until well into the twentieth century.

The first steel-framed building in Britain witha non-loadbearing external wall was the RitzHotel, of 1903-06, in London's Piccadilly. It isoften said that the reason for the adoption ofthe elaborate stone facade, which did not bearany load, was that the London Building Bye-Laws did not permit otherwise but, given theconservativeness of the architectural establish-ment in England at the time, it is difficult toimagine that a building of less conventionalappearance would have been erected on thisvery prominent London site.

The constructional system which was devel-oped in the nineteenth century in connectionwith metal-framed buildings, that of thecomplete separation of functions betweenstructural framework and non-loadbearingwalls, was to become the standard pattern forthe steel-frame architecture of the twentiethcentury. Structural skeletons would be made tosupport external cladding of many differentmaterials and the building-type would offerarchitects new opportunities for architecturalexpression. Masonry, which has ideal proper-ties as an external walling material, would stillbe used but other, more fragile materials, suchas sheet metal, glass and plastics, would alsobecome part of the architectural vocabulary.

The promotion of the steel framework fromthe status of a purely supportive role to that ofa major contributor to the aesthetics of abuilding was accomplished by the Modernistarchitects of the early twentieth century. It wasinevitable that they would find the materialexciting. In the 1920s and 1930s they wereengaged in the project of inventing a newarchitectural vocabulary for the modern worldof industry and technology, and, from theirpoint of view, this new structural material hadmany virtues. Perhaps the greatest was that itwas new, but it was also appealing because itwas a product of a complex and sophisticatedindustrial process which made it an appropri-ate medium with which to develop an architec-tural vocabulary celebrating industry andtechnology. Steel also had excellent structuralproperties, especially high strength in both 51

Steel structures


Fig. 3.3 Part cross-section of Selfridge's Department Store, London, England, 1907.The structure of this building is a skeleton framework of steel beams and columns whichsupport reinforced concrete floor slabs. Unlike in the mill building of Fig 3.2, the exteriorwalls do not support the floors and consist principally of glazing. The elaborate giant-order columns with their bases and stylised entablature are a non-structural screendesigned to give the appearance of a Beaux Arts style 'palace'. [ Illustration, Milchell'sHistory of Building]

52

1 0 5 10 15 metres

10 0 10 20 30 40 50 feet

ESCALATOR MACHINE ROOM

Concretewall

Steel beams insidefalse stonework

Stone columns onOxford Street front

Steel beams insidefalse stonework

Canopy

OXFORD STREET

PavementRevolving

door

Void

Window

Steel beams

GALLERY

GROUND FLOOR

BASEMENT

Steel structures

tension and compression. Its deficiencies -poor durability, poor performance in fire, thedifficulty of shaping it into useful components,the high weight of the components - were notsignificant, and would be overcome by employ-ing other technologies. The large ecologicalcost, in terms of transportation of raw materi-als and of the high energy consumption andpollution associated with its manufacture,were not an issue at the time. So far as themodernists were concerned, here was a newand exciting material whose expressive possi-bilities, if explored, might lead to a trulyappropriate architectural vocabulary for thetwentieth century.

The technology of the steel frameworkcontributed to the aesthetics of architecture inthe 1920s and 1930s in two quite separateways. Firstly, it was crucial to the developmentof the glass-clad building; secondly, it madetenable the overt use of structural elements asconstituents of a modern visual vocabulary.These two aspects of the aesthetics ofModernism were, and still are, often combinedby architects and confused by critics. They are,however, different and distinct aspects of therelationship between architecture and thestructural technology of steel frameworks.

The aesthetic programme of the glass-cladframework is concerned with 'transparency'.This, and the use of 'crystalline' form, weregiven symbolic meaning in the 1920s by theExpressionists, notably Bruno Taut. One of themost striking images to be published by thisgroup, however, was the well-known glassskyscraper project of Mies van der Rohe (Fig.3.4). Despite its Expressionist genesis thisbuilding form survived the gradual triumph ofRationalism - the Rationalists could regard theglass-clad frame as a logical and honest reduc-tion of the elements of a tall building to itsbare essentials - and was used at every scalein the architecture of Modernism, from smalldomestic buildings to the corporate sky-scrapers which dominate the skylines of mostcapital cities.

Among the visual sources of this architec-tural vocabulary of glass and steel were theiron-framed warehouses, factories and railway

Fig. 3.4 Glass skyscraper project, 1922, model. Mies vander Rohe, architect. In the early twentieth century manyvisionary architects considered mass-produced highquality glass to be the ultimate modern building material,principally because it allowed the exterior of a building tobe 'dematerialised'. The almost featureless high-rise struc-ture faced in glass was an important architectural innov-ation of the early modern period.

53


Fig. 3.5 Crystal Palace,London, England, 1851.Joseph Paxton,architect/engineer. Thisglass-clad framework madea significant contributionto the visual vocabulary ofthe architecture of thetwentieth century. Unlikethe later buildings which itinspired, the arrangementwas fully justified here ontechnical grounds.

54

stations of the nineteenth century. These hadlargely been ignored in their own time by thearchitectural mainstream but were destined toexert great influence on the visual vocabularyof twentieth-century Modernism - rather as themotifs of the architecture of Roman antiquity,as interpreted by the scholars of the ItalianRenaissance, had laid the foundations forNeoclassicism and the Beaux-Arts school. Awell-known example was the famous BoatStore at Sheerness Dockyard in England(1858-66) in which aspects of twentieth-century Modernism were anticipated, but othertypes of industrial building, especially thelarge train-sheds of the railway termini, werealso important. Extreme versions of this typeof building were the glasshouses - glass-cladframeworks of timber and iron - which hadbeen developed for horticultural purposes inthe late eighteenth and early nineteenthcenturies. This particular tradition of framebuilding reached its climax with the CrystalPalace in London (Fig. 3.5), which was built tohouse the Great Exhibition of 1851.1 It wasvirtually a cathedral in glass and iron and was

1 The designer of the building was the landscapegardener loseph Paxton.

destined to exert a profound influence on thearchitecture of the twentieth century.

It is appropriate, in a book which isconcerned principally with technical matters,to dwell briefly here on the relationshipbetween technology and aesthetics in thecontext of the glass-clad framework. It was, ofcourse, principally the look of the frame build-ings of the nineteenth century, and the CrystalPalace in particular, rather than the novelty oftheir technology of construction, whichappealed to the architects of the ModernMovement. The buildings did, in fact, have anumber of technical deficiencies. The poorthermal and acoustic properties of the externalwalls and the poor durability of the junctionsbetween the individual panes of glass in thetransparent skins were problematic. The lattercaused leaks to develop, which in turn broughtabout a deterioration in the condition of thestructural frameworks of timber and cast iron.None of these was a particularly seriousconsideration in the context of most of theglass-clad frameworks of the nineteenthcentury. The Crystal Palace, for example, was atemporary building designed to house a verylarge exhibition of short duration, and, from atechnical point of view, the glass-clad frame-work was actually an ideal, and probablyinevitable, solution to the problem posed,

Steel structures

especially as there was a requirement that thebuilding be erected very quickly. The glass-cladmetal framework, as a genus of building type,also performed well in the context of train-sheds, where generous provision for daylight-ing and natural ventilation were essential andwhere high levels of thermal and acousticinsulation were not required. The indifferentweathertightness of the external envelope wastolerable in the context of a large railwaystation where a programme of maintenancewas accepted as normal.

The technical deficiencies of the glass-cladframework were serious drawbacks in thecontext of other types of buildings, however,and especially of those within which a 'welltempered environment' was a reasonableexpectation. The buildings in which peoplelived, worked, became educated or were ill -the glass-clad houses, offices, schools andhospitals of the modernist twentieth century,in other words - were seriously deficient intheir technical performance. Glass, if used asthe sole covering for a building which isintended to be occupied by humans, actuallyperforms rather badly in a technical sense. Amasonry wall, which is pierced with glasswindows for light and ventilation, is a muchbetter technical solution to the problem ofcladding a frame building.

The early Modernists, however, despite theirdeclared allegiance to the idea of celebratingtechnology, were more interested in aestheticsthan in technical performance. Both theRationalists2 and the Expressionists, weredealing in metaphor and the symbolic attrac-tiveness of the glass-clad framework ensuredthat mundane, practical considerations wereoverridden. The glass-walled building becameone of the cliches of twentieth-centuryModernism - a triumph of ideas and aesthetics

2 In the context of architecture the terms Rational orRationalism should never be taken literally. They didnot mean 'that which is logical, sensible and practical'.The Rationalists produced an architectural vocabularywhich symbolised the idea of being logical, sensible andpractical.

Fig. 3.6 The Seagram Building, New York, USA, 1957.Ludwig Mies van der Rohe, architect. This was the arche-typal glass-clad framework of the late modern period.

over that which was practical in building terms(Fig. 3.6). That triumph was to make a signifi-cant contribution to the alienation whichsubsequently developed between the world ofarchitecture (architects and their apologistcritics) and the mainstream of society.

It is worth noting that, although the tech-nical efficacy of using glass cladding may be 55


56

Fig. 3.7 German Pavilion, World Exhibition, Barcelona,1929. Ludwig Mies van der Rohe, architect. The slendersteel structural columns, faced in thin coverings of stain-less steel, made a significant contribution to the aestheticsof this seminal building.

questioned, the structural performance of theglass-clad steel frameworks of the twentiethcentury was normally satisfactory. Most of thebuildings fall into the category of 'structureaccepted'3 in which the technical performanceof the structure was not compromised forvisual reasons.

In many of the buildings discussed above,and especially the multi-storey buildings, thesteel which formed the loadbearing structure

3 See Section 2.2, and Macdonald, Structure andArchitecture, Chapter 7.

was not actually visible. It was hidden, encasedin fire-proofing materials for good technicalreasons, and its presence was acknowledgedonly in the architectural treatment. It is for thisreason that a distinction is made here betweenthe role of steel in the glass-clad framework,where it served a purely structural functionwhich happened to make a new aestheticpossible, and the other aesthetic part which ithas played in modern architecture, that of adistinctive element in a visual vocabulary.

Metal components can be shaped with greatprecision and this, together with the lowvolume of the elements in skeleton frameworksof steel, can be used to create structures ofgreat elegance. The exposure of steelwork andits incorporation into the visual vocabulary isan aesthetic device which has been usedextensively in the twentieth century. The archi-tectural intention in this case was more-or-less

Steel structures

the same as with the glass-clad framework -the creation of an aesthetic which celebratedthe idea of progress, of industrial technologyand of the modern lifestyle which these madepossible - but the method of expression wasmore overt. The visual and tactile qualities ofsteel became important factors in the aestheticmake-up of the building.

Again, Mies van der Rohe is one of theprincipal exemplars. In the Barcelona Pavilionof 1929 (Fig. 3.7) the steel columns act inconjunction with new treatments of traditionalmaterials, such as marble, to create a newaesthetic. In the Farnsworth House of 1946-50(Fig. 3.8) the l-shaped structural elements of aglass-clad framework are exposed and bringthe visual vocabulary of the engineeringworkshop into the world of the country retreat.

The exposure of steel frameworks foraesthetic reasons was one of the favouredstylistic devices of the so-called 'high-tech'architecture of the 1970s and 1980s (Fig. 3.9).These buildings, which also were frequentlyglass-clad frameworks, were visually spectacu-lar and among the most memorable architec-tural images of the late twentieth century.

The relationship between the structure andthe architecture in all of these cases was oneof 'structure symbolised'4 and this way of treat-ing structure usually had the paradoxical effectof compromising its technical performance,because it resulted in structural forms beingmanipulated predominantly in accordance withvisual rather than with technical criteria.

Other problems associated with theexposure of steelwork were those of mainten-ance of the structure and of fire protection. Thesecond of these arose only in the case ofmulti-storey structures and was the principalreason why steel structures were rarelyexposed other than in single-storey buildings.A notable exception was the Centre Pompidouin Paris, where fairly elaborate fire-protectionsystems were provided to ensure that thesteelwork which was exposed on the exterior ofthe building would meet the required fire

4 See Section 2.2.

Fig. 3.8 The Farnsworth House, Illinois, USA, 1951.Ludwig Mies van der Rohe, architect. The mass-producedstructural steel beam and column I-sections form impor-tant elements in the aesthetic vocabulary of this building.

Fig. 3.9 Sainsbury supermarket, London, England,1986-89. Nicholas Grimshaw & Partners, architects.Exposed steel structures with exaggerated connectionswere one of the prominent features of the 'high-tech' style.[Photo: E. & F. McLachlan] 57


Fig. 3.10 Hong KongStadium, Hong Kong,1995. Ove Arup &Partners, structuralengineers. The principalstructural elements ofthis long-span structureare steel arches whichhave a space-frameworkconfiguration. |Photo:Ove Arup Partnership)

58

performance. All of the structural steelwork inthe interior of this building was covered withinsulating material and was therefore notexposed and visible.

Steel has made an important contribution tothe development of two other categories ofbuilding, namely, the very tall building and thevery long-span enclosure. Where steelwork hasbeen used in these contexts the limits of whatwas technically possible were sometimesapproached. In such situations the formswhich were adopted had to be determinedprincipally from technical considerations. Inthese two types of steel building, structuralrequirements dominated the form which wasadopted and the resulting building was oftenspectacular and exciting in appearance. Thesebuildings demonstrate the features of truestructural high tech'5 which should not be

5 See Section 2.2.

confused with the technically misleadingversion described above, which might betermed 'stylistic' high tech.

In the long-span structure (Fig. 3.10) thelevel of efficiency (strength-to-weight ratio)must be high. The types of steel structurewhich have been developed to achieve this arethe space framework, often in arched or barrel-vaulted configuration, and the cable network(see Sections 3.6.2.3 and 3.6.4).

In the very tall building (Fig. 3.11) the criticalstructural problem is that of providingadequate resistance to lateral loading. Themost efficient solution to this is to place thestructure on the perimeter of the plan, andthus treat the building as a hollow,cantilevered tube. In such circumstances thestructure is concentrated in the exterior wallswith the result that it inevitably affects theappearance of the building.

In the late twentieth century, the introduc-tion of the ideas of Deconstruction into archi-

Steel structures

tecture has led to the extension of the visualvocabulary of glass and steel. In the buildingsof architects such as the Coop Himmelblaugroup (Fig. 2.1) and Behnisch (Fig. 2.2) thismost 'technical' of materials has been used inthe 'structure ignored' type of relationship tosupport buildings with highly complex andirregular geometries. It is the very highstrength of steel in both tension and compres-sion and its ability to sustain high levels ofconcentrated load which makes these free-formnetworks of elements possible.

To sum up, the use of steel as the structuralmaterial creates a range of aesthetic opportun-ities for the designer of a building, any one ofwhich might provide the justification for itsselection. Firstly, steel provides a skeleton-framestructure which allows the freedom in internalplanning and external treatment which is associ-ated with buildings in which the walls are non-structural. An extreme instance of this is thebuilding with transparent glass walls, but thefreedom can be exploited in other ways in thetreatment of both the interior and the exterior ofthe building. Secondly, the great strength ofsteel allows the creation of very slender struc-tural elements. This, together with walls of light-weight appearance, can be used to create afeeling of lightness, openness and structuralelegance. Thirdly, the variety of precisely shapedelements which steel makes possible and the'technical' ambience which surrounds theseallows them to be used to convey symbolicmeaning. This possibility has so far been used ina straightforward manner to express the idea oftechnical progress. It could, of course, be usedironically to convey a quite different meaningand there is perhaps evidence that this ishappening in the architecture of Deconstruction.

Thus, the aesthetic of steel is the aestheticof the structural framework in which theelements are slender and in which the problemof providing support is solved in an apparentlyeffortless manner. It has normally also beenthe aesthetic of the celebration of technology,the machine and the production line as theyare symbolised by straight-edged forms,regular and often rectilinear grids and overtlymachine-like detailing.

Fig. 3.11 Sears Tower, Chicago, USA, 1974. Skidmore,Owings & Merrill, architects. The Sears tower, which wasfor over two decades the tallest building in the world, hasa steel framework structure. [Photo: I. Boyd Whyte].

3.2.2 The technical performance of steel asa structural material

3.2.2.1 IntroductionSteel is the strongest of the four commonly usedstructural materials but has a strength-to-weightratio which is similar to that of timber (i.e. very

59


high). It is ideally suited to the form of skeletonframeworks, where the principal alternative inmulti-storey buildings is reinforced concrete. Itshigh ratio of strength to weight also makes itsuitable for lightweight frameworks such as areused in roof structures. In this application theprincipal alternative is timber. The advantagesand disadvantages of steel in relation to thesematerials is reviewed briefly here.

3.2.2.2 Advantages of steel

StrengthThe high strength of steel, and its high ratio ofstrength to weight, makes it suitable for use insingle- and multi-storey skeleton frames, overa large range of spans and building heights.The material therefore offers the advantages ofskeleton-frame construction in all of itsmanifestations. In addition to freedom fromthe restrictions of loadbearing walls in theplanning of both the interior and the exteriorof the building, as mentioned above, theseinclude the subsequent flexibility to alter plansand elevations when required.

Considerable flexibility is also possible inthe planning of the building services and intheir subsequent maintenance and replace-ment when this becomes necessary. Theseadvantages are, of course, also present withreinforced concrete frameworks but becausethe structural elements in steel frames aremore slender and less obtrusive than those inequivalent reinforced concrete frames, the useof steel allows interiors with lighter and moreopen aspects to be created.

Ratio of strength to weightSteel frames are lighter than reinforcedconcrete frames of equivalent strength, partic-ularly if efficient types of element such astriangulated girders are used. This makes themmore suitable than reinforced concrete framesfor use in single-storey buildings and the roofstructures of multi-storey buildings. In this rolethe principal alternative to steel is timber.

Quality controlSteel is manufactured under conditions ofstrict quality control and its properties can be

relied upon to be within narrow specifiedlimits; this allows relatively small factors ofsafety to be adopted in the structural designcalculations and is a further reason why lightand slender elements are possible.

AppearanceDue to the strict quality control which isexercised during its manufacture and to themethods which are used in the final shaping ofsteel components, the finished structure has adistinctive appearance which is characterisedby slender elements, smooth surfaces andstraight, sharp edges.

PrefabricationSteel structures are assembled from prefabri-cated components which are produced off-siteand this allows their dimensions and generalquality to be carefully controlled. It also resultsin fast erection of the structure on site andenables a relatively simple erection process tobe adopted, even on difficult sites.Prefabrication with site-jointing also meansthat it is relatively easy for the designer toexercise control over whether or not the struc-ture is statically determinate.

3.2.2.3 Disadvantages of steel

IntractabilitySteel is a very tough material which is difficultto work and shape in the solid form and thishas a number of consequences. It means that,in most steelwork design, it is necessary tospecify elements from a standard range ofcomponents which are produced by steelmanufacturers and to carry out the minimumamount of modification to these. The standardelements, which are produced by a hot-rollingor cold-forming process (see Section 3.4), arestraight and parallel sided with the result thatsteel frameworks must normally have a regular,rectilinear, or at least straight-edged geometry.

The production of 'tailor-made' cross-sections, or of geometries which are curvilin-ear, is difficult and the use of steelwork tendstherefore to place more restrictions on theoverall forms of structures than does the use ofreinforced concrete. Also, the final adjustment60

Steel structures

of the dimensions of elements on site is diffi-cult, if not impossible, so that a much higherstandard of quality control is required in suchprocesses as the initial shaping and cutting tolength, than is necessary with, for example,timber.

WeightThe density of steel is high and this makesindividual components fairly heavy. Elementssuch as beams and columns are difficult tomove around on site and cranes are normallyrequired for the assembly of steel structures.

CostThe basic cost of a steel structure is normallygreater than that of its timber or reinforcedconcrete equivalent. The shorter on-siteconstruction time can be a compensatoryfactor, however.

DurabilityMost steels are relatively unstable chemicallyand a corrosion-protection scheme is normallyrequired for a steel structure.

Performance in fireSteel loses its ability to carry load at arelatively low temperature (around 500 °C) andthis means that, while a steel structure doesnot actually burn, it will collapse in fire unlessit is kept cool. This is normally achieved byprotecting the steelwork with a suitably thicklayer of fire resistant insulating material butsometimes more sophisticated methods, suchas water cooling systems, are used. The trad-itional fireproofing material was concrete - theelements of a steel frame were simply encasedin concrete - but much lighter materials whichare easier to apply have been developed. Theneed to provide fireproofing for steelworknevertheless increases the complexity of asteel-frame building and adds to the cost ofthe structure.

The need to provide fire protection isobviously particularly problematic if there is anintention to expose the structure as part of itsarchitectural expression. A prominent examplealready mentioned is the Centre Pompidou inParis where an elaborate system of shutters,which would operate automatically in the event

of a fire, had to be installed adjacent to theglass external wall in order to protect the steel-work on the exterior of the building should afire occur. The steel trusses in the interior wereencased in fireproofing material which waswrapped in a thin skin of sheet metal topreserve the appearance of a steel structure. Asimilar architectural language was proposed forthe Lloyd's Headquarters Building in London.In this case, however, no satisfactory way ofmeeting the fire-resistance requirements couldbe found and reinforced concrete was finallyadopted as the structural material.

3.2.2.4 ConclusionSteel is a material whose properties make itparticularly suitable for skeleton-frame struc-tures. The principal advantage which resultsfrom its adoption is that it releases thedesigner from the restrictions on internalplanning and the aesthetic treatment of theexterior which are imposed by the use ofloadbearing walls. In addition, the highstrength of steel makes possible a very widerange of frame types; both single-storey andmulti-storey frames can be constructed over avery wide range of spans and very tall multi-storey structures are also possible. In all casesthe structures will be lighter than equivalentreinforced concrete frames. Other advantagesstem from the fact that steel structures areprefabricated: these include speed of erection,ease of assembly on difficult sites and, ifrequired, statical determinacy. Restrictions onthe overall form of a building must normally beaccepted, however, due to the limited range ofcomponents which are readily available, andthe detailing of the structure is likely to becomplicated by the need to provide fire protec-tion. In addition, the cost of the steel-framestructure is likely to be higher than that of anequivalent reinforced concrete frame.

3.3 The properties and compositionof steel

Steel is a ferrous metal (its principalconstituent being iron), but it contains a 61


number of other chemical elements which actas alloying agents and which have a criticaleffect on its properties. The most important ofthese is carbon; steel is defined as ferrousmetal with a carbon content in the range of0.02% to 2%. Low-carbon steels are relativelysoft and ductile while those with a high carboncontent are hard and brittle. The range ofproperties is fairly wide, depending on theprecise levels of carbon and of other traceelements which are present. The term 'steel'refers therefore not to a single metal but to arange of alloys.

Structural steels are 'mild steels', which havea carbon content of around 0.23%; the otherprincipal alloying agent is manganese which ismaintained at around 1.6%; sulphur andphosphorus are also present. A range of struc-tural steels is available in most countries.Those which are used in the UK are specifiedin BS 4360 'Weldable Structural Steels' whichis currently being superseded by a EuropeanStandard EN 10 025. The latter incorporatesthe provisions of BS 4360.

BS 4360 specifies four grades of structuralsteel: 40, 43, 50 and 55. The grade numbersrefer to the tensile strength values (400, 430,500 and 550 newtons per square millimetre).6

Within each grade there are various sub-grades: A, B, C, etc. determined by minor vari-ations in chemical composition, principallycarbon content. The higher sub-grades (lowercarbon content) have slightly improvedmechanical properties and perform better inrespect of welding.

The properties of steel can be manipulatedby heat treatment. If the metal is cooled veryrapidly (quenched) the crystalline structure isquite different from that which results fromgradual cooling. Quenched steel is extremelyhard and brittle and is not used in structuralengineering. Following re-heating, however,the metal regains its ductility and the level of

6 It is intended to replace these designations with gradesbased on yield stress values. Thus grade 43A willbecome 275A as the yield stress for this type of steel is275 N/mm2.

7 Blanc, McEvoy and Plank, Architecture and Construction in

Steel, Chapter 3.62

Fig. 3.12 The relationship between stress and strain intypical structural steels. The short section of the graphbetween the elastic and plastic ranges, in which the graphis more-or-less horizontal, is of fundamental importance indetermining the excellent structural behaviour of steel.

brittleness/ductility (and therefore yieldstrength) which is achieved can be accuratelycontrolled in this process (which is calledtempering). Heat treated steels are used instructural engineering only for very specialisedapplications, the most common of which is inthe manufacture of high strength friction-gripbolts.7

Steel is a high-strength material which hasequal strength in tension and compression;the ultimate strength and design strengthvalues which are used in the UK are given inTable 3.1, which is reproduced from BS 5950The Structural Use of Steelwork in Building'.

The relationship between stress and strain ofa typical structural steel is shown in Fig. 3.12and it will be seen that 'elastic' behaviour(linear behaviour in which the graph of stressagainst strain is a straight line) occurs in thelower part of the load range. In the higher loadrange the relationship is curved (inelastic, non-linear behaviour) and a larger increase in strainresults for a given increase in stress. Thelocation in the graph at which the transition to

StressN/mm2

600

450

30C

150

5 10 15 20 25%

Strain

non-linear behaviour occurs is termed the'yield point'.

The portion of the stress-strain graph whichimmediately follows the yield point is more-or-less horizontal. This feature illustrates a veryimportant property of steel which is themechanism for the relief of stress concentrat-ions, especially in the vicinity of connections.If, for example, one part of a cross-sectiontends to be highly stressed relative to otherparts (this might occur in the vicinity of a bolt),the highly stressed material could reach itsyield point while the average stress was stillrelatively small (Fig. 3.13). If more load wereapplied the strain would increase equally in allparts of the cross-section but the level ofstress in the most highly stressed area wouldtend to remain constant because the horizon-tal portion of the stress-strain graph wouldhave been entered by the material in that area.In other parts of the cross-section, where thestress was within the elastic range, the stresslevel would continue to increase with increas-ing strain, however, and the distribution ofstress would therefore tend to become moreeven. The existence of the short horizontalportion of the stress/strain graph is therefore astress-relieving feature. It is a very importantfactor in determining the good structuralperformance of steel.

Another significant aspect of thestress-strain graph is the amount of deform-ation which occurs before failure. A very large

Fig. 3.13 The stress-relieving mechanism. Each of thefour diagrams in the lower half of this illustration showsthe distribution of stress across the cross-section X-X. Inthe first diagram all of the material is stressed within theelastic range and the material which is closest to the bolthole is the most highly stressed. As the load rises the levelof stress also rises. Once the yield stress is passed at themost highly stressed locations the horizontal portion ofthe stress-strain graph is entered and it is possible for thestress level to remain constant while stress levels in otherparts of the cross-section continue to rise. The distributionof stress then becomes more even. This is an example ofthe stress-relieving mechanism which is responsible forthe ability of steel to resist high levels of tensile load.

amount of deformation is in fact requiredbefore steel fractures. This too is a safetyfeature because it means that an overloadedstructure will suffer a large deflection whichgives warning of impending collapse.

3.4 Structural steel products

3.4.1 IntroductionThree basic techniques are used to form metalcomponents. These are casting, in whichmolten metal is poured into a suitably shapedmould, forging, where solid metal is beaten orotherwise forced into a particular shape, andmachining, where material is cut away byvarious means from a basic block of metal, toform a component with a particular shape. All 63

Table 3.1 Basic design strengths for steel

BS 4360 Grade Thickness, Strength of sections,

less than or plates and hollow

equal to (mm) sect ions (N/mm2)

43 16 275A, B and C 40 265

100 24550 16 355B and C 63 340

100 32555 16 450C 25 430

40 415

Steel structures

Yield stress

Increase in load


64

three methods are used in the formation ofsteelwork elements. The greatest tonnage ofelements for structural steelwork is producedby forging; two processes are used, hot-rollingand cold-forming8 and large ranges of steelcomponents are made by both of theseprocesses. The casting of major steel com-ponents is rare and the principal use of thisprocess is as the preliminary stage in theforming of small jointing components (Fig.3.14). Machining is the most precise of theshaping processes and is used both in the finalstages of production of small components,such as nuts and bolts, and in the final prepar-ation of large steel elements for a particular

8 See Blanc, McEvoy and Plank, op.cit., Chapters 4 and 5for a description of the manufacturing processes ofsteel components.

Fig. 3.14 Renault Sales Headquarters, Swindon, England,1983. Foster Associates, architects; Ove Arup & Partners,structural engineers. Steel castings are used for the joint-ing components of complex shape which occur at the endsof the tie bars in this connection. [Photo: Alastair Hunter]

structure. The drilling of bolt holes, forexample, is a machining process.

3.4.2 Hot-rolled sectionsHot-rolling is a forging process in which hotbillets of steel are passed repeatedly betweenprofiled rollers to produce straight elementswhich have particular shapes and sizes ofcross-section (Fig. 3.15). Several sets of rollersare normally required to transform a roughbillet into a finished element with a cross-section which has satisfactory structuralproperties. Most sets of rollers are capable of

Fig. 3.15 Hot-rolled steel sections. These mass-producedelements are formed by a rolling process which results inthe characteristic parallel-sided arrangement with constantcross-section. The thickness of the metal is relatively highand results in correspondingly high load-carrying capacity.

producing only one cross-section-, some can beadjusted within narrow limits to produce asmall range of cross-section weights within aparticular overall cross-section size (serialsize).

Steel manufacturers produce a limited rangeof cross-sectional shapes by this process; theprincipal ones are as follows:

I-sectionsThese are typified by the UK universal beamseries (Table 3.2); they have cross-sectionalshapes with a high second moment of area9

about one principal axis and a much lowersecond moment of area about the other princi-pal axis. They are intended for use as beamelements, which are normally subjected tobending in one plane only. The finishing rollerswhich are used to produce universal beams

9 See Macdonald, Structure and Architecture, Appendix 2, foran explanation of this term.

consist of a complex arrangement of two pairsof rollers, one with horizontal and one withvertical axes. This allows the thickness of theflanges and webs to be varied within smalllimits and accounts for the different weights ofsection which are produced within each serialsize.

H-sectionsThese are typified by the UK universal columnrange (Table 3.3) and have more-or-less thesame second moment of area about bothprincipal axes. They are designed to resist axialcompression, which requires a cross-sectionwith more-or-less equal rigidity about all axesto prevent buckling.

Channel and angle sectionsThese are much smaller than the I and Hranges and are used mainly for secondaryelements, such as purlins (secondary elementsin roof structures), and for the sub-elements ofsteel trusses and built-up cross-sections.

Hollow sectionsThese ranges are produced in circular, squareand rectangular form. The lighter sections areused mainly for trusses and space frames.Heavier sections are used as beams andcolumns.

Flat plateThis is available in thicknesses from 5 mm to65 mm and is used for elements with built-upcross-sections (plate girders and compoundbeams) and for jointing components.

StripThis is produced in the form of thin flat sheet.Its principal use in structural engineering is asthe raw material for cold-formed sections.

Wire, rope and cable

Steel wire is manufactured by drawing solidbar through a series of small diameter diesaccompanied by heat treatment. Wires (usuallyaround 50) are spun into strands. A ropeconsists of a number of strands (usually 6)spun around a core of steel or fibre. A cable ismade up of six ropes. The largest ropes are100 mm in diameter and cables may be up to250 mm in diameter.

Steel structures

65

Table 3.2 Dimensional properties of universal beam sections

To: BS4 Part 1

Designation

Serial

size

mm

914 X 419

914 X 305

838 X 292

762 X 267

686 X 254

610 X 305

610 X 229

533 X 210

Mass per

metre

kg

388

343

289

253

224

201

226

194

176

197

173

147

170

152

140

125

238

179

149

140

125

113

101

122

109

101

92

82

Depth

ofsection

D

mm

920.5

911.4

926.6

918.5

910.3

903.0

850.9

840.7

834.9

769.6

762.0

753.9

692.9

687.6

683.5

677.9

633.0

617.5

609.6

617.0

611.9

607.3

602.2

544.6

539.5

536.7

533.1

528.3

Width

ofsection

B

mm

420.5

418.5

307.8

305.5

304.1

303.4

293.8

292.4

291.6

268.0

266.7

265.3

255.8

254.5

253.7

253.0

311.5

307.0

304.8

230.1

229.0

228.2

227.6

211.9

210.7

210.1

209.3

208.7

Thickness

Web

t

mm

21.5

19.4

19.6

17.3

15.9

15.2

16.1

14.7

14.0

15.6

14.3

12.9

14.5

13.2

12.4

11.7

18.6

14.1

11.9

13.1

11.9

11.2

10.6

12.8

11.6

10.9

10.2

9.6

Flange

T

mm

36.6

32.0

32.0

27.9

23.920.2

26.8

21.7

18.8

25.4

21.6

17.5

23.7

21.0

19.0

16.2

31.4

23.6

19.7

22.1

19.6

17.3

14.8

21.3

18.8

17.4

15.6

13.2

Area

ofsection

cm2

494.5

437.5

368.8

322.8

285.3

256.4

288.7

247.2

224.1

250.8

220.5

188.1

216.6

193.8

178.6

159.6303.8

227.9

190.1

178.4

159.6

144.5

129.2

155.8

138.6

129.3

117.8

104.4

Moment

Axis

x-x

cm4

718742

625282

504594

436610

375924

325529

339747

279450

246029

239894

205177

168966

170147

150319

136276

118003

207571

151631

124660

111844

98579

87431

75720

76207

66739

61659

55353

47491

of inertia

Axis

v-ycm4

45407

39150

15610

13318

11223

9427

11353

9069

7792

8174

6846

5468

6621

5782

5179

4379

15838

11412

9300

4512

3933

3439

2912

3393

2937

2694

2392

2005

Radius of gyration

Axis

x-x

cm

38.1

37.8

37.0

36.8

36.3

35.6

34.3

33.6

33.1

30.9

30.5

30.0

28.0

27.8

27.6

27.2

26.1

25.8

25.6

25.0

24.9

24.6

24.2

22.1

21.921.8

21.7

21.3

Axis

y-ycm

9.58

9.46

6.51

6.42

6.27

6.06

6.27

6.06

5.90

5.71

5.57

5.39

5.53

5.46

5.38

5.24

7.22

7.08

6.99

5.03

4.96

4.88

4.75

4.67

4.60

4.56

4.51

4.38

Elastic modulus

Axis

x-x

cm3

15616

13722

10891

9507

8259

7210

7986

6648

5894

6234

5385

4483

49114372

3988

3481

6559

49114090

3626

3222

2879

2515

2799

2474

2298

2076

1798

Axis

y-ycm3

2160

1871

1014

871.9

738.1

621.4

772.9

620.4

534.4

610.0

513.4

412.3

517.7

454.5

408.2

346.1

1017

743.3

610.3

392.1

343.5

301.4

255.9

320.2

278.8

256.5

228.6

192.2

Plastic

Axis

x-x

cm3

17657

15474

12583

10947

9522

8362

9157

7648

6809

7167

6197

5174

5624

4997

4560

3996

7456

5521

4572

4146

3677

3288

2882

3203

2824

2620

2366

2056

modulus

Axis

y-ycm3

3339

2890

1603

1372

1162

982.5

1211

974.4

841.5

958.7

807.3

649.0

810.3

710.0

637.8

542.0

1574

1144

936.8

612.5

535.7

470.2

400.0

500.6

435.1

400.0

356.2

300.1

Structural D

esign for Architecture

66

457 X 191

457 X 152

406 X 178

406 X 140

356 X 171

356 X 127

305 X 165

305 X 127

305 X 102

254 X 156

254 X 102

203 X 133

9889

82

74

67

82

74

67

60

52

74

67

60

54

46

39

67

57

51

45

39

33

54

46

40

48

42

37

33

28

25

43

37

31

28

25

22

30

25

467.4463.6

460.2

457.2

453.6

465.1

461.3

457.2

454.7

449.8

412.8

409.4

406.4

402.6

402.3

397.3

364.0

358.6

355.6

352.0

352.8

348.5

310.9

307.1

303.8

310.4

306.6

303.8

312.7

308.9

304.8

259.6

256.0

251.5260.4

257.0

254.0

206.8

203.2

192.8192.0

191.3

190.5

189.9

153.5

152.7

151.9

152.9

152.4

179.7

178.8

177.8

177.6

142.4

141.8

173.2

172.1

171.5

171.0

126.0

125.4

166.8

165.7

165.1 .

125.2

124.3

123.5

102.4

101.9

101.6

147.3

146.4

146.1

102.1

101.9

101.6

133.8

133.4

11.410.6

9.9

9.1

8.5

10.7

9.9

9.1

8.0

7.6

9.7

8.8

7.8

7.6

6.9

6.3

9.18.0

7.3

6.9

6.5

5.9

7.7

6.7

6.1

8.9

8.0

7.2

6.6

6.1

5.8

7.3

6.4

6.16.4

6.1

5.8

6.3

5.8

19.617.7

16.0

14.5

12.7

18.9

17.0

15.0

13.3

10.9

16.0

14.3

12.8

10.9

11.2

8.6

15.7

13.0

11.5

9.7

10.7

8.5

13.7

11.8

10.2

14.0

12.1

10.7

10 8

8.9

6.8

12.7

10.9

8.6

10.0

8.4

6.8

9.6

7.8

125.3113.9

104.5

95.0

85.4

104.5

95.0

85.4

75.9

66.5

95.0

85.5

76.0

68.4

59.0

49.4

85.4

72.2

64.6

57.0

49.4

41.8

68.4

58.9

51.5

60.8

53.2

47.5

41.8

36.3

31.4

55.1

47.5

40.0

36.2

32.2

28.4

38.0

32.3

4571741021

37103

33388

29401

36215

32435

28577

25.464

21345

27329

24329

21508

18626

15647

12452

19522

16077

14156

12091

10087

8200

1 1710

9948

8523

9504

8143

7162

6487

5421

4387

6558

5556

4439

4008

3408

2867

2887

2356

234320861871

1671

1452

1 143

1012

878

794

645

1545

1365

1199

1017

539411

1362

1109

968

812

357

280

1061

897

763

460

388

337

193

157

120

677

571

449

178

148

120

384

310

19.119.0

18.818.7

18.5

18.6

18.5

18.3

18.3

17.9

17.0

16.9

16.816.5

16.3

15.9

15.1

14.9

14.8

14.6

14.3

14.0

13.1

13.0

12.9

12.5

12.4

12.3

12.5

12.2

11.8

10.9

10.8

10.5

10.5

10.3

10.0

8.72

8.54

4.334.28

4.23

4.19

4.12

3.31

3.26

3.21

3.23

3.11

4.03

4.00

3.97

3.85

3.02

2.89

3.99

3.92

3.87

3.78

2.69

2.59

3.94

3.90

3.85

2.75

2.70

2.67

2.15

2.08

1.96

3.51

3.47

3.35

2.22

2.14

2.05

3.18

3.10

19561770

1612

1461

1296

1557

1406

1250

1120

949.0

1324

1188

1058

925.3

777.8

626.9

1073

896.5

796.2

686.9

571.8

470.6

753.3

647.9

561.2

612.4

531.2

471.5

415.0

351.0

287.9

505.3

434.0

353.1

307.9

265.2

225.7

279.3

231.9

243.0217.4

195.6

175.5

152.9

149.0

132.5

1 15.5

103.9

84.6

172.0

152.7

134.8

114.5

75.7

58.0

157.3

128.9

112.9

95.0

56.6

44.7

127.3

108.3

92.4

73.5

62.5

54.6

37.8

30.8

23.6

92.0

78.1

61.5

34.9

29.0

23.6

57.4

46.4

22322014

1833

1657

1471

1800

1622

1441

1284

1094

1504

1346

1194

1048

888.4

720.8

1212

1009

894.9

773.7

653.6

539.8

844.8

722.7

624.5

706.1

610.5

540.5

479.9

407.2

337.8

568.2

485.3

395.6

353.4

305.6

261.9

13.3

259.8

378.3337.9

304.0

272.2

237.3

235.4

209.1

182.2

162.9

133.2

266.9

236.5

208.3

177.5

118.3

91.08

243.0

198.8

174.1

146.7

88.68

70.24

195.3

165.8

141.5

115.7

98.24

85.66

59.85

48.92

37.98

141.2

119.6

94.52

54.84

45.82

37.55

88.05

71.39

Steel stru

ctures

67

Table 3.3 Dimensional properties of universal column sections

To: BS4 Part

Structural D

esign for Architecture

Designation

Serial

size

mm

356 X 406

Co lumn core356 X 368

305 X 305

254 X 254

203 X 203

152 X 152

Mass per

metre

kg

634

551

467

393

340

287

235

477

202

177

153

129

283

240

198

158

137

118

97

167

132

107

89

73

86

71

60

52

46

37

30

23

Depth

ofsection

D

mm

474.7

455.7

436.6

419.1

406.4

393.7

381.0

427.0

374.7

368.3

362.0

355.6

365.3

352.6

339.9

327.2

320.5

314.5

307.8

289.1

276.4

266.7

260.4

254.0

222.3

215.9

209.6

206.2

203.2

161.8

157.5

152.4

Width

ofsection

B

mm

424.1

418.5

412.4

407.0

403.0

399.0

395.0

424.4

374.4

372.1

370.2

368.3

321.8

317.9

314.1

310.6

308.7

306.8

304.8

264.5

261.0

258.3

255.9

254.0

208.8

206.2

205.2

203.9

203.2

154.4

152.9

152.4

Thickness

Web

t

mm

47.6

42.0

35.9

30.6

26.5

22.6

18.5

48.0

16.8

14.5

12.6

10.7

26.9

23.0

19.2

15.7

13.8

11.9

9.9

19.2

15.6

13.0

10.5

8.6

13.0

10.3

9.3

8.0

7.3

8.1

6.6

6.1

Flange

T

mm

77.0

67.5

58.0

49.2

42.9

36.5

30.2

53.2

27.0

23.8

20.7

17.5

44.1

37.7

31.4

25.0

21.7

18.7

15.4

31.7

25.3

20.5

17.3

14.2

20.5

17.3

14.2

12.5

11.0

11.5

9.4

6.8

Area

ofsection

cm2

808.1

701.8

595.5

500.9

432.7

366.0

299.8

607.2

257.9

225.7

195.2

164.9

360.4

305.6

252.3

201.2

174.6

149.8

123.3

212.4

167.7

136.6

114.0

92.9

110.1

91.1

75.866.4

58.847.4

38.2

29.8

Moment of inertia

Axis

x-x

cm4

275140

227023

183118

146765

122474

99994

79110

172391

66307

57153

48525

40246

78777

64177

50832

38740

32838

27601

22202

29914

22575

17510

14307

11360

9462

7647

6088

5263

4564

2218

1742

1263

Axis

y-ycm4

98211

82665

67905

55410

46816

38714

31008

68056

23632

20470

17469

14555

24545

20239

16230

12524

10672

9006

7268

9796

7519

5901

4849

3873

3119

25362041

1770

1539

709

558

403

Radius of gyration

Axis

x-x

cm

18.5

18.0

17.5

17.1

16.8

16.5

16.2

16.8

16.0

15.9

15.8

15.6

14.8

14.5

14.2

13.9

13.7

13.6

13.4

11.9

11.6

11.3

11.2

11.19.27

9.16

8.96

8.90

8.81

6.84

6.75

6.51

Axis

y-ycm

11.0

10.9

10.7

10.5

10.4

10.3

10.2

10.6

9.57

9.52

9.46

9.39

8.25

8.14

8.02

7.89

7.82

7.75

7.68

6.79

6.68

6.57

6.52

6.46

5.32

5.28

5.19

5.16

5.11

3.87

3.82

3.68

Elastic modulus

Axis

x-x

cm3

11592

9964

8388

7004

6027

5080

4153

8075

3540

3104

2681

2264

4314

3641

2991

2368

2049

1755

1442

2070

1634

1313

1099

894.5

851.5

708.4

581.1

510.4

449.2

274.2

221.2

165.7

Axis

y-ycm3

4632

3951

3293

2723

2324

1940

1570

3207

1262

1100

943.8

790.4

1525

1273

1034

806.3

691.4

587.0

476.9

740.6

576.2

456.9

378.9

305.0

298.7

246.0

199.0

173.6

151.5

91.78

73.06

52.95

Plastic modulus

Axis

x-x

cm3

14247

12078

10009

8229

6994

5818

4689

9700

3977

3457

2964

2482

5101

4245

3436

2680

2298

1953

1589

2417

1875

1485

1228

988.6

978.8

802.4

652.0

568.1497.4

310.1

247.1

184.3

Axis

y-ycm3

7114

6058

503.8

4157

3541

2952

2384

4979

1917

1668

1430

1196

2337

1947

1576

1228

1052

891.7

723.5

1132

878.6

695.5

575.4

462.4

455.9

374.2

302.8

263.7

230.0

140.1

111.2

80.87

Fig. 3.16 The castellated beam is a modified universalI-section beam. It is more efficient in resisting bendingthan the parent beam because a higher bending momentcan be resisted by the same area of cross-section andtherefore volume of material.

Steel structures

Fig. 3.18 I-section girder assembled by welding from flatplate. These are used when the required size of cross-section is larger than the largest available in the standardrange.

non-standard sections can be built up from thestandard sections. The 'castellating' process,by which universal beam sections are modified(Fig. 3.16), is one such process and provides arange of beam sections which are lighter thanthe standard hot-rolled sections (Tables 3.2and 3.3). Light beam sections can also bemanufactured from combinations of channeland angle sections (Fig. 3.17). Where veryheavy cross-sections are required, these can bebuilt up from flat plate as welded plate girders(Fig. 3.18).

3.4.3 Cold-formed sectionsCold-formed sections are fabricated from thinsheet steel (strip) by rolling or by folding(press braking), both of which are types offorging. As with hot-rolled sections, manufac-turers produce standard ranges, such asangles, channels and l-sections, which areintended to be generally useful for a variety ofstructural purposes. The capital cost of theequipment required to produce these sectionsis less than that for hot-rolled products andthe ranges which are readily available aremuch wider. The ease with which steel sheetcan be bent into complicated shapes allows 69

Fig. 3.17 Composite beams, built up by bolting orwelding standard sections together, are used where nosuitable standard section is available.

The structural sections described above arethe basic components from which steel struc-tures are assembled. Elements for particularapplications are normally selected from thestandard ranges; it is not considered economicto produce 'tailor made' cross-sections by thehot-rolling process due to the high capital costof the special rolling-mill equipment whichwould be required. If none of the standardsections is suitable for a particular application


Fig. 3.19 Cold-formed sections. Structural sections canbe formed by bending thin sheet. This allows morecomplex shapes of cross-section to be created than ispossible with the hot-rolling process. The metal must berelatively thin, however, so these are lighter sections with alower carrying capacity than hot-rolled equivalents.

the fabrication of sections with complexgeometries (Fig. 3.19) and 'tailor-made' cold-formed sections can be produced for specialpurposes at a reasonable cost. The sectionswhich are most commonly employed in build-ing are Z-purlins, which are frequently used inconjunction with frames of hot-rolledelements, and profiled-sheet decking. Manykinds of the latter are produced and they areemployed both as roof decks and, in conjunc-tion with in situ reinforced concrete, as floorstructures. Another common use of cold-formed sections is for the flanges of lightlattice beams and a number of proprietaryranges of these are available. An example isthe Metsec range (Figs 3.20). These aresuitable for use as joists in conjunction withloadbearing masonry or with main frames ofhot-rolled sections. They are also used to formcomplete frames in situations where only lightloads are carried.

Cold-formed sections are much lighter thanhot-rolled sections and they have considerablyless strength than the latter. Their principalapplication has been in single-storey struc-tures; most cold-formed sections do not have

Fig. 3.20 Metsec beam.This is an example of aproprietary beam inwhich cold-formedsections of relativelycomplicated shape havebeen used for the flanges.Elements like this arehighly efficient in resist-ing load and are thereforelight. Their limited load-carrying capacity makesthem most suitable foruse in roof structures.

70

Steel structures

Fig. 3.21 Cast steel 'Gerberette', Centre Pompidou, Paris,France, 1977. Piano and Rogers, architects; Ove Arup &Partners, structural engineers. Casting was the only feas-ible method by which these large steel components couldbe manufactured, given their complex shapes in profileand cross-section. Recent developments in castingtechnique have resulted in an increase in the use of thistype of component. [Photo: A. Macdonald]

sufficient strength to be used effectively inmulti-storey structures where floor loads mustbe carried.

3.4.4 Cast steel componentsAs with other metals, individual componentsof complex geometry can be produced in steelby casting. In structural engineering thistechnique is normally confined to the produc-tion of small components for joints in struc-tures. Exceptionally, it is used to produce large

structural elements of complex geometry; anexample is shown in Fig. 3.21.

3.4.5 BoltsTwo types of bolt are used to make connec-tions in structural steelwork, ordinary boltsand friction-grip bolts. In the case of ordinarybolts the load is transmitted between com-ponents through the shanks of the boltsthemselves (Fig. 3.22), which are normallyloaded in shear but which can be made to 71


3.5 Performance of steel in fire

The performance of a building in a fire is animportant design consideration, the principalconcern being the safety of people who are inor near the building at the time. There aremany aspects to this including the provision ofadequate means of escape and the control ofsmoke. So far as the structure is concerned,the principal concern is the prevention ofinstability due to the lowering of the yieldstrength of the material as a result of increasein temperature. In the case of steel, seriousdeterioration in strength begins at tempera-tures of around 500 °C. The most vulnerableparts of the structure are compressionelements and the compression flanges ofbeams.

Most regulatory authorities specify minimumrequirements for performance of a building infire depending on the type of occupancy forwhich the building is intended, the size of thebuilding, and the extent to which it is compart-mentalised. So far as the structure isconcerned the performance is measured interms of a minimum period of fire resistance(from 30 minutes to 4 hours, depending on thesize of the building and the type of occupancy).Roof structures and single-storey structures arenormally exempted.

In the case of steel structures, two types ofstrategy may be employed to meet the fireperformance requirements, namely the 'fireprotection' strategy and the 'fire resistance'strategy. 'Fire protection' is by far the morecommon. This involves the encasing of thestructure in a layer of insulating material inorder to reduce the rate at which the tempera-ture of the steel increases, so that the criticaltemperature is not reached within the requiredfire-rating period.

Where the strategy is one of fire resistancethe objective is to minimise or entirely elim-inate the need for protection. To comply withfire regulations by this method the designer ofa structure is required to demonstrate, bycalculation or some other simulation method,that the structure will not lose its integrity orbecome unstable in a fire before the required

Fig. 3.22 The actions of the two principal types of boltused in structural engineering are illustrated here.(a) The 'ordinary' bolt transmits load by bearing and shear.This causes stress concentrations and is not particularlyefficient. Connections are simple and cheap to constructby this method, however.(b) The 'high strength friction grip bolt' clamps togetherthe elements being joined with sufficient pressure to allowthe load to be transmitted by friction between thesurfaces. This reduces stress concentrations and is a moreefficient way of transmitting load through the connection.It is more expensive to construct, however, because boththe bolt and the elements being connected must bemanufactured to closer tolerances.

carry tension. Ordinary bolts are made inseveral grades of steel and the two gradeswhich are most commonly used in the UK are4.6 and 8.8. The first figure in the gradenumber represents the tensile strength of thesteel in kgf/mm2 X 10-1; the second figure is afactor by which the first is multiplied to givethe yield stress of the steel.

Friction-grip bolts operate by clampingcomponents together with such force that theload is transferred by friction on the interfacebetween them. Shear load therefore passesdirectly from one component to the other andis not routed through the bolts (Fig. 3.22).Friction-grip bolts are also used to carry axialtension through joints, however, in which casethe load passes along the bolt shank.

72

Steel structures

form an insulating layer. Intumescent coatingsare particularly effective in situations in whichthe steelwork forms part of the architecturallanguage of the building.

A further fire protection strategy is the use ofwater cooling as a means of maintaining thetemperature of a steel structure at an accept-able level. This strategy normally requires thathollow-section elements be adopted and inpractice has only very rarely been utilised. Aprominent example is the Centre Pompidou inParis in which the circular cross-sectioncolumns are protected in this way. Even in thisbuilding, however, conventional fire protectionsystems were used for all other structuralelements.

3.6 Structural forms

3.6.1 IntroductionSteel frames are assemblies of structuralelements which are selected or built up fromthe standard range of components. They areprefabricated structures whose constituentsare prepared to a semi-finished state in thefactory or workshop before being transportedto the site for erection.

The design of the connections between theelements is an important aspect of theplanning of steel frames. Two types of fasten-ing element are used in steelwork: bolts andwelding. In most cases, welding provides abetter structural junction than bolts, but boltsare easier to connect on site. Most steelframes are therefore designed so that they canbe prefabricated by welding into parts whichare small enough to be transported easily (notmore than 10 m in length) and these are thenconnected together on site by bolting; moststeel frames therefore have joints in which acombination of welding and bolting is used.

So far as the structural performance isconcerned there are two basic types of joint,hinge-type joints and rigid joints. Hinge-typejoints transmit shear and axial force betweenelements but not bending moment. They are,in other words, incapable of preventingelements from rotating relative to one another 73

fire-rating time has elapsed. Various strategiescan be adopted to achieve this. For example,the structure can be designed such that theactual stresses under service loads aremaintained at a low level in parts of the build-ing which are vulnerable to fire (e.g. columns).Bearing in mind that the problem with fire isthe reduction in yield strength due to tempera-ture increase, and that instability will not occuruntil the yield strength becomes reduced tothe level of the actual stress in an element, thelower the stress in the material the longer willit take for that point to be reached followingexposure to fire. If this method is applied,therefore, the need for protection of steelworkby insulating material can theoretically beeliminated or the amount of the requiredprotection reduced. The problem with the fireresistance strategy is the lack of data on whichto base the calculations required to prove thatthe necessary resistance time will be achieved.This is currently an active field of research,however, and the situation may be expected toimprove. The protection of steelwork ('fireprotection') is, however, the strategy which isnormally adopted at present.

Three types of insulating material are usedto protect steel structures. The first groupcomprises the traditional materials ofconcrete, masonry and insulating board, suchas plasterboard. Concrete is perhaps the mosteffective but is expensive and can add signifi-cant dead load to the structure. Insulatingboards are also expensive because the fixing ofthem is labour intensive. The second type ofsystem consists of spray-on material based onrock-fibre or vermiculite. This is currently thecheapest method and is particularly conven-ient for complex shapes or connections but itis generally regarded as unsightly and onlysuitable in situations in which the structurewill not be exposed to view, such as abovefalse ceilings. A third type of product is theintumescent coating. With this system theprotecting material is very thin (treated steel-work has the appearance of having beenpainted) but on exposure to fire, a chemicalreaction occurs which generates gas andcauses the material to become a foam and


and therefore behave in a similar way to ahinge. Rigid joints transmit shear and axialforce and bending moment between elementsand therefore provide a fully fixed connection.Both types of joint can be made either bywelding or by bolting but are normally madeby a combination of both; rigid joints are morecomplex and more expensive to design and toconstruct. Examples of both types of joint areshown in Figs 3.23 and 3.24.

The selection of the joint types and theirdisposition in the frame is an importantconsideration in the design of a steel structureThe principal factors which are affected by thejoints are the stability of the structure and itsstate of statical determinacy.10 Frames in whichthe majority of joints are of the hinge type(simple frames) are usually statically determin-ate and are easier and cheaper to design andconstruct than those with rigid joints, whichare normally statically indeterminate. Frameswith hinge-type joints are usually unstable intheir basic form and require separate bracingsystems for stability. Rigidly jointed frames,due to their statical indeterminacy, are moreefficient types of structure and therefore havemore slender elements than hinge-jointedequivalents; they are also self-bracing.

10 The question of statical determinacy versus staticalindeterminacy is an important consideration in struc-tural design because it affects a number of aspects ofthe performance of structures. For a discussion of theissues involved see Appendix 3 of Macdonald, Structureand Architecture.

Fig. 3.23 Hinge-type joints. The joints illustrated hereare capable of transmitting shear force but not bendingmoment and are therefore hinge-type connections. Notethat both welding and bolting are used. Welding is moreefficient in transmitting load but bolting is easier on thebuilding site. The joints are detailed so that the site-madepart of the connection is made by bolting.

Fig. 3.24 Rigid joints. These joints allowthe transmission of both shear force andbending moment between elements. Notethat they too are executed by a combin-ation of bolting and welding.

74

Steel structures

Secondary beam

Primary beam

Steel frames can therefore be thought of aseither 'simple' frames, that is frames in whichthe majority of the joints are of the hinge type,or as 'rigid' frames, in which case the joints arerigid. Most frames are of the 'simple' type andtherefore require bracing systems for stability(see Sections 3.6.2 and 3.6.3).

As with other types of structure, the config-uration which is adopted for a steel frame isinfluenced by the principal types of load whichit will carry and, in particular, by the characteris-tics of whichever of the applied loads isdominant. In the descriptions which followframes are subdivided into three categoriesdepending on the dominant type of appliedload. The three main types of applied load onarchitectural structures are imposed roof load,imposed floor load and wind load. The threeresulting categories are single-storey frames, inwhich imposed roof loading is the dominantload, low- and medium-rise multi-storey frames,in which imposed floor load is the dominantform of load, and high-rise multi-storey frames,in which wind load is a significant structuralfactor in the design. A very large number ofdifferent geometries and configurations ispossible within each category. Only the mostbasic forms are actually described here.

3.6.2 Single-storey frames

3.6.2.1 IntroductionThe variety of different possible arrangementsfor single-storey frames is very large and theyare placed here into the two broad categoriesof one-way- and two-way-spanning systems.Each of these categories is then further sub-divided.

3.6.2.2 One-way-spanning systemsThere are two basic types of one-way-spanningsingle-storey frame, the main differencebetween them being that in one the principalelements are spaced close together and carrythe roof cladding directly, and in the other theyare located at a fairly wide spacing and arelinked by a secondary system of elements towhich the cladding is attached. The second ofthese arrangements is more versatile and the

Fig. 3.25 This is a typical layout for a single-storey framein which lightweight steel elements are used. In theprimary-secondary beam system illustrated, closely spacedtriangulated joists are arranged parallel to one anotherand are supported by primary beams, which are in turnsupported by columns. The column grid is rectangular withthe secondary beams running parallel to the long side ofthe rectangle.

majority of single-storey steel frames are inthis category.

Figure 3.25 shows a typical arrangement ofelements in the first type of frame. It is normalfor only every third or fourth element to besupported directly on a column with theremaining elements being carried on beamsrunning at right angles to the principal direc-tion to produce a rectangular column grid. Theroof geometry can be given a wide variety ofelevational forms including flat, mono-pitch,duo-pitch and curved, but a regular-plangeometry consisting of equally spaced parallelelements is normally adhered to.

The principal elements in this type of framecarry relatively small areas of roof and aretherefore lightly loaded. This favours the use oflight, efficient types of structural element suchas proprietary lightweight lattice joists, basedon cold-formed sections (Figs 3.20 and 3.26).'Improved' hot-rolled sections, such as castel-lated beams (Figs 3.16 and 3.27), are alsosuitable but the use of standard hot-rolledsections such as the universal beam is rarelyeconomic. 75


Fig. 3.26 A single-storey steel framework isshown here comprisinglightweight Metsec joistsbased on cold-formedsections. [Photo: Photo-Mayo Ltd ]

Fig. 3.27 Castellatedbeams are arranged herein one of the standardconfigurations for single-storey steel frameworks.[Photo: Pat Hunt]

Fig. 3.28 Single-storey frameworksbased on lightweight joists can bebraced by the rigid-joint technique.The rigid joints can be either at thebases or at the tops of the columns.All columns must normally bebraced, however.

The size of the column grid depends on thetype of structural element which is specified;the normal range is from approximately 6 m X6 m to 10 m X 25 m with element spacings ofaround 1.5 m to 2 m. The columns are usuallyhot-rolled sections, either universal columnsor rectangular hollow sections. Approximatesizes for the normal span range of this type ofstructure are given in Table 3.4.

Both self-bracing and braced versions of thisconfiguration are possible." Self-bracingframes have either rigid column base connec-tions or rigid beam/column joints (Fig. 3.28).

Where hinge-type connections are used,bracing in both horizontal and vertical planesis required and diagonal bracing can be usedfor both (Fig. 3.29). Diaphragm bracing is analternative, however. In the vertical plane thiscan be provided by masonry walls, if these areadequately tied to the columns, and in thehorizontal plane the roof decking can performthis function, where it has sufficient strength.

Figure 3.30 shows a typical plan layout, withtypical span ranges, for a single-storey frame inwhich the primary elements are arranged atfairly wide spacings and a secondary structure,on which the roof cladding is mounted, isprovided to link them together. The primaryelements must be stronger than in the previoustype of frame and hot-rolled sections are

11 For a more detailed explanation of the bracing require-ments of frameworks see Macdonald, Structure andArchitecture, Section 2.2. 77

Secondary beam Secondary beam Secondary beam Column spacing. Column width Column height

span, L (m) spacing, P (m) depth (mm) S (m) (mm) (m)

4 1.0 200 3 50 3.05 1.0 200 3 60 3.06 1.5 250 4.5 70 3.57 1.5 300 4.5 70 3.58 2.0 350 6 90 4.09 2.0 350 6 90 4.0

10 2.0 400 6 90 4.012 2.5 550 9 100 5.014 2.5 600 10 150 5.016 3.0 700 12 150 5.018 3.0 800 12 150 5.020 3.0 900 15 180 5.022 3.0 1100 15 180 6.024 3.0 1300 18 200 6.026 3.0 1500 18 250 6.028 3.0 1600 24 250 6.030 3.0 1800 24 250 6.0

Table 3.4 Span ranges and typical element sizes for lightweight steel joists in single-storey framestructures

Steel structures


Fig. 3.30 Basic plan arrangement for a single-storey steelframework in which the primary elements are stronggirders based on hot-rolled sections.

Fig. 3.29 Diagonal bracing schemes for single-storeyframeworks with lightweight elements,(a) Diagonal bracing in both horizontal and vertical planes.(b) Diagonal bracing in the vertical planes; roof claddingused as.diaphragm bracing in the horizontal plane.

normally used. A wide range of element typescan be adopted depending on the spaninvolved (Fig. 3.31). For medium spans, univer-sal or castellated beams are feasible for theprincipal elements and if the portal frameconfiguration is adopted a very wide span rangeis possible with these (15 m to 60 m). Latticegirders of various geometries, including planeand space trusses, in which the individual sub-elements have hot-rolled sections (rectangularor circular hollow sections, angles, channels,etc.), are also used over a wide range of spans.The spacing of the primary elements mustusually be restricted to a maximum of around6 m to 8 m, regardless of the primary span, sothat the secondary elements are not excessivelylarge. As in the case of frames with closelyspaced primary elements, a wide range of78

Fig. 3.31 Typical examples of primary elements in hot-rolled steelwork.

Portal frame

Parallel chord truss

Pitched truss

Direction spanof cladding

1.5-4.0 m

Mainframe

Secondary structure or purlin

4.0-6.0 m

15-60m

Depth of main frame (mm)

Span (m) Solid web Plane truss Space truss

10 450 1000 100015 600 1200 120020 700 1400 140030 900 1800 160040 1200 2500 220050 - 3000 280060 - 4000 380070 - 5000 480080 - 6000 5500

100 - 8000 6000

Main Frame

Span, Spacing, Height, Rafter Depth

L S H depth at knee

(m) (m) (m) (mm) (mm)

10 3 6 300 60015 4 720 5 8 450 80025 6 930 7 10 550 105035 8 1140 9 12 600 120045 10 1350 10 14 700 135055 12 1560 12 16 750 1600

Table 3.5 Approximate depths and span rangesfor trusses of hot-rolled sections in single-storeysteel frames

Table 3.6 Approximate span ranges and elementdimensions for steel portal frameworks

Steel structures

elevational roof geometries can be adopted buta regular plan layout must normally be adheredto. Table 3.5 gives an indication of the depthsrequired for the primary elements in parallelchord arrangements. Approximate sizes forportal frameworks are given in Table 3.6.

Frames of this kind are constructed with bothrigid and hinge-type joints and the bracingrequirements are dependent on the particularconfiguration of joints which is adopted. Acommon arrangement is illustrated in Fig. 3.32.

3.6.2.3 Two-way-spanning systemsIn two-way-spanning systems primary elementsare provided in two principal span directions.The structure can take the form of intersectingsets of single-plane elements, such as triangu-lated trusses (Fig.. 3.33), or of a fully three-dimensional triangulated space framework(Fig. 3.34).

Fig. 3.32 Bracing of single-storey frameworks, based on strong, hot-rolled primary elements.(a) This is a typical bracing arrangement for a portal-frame structure. The frame is self-bracing, by rigid joints, in theacross-building direction but triangulated bracing girders are required in the vertical and roof planes adjacent to the endsof the building to guarantee stability in the along-building direction.(b) Where a framework is not self-bracing in the across-building direction an arrangement of bracing girders must be

79provided in both horizontal and vertical planes.


(a)

Fig. 3.33 (a) Two-way-spanning roof structure formed byintersecting plane frameworks.(b) Willis, Faber & Dumas office, Ipswich, England, 1974.Foster Associates, architects; Anthony Hunt, structuralengineers. (Photo: Pat Hunt).

80

In the first type of system the elements mayintersect at right angles or obliquely. Theoblique grid pattern, which is called a 'diagrid'(Fig. 3.35), is the more efficient configurationbecause the girders which span across thecorners are relatively short and can provide ameasure of support for the longer spanninggirders which run between the corners.

True space frames, that is fully three-dimen-sional structures whose fundamental units arepyramids or tetrahedra (Fig. 3.34), allow amore efficient use to be made of material thanin the case of the intersecting plane-frame typebecause the disposition of the elementsproduces a more satisfactory distribution ofinternal forces. True space frameworks also can

(b)

Fig. 3.34 Fully three-dimensional two-way-spanningspace framework.

be given the rectangular or diagrid formdepending on the orientation of the elementswith respect to the principal span directions.

Space frameworks of both of the kindsdescribed above are highly statically indeter-minate and this, together with the fact thatthey are triangulated, allows very efficient useof structural material. They are, however,complicated to design and construct and aretherefore more expensive than the one-way-spanning systems described above. Trulythree-dimensional frames are economicallycompetitive with conventional structures atspans greater than 20 m; space frames whichconsist of intersecting plane frames are usedmainly in the span range 15 m to 20 m.

Fig. 3.35 Plan and cross-sectional arrangements forspace frameworks.

Because space frames are two-way-spanningsystems, the most suitable overall plan-form isa square in which the frame is supportedaround its entire perimeter on regularly spacedcolumns or walls (Figs 3.36 and 3.37). A squareplan-form with supports at the corners isanother common arrangement. The high degreeof statical indeterminacy allows an irregularpattern of support to be used, however, if this

Fig. 3.36 Support arrangementsfor space frameworks. Space frame-works are normally supported eitherby regularly spaced perimetercolumns around the entire perime-ter or by a symmetrical arrangementof columns or piers set back fromthe perimeter. The two-way-spanning capability of the spaceframe allows, however, an irregularpattern of support. This issometimes the reason for adoptinga space frame as the horizontalstructure in a building.

Steel structures

Double layer

Single layer Span up to 30 m

Span greater than 30 m

Diagrid

OrthogonalSpanDepth

= 30(approx.)

SpanDepth

= 40 (approx.)

Lattice beams

Square space frames supportedaround their perimeters bycolumns and deep lattice beams

Centralquadrupalcolumn

81

Support around perimeter Minimal support


82

Fig. 3.37 A Mero space frame, supported around itsperimeter, acts as the roof structure in this single-storeybuilding. Note the vertical-plane diagonal bracingelements which must be provided in each external wall ofthe building. The fully triangulated space framework actsas horizontal-plane bracing. [Photo: BICC]

is required, and space frames can even be givenan irregular overall form so long as the internalgeometry is fully triangulated.

One of the advantages of the space frame isthat the individual elements are small andlight. The parts can therefore be easily trans-ported to the site and no large equipment isrequired for erection; often the structure canbe assembled at ground level, by bolting orwelding, and subsequently lifted by a jackingprocess into its final position. This greatlysimplifies the construction.

The principal disadvantage of the spaceframe is its complexity. In particular, the highdegree of geometric complexity makes difficultthe design and manufacture of the joints, and

the high degree of statical indeterminacyaggravates this problem by requiring thatcomponents be manufactured to small toler-ances so that the 'lack-of-fit'12 is minimised.

The difficulties which are associated withconstruction can be reduced if a high degree ofstandardisation is adopted and the spaceframe is a type of structure which can be mosteconomically produced in the form of a propri-etary system. A number of these are currentlyavailable and they exploit the potentially highstructural efficiency of the space-frame config-uration while at the same time allowing areasonable level of cost to be achieved throughthe standardisation of components and bymeans of the economies of production-linemanufacture (Figs 3.38 to 3.41).Standardisation reduces the structural

12 See Macdonald, Structure and Architecture, Appendix 3 foran explanation of the 'lack-of-fit' problem.

Steel structures

Fig. 3.38 The elements of the Unistrut space deck system. This is based on lightweight cold-formed sections and aningenious jointing component. The efficiency with which load is transmitted through the joints is low, however, and thesystem is capable of relatively short spans only.

efficiency, but proprietary space-frame systemsare nevertheless normally significantly lighterthan conventional single-plane, one-way-spanning arrangements. They are also cheaperthan one-off space-frame systems and the

precision which can be achieved with economyby the use of production-line techniques infact makes the proprietary space frame one ofthe cheapest of the more sophisticated formsof structure.

Fig. 3.39 Spacedeckspace frame systemsupported around itsperimeter at the AberfanCentre, Wales. [Photo:Henry Morgan]

83


Fig. 3.40 (a) Basic components of the Mero space frame.This highly versatile system allows great freedom to thebuilding designer because the lengths of the elements arenot predetermined. It is possible to use this system tocreate a wide variety of structural forms including domesand vaults as well as flat space decks,(b) Church, Milton Keynes, England, with Mero spacedeckroof. [Photo BICC]

Standardisation normally imposes fairlysevere geometric constraints on the design,however, both on the overall form which canbe adopted and in the positioning and spacingof the supports. Some proprietary systems canbe used for flat decks only (Figs 3.38 and 3.39)but others are more versatile and can havemore complicated geometries (Figs 3.40 and3.41). Most proprietary systems are designedfor the short-span range (to cover areas of84

Fig. 3.41 Nodus joint. TheNodus joint is designed foruse with hot-rolled steelhollow sections. It isbasically a clamping devicewhich is used in conjunctionwith end pieces which arewelded to standard hollowsections. It allows a widerange of space frame geom-etries to be constructed andgives the designer freedom,not only to select differentframe configurations, butalso to vary the sizes of theconstituent elements.

Steel structures

around 10 m by 10 m to 30 m by 30 m on plan).Requirements for larger spans, or for morecomplex geometries, can frequently only besatisfied by use of a one-off, specially designedsystem. Such systems are usually all-weldedstructures with hot-rolled hollow sectionelements. An indication of the span capabil-ities and depths required for the systems illus-trated in Figs 3.38, 3.40 and 3.41 is given inTables 3.7, 3.8 and 3.9.

85

Table 3.8 Approximate depths and span rangesfor the Mero space framework

Span (m) Module (m) Frame depth (m)

up to 15 2 to 3 up to 1.515 to 27.5 2.4 to 3 1.5 to 2.127.5 to 36 2.4 to 3.6 2.1 to 2.536 to 50 3.6 to 4.8 2.5 to 4.050 to 100 4.8 to 6.0 3.6 to 4.8

Table 3.7 Dimensions and span ranges for theUnistrut space framework

Module Depth Span range

(m) (m) (m)

1.22 1.0 8.5 to 15.2 Column support at corners8.5 to 31.7 Support around perimeter

1.52 1.26 7.6 to 18.2 Column support at corners7.6 to 36.6 Support around perimeter

Table 3.9 Approximate depths and maximumspan ranges for the Nodus space framework

Module 2 to 3 m

Frame depth span Column support at corners20

span Support around perimeter15

Maximum span 50 m Column support at corners60 m Support around perimeter


Fig. 3.42 Terminalbuilding at StuttgartAirport, Germany. Theroof structure hereconsists of a series ofcomplex 'trees',constructed from steelhollow sections, whichsupport a regular grid ofsecondary elements onwhich the cladding ismounted. The sub-elements of the trees aresubjected to high levelsof bending load and it isthe great strength ofsteel which makes thistype of arrangementpossible. [Photo: A.Macdonald]

86

3.6.2.4 Frames with special geometriesThe overwhelming majority of single-storeysteel frameworks for buildings can be placedinto one or other of the categories describedabove-, these represent the most sensible,straightforward and economic ways of usingthe material. There are, of course, exceptions

which arise due to the existence of unusualdesign requirements or simply as a responseto the desire to produce an unusual orspectacular structure.

One example of a departure from the normalframe arrangement is a two-way-spanningsystem based on a series of structural 'trees'.

Steel structures

The basic elements of this type of structureconsist of a column unit (the 'trunk') which hasa rigid joint at its base, and a series ofcantilevering horizontal members (the'branches'). In most cases this primary struc-ture supports a secondary system of parallelbeams to which the roof cladding is attached.The plan of each 'tree' unit can have any shapebut is usually square with plan dimensions inthe range 15 to 40 m. A number of units areplaced alongside one another to give a build-ing of the required size. The advantage of thesystem is that each unit is entirely independ-ent structurally from its neighbours. This situa-tion allows for simple extension or alterationof the building, which is frequently the reasonfor its adoption. This is also a form ofconstruction which lends itself to very fasterection and the rapid production of aweathertight envelope.

Two notable buildings which have beenconstructed in this way in recent years are theterminals at Stuttgart (Germany, Fig. 3.42) andStanstead (England) Airports. The RenaultWarehouse building at Swindon, England, byFoster Associates (Fig. 2.6), and the Fleetguard

Fig. 3.43 Inmos Microprocessor Factory, Newport, Wales,1982. Richard Rogers Partnership, architects; AnthonyHunt Associates, structural engineers. A mast-and-tiearrangement is used here to achieve a long span with aone-way-spanning structural system. [Photo: AlastairHunter]

Factory at Quimper, France, have structuralarrangements which are a combination of thestructural tree system and the continuous two-way-spanning frame. In each of these cases thestructural trees are not independent units.

In the Fleetguard Factory the sizes of thehorizontal elements were kept low by the useof mast-and-tie systems. Mast-and-tie systemscan also be used in the context of one-way-spanning frame arrangements. Examples ofthis are the Inmos Factory, England by RichardRogers (Fig. 3.43) and the Ice Rink at Oxford,England by Nicholas Grimshaw (Fig. 3.44). Inall of these cases the tie rods or cables serveto provide a regular pattern of vertical supportfor the horizontal elements and may beregarded as substitutes for columns at these 87


Fig. 3.44 Ice Rink, Oxford, 1984. Nicholas Grimshaw &Partners, architects; Ove Arup & Partners, structuralengineers. The mast-and-tie system was used here toachieve a large column-free interior. The prevailing groundconditions were a factor in the selection of this particularstructural configuration. [Photo: Ove Arup & Partners]

points. The justification for the elaborate mast-and-tie system is normally that it allows alarge column-free interior to be achieved,either because this is a space-planningrequirement (as was the case at the Oxford IceRink) or because great flexibility is needed inthe planning of the interior arrangements ofthe building.

3.6.3 Low- and medium-rise multi-storeyframes

3.6.3.1 IntroductionIn low- and medium-rise multi-storey frames,that is frames with up to 30 storeys, the princi-pal structural factor which affects the planarrangement is the need to provide a suitablesystem of beams and columns to support the

gravitational load on the floors. The need forthe structure to have adequate lateral strengthto resist wind load is, of course, also an impor-tant consideration, particularly with the higherstructures.

3.6.3.2 The floor gridIn most multi-storey frames the floors are ofreinforced concrete. A number of differenttypes of floor slab are used including in situ orprecast concrete slabs with simple rectangularcross-sections, voided precast slabs of variouskinds, and in situ slabs which are cast on apermanent formwork of either steel or precastconcrete (Fig. 3.45). The concrete slabs arenormally made to act compositely with thefloor beams to improve the overall efficiency ofthe structure (Fig. 3.46). The floor slabs arealmost invariably of the one-way-spanning typeand must be supported on a parallel system ofbeams spaced 2 m to 6 m apart depending onthe slab type.

The column grid can take various formsdepending on the space-planning require-ments of the building. A very common basic88

Fig. 3.45 Typical floor-slab systems for steel frameworks.(a) In situ reinforced concrete flat slab.(b) Precast concrete floor units.(c) In situ concrete on profiled steel permanent formwork.(d) Composite precast and in situ concrete.All of these systems are normally one-way-spanning andrequire to be supported on a parallel arrangement of steelbeams. System (a) can be two-way-spanning.

arrangement is one in which the columns arepositioned at every third or fourth beam andintermediate beams are carried on a secondsystem of beams which run at right angles tothe direction of the main beams (Fig. 3.47).This produces a column grid which is rectangu-lar. A variation of this is the grid in whichcolumns are provided at every main beamwhere these meet the perimeter of the build-ing, and at every third or fourth beam in thebuilding's interior (Fig. 3.48). This eliminates

Fig. 3.46 Shear studs welded to the top flanges of steelfloor beams allow composite action to be developedbetween the beams and the floor slabs. This reduces thesize required for the beam. As is shown, such action ispossible with both in situ and precast floor systems. 89

Steel structures


Fig. 3.47 Typical layout of floor beams in a steel framework(a) The basic unit consists of a series of secondary beamsin a parallel arrangement at equal spacing. The spacingdepends on the span capability of the floor-slab system.The secondary beams are used in conjunction with primarbeams to give a column grid which is more-or-less square.(b) Variations to the basic arrangement are used to accommodate space-planning requirements.

Fig. 3.48 A common variation on the basic floor gridarrangement is the provision of columns for eachsecondary beam at the building' perimeter. The closelyspaced columns eliminate the need for a separate claddingsupport system.

the need for a secondary system of mullions inthe external wall on which to mount thecladding.

The types of steel sections which are usedfor the elements in the floor grids depend onthe span and on the intensity of the load. Forgrids in the span range of 4 m X 6 m t o 6 m X12 m, carrying office-type loading, universalbeam or castellated beam sections arenormally adopted; an indication of the depthsof beam which are required is given in Table3.10.

For economy in construction and simplicityin the detailing of the structure, standardisa-tion of elements is desirable. This is facilitatedif the spans of the principal elements are keptconstant. Where beams with different strengths

are required, a constant floor depth issometimes achieved without compromisingefficiency by the use of different weights ofbeam within a given serial size.

Where a larger column grid is required thanis possible with standard sections, that is agrid with dimensions in excess of around 20 m,various forms of built-up sections can be usedto provide stronger beams. Plate girders andcompound beams allow the span range to beextended but for very large spans latticegirders are required. In cases where very longspans must be achieved it is frequently advan-tageous to maintain a basic floor grid, basedon standard sections, and use a tertiary systemof girders to achieve the long span (Fig. 3.49).

The depth of long-span structures tends tobe large and various measures can be adoptedto minimise the amount of unusable volume90

Primary beam

Secondarybeam

Normally around 8 m(maximum 30 m)

Normally around 3 m(maximum 6m)

Table 3.10 Typical element sizes in multi-storey steel frameworks

Floor Floor beam depth (mm) Column cross-section

beam (H-section or square

span I-section I-section in Castellated hollow section) (mm)

(m) compositeStorey height (m)

structure

3 4 5 6

468

101214

250400530760900900

200380450600760830

550600900

11501370

150150200200250250

150 150200 200200 200250 250250 250250 300

150200250250250300

building so that the girders can be incorpor-ated into walls, as in the staggered-truss andinterstitial-truss systems (Fig. 3.50).

3.6.3.3 Bracing systems for medium-rise, multi-storey framesAs in the case of single-storey frames multi-storey frames can be of either the 'simple' orthe rigid type. Where the joints are rigid, theframe is self-bracing and the need to providelateral stability is not a major factor in thedetermination of the overall form. Wherehinge-type joints are used it is normal practicefor all beam/column connections to be of thehinge type and for the columns to be jointed atevery second storey level, also with hingejoints (Fig. 3.51). This fully hinged configur-ation has a number of advantages: it simplifiesboth the analysis and the erection of the struc-ture, and it also allows the accommodation ofthermal expansion and minor foundationmovements without the introduction of stressinto the steelwork. In its basic form this type offrame is highly unstable, however, andadditional bracing must be provided.

To render the 'simple' type of frame stable, alimited amount of vertical-plane bracing, of thediagonal or diaphragm types, must be incorp-orated into the arrangement in two orthogonal(mutually perpendicular) directions and thismust be linked to all other parts by horizontal-plane bracing at every level. The action of amulti-storey frame in response to wind loadingis shown diagrammatically in Fig. 3.52.

91

Steel structures

Fig. 3.49 Where long spans are required in multi-storeyframeworks it is normal to provide a tertiary system ofbeams. In the example illustrated, triangulated girders,spanning the entire width of the building (shown dottedon the plan), are substituted for the columns at points A.[after Hart, Henn and Sontag]

which is created within the building. Wherevery deep lattice girders are involved, forexample, the depth of these can sometimes bemade the same as the storey height of the

Fig. 3.50 Where long spansrequire the use of deep tri-angulated girders this cangreatly increase the struc-tural depth of the floors. Thetwo systems shown hereallow deep girders to beaccommodated without anincrease in storey height.The penalty which is paid forthis is that it places restric-tions on space-planning.


provided in the form of self-supporting coresof masonry or reinforced concrete (Fig. 3.54),which are normally constructed in advance ofthe erection of the steelwork and which areused to carry vertical as well as horizontalloads.

The reinforced concrete floors of the struc-ture are usually capable of acting as horizon-tal-plane bracing and the need for this doesnot usually affect the general planning of thebuilding. Where a lightweight roof cladding isused, a horizontal-plane girder (normally of thetriangulated type) is provided at the topmostlevel but this can be accommodated within theroof structure and does not usually present aplanning problem.

To sum up, two factors must be consideredin the planning of low- or medium-rise multi-storey frames. These are firstly, the geometry ofthe floor grid and its relationship to thecolumn grid and secondly, the provision ofbracing. A general arrangement must bedevised which both produces a satisfactorystructure and allows the space-planningrequirements of the building to be met.

3.6.3.4 Multi-storey frames with special geom-etriesAlthough most of the steel frames which areconstructed in practice have characteristicswhich are similar to those which are describedabove, the need for a frame which has quitedifferent characteristics does sometimes arise.

Fig. 3.51 Typical arrangement of joints in a hinge-jointedmulti-storey framework.

The disposition of the vertical-plane bracingis a factor which affects the general planningof the building. It is normally positioned in assymmetrical an arrangement as possible andconvenient locations for diagonal bracing arethe perimeter walls or around stair wells, liftwells and service cores (Fig. 3.53). Ifdiaphragm bracing is used it is frequently92

Staggered Interstitial

Steel structures

Fig. 3.52 Resistance of a multi-storey framework to wind load.(a) The wind loading acts on the cladding and is transmitted by the cladding support system to the floors of the building.(b) In this plan view the wind loading appears as a distributed load on the edge of the floor slab, which is restrained bythe vertical-plane bracing.(c) The vertical-plane bracing is shown here independently of the rest of the frame. The loads received from each floor areindicated. These are transmitted to the foundations by the bracing girder (which consists of the columns forming the bay,together with the diagonal bracing elements).

Fig. 3.53 The locations of vertical-plane bracing must be compatiblewith space-planning requirements.Typical arrangements, in which verti-cal-plane bracing is located in theperimeter walls or around servicecores, are shown here.

93Vertical-plane bracing

Vertical-plane bracing

(b) Floor slabs transmit wind load tovertical-plane bracing

(a) Cladding transfers wind load tofloor slabs

Wind load

(c) Vertical-plane bracing transmitswind load to foundations


Fig. 3.54 Where diaphragm-type vertical-plane bracing isused it is normally constructed in either masonry orreinforced concrete. It can consist simply of infill walls inthe steel framework or, as is shown here, it can take theform of free-standing cores to which the steelwork isattached. In the latter case the cores are normallyconstructed to the full height of the building in advance ofthe steelwork being erected. As these cores normallycontain stairs they provide access to all parts of the build-ing throughout the construction sequence.

An example of a special requirement is thesituation in which the lowest storey of a build-ing must have a different plan geometry fromupper floors because it contains a radicallydifferent type of accommodation. This cancreate the need for a special frame, especiallyif, as is often the case, there is a requirementto reduce the number of columns in the loweststorey. One way in which this is achieved is bylocating a number of very large girders at first-floor level so as to transmit the loads from thecolumns in the upper storeys to a smallernumber of columns in the ground floor.Another arrangement which has been used isthe suspended frame (Fig. 3.55). This has theadditional advantage that it reduces the size ofthe majority of the vertical elements becausetension elements have a smaller cross-sectionthan equivalent columns. The adoption of thistype of system therefore increases the usablefloor area per storey.

3.6.3.5 High-rise multi-storey framesThe dominating factor in the design of high-rise frames is the need to provide sufficientlateral strength and rigidity to resist wind loadeffectively. As in the case of low- and medium-rise multi-storey buildings, the floor must bedesigned to provide effective resistance togravitational loads and to act as a horizontal

Fig. 3.55 Suspended frame arrangement. The floors hereare suspended from one or other of two massive trusseswhich are supported on a central core. The total volume ofvertical structure is theoretically lower than with the trad-itional column grid arrangement because the concentrationof the compressive function into one massive centralelement allows higher compressive stresses to be used(because the slenderness of the single compressive elementis relatively low). The structure of the HongkongBankHeadquarters (Fig. 1.3) is based on this principle.

94

Fig. 3.56 Belted trusssystem.(a) The diagonally braced coreis here carrying all of thehorizontal load.(b) The introduction of therigid girder at the top of thebuilding allows the perimetercolumns to act compositelywith the core to resist thehorizontal load.

Steel structures

(a) (b)

diaphragm to conduct wind forces to the verti-cal-plane bracing. In the case of high-risestructures more attention is given to the latterfunction but the floor grids which are adoptedare nevertheless similar to those which areused in lower structures.13

For frames with more than 30 storeys,neither rigid-frame action nor diagonal ordiaphragm core-bracing are capable of provid-ing sufficient lateral strength in the verticalplane to resist wind load effectively. A combin-ation of core diaphragm bracing and rigidjoints provides sufficient lateral strength forbuildings of up to 60 storeys in height; thebelted truss system (Fig. 3.56), which allowsthe core to act compositely with the perimetercolumns of the building and which reduces thebending moment on the core due to the lateralloads, is another arrangement which issuitable for buildings in this height range. Forhigher buildings the lateral strength isincreased by making use of the full cross-sectional width of the building, so that itbehaves like a single vertical cantilever.

Various techniques have been used to achievethis. In the framed-tube system (Fig. 3.57),which was used in the World Trade Centrebuilding in New York, the closely-spacedperimeter columns form a cantilever tube;interior columns play no part in the resistanceof lateral load in this system and the whole ofthe wind load is resisted by the external walls.The walls which are parallel to the direction ofthe wind are rigid frames and form a shearconnection between the walls which arenormal (at right angles) to the wind direction,which act as flanges. The building is thereforea box girder which behaves as a verticalcantilever in response to lateral load. Theaction of the trussed tube (Fig. 3.58) is similar.In the case of the tube-in-tube structure (Fig.3.59) the stiffness of the tube, which is formedby the perimeter walls, is augmented bycomposite action with a core tube. Thisrequires that floor structures be stiff enough toprovide a shear connection between the twotubes. Bundled-tube structures (Fig. 3.60),which were developed to reduce the shear-lageffect which occurs in single-tube construction,produce the stiffest buildings for a given size ofcross-section.

9513 See, Schueller, High Rise Building Structures, John Wiley,

London, 1977.


Fig. 3.57 In the framed-tube system the columns on theperimeter of the building are closely spaced. The buildingacts as a vertical cantilever in response to wind loading.The cross-section of the cantilever is a rectangular tubeformed by the closely spaced perimeter columns. Those onthe windward and leeward faces of the building act asflanges while the walls which are parallel to the windprovide a shear connection between these by rigid frameaction. The World Trade Centre buildings in New York areperhaps the most well-known buildings to be based onthis principle. The lower diagram here gives an indicationof the variation in the level of load which occurs betweencolumns. It will be seen that the effectiveness of columnsin the windward and leeward walls is affected by theirproximity to the walls which provide the shear connection.The phenomenon is known as shear lag (see Fig 3.60).

3.6 .4 Cable structures

The most efficient forms of structure arethose which are stressed in pure tension.

Fig. 3.58 The braced-tube system. The action of this issimilar to the framed-tube but in this case the shear loadbetween the windward and leeward walls is carried bydiagonal bracing elements.

The steel cable structure, in which theprimary load carrying elements are flexiblecables acting in pure tension, is an exampleof this type. Because these are very efficientstructures they are capable of very long spans(Fig. 1.13).

The key feature of cable structures is highflexibility which allows changes in geometry tooccur in response to variations in load andmaintains the state of purely tensile stresses.The extent to which a geometric change can bepermitted in a building is limited, however,96

Steel structures

Fig. 3.60 The bundled-tube system. In this system,closely-spaced columns on the perimeter of the buildingact in conjunction with rows of closely-spaced columnswhich cross the interior. These provide additional shearconnections between the windward and leeward walls andreduce the phenomenon of shear lag.

(Figs 3.61 and 3.62). Multi-cable systems havethe advantage that the cables can be stressedagainst each other (prestressed) which furtherrestricts the amount of movement which canoccur in response to variations in load. Thehigher the level of prestress the more rigiddoes the structure become.

Although a cable structure will always takeup the form-active shape14 of the loads whichact on it, this does not mean that the designerhas no control over its shape, because differ-ent geometries can be produced by altering the

9714 See Appendix 1 for an explanation of the term 'form-

active'.

Fig. 3.59 The tube-in-tube system. In this system closely-spaced columns on the perimeter of the building actcompositely with the core to resist horizontal load.

and cable structures must be designed so thatdrastic changes in shape, such as might occurdue to a complete reversal of load under theaction of high wind suction, must beprevented. A number of techniques are used toachieve this. Most are based on the use of twosets of cables, which are arranged so that thestructure can resist forces from all directions


Fig. 3.62 Double-skin cable system with two sets ofcables separated by short compressive elements. This typeof arrangement is an alternative to single-skin systemswith anticlastic shapes.

Fig. 3.61 These cable networks can resist load fromdifferent directions without undergoing gross changes inform due to their having two sets of cables which arecurved in opposite directions to form an anticlasticsurface. They are each supported on arch-type compressiveelements. This type of cable system can also be supportedon an arrangement of masts (see Fig 1.13).

indeterminacy which is present, due, in part, tothe problem of predicting their final shapeunder the action of a particular load. Analysisby computer must frequently be supplementedby model analysis. Constructional detailingproblems arise from the difficulty of providingsuitable end anchorages for the tensionmembers (one of the classic problems ofengineering) and from the need to constructthe structure in such a way that it can accom-modate the small changes in geometry whichwill occur in response to changes in load. Thisnormally requires the use of complicated jointcomponents, which must be speciallymachined and are therefore expensive. Thecladding system also must be able to adjustitself without damage to the movements whichwill occur and this involves the design of jointswhich are both weathertight and of adequatestrength, and which can accommodatemovement. Cladding systems tend therefore tobe fairly sophisticated and to have poordurability. The consequences of all of thesecomplexities is that cable structures are veryexpensive; they have been most often used forhigh-prestige temporary buildings such asexhibition halls.

support condition. Networks which aresupported on masts, for example (Fig. 1.13),have a different geometry from those which arecontinuously supported on arches. Variationsin geometry are also achieved by making use ofdifferent cable arrangements (multi-layersystems as opposed to single-layer systems,for example). The choice of geometriesprovided by the use of cable structures islimited, however, because all of them must bebased on doubly-curved forms; all cable struc-tures tend, therefore, to have a similar appear-ance.

Cable structures are highly complex, both toanalyse and to construct. The difficulty in theiranalysis results from the high degree of statical

98

Chapter 4

Reinforced concrete structures

4.1 Introduction

Concrete is an extremely versatile structuralmaterial. It is moderately strong in compres-sion but weak in tension; it has good resis-tance to fire and good durability. Perhaps itsmost distinctive characteristic, however, is that

it is available to the builder, on the buildingsite, in semi-liquid form. This has two veryimportant consequences. First, it allowsconcrete to be cast into a wide variety ofshapes; the material itself places little restric-tion on form. Secondly, it makes possible theincorporation into concrete of other materials,

Fig. 4.1 Church, Vienna,Austria, 1965-76. FritzWotruba with Fritz G. Mayr,architects. The expressivepossibilities of reinforcedconcrete are well illustratedhere. [Photo: E. & F.McLachlan]

99

Fig. 4.2 Hennebique's system for reinforced concrete,which was patented in 1897, was one of a number whichwere developed in the late nineteenth century and whichled to the subsequent widespread use of the material formulti-storey structures.

with which its properties can be augmented.The most important of these is steel, in theform of small-diameter reinforcing bars, andthis produces the composite materialreinforced concrete, which possesses tensileand flexural strength as well as compressivestrength. Reinforced concrete can therefore beused to make any type of structural element.

Concrete can be either cast directly into itsfinal location in a structure, in which case it issaid to be in situ concrete, or used in the formof elements which are cast at some otherlocation, usually a factory, and simply assem-bled on site, in which case it is referred to asprecast concrete. The relative advantages ofthe two types are reviewed in Section 4.4. Bothin situ and precast concrete can be produced inordinary reinforced form or in pre-stressedform. The distinctions between these types arediscussed in Section 4.3.2.

4.2.1 The aesthetics of reinforced concreteThe opportunities which reinforced concreteoffers in the matter of architectural form canbe seen by examining the range of buildingtypes for which it has been used during therelatively short period in which it has beenavailable as a structural material.

Although a type of concrete was used togood effect by the architects and engineers ofRoman antiquity the modern material,reinforced concrete, dates from the nineteenthcentury. 'Roman' cement, the forerunner of thepresent-day Portland Cement, was patented inthe UK by ). Aspden in 1824, but the earliestuses of reinforcement in concrete appears tohave occurred more or less simultaneously inFrance and USA where, in the 1880s and 1890s,Francois Hennebique (Fig. 4.2) and ErnestRansome, respectively, each developed framingsystems for buildings based on this principle.Another early innovator was Robert Maillart,whose development of the two-way-spanning

4.2 The architecture of reinforcedconcrete - the factors which affectthe decision to select reinforcedconcrete as a structural material

100


flat-slab structure was particularly notable(Fig. 4.3 ). All of these early pioneers wereconcerned principally with the application ofreinforced concrete framing systems to indus-trial buildings.

In the late nineteenth century reinforcedconcrete was a 'new' structural material,capable of producing durable and fire-proofskeleton frameworks and therefore buildingswith open interiors free from structural walls. Itarrived on the architectural scene at a timewhen the precursors of the Modern Movementwere exploring the possibilities of creating anew architectural language which would beappropriate for the twentieth-century world.These architects were anxious to make use ofthe new materials which industry was produ-cing and the most innovative of them were notslow to appreciate the potential of reinforcedconcrete.

Among the earliest of designers to under-stand the purely architectural qualities of thenew material was August Perret. In the apart-ment block at 25 bis Rue Franklin, Paris, 1902(Fig. 4.4), the adoption of a reinforced concreteframe structure was used to produce an open-plan interior with light non-loadbearing parti-tion walls. Large areas of glazing were a featureof the exterior and the reinforced concretecolumns of the building, although not actuallyexposed (a tile cladding system was used),were expressed on the facade. The later garageat 51 Rue de Ponthieu by Perret was alsobased on reinforced concrete and in this build-ing the concrete framework was left entirelyexposed, apart from a thin layer of paint. Thesebuildings were very important precursors of theModern Movement. An important aspect oftheir novelty was the role played by the struc-ture in liberating the organisation of spacefrom the tyranny of the loadbearing wall. Thiswas exploited by Perret in his design for theRue Franklin flats, producing a version of theParis apartment in which a new free-flowingspace could be enjoyed. The other significantaspect was the re-establishment of structure,and in particular the column, as a part of thearchitectural expression, something which hadnot happened since the eclipse of Neo-

101

Fig. 4.3 Cross-section through five-storey flat-slab struc-ture by Robert Maillart, c. 1912. Flat-slab systems, whichare sophisticated two-way-spanning structures whichderive much of their strength from the high degree ofstructural continuity which is present, were developedsurprisingly early in the history of reinforced concreteconstruction. Because there are no downstand beams thesystem allows a very economical use to be made of bothmaterial and labour. Local thickening is required in thevicinity of the columns, where shear forces are high. This isaccomplished by the use of 'mushroom' column heads.



Fig. 4.4 Apartment block, 25 bis rue Franklin, Paris,1902-05, Auguste Perret, architect. By using a reinforcedconcrete frame structure Perret was able, with this build-ing, to redefine the layout and fenestration of the tradi-tional Paris apartment. The building also made asignificant contribution to the development of a new archi-tectural aesthetic based on the technology of the skeletonframe.

legacy from his master, but reinforced concretewas, for Le Corbusier, the ideal structuralmaterial. Its excellent structural propertiesplaced little restriction on architectural formbut above all other considerations were thenewness of the material and the fact that itwas a product of an industrialised process. Itwas therefore an appropriate medium withwhich to develop an architectural language forthe twentieth century - the age of rationalism,industry and social idealism.

That Le Corbusier appreciated the physicalproperties of reinforced concrete is welldemonstrated by his famous drawing of thestructural core of the Domino House of 1914(Fig. 4.5). This showed that he understood thetwo-way-spanning capability of the materialand knew of its ability to cantilever beyond theperimeter columns. His realisation that theintegral stair of this small structure could beused to brace it in the two principal directionsled to the adoption of very slender supportingcolumns. The structure therefore causedminimal interference to the layout of theinterior of the building or to the treatment ofthe exterior. The stair could, of course, havebeen positioned anywhere in the plan.

102

classicism by Romanticism in the earlynineteenth century.

Le Corbusier, who was for a time a pupil ofPerret, was another early pioneer of the archi-tectural use of reinforced concrete. His enthu-siasm for the material was, to some extent, a

Fig. 4.5 This drawing depicts the structural carcass of LeCorbusier's Domino house of 1914 and demonstrates thathe had a complete, probably intuitive, understanding ofthe structural potential of the material. The renowned 'fivepoints', which were to have such a profound influence onthe architecture of the twentieth century, were developedfrom this.


The planning freedom which such astructural system offered the architect wassummarised by Le Corbusier in his well-known'Five points towards a new architecture'1 of1926 and exploited by him in the designs forthe houses which he built in the 1920s, cul-minating in the Villa Savoye of 1929 (Fig. 4.6).Most of these buildings did not in fact havethe two-way-spanning ribless slab of theDomino House but were based onbeam/column frame arrangements with one-way-spanning slabs. The planning freedom

1 The renowned 'five points' were first declared inAlmanack de I'Architecture moderne, Paris, 1926. The most

significant of them were: separation of structure andskin, freedom in the plan and freedom in the treatmentof the facade.

Fig. 4.6 Villa Savoye, Poissy, France, 1929. Le Corbusier,architect. The technology of the reinforced concrete frame-work contributed significantly to the visual language devel-oped by Le Corbusier in the 1920s. [Photo: AndrewGilmour]

offered by the use of a concrete framework wasnevertheless fully exploited.

The origins of the visual vocabulary whichwas evolved by Le Corbusier in the 1920s havebeen well documented elsewhere.2 Amongthese was the idea that Modern architectureshould be tectonic (in other words that thestructure should influence the architecture)and that it should symbolise rationalism. It is

2 For example, Curtis, W. J. R. Le Corbusier: Ideas and Forms,

Oxford, 1986. 103


104

Fig. 4.7 The Bauhaus, Dessau, 1926, Walter Cropius,principal architect. This building had a reinforced concreteframe structure which provided the same opportunities fornew architectural expression as had been recognised by LeCorbusier. The glazed curtain wall which rises through theupper floors of the building represents an exploitation ofone of those opportunities. [Photo: Paul H. Gleye]

perhaps as a result of the influence of thesecond of these that Le Corbusier restrictedhimself, in these early buildings, to rectilinearforms and did not avail himself of the fullpotential of reinforced concrete to providealmost unlimited freedom in the matter ofform. The rectilinearity not only symbolisedreason; it also represented a sensible andtherefore rational way of using reinforcedconcrete to provide the structure for a build-ing. This is an example of the 'structure

accepted' relationship between structure andarchitecture (see Section 2.2).

Another of the icons of early Modernism, theBauhaus building in Dessau (1926) (Fig. 4.7) bya team of architects led by Walter Gropius,also had a reinforced concrete frame structure,the layout of which was very straightforward. Inthe principal block, a rectangular grid ofcolumns supported a series of deep beamswhich spanned across the building and carrieda one-way-spanning hollow-block floor.3 Thefloor deck was cantilevered slightly beyond theperimeter columns and this allowed the steeland glass curtain wall on the exterior to run

3 The hollow-block floor is a form of reinforced concreteribbed slab in which ceramic blocks act compositelywith in situ reinforced concrete.


continuously through the three upper storeysof the building. The steel mullions whichsupported the curtain wall were mounteddirectly on the edge of the floor slab. Theopaque parts of the exterior wall wereconstructed in masonry which was rendered togive a smooth white finish. The columns andbeams of the reinforced concrete frameworkwere left exposed throughout the interior.

The architectural vocabulary of the Bauhausbuilding was very similar to that which wasbeing developed contemporaneously by LeCorbusier. Its characteristics were due, in nosmall measure, to the use of a reinforcedconcrete framework which allowed greatfreedom to be exercised in both the internalplanning and the treatment of the exterior.

An early demonstration in Britain of thearchitectural qualities of reinforced concreteoccurred in the Boots Warehouse at Beestonby Owen Williams (1930) (Fig. 4.8). In this casea two-way-spanning slab was adopted,supported on a square grid of columns withexaggerated mushroom heads, but visually thebuilding was similar to the earlier examplesdescribed above. A further example was theHighpoint 1 apartment block by BertholdLubetkin of the Tecton Group and Ove Arup(1933). Lubetkin and Arup had earlier demon-strated their awareness of the capabilities ofreinforced concrete in the buildings which theydesigned for London Zoo - most notably andeloquently the Penguin Pool of 1933 (Fig. 4.9).In the Highpoint block a loadbearing-wall typeof structure was adopted with one-way-spanning floor slabs. This arrangement wasused to provide variations in the floor plans atdifferent levels. Where wall-free spaces wererequired, the loadbearing reinforced concretewalls in the floor above were made to act asstorey-height beams spanning betweencolumns. The device is seen at its mostextreme in the ground floor where theloadbearing walls were eliminated entirely andthe vertical support was provided solely by agrid of columns.

The visual language developed in thesehighly influential early examples of twentieth-century rationalist architecture owed much to

Fig. 4.8 Boots Warehouse, Beeston, England, 1930-32,Owen Williams, architect/structural engineer. The structurehere is a two-way spanning flat slab with mushroom-headcolumns. Its prominence and the transparency of the wallsproduce a similar tectonic quality to that seen in theBauhaus (Fig. 4.7) and place the building in the forefrontof contemporary architectural development.[Photo: BritishArchitectural Library Photographs Collection]

the properties of the new structural material bywhich they were supported. In all of them therelationship between structure and architec-ture was that of 'structure accepted' (seeSection 2.2), in which the visual and structuralprogrammes were allowed to co-exist withoutconflict. The forms were predominantly recti-linear; this made easy the construction of theformwork in which the liquid concrete wouldbe cast and, given the spans involved, was asensible choice of structural form.

Rectilinearity was also an appropriateaesthetic choice given the ideas which thevisual vocabulary was intended to express. In 105


106

Fig. 4.9 Penguin Pool, London Zoo, London, England,1933. Berthold Lubetkin and Tecton, architects; Ove Arup,structural engineer. The expressive possibilities of continu-ous structures of reinforced concrete are most eloquentlydemonstrated here. [Photo: E. & F. McLachlan]

all of the buildings some parts of the structurewere visible; in some buildings the structure inits entirety was visible. This option was madetechnically possible by the fire-resistingproperties of concrete and also by its durabil-ity, and was an important aspect of the visualquality of the buildings. In addition to the factof its exposure, the tactile nature of theexposed structure was eminently suited to theexpression of the Modernist ideals of the'honest' portrayal of the constituents of abuilding. Similar developments were occurringsimultaneously in steel-framed buildings, aswas shown in Chapter 3, but the architectural

language of reinforced concrete was subtlydifferent from that of steel, due mainly to itsdifferent structural properties, to the fact thatit could be left exposed in the finished build-ing and to the quite different ways in whichsteel and reinforced concrete buildings wereconstructed. Concrete therefore made adistinct contribution to the developinglanguage of early architectural Modernism.

In his buildings of the 1940s and 1950s, LeCorbusier introduced a new element into thevocabulary of reinforced concrete. This wasboard-marked exposed concrete ('beton brut')4

All concrete does, of course, bear the marks ofits formwork, but in the case of beton brut the

4 Exposed concrete which bears the marks of theformwork in which it was cast (literally 'rough', 'raw','unfashioned', 'unadulterated').


Fig. 4.10 Monastery ofLa Tourette, Eveux,France, 1955. LeCorbusier, architect.Exposed structuralconcrete, which bears themarks of the roughformwork against whichit was cast, was a notableaspect of the architec-tural vocabularyemployed by LeCorbusier in this period.It was compatible withhis architectural thinkingat that time but was alsoideally suited to therequirements of thisparticular client andproject. [Photo: AndrewGilmour]

formwork was very crudely constructed andproduced textures which were rough anduneven. The device was first used by LeCorbusier in buildings such as the 'PetiteMaison de Weekend' of 1935 but the mostnotable of his buildings in which it wasemployed were the 'Unite d'Habitation',Marseilles, of 1947-52 and the monastery of LaTourette, near Lyons, of 1957-60 (Fig. 4.10). Inthese later buildings the crudity of theconstruction was to some extent caused by the

inexperience of the building contractors butthere is no doubt that Le Corbusier himselffound that the suggestions of 'savageness' and'the primitive' which it produced were compat-ible with his architectural thinking at that time.

The resulting style was to be much imitated,a prominent British example being theNational Theatre of Great Britain (1967-76),London (Fig. 4.11) by Denys Lasdun, and wasentirely compatible with the fashion for 'NewBrutalism' which had emerged in Britain in the 107


108

Fig. 4.11 National Theatre of Great Britain, London,1967-73, Denys Lasdun, architect. This building is a veryprominent example of the type of Corbusier-inspiredarchitecture which was fashionable in the 1950s and 1960sin which the expressive possibilities of roughly texturedexposed reinforced concrete were explored. (Photo: DensyLasdun Peter Softley and Associates).

1950s. Another British example was theEconomist Building, again in London (1964),by Alison and Peter Smithson. Frequently, thetextures involved were deliberately contrived tobe extremely rough and thus was a new andsavage element introduced into the otherwisesmooth and refined vocabulary of architecturalModernism.

The architectural language developed by LeCorbusier, Gropius, Lubetkin, Williams andothers in the 1920s and 1930s, and extended inthe 1940s and 1950s by Le Corbusier, Alisonand Peter Smithson and others, was destinedto be imitated in almost every city in the world.In virtually all of the buildings involved,whether the seminal buildings in which thelanguage was evolved or the many imitations,the relationship between the structure and thearchitecture fell into the category of 'structureaccepted'. The properties and requirements ofthe reinforced concrete structure were, in other

words, given an important role in influencingthe visual language of the architecture.

Not all of the great reinforced concretebuildings of the twentieth century fall into thecategory described above, however. LeCorbusier himself did occasionally make fulland exaggerated use of the expressive possibil-ities of reinforced concrete; perhaps the mostfamous example of this occurred at the chapelat Ronchamp (Fig. 4.12). Here the relationshipbetween the structure and the architecture wasone of 'structure ignored'.5 Structural consider-ations played little part in the evolution of theform of the building. The structural systemwhich was adopted was nevertheless relativelystraightforward, largely due to the excellentstructural properties of reinforced concrete.The distinctively shaped roof is nothing morethan a beam/column framework with a one-way-spanning slab. In this case the mouldabil-ity of the concrete and the structural continuityit offered were used to produce a slab withdouble curvature to form the remarkablecanopy. The beams which support this were

5 See Section 2.2.


concealed by the upward sweep of the roof andthe columns were buried within the self-supporting masonry walls.

It should be noted, however, that the relativeease with which this complex form wasachieved, using a very basic type of structure,was due to the small span of around 20 m. Hadthe building been significantly larger, a moreefficient type of structure would have beenrequired ('semi-form-active' or 'form-active'). Inthat case it might, for structural reasons, havebeen necessary to adopt a slightly differentshape of roof and Le Corbusier would havefound himself in the position of having toaccept restrictions on his total freedom toinvent the form of the building.

Another famous building which illustrateswell the excellent structural properties ofreinforced concrete is the 'Falling Water' houseby Frank Lloyd Wright (1936) (Fig. 4.13). Aswith the chapel at Ronchamp this is anexample of 'structure ignored'. The oversailingand apparently free-floating horizontal

Fig. 4.12 Notre-Dame-du-Haut, Ronchamp, France, 1954.Le Corbusier, architect. This form, with its elaborate roofcanopy which does not actually touch the tops of the walls(there is a gap between roof and walls which is used togood effect to introduce light into the interior) could onlyhave been constructed in reinforced concrete. It illustratesvery well the freedom to invent form which the materialconfers on the architect. [Photo: P. Macdonald]

platforms which are one of the principalfeatures of this architecture, and which areessential for conveying the ideas of free-flowing space, the interpenetration of solidand void and the blurring of the boundarybetween exterior and interior, were not idealfrom a structural point of view. Their feasibilitydepended on the use of two-way-spanningreinforced concrete slabs supported on anetwork of columns. As at the Villa Savoye,the decks are not solid slabs but networks ofbeams and cantilevers supporting a thin floorslab. Some of the cantilevered beams are ofmassive proportions and, as at Ronchamp, the 109


Fig. 4.13 Falling Water,the Bear Run,Pennsylvania, 1936. FrankLloyd Wright, architect.The cantilevered balconiesof this building requiredthe use of a high-strengthstructural material whichwould allow a high degreeof structural continuity tobe developed. Reinforcedconcrete was the idealmaterial in which torealise this form. [Photo:Andrew Gilmour]

110

size of the building is close to the limits ofwhat is feasible with the method of construc-tion employed.

It is interesting to note that, unlike the caseof the rationalist buildings of Le Corbusier,Gropius, Lubetkin and others, in which thelogic of the structure was accepted andallowed to influence the architecturallanguage, neither of these buildings in which

structure was virtually ignored when the formwas determined was destined to be imitated toany significant extent.

In the building boom which followed the warof 1939-45 many buildings which conformed tothe mainstream of architectural Modernism, asit had been developed in the 1920s and 1930s,were erected with reinforced concrete struc-tures. The concrete was often left exposed and


expressed on the exterior, especially followingthe advent of the fashion for 'Brutalism'. Muchof this architecture was of indifferent quality,however, and as a consequence, the term'concrete jungle' was added to the language ofarchitectural criticism, albeit not bymainstream architectural critics who were, andremain, largely apologists for this type of archi-tecture. Even in the best examples, the fullpotential of reinforced concrete as a materialwhich allows the creation of almost every typeof form was not realised. The preferred shapeswere almost invariably rectilinear.

The language of reinforced concrete architec-ture continued to develop, however. Thedistinctive buildings designed by lames Stirlingin the late 1950s and 1960s, for example theLeicester Engineering Faculty Building(1959-63) (Fig. 4.14), The History FacultyBuilding at Cambridge University (1964-67)and the Florey Building at Oxford University(1966-71) (Figs 4.15 and 4.16), all in the UK,represent a more adventurous exploration ofthe expressive potential of reinforced concretefor multi-storey institutional buildings thanwas seen in the period of early Modernism.

The form of the Florey Building (Figs 4.15 and4.16) is distinctive in both plan and elevation.The structure is a reinforced concrete frame-work and consists of a series of identicallyshaped main frames each with canted columnsand horizontal beams carrying the one-way-spanning floor decks. Several design features,namely the crescent-shaped plan composed ofa series of rectangles and triangles whoseboundaries are defined by the main frames (Fig.2.9), the distinctive cross-section and theingeniously moulded stairs which penetrate thesloping exterior skin of the building (Fig. 4.16),create many junctions between the elementswhich would have been very difficult to fashionin any material other than in situ reinforcedconcrete. The structural continuity which thematerial allows was also an essential require-ment for the complex geometry of this building.The structural system is working here wellwithin its capabilities and the fact that it hasbeen left exposed to form a prominent part ofthe architectural language, places the building

Fig. 4.14 Engineering Building, Leicester University,England, 1965-67. James Stirling, architect, F. J. Samuely &Partners, structural engineers. This building has areinforced concrete framework structure and demonstratestwo of the principal advantages of the material. Its highdurability has allowed the structure to be exposed on boththe exterior and interior, and thus to contribute to thearchitectural language being used, and its mouldabilityhas been employed to produce a relatively complex build-ing form (Photo: P. Macdonald).

firmly within the rationalist tradition of Gropiusand early Le Corbusier.

A more recent building in which the struc-tural properties of reinforced concrete havebeen fully exploited is the Willis, Faber andDumas Building at Ipswich by Foster I l l


Fig. 4.15 Accommodation plan andcross-section of the Florey Building,Oxford, England, 1967-71. lamesStirling, architect, F.J.Samuely &Partners, structural engineers. Thedistinctive plan and cross-section ofthis building were easily realised witha structure of in situ reinforcedconcrete.

Associates (1975) (Fig. 4.17). This building is ofthe same basic constructional type as the earlybuildings of Gropius and Le Corbusier. Itconsists of a reinforced concrete structuralarmature giving a plan which is free from struc-tural walls. Two-way-spanning coffered slabs,supported on a square column grid, define thethree principal floors. These are readily visiblefrom the outside through the transparent skinwhich is entirely of glass and which is mounteddirectly on the thin edges of the floor slabs.Both the distinctive curvilinear plan of thisbuilding and the cantilevering of the floorslabs beyond the perimeter columns areexpressive of the unique properties of struc-tural concrete and recall Le Corbusier'sdrawing of the structural armature of theDomino house (Fig. 4.5).

The architectural language of reinforcedconcrete was greatly extended in the 1980s,

particularly by the work of the New York Five.6

The Museum of Decorative Arts, Frankfurt onMain (1979-85) by Richard Meier (Fig. 4.18),will serve as an example. In this building Meieraccepted the vocabulary of early Modernismbut not the grammar or syntax, in the sensethat he did not subject himself to a dogmaticbelief in a quest for a universally applicablearchitecture, as Gropius and Le Corbusier haddone. Instead he allowed the requirements ofthe building's situation and of the very complexbrief to affect the development of the form.

The plan of the building was based on tworectangular grids set at an angle of 3.5° to eachother (Fig. 4.19). Its essential features are agrouping of distinctively shaped masses aroundcourtyards, giving a juxtaposition of solid andvoid which is seen also in the alternation ofsolid and transparent sections in the externalskin. The structure is basically a post-and-beamframework but the geometries of the plan andelevations are complex. These, and the boldtreatment of small-scale features such asdoorways and balconies, are all readily accom-

6 Richard Meier, Peter Eisenmann, Michael Graves,Charles Gwathney, John Hejduk.112


Fig. 4.16 Florey Building, Oxford, England, 1971. lamesStirling, architect, F.J.Samuely & Partners, structuralengineers. The stairs which protrude from the inwardlycanted rear wall of the Florey Building provide a goodillustration of the type of complexity which the mouldabil-ity and structural continuity of a reinforced concrete struc-ture allow (Photo: P. Macdonald).

Fig. 4.17 Willis, Faber and Dumas office, Ipswich,England 1974. Foster Associates, architects, Anthony Hunt,structural engineer. The curvilinear plan, cantilevered floorslabs and transparent walls of this building give architec-tural expression to the structural properties of reinforcedconcrete (Photo: Alastair Hunter).


114

Fig. 4.18 Museum of Decorative Arts, Frankfurt on Main,Germany, 1979-85. Richard Meier, architect. The juxtapos-ition of solid and void and of rectilinear and curvilinearelements is typical of the work of Richard Meier. In situreinforced concrete is an ideal structural material for thistype of architecture [Photo: E. & F. McLachlan]

modated by the continuity and mouldability ofthe reinforced concrete structure. In both itsoverall form and in its detailing the buildingproduces constant reminders of the sculpturalpossibilities of the structural material and istherefore another example of the contributionwhich reinforced concrete has made to thedeveloping language of mainstream modernarchitecture. Meier produced here a buildingwhich functioned well for its intended purpose,which was a positive addition to the city ofFrankfurt, and which extended the architecturallanguage of reinforced concrete.

Another example of the use of reinforcedconcrete to produce a highly sculptural form isthe building for the Vitra Design Museum

(1988-89) (Fig. 4.20) by Frank Gehry. As withearlier examples of this genre, such as LeCorbusier's chapel at Ronchamp, this is arelatively small building and the ability ofreinforced concrete to allow the architectalmost unlimited freedom in the matter ofform, provided that a building is not too large,is well demonstrated.

Buildings such as Le Corbusier's chapel atRonchamp and Gehry's Vitra Design Museummight be described as belonging to the'irrational school' of curvilinear concrete archi-tecture. There has also been a 'rationalistschool' which has produced structures whichare entirely justifiable in a purely technicalsense as well as being striking visually. Suchstructures have normally been the designs ofstructural engineers (Robert Maillart, Pier LuigiNervi, Felix Candela, Eduardo Torroja,Santiago Calatrava) rather than of architects.

This group of buildings is characterised byforms which allow a very high ratio of strengthto weight to be achieved, principally by

conforming to 'form-active'7 shapes. Many, forexample, are based on the parabola, which isthe 'form-active' shape for uniformly distrib-uted gravitational load. They are also shapeswhich can be easily described mathematically,and this makes both their design and theirconstruction much simpler than those ofequivalent irregular forms such as the chapelat Ronchamp or the Vitra Design Museum. Thedesign is simpler because the analysis of thestructure, in which internal forces and stressesare calculated, is straightforward if the formcan be described mathematically by equationsbased on a Cartesian co-ordinate system. Theconstruction is straightforward when the form

can be described mathematically because thisgreatly eases the problem of setting out andbuilding the formwork on which the concrete iscast. Thus, the buildings which have beendesigned by the architect/engineers fall intothe category of 'true structural high tech' (seeSection 2.2). They perform well when judged bypurely technical criteria concerned with struc-tural efficiency and 'buildability' in present dayindustrialised societies because their shapesare both 'form-active' and part of the world ofprecise mathematical description rather thanof personal preference.

To sum up, the buildings which have beendescribed above serve to illustrate both thecontribution which reinforced concrete hasmade to the development of twentieth-centuryarchitecture and the range of architecturalforms made possible by the material. 1157 See Appendix 1.


Fig. 4.19 Plan of the Museum of Decorative Arts,Frankfurt on Main, Germany, 1979-85. Richard Meier,architect. Features such as the 3.5° skew which is presentin this plan are readily accommodated by an in situreinforced concrete structure. Note that regular columngrids are used where possible as this gives rise to the mosteconomical forms of construction.


116

Fig. 4.20 Vitra Design Museum, Basel, Switzerland,1987-89, Frank Gehry, architect. This highly sculptural formcould only have been realised in reinforced concrete[Photo: E. and F. Mclachlan].

Reinforced concrete has been used principallyin the form of skeleton frameworks where ithas allowed architects to exploit the opportu-nities articulated in Le Corbusier's 'Five pointstowards a new architecture' of 1926, in particu-lar, the '... free designing of the ground plan ...'and the '... free designing of the facade ...'.Examples of this type of building can be foundin every decade of the Modern period, from LeCorbusier and Gropius to Foster and Meier. Inmost cases the relationship between structureand architecture has been one of 'structureaccepted' (see Section 2.2) in which thereinforced concrete armature of the building,though playing a significant role in determin-ing the form and general arrangement, was not

emphasised in the aesthetic treatment, whichwas primarily, if not exclusively, orthogonal.Some notable exceptions to the rectilineargeometry which characterised the mainstreamModern Movement have also been shown. Inthese the mouldability of concrete wasexploited to produce distinctive irregular orcurvilinear structural shapes. This type ofbuilding, which has become more common inrecent decades as a result of changing archi-tectural fashion, illustrates better the fullpotential of reinforced concrete as a structuralmaterial.

4.2.2 The technical performance of concreteas a structural material

4.2.2.1 IntroductionThe technical advantages and disadvantages ofconcrete in relation to those of alternativematerials are reviewed in this section. These

are obviously an important consideration inrelation to the selection of concrete as thestructural material for a building.

4.2.2.2 Advantages

StrengthOf the four principal structural materialsreinforced concrete is one of the strongest. Itperforms well in skeleton-frame-type structuresand is therefore best used in situations inwhich the properties of a frame are required(planning situations in which the restrictionsof loadbearing walls cannot be accepted). It isparticularly suitable for frameworks on whichthe level of imposed loading is high.

Reinforced concrete has tensile, compressiveand flexural strength and can be used to makeall types of structural element. It is capable ofresisting the internal forces which result fromevery possible combination of applied loadand structural geometry and is thereforecapable of producing elements with anygeometry. The material itself places no restric-tion on structural form.

MouldabilityThe fact that reinforced concrete is available insemi-liquid form means that it can be cast intoan almost infinite variety of shapes. Thisproperty, together with its strength characteris-tics, means that virtually any form can becreated relatively easily in reinforced concrete.The moulding process also allows structuralcontinuity between elements to be achievedrelatively easily and the resulting staticalindeterminacy is another factor which easesthe production of complicated forms. In par-ticular, it allows the adoption of irregularpatterns of vertical support for floor and roofstructures, the cantilevering of floor and roofstructures beyond perimeter columns and theomission of areas of floor to create voidsrunning through more than one storey in theinteriors of buildings. It also allows reinforcedconcrete structures to be self-bracing.

DurabilityReinforced concrete is a durable materialwhich can be left exposed in relatively hostile

environments. A wide variety of surfacetextures can be achieved, depending on thetype of mould treatment which is specified.Finishing materials can therefore be eliminatedwhere concrete structures are used.

Fire resistanceReinforced concrete performs well in fire; it isincombustible and it retains its structuralproperties when exposed to high temperatures.

Cost

Reinforced concrete is relatively cheap and,when used for frame structures, will usually becheaper than steel. It is, however, normallymore expensive than masonry for loadbearing-wall structures.

4.2.2.3 Disadvantages

WeightReinforced concrete structures are heavy. Thematerial has a relatively low ratio of strengthto weight and a reinforced concrete frame isnormally significantly heavier than an equiva-lent steel frame. It is because of the high self-weight that reinforced concrete performs bestin situations in which relatively high imposedloads are involved - for example, for the floorstructures of multi-storey commercial or indus-trial buildings. It is rarely used where imposedloads are small, such as in single-storey build-ings or the roof structures of multi-storeybuildings, except in forms in which the level ofstructural efficiency is high, such as form-active shell roofs.

The relatively high self-weight of reinforcedconcrete structures does have some advan-tages, however. It gives the building a highthermal mass which eases the problemsassociated with environmental control. It alsomeans that concrete walls and floors arecapable of acting as effective acoustic barriers.

ConstructionThe construction of a reinforced concrete struc-ture is complicated and involves the erectionof formwork, the precise arrangement of intri-cate patterns of reinforcement and the carefulplacing and compacting of the concrete itself.The construction process for a reinforced 117



concrete structure therefore tends to be bothmore time consuming and more costly thanthat of an equivalent steel structure and thesefactors can mitigate against its use in a par-ticular situation. Another disadvantage is therequirement for sufficient space for storage offormwork and for the assembly of reinforce-ment cages. This can be problematic if thebuilding site is very tight and congested.

StrengthAlthough, as stated above, reinforced concreteis one of the strongest of the four primarystructural materials it is nevertheless weakerthan steel, with the result that elements inreinforced concrete structures tend to bebulkier than steel equivalents. The relativeweakness also places restrictions on the spansfor which reinforced concrete is suitable. Apractical maximum span for floor structures isaround 20 m and spans greater than 15 m arein fact rare.

4.2.3 The selection of reinforced concreteA structural material is selected on the basis ofthe aesthetic opportunities which it allows andthe technical performance which it provides.Many of the former are, of course, conse-quences of the latter.

So far as the overall form of a building isconcerned, most of the possibilities whichreinforced concrete allows result from both itsrelatively high strength and the high level ofstructural continuity which it makes possible.The high strength makes it suitable for skele-ton-frame-type structures in which the internalforces are relatively high. It is most suitable,however, for structures in which fairly largeimposed loads are carried over relatively shortspans, that is spans in the range 6 m to 15 m,and is therefore employed mainly for multi-storey buildings where floor loads have to becarried.

The high level of structural continuity whichthe use of reinforced concrete allows gives thedesigner considerably more freedom to ma-nipulate the overall form of a framework thanis possible with steel (the principal alterna-tive). The two-way-spanning capability of the

floor structure is especially significant andallows both the adoption of an irregularpattern of vertical support and the omission ofsections of floor to create volumes which runthrough more than one storey. It also makespossible the cantilevering of floors beyondperimeter columns and the simple creation oframps or stepped changes in the levels offloors. The structural continuity of reinforcedconcrete also facilitates the creation of curvi-linear plan-forms, either in the complete build-ing or in the form of internal breaks in the floorstructure, and the adoption of complex, non-rectilinear column grids. Reinforced concrete isalso ideally suited to the creation of form-active or semi-form-active types of structuresuch as shelis, vaults, domes and arches.Reinforced concrete therefore offers the archi-tect very great freedom in the matter of form.

The durability and fire resistance of concreteare two very significant properties if theexpression of structure is an aspect of thearchitectural programme. The range of surfacetextures with which concrete can be producedis an added benefit where this is done. Evenwhere the structure is not exposed, thesequalities will normally simplify both the initialconstruction of a building and its subsequentmaintenance by removing the necessity forfireproofing and corrosion protectionschemes.

4.3 A brief introduction to concretetechnology

4.3.1 IntroductionReinforced concrete is a composite materialwhose constituents are concrete, which formsthe main bulk of the material, and steel, in theform of reinforcing bars. Concrete itself is alsoa composite material being composed ofcement and aggregate (fragments of stone),The properties of concrete depend on those ofits constituents and on the proportions inwhich these are mixed. Concrete can actuallybe manufactured on the building site althoughin modern practice it is normal for even in situconcrete to be mixed in a separate factory and118

delivered to the site in liquid, ready-mixedform. The mix proportions can, however, bespecified by the building's designer andconcrete is therefore one of the few materialswhose properties can be controlled directly bythe designers of buildings. The properties ofreinforced concrete depend on those of theconstituent concrete and of the reinforcementand on the location of the reinforcement in thestructural elements.

Reinforced concrete is therefore a complexmaterial which places certain demands on thedesigner who wishes to exploit its potentialfully. The successful use of such a materialmust be based on a knowledge of its basicproperties and its behaviour in response toload. These aspects of concrete technology arenot considered here in detail, although theprincipal issues are briefly described.

4.3.2 The terminology of concreteThe constituents of concrete are cement paste,which acts as a binding agent, and aggregate,which is an inert filling material. The aggregateusually consists simply of small pieces ofnatural stone and it acts both as a bulkingagent and to impart dimensional stability tothe concrete. Concrete is made by mixingtogether appropriate quantities of cement andaggregate in the dry state and adding sufficientwater to hydrate the cement. After the water isadded, a chemical reaction occurs whichcauses the concrete to become solid within afew hours in what is called the 'initial set'. Aconsiderable period of time is required beforeit develops its full strength, however, and thislatter process is called the 'final setting' or'hardening' of the concrete. The time requiredfor final setting varies, depending mainly onthe type of cement which is used, but a typicalconcrete will have developed about 80% of itsfull strength within three months of the initialset.

The properties of concrete which are ofprincipal interest to the building designer areits liquidity when it is in the 'fresh' state andits strength when in the hardened state. Theliquidity of fresh concrete is referred to as its'workability' and this property affects the ease

Fig. 4.21 The principal reinforcement in reinforcedconcrete is placed in the locations where tension occursdue to the bending effect of the load. Secondary reinforce-ment (not shown) is placed throughout the structuralelements to resist the tensile stress which occurs due tothe shrinkage associated with the curing process and withthermal movement.

with which it can be 'placed' and 'compacted'to form a dense solid. This is an especiallyimportant consideration where elements ofcomplex geometry are being constructed orwhere the concrete must be compacted aroundcomplicated patterns of reinforcement. Highworkability is required in these situations andthis can adversely affect the final strength ofthe concrete (see Section 4.3.3.3).

Plain concrete in the hardened state is amaterial which has moderate compressivestrength (typically between 20 N/mm2 and60 N/mm2 depending on the mix proportions)but very low tensile strength (usually aboutone tenth of the compressive strength). Whensteel bars are incorporated into concrete theresulting composite material is calledreinforced concrete. This has greatly improvedtensile and bending strength, the precisestrength properties being dependent on theamount of reinforcement which is used and itslocation in element cross-sections. In elements 119


Fig. 4.22 Cracking of the concrete, shown exaggeratedhere, is inevitable in the parts of reinforced concreteelements in which tensile stress occurs. The width of thecracks which form must be controlled to prevent theingress of water, which would cause corrosion of thereinforcement. This places a limitation on the amount ofstrain, and therefore stress, which can be tolerated in thereinforcement. Allowable stresses in reinforcement arenormally lower than the strength of steel would in othercircumstances allow.

which carry bending-type load the reinforce-ment is concentrated in locations wherestretching, and therefore, tensile stress, occurs(Fig. 4.21).

The tensile internal forces which are presentin reinforced concrete as a result of bendingaction, and which are resisted by the reinforce-ment, cause tensile strain8 to occur. Thisproduces cracking in the concrete surroundingthe reinforcement which affects the appearanceand durability of the elements (Fig. 4.22). Theextent of the cracking must be limited bysuitable design and is usually controlled byrestricting the amount of stress which ispermitted in the reinforcement. This limits theamount of stretching which can occur andtherefore the width of any cracks which formbut it has the disadvantage of requiring thatthe steel be greatly understressed and there-fore inefficiently used.

Pre-stressing of concrete allows the fullpotential strength of the steel reinforcement tobe realised. In pre-stressed concrete, crackwidths are limited by deliberately introducingan axial compressive load into the concrete so

8 For definitions of stress and strain see Macdonald,Structure and Architecture, Appendix 2.

Fig. 4.23 The principle of pre-stressed concrete. In thetop diagram the steel bar is tensioned by the bolts at theends of the beam and introduces a compressive 'pre-stress' into the concrete which is uniformly distributedwithin each cross-section. This neutralises the tensilebending stress which occurs in the lower half of the beamwhen the load is applied and thus eliminates tensile crack-ing of the concrete. A quite different relationship betweensteel and concrete exists here than occurs in plainreinforced concrete.

as to limit the amount of stretching which candevelop when a bending-type load is applied.The pre-stressing is carried out by stretchingthe reinforcement during the construction ofan element and then connecting it to theconcrete in such a way that it imparts a per-manent compressive stress into the concrete(Fig. 4.23). If a load occurs on the structurewhich tends to cause tension to develop, thecompressive pre-stress must first be overcomebefore any stretching can occur. The pre-stress-ing technique thus uses the concrete-steelcombination in such a way that the full tensilestrength of the steel is utilised without thepenalty of excessive stretching and therefore ofcracking of the concrete being incurred.

Pre-stressed concrete is either 'pre-tensioned' or 'post-tensioned'. In pre-tension-ing the reinforcing bars, which are usually120


small-diameter wires of high-tensile steel, arestretched to a predetermined tensile stress in apre-stressing bed (Fig. 4.24). The fresh concreteis then placed around these and allowed to setand harden. The pre-tension is then releasedand the friction (bond) between the concreteand the surface of the reinforcement transfersa permanent compressive load into theconcrete. Pre-tensioned elements are normallyprecast and the technique is commonly usedfor producing a range of proprietary compon-ents such as lintels, beams of various kinds,and floor slabs. These are used in a number ofdifferent types of structure. The lintels arenormally used in conjunction with masonry.

Pre-stressed floor slab units are used in manytypes of structure including steel andreinforced concrete frames and masonryloadbearing-wall structures.

In post-tensioned pre-stressed concrete,which is usually cast in situ, the concrete isplaced into moulds which contain narrowtubes (Fig. 4.25). After the concrete has set andhardened the reinforcement is placed in thesetubes and tensioned against plates or jackswhich are permanently fixed to the ends of theelements. This has the effect of introducing apermanent compressive pre-stress.

Pre-stressing allows much higher stresses tobe used in the reinforcement in concrete and 121

Fig. 4.25 Post-tensionedpre-stressed concrete. Whereconcrete is post-tensioned,ducts for pre-stressingtendons are cast into theelements concerned. Thetendons are fed into theducts once the concrete hashardened and then stretchedby external jacking devices.The resulting pre-stress isthen introduced to theconcrete via special anchor-ing devices at the ends of theelement.

Fig. 4.24 Pre-tensionedpre-stressed concrete. In pre-tensioned pre-stressedconcrete the pre-stressingwires are stretched in a pre-stressing bed prior to theplacing of the concrete. Thewires are released from thepre-stressing bed once theconcrete has hardened andthe pre-stress is transferredto the concrete by frictional(bond) stresses actingbetween the concrete andthe surfaces of the wires.



permits much lighter and more slenderelements to be achieved than is possible withordinary reinforced concrete. The quality ofboth the concrete itself and the reinforcementmust, however, be higher than in ordinaryreinforced concrete and the components aregenerally more sophisticated. Pre-stressedconcrete is usually more expensive toconstruct than equivalent reinforced concreteand, with the exception of the range of pre-tensioned proprietary components which hasbeen mentioned above, the technique tends tobe used for architectural structures only wherea special requirement exists, such as the needto achieve a very long span.

4.3.3 The constituents of reinforcedconcrete

4.3.3.1 CementCement is the binding agent in concrete whichpossesses the cohesive strength to hold theaggregate and reinforcement together into asolid, composite material. A considerablenumber of types of cement are used in building;most of these are varieties of Portland cement.9

All of the cements which are used for structuralconcrete are dependent on water for the devel-opment of strength; when water is added to drycement a fairly complex series of chemicalreactions takes place in a process which iscalled hydration and which causes the resultingpaste to stiffen. The subsequent development ofstrength takes place in two stages: initial settingoccurs fairly quickly, usually within a few hours;a much longer period, of up to a year, isrequired for the development of full strength,although a reasonable proportion of the finalstrength will normally have been developedwithin seven days. The hardening of the cementis simply a continuation of the hydrationprocess which produced the initial set and thecement must be kept wet if this is to proceedsatisfactorily to compensate for water whichmay be lost due to evaporation.

One of the principal factors which affects thestrength of hardened cement is thewater-cement ratio, which is the ratio of theweight of water to the weight of cement in thewet mix. Only a small quantity of water(around 25% by weight) is required to bringabout sufficient hydration of the cement tocause the initial set; if more is present theextra water remains as a separate phase in thecement paste and becomes incorporated as aseparate phase into the cement matrix whensetting occurs. The extra water subsequentlyevaporates to leave voids in the hardenedcement and this reduces its strength. It istherefore necessary to ensure that no morewater than the minimum required to produceinitial setting is present in a mix of concrete.10

Figure 4.26 shows the relationship betweenwater-cement ratio and final compressivestrength for a typical concrete.

The use of a low water-cement ratio has theeffect of making the concrete very stiff anddifficult to compact. This is a factor which hasto be considered by structural designers whenthey specify the water-cement ratio which is tobe used for a concrete, because the entrain-ment of air into the concrete, due to poorcompaction, will reduce its strength. Theminimum water-cement ratio which is prac-ticable in a particular case depends on individ-ual circumstances, such as the complexity ofthe element in which the concrete is to beplaced and the type of equipment which willbe available for use in the compaction process.The practical minimum for water-cement ratiois usually in the range 0.4 to 0.5 (40-50% byweight).

Two phenomena which are associated withthe setting and hardening of cement and whichaffect the specification of concrete are shrink-age and heat gain. The cement matrix which isformed during the initial setting process

10 More water must in fact normally be present toproduce sufficient workability (liquidity) to allow thecement to be compacted effectively. The practicalminimum amount of water is around 40% to 50%,depending on the system being used for compaction.

9 So called because of its supposed resemblance toPortland stone.122

Fig. 4.26 The relationship between the ultimate strengthof concrete and the water/cement ratio. This graph demon-strates that the more fluid the wet concrete is the lowerwill be the ultimate strength. This is because relativelylittle water is required to cause the initial set of concrete.If more is present it forms separate pockets of pure waterin the setting concrete which ultimately evaporate to leavesmall voids which weaken the concrete.

shrinks subsequently as the hardening processproceeds. Shrinkage, by itself, is not necessar-ily harmful to a concrete component,especially if the element is not restrainedduring the hardening process (a small precastconcrete lintel, for example). Where anelement is restrained, however, such as an insitu reinforced concrete beam or slab which ispart of a large frame, then the effect of shrink-age is to generate tensile stress in the elementwhich may cause it to crack. Shrinkage is there-fore a phenomenon which must be consideredduring the design of a concrete structure.Normally it is controlled by the incorporationinto the structure of suitable reinforcement;this simply carries the stress which resultsfrom the shrinkage. Occasionally, movementjoints are used to control shrinkage cracking. Atypical example of this is found in lightlyreinforced non-structural ground floor slabs.

Heat gain is a phenomenon which occursbecause the chemical processes which areinvolved in the hydration of cement are

exothermic. The increase in temperature whichresults from this tends to accelerate theprocess of hydration and this can producecracking of elements because it also causeshigh rates of shrinkage. The heat of hydrationrarely causes a problem in architectural struc-tures, however, because the elements arenormally slender enough to allow sufficientcooling to occur to maintain temperatures atan acceptable level. Where an element is oflarge bulk the use of special 'low-heat' cementsis sometimes required.

4.3.3.2 AggregateAggregate, being cheaper than cement, is usedas a bulking agent in concrete and, typically,will account for 75% to 80% of its volume. Italso serves to control shrinkage and toimprove dimensional stability. It normallyconsists of small pieces of stone, of varioussizes in the form of either naturally occurringsand or gravel, or crushed rock fragments;other materials, such as crushed brick, blastfurnace slag or recycled building materials, aresometimes used. Aggregate must be durable,of reasonable strength, chemically and phys-ically stable, and free of constituents whichreact unfavourably with cement.

The proportions which are present of thedifferently sized particles which occur in anaggregate are referred to as its grading and ifthe aggregate is to be effective as a bulkingagent the grading must be such that a particu-lar distribution of particle sizes occurs. Thesmaller particles should ideally be of such asize and quantity as to fully occupy the voidsbetween the larger particles and leave aminimum volume to be filled by cement (Fig.4.27). The grading of aggregate also affects theworkability of the concrete, which is betterwhen a higher proportion of fine particles ispresent. This is a factor which indirectly affectsstrength because a well-graded aggregateallows a required workability to be achievedwith a lower water-cement ratio than a poorlygraded aggregate.

If a naturally occurring aggregate is used forconcrete, the designer has no control over thegrading but must ensure that the range and 123


Vibration

Hand compaction

Fully compacted concrete

Insufficientlycompacted concrete

Water/cement ratio

Co

mp

ress

ive

stre

ngth


Fig. 4.27 If aggregate is to form an effective bulkingagent the relative sizes of the particles present must bewell matched. This has occurred in the sample on the lefthere, where only small voids are left to be filled withcement, but not in that on the right.

distribution of particle sizes is satisfactory.This is determined by carrying out a sieveanalysis, which involves passing a sample ofthe aggregate through a series of progressivelysmaller sieves and noting the fraction which isretained on each (Fig. 4.28). The percentage ofthe whole sample which passes each sieve isthen plotted to give a grading curve (Fig. 4.29).The relevant British Standard (BS 882) sub-divides aggregates into coarse aggregate, in

which particles larger than 5 mm in diameterpredominate, and fine aggregate or sand, inwhich particles smaller than 5 mm in diameterpredominate. The grading limits for each typewhich are considered to be suitable for makingconcrete are specified in the Standard by defin-ing zones within which the grading curves ofthe fine and coarse aggregates must fall.

Where a naturally occurring aggregate with asuitable grading is not available, an aggregatewith a specific grading can be produced artifi-cially by mixing together appropriate quan-tities of screened, single-sized batches ofcrushed rock fragments.

Aggregate particles are classified into threecategories according to their shape and aresaid to be either rounded, irregular or angular.

Fig. 4.28 The grading of anaturally occurring aggre-gate is determined by asieve analysis. In the nestof sieves shown the meshsize diminishes down thestack. To carry out theanalysis the sample isintroduced at the top andthe weight of particleswhich are retained at eachlevel determined. Thisallows grading curves to beplotted.

124

particles is larger than if the same bulk is pro-vided with a smaller number of larger particles.

A final consideration in relation to theaggregate which is used in concrete is themaximum particle size of the coarse fraction.This must usually be restricted, depending onthe size of the elements in which the concretewill be cast and on the complexity of thereinforcement pattern. The maximum size ofaggregate is normally no greater than 25% ofthe minimum thickness of elements and atleast 5 mm less than the clear distancebetween reinforcing bars.

4.3.3.3 Specification of concreteThe specification of concrete has two aspects,which are the selection of the constituents,that is of the type of cement and type of aggre-gate, and the determination of the mix propor-tions. The latter process is called the 'mixdesign'.

So far as the selection of materials isconcerned the type of cement is determined bythe requirements in respect of rate of harden-ing, heat gain and resistance to chemicalattack; all Portland cements achieve approxi-mately the same final strength for a given setof mix proportions. It is usual to selectnaturally occurring aggregates for concrete, ifthese are available; they are usually specifiedin two batches, fine aggregate (sand) andcoarse aggregate, and the grading of thesemust normally conform to the limits which arespecified in the relevant standard (e.g. BS 882).Where a crushed rock type is used this toomust be correctly graded. Crushed rock aggre-gate will normally require a higher watercontent to achieve a given workability and willtherefore produce a weaker concrete than anaturally occurring aggregate. The workabilityof a concrete which is made with crushed rockcan be improved by increasing the proportionof fine particles present. This allows the use ofa lower water content, which improves thestrength, but it results in a higher cementcontent being required, which increases thecost of the concrete.

The mix proportions of a concrete are speci-fied in terms of the weights of materials which 125


Fig. 4.29 Aggregate grading curves. Following sieveanalysis a sample of aggregate will produce a gradingcurve similar to those depicted. Its suitability as a bulkingagent in combination with other aggregates can then bejudged.

Naturally occurring aggregates tend to berounded while crushed rock types are predom-inantly angular. The significance of this factoris that it affects the workability of the concrete:aggregates with rounded particles produce amore workable mix for a given water-cementratio than do those with angular particles. It istherefore possible to achieve a particularworkability with a lower water-cement ratio ifthe aggregate is rounded than if it is angular,and this in turn produces a stronger concrete.Aggregate shape, like grading, therefore has anindirect effect on concrete strength and, ifother things are equal, naturally occurringaggregates will produce stronger concrete thanartificially produced aggregates.

The grading of an aggregate can affect thecost of concrete because it affects the quantityof cement which must be provided to producea given volume of concrete. If a large numberof fine particles are present, the total surfacearea of the aggregate is high and the quantityof cement which is required to coat all of the

BS sieve sizes

Per

cent

age

pass

ing


are required to produce a unit volume of fullycompacted fresh concrete and the principalfactors which affect the mix proportions are therequired strength and the required workability.The required strength is determined during thestructural design calculations; the requiredworkability depends on the nature of the struc-ture, specifically on the sizes of the elements,the complexity of the pattern of reinforcementand the type of equipment which will be avail-able to assist with the compaction. Mix designinvolves following a set of procedures in whichthe various properties which are required ofthe fresh and hardened concrete are used toderive systematically the specification for theconcrete: a suitable water-cement ratio,cement-aggregate ratio and aggregate grading.The object is to obtain optimum mix propor-tions for the requirements of a particular struc-ture. The process is a fairly complicated oneand will not be described in detail here.

4.3.3.4 ReinforcementThe reinforcement which is used in concrete isnormally in the form of steel bars, either ofplain circular cross-section or with varioussurface treatments which increase the bondwith the concrete (Fig. 4.30). The preferreddiameters are 6, 8, 10, 12, 16, 20, 25, 32 and40 mm, and the normal maximum length is12 m. Reinforcement for slabs is produced inthe form of square mesh, in individual piecesor in rolls, in which the bars are weldedtogether at their crossing points. Meshes withmost combinations of bar sizes and spacingscan be readily obtained.

Reinforcement is produced in both mildsteel and high-yield steel. The latter allowsmuch higher tensile stress to be specified forthe reinforcement; its use is limited, however,because the critical factor which determinesthe amount of stress which can be permitted inreinforcement is frequently the need to controlthe amount of strain which occurs so as toprevent cracking of the concrete. The fulltensile strength of high-yield steel cannottherefore be used in the design of reinforcedconcrete structures. The reinforcement which isused in pre-stressed concrete is hard-drawn

Fig. 4.30 Reinforcing bars are produced with a number ofsurface treatments to improve the bond with concrete.

steel wire; the preferred sizes are 4, 5 and7 mm.

4.3.4 The behaviour of reinforced concreteReinforced concrete is a composite materialwhose structural action is both subtle andcomplex. The basic relationship betweenconcrete and reinforcement is, however, fairlystraightforward and is explained here inrelation to elements which are subjected tobending-type loads.

The effect of bending on any structuralelement is to place part of its cross-section incompression and part of it in tension. In thesimplest types of reinforced concrete element,the compression on one side of the cross-section is resisted by the concrete and thetension on the other side by the126

Ribbed bars

Stretched and twisted ribbed bar

Ribbed and twisted bar

Square twisted bar

reinforcement." To appreciate the way in whichthe two materials interact it is necessary tovisualise the distribution of stress whichoccurs within a beam when a bending-typeload is applied.

The distribution of stress within a beam ofany material depends on the configuration ofthe applied load. The action of a uniformlydistributed load is to bend the beamdownwards and cause tensile stress to be setup in the lower half of the cross-section andcompressive stress in the upper half. The exactdistribution of the stress is complex but a wayof visualising it is shown in Fig. 4.31.

In Fig. 4.31a a series of circles has beendrawn on the side of the beam. Following theapplication of the load these change theirshape to become ellipses (Fig. 4.31b). Theminor and major axes of the ellipses coincidewith the directions of maximum compressionand tension at each location. This allows thedirections of the maximum tensile andcompressive stresses (the principal stresses) tobe plotted (Fig. 4.31c). Thus, at mid-span thedirection of stress is parallel to that of thebeam. Towards the ends of the beam the direc-tion of the tensile and compressive stressesbecome progressively more inclined to the axisof the beam and the tensile and compressivestress lines cross each other. The material inthese regions is stressed simultaneously intension and compression in two orthogonaldirections.

The diagram of principal stresses (Fig. 4.31c)shows only the directions of maximum tensionand compression. The magnitudes of thestresses vary along each line from a maximumat the mid-span position, where they are paral-lel to the axis of the beam, to zero where thecurved portions of the lines approach the topor bottom of the beam. The magnitudes of thestresses also vary between lines. They aregreatest in the lines which run closest to the

Fig. 4.31 The pattern of stresses in a beam. Circlesdrawn on the side of an unloaded beam (a) becomeellipses as a result of the strain caused by the load (b).This gives an indication of the directions of maximumtension and compression at every location and allowsprincipal stress lines to be drawn (c). The magnitude of thestress varies along each principal stress line from zero atthe point at which it crosses the top or bottom surface ofthe beam to a maximum at the mid-span position. Themagnitude of stress also varies between different lineswith those closest to the top and bottom surfaces at themid-span position carrying the greatest stress.

top and bottom of the beam at mid-span anddecrease towards the interior of the beam.

The highest levels of stress occur on thecross-section at the mid-span point. The stressis tensile in the lower half of the cross-sectionand compressive in the upper half. Its magni-tude varies from a maximum tensile stress atthe lower extreme fibre to maximum compres-sive stress at the upper extreme fibre with zerostress half-way up at what is called the neutralaxis. 127

11 In complex reinforced concrete elements compressionreinforcement is also provided.


(b)

(a)

(c)

CompressionTension

Fig. 4.32 Failure of an unreinforced concrete beam. Theunreinforced beam fails as a result of the formation of acrack situated at the location of the highest tensile stressand at right angles to the direction of maximum tension.

Fig. 4.33 Crack pattern in a reinforced concrete beam.The reinforced beam carries significantly more load thanthe unreinforced beam. Tensile cracks form, always at rightangles to the direction of maximum tension (given by theprincipal stress lines) in those locations at which thetensile strength of the concrete is exceeded. The beamcontinues to carry load, however, so long as each crack iseffectively crossed by the reinforcement.

Figure 4.32 shows the behaviour of anunreinforced concrete beam which is subjectedto increasing load. The symmetrical pattern oftensile and compressive stresses shown in Fig.4.31c becomes established when the load isapplied, with the maximum stresses occurringat the mid-span position. As the loadincreases, the magnitudes of the stressesincrease and eventually the tensile strength ofthe concrete is exceeded at the lower extremefibre (the location of the highest level oftensile stress). A crack forms at right angles tothe direction of maximum tension and thisrapidly propagates up through the beam whichbreaks into two halves. Because the compres-sive strength of the concrete is at least tentimes greater than its tensile strength, theconcrete in the top of the beam is relativelyunderstressed when the tensile failure occurs.

The failure can be prevented if steelreinforcement is placed close to the lowersurface of the beam (shown diagrammaticallyin Fig. 4.33). In the reinforced beam the tensilefailure of the concrete at the mid-span positionstill occurs but the beam as a whole does notfail because the crack which forms is nowcrossed by the reinforcement. The reinforcingbar carries the tensile load and the beamcontinues to function so long as the cross-sectional area of the steel bar is sufficient toprevent it from becoming over-stressed intension and it does not pull out of theconcrete on either side of the crack. Thefrictional stress between the concrete and thesurface of the bar (bond stress), is therefore an

important consideration in determining theability of the two materials to act compositely;this is the reason that special surface texturesare often applied to reinforcement (Fig. 4.30).

If more load is applied to the beam than wasrequired to cause the first crack to appear, thestress everywhere will increase and the pointon the principal tensile stress line at which thetensile strength of the concrete is exceededprogresses along the line, in both directions,from the mid-span position. More cracks form(Fig. 4.33), always at right angles to the direc-tion of maximum tension, but failure of thebeam as a whole does not occur so long as thecracks are effectively crossed by the steel barand the compressive strength of the concreteis not exceeded in the top half of the beam. Asthe tensile failure points move towards theends of the beam the cracks, which alwaysform at right angles to the direction ofmaximum tension, become more and moreinclined to the beam axis and are less effec-tively crossed by the reinforcement.

Failure eventually occurs due to the forma-tion of a diagonal crack which is not crossed bythe reinforcement (Fig. 4.34). This type offailure is called a shear failure because thedegree of inclination of the principal stresslines causes a shearing action on the cross-section rather than a simple bending action.Shear failure can be prevented by shaping thereinforcing bar so that it conforms to theprofile of the line of the principal tensile128


Fig. 4.34 Shear failure of a reinforced concrete beam.The beam depicted in Fig. 4.33 eventually fails due to theformation of an inclined crack at the end of the beamwhich is not effectively crossed by the reinforcement. Thistype of failure is called a shear failure.

stress. This is, however, difficult to carry out inpractice and the solution which is normallyadopted is either to bend up the mainreinforcement bars in straight sections towardsthe ends of the beam (Fig. 4.35a) or to form acage of reinforcement with links of steel (Fig.4.35b). Both of these methods ensure thatevery potential tensile crack in the concrete iscrossed by some reinforcing bars.

Figure 4.36 shows diagrammatically a typicalpattern of reinforcement for a reinforcedconcrete beam. This consists of primaryreinforcement in the tensile half of the beam tocarry the tensile component of the bendingstress, and shear reinforcement to carry theshear load caused by the inclination of theprincipal stresses. Secondary reinforcement isalso normally provided (in this case longitudi-nal bars in the top half of the beam) to holdthe primary reinforcement in position while theconcrete is being placed around it. Thissecondary reinforcement also helps to controlshrinkage and to carry secondary tensile stresswhich could occur, for example, due to thermalmovement.

The behaviour described above, by whichbending-type load is resisted by the compositeaction of concrete and steel, illustrates the wayin which reinforced concrete functions. In largecomplex structures, such as multi-storeyframeworks, the direction of bending within anindividual beam or slab may vary due to thecombined effects of structural continuity andload configuration. The distribution of themain reinforcement within the structure must

Fig. 4.35 Beam with shear reinforcement. Bent up barsor links prevent shear failure by ensuring that all possibletensile cracks are effectively crossed by reinforcement.

be such as to allow the effective resistance ofall tension which occurs within the concrete.The pattern of reinforcement in real structuresis therefore often considerably more complexthan that shown in Fig. 4.36.

Fig. 4.36 The pattern of reinforcement in a fullyreinforced concrete beam. 129


(a)

Shear reinforcement

Small-diameter bars to holdshear reinforcement in position

Tensile reinforcementresisting negative bendingmoment at support

Tensile reinforcementresisting positive bendingmoment at mid-span

4.4 Structural forms for reinforcedconcrete

4.4.1 Basic elements

BeamsReinforced concrete beams are normally eitherrectangular in cross-section or combined withthe slab which they support to form T- and L-shaped cross-sections (Fig. 4.37). The normalspan range for reinforced concrete beams is 4.5to 10 m. Spans of up to 20 m are occasionallyused but depths of around 1.5 m and upwardsare required for this, and a large volume ofconcrete is therefore involved. Spans greaterthan 20 m are possible but other types ofstructure will normally perform better in thisspan range. The depth which is required forbeams depends on the span and the loadwhich is carried. It is frequently determinedfrom deflection rather than from strengthrequirements and in the normal span range adepth of around one twentieth of the span isrequired for a simply supported beam and onetwenty-sixth of the span for a beam which iscontinuous across a number of supports. Thebreadth of a rectangular beam is usuallyaround one third to one half of its depth.

ColumnsColumns are constructed with a range of cross-sectional shapes, the most common beingsquare, rectangular and circular. The primaryreinforcement in columns is longitudinalreinforcement which contributes to the resis-tance of the compressive load and thereforereduces the required size of the cross-section.Transverse reinforcement, in the form of links,is also provided to give lateral support for thelongitudinal reinforcement to prevent a burst-ing-type compression failure (Fig. 4.38). As inall compression elements, the principal factorwhich determines the dimensions which must

Fig. 4.37 T- and L-beams. Where reinforced concretebeams support a reinforced concrete slab the elements arenormally cast together and act compositely to form T- andL-beams.

Fig. 4.38 Reinforcement in a column. Columns areprovided with longitudinal reinforcement, to increase theircompressive strength, and links to prevent buckling failureof the very slender reinforcing elements.130



be adopted for columns is the need to avoidhigh slenderness, which would make thecolumn susceptible to a buckling type offailure. The slenderness ratio of a concretecolumn is its effective length12 divided by itsleast width and the British Standard (BS 8110,'Structural Use of Concrete') recommends amaximum value of 60; 30 should be regardedas a practical maximum for most forms ofconstruction, however, and to achieve reason-able load-carrying capacities, slendernessratios in the range of 20-25 are normallyadopted. Reinforced concrete walls perform ina similar way to columns and are made toconform to the same slenderness ratio require-ments. In the case of walls the slendernessratio is the effective height of the wall dividedby its thickness.

SlabsReinforced concrete slabs can be either one-way- or two-way-spanning structures. Thebehaviour of a one-way-spanning slab issimilar to that of a rectangular beam, andthese elements can in fact be regarded simplyas very wide beams. The primary reinforcementconsists of longitudinal bars placed parallel tothe direction of span on the tension side of thecross-section (that is the lower part in the mid-span position of slabs and the upper part nearthe supports where slabs are continuous over anumber of supports).

Secondary reinforcement, in the form ofstraight bars which run at right angles to thedirection of the span, is provided to controlshrinkage and to distribute the effects ofconcentrated load (Fig. 4.39). The reinforce-ment for a solid slab normally takes the formof a mesh in which the bars of the primary andsecondary reinforcement are welded togetherat the points where they cross. Shear stressesin slabs are very low, except close to columns,and primary reinforcement is not normallyprovided to resist shear.

Fig. 4.39 Reinforcement in a slab. Slabs are providedwith mesh reinforcement close to the top and bottomsurfaces. If the bars are of equal thickness and spacing inboth directions the slab will be capable of spanning simul-taneously in both directions (two-way-spanning slab). Ifthe degree of reinforcement in one of the directions isgreater than in the other (larger bars at closer spacing)then the slab will span more effectively in that direction(one-way-spanning slab). In the latter case reinforcementis nevertheless provided in the non-span direction todistribute concentrated load and to control shrinkage ofthe concrete.

The depth of one-way-spanning slabs isnormally approximately one thirtieth of thespan for a simply supported slab and onethirty-fifth of the span for a slab which iscontinuous over a number of supports. Theeconomic span for a solid slab is in the range4 m to 8 m but this can be extended by using aribbed form in which a proportion of theconcrete in the lower half of the cross-sectionis removed. This type of slab is economic inthe range 6 m to 12 m. If pre-stressing isapplied the maximum spans are increased to13 m for solid slabs and 18 m for ribbed slabs.

A two-way-spanning slab spans simultane-ously in two directions and must be supportedon beams, walls or rows of columns around itsperimeter or at its corners. It should ideallyhave a square plan and the primary reinforce-ment is a square mesh of longitudinal bars; nosecondary reinforcement is required. Two-wayslabs are statically indeterminate structuresand allow a more efficient use of material thanone-way slabs. Similar span-to-depth ratios are 131

12 Effective length is based on the distance betweenpoints at which the column is supported laterally byother parts of the structure. Lateral support normallyoccurs at each storey level.

Secondary reinforcementto distribute the effect ofconcentrated load

Main reinforcement in base of slabresisting positive bending momentat mid-span

main reinforcement in topof slab over supports


Fig. 4.40 Coffered slab.The coffered slab is a two-way-spanning flat-slabstructure (i.e. a slab whichis supported directly oncolumns). The coffersimprove the efficiency byremoving concrete from thetensile areas where itcontributes little to thestructural performance[Photo: Pat Hunt].

132

used but the maximum economic spans arehigher; solid slabs are used for spans of up to8 m and this can be extended to 16 m if theweight is reduced by removing some of theconcrete on the tension side of the cross-section, and to 20 m if pre-stressing is used. Asystem of intersecting ribs, which form asquare grid, is then created on the soffit of theslab; the resulting form is said to be a cofferedor 'waffle' slab (Fig. 4.40).

StairsStairs are designed as reinforced concreteslabs whose depths are equal to the waistthickness of the stair, and which are normallyan integral part of the structure (Fig. 4.41). Therequirements in respect of the ratio of span todepth which is specified are the same as forone-way-spanning slabs.

Precast components

The precasting technique, in which concretecomponents are manufactured in a factory andtransported subsequently for erection on site,allows higher quality control to be achievedthan is possible with in situ concrete. Amongthe advantages which this offers are higherconcrete strength for given mix proportions,better quality of finish, greater dimensionalaccuracy and the economies of scale which areassociated with the factory process. The factorymethod of manufacture also allows morecomplicated shapes of cross-section to beachieved than are possible with in situconcrete. This, in turn, makes possible higherlevels of structural efficiency and can also fa-cilitate the use of structural elements, such asbeams and columns, as ducts for services.Precast concrete has therefore been widely

Fig. 4.41 Reinforced concrete stairs are simply one-way-spanning slabs which span between the landings.

Fig. 4.42 Precast concrete flooring units. Precasting ofconcrete allows higher levels of quality control than arepossible with in situ casting, which allows greater strengthto be achieved from a particular set of mix proportions.This permits the use of reduced thicknesses for a givenload-carrying capacity and the use of more complex cross-sections. The efficiency of the units depicted here isimproved by the introduction of voids into the core.

used in the context of heavily serviced build-ings where the combination of structure andservices is desirable.

A wide variety of precast concrete propri-etary components is currentlyavailable rangingfrom simple rectangular-cross-section beamsor slabs (Fig. 4.42) to complete framingsystems for buildings (Fig. 4.43). In caseswhere no proprietary system is available,bespoke elements are used provided the scaleof the project is such as to justify the settingup of the necessary production process.

Among the disadvantages of precasting arethe difficulties associated with connectingelements together satisfactorily. Precastingalso favours the adoption of building formswhich are regular and repetitive because thissimplifies the erection process and allowsmaximum advantage to be taken ofeconomies associated with the mass produc-tion of identical units. The restrictions onform which are associated with steel frame-works are therefore also a feature of precastconcrete structures.

Fig. 4.43 Precast concrete framework. The precastingtechnique can be applied to all types of structural elementand makes possible the construction of complete precastframeworks.

Precast components are frequentlycombined with in situ concrete and this allowsthe advantages of both forms of constructionto be enjoyed (Fig. 4.44). The use of in situconcrete at the junctions of precast frameworkscan solve the problems associated with joint-ing the units together, particularly the 'lack-of- 133


Fig. 4.44 Composite in situ and precast concrete. Precastunits of complex shape are used here in conjunction within situ concrete to form a composite floor deck with anefficient 'improved' cross-section.

fit' problem and the provision of structuralcontinuity. In situ concrete can also be used toallow more complex or irregular overall formsto be adopted.

4.4.2 Structural forms - cast-in-situ forms

4.4.2.1 IntroductionReinforced concrete cast-in-situ forms are usedprincipally for multi-storey buildings in which aframe-type structure is required. The form ofthe structure is determined by the same factorsas influence the design of all structures, whichare the need to provide effective resistance to

Fig. 4.45 Four basic types of reinforced concrete struc-ture.(a) Two-way-spanning ribless flat-slab.(b) One-way-spanning ribbed flat-slab.(c) One-way-spanning slab supported on rigid frame.(d) One-way-spanning slab supported on loadbearingwalls.

both gravitational and lateral load, and toachieve reasonable economy in the use ofmaterial. Although the mouldability ofconcrete theoretically gives the designer a widechoice of frame geometries, the need tominimise costs normally favours the use ofrectilinear arrangements which require simplepatterns of reinforcement and formwork. Thevariety of multi-storey forms which are used in


134

Table 4.1 Span ranges and size requirements for the four basic types of in situ reinforced concretestructure

Structure type Span range (m) Span/Depth Column/wall

Reinforced Pre-stressed Reinforced Pre-stressed slendemess

One-way-span slab 4X6 to 8X11 8X11 to 8X14 25 36 15 to 20Two-way-span slab: solid 4X4 t o 6 X 6 6X6to10X10 25 to 30 30 to 35 15to20

coffered 6X6 to 18X18 8X8 to 20X20 35 30 15 to 20Beam/Column frame:

one-way-span slab 3X6 to 6X12 6X12 to 8X15 slab 36 36 15 to 20beam 15 to 20 20 to 25

two-way-span slab 4X4 to 8X8 8X8 to 15X15 slab 36 36 15 to 20beam 15 to 20 20 to 25

Loadbearing wall:one-way-span slab 3 to 12 36 15 to 20two-way-span slab 3X3 to 12X12 36

(b)

(a) (c)

(d)

practice is nevertheless considerable and theyare categorised here into the four basic typesof one-way-spanning flat-slab structures, two-way-spanning flat-slab structures, beam-and-slab structures and loadbearing-wall structures(Fig. 4.45).

All four basic types are beam-and-postarrangements and can be considered to consistof a floor deck system supported on a verticalstructure of either columns or walls. Theirproperties, span capabilities and sizingrequirements are summarised in Table 4.1.

In all cases, the design of the floor deck isdetermined principally by the requirements ofgravitational load and this in turn dictates thepattern of the column or wall grid which isprovided to support it. One-way-spanning floorsystems are best carried on rectangular gridswhile two-way systems perform better oncolumn or wall grids which are square. Ineither case the grid is normally kept as regularas possible for reasons of economy but it neednot be perfectly regular.

The principal effect of lateral loads is uponthe design of bracing systems. In many casesthis will have no influence on the overall formof the structure because the beam-columnarrangement will be self-bracing due to thehigh level of structural continuity which ispossible with reinforced concrete. Somereinforced concrete structures require bracingwalls for stability, however, and where theseare necessary the internal planning of thebuilding is affected.

Only the most basic, regular forms of eachtype of structure are described here to give anindication of the general arrangements andspan ranges for which they are suitable. Oftenthese basic forms are manipulated anddistorted to produce more complex structuralgeometries (see Section 2.5).

4.4.2.2 One-way-spanning flat-slab structuresIn this system the floor slab spans one waybetween rows of columns (Fig. 4.46). Thearrangement is also known as ribbed-slabbecause the slab is frequently given a ribbedcross-section in order to improve its efficiencyby removing concrete from the tensile side of

Fig. 4.46 One-way-spanning flat-slab system. The ribbedversion depicted is used at the long span end of the spanrange for improved efficiency. For short spans, one-way-spanning flat-slabs have a simple rectangular cross-section.

the cross-section (Fig. 4.46). The ribs follow thespan direction of the slab. The voids betweenthe ribs are stopped short of the strip of slabbetween the columns and this solid area actsas a beam spanning between the columns. Asthis is a one-way-spanning system the columngrid is rectangular with the slab spanningparallel to the long side of the rectangle. Thenormal span range i s 4 m X 6 m t o 8 m X l l mbut the maximum span can be increased to8 m X 14 m if pre-stressing is used. Thespan/depth ratio is normally around 25 but canbe as high as 36 in pre-stressed versions.Ribbed-slab structures are braced either byrigid frame action or by walls acting asdiaphragm bracing in the vertical plane.

The particular advantages of the ribbed-slabsystem are the relative simplicity of theconstruction and high structural efficiency,which allows relatively long spans to beachieved with a small volume of concrete anda small depth of structure.

4.4.2.3 Two-way-spanning flat-slab structuresIn two-way-spanning slab structures the floordeck system consists of a two-way-spanningflat plate of reinforced concrete which issupported directly on a grid of columns (Fig.4.47). The column grid is normally square tomaximise the two-way-spanning action. The 135


136

Drop panel

structure is highly statically indeterminate andthe resulting structural continuity allows thenecessary strength and rigidity to be achievedwith great efficiency. The economic span rangefor a solid slab is 4.5 m to 6 m which isincreased to 10 m if pre-stressing is used. Thespan/depth ratio is typically in the range 25 to30 for reinforced slabs and 30 to 35 for pre-stressed slabs. An indication of typical slabdepths is given in Table 4.2.

Fig. 4.47 Two-way-spanning flat-slab system.


The critical stresses in solid slabs are shearstresses which are very high in the vicinity ofthe supporting columns. The strength of thefloor can be increased by thickening the slablocally at these points by a system of 'drop'panels or by using 'mushroom-type' columncaps (Fig. 4.48); alternatively the slab can bestrengthened in shear in the vicinity of thecolumns by the use of extra reinforcement inthe form of steel shearheads (Fig. 4.49). Thesevariations allow heavier loads to be carried orhigher span-to-depth ratios to be achieved.The use of shear heads also allows voids forservices ducts to be located close to thecolumns.

Fig. 4.48 Column caps and drop panels. The column capand drop panel can be used independently or together toincrease the shear strength of flat-slabs in the vicinity ofcolumns where high shear forces require a greater thick-ness of slab.

Table 4.2 Overall slab thickness and columnwidths for two-way-spanning flat-slab structures

Column spacing Slab thickness Column width

(square column (mm) (mm)

grid)

(m) Solid Coffered

4 150 - 2005 175 - 2006 200 300 2507 250 300 2508 275 300 2509 300 400 300

10 400 30012 - 500 40014 - 500 50016 - 600 60018 - 700 700

Fig. 4.49 Steel shearheads, which are cross-shapedarrangements of steel plates joined by welding, are analternative method to column caps or drop panels forincreasing the shear strength of flat-slabs in the vicinity ofcolumns.

The maximum size of column grid of thistype of structure can be increased to around18 m for reinforced slabs and 20 m for pre-stressed slabs if a coffered system is used toreduce the weight of concrete present in theunderside of the slab. The span/depth ratiosfor coffered slabs are around 25 forreinforced slabs and 30 for pre-stressedslabs.

The two-way-spanning slab system is de-pendent, to a large extent, on structural conti-nuity for its strength. It performs best (i.e.allows the highest ratios of span-to-depth)with column grids which have at least threebays in each direction and in which the varia-tion in the sizes of the bays is kept to aminimum. The efficiency of the system is such,however, that slab thicknesses are approxi-mately the same as those in frame structuresof equivalent span, i.e. a system in which theslab is supported on downstand beams.

Fig. 4.50 Typical arrangements of vertical-plane bracingfor flat-slab structures, which should be as symmetrical aspossible, are shown here.

Where a thin, solid slab is used its stiffnesscan be insufficient to allow rigid-frame actionto develop with the columns in response tolateral load. Extra bracing must therefore beprovided and in situ concrete walls are usuallyincorporated into the structure for thispurpose. These must be arranged in twomutually perpendicular directions and canusually be accommodated around lift or stairtowers (Fig. 4.50). In the long-span part of therange, that is for spans greater than around10 m, the depth of the coffered slab which isrequired will be greater than 350 mm and theslab is usually sufficiently stiff to allow rigid-frame action to develop between the floor andcolumns, in response to lateral load; in thesecases no additional bracing is required.

The high degree of statical indeterminacywhich is associated with the two-way-spanningslab allows greater flexibility in the column gridthan is possible with frame structures. The smallconstruction depth of flat-slab structures,compared to frame structures, also facilitates theeasy accommodation of a services zone and willusually result in a lower overall storey heightthan is possible with a beam-column frame. 137


Fig. 4.51 Plan arrangements for reinforced concretebeam-column frames.(a) One-way-span slab spanning between parallel beams.(b) Two-way-span slab on a grid of beams.

Due to the simplicity of the formwork andthe high level of structural efficiency which isachieved as a result of the high degree of stat-ical indeterminacy, two-way-spanning slabstructures of both the plain and the cofferedtypes provide very economical support systemsfor multi-storey buildings in which large wall-free areas are required. They are particularlysuitable where the imposed load is high anduniformly distributed but are less suitablewhere concentrated loads are high, forexample in buildings in which machinery hasto be housed. They are also unsuitable forbuildings in which the benefits of structuralcontinuity are limited, either because largebreaks in the continuity of floors occurs (Fig.2.14) or because the building has a compli-cated geometry in plan and section (Fig. 4.15).In such cases a frame structure is normallyrequired.

4.4.2.4 Frame structuresThe distinctive characteristic of a frame is thatit consists of an arrangement of beams andcolumns which supports slab floors. There aretwo basic types of reinforced concrete frame,the distinguishing feature being whether thefloor is a one-way-spanning or a two-way-spanning system (Fig. 4.51).

In the frame with a one-way-spanning floorthe floor-slab spans as a continuous system

Fig. 4.52 Beam-column frame with one-way span slab.One-way span slabs are normally carried on parallelarrangements of beams supported by a rectangular columngrid. No beams are provided in the direction of the slabspan.

across a series of beams which are supportedindividually on columns. The beams act withthe slab to form flanged T- or L-beams (Figs4.37 and 4.52). Linking beams, running in thesame direction as the slab, are not normallyprovided between the columns because theslabs themselves provide an adequate struc-tural connection. The whole structure ofbeams, columns and slabs is cast in stages toform a continuous monolithic unit. Typical138


5mto10m

3 m to 6 m

Table 4.3 Span range and principal dimensions of reinforced concrete frame structures

Slab span Beam span Slab thickness Beam depth Column widthfrom top of slab)

(m) (m) (mm) (mm) (mm)

3 4.5 125 350 2004 6.0 150 420 2505 7.5 175 520 2756 9.0 200 670 2757 10.5 225 780 2758 12.0 275 900 3009 13.5 300 1060 300


dimensions for this type of arrangement aregiven in Table 4.3.

Frames of this kind can extend to a largenumber of bays in each direction and are bestplanned on a rectangular column grid with theslabs spanning parallel to the short side of therectangle. The normal span range is 3.5 m X6 m to 6 m X 12 m for reinforced concrete andthis can be extended to 8 m x 15 m if pre-stressing is used. The most economic gridratios vary between 1:1.5 at the short-span endof the range to 1 : 2 for longer spans. Thespan/depth ratio is around 36 for the slab and15 to 20 for the beam.

So far as resistance to lateral load isconcerned the one-way-spanning frame is rigidand self-bracing in the plane of thebeam/column frames, due to the relatively highstiffness of the beams and columns and therigid joints which exist between them, but isnot stable in the direction of the span of theslab, because the stiffness of the slab isnormally insufficient for effective rigid-frameaction to be possible (Fig. 4.53). Additionalbracing is therefore necessary and this isnormally provided in the form of in situconcrete walls which act as vertical-planediaphragm bracing (Fig. 4.51). These areconstructed by simply extending pairs ofcolumns to fill the space between them. Aswith other types of vertical-plane bracing itperforms best if it is disposed around thebuilding in a symmetrical arrangement and canusually be conveniently located at stairs andservice ducts. The need to position the verti-

cal-plane bracing correctly is a factor whichaffects the internal planning of buildings whichhave this type of structure.

Where a two-way-spanning slab is used inconjunction with a grid of beams it is neces-sary that the slab spans should be more-or-less the same in each direction. The columngrid must therefore be more-or-less square(Fig. 4.51). The two-way-spanning floor struc-tures have higher degrees of statical indeter-minacy than one-way systems and this allowsthinner slabs and shallower beams than theequivalent one-way system; it also increasesthe maximum economic span for a solid slabto about 8 m. Because frames with two-way-spanning floor systems have beams running intwo mutually perpendicular directions they arecompletely self-bracing and do not require any

Fig. 4.53 Bracing of one-way-span frame systems by rigidjoints is ineffective in the direction of the slab span due tothe flexibility of the slabs. 139


Fig. 4.54 Plan arrangements for loadbearing-wall struc-tures of reinforced concrete.

structural walls for stability. This is sometimesthe reason why a frame, as opposed to a flat-slab structure is adopted.

The frame types which have been describedhere are the most basic forms. Considerablevariation from these is possible although thiswill normally increase the cost of the structure.The simplest variation is the displacement ofindividual columns from a strictly regular grid,in order to accommodate some aspect of thespace-planning of the interior. If the displace-ment is kept within one quarter of the span itcan be accommodated easily by strengtheningthe structure locally. Another common vari-ation is a small change in the level of a floorover a short area of the plan; this too can beeasily accommodated in in situ reinforcedconcrete, as can the ramps and stairs whichare required for access. More significant varia-tions from the standard forms are illustrated inFigs 2.8, 2.10 and 2.14.

Fig. 4.55 Plan arrangement for reinforced concreteloadbearing-wall structure with one-way-spanning floorslabs. Note that this conforms to the parallel-wall arrange-ment which is typical of loadbearing wall-structures in allmaterials.

4.4.2.5 Loadbearing-wall structuresThe planning principles for reinforced concreteloadbearing-wall structures are similar to thosewhich are used for masonry structuresalthough the arrangement of walls is normallysimpler because the greater flexural strength ofconcrete eliminates the need for local stiffen-ing to combat buckling. Three plan-forms arecommonly used: cross-wall and spine-wall,with one-way-spanning floors, and cellular,with two-way-spanning floors (Fig. 4.54 and4.55). The spacing between the walls is kept asuniform as possible so that the slab spans areequal and maximum benefit is obtained fromcontinuity. The most economic spans arearound 5 m to 6 m. In cross-wall and spine-wall structures, in which the loadbearing walls140

Plan showinglayout of rooms

Plan showing

structural elements

Non-structural face panels

Non-structural partitions 1m 6.9 m

5m

2.5m

2.5 m

3.7 m

2.5 m

2.5 m

5m

are parallel to one another, it is necessary tobrace the structure with a limited number ofwalls running in the orthogonal direction;these can normally be located around stairsand service cores.

The advantages of reinforced concrete overmasonry for this type of structure is that itallows the structural plan to be simpler andgives much more planning freedom generally:it is, for example, possible to omit some of theloadbearing walls at occasional locations andbridge the gap by reinforcing the wall above sothat it acts as a deep beam.13 The disadvantageof reinforced concrete is that it is almostinvariably more expensive than the equivalentmasonry structure.

4.4.2.6 Summary of cast-in-situ forms\n situ reinforced concrete structures performbest in circumstances in which imposed loadsare high and they are therefore used principallyto support multi-storey commercial and indus-trial buildings and only rarely as the structuresfor low-rise domestic-type buildings or single-storey buildings.

The main alternative to reinforced concretefor the multi-storey building is steel and theparticular advantages of reinforced concreteare that it allows complex and irregular formsto be achieved more easily than with steel andthat it is both durable and fire resistant, whicheliminates the need for finishing materials.

4.4.3 Structural forms - precast forms

4.4.3.1 IntroductionIn precast concrete construction the structuralcomponents are cast in a location which isdifferent from that which they will finallyoccupy in the structure. They are normallymanufactured in a precasting factory which isremote from the site but they may sometimesbe cast on the site. The advantage of the latteris that it removes the constraints imposed by

the need to transport the components to thesite. In either case the main advantage ofprecasting is the achievement of higher qualitycontrol, which results in higher strength, betterdurability and better surface quality than ispossible with equivalent in situ concrete. It alsoreduces the time required to construct thebuilding on site.

The particular advantages and disadvantagesof precast concrete are summarised below.

Advantages1 Almost all of the advantages of in situ

reinforced concrete are obtained. The mater-ial has relatively high strength in compres-sion, flexure and tension and is thereforesuitable for all types of structural elementand for skeleton-framework arrangements. Itis also durable and fire resistant, whichfacilitates the exposure and expression ofthe structure.

2 The precasting of concrete allows betterquality control to be achieved. The materialis therefore stronger than equivalent in situconcrete and has a better standard ofsurface finish. Structural elements cantherefore be more slender and are less likelyto require the application of a finishingmaterial to bring them to a satisfactoryvisual standard.

3 Precasting allows greater complexity of theform of individual components to beachieved. This characteristic can beexploited in several ways. In heavily servicedbuildings, in which a large number ofservice ducts are required, it allows precastconcrete columns and beams of complexcross-section to be used as frame elementsand as service ducts. Precast concreteframes have therefore been fairly widelyused for building types, such as hospitalsand laboratories, in which the provision ofservices is a major factor in the design. Theprecasting technique has also been widelyused for the manufacture of proprietarycomponents with complex cross-sectionalshapes.


141

13 The most famous building in which this technique wasexploited was perhaps the Highpoint flats by Lubetkinand Arup.


4 Precasting of components leads to a muchsimpler site operation than is possible within situ concrete. The building can thereforebe erected more quickly with fewer tem-porary structures.

Disadvantages1 One-off precast concrete systems tend to be

more expensive than in situ equivalents. Thisfavours their use for large-scale projectswhere economy of scale reduces the differ-ential. It is claimed by the precast concreteindustry, however, that the overall buildingcost is reduced if a precast structure is usedinstead of in situ concrete, due to thesimpler and shorter site operation.

2 Inflexibility. Precast concrete units mustnormally be manufactured some time inadvance of their being transported to thesite and installed in a building to give timefor the concrete to gain adequate strength.(Sometimes the highest load to which acomponent will be subjected will occurduring transportation or installation.) Latechanges to the design of the buildingcannot therefore be readily accommodated.

3 Standardisation. Another aspect of inflex-ibility is that, because precast structures areassembled on site from components whichare fairly large, it is convenient to maintainthe overall geometry of the structure in assimple a form as possible.

There is considerable incentive tostandardise components so as to obtain themaximum re-use of moulds and to facilitateerection. Thus, while precasting offers thepossibility that individual elements canhave complex cross-sections and profiles,the overall form of the building mustnormally be relatively simple and repetitive.This is a disadvantage which precastconcrete shares with steel.

4.4.3.2 Precast frame structuresPlanning principles for precast concrete framesare the same as for in situ concrete frames:floors are normally of the one-way-spanning

type, in which case a rectangular beam-column grid is used but two-way-spanningfloors, on a square column grid, are also possi-ble. Normally the structure consists of individ-ual beam, column and slab units which areerected on site in a similar way to a steel frame(Fig. 4.56); the joints between elements can beof the hinge or rigid type, depending on thedetailing. If hinge-type joints are used (Fig.4.57), additional bracing components arerequired and these can take the form of infillwalls, either in situ or precast, or diagonalbracing. Where the beam-column joints arerigid the frames are self-bracing. Frequently inprecast frame construction the joints betweenthe individual units do not coincide with thebeam-column junctions (Fig. 4.58). The unitsthen have a fairly complex geometry whichmakes transport and stacking on site moredifficult; the advantage is that it makes pos-sible the achievement of a self-bracing framewithout the need for site-made joints whichare of the rigid type.

Fig. 4.56 Layout of a basic precast concrete framework.(a) Typical plan arrangement in which a beam-columnframe supports a one-way-spanning floor slab.(b) and (c) Slab units can be of rectangular or ribbed cross-section depending on the span.142

Slab span

Beam span (a)

(b) (c)

Table 4.4 Span range and principal dimensionsof basic precast concrete frames

Slab span Beam span Slab depth Beam depth(m) (m) (mm) (mm)

4 6.0 140 4505 7.5 140 6006 9.0 150 7007 10.5 190 8008 12.0 190 10009 13.5 190 1150

10 15.0 250 130011 16.5 250 140012 18.0 250 1500

It is normal for precast concrete frameworks tohave a regular, rectilinear form so as tomaximise the standardisation of componentsbut it is possible to adopt irregular grids. Typicalelement sizes for the normal span ranges ofrectilinear frames are given in Table 4.4.

Columns are normally rectangular in cross-section but other shapes can be provided ifthis is necessary to accommodate irregularbeam layouts or for architectural effect. Themost favoured beam cross-section is theinverted T, as this facilitates the carrying ofsimple slab types. More complex shapes areused to accommodate changes in floor level.

Precast floor slabs are normally of the one-way-spanning type and are either solid,hollow-core or with a T-profile (Fig. 4.56c). Allof these are ideally suited to a rectangularbeam layout. Rhomboid plan-forms are fairlystraightforward to produce but where the planshape is highly irregular the section of floorinvolved is cast in situ.

Where vertical-plane bracing is provided bystructural walls these can be precast units andcan be used to provide support for floors aswell as to resist lateral load. They are normallyarranged as bracing cores around lift or stairwells. A reasonable number of bracing wallsmust be provided in two orthogonal directionsand these should be arranged as symmetricallyon plan as practicable.

4.4.3.3 Hybrid in situ and precast formsHybrid structures, in which both in situ and

Fig. 4.57 Joints in precast concrete frameworks. All of thejoints shown are capable of transmitting shear and axialforce only. They are therefore hinge-type joints.

precast concrete are used, allow the advan-tages of both forms of construction to berealised. The precast components will bringthe benefits of factory production (highstrength and efficiency, durability, goodappearance, complex element cross-sections,dimensional accuracy, rapid erection) and thein situ parts allow complex or irregular overallforms and structural continuity betweenelements to be easily achieved.

The in situ and precast components in hybridstructures can be combined in two principal 143



Fig. 4.58 Where rigid joints are required in precastarrangements, so that the structure is self-bracing, this isfrequently accomplished by positioning the junctionsbetween the elements at different locations from thebeam-to-column connections.

Fig. 4.59 Hybrid in situ/precast structure. In this structureprecast concrete perimeter units are used in conjunctionwith an in situ concrete framework. The superior standardof finish which precasting allows is exploited to allow theunits to be exposed on the exterior of the building. [Photo:British Cement Association]

ways. In the first of these the structure consistsof a mixture of element types each of which iseither precast or in situ (Fig. 4.59). The ratio ofprecast to in situ components can vary widely.At one extreme an in situ framework of beamsand columns, with rectangular shapes of cross-section, might be combined with precast stairand ribbed-slab elements whose more compli-cated profiles are more easily achieved underfactory conditions. At the other extreme, theuse of in situ concrete might be confined to themaking of continuous joints in a structure inwhich all of the principal elements wereprecast.

A second type of hybrid structure is one inwhich the individual elements are formed by acombination of precast and in situ concreteacting compositely. With this type of arrange-ment the precast parts of the structure areinvariably used as permanent formwork onwhich the in situ parts are cast (Fig. 4.60). Thusa ribbed slab may be formed from a slender

precast soffit of complex shape on to which anin situ top is cast. Beams of complex shape canbe formed in the same way. Composite hybridstructures are particularly well suited to therealisation of the advantages of both precastand in situ concrete.

The general arrangements of hybrid struc-tures are similar to those for precast forms.They are either beam-column frameworks orflat-slab arrangements with column gridswhich are either square or rectangular depend-ing, respectively, on whether the floor slabs aretwo-way or one-way-spanning systems.144

Fig. 4.60 In this hybrid struc-

ture precast beam and slab units

act compositely with an in situ

concrete topping slab.

4.4.4 Curved forms and structures ofcomplex geometryThe mouldability and strength of concreteallow it to be cast into a wide variety of shapesand it has been extensively used to createenvelopes which are based on curved forms(Fig. 4.61). The fact that these are moreefficient structural forms than the post-and-beam structures which have already beendescribed, either because they are closeapproximations to the form-active14 shape forthe loads which are applied to them orbecause they allow the proportion of under-stressed material which is present to bereduced in some other way, means that largespans can be achieved with very small volumesof structural material. Spans of up to 70 m arepossible with shell thicknesses in the range40 mm to 250 mm (Tables 4.5 and 4.6) and thisrepresents a considerable saving comparedwith the volume of material which would berequired to achieve the same span by usingpost-and-beam types of structure. A highdegree of expertise is necessary, however, bothin the design and in the construction of struc-tures which have a complicated geometry andthis tends to make them more expensive thanforms with simpler geometries, despite the

14 See Appendix I for an explanation of the form-active

concept.

large saving in material which is involved.Considerable reductions in the cost of thedesign and construction of shells is possible ifsome degree of regularity is introduced. Forthis reason most curvilinear or folded forms ofstructure are given a very simple basic geom-etry. The most favoured shell forms are those

Table 4.5 Typical thicknesses of hyperbolicparaboloid shells in reinforced concrete

Span Shell thickness (mm)

(m) At crown At edges

10 40 50

20 40 75

30 40 100

40 75 130

Table 4.6 Typical thicknesses of ellipticalparaboloid shells In reinforced concrete

Span Shell thickness (mm)

(m) At crown At edges

10 40 50

20 50 80

30 60 120

40 70 170

50 80 200

60 100 220

70 130 250


Pre-stressed concreteedge beam

Precast trough units

Shear links

Pre-stressed soffit slabto main beam

In-situ concrete structural topping

Bars in edge oftrough unit acting asreinforcement insecondary beams

145


Shell forms have other disadvantages besidesthe high cost of construction. They do notperform well when subjected to concentratedloads and their use must therefore normally berestricted to that of providing a free-standingenvelope to which no other substantial compo-nents, such as services ducts or servicesmachinery, are attached. The fact that the struc-tural envelope of a shell-type building is verythin can produce a number of difficulties for thedesigner, which are not present when moreconventional types of structure are used. If, as isusually the case, the shell is used as the soleexternal envelope an obvious deficiency is itspoor thermal insulation and low thermal mass.Another disadvantage arises from the fact that itis not possible to accommodate the largenumber of systems which occur within a build-ing, such as wires, pipes, etc., within the verysmall structural zone of a shell. This problemdoes not arise in buildings which have moreinefficient post-and-beam type structuresbecause the fairly large structural volumeinvolved provides a zone within which manyrequired components (ducts, cable runs) can beaccommodated; where a very efficient structuralform is used these components must be locatedelsewhere. The high complexity of the shellform, combined with its other disadvantagestherefore make it suitable for a rather limitedrange of building types.

4.4.5 ConclusionThe advantages of reinforced concrete as a struc-tural material stem from its physical propertiesand its mouldability. Its properties allow it to beused in both compressive and bending-typestructural components and its strength is suchthat it can be used to make frameworks in whichfairly large spans are involved. Of the non-struc-tural physical properties, the good durability andfire resistance of concrete are probably the mostimportant; these make it fit for use as a buildingsupport system without the need for extra finish-ing materials or components. The mouldabilityof concrete allows good structural continuity tobe achieved, through the use of cast-m-situ rigidjoints, and makes possible a considerablevariety of geometries.

Fig. 4.61 Reinforced concrete lends itself to use incompressive form-active structural elements. The greatefficiency of this type of structure allows the strengthrequired for long spans to be achieved with very thin shells- typically between 50 mm and 150 mm. Various types ofhyperbolic paraboloid shell are depicted here.

which can be described by an elementarymathematical relationship based on theCartesian co-ordinate system. Examples arethe cylinder, the sphere, the elliptical parabol-oid, the hyperbolic paraboloid and the conoid(see Fig. 4.61 and also Joedicke, Shell ArchitectureKarl Kramer Verlag, Stuttgart, 1963).146

Chapter 5

Masonry structures

5.1 Introduction

Masonry has been used for the construction ofbuildings from earliest times and the history ofits use comprises buildings in virtually everyarchitectural style. It is a composite material inwhich shaped natural stone or manufacturedbricks or blocks are bedded in mortar. It isbrittle, with moderate compressive strength butlittle strength in tension and it thereforeperforms well in structural elements in whichaxial compression predominates, such as walls,piers, arches, domes and vaults, but will crack ifsubjected to tensile stress resulting from theapplication of bending. Where masonryelements carry bending-type loads they must bedesigned such that tensile stress is maintainedat a very low level. If the element is subjected topure bending, as in the case of a lintel or archi-trave, the tensile bending stress has to beminimised by the adoption of generous dimen-sions for the cross-section and an acceptance ofvery short spans. If the element is subjected toa combination of bending and axial compres-sion the arrangement must be such that thetensile bending stress never exceeds the level ofaxial compressive stress. This is done by ensur-ing that the elements are sufficiently thick.1

Where the intensity of bending is high, ascan occur in long or high stretches of wallsubjected to wind loading or where high wallssupport vaults or domes, the overall thicknessof element required to contain the bending

stress within reasonable limits can be large. Insuch cases various measures can be adoptedto achieve high overall thickness without theneed for a large volume of masonry. Often, thestructural devices which make this possibleprovide the stimulus for architectural innov-ation (see Section 5.2).

The forms which are appropriate for masonrystructures are suitable for any material withsimilar properties (i.e. materials with moderatecompressive strength but little strength intension or bending). Baked earth, mud andunreinforced concrete are examples of suchmaterials.

5.2 The architecture of masonry -factors which affect the decision touse masonry as a structural material

5.2.1 The aesthetics of masonryThe architectural vocabulary of masonry is thatof compression: the structural forms are thoseof the loadbearing wall, the buttress, the arch,the vault and the dome. The two great trad-itions of Western architecture, the Gothic andthe Classical, were each developed through themedium of masonry structures and each foundits greatest architectural expression inmasonry. Masonry accounts for some of themost spectacular of the world's buildings, bothin terms of physical scale and visual qualities.

The architectural use of masonry is herereviewed under the three headings of vaultedhalls, domes and post-and-beam structures.Much of the architecture described is pre-twentieth century but this is necessary so thatthe full range of the architectural potentialoffered by masonry can be illustrated. 147

1 The factors on which the relationship between bendingstress and the dimensions of structural elementsdepends are explained in Macdonald, Structure andArchitecture, Chapter 2.


Fig. 5.1 CathedraleSte-Cecile, Albi, France,13th century. Thisbuilding, which has thewidest vaulted nave ofall the French Gothiccathedrals (18 m span),is constructed of brick-work masonry anddemonstrates the scaleand complexity of archi-tectural form which ispossible in thismaterial. [Photo: A.Macdonald]

148

Vaulted HallsVaulted hall is the generic term for a type ofbuilding which consists of a large interiorspace constructed entirely of masonry; aspace, in other words, which is roofed by amasonry vault supported on masonry walls.This group of structures encompasses some

of the largest and most spectacular ofmasonry buildings, including the large basili-cas and bath houses of Imperial Rome andthe Gothic cathedrals of medieval Europe(Fig. 5.1).

Where a large enclosure is made entirely ofmasonry the use of a compressive form-active

Fig. 5.2 Failure of a vault. Vaults exert horizontal forceon their supports. If the supports are high walls, thesemust be sufficiently thick not to overturn and have suffi-cient bending strength not to be overstressed by thebending moment caused by the horizontal force at theirtop. If significant horizontal movement occurs at the topsof the walls the vault is likely to fail in the manner shownhere.

structure,2 such as a vault or a dome, is neces-sary to achieve the horizontal span withoutincurring high bending and therefore tensilestress. The action of a vault produces horizon-tal thrusts at the points of support, however,which, in the case of a vaulted hall, are thetops of the supporting walls (Fig. 5.2). Thewalls are therefore subjected to a combinationof axial compression, due to the weight of thevault, and bending, due both to the lateralthrust of the vault and to any eccentricitywhich is present in the transfer of its weight tothe wall. The configuration of the structuremust be such that the axial compressive stressin the walls is always greater than the tensilecomponent of the bending stress. This canonly be achieved if the wall is of adequatethickness. The very thick walls required to

Fig. 5.3 Evolution of the form of the vaulted hall.(a) The earliest vaulted halls relied on very thick walls forstability. The volume of masonry involved was very large.(b) The use of the concept of the buttress allowed a degreeof stability to be achieved which was equivalent to thatprovided by the solid wall but with a greatly reducedvolume of masonry.(c) The cross-vault concentrates the load at the buttressesand creates flat areas of wall above the springing point ofthe vault which may be used as fenestration. 1492 See Appendix 1.

Masonry structures

(a)

(b)

(c)

Fig. 5.4 Basilica Nova, Rome, 4thcentury CE. The architects and engineersof Imperial Rome exploited the architec-tural opportunities offered by thebuttressed, cross-vaulted hall to createinteriors of high architectural quality.


support high-level vaults do not have to besolid, however, and much of the architecturalexpression of large vaulted hall structures hasbeen derived from the manner in which thevoiding of the walls was carried out.

The earliest of the large vaulted halls werethe basilicas and bath houses of ImperialRome. The evolution of the form is showndiagrammatically in Fig. 5.3. The simplest formof the building consisted of a barrel vaultsupported on very thick, solid walls (Fig. 5.3a).The efficiency of this structural arrangement islow and is improved if voids are incorporatedinto the walls (Fig. 5.3b). The solid areasbetween the voids are, in effect, buttresses; thevoids are also vaulted. A logical progressionfrom this configuration is the adoption of across-vault arrangement (Fig. 5.3c) as thisconcentrates the load from the vault on to thesolid areas of the walls. It also creates flatareas of wall above the original wallhead level(i.e. in the zone of the vault) in whichclerestory lighting can be inserted.

The Basilica Nova (4th century CE) (Fig. 5.4)and the main building of the Baths of Caracalla(3rd century CE) are examples of this type ofstructure. In both cases the structural armatureconsisted of the vaulting system describedabove, constructed in mass concrete, encasedin a thin skin of brickwork which served aspermanent formwork (Fig. 5.5). In both of theseexamples this structural armature was clad inmarble to create a sumptuous interior.150

Fig. 5.5 Detail of Basilica Nova, Rome, 4th century CE.The largest interiors in Rome were constructed in unrein-forced concrete, which was placed in a thin skin of brick-work which acted as permanent formwork. The structuralcarcass was then faced in marble to create a sumptuousinterior. The system of construction allowed very large-scale buildings to be constructed economically.

The architectural opportunities offered bythe structural vocabulary of masonry weretherefore fully exploited by the Roman builderswho used them to create sequences of interiorspaces which varied in height, volume andlighting intensity. Buildings of this type playeda significant role in the development of theclassical language of architecture. The stylistic

devices which evolved in response to the formswhich were adopted out of structural necessitywere to be much imitated at the time of theItalian Renaissance, and subsequently.

The Gothic cathedrals of the medievalperiod, in which vaulted roofs were balancedon highly complex buttressed wall systems,belong to the same generic type of building asthe vaulted basilicas and bath houses ofImperial Rome. It is interesting to note,however, that, although the architects of theGothic period were faced with very similarstructural problems to those of their Romanpredecessors and that their structural solutionto it was similar, they produced an architecturewhich was quite different in style and characterfrom that of the Romans.

The structural solution adopted by theGothic builders was, therefore, similar struc-turally to that of the largest Roman buildings.The vaulted roof of a large hall (the nave ofthe church) was supported on vertical struc-tural systems of great thickness but fromwhich much mass was extracted to improvethe efficiency with which material was used(Fig. 5.6). In the case of the Gothic buildingsthe sculptural effect occurred on the exteriorof the building and took the form ofbuttresses, flying-buttresses, finials and all ofthe other elements of the Gothic architecturalvocabulary. The basic structural arrangementwas the same as in the Roman buildings butthe architectural treatment of it was entirelydifferent.

In some of the later buildings of the Gothicperiod the degree of refinement was extreme(Fig. 5.7). In these buildings the loads from thevaulting were concentrated into slendercolumns braced by stabilising buttresses. Thewalls linking these were extensively piercedwith traceried windows. The walls, in fact, hada minimal structural function and had becomevirtually non-loadbearing curtain walls. Thestructural system of these late Gothic cath-edrals was really that of the rigid frame ratherthan of the loadbearing wall - a remarkableachievement in a material with minimal tensilestrength. The relationship between architec-tural aspiration and structural technology was

Fig. 5.6 Notre-Dame Cathedral, Paris, France, 12thcentury. As in the case of the Roman vaulted halls thevault here is supported on a wall system of great thickness.The architectural treatment of the buttressing system isquite different however.

perhaps closer in these buildings than has everbeen achieved subsequently.

Present-day tall single-storey buildings,which are used for single-space large enclos-ures such as sports halls, normally haveroofing systems which do not exert significantlateral thrust at the tops of the walls. They arenot therefore true vaulted halls. The walls arenevertheless subjected to significant out-of-plane forces due to the action of wind pressureand must have adequate strength in bendingto resist this. The strategies which are adoptedin modern practice to achieve this are similarto those which were used in historic times -that of providing high overall thickness withoutthe use of an excessive volume of masonry.The fin and diaphragm wall systems areexamples of this (Fig. 5.41). A notable alterna-tive method is the use of corrugations as isseen in the distinctive work of the Uruguayanarchitect Eladio Dieste (Fig. 5.8). 151

Masonry structures


Fig. 5.7 Notre-Dame Cathedral, Chartres, France, 12thand 13th centuries. The main structural elements here arethe quadripartite vaulting of the roof and the supportingpiers which act in conjunction with external flyingbuttresses. The sections of wall between the piers can beregarded as non-structural cladding. The structuralarrangement is therefore that of a skeleton framework - aremarkable achievement in a material such as masonry.[Photo: Courtauld Institute]

Fig. 5.8 Church at Atlantida, Uruguay, 1958. EladioDieste, architect. The plan-form of the walls of this build-ing is that of a sinusoidal curve whose amplitude is great-est at the level where the bending moment is highest. Thearrangement allows a large overall thickness of wall to beachieved with a minimal volume of masonry. This is amodern equivalent of the flying buttress. |Photo: Vicentedel Amo|

152

DomesThe dome is a similar structural form to thevault. In the history of the architectural treatmentof the dome, a similar gradual progression tothat which occurred with the vault may beobserved, that is a progression from fairly simple,solid arrangements to gradually increased levelsof refinement (Fig. 5.9). The structural action ofthe dome is slightly different from that of thevault, however, because although, in theory aperfectly intact dome exerts no horizontal thruston the supporting structure, which can be acylinder of masonry or even a ring of columns, inpractice, horizontal thrusts do frequently occur atthe supports. The precise behaviour depends onthe profile of the dome.

The easiest type of dome to construct is thehemisphere because the constant radius ofcurvature allows simple formwork to be used.The hemisphere is not an ideal structural form,however, because the lower parts ofhemispherical domes are subjected to horizon-tal tensile stresses in the circumferential direc-tion (hoop stresses), which can give rise tomeridional cracking (Fig. 5.10). The dome then

Fig. 5.9 Evolution of the form of the domed enclosure.(a) The earliest domes were carried on very thick solid walls.(b) At the Pantheon in Rome (2nd century CE) niches wereprovided in the thick walls and the underside of the domewas coffered to reduce the volume of material required.Both of these devices were exploited in the architecturaltreatment of the interior.(c) The dome of the Hagia Sophia, in Istanbul (6th centuryCE), and those of several other large Byzantine buildingswere carried on an arrangement of arches and pendentiveswhich transmitted the weight to four massive piers.

Fig. 5.10 Internal forces in a hemispherical dome. Ahemispherical dome which is intact exerts no lateral forceon the supporting structure. The lower parts of hemispher-ical domes are subjected to horizontal, tensile hoop'stresses, however, which can cause cracks to develop asshown. If this occurs the individual segments of the domebehave as arches and do exert horizontal force on thesupports. Most domes of masonry or unreinforcedconcrete exhibit cracking caused by hoop stress and dotherefore require support structures which are capable ofresisting horizontal load. 153

Masonry structures

(a)

(b)

(c)


Fig. 5.11 ThePantheon, Rome, 2ndcentury CE. This wasthe largest domedstructure of Romanantiquity. The coffer-ing of the inner faceof the dome andvoiding of the wallssuggest that the'architects'/'engineers'responsible for thisbuilding had a goodintuitive or empiricalunderstanding of thestructural behaviourinvolved. They usedthese features toexcellent architecturaleffect. [Copyright:RIBA Photo Library]

154

behaves as a series of arches, arranged arounda central axis, which exert a significanthorizontal thrust on the supporting structure.The effect is lessened if the profile of the domeis pointed rather than hemispherical. Cracking,and the resultant horizontal thrusts, can beminimised if material capable of resisting thetensile hoop stresses is incorporated into thedome structure. Iron chains have been used forthis purpose.

The earliest of the great domes of theWestern architectural tradition was that of thePantheon in Rome, constructed in the secondcentury CE (Fig. 5.11). This is 43 m in diameterand was made of Roman concrete. The internalprofile was hemispherical and the dome wassupported on a cylindrical drum of concretefaced internally in marble and externally inbrickwork. Meridional cracks developed in thisdome, in the manner described above, but theresultant outward thrusts were absorbed by thevery thick walls of the supporting drum. Thesupporting system of voided walls was of a

similar configuration to that of the systemswhich were used to support vaulted halls,which suggests that the Roman architects werewell aware that the drum would be subjectedto bending as a result of outward thrusts fromthe dome.

The dome of the Hagia Sophia in Istanbul(6th century CE) (Fig. 5.12) has a slightlysmaller span than that of the Pantheon (31 m)but the design is considerably more audacious.This dome is much thinner than that of thePantheon and it is supported on four piers, towhich the weight is transferred by four archesand pendentives (in a square arrangement onplan) rather than a cylindrical drum ofmasonry. The piers are heavily buttressed bythe surrounding parts of the building butnevertheless show signs that outwardmovement occurred due to the lateral thrustsproduced by the dome and its supportingarches. The configuration of the structuresuggests that the architects were aware of thecrack pattern which had developed in the

Fig. 5.12 Hagia Sophia, Istanbul, Turkey, from the sixthcentury CE, The dome which spans the central space iscarried on four arches which transmit the weight to fourmassive piers. These are buttressed by the surroundingparts of the building. The combination of dome and archesallowed a complex interior of very large volume to becreated. The architectural effect was one of a feeling ofimmense space. [Copyright: Rowland Mainstone]

advantages of this were that it allowed greatoverall thickness to be achieved without theoccurrence of a severe weight penalty, that theweathertightness of the structure wasimproved and that it allowed greater freedomin the choice of the shapes for the inner andouter visible surfaces of the dome.

The first very large structure to which thisidea was applied was the fifteenth-centurydome of Florence Cathedral (S. Maria del Fiore)by Brunelleschi (Fig. 5.13).4 The dome wasmuch higher in relation to its span than thoseof the Pantheon or the Hagia Sophia and thiswould have reduced the hoop stresses.

4 An account of the construction of this very significantstructure has been given elsewhere - see Mark (op. cit.)and Mainstone, Developments in Structural Form, London,1975, Chapter 7.

3 See Mark (ed.), Architectural Technology up to the Scientific

Revolution, Cambridge, MA, 1993, Chapter 4.

Pantheon dome3 and that they had a soundqualitative awareness of the behaviour of form-active masonry structures.

A significant development in the history ofthe masonry dome was the use of two struc-tural skins separated by a voided core. The

155

Masonry structures


156

Fig. 5.13 Dome of S. Maria del Fiori, Florence, 15th century. Filippo Brunelleschi, architect/master builder. This brick-work dome has two skins separated by structural diaphragms. This significant innovation - one of several in this veryimportant structure - allowed a measure of independence to be achieved between the internal and external profiles ofthe structure. [Copyright: RIBA Photo Library]

Masonry structures

Fig. 5.14 Dome of St Paul's Cathedral, London, England, 17th century. Sir Christopher Wren, architect. In this dome theprinciple of twin structural skins, which was first used in Florence by Brunelleschi, was extended to allow completeindependence to be achieved between the interior and exterior profiles of the dome. Here, the outer structural skin - acone of brickwork - is not actually seen but is used to support a timber outer dome of non-structural cladding.[Copyright: RIBA Photo Library] 157


It nevertheless suffered slight meridionalcracking but this could have been due tothermal movement rather than primary load-carrying stress. The walls of the hexagonaldrum which supports the dome are notespecially thick and the lack of outwardmovement of these suggests that no significantoutward thrust occurred at the base of thedome.

The dome of St Peter's in Rome (early seven-teenth century) was designed by MichelangeloBuonarotti and modified by Giacomo dellaPorta. This too was a double-skin masonrydome but it was much flatter than theFlorentine dome. Two iron chains were builtinto it to prevent meridional cracking butcracking which was extensive enough tothreaten the stability of the structure neverthe-less occurred. The dome was stabilised in theearly eighteenth century, under the direction ofGiovanni Poleni, by the insertion of fiveadditional chains, at which time it was discov-ered that one of the original chains hadbroken. Poleni's measures were successful andno further intervention has been required.

The architectural possibilities offered by thedouble-skin dome were extended in the seven-teenth century by the configuration adoptedby Christopher Wren in his design for thedome of St Paul's Cathedral in London (Fig.5.14). Wren was aware of the problems whichhad occurred at St Peter's and adopted a verylightweight double-skin structure of brickworkmasonry. The inner skin formed the innervisible profile of the dome and was more-or-less hemispherical. The outer structural skinwas a cone of brickwork which, with straightsides, rose from the same springing as theinner dome and provided direct support forthe masonry lantern which formed the climaxof the building's exterior. This outer structuralskin was bound with chains but the profile ofthe cone was such that hoop stresses wereminimised. A lightweight secondary structureof timber was built on top of the outer skinand supported the visible external profile ofthe dome. Thus did Wren create one of themost spectacular examples of the architecturalscenic effect - one of the several examples of

ingenious architectural artifice which heemployed at St Paul's.5

The brilliance of Wren's configuration wasthat it allowed a dome to be constructed fromcomponents of relatively light weight, whichcould be supported on an underlying system ofslender columns and arches, but which never-theless provided internal and external profileswhich were compatible with their respectivearchitectural schemes. A similar system wasadopted by Jules Hardouin-Mansart for theDome Church of Les lnvalides in Paris. Thistype of arrangement was to be adopted forvirtually all subsequent large masonry domesin the architecture of the West.

No significant developments in masonrydome construction have in fact occurred sincethe time of Wren and Mansart. The greatdomes of the eighteenth century had similarstructural schemes and, in the nineteenthcentury, masonry was displaced by metal andreinforced concrete for long-span structures.

Post-and-beam structuresThe most common architectural use ofmasonry has always been as the verticalelements in post-and-beam structures in whichtimber or reinforced concrete was used for thehorizontally spanning elements. Buildings ofthis type have been constructed from the verybeginning of the Western architectural trad-ition. The majority are rectangular in plan withflat or pitched roofs. The horizontal structuralelements are carried on sets of loadbearingwalls which are parallel to each other on planbut the plan, for structural stability, must alsoinclude some walls running at right angles tothese. The resulting buildings are thereforemulti-cellular. This generic plan and construc-tional arrangement for loadbearing-wall build-ings has remained unchanged for centuries.The variety of architectural treatments whichhave been possible has nevertheless been verywide and the building type has been used in

5 A similar system of timber exterior supported on amasonry core - in this case a single skin - had beenused at St Mark's in Venice in the eleventh century butWren may not have known of this.158

Masonry structures

the context of most of the stylistic types andvariations of the Western architectural trad-ition.

The classical language of architecture orig-inated in a system of ornamentation for post-and-beam structures. Although the earliestversions may have been constructed in timberthe primary sources of the classical orderswere the masonry temples of Greek antiquity(Fig. 5.15). These were post-and-beam struc-tures with vertical elements of masonry andhorizontal elements of timber or masonry. Thesmallest of these buildings were single-cellstructures but, due to the need to limit spans,larger versions were subdivided internally byparallel sets of walls or columns whichconformed to the generic plan for theloadbearing-wall building described above.

Versions of the generic plan occurred inevery period of Western architecture. A fewnotable examples serve to illustrate this. Thevillas of Andrea Palladio (Fig. 5.16), datingfrom the sixteenth century, show the normalpattern of parallel loadbearing walls as do themany subsequent large houses of Europe andNorth America whose forms were inspired bythose of Palladio and other architects of theItalian Renaissance (Fig. 5.17). The much

Fig. 5.15 The Parthenon, Athens, 5th Century BCE. Thisis the most refined example of post-and-beam construc-tion in the Western architectural tradition. The short spansinvolved are a consequence of the inherent weakness ofmasonry in tension and therefore in bending. [Copyright:RIBA Photo Library]

humbler tenement housing which was built inthe nineteenth century to accommodate thegrowing urban populations of the cities ofindustrial Europe shows a different plan-formbut one which nevertheless conforms to thebasic arrangement of parallel loadbearing walls(Fig. 5.18). The variety of architectural treat-ments of masonry structures in terms ofmassing, ornamentation and fenestration istherefore very great despite the similaritiesexhibited by the plan-forms of loadbearing-wall buildings through the centuries.

In the twentieth century, the vast majority ofmasonry loadbearing-wall buildings haveconformed to the traditional plan-form and theversatility of the medium is such that thevocabulary of Modernism, with, for example,large areas of wall glazing in the external walls,has been accommodated without difficulty.

The development of new constructionalmethods, especially the introduction of steel 159


Fig. 5.16 Plan and elevation of the Villa Emo, Fanzolo,Italy, 1564. Andrea Palladio, architect. The vaultedbasement and timber upper floor of this building arecarried on the parallel cross-walls of a typical loadbearingmasonry plan. [Illustration from Palladio's Quattro Libri dell'Architettura of 1570].

Fig. 5.17 Blenheim Palace, Oxfordshire, 18th century,John Vanbrugh and Nicholas Hawksmoor, architects. Theplan of this very large house shows an arrangement ofparallel walls which is typical of multi-storey loadbearing-wall construction.

160

Fig. 5.18 Plan of a

tenement building,Glasgow, 19th century.This much humbler build-ing also shows the typicalparallel-wall arrangementwhich is required toensure that adequatesupport is provided for allareas of wall and roof.

Masonry structures

161

Fig. 5.19 Brick Country House Project, 1923-24. Ludwig Mies van der Rohe, architect. The plan of this projected build-ing does not conform to the traditional parallel-wall arrangement seen above. Plans like this became possible only aftersystems of horizontal structure which were capable of spanning in more than one direction (reinforced concrete slab,steel space framework) had been developed.


and reinforced concrete, has made possiblethe creation of quite different plan-forms inmasonry, however. The 'brick country houseproject' of 1923-24 by Mies van der Rohe mayserve as an example (Fig. 5.19). Here an irregu-lar pattern of loadbearing walls was adopted toaccommodate the Modernist preoccupationwith free-flowing space. This building wasnever actually constructed but if it had been,the organisation of the horizontal structure ofthe roof would have been necessarily morecomplex than was necessary with the genericparallel-wall type of plan. The choices for thehorizontal structure system would have beeneither a two-way-spanning system (reinforcedconcrete-slab or steel space framework) or onebased on a complex system of primary andsecondary beams.

This project demonstrates that with modernconstruction techniques virtually any plan-formis possible in loadbearing masonry. Non-regular arrangements of walls produce arequirement for more complex systems ofhorizontal structure than those which arebased on the traditional, and more sensible,parallel-wall type of plan.

5.2.2 The technical performance of masonryas a structural material

5.2.2.1 IntroductionThe decision by an architect to use masonryas the structural material for a particularbuilding is based on a knowledge of itsproperties and capabilities in relation tothose of the alternative structural materials.The proposed form of the building, in relationto those for which the material is best suited,will normally have a major bearing on this.Any design decision has consequences andone of these is that the constraints andlimitations of the material must thereafter beaccepted. The adoption of masonry willnormally result in the selection of a loadbear-ing-wall, panel-type structure (see Section1.3.1). The advantages and disadvantages ofmasonry in relation to the other structuralmaterials are summarised below.

5.2.2.2 Advantages

StrengthMasonry obviously has sufficient strength toperform as a structural material. It has moder-ate compressive strength but very low tensilestrength and therefore has a limited capacityto resist bending. It performs best in situationsin which compression predominates, as in thewalls and piers of buildings of moderateheight. It is also suitable for use in compres-sive form-active structures such as arches,domes and vaults.

DurabilityMasonry is a durable material both physicallyand chemically. Except in the most exposedsituations externally, and often even then, itcan be left free of finishing materials. This bothsimplifies the detailing of buildings andprovides a carcass which is virtually mainten-ance free.

Low costDue to the fact that relatively little energy isconsumed in their production and due to theirgood availability at most locations, the basiccost of masonry units and the cost of trans-porting them to particular sites are low.Masonry structures can be built using simple,traditional techniques which do not requirecomplicated plant or machinery. The construc-tion process is therefore also relatively cheap;low cost is one of the principal advantages ofmasonry construction.

AppearanceMost masonry has a pleasing appearancewhich matures rather than deteriorates withage. A considerable variety of colours andsurface textures is usually available and maybe used to create a range of architecturaleffects.

Design flexibilityThe fact that large structures can be assembledfrom small basic units allows complicatedgeometries to be achieved relatively easily withmasonry and the material therefore offersconsiderable scope for imaginative design,subject always to the technical constraints (the162

loadbearing wall, the arch, the vault and thedome). Another consequence of the method ofconstruction is that other materials can beincorporated into masonry so as to augmentits properties. The placing of steel reinforce-ment in the bedding planes to give masonryflexural strength is an example of this.

Fire resistanceMasonry performs well in fire; it is non-combustible and retains its structural proper-ties at high temperatures.

Acoustic performanceMasonry walls form good acoustic barriers.

Thermal performanceMasonry walls provide a reasonable level ofthermal insulation. Their high thermal mass isa further advantage which allows the creationof enclosures with good levels of passiveenvironmental control.

5.2.2.3 Disadvantages

Lack of tensile strengthThe lack of tensile and therefore of flexuralstrength is masonry's principal structuraldisadvantage and has restricted its use, inmodem times, to loadbearing-wall-type struc-tures. The selection of masonry thereforenormally implies that the constraints of thistype of structure be accepted in the planningof the building. The most obvious of these isthe requirement that a plan consisting ofparallel loadbearing walls be adopted and that,in multi-storey buildings, the plan be more-or-less the same at all levels.

The lack of tensile strength also makes diffi-cult the construction of high-strength struc-tural connections between masonry and otherstructural elements. This has tended to restrictthe use of masonry in modern times to multi-cellular buildings in which none of the interiorspaces is large.

WeightIn comparison to timber, masonry is relativelyheavy and while this has advantages (it is

responsible for the good performance ofmasonry as an acoustic barrier) it also hasdisadvantages. In particular, it results in highdead loads being imposed on supportingstructures, such as foundations. It also affectsthe cost of transport to the site.

PorosityMost masonry is porous and while this maynot be a serious disadvantage so far as its useas a structural material is concerned, it doesaffect the structural design. In particular, wherethe material is used for external walls, itcomplicates the detailed design, which mustbe such as to prevent water from penetratingthe building.

5.2.2.4 ConclusionTo sum up, the principal advantage of masonryis that it possesses a good combination ofproperties rather than that it performsoutstandingly well with respect to any particu-lar criterion. Its good appearance and durabil-ity, its reasonably high compressive strength,its low cost, its thermal and acoustic proper-ties and its performance in respect of fire makeit an ideal material for structural walls and thisis, of course, its principal use. In modempractice masonry structures are usually basedon post-and-beam arrangements and theparticular properties of the material affect theforms which are adopted. The general prin-ciples which are followed in the planning ofmasonry buildings, and which take these intoaccount, are outlined in Section 5.3.

The treatment of masonry which is givenhere is confined almost entirely to the use ofbrick and block manufactured components.These are by far the most commonly usedconstituents in the developed world, butmasonry is also constructed in a variety ofother materials such as natural stone andmud. Although these materials are not dealtwith here specifically, the general principles ofloadbearing masonry which are outlined in thischapter apply equally to construction in allbrittle building materials.

163

Masonry structures


5.3 The basic forms of masonrystructures

5.3.1 IntroductionThere are two basic types of masonry: plainmasonry, in which bricks or blocks are simplybedded in mortar to form walls and piers, andreinforced masonry, which contains steelreinforcement bars in addition to the

constituents of plain masonry (Fig. 5.36).Reinforcement bars provide tensile strengthand produce a material whose structuralproperties are similar to those of reinforcedconcrete. Reinforced masonry tends to be usedin conjunction with plain masonry in situationswhere walls are required to have flexuralstrength, such as where they are exposed tolarge out-of-plane loads. Its presence tends tohave only a modifying effect on the forms ofmasonry structures, whose basic geometrieshave been determined by the properties ofplain masonry.

In common with all buildings, loadbearingmasonry structures must be capable of resist-ing two principal types of load: gravitationalload, caused by the weight of the building andits contents, and horizontal load, caused bythe action of the wind or of earthquakes. Anumber of form-determining factors resultfrom this; the influences of each load type areconsidered separately.

5.3.2 Resistance to gravitational load

5.3.2.1 IntroductionThe strategy for resisting gravitational loads inmasonry structures is to use a post-and-beamarrangement in which the masonry forms thevertical elements and the horizontal elementsare made from some other material, usuallytimber or reinforced concrete (Fig. 5.20). Twofactors play a dominant role in determining thearrangement of walls which must be adopted.Firstly, the plan-form must be such as toprovide support for all areas of floor and roofwhich have to be carried; this is the moreimportant factor. Secondly, the walls musthave adequate strength; they must not beover-stressed by the loads which they carry andneither must they become unstable bybuckling. The second set of considerationstends to affect the thickness of walls ratherthan the plan-form of the building, but theneed to provide lateral support to prevent wallbuckling can affect the latter.

5.3.2.2 Provision of support for floors and roofsThe wall arrangements which are adopted to

Fig. 5.20 Typical arrangement of structural elements in aloadbearing masonry building. The timber floor structuresare carried on a parallel arrangement of loadbearing walls.The trussed-rafter roof, which is a more efficient type ofstructure than the rectangular cross-section floor joistsand which carries less load, spans across the entire widthof the building between the exterior walls. Additional non-loadbearing walls (not shown) are required in the across-building direction for stability.164

provide adequate vertical support for all areasof floor and roof depend on the type of floorand roof system which is used and, in particu-lar, on whether one-way- or two-way-spanningsystems are employed. Typical one-way-spanning systems for masonry buildings arethe traditional timber joisted floor and thetraditional rafter or trussed-rafter roof (seeSection 6.7.2); precast concrete slabs and one-way-spanning in situ reinforced concrete slabsare two other common examples (Fig. 5.21).

Examples of two-way-spanning systems are thein situ reinforced concrete floor slab and thesteel space-deck roof structure (Fig. 5.22).

Four basic plan types are employed formasonry buildings: spine-wall, cross-wall,cellular and core-type (Fig. 5.23). Note thatthese are the basic forms in which only theprimary loadbearing walls are shown. Extrastructural walls are usually required for lateralstability, particularly in the spine- and cross-wall forms, and most building plans alsocontain a number of non-structural walls inaddition to the structural walls, because theseare necessary to achieve the required arrange-ment of internal spaces. The plans of realbuildings, of which those shown in Fig. 5.24are typical, are therefore much more compli-cated than the basic forms, which are shown

Masonry structures

Fig. 5.21 Typical one-way-spanning floor structures inloadbearing-wall buildings.(a) Timber joists supported on joist hangers.(b) Precast concrete slabs.(c) Lightweight steel joists.

Fig. 5.22 Two-way-spanning structures.(a) In situ reinforced concrete slab.(b) Space framework. 165

(a)

(b)

(c)

(b)

(a)


Fig. 5.23 The four basic plan types for loadbearing-wallstructures. Only the loadbearing walls are shown,(a) Spine-wall.(b) Cross-wall.(c) Cellular.(d) Core-type.

diagrammatically in Fig. 5.23. All buildingplans must, however, contain some wallswhich conform to one of these basic plan-forms. Typical element dimensions for thistype of building are given in Table 5.1.

Where a one-way-spanning floor system isused it is best supported on a parallel arrange-ment of loadbearing walls; the cross-wall andspine-wall arrangements are therefore particu-larly suitable for this type of construction. Thespacing between the walls depends mainly on

Fig. 5.24 Typical wall arrangements in actual buildings inwhich numerous non-structural partition walls are requiredfor space-planning purposes.(a) Spine-wall arrangement.(b) Cross-wall arrangement.166

Principal loadbearing walls (a)

(b)Principalloadbearingwalls

(c)(d)

(a)

(b)

Table 5.1 Data for elements in low-rise loadbearing masonry structures

Walls single and >2 storey 2 storey

External wall (cavity) Inner leaf 102.5 mm 215 mmOuter leaf 102.5 mm 102.5 mm

Internal wall Loadbearing 102.5 mm 215 mmNon-loadbearing 102.5 mm 102.5 mm

Floors Timber joist Depth range 75 mm to 300 mm (see Table 6.10)Reinforced concrete slab Depth range 150 mm to 440 mm (see Table 4.1)

Roof Rafters and joists Depth range 75 mm to 300 mm (see Table 6.11)Trussed rafter See Table 6.8Truss See Table 6.9

Masonry structures

the span capacity of the floor system. A spanrange of 3 to 5.5 m is normal for timber floorsand 3.5 to 8 m for reinforced concrete. It isadvantageous to keep all principal floor spansin a building (i.e. those of the main spaces)more-or-less the same by spacing theloadbearing walls equal distances apart,because this avoids the need for floor unitswhich are of different depth and thereforesimplifies the detailing of the building. It alsomaintains the stresses in the walls at aconstant level and avoids the need to specifydifferent strengths of masonry. Some spaces,such as corridors, in which spans are markedlydifferent from the principal spans of a buildingare, of course, unavoidable. In buildings whichhave one-way-spanning floors and roofs,openings in walls are spanned by lintels; theseare normally of precast concrete or steel.

The two-way-spanning floor system which ismost commonly used in masonry buildings isthe in situ reinforced concrete two-way-spanning slab, which is sometimes called aflat-slab. Individual slabs are supported by anarrangement of walls which is square on planand this type of floor structure therefore worksbest in conjunction with cellular plan-forms inwhich each cell is approximately square. Thehigh statical indeterminacy6 of this form ofstructure allows it to be used with irregularplan-forms, however, and this is one of its

advantages over one-way systems. The highstatical indeterminacy also means that thestructural material is more efficiently usedthan in the one-way system, particularly ifcontinuity is achieved over a number ofapproximately equal spans. This system istherefore capable of larger spans than are one-way systems and gives smaller slab depths forequivalent spans. The statical indeterminacyalso makes it possible for walls to be designedwith fairly large numbers of openings, withoutthe need for local thickening of the slab or forlintels to be used to bridge the gaps.

It is desirable that the structural walls inmulti-storey masonry buildings should becontinuous throughout the height of the build-ing and so the same basic plan-form must beadopted at every level. Where this is not feas-ible, as in certain building types in whichlarger interior spaces are required at groundfloor level (for example, ground floor shopswith houses above or an hotel with a restau-rant and bar on the ground floor andbedrooms above) a special structural arrange-ment must be used at the level where thefloor-plan changes. The most commonsolution to this problem is to use a reinforcedconcrete or steel frame at ground floor level;the reinforced concrete floor slab which formsthe top of this is then used as a base for theloadbearing-wall structure above (Fig. 5.25).Alternatively the masonry itself can bereinforced by incorporating steel reinforcingbars into the bedding planes. This gives it 1676 See Macdonald, Structure and Architecture, Appendix 3.


similar properties to those of reinforcedconcrete and allows the walls themselves toact as beams spanning across the voids atground level.

Where reinforced concrete floors are usedthey will normally be strong enough to supportany non-loadbearing walls which are providedfor space-planning purposes. There is thereforeno structural requirement for non-loadbearingwalls to be positioned vertically above oneanother and this makes possible minor vari-ations in the plan between levels. Wheretimber floors are used, these will not normallybe capable of supporting the weight of non-loadbearing masonry walls and it is thereforenecessary that the plans at different levelsshould be more-or-less the same. Where minorvariations occur, provision must be made totransfer the weight of non-continuous walls tothe loadbearing walls on steel or reinforcedconcrete beams.

Roof structures are usually capable of longerspans than floor structures, especially iftrusses or trussed rafters are used. In multi-storey structures the strategy which is adoptedfor supporting the roof is therefore frequentlydifferent from that for the floors in the samebuilding. For example, in spine-wall construc-tion the roof is often made to span betweenthe external walls of the building and does notrequire support from the spine-wall (Fig. 5.26).The same system of supporting the roof on theexternal walls is frequently used when thefloors span between cross-walls; the roof spanis then at right angles to the span of the floors.Alternatively a purlin system can be used tosupport the roof as well as the floors on thecross-walls (Fig. 5.27). This relieves the exter-nal walls of any load-carrying function, whichcan be an advantage if large areas of glazingare required.

The thickness of the walls which are requiredin masonry structures depends on the indi-vidual circumstances of the building concernedand, in particular, on the load which each wallmust carry. Any special requirements withregard to bond will also affect this. It is normalpractice to specify cavity construction for theexterior walls, to prevent damp penetration,

Fig. 5.25 Where large open-plan areas are required onthe ground floor of a building in which the upper floors aremulticellular, this can be achieved by the provision of aframe structure of steel or reinforced concrete to act as thebase of a loadbearing-wall building.

Fig. 5.26 Typical cross-section of a small spine-wallbuilding in which the roof structure spans betweenthe external walls.168

Roof truss spans betweenexternal walls

3m

3 m

4.5 m4.5 m

Fig. 5.27 Cross-wall arrangement with roof carried onpurlins.

and for internal walls to be solid. In low-risebuildings of four storeys or less, the externalwalls will normally consist of two leaves whichare one half-brick-length thick or its equivalentin blockwork (i.e. approximately 100 mm)separated by a 60 mm wide cavity to give atotal wall thickness of approximately 260 mm.Normally only one of the leaves is loadbearing(usually the inner leaf), and frequently brick-work is used for the outer leaf and blockworkfor the inner leaf. Interior walls can be onehalf-brick-length thick or its equivalent inblockwork unless, for reasons of appearance, abond which requires headers is used. For high-rise buildings the inner, loadbearing leaf ofexternal walls and the solid internal walls mustusually be one brick-length thick or its equiva-lent in blockwork (i.e. approximately 200 mm).

5.3.2.3 Resistance to bucklingMasonry walls, being compression elements,are susceptible to buckling failure and theplans of masonry buildings have to bearranged so as to avoid the creation of wallswhich are excessively slender. The slendernessratio of a wall is defined as the ratio of itseffective length or its effective height to itswidth, whichever is the smaller (Fig. 5.28) andthe British Standard for Loadbearing Masonry,BS 5628, requires that this should never be

Fig. 5.28 The slenderness of masonry walls, which isdetermined by the distance between locations at which thewall is provided with lateral support, must be limited tocontrol buckling. Lateral support can be provided either bybuttresses or orthogonal walls, in which case the slender-ness is determined by the horizontal distance betweenthese, or by floor structures or horizontal-plane bracinggirders, in which case the slenderness is determined by thevertical distance between these. In either case the measureof slenderness is simply the distance between thelocations of lateral restraint divided by the thickness of thewall. The need to limit slenderness can sometimes affectthe internal planning of buildings by requiring thatbuttresses or bracing walls be located in particularpositions. 169

Masonry structures


greater than 27. In practice, however, it isusually desirable to achieve a value which isconsiderably smaller than this (a maximum ofaround 20).

Given that the range of wall widths which isused in masonry is limited by the availabledimensions of the masonry units, it is theeffective lengths and effective heights of wallswhich must be adjusted by the designer of thebuilding in order to control their slendemess.These are the distances between the horizontaland vertical lines along which the walls arerestrained against lateral movement(movement out of their own plane (Fig. 5.28)).In the case of height the distance betweenlateral restraints will normally be the same asthe storey height of the building because theroof and intermediate floor structures cannormally be regarded as being capable ofproviding effective lateral restraint. The effec-tive length of a wall is the distance betweenlateral support provided either by walls,bonded at right angles to the wall in question,or by substantial piers.

In most buildings, the storey height will besuch that the slenderness ratio, calculatedfrom the effective height, will be within therequired limit. Where this is not the case it willnot normally be possible to adjust the slender-ness ratio by altering the storey height becausemany factors besides those concerned withstructural performance influence the choice ofstorey height. In these cases the requiredslenderness ratio must be achieved by alteringthe plan so as to reduce the effective length ofthe wall, either by inserting buttresses or byaltering the positions of walls which areperpendicular, on plan, to the wall in question.

It is important that the structural elementswhich are considered to provide lateral supportfor a wall are in fact capable of doing so. Twofactors affect this, the strength of the support-ing elements and the way in which they areattached to the wall. Buttressing walls andconcrete floors and roofs will normally haveadequate strength. The ability of timber floorsor roofs which are sheathed with boarding toact as horizontal diaphragms may be in doubt,however, but where this occurs, they can be170

Fig. 5.29 A bracing girder is located at eaves level andacts in conjunction with the gable walls to provide lateralsupport for the top of the side wall.

strengthened either by sheathing in structuralplywood, so as to form a strong horizontal boxbeam, or by incorporating into the floor or roofstructure a system of diagonal bracing (Fig.5.29).

5.3.2.4 Summary of the form-determiningfactors which are derived from the need toresist gravitational loadThe need to resist gravitational load exerts twomajor influences on the plan-forms which areadopted for loadbearing masonry buildings.Firstly, the requirement to provide support forall areas of floor and roof favours the use ofcertain plan geometries; secondly, the need toprevent buckling requires that the slendernessratios of walls should not be unduly large. Tosatisfy these requirements the followinggeneral rules are observed when planningmasonry buildings.

1 The arrangement of supporting walls istailored to the type of roof or floor systemwhich they will support. For one-way-spanning systems parallel sets of walls areused and placed at approximately equaldistances apart. For two-way-spanningsystems a cellular arrangement is adopted in

Extra members are introduced into the ceiling plane toform a horizontal-plane girder which provides lateralsupport for the top of the wall

which the individual cells are approximatelysquare and approximately the same size.

2 The plan geometries of multi-storey build-ings are kept more-or-less the same at everylevel. It is particularly important thatloadbearing walls should be continuousthroughout the entire height of a building.

3 Attention is paid to the slenderness ratiosof walls and columns. Storey heights are nolarger than is necessary and long sections ofplane wall are avoided. A wall thickness isselected which gives a slenderness ratio ofless than 20.

5.3.3 Provision of overall stability and lateralstrength

5.3.3.1 IntroductionAs with most types of structure, the issue ofthe overall stability of a masonry building canbe considered together with that of its abilityto resist horizontal load. If, in addition togravitational load, the building can resisthorizontal load from two orthogonal directionsthen it will also be geometrically stable.

The action of horizontal load on a masonrybuilding has two consequences which affect itsplanning. Firstly, it requires that the building,taken as a whole, should have the ability toresist its effect (Fig. 5.30a). The building must,

Fig. 5.31 Typical wind-bracing system for aloadbearing-wall structure.Wind pressure loads areapplied initially to thesurfaces of the externalwalls and are transmitted bythe horizontal-plane floorsand wind girders to wallswhich are parallel to thewind direction and whichact as diaphragm bracing inthe vertical plane.

171

Masonry structures

Fig. 5.30 The effect of wind loading on a masonrystructure.(a) The structure, taken as a whole, must be capable ofwithstanding the horizontal load.(b) Individual walls must have sufficient bending strengthto resist the out-of-plane pressure loads.

(b)

(a)

Floor acting as horizontal-plane bracing

Wind girder incorporated into plane of ceiling


Fig. 5.32 Individual plane walls are ineffective at resist-ing out-of-plane loads but can resist in-plane loads.

in other words, contain a structural systemwhich is capable of conducting the lateralloads from the points at which they areapplied, to the foundations, where they canultimately be resisted (Fig. 5.31). This propertywill also guarantee stability, as noted above.Secondly, the horizontal load subjects theexternal walls to out-of-plane forces which theyare ill equipped to resist, due to their lowstrength in bending (Fig. 5.30b). Provisionmust therefore be made either to give themthe necessary strength to resist these loads orto provide them with lateral support.

5.3.3.2 The lateral strength of the building,taken as a wholeThe strategy which is adopted to provide amasonry building with lateral strength, andtherefore also stability, is usually based on thevery good performance of walls in resistance toin-plane forces (Fig. 5.32). Buildings are there-fore planned so that wind forces are conductedto the foundations by walls which are more-or-less parallel to their direction.7 This meansthat walls must be provided in two orthogonaldirections on plan. In practice, it is rarely diffi-cult to achieve this because the space-

7 The building must obviously be capable of resistingwind load from any direction. This condition is satisfiedif resistance is provided in two orthogonal directions.172

Fig. 5.33 This cross-wall arrangement required non-loadbearing bracing walls for stability.

planning requirements of masonry buildingswill normally produce a plan which containswalls which run in the two principal directions,even though the primary loadbearing wallsmay be parallel to each other in only one ofthese. The plans in Fig. 5.24, for example, aresatisfactory in this respect. Figure 5.33a showsan arrangement which is not satisfactorybecause there is no wall which can resist windload as an in-plane force in one of the build-ing's principal directions. Additional wallsmust therefore be added to render the buildingstable (Fig. 5.33b).

The horizontal parts of masonry structuresplay an important part in the resistance ofhorizontal load and they can act in two capaci-ties. For a particular direction of horizontalload they provide a direct tensile or compres-sive link between bracing walls and other wallswhich require to be stabilised. They can alsoact as horizontal-plane bracing (Fig. 5.31).Normally the horizontal structure in a masonry

(b)

(a)

building will be required to carry out both ofthese functions and it must be designedaccordingly, and suitably attached to all of thewalls. Reinforced concrete floors provide thebest horizontal structural elements for thesepurposes. Where timber elements are used,their ability to act collectively as a horizontaldiaphragm must always be considered. Incases where they cannot do so, a horizontalbracing girder must be incorporated into theplane of a roof or floor to give it the ability toact as horizontal-plane bracing (Fig. 5.31).

In low-rise buildings of up to four storeysthe magnitudes of the lateral loads due towind are not high and the resistance of theseis rarely a critical factor in the design. Thebuilding will be satisfactory so long as itcontains some walls in two orthogonal direc-tions and any single-plane walls which arepresent are connected, by an adequatehorizontal system, to other wall groups whichare capable of bracing them. The exact disposi-tion of the loadbearing and bracing walls is notcritical although a symmetrical arrangement ofbracing walls will produce the most satisfac-tory structure for the resistance of lateral load.

In tall multi-storey buildings of five storeysor more, the resistance of wind load can,however, be a critical consideration. Theadditional factors which affect their planningare reviewed in Section 5.3.4.

5.3.3.3 Resistance of individual walls to out-of-plane forcesWind pressure is an out-of-plane load on theexternal walls of buildings. It producesbending-type internal forces in walls and canresult in the generation of levels of tensilestress which the masonry cannot resist (Fig.5.30b). The task for the designer is to containthe tensile stress which develops. This can bedone by adopting a structural form which eitherlimits bending action or which neutralises,using axial compressive stress, any tensilebending stress which does develop. Bendingaction is countered by minimising span. Theexternal walls of masonry buildings span eithervertically, between floors, or horizontally,between buttressing walls (Fig. 5.34). If the

Masonry structures

Fig. 5.34 In response to out-of-plane wind loads, theexternal walls of loadbearing wall buildings span eithervertically between horizontal-plane bracing elements orhorizontally between buttressing walls.

strategy being adopted to limit bending actionis that of providing buttressing walls, thenthese have to be spaced closely. This obviouslyaffects the planning of the building.

If the strategy for limiting the bending stresscaused by out-of-plane wind pressure on exter-nal walls is one of neutralising the tensilestress with compressive stress, the walls mustbe arranged to span vertically between floorsin response to wind loading. This ensures thatany tensile stress acts on the horizontal cross-sectional planes where it can be neutralised bythe compressive stress produced by the gravi-tational loads. The magnitude of the tensilebending stress can be minimised by ensuring

8 See Macdonald, Structure and Architecture Appendix 2 foran explanation of this term. 173

(a)

(b)

Fig. 5.35 Plans of walls showing various arrangements for incorporating vertical reinforcing bars. These give the wallsthe ability to span vertically between floor and roof structures in response to out-of-plane loads.

that the second moment of area8 of the wallcross-section is maintained high, by makingthe wall reasonably thick.

Masonry, as mentioned above, can be giventhe bending strength which is necessary to resisthigh out-of-plane load by reinforcing it withsteel bars. Its behaviour is then similar to that ofreinforced concrete. The technique is usuallyused to augment the strength of walls whichmust carry very large bending loads, such asretaining walls or the external walls of highbuildings. Vertical bars provide the bendingstrength to span vertically in response to out-of-plane forces. These must either be incorporatedinto the voids of special bricks or blocks, oraccommodated in vertical channels provided byspecial bonds (Fig. 5.35). Walls can also be giventhe strength to span vertically by incorporatingreinforced concrete or steel columns into them.Horizontal bars in the bedding joints of masonryor within special blocks (Fig. 5.36) give walls thestrength to span horizontally. These techniquesdo not affect the overall form which must beadopted for a building, except insofar as theypermit large areas of plane wall to be adopted.

5.3.3.4 Summary of form determining factorswhich result from the need to resist horizontalloadThe principal features which are incorporatedinto the geometries of masonry buildings inorder that they can resist horizontal load effec-tively are as follows.

Fig. 5.36 Incorporation of reinforcement into thebedding planes gives masonry the ability to span horizon-tally between buttressing walls in response to out-of-planeloads.

1 The plan must contain walls orientated intwo orthogonal directions.

2 The horizontal parts of the structure mustprovide an effective structural link betweenthe walls. They must therefore haveadequate strength and must be adequatelyfixed to the walls.

3 The walls are disposed in as symmetrical anarrangement as possible on plan. This is notessential in low-rise buildings, where windloads are small, but becomes increasinglyimportant as building heights increase.174


Reinforced, tiedand groutedcavity wall

Reinforcedpocket

Quetta bond Vertical slotwall wall

Pocket wall Cavity wall

4 The bending action which results from theapplication of out-of-plane forces to theexternal walls is minimised by providinglateral support at frequent intervals andthus lowering the span. Attention is alsogiven to the selection of the cross-sectionaldimensions of external walls to ensure thattensile bending stresses in response to out-of-plane loads do not become excessive.

5.3.4 Tall multi-storey buildingsTall multi-storey buildings may be regarded asbehaving like 'bundles' of vertical cantileverswhen subjected to wind loading (Fig. 5.37).These perform best if the walls are arranged

Fig. 5.37 Tall multi-storey loadbearing-wall structuresbehave as 'bundles' of cantilevers in response to windloading. The structural action is most effective whenindividual wall groups are as large as possible.

into groupings which are as large as possible,and which are bonded together at their verticaljunctions (Fig. 5.38). Symmetry in the dis-position of walls is also desirable, to preventtwisting of the building under the action ofasymmetrically applied horizontal load.

5.3.5 Tall single-storey buildingsTall single-storey buildings are buildings whichconsist predominantly of a single large interiorspace (sports hall, swimming pool). If these areconstructed in loadbearing masonry the provi-sion of walls which can withstand gravitationalload is rarely a problem. A post-and-beam formcan be adopted in which masonry walls support

Fig. 5.38 Typical planforms for tallmulti-storey loadbearing-wall struc-tures. Each contains several very largewall groupings which have complexplans. These will make the largestcontribution to the resistance of windloading.

175

Masonry structures

(a) (b)

Fig. 5.39 Tall single-storey building. A roof structure oftimber or steel girders is supported on high masonry walls.Such buildings are the modern equivalent of the vaultedhall (see Section 5.2.1).

a roof structure of steel or timber-truss primaryelements (Fig. 5.39). Typical depths for varioustypes of roof structure are given in Table 5.2.

The strategy which is normally adopted toresist horizontal load is fairly standard.Vertical-plane bracing is provided by the wallswhich are parallel to the direction of the lateralload and these are linked to all other parts ofthe building through horizontal-plane bracing(Fig. 5.40). In the case of these single-storeybuildings the horizontal-plane bracing mustnormally be provided by incorporating a wind-girder into the plane of the roof.

External walls which are exposed to out-of-plane wind pressure span vertically betweenthe wind girder and the foundations and, asthe walls are high, the resulting bending actioncan be large (Fig. 5.40). Because the compres-sive stress due to gravitational load in the roof

will be only moderate in a single-storey build-ing, it is necessary that the tensile bendingstress should not be high. The latter iscontrolled by the expedient of increasing thesecond moment of area of the wall cross-section, and this is a factor which can have amajor effect on the planning of the building.

If the wall is relatively low (say around 3 m),normal cavity construction, with perhaps athicker than normal loadbearing leaf or withpier stiffening of the loadbearing leaf (Fig.5.41a), is usually sufficient to prevent thebending stress from being excessive, but if it ishigher, then relatively sophisticated systemsmust be used. In the range 4 m to 5 m one ofthe most economical forms of construction isthe fin-wall (Fig. 5.41b) in which one of theleaves is stiffened with very large piers. Thewall then behaves as a series of T-beams inresponse to lateral load. Fins are normallyone-half to two brick-lengths wide and four toeight brick-lengths deep and are spaced 3 m to5 m apart. For wall heights in the range 5 m to10 m a form of construction called diaphragmwalling is necessary (Fig. 5.41c). This involvesseparating the leaves of the wall by a gapwhich is much larger than is used in normalcavity construction, that is by one or two brick-lengths, and providing a shear connectionbetween the leaves in the form of diaphragmsat closely spaced intervals (1 m to 1.5 m). Thewall then acts as a box-beam in response to

176


Table 5.2 Spans and dimensions for roofelements in tall single-storey masonry structures

Beam Beam depth (mm)

Span

(m) Universal Castellated Parallel Laminated

beam beam chord truss timber

(hot-rolled beam

elements)

10 450 500 1000 70015 600 700 1250 90020 700 800 1500 120025 800 900 175030 900 1200 200035 - 1300 225040 - 1300 2500

Masonry structures

Fig. 5.40 Action of tall single-storey building in responseto wind loading. The high walls span vertically between thefoundation and the wind girder at roof level, which trans-mits the load to the end walls. The side walls here act aspropped cantilevers and can be subjected to high levels ofbending moment.

bending loads. Typical dimensions for all ofthese types of wall are given in Table 5.3.

5.3.6 Provision for accidental damageThere are certain types of very high intensityload whose occurrence is so rare that it is noteconomic to make walls strong enough toresist them. Examples are the collision of aheavy vehicle with a building or an explosionwhich results from the ignition of leaking gas.Because buildings are not sufficiently strong toresist these, damage or partial collapse results

Fig. 5.41 Various strategies can be adopted to give thewalls of tall single-storey buildings sufficient bendingstrength to resist out-of-plane load due to wind.(a) Buttressed cavity wall (suitable for heights up to 3 m).(b) Fin wall (suitable for heights up to 5 m).(c) Diaphragm wall (suitable for heights up to 10 m).

if they occur, but the risk of this happening istaken to be sufficiently low to be acceptable. Itis necessary, however, that the stability of thewhole building should not be endangered if asingle structural element, such as a wall or anarea of floor, should collapse due to the occur-rence of a freak load. In other words, a progres-sive collapse should not result from anincident in which a part of the building only isdirectly involved.

The UK Building Regulations require thatstructures should be designed so that progres- 177

(a)

(b) (c)


sive collapse cannot occur and specify limitswithin which the collapse which is caused by asingle incident must be contained. Theseregulations apply to all types of building butare particularly relevant to the overall planningof masonry structures which are more vulner-able to this form of collapse than other types.As with all of the Building Regulations,however, the requirement is no more difficultto meet than that which a responsible designerwould in any case wish to impose.

In practice the satisfaction of the require-ments in respect of progressive collapse isnot difficult with a masonry structure. Bothcommon sense and the Building Regulationsdictate that, in these exceptional circum-stances, the concern should be with thepreservation of life rather than with thepreservation of the building itself. A consid-erable amount of damage to the building istherefore acceptable, following an incident,so long as the occupants are neither injurednor trapped as a consequence of it. Thefactor of safety against the total collapse of adamaged building can be very low, and afigure of 1.05 is specified in the BuildingRegulations.

Most masonry structures are capable of'bridging' a gap in the structure with thisdegree of safety if certain rudimentary precau-tions are taken during the design. Theseusually amount to the incorporation of certaindetails into the wall-to-wall and wall-to-floorjunctions so as to improve the integrity of thestructure, and the insertion of a certainamount of reinforcement into key walls andpiers. It is not usually necessary for any majoralteration to be made for this purpose to thestructural form which is required to resist thenormal gravitational and horizontal loads. Forthis reason the specific design requirements inrespect of the prevention of progressivecollapse are not dealt with here.

Fig. 5.42 Plan dimensions of walls.(a) Plain cavity wall.(b) Fin wall.(c) Diaphragm wall.

178

Table 5.3 Wall dimensions for tall single-storeymasonry structures

Wall Plain

height cavity Fin wall Diaphragm wall

wall

(m) ti(mm) tp (mm) T (mm) D (mm) S (mm)

3 215 - - - -4 327.5 665 440 - -5 440 890 440 - -6 - 1012.5 440 557.5 13607 - 1340 440 557.5 11358 - 1340 440 557.5 9109 - - - 665 113510 - - - 665 910

Fig. Fig. 5.42b Fig. 5.42c5.42a

Chapter 6

Timber structures

6.1 Introduction

Timber is a structural material with a usefulcombination of physical properties. Althoughits strength is not high (typical design stressvalues are in the range 5 to 20 N/mm2

compared to equivalent values for steel of 150to 250 N/mm2), timber is, like steel, more-or-less equally strong in tension and compres-sion. It can therefore withstand bending andcan be used to make every kind of structuralelement. Due to the origin and nature of thematerial, timber is available normally in theform of slender, linear elements and this

favours its use in framework arrangements.The ratio of strength to weight of timber is

high, and is comparable with that of steel, withthe result that, although, for a given size ofcross-section, timber elements are not sostrong as those of steel, they are much lighter.Timber is therefore a lightweight material,capable of providing structural elements whichare of low dead weight, but which are never-theless reasonably strong and tough.

One of the problems associated with thestructural use of timber is that due to itsarboreal origins individual elements arerelatively small. The construction of large

Fig. 6.1 Timberhouses with skeleton-frame structures inwhich the principalelements are A-framesof plywood box-section. A characteris-tic of timber skeletonframeworks is that theprimary structuralelements are normallyfairly large and mustbe incorporated intothe aesthetic schemeof the building [Photo:Finnish Birch PlywoodAssociation].

179


structures therefore involves the joiningtogether of many separate components. Jointsand connections are therefore an importantaspect of the technology of structural timber.

6.2 Timber and architecture

6.2.1 IntroductionOf the four principal structural materials,timber is one which is not directly associatedwith a major architectural style or movementin the Western architectural tradition, althoughsignificant timber building traditions (forexample, the 'stick' and 'shingle' styles ofNorth America) have occurred. Other architec-tural traditions, such as those of China orJapan, have, however, produced significanttimber styles.

Although no major Western architecturalstyle is associated with timber, the contribu-tion of the material to the development ofWestern architecture has nevertheless beenconsiderable. Its principal structural use hasbeen as the horizontally-spanning elements inpost-and-beam structures in which the verticalelements were of masonry. As in the case ofstructures constructed entirely of masonry, itwill be necessary here to consider historicexamples in order to review the full contribu-tion which timber has made to architecturalexpression.

Before considering the architectural qualitiesof timber it is necessary to say something ofthe range of structural arrangements for whichit is suitable. Although there is considerablevariation in the forms and layouts which havebeen created, almost all timber floor and roofstructures conform to a similar basic layout.The roof or floor covering of boards, slates,tiles, etc. is normally supported on a series ofclosely spaced parallel elements. The timberbeam floor (Fig. 6.2) is perhaps the simplestexample of this. The modern trussed-rafter roofis a more sophisticated version of it (Fig. 6.27).The common characteristic of these arrange-ments is that the individual elements, be theybeams or trussed rafters, carry relatively smallareas of floor or roof and are therefore lightlyloaded. They span one way between parallelsupports which may be loadbearing walls ormay be primary beams in a skeleton frame.

The spacing of the structural elements inthis type of arrangement is close and is deter-mined principally by the span capability of thematerial which the timber elements support(floor or roof boarding) but also by the need tomaintain the load carried by each timberelement at a suitably low level.

Where large spans are involved and the sizesof the main elements are therefore large itbecomes uneconomic for the spacing of theseto be close. For large spans, therefore, a hierar-chical arrangement of primary, secondary andeven tertiary elements provides a bettersolution. An example of this is the purlin roof(Figs 6.44 and 6.45). In the arrangement shownin Fig. 6.44 the primary elements, which span

Fig. 6.2 In this typical timber floor structure theelements are at close spacings and each carries a relativelysmall amount of load.180

Wall support(masonry ortimber framing)

Beam support(timber or steel)

Timber structures

Fig. 6.3 Primary-secondarysystem. The primaryelements here are laminatedtimber portal frames. Thesecarry purlins which in turnsupport the roof and wallcladding.

the full width of the space, are trusses. Purlins(secondary elements) span between these andsupport rafters, which are tertiary elements. Acharacteristic of the hierarchical arrangementof elements is that the primary elements carrythe greatest amount of load and must normallytherefore be relatively complex built-uparrangements such as triangulated trusses. The

secondary and tertiary elements are closelyspaced and carry relatively small amounts ofload across small spans, as in the case of thepurlins in Fig. 6.3 and the rafters in Fig. 6.44.

The elements of which timber structures arecomposed may take a number of forms (Fig.6.4). The simplest of these is the solid beam.This carries load by pure bending action and is 181


the least efficient type of structural element.1

The maximum span which is possible is deter-mined by the largest size of cross-sectionwhich is available. In the present day themaximum span possible with timber floorjoists is around 6 m unless larger than normalsizes of cross-section, which are now difficultto obtain and therefore expensive, are used. InGreek and Roman antiquity, when larger sizeswere apparently readily available, beam spansof up to 12 m were used.

The rafter arrangement (Fig. 6.42) is moreefficient than that of the beam because theelements carry a combination of axial internalforce and bending. This, together with the factthat the individual elements do not cross theentire span, allows a greater span to beachieved but has the disadvantage of imposinghorizontal thrust at the points of support. Therafter system therefore carries with it the sameproblem as the masonry vault, which is thegeneration of horizontal forces at the tops ofsupporting walls. The magnitudes of these canbe reduced if a tie is introduced (Fig. 6.43) andif the tie is placed at the level of the supportsthe lateral thrust is eliminated. The tie neednot be a single piece of timber but can bejointed. It can also be supported from the apexof the rafters by a second, vertical tie (Fig. 6.43).

The tied rafter with the tie at support level,and secondary vertical tie, is in fact a simpleversion of a fully triangulated truss or trussedrafter (Figs 6.27 and 6.28), which is moreefficient than any of the various rafter systemsbecause the majority of the constituent sub-elements carry axial internal forces rather thanbending.

The fully triangulated truss is a type ofelement which is frequently used for long spansin the present day. The advantages of triangula-tion seem not to have been fully understooduntil quite late in the history of Western archi-tecture, however, and it was not until thenineteenth century that widespread use wasmade of fully triangulated, highly efficient trusssystems, which exert no horizontal loads on

1 See Macdonald, Structure and Architecture, Chapter 4.

Fig. 6.4 A selection of timber element types.(a) Sawn-timber joist - the simplest type of element whichis produced in a wide range of sizes.(b) Plyweb beam. The web of this built-up-beam is ofplywood and the flange elements of sawn timber.Rectangular box sections are another common type ofplyweb beam.(c) Laminated timber beam. Large cross-sections can bebuilt up by the laminating process which allows curvedarch and portal frame elements to be produced as well asstraight or tapered beams.182

(c)

(b)

(a)

supporting walls. Most earlier truss systemswere hybrid arrangements in which some trian-gulation was used but which also relied onrigid joints or continuity of elements throughjoints for satisfactory structural performance(see Fig. 6.29). Both of these expedients inducebending into the sub-elements of the truss andreduce their efficiency. The consequence wasthat larger cross-sections had to be used thanwould have been necessary had full triangula-tion been adopted.

6.2.2 The architectural significance of theuse of timber as the horizontal elements inlarge-scale loadbearing-wall masonrystructuresTimber beams were used in the roofs of thetemples of Greek antiquity - probably in theform of simple beam systems (Fig. 6.5). It seemsunlikely that their presence affected the externalappearance of the buildings, which is the chieflegacy of Greek builders to the vocabulary ofWestern architecture. The use of timber wouldhave allowed larger interior spaces to be createdthan would have been possible had the build-ings been constructed entirely in stone but, asthe Greeks seem not to have used any form ofbuilt-up-beam or truss, the maximum span(around 12 m) was determined by the largestsize of timber which was available. There issome evidence that the Greeks were aware ofthe principle of trussing but that, as in the caseof the arch, they chose not to use it.2

The Romans, who were more adventuroustechnically than the Greeks, did adopt theprinciple of trussing. A basic triangulatedarrangement was described by Vitruvius and healso provided a description of a timber-roofedbasilica with a span of 60 Roman feet (approx.20 m). A further example from the first centuryBCE is that of the basilica at Pompeii. Theconjectured reconstruction of this building byLange3 (Fig. 6.6) shows a truss-and-purlinarrangement.

2 See Mark ( ed ) , Architectural Technology up to the Scientific

Revolution, Cambridge, MA, 1993, Chapter 5.3 Lange, Basilica at Pompeii, Leipzig, 1885.

Fig. 6.6 Cross-section of Basilica at Pompeii, 1st centuryBCE. A form of semi-trussed timber structure is used hereto span the wide central space (after Lange). 183

Fig. 6.5 Cross-section of the Parthenon, Athens, 5thcentury BCE. In this conjectured reconstruction of theParthenon the roof structure consists of timber beamelements (shown shaded). The spans are kept short by thesubdivision of the cross-section by walls and rows ofcolumns. Note the very deep beam which is required tocross the central part of the interior, between the rows ofcolumns which flank the statue, and which carries theposts which support the ridge beam (after Coulton).

Examples of later Roman timber-roofedbasilicas are that of the original St Peter's inRome (Fig. 6.7) and the church of St Paul'sOutside the Walls, also in Rome (Figs 6.8 and

Timber structures


Fig. 6.7 St Peter's Basilica,Rome, 4th century CE. A fullytriangulated timber structure,which exerts no lateral force onthe walls, spans the central partof the building. The use of thistype of timber structure hasallowed the creation of a largeinterior space without the needfor the elaborately buttressedsupporting walls which wouldhave been required had amasonry vault been used (afterMark).

Fig. 6.8 Left. St Paul's Outside the Walls, Rome, 4thcentury CE. This perspective drawing illustrates well theadvantage of using a triangulated timber roof structurerather than a vault. The thinness of the walls is well shown(after Piranesi).

Fig. 6.9 Below. St Paul's Outside the Walls, Rome, 4thcentury CE. Cross-section (after Rondelet).

184

6.9). Both of these buildings are of the fourthcentury CE. They each consisted of a highcentral nave flanked by lower aisles. The naves,which were each approximately 24 m wide,were roofed in timber by truss-and-purlinsystems and the flanking aisles by lean-tostructures, also of timber. The use of tiedtrusses over the naves eliminated side thrustat wallhead level and enabled the builders toadopt very slender walls. The buildings were oflight and delicate appearance, as is well illus-trated by the seventeenth-century drawing ofSt Paul's Outside the Walls by Piranesi (Fig.6.8).

The importance of the timber-roofed basilicawas that it was a building with a relativelylarge interior which did not have the massivebuttressing walls which were required tosupport a masonry vault (the only other way ofachieving a large span at the time). It was anearly example of a building type which wouldmake a very significant contribution to thedevelopment of Western architecture - thelarge building with large interior spaces andthin walls. It was the development of thetechnology of the timber truss, from theRoman period onward, which made this possi-ble.

Examples of this type of building can befound in all subsequent periods of westernarchitecture. The palaces and large houses ofthe Italian Renaissance, the country housesand public buildings of the classical period inBritain, Northern Europe and America (Fig.6.10) and churches from all subsequentperiods have structural armatures of this type.In most of these buildings the timber struc-tures which spanned the large interior spaceswere entirely hidden from view and made noobvious contribution to the architecture. Theforms of the buildings would have been impos-sible to achieve, however, without the technol-ogy of the timber truss.

In the nineteenth century timber gave way toiron and then to steel as the principal materialfrom which large trusses for long-span roofswere constructed. The tradition of using largetimber trusses did not die out, however.Twentieth-century examples have frequently

Fig. 6.10 Banqueting House, London, 1619-22. InigoJones, architect. The roof trusses which span the width ofthe Banqueting House in London (17m approx.) aretypical of the structural arrangements which were used tocreate large interior spaces in the buildings of theEuropean Renaissance.

been exposed in the interior of buildings andused as part of the architectural language.

6.2.3 Timber loadbearing-wall structuresA timber loadbearing-wall structure is a post-and-beam arrangement in which both the verti-cal and the horizontal structural elements areof timber. The horizontal structures are similarto those which are used in conjunction withmasonry and are beamed floors or rafter ortrussed roofs. In timber loadbearing-wall struc-tures the vertical elements are also of timberand consist of loadbearing walls formed byclosely spaced timber sub-elements tiedhorizontally at regular intervals by furthertimber sub-elements. The building type hasexisted from the beginning of the Westernarchitectural tradition and has taken twodistinct forms - the half-timbered building andthe timber wallframe building.

The origins of the half-timbered building liein prehistory but the building type did notbecome significant architecturally until theeleventh century. It remained so until theseventeenth century. In the half-timberedbuilding the walls consist of large timberelements, frequently roughly hewn and withcross-sections which were more-or-less square,tied at each storey level by horizontal elementsof similar dimensions (Fig. 6.11). A third set ofelements, inclined to the vertical at between45° and 60°, was provided to brace thestructure. The joints between the elements(Fig. 6.12) were of the type which required the 185

Timber structures


Fig. 6.11 The medieval half-timbered house was basically aloadbearing-wall structure. Thewalls consisted of verticalelements at relatively closespacing tied by horizontalelements at each storey level.The structure is stabilised by themassive masonry chimney and bythe inclined, often curved, cornerelements. The sizes of the cross-sections of the main elementwere large.

cutting away of material (mortice-and-tenon,halved, dovetail) and the resulting inefficiencywas one of the reasons why timbers of largecross-section were used. The loadbearing wallswere arranged parallel to each other andconnected by horizontal beamed floors andrafter or trussed-roof structures. The buildingwas completed by non-structural elements.The wall infilling was entirely non-structuraland was made from a variety of materials.Brickwork or plaster on a woven mesh of thintimbers were commonly used. Normally, thestructural timber was exposed on the exteriorof the buildings, but occasionally the wallswere rendered with plaster or lime.

The European tradition of the half-timberedbuilding occurred north of the Alps and was

associated with the rise of the mercantilemiddle class which occurred after theReformation. The building type was usedextensively for merchants' houses and publicbuildings in the trading towns of northernEurope.

The form of the buildings was predominantlyrectilinear and the plan was that of theloadbearing-wall structure - parallel arrange-ments of walls supporting one-way-spanningfloors and pitched roofs. Often, the buildingswere of considerable height (five or six storeys)and the patterns formed by the exposed struc-tural timbers on the external walls werefrequently used as ornamentation andregularised and formalised for aestheticpurposes (Fig. 6.13).186

The construction of half-timbered buildingsdeclined in the seventeenth century foreconomic reasons. The very large sections oftimber which were required became increas-ingly scarce, and therefore expensive, and thelabour-intensive construction process, with itsrequirement for complicated joints, becameuneconomical in the context of a growingmaterialistic, industrialised society in whichlabour costs were destined inexorably to rise.Thus did a building system which had consid-erable architectural potential pass out ofexistence. Its revival, in a modern or post-modern context, should not be ruled out, inview of its potential as a constructional systemfor a sustainable architecture.

The same economic forces which caused thedemise of the half-timbered building ledeventually to the development of a new type oftimber loadbearing-wall structure, the timberwallframe. The wallframe type of building,which was developed in the nineteenth centuryin North America, can be seen as an inevitableconsequence of the increasing use of manufac-tured products in building (Fig. 6.14).

The wallframe building consisted of a seriesof timber loadbearing walls carrying timberfloor and roof structures in a conventionalrectilinear layout not dissimilar to that of thehalf-timbered building. The loadbearing wallswere formed by closely spaced timberelements tied horizontally at the levels of theroof and floors. In the wallframe building,however, the timber elements were produced Fig. 6.12 Joints in half-timbered buildings were of the

traditional carpentry type which involved the removal oflarge areas of cross-section. This is one of the reasons whysuch large overall cross-sections were required.

Fig. 6.13 The exposed structure contributesto the aesthetic appeal of this modestexample of a half-timbered building.

187

Timber structures


Fig. 6.14 The timber wallframeis a loadbearing-wall structurebuilt up from sawn-timberedelements. It is the industrialisedequivalent of the half-timberbuilding. The fastening elementsare nails which allow the joints tobe made without removing signifi-cant parts of the cross-sections.The sizes of the cross-sections aretherefore considerably smallerthan those used in half-timberedbuildings.

in mechanical sawmills. They were small andof uniform, rectangular cross-section and wereheld together by nails - another mass-produced product - rather than beingconnected by traditional carpentry joints.Lateral stability was provided by the boardingwhich formed the surfaces of the walls, floorsand roof. The structural frameworks could beerected quickly and therefore cheaply byrelatively unskilled labour. There were nocomplicated joints. The building type is still inuse in the present day.

6.2.4 Timber skeleton-frame structuresIn skeleton-frame structures the volume ofstructural material which is present is consider-ably smaller than in loadbearing-wall structures

and the structural loads are concentrated intoslender beams and columns. Stress levels aretherefore high, and, because the strength oftimber is only moderate compared to materialsuch as steel, it is frequently considered unsuit-able for skeleton frames. There is, however, atradition of timber skeleton-frame architecture.This has received fresh impetus in the twenti-eth century by the development of laminatedtimber and other types of built-up beam whichhave allowed the creation of beams andcolumns with larger cross-sections than arepossible with sawn timber. Skeleton frames intimber are characterised by elements with fairlylarge cross-section in relation to their length.The structure is normally allowed to dominatethe architectural language (Fig. 6.15).188

Timber structures

Fig. 6.15 Ragnitz ChurchHall, Graz, Austria,Szyszkowitz and Kowalski,architects. The roof of thisbuilding is supported by atimber skeleton frameworkwhich has a significantarchitectural presence.[Photo: E. & F. McLachlan]

6.2.5 Twentieth-century developmentsAll three structure types which have beendescribed in the preceding sections are incurrent use. Large timber trusses are used asthe horizontal structures in wide-span build-ings, timber wallframe houses account for asignificant proportion of domestic buildingsand the distinctive skeleton frame in timber is

still part of the architectural vocabulary. Thestructures of the present day are significantlydifferent from those of previous centuries dueto recent developments in the technology oftimber. In particular, modern timber structuresare lighter and they are also more preciselycrafted. The two very significant developmentswhich have occurred in the twentieth century 189


are the practice of stress grading and theevolution of greatly improved jointing techniques.

The stress grading system, which isdescribed briefly in Section 6.5, overcomes oneof the fundamental problems of timber whichis the variability of the material. It allows thestrength of timber elements to be specifiedwithin fairly precise limits. This in turn allowsfactors of safety which are used in the designcalculations to be low. The material is there-fore used efficiently.

As was seen in Section 6.2.3, traditionalcarpentry joints involve the cutting away of aconsiderable amount of material - sometimesas much as half the cross-section of a struc-tural element. The joint was therefore theweakest part of the element and the resultingefficiency was very low. In modern timberjoints the quantity of material which isremoved is minimised. In addition, fasteningtechniques, based on improved glues andmechanical components, have been developedwhich minimise the concentration of stresswhich occurs at joints (see Section 6.6.6).

One of the main consequences of the avail-ability of these improved jointing techniqueshas been the development of a wide variety ofbuilt-up elements such as laminated timberbeams, portal frames and arches, ply-webbeams and trussed rafters of various configur-ations. The availability of efficient types ofjoint has also allowed timber to compete effec-tively with steel for triangulated plane andspace frameworks for roof structures.

6.3 The material, its properties andcharacteristics

6.3.1 IntroductionThe technical advantages and disadvantages oftimber in relation to other structural materialsare reviewed in this section.

6.3.2 Advantages

StrengthTimber possesses tensile, compressive andflexural strength and is therefore suitable forall types of structural element.

LightnessTimber is a lightweight material with a highratio of strength to weight. It thereforeproduces lightweight structures with com-ponents which can be easily transported andhandled on site.

TractabilityTimber can be easily cut and shaped withsimple tools and the erection of timber struc-tures is therefore straightforward. Othercomponents can be easily attached to timberwith simple fasteners, such as nails or screws,and this simplifies the detailing of timberbuildings.

Performance in fireTimber is a combustible material but the rateat which it is consumed in a fire is relativelylow and it does not lose its structural proper-ties when it is subjected to high temperatures.It therefore continues to function in a fire untilthe cross-section of elements become reducedto the point at which excessive stress occurs(other materials, particularly metals, lose theirstrength at relatively low temperatures). Theperformance of timber structures in fire istherefore good.

DurabilityThe constituents of timber are relatively stablechemically and the material does not sufferchemical degradation in environmental condi-tions (such as high humidity levels) whichmight prove detrimental to metals. It is,however, susceptible to insect infestation andfungal attack (see Section 6.4.4.4).

AppearanceTimber is a material which has a pleasingappearance which matures rather than deteri-orates with age. It can therefore serve in thecombined role of a structural material and afinishing material.

6.3.3 Disadvantages

Lack of strengthAlthough the strength-to-weight ratio of timberis high, its actual strength is low compared toother structural materials such as steel and190

reinforced concrete. This restricts the size ofspan which can be achieved and the number ofstoreys which can be constructed in an all-timber structure.

Jointing difficultyAlthough the material is tractable it is difficultto make joints in timber which have a goodstructural performance. Joints which are madewith mechanical fasteners, such as nails,screws and bolts, suffer from the problems ofstress concentration. They also tend to workloose due to the phenomenon of moisturemovement (see Section 6.4.3). Although verysignificant advances have been made in recentdecades in the development of new types ofmechanical fastener, these problems have notbeen entirely overcome.

The development, in the twentieth century,of reliable glues has made possible the virtualelimination of both stress concentration andthe development of looseness at joints. Toachieve a satisfactory connection with glue,however, it is necessary that the contactsurfaces should be very carefully prepared andalso that the glue should be properly cured.4

Both of these are difficult to achieve on siteand this requires that gluing be carried outunder factory conditions only. The size ofcomponent which can be assembled withglued joints is therefore restricted by consider-ations of transport.

Component sizeThe sizes of individual timber planks or boardsare obviously determined by the size of avail-able trees. A consequence of the depletion ofthe world's resources of timber is that lengthsgreater than around 6 m and cross-sectionaldimensions in excess of 300 mm by 100 mmare difficult to obtain in most of the specieswhich are used for structural purposes. This,together with the jointing problems describedabove and the low relative strength, restrictsthe overall size of the structures which can beconstructed in timber in the present day.

4 See Gordon, The New Science of Strong Materials,Harmondsworth, 1968.

Susceptibility to rot and decayTimber components are susceptible to variouskinds of infestation, notably by fungi andinsects. The likelihood of fungal attack isminimised if the timber is kept dry and thiscan be achieved through suitable detailing ofthe structure. Insect attack is inhibited byapplication of insecticides but the effective-ness of these is limited.

VariabilityTimber, being a natural material, exhibitsconsiderable variability in its properties.Different species produce timbers with quitedifferent properties and even within a singlespecies the variability in properties such asstrength, elasticity, durability and appearancecan be high. The problem has been overcomeby the adoption of a grading system forcommercial timber in which individual planksand boards are inspected and placed intocategories according to their particular charac-teristics. Timber which is used in constructionis specified by grade and this ensures that itsproperties can be relied upon to be withinknown limits.

6.3.4 ConclusionTo sum up, timber is a material which offersthe designers of buildings a combination ofproperties which allow the creation of light-weight structures which are simple toconstruct. Its relatively low strength, the smallsizes of the basic components and the difficul-ties associated with achieving good structuraljoints tend to limit the size of structure whichis possible, however, and the majority oftimber structures are small in scale with shortspans and a small number of storeys.Currently, the most common application oftimber in architecture is in domestic buildingwhere it is used as a primary structural mater-ial, either to form the entire structure for abuilding, as in timber wallframe construction(also called timber-frame construction) (Fig.6.14), or to make the horizontal elements inloadbearing-masonry structures (Fig. 6.2).

The high ratio of strength to weight which isfound in timber also makes it particularly 191

Timber structures


suitable for structures of moderately largespan, especially if the level of imposed load islow. Timber is therefore used for the construc-tion of large, single-storey enclosures and formedium- to long-span roof structures.

Trees may be classified into two types,narrow-leaved trees and broad-leaved trees.Narrow-leaved trees are coniferous and mostlyevergreen - an exception is larch; broad-leavedtrees are mainly deciduous - an exception tothis is holly. Many differences exist betweenthe physiologies and anatomical structures ofthese two types of tree. The most significant ofthem is that the narrow-leaved species tend togrow much faster and to produce timberswhich are less dense and less strong than thebroad-leaved species.

Commercial timbers are subdivided into thetwo broad categories of softwoods andhardwoods, and these correspond approxi-mately to the botanical classifications. Thesoftwoods are derived from the narrow-leavedspecies and the hardwoods from the broad-leaved species. The terms are misleading,however, because in some cases they do notcorrespond to the physical properties of thetimber. Balsa wood, for example, which is oneof the softest of timbers, is classified as ahardwood because it is derived from a broad-leaved tree. In general, however, the coniferousspecies from which the softwoods are derivedare fast-growing and produce timbers with ahigh proportion of spring wood. They aretherefore less dense and less rigid than thehardwood timbers which are derived fromslow-growing, broad-leaved varieties.

6.4.3 Moisture content and moisturemovementThe moisture content of a specimen of timberis defined as the ratio of the weight of waterwhich it contains to its dry weight. It is alwaysexpressed as a percentage. In the living treethe moisture content is around 150%. After thetree is cut down, most of the water evaporatesand the timber eventually reaches an 'equilib-rium moisture content' whose exact valuedepends on the species and on the conditionsof the environment in which it is placed; itdepends principally on temperature andrelative humidity and is usually in the region of15 to 20%.

When timber dries out following the cuttingdown of a tree, moisture is lost initially from

6.4 Properties of timber

6.4.1 IntroductionIn this short section only the properties oftimber which are relevant to its use as a struc-tural material are reviewed. These are themechanical properties (i.e. strength andelasticity), fire resistance and durability. Twointroductory sections, in which the internalstructure of timber and the phenomenon ofmoisture movement are described briefly, arealso included. More detailed accounts of all ofthese topics will be found in Desch, H. E.,Timber, 6th ed. (revised Dinwoodie), London,1981.

6.4.2 Internal structureThe parts of the tree which are used as thesource of structural timber are the trunk andsometimes the major branches. These, incommon with other parts of a tree, arecomposed of cells; in the trunk these are longand thin and are aligned parallel to its direc-tion. The mechanical strength of timber isderived mainly from the fibrous tissue whichforms the walls of the cells together withdeposits within the cells.

The most active part of a living tree trunk isthe layer which lies immediately under thebark. This is called the cambium layer and it ishere that growth by cell division occurs. Inseasonal climates the rate of growth variesthroughout the year. It takes place mostly inthe spring and summer and is more rapid inspring than in summer. Spring wood is softerthan summer wood, and this differenceaccounts for the annual rings which are aprominent feature of the cross-section of a treetrunk and which are responsible for thephenomenon of 'grain'. The denser, harderparts of the grain are formed by summer woodand the softer parts by spring wood.192

the interiors of the cells and the only physicalchange which this causes is a reduction indensity. At a moisture content of around 30%(the exact figure depends on the species), allof the moisture in the cell interiors will havebeen lost and only the water in the cell wallswill remain. This condition is known as the'fibre saturation point' of the timber. Furtherreduction in the moisture content causes waterto be lost from the cell walls and this gives riseto shrinkage as well as to loss of weight. Theshrinkage is greater in the across-grain direc-tions than in the along-grain direction and, ofthe across-grain directions, it is greatertangentially than radially (Fig. 6.16). Theshrinkage is reversible; if the equilibriummoisture content of a timber becomes higher,due to a change in environmental conditions,then the timber will absorb water from theatmosphere and a corresponding expansionwill occur. The phenomenon is known as'moisture movement' and the extent to whichit occurs depends on the species.

The greatest amount of 'movement' takesplace during the initial drying process whichmust be regulated carefully because differen-tial shrinkage, due to variations in the rates atwhich moisture is lost from different parts,could cause damage in the form of cracks andwarping. The controlled drying out of timberafter felling is called 'seasoning'. It involves theadoption of measures to restrict the rate atwhich moisture is lost from the timber and torestrain the latter physically so as to preventexcessive deformation from occurring. Varioustechniques for seasoning have been developedbut most timber which is used in building iseither 'air-seasoned' or 'kiln-seasoned'. In air-seasoning the timber is stacked in well-venti-lated sheds and allowed to dry out underatmospheric conditions; in kiln-seasoning thetimber is dried more quickly under conditionsof close environmental control. In both casesthe stacking arrangement which is adopted isdesigned both to allow air to circulate freelybetween timbers, so as to promote drying, andto restrain the timber to prevent it fromwarping. A certain amount of deterioration inthe quality of the timber, due to the formation

of cracks and to warping (Fig. 6.17), isinevitable during seasoning, however, becausedifferences in the rates at which moisture islost from different parts of each timber cannot 193

Timber structures

Fig. 6.16 The shrinkage which occurs to timber duringthe initial drying out is greater in the tangential than in theradial direction.

Fig. 6.17 Examples of seasoning defects.

Tangential movement Radial movement (abouthalf the tangential movement)

Twist in board which is not cutparallel to the heart of the tree

Twist

Cuppingof plain sawnboard away from heart

Spring

In same direction as curvature of grain

Bow


6.4.4 Mechanical properties

6.4.4.1 IntroductionDue to the fibrous nature of the material andthe phenomenon of grain, the strength oftimber and its behaviour in respect of deform-ation are non-isotropic (not the same in alldirections).5 Many of the features of themechanical properties of timber can in fact beunderstood by imagining its internal structureto be similar to that of a bundle of drinkingstraws.

6.4.4.2 StrengthIn the direction parallel to the grain thestrength of timber is slightly less in compres-sion than it is in tension. This is due to thetendency for the cell walls, which are alignedparallel to the grain, to buckle under the actionof compressive loads. The strength at rightangles to the grain is considerably less thanthat parallel to the grain. This is due to separ-ation of the fibres when a tensile load isapplied and crushing of the cells when acompressive load is applied. A number ofdifferent values are therefore normally quotedfor the strength of a timber depending on thedirection in which a load is applied withrespect to the grain.

The grade stress values which are given inthe British Standard (BS 5268), are reproducedhere in Table 6.1. These are derived from theresults of tests carried out on structural-sizedelements rather than small specimens of thespecies concerned, but they are subject tomodification to allow for differences betweenthe test conditions and the conditions whichare likely to prevail in a real structure inservice. Examples of the latter are the durationof the load (the ability of timber to sustain aload depends on the length of time for which itis applied), the moisture content of the timber,the particular dimensions of the cross-sectionof the element and a number of other factors.

5 The strength and elasticity of a specimen of timberdepend on the orientation of the load with respect tothe grain. Strength and modulus of elasticity aregreater parallel to the grain than normal to the grain.

Fig. 6.18 Seasoning defects which occur due to thedifference between the rate of radial and tangential shrink-age.

be entirely eliminated. The extent andincidence of seasoning defects is reduced ifsmall cross-sections (i.e. cross-sections inwhich one of the dimensions is 75 mm or less)are used; these factors are also affected by theoriginal location and orientation of a plankwithin the tree (Fig. 6.18)

The equilibrium moisture content of thetimber in a building depends on the internalenvironment of the building. In the UK it isusually around 15% and is sometimes less inbuildings in which high temperatures aremaintained. Air-seasoned timber is suppliedwith a moisture content of around 22% and willtherefore suffer further drying and shrinkageafter installation in a building. Kiln-seasonedtimber can be supplied with a lower moisturecontent and is less likely to suffer a large shiftin moisture content after installation. Alltimber is likely to undergo small shifts inmoisture content, with corresponding dimen-sional changes, during the life of a building,because of changes in the environmentalconditions.194

Cupping of plain sawnboards caused by greatershortening of longergrowth rings during drying

Radial sawn

Diamonding

Plain sawn

Timber structures

195

Table 6.1 Grade stresses for softwood timbers (after BS 5268)


The allowable stress which is used in thedesign of a timber element is the grade stressmultiplied by one or more stress modificationfactors, which allow for the effects notedabove.

Because timber is much stronger parallel tothe grain than in other directions most timbercomponents are designed so that the stresswhich results from the principal load occursparallel to the grain.

6.4.4.3 Fire resistanceAlthough timber is combustible its perform-ance in fire is good, compared to that of othermaterials such as steel, aluminium or plastic.There are two reasons for this. Firstly, thestructural performance of timber is more-or-less independent of temperature (within therange of temperatures likely to be experiencedin a building fire) so that a timber structurewill continue to support load during a fire untilthe cross-sections of the elements are reducedby combustion to the point at which thestresses in them become excessive. It does notcollapse as a result of the effects of hightemperature alone. Secondly, the rate at whichtimber is consumed in a fire is low andrelatively constant. It is therefore possible todesign a timber structure to withstand theeffects of fire for a predetermined period, byoversizing the elements to allow for the mater-ial which will be consumed in that period. Thismethod of design for fire resistance is known

Table 6.2 Charring rates of timber

Loss of timber from one

face of element (mm)

Minutes

15 30 45 60 90

Softwood includinglaminated timber butexcluding western red cedar 10 20 30 40 60

Western red cedar 12.5 25 37.5 50 62.5

Hardwoods 7.5 15 22.5 30 45

as the 'sacrificial timber' or 'residual section'design method. Charring rates for timber aregiven in Table 6.2.

6.4.4.4 DurabilityTimber structures are likely to suffer from twotypes of decay: insect attack and fungal attack.Fungal attack is the more destructive of these,but this will only occur if the moisture contentof the timber is greater than 20% and can beprevented by suitable design. Timber structuresare therefore detailed so that direct contact withthe ground is avoided and in such a way as tominimise the risk of condensation forming ontimber elements. Features which will allowwater to collect and lie on timber are alsoavoided. The likelihood of fungal and insectattack can also be reduced by impregnating thetimber with various preservative materials.

6.5 Grading of timber

The properties of timber can exhibit consider-able variability, even within samples from asingle species, and where the material is usedfor structural purposes it is necessary that thisvariability should be quantified. The systemwhich is adopted for dealing with variability isone of grading. Processed timber is placed intogrades, which are simply categories, andindividual pieces of timber must comply withcertain performance requirements to qualify forinclusion in a particular grade. All timber istherefore inspected or tested at some stageduring its processing and allocated a gradedepending on its performance. The user oftimber specifies the grade which must be usedfor a particular component as well as itsspecies, and can then be confident that itsproperties will be within defined limits.

Most timber is graded in its country of originand, while the basic principles which areadopted for grading are similar for all majorproducing countries, there are differencesbetween countries in the details of themethods. For softwoods, which account formost structural timber in contemporary build-ing practice, most countries operate what is196

Timber structures

known as the 'appearance' or 'defect' system ofgrading in which limits are set for the sizes andfrequency of occurrence of various types ofdefect such as knots, grain (its width, straight-ness, angle of inclination to the principaldirection of the member, etc.) and the defectswhich arise from conversion6 and seasoning,such as wane, splits, twist, cupping andbowing (Fig. 6.17). Each piece of timber issimply inspected and placed in a grade accord-ing to the sizes and frequency of defects whichit contains. For each grade the limits on thesizes of defects are related to the dimensionsof the piece of timber being considered. Thus,large defects in a large plank can result in itbeing placed in the same grade as a smallerplank with smaller defects. If a large plank issawn up into two or more smaller planks thetimber must be regraded. Most countries haveadopted either four, five or six grades. Table6.3 shows the relationship between the gradeswhich are used by a number of major sourcesof structural timber.

The appearance or defects method of gradingis not entirely satisfactory as a means of classi-fying structural timber because the appearanceof the surface of a specimen is not a reliableenough guide to its strength properties. Forthis reason a system of stress-grading has beenadopted in the UK, and all timber which is usedin structures which have been designed tocomply with BS 5268 must be graded by thismethod, even though it may already have beengraded by the appearance method. The BritishStandard in which the stress grading proced-ures are specified is BS 4978.

The stress-grading system does in fact makeuse of visual criteria but these are based onthe effects which knots and other defects haveon the cross-section of the member rather thanon the appearance of its surface and thereforegive a more realistic assessment of the effectof defects on strength. Two visual stressgrades, general structural (GS) and specialstructural (SS) are specified for each species ofsawn timber and three visual stress grades arespecified for laminated timber (LA, LB and LC).The visual stress-grading system is now beingused in conjunction with 'machine-grading' inwhich the timber is passed through a devicewhich imposes a bending load on it and simul-taneously measures the resulting deflection. 197

6 Conversion is the name given to the process by whichfelled timber in the form of tree trunks is sawn up intoplanks.

Table 6.3 Relationship of 'appearance' grades for softwood timber

Norway, Sweden, I, II, III, IV V VIFinland, Poland unsortedand Eastern Canada

USSR I, II, III IV Vunsorted

Brazil No. 1 andNo. 2

British Columbia No. 1 Select No. 2 No. 3and Pacific clear merchantable merchantable commonCoast of No. 2 No. INorth America clear merchantable(R list) No. 3

clear

UK BS 3819 I Clear I,II III IV

Table 6.4 Basic sawn sizes of softwood (from BS 4471 Part 1: 1978)

Thickness (mm) Width (mm)75 100 125 150 175 200 225 250 300

161922253236384447506375

100150200250300 X

XXXXXXXXXX

XXXXXXXXXXXX

X

XXXXXXXXXXXX

XXXXXXXXXXXX

XX

XX

XXXXXX

XX

XXXXXX

XXX

XX

XXXXXX

XX

XXX

X

X

X

XX

XXX

X

XX


This system is based on the fact that thebending deformation which occurs to a speci-men of timber is a reliable guide to itsstrength.

The stress values which are quoted here inTable 6.1 are the grade stresses which arecurrently used in the UK. They are reproducedfrom BS 5268. The prefix M indicates that thematerial must be machine graded.

Table 6.5 Basic lengths of sawn softwood (m)(From BS 4471 Part 1: 1978)

1.80 2.10 3.00 4.20 5.10 6.00 7.20

2.40 3.30 4.50 5.40 6.30

2.70 3.60 4.80 5.70 6.60

3.90 6.90

6.6 Timber components

6.6.1 Solid timber

6.6.1.1 Sawn timber and its derivativesThe simplest timber components are solidelements of rectangular cross-section on whichno finishing was carried out after conversion.These are referred to as 'sawn-timber'elements. The range of cross-sectional sizes(basic sawn sizes) and lengths in whichsoftwood timber is available in the UK aregiven in Tables 6.4 and 6.5 and the dimen-sional properties of their cross-sections aregiven in Table 6.6. Note that the terminologywhich is used to describe the cross-section is198

to call the smaller of the two dimensions thethickness and the larger the width.

Basic sawn-timber elements tend to have aslightly irregular geometry and in situationswhere this is undesirable, such as in trusses inwhich the elements are joined by gusset platesand must therefore be of equal thickness, thetimber is 'regularised'. This is a machineprocess by which the thickness and/or width ofan element is made uniform throughout itslength. It is normally carried out on the criticaldimension only: that is on the thickness, forelements which are to be used in trusses, andon the width, for elements, such as floor joistsor wall studs, where this is the critical dimen-sion. Regularising involves a reduction ofaround 3 mm in the dimension which isaffected.

Table 6.6 Geometrical properties of sawn softwoods

Timber structures

199


Where a high quality of finish is required,timber elements are planed on all four sides(planed-all-round). This results in a largerreduction in the cross-sectional dimensionsthan occurs with regularising. Where a timberelement has been regularised or planed-all-round the structural calculations are based onthe modified dimensions.

6.6.1.2 Solid timber deckingTimber decking is machine-finished timber ofrectangular cross-section with edges which aretongued-and-grooved so that boards can beassembled into a flat deck in which adjacentelements give each other mutual support (Fig.6.19). Traditionally, this was produced in smallthicknesses, to span short distances of around300 mm to 450 mm, and was used in conjunc-tion with sawn-timber beams (joists) to formfloor or roof decks. The range of thicknesseshas been extended in recent years to allowwider spacing of beams. (See Table 6.7 and Fig.6.20.)

Fig. 6.19 The tongues and grooves in manufactureddecking elements allow concentrated loads to be distrib-uted.

Table 6.7 Span-load table for tongued and grooved solid timber decking

200

Timber structures

Fig. 6.20Tongued-and-grooved deckingelements are usedhere in conjunctionwith laminatedtimber frames. Theuse of thick cross-sections for thedecking elementsallows the elimin-ation of secondarystructural elementssuch as purlins.

6.6.2 Composite boards

6.6.2.1 IntroductionComposite boards are manufactured productscomposed of wood and glue (Fig. 6.21). Theyare intended to exploit the advantages oftimber while at the same time minimising theeffects of its principal disadvantages, which arevariability, dimensional instability, restrictionsin the sizes of individual components andanisotropic behaviour.

There are two basic types of compositeboard: laminated boards (also calledplywoods), in which the constituent parts arerelatively large and are either veneers or stripsof wood, and particle boards, in which theconstituent pieces of timber are very smallfibres or particles.

6.6.2.2 Laminated boardsThe components of laminated boards areassembled by gluing such that the direction ofthe grain is different in adjacent layers. This,together with the high level of glue impregna-tion, reduces variability and improves dimen-sional stability in the plane of the board. It

also reduces the tendency of the board to splitin the vicinity of nails and screws.

Laminated boards are subdivided into twocategories: veneer plywood and core plywood.Veneer plywood is defined as plywood in whichall the plies are made up of veneers, up to7 mm thick, orientated with their grain parallelto the surfaces of the panel. Core plywood isplywood which has a relatively thick core andwhich may contain layers with grain normal tothe plane of the board. Batten board, block-board and laminboard are included in thissecond group (Fig. 6.21).

The plywood which has been most exten-sively used as a structural material is veneerplywood. A large number of types of this areproduced by the timber industry but few areconsidered suitable for structural use. Theprincipal structural plywoods which are used inthe UK are listed in BS 5268. These are usuallyavailable in sheets measuring 2440 mm by1220 mm and in a range of thicknesses andsurface qualities. The good in-plane rigidityand out-of-plane bending strength of veneerplywood make it suitable for a variety of struc-tural applications, usually in some form of

201


202

composite construction. It is ideal for shearpanels such as the webs of built-up-beams(Fig. 6.25) or the skins of wall or floor panelswhich are used for diaphragm bracing. It isalso used as the flanges in stressed-skin panelstructure (Fig. 6.40).

6.6.2.3 Particle boardsParticle boards have traditionally been littleused as structural components and no data forthem are given in BS 5268. There is an increas-ing tendency for these to be used in the struc-tural role, however, particularly as a substitutefor tongued-and-grooved floor and roofdecking.

6.6.3 Built-up-beams

6.6.3.1 IntroductionBuilt-up-beams are components in which anumber of separate planks or boards are

assembled, by gluing, nailing, screwing, boltingor a combination of these, to form beams ofcomposite cross-section. Various types ofbuilt-up-beam are produced and these are allintended to exploit the good structural proper-ties of timber and to overcome the size restric-tion which is inherent in the use of solidsawn-timber elements. They are normally usedin forms of construction which are similar tothose which are suitable for sawn timber (i.e.as joists in the traditional forms of floor androof structures - see Fig. 6.41) and they allowlonger spans to be achieved than are possiblewith solid sawn-timber. Some of the morecommonly used forms of built-up-beams arereviewed briefly here.

6.6.3.2 Laminated timberLaminated timber, which is also called'gluelam', is a product in which elements withlarge rectangular cross-sections are built up by

Fig. 6.21 A selection of timber composite boards.

Traditional plywood

Four-ply plywood

Six-ply plywood

Three-ply blockboard

Five-ply blockboard

Laminboard

Composite panels

means of finger joints (Fig. 6.23). The laminat-ing process also allows the construction ofelements which have a curved profile (Figs 6.3,6.20 and 6.24).

The general quality and strength oflaminated timber is higher than that of sawn-timber for two principal reasons. Firstly, theuse of basic components which have smallcross-sections allows more effective seasoning,with fewer seasoning defects, than can beachieved with large sawn-timber elements.Secondly, the use of the finger joint, whichcauses a minimal reduction in strength in theconstituent boards, allows any major defectswhich are present in these to be cut out. Theallowable stresses of laminated timber aretherefore higher than for equivalent sawntimber. Laminated timber elements areproduced in three standards of finish: architec-tural (machined, sanded and free fromblemishes), industrial (machined and sandedbut with blemishes permitted) and economy(untreated after manufacture), but there is nostructural difference between these categories.

6.6.3.3 Plyweb beamsPlyweb beams are built-up-beams of rectangu-lar-box or l-cross-section in which solid timberflanges are used in conjunction with a web of

gluing together smaller solid timber membersof rectangular cross-section, usually either38 mm or 50 mm in thickness (Fig. 6.22). Thedirection of the grain is the same for alllaminations and is usually parallel to the axisof the element. The obvious advantage of theprocess is that it allows the manufacture ofsolid elements with much larger cross-sectionsthan are possible in plain sawn timber. Verylong elements are also possible because theconstituent boards are joined end-to-end by

Fig. 6.23 The glued finger jointis an essential part of the technol-ogy of laminated timber. It allowsthe creation of very long sub-elements from which majordefects, such as knots, have beenremoved [Photo: TRADA].

203

Timber structures

Fig. 6.22 Typical cross-section of a laminated timberelement [Photo: TRADA].


204

Fig. 6.24 St loseph's Chapel, Ipswich, J. Powlesland,structural engineer. The timber roof structure here is basedon large cross-section laminated timber elements, whichare used in conjunction with laminated timber secondaryelements and solid timber decking [Photo: SidneyBaynton]

veneer plywood (Fig. 6.25). Plyweb beams arenormally considerably cheaper than gluelamequivalents. They are used in similar types ofconstruction to solid sawn-timber beams butthey allow wider spacings (around 1.3 m) andlonger spans (up to 20 m ) to be achieved.

6.6.3.4 Other forms of built-up-beamsA number of other types of built-up-beamare produced in timber. One example is thelattice-type beam, which is illustrated inFig. 6.26.

6.6.4 Trussed raftersTrussed rafters are lightweight timber trusses,Fig. 6.27; they are usually factory-made assem-blies of small, sawn-timber sections,regularised on thickness and jointed withpunched-metal-plate fasteners, but they can beassembled on site with bolted joints. Typicalindividual element sizes are 38 mm X 75 mmfor short-span versions (6 m) and 50 mm X100 mm for the larger versions (10 m). Mostmanufacturers produce a range of sizes in100 mm increments of span and 2.5° incre-ments of pitch between 15° and 35°.

Trussed rafters are designed to be used inpitched roof construction and spaced at closecentres (around 600 mm) so as to carry theroof cladding without the need for a secondarystructural system, other than tile battens orsarking boards. They therefore perform asimilar role to that of the rafters in traditional

Timber structures

Fig. 6.25 Plywebbeams are built-up-beams with sawn-timber flanges andplywood webs.Composite elementssuch as this are used toextend the span rangeof traditional joistedforms of construction.Note the web stiffenersin the beams herewhich are positioned atthe support pointswhere shear loads arehigh.

Fig. 6.26 Built-up lattice beams. Proprietary components such as these are used, like plyweb beams, to extend the spanranges of traditional forms of timber construction. 205


forms of roof construction, from which theyderive their name, but they allow a much moreeconomical use of timber to be made due tothe use of the triangulated form. They aregenerally considered to be most suitable forspans in the range 5 m to 11 m. Typical spansfor which trussed rafters are suitable are listedin Table 6.8.

6.6.5 Large one-off trussesLarge trusses are used as the horizontalelements in post-and-beam structures of boththe skeleton-frame and the loadbearing-walltypes. They are normally only specified forsituations where very high strength or rigidityis required, either because a long span mustbe achieved or because heavy loads are carried

Fig. 6.27 Trussed rafter. Trussedrafters are lightweight triangulatedelements which are designed to beused at close spacings to carry roofcladding directly.

Toothed plate fastener

Main ties acting as ceiling joists

206

Rafter size Span

Pitch (degrees)

15 17½ 20 22½ 25 27½ 30 32½ 35

5 0 X 1 2 5 10.01 10.24 10.44 10.62 10.77 10.94 11.00 11.00 11.00

Table 6.8 Typical spans for fully triangulated trussed rafters

mm

38 X 75

38 X 100

38 X 125

44 X 75

44 X 100

44 X 125

50 X 75

50 X 100

m

6.03

7.48

8.80

6.45

8.05

9.38

6.87

8.62

m

6.16

7.67

9.00

6.59

8.23

9.60

7.01

8.80

m

6.29

7.83

9.20

6.71

8.40

9.81

7.13

8.97

m

6.41

7.97

9.37

6.83

8.55

9.99

7.25

9.12

m

6.51

8.10

9.54

6.93

8.68

10.15

7.35

9.25

m

6.60

8.22

9.68

7.03

8.81

10.31

7.45

9.38

m

6.70

8.34

9.82

7.14

8.93

10.45

7.53

9.50

m

6.80

8.47

9.98

7.24

9.09

10.64

7.67

9.66

m

6.90

8.61

10.16

7.35

9.22

10.81

7.78

9.80

due to the infrequent spacing of the primarystructural elements. So far as the overallgeometry of trusses is concerned, any shape istheoretically possible so long as a stableconfiguration is adopted. The most favouredshapes are the traditional pitched truss andthe parallel-chord truss (Fig. 6.28). Themansard shape is a commonly used variationon the latter. Bowstring shapes are also usedbut these are more expensive to construct(especially if they are at the higher end of thespan range). Typical span ranges for all ofthese types are given in Table 6.9.

The internal geometry of a truss must bestable and this is achieved by adopting a fullytriangulated arrangement. Where, for anyreason, these types of geometry cannot beadopted, stability can be provided by continu-ity of elements through joints (Fig. 6.29). Forefficiency in the use of material, internalangles of around 60° are best and angles whichare less than 30° are avoided. Economy inconstruction results if the total number ofjoints is minimised, but this must not beachieved at the expense of ill-conditionedtriangles (i.e. triangles with angles less than30°) or the creation of excessively longcompression elements. Adoption of a smallnumber of joints also reduces the deflectionwhich results from slipping at the fastenings.

Timber structures

Fig. 6.28 Typical configurations for large one-off timbertrusses.

Such slipping occurs with all types of mechan-ical fastener and can increase the total deflec-tion of the truss significantly.

The overall depth of a timber truss must besuch as to give a fairly small ratio of span todepth (Table 6.9) and this is particularlyimportant where the span is long (greater than

Pitched

Parallel chord

Bowstring

10-30

15-45

15-75

5

8-10

6-8

207

Table 6.9 Typical span ranges and proportions for one-off timber trusses

Configuration Span range (m) Span/depth

Pitched Pratt truss

Parallel chord Pratt truss Mansart truss

Pitched fink truss

Bowstring truss

Fig. 6.29 Traditional forms of semi-truss such as theseare not fully triangulated and depend on continuity of thesub-elements through the joints for stability.

20 m). Timber trusses must normally be signifi-cantly deeper than equivalent steel trusses andprovision must be made for this in the designof buildings which are to have a primary struc-ture of timber.

Most one-off trusses are assemblies ofmultiple-component elements which areconnected together at multiple lap-joints inwhich the fastening components are bolts,usually with timber connectors (Figs 6.30 and6.31). The critical factor which determines thefeasibility or otherwise of a proposed arrange-ment is frequently the design of the joints(Fig. 6.32). This must be such as to allow suffi-cient area of overlap between elements toaccommodate the number of connectorswhich are required to transmit the load. It isfrequently found advantageous to use low-grade timber for large trusses because the

Fig. 6.30 Typical joint arrangements for large, one-offtrusses. The fastening elements are bolts which are used inconjunction with timber connectors.

larger elements which result from this providemore space for the joints. This also reducesthe slenderness ratios of the compressionelements.208


Three members

Five members

Seven members

(a)

(b)

(c)

(d)

Timber structures

209

Fig. 6.31 Timber connectors. These components are usedin conjunction with bolts to make lap-type connectionsbetween timber elements. Their function is to reduce thestress concentration which occurs at the bolt.(a) Toothed-plate connector.(b) Shear-plate connector.(c) Split-ring connector.

Fig. 6.32 Typical arrangement of joints in a multiple-element large timber truss.(a) Configuration of truss.(b) Section through joint B.(c) Section through joint A.(d) Section through joint C.

(a)

Single-sided

Double-sided

(c)(b)


Fig. 6.33 Detail oflarge one-off timbertruss.

6.6.6 Joints

6.6.6.1 IntroductionJoints in structural timber are made either bydirect bearing of one element on another, as inthe case of joists bearing on a wall-plate or onjoist hangers (Fig. 6.34) or with the aid of

fastening components which transmit the loadbetween elements in lap-type arrangements(Fig. 6.35). The design of the direct-bearingtype of joint is straightforward and is rarelycritical in determining the feasibility of a struc-ture. Lap-type joints between elements intrusses, skeleton frames or built-up-beam

210

Fig. 6.34 Timber joists are normally supported by directbearing, either on a wall-plate, as in (a) or on joisthangers, as in (b).

Oversite concrete Oversite concrete

(a) (b)

Timber structures

(b)(a)

(c)

Fig. 6.36 The three principal types of fastening com-ponent which are used in timber engineering are nails,screws and bolts.

Two types of fastening medium are used fortimber joints: mechanical fasteners (nails,screws, bolts, etc. (Fig. 6.36)) and adhesives.Additional jointing components such as gussetplates of plywood or steel are also frequentlynecessary and joints are normally designed sothat the fastening medium is loaded in shear.A fairly wide choice of different possiblearrangements will usually be available for aparticular joint due to the range of fasteningcomponents which are available and to thepossibility of placing elements either in asingle plane and joining them through gussetplates, or making them overlap and joiningthem directly together in lap-type joints. 211

Fig. 6.35 Lap-type joints. In each case the load is trans-mitted between elements by shear in the fastening com-ponents, which are either nails, screws or bolts.

components, can affect the feasibility of aproposed structure and are therefore an im-portant consideration in the early stages of adesign.


Fig. 6.37 Toothed-plate and shear-plate connectors areavailable in single- and double-sided formats and can beused either for timber-to-timber or timber-to-steeljunctions.

achieved by using a large number of smallfastening elements (Fig. 6.36a) (although thereare limits to the number of fasteners which canbe placed in a single line due to the uneven-ness in load distribution which tends todevelop between them) or by using individualcomponents which are specifically designed toreduce concentrations of stress, such as timberconnectors (Fig. 6.31). The required overallwidth of a timber element is frequently deter-mined by the need to provide sufficient spaceto accommodate the fastening componentswhich are required at the joint rather than bythe need to provide a cross-section which islarge enough to resist the internal force whichthe element carries.

6.6.6.2 loints with mechanical fastenersThe principal types of mechanical fastener arenails, screws, bolts, timber connectors andpunched metal plates. Nails and screws, whichhave similar dimensions, are capable of carry-ing similar amounts of load. Nails are normallypreferred due to their lower primary cost andto the lower labour cost involved in drivingthem into the timber. Screws are used onlywhere high resistance to withdrawal isrequired. Individual bolts carry higher loadsthan nails but the safe working load for a boltis usually considerably lower than the strengthof the bolt itself, due to the inability of thetimber to accommodate the high concentrationof stress which occurs at a bolted connection.Thus, although fewer bolts than nails may berequired for a particular connection the total

The selection of the joint type is usuallybased on a number of considerations, such aseconomy (both in the cost of the fasteningcomponents and in the manufacture of thejoint), the type of workshop facilities which willbe available, the appearance of the joint, andany special requirements which are imposedby the way in which the structure will beassembled. A fundamental decision is whetherto use mechanical fasteners or adhesive. Gluedjoints produce more satisfactory structuralconnections but are difficult to make on site.Components which are fastened with gluedjoints must normally therefore be producedunder factory conditions and considerations oftransport dictate the maximum size of these.Very large elements must be assembled on siteand this will normally mean that mechanicalfasteners will have to be used.

A critical factor in the design of timberjoints, especially mechanical joints, is theweakness of the material in crushing andshear; the basic strategy which is adopted toovercome this is to reduce the concentration ofstress which occurs by bringing the largestpossible area of timber into play. In the case ofa glued joint this is achieved by providing alarge contact area on which the glue can act.Where mechanical fasteners are used it can be212

Fig. 6.38 Punched-metal-plate fasteners, which formcombined nail-and-gussetplate units, are used as thefastening elements in thesetrussed rafters.

Timber structures

size of the connection may be greater whenbolts are used due to the larger spacingrequirements combined with the relatively lowload-carrying capacity of individual bolts intimber.

Timber connectors are the mechanicalfasteners which can carry the highest loads andthere are three types of these: toothed-plateconnectors, shear-plate connectors and split-ring connectors (Fig. 6.31). Toothed-plate andshear-plate connectors are available in singleor double-sided versions. The former are usedfor timber-to-plywood and timber-to-steeljoints and the latter for timber-to-timber joints(Fig. 6.37). All timber connectors are used inconjunction with bolts, which carry a propor-tion of the load which passes through thejoint. The principal function of the connector isto reduce the concentration of stress.

The detailed design of a mechanical jointwhich is made with nails, bolts or timberconnectors involves decisions on the type offastener to be used, the number which arerequired to provide the necessary strength, andthe spacing which must be adopted betweenthese. The choice of fastener type is affected byconstructional factors and considerations ofstrength. Normally, nails and bolts (withoutconnectors) are used for structures of modestsize and timber connectors for the larger struc-

tures such as one-off trusses. The number ofindividual fasteners which are required for ajoint depends on the load-carrying capacity ofthe fastener chosen in relation to the load whichpasses through the joint. This depends in turnon the size of the individual fastener, the speciesof the timber, and the dimensions of the timber.Individual fasteners must then be positioned soas to satisfy the minimum spacing requirementsand in some cases it is necessary to increase thesize of the element which is used in order toaccommodate the fasteners safely.

Punched metal plates (Fig. 6.38) combinethe functions of nails and gusset plates andare used to make connections in trussedrafters and trusses in which all sub-elementsare in a single plane. They are not suitable forthe largest trusses, however, and perform bestin elements of relatively short span (up to12 m). A range of products of this type is avail-able and load values for design must beobtained direct from manufacturers.Calculations are rarely performed for this typeof joint. Punched-metal-plate fasteners are notsuitable for joints which must be assembledon site.

6.6.6.3 Glued jointsA range of adhesives which are suitable formaking structural joints in timber and which are 213


resistant to a variety of environmental condi-tions is available. Both timber-to-timberconnections and joints between timber andother materials can be made with glue. Gluedjoints are designed so that the glue is loaded inshear only and tests have indicated that whenthis type of joint is loaded to breaking point,the failure occurs in the timber adjacent to theglued surface and not in the glue itself. Theallowable stresses on which the design of gluedjoints are based are therefore the shearstresses, parallel to the grain of the timberconcerned; they are subject to the normalmodification factors to allow for the effects ofduration of load and the direction of load inrelation to the grain. Although the rigorousdesign of a glued joint is a complicated process,the basic requirement is to provide an area ofcontact between the components which is largeenough to maintain the shear stress at anacceptable level. The feasibility of a proposedjoint, and therefore of the structural element ofwhich it forms part, is frequently determined bythe required size of the contact area. This must

obviously not be excessive in relation to thesizes of the sub-elements being joined.

An additional factor which must be consid-ered in relation to the feasibility of specifying aglued joint is the condition of the environmentin which the joint will be made. This shouldideally be that of the workshop or factory. Mostglues have to be cured in conditions of closeenvironmental control and surfaces which areto be glued must be carefully prepared toachieve a glue line which is as thin as possible.Also, the surface to which the glue is appliedshould be freshly machined (not more than 48hours old); this is the case because the openends of cut cells, whose presence is essentialto achieve effective bond with glue, tend toclose with time. The standards of both environ-mental control and the type of supervisionwhich are necessary for the manufacture ofreliable glued joints are frequently difficult toachieve on site, and it is normal practice tospecify mechanical fasteners rather than gluefor site-made joints.

Fig. 6.39 Typical timber floor deckwith closely spaced joists supportinga floor consisting of boarding.

214

Herring-bone strutting

Joist ends on steelbearing plate

Block partition

Chipboard orplywood flooring

Brick outer leafto cavity wall

Wall-plate

Timber structures

Fig. 6.40 Stressed-skindeck in which compositeaction takes placebetween joists andboarding. This extendsthe span range of thetraditional joisted floor.

6.7 Structural forms for timber

6.7.1 Floor and roof decksPerhaps the simplest type of timber structureis the traditional floor or roof deck in whichsawn-timber beams (joists) are spaced at closecentres (450 mm to 600 mm) supporting a deckof relatively thin boarding (Fig. 6.39).Traditionally, the boarding was of machinedsolid timber boards which had tongued-and-grooved edges so as to spread the effect ofconcentrated load. In recent years plywood,blockboard and even chipboard have beenused in this role. These provide the deck withgreater resistance to in-plane loads thantongued-and-grooved forms of boarding, andgive it the ability to act more effectively asdiaphragm bracing, and therefore to contributeto the general stability of the structure.

Timber decks are normally used in conjunc-tion with loadbearing-wall vertical elements(either of timber or masonry) but can also beused with skeleton frames. The range of spansfor which the traditional joisted deck isnormally used is 3.5 m to 6 m. Tables 6.10 and

6.11 give the spans which are possible for thestandard range of sawn-timber sizes.

Two comparatively recent developments inthis form of construction are composite decks,in which the joists and sheeting act together inT- or box-beam configurations, and the use ofbuilt-up-beams in place of sawn timber to formthe joists. If plywood is used to form a floor orroof surface and is nailed and glued to thejoists, a T- or box-beam structure is createdwhich is stronger, and can span further, than atraditional boarded deck of equivalent dimen-sions (Fig. 6.40). The resulting element iscalled a 'stressed-skin' deck and this form ofstructure allows spans of up to 9 m to beachieved with the largest sizes of sawn-timberjoist. Its use therefore extends the span rangeof flat timber decks made with sawn-timberelements. A good structural connectionbetween the web members (internal stringers)and the skin is required if effective compositeaction is to develop, however, and the jointshould preferably be made by dense nailingand gluing. The assembly of the deck is bestcarried out under factory conditions and 215

Plywood splice

Upper skin

Lower skin

End blocking

Splice plate

NoggingInternallongitudinalweb member

External longitudinalweb member


Table 6.10 Span ranges for sawn-timber joists in domestic floors (section sizes under the line are notin common use)

216

Table 6.11 Span ranges for sawn-timber joists in flat roof decks to which pedestrians do not have access(section sizes under the line are not in common use)

Timber structures

217


Table 6.12 Simple rules for preliminary sizing ofstressed-skin panels

Parameter

Minimum overalldepth

Minimum stringerwidth

Maximum stringerspacing

Tension skinCompression skin

0.75 kN/m2

span4050 mm

600 mm

8 mm9 mm or 12

\mposed load

1.5kN/m2

span3550 mm

400 mm

9 mmmm 12 mm, 15 mm

or 18 mm

stressed-skin panels are normally regarded ascomponents which must be prefabricatedrather than made on site. They are used wherelonger spans are required than are possiblewith traditional forms of timber deck or wherea smaller constructional depth must beprovided. Data for stressed-skin panels aregiven in Table 6.12.

The second development of the traditionaldeck is the use of built-up-beams in place ofsawn-timber joists for the principal elements(Fig. 6.41). Plyweb and laminated timberbeams, as well as various proprietary lattice-beam types have been used in this role andthese allow much larger spans (up to 20 m) to

be achieved than are possible with sawn-timber elements. Economy in the use of thesemore sophisticated components requires thatthey be spaced further apart than traditionaljoists and beam spacings of around 1.5 m to3.5 m are usually adopted. This requires that afairly thick form of boarding be used to formthe deck skin.

6.7.2 Pitched roof structuresIn the traditional pitched roof structure, sawn-timber rafters are positioned at close centres(450 mm to 600 mm) supporting a roof skin ofsarking boards or tile battens (Fig. 6.42). Thevertical support structure is usually of theloadbearing-wall type and, because the raftersproduce an outwards thrust at the wallhead,the system is suitable for small spans only (upto 3.5 m). The span range can be extended byuse of tie elements, and in traditionalconstruction various arrangements of semi-trussing were devised to allow longer spans tobe achieved (Fig. 6.43). Traditional roof formsof this type require fairly large sizes of timber,however, because the principal elements arestressed in bending as well as axially, and somust have large cross-sections to resist this.Also, many of their geometries are fundamen-tally unstable and depend on the continuity ofelements through joints for stability. Elements

Fig. 6.41 Plyweb built-up-beamsat close spacing are used here toform the principal structuralelements of a roof deck. Thearrangement is similar to that of atraditional joisted deck with sawn-timber elements. The use of built-up-beams allows longer spans tobe achieved.

218

Fig. 6.42 Traditional pitched roof arrangement of closelyspaced rafters and ridge board. The rafters exert a horizon-tal load on the tops of the walls.

must therefore be long as well as of largecross-section. This method of construction hasnow been largely superseded by the trussed-rafter system in which rafters are incorporatedinto a fully triangulated arrangement (Fig.6.27). Very small timber sections may beemployed with this arrangement and savingsare possible of up to 30% in the volume oftimber which is used, as compared with trad-itional methods. Trussed rafters can be builtup on site using bolts or nails but morecommonly are proprietary components jointedwith punched-metal-plate fasteners. They aremost suited to spans in the range 5 m to 11 m.A disadvantage of the trussed-rafter roofsystem is that the structure occupies the wholeof the roof space.

Fig. 6.43 Semi-trussed arrangements reduce or eliminatethe horizontal force on the wallhead but require long sub-elements.(a) The tie at eaves level eliminates horizontal thrust at thewallhead. The tie is here spanning the entire width of thebuilding, which limits the span.(b) The collared roof leaves some horizontal thrust ateaves level but requires a shorter length of tie.(c) A longer span is possible if the tie is supported fromthe ridge.(d) This semi-trussed arrangement is dependent on con-tinuity of elements through joints for stability.

Timber structures

219

(d)

Binder

Ceiling joistsacting as ties

Wall plate

(c)

Ridge board

Hanger

Rafters

(b)

(a)

Ridge board

Wall-plate

Rafters


Fig. 6.45 Section through truss-and-purlin roof. Thepurlins are positioned so that the 'common' rafters liein the same plane as the top elements of the trusses.

An alternative to trussed-rafter constructionis purlin construction in which plain rafters aresupported on a series of secondary beams,called purlins, which run parallel to the direc-

tion of the ridge. The purlins can be supportedon trusses or cross-walls of timber-stud ormasonry (Figs 6.44 to 6.47). Where trusses areused they are usually spaced 2 m to 3 m apart.220

Fig. 6.44 Truss-and-purlin roof. In the truss-and-purlinsystem fully triangulated trusses support purlins which inturn support common' rafters.

Purlin

Binder

Tie

Wall plate

Truss

Common rafter

Purlin

Rafters

Pack nailedin position

HangerCeiling joists

Strut

Ceilingbinder

Table 6.13 Approximate span ranges fordifferent types of timber pitched-roof structures

Span

Spanning across Timber joist (flat), 2.5 m to 6 mbuilding plain rafter, close

couple

Tied rafter 3 m to 7 mTrussed rafter 6 m to 20 mTruss and purlin 10 m to 30 m

Spanning between Sawn-timber purlin 2.5 m to 3.5 mcross-walls Plyweb, laminated or 2.5 m to 8 m

steel purlin

Timber structures

Fig. 6.46 Purlin roof supported on cross-walls. Thisarrangement leaves the roof space free of structure. Thepurlins carry a relatively large area of roof, however, andtherefore significant load. Built-up-beam sections arerequired to achieve spans greater than 3 m.

Fig. 6.47 Typical cross-section through a purlin roof.

Cross-wall spacing is normally wider (5 m to6 m). The trusses which are used in truss-and-purlin construction are much stronger andheavier than trussed rafters and usually havesub-elements which consist of several boardsjointed by timber connectors. Alternatively, atruss can be made by bolting together anumber of trussed rafters.

Purlin construction has the advantage ofleaving the roof space relatively clear of struc-ture. It also allows the longitudinal walls of abuilding to be non-loadbearing, which can bean advantage with certain types of planning.The purlins themselves tend to be fairly heavilyloaded, however, and must normally betimbers of large cross-section. The size ofcross-section which is required depends on thepurlin spacing and on the weight of the roofwhich is carried, but the largest sizes of sawn-timber elements are rarely capable of spanningfurther than around 2.5 m when acting aspurlins. Where larger spans are required, built-up timber beams or lightweight steel elementsare used. The depth of built-up timber purlinsis likely to be large (800 mm for spans in theregion of 5 m to 6 m and provision for thismust be made in the planning of the building.

In truss-and-purlin structures it is commonpractice for the top element of the truss to belocated in the same plane as the rafters as thissimplifies the detailing of the roof. The arrange-ment requires that purlins be suspended fromthe top elements of the trusses rather thanlocated on top of them (Fig. 6.45) as was thenormal practice in traditional roofs (see Section6.2). Span ranges for the types of timber roofdescribed here are given in Table 6.13.

221

Jack rafters birdsmouthedover purlins and wall-plate

Floor joists

Purlin

Non-loadbearingpartition

Externalloadbearingwall

Internal loadbearing wallor beam to supportintermediate floor

Ridge purlin


Fig. 6.48 Timber wallframe - platform frame type. In theplatform frame the individual wall and floor panels areindependent of each other. The sequence of constructionis that the ground-floor walls are erected first, followed bythe floors. These form a 'platform' on which the first-floorwalls are then erected. The system lends itself to prefabri-cation and has the additional advantage that none of theindividual elements is of great length. Its disadvantage isthat the level of structural integrity is lower than with othertypes of arrangement.

6.7.3 Timber wallframe structuresTimber wallframes are a type of loadbearing-wall structure. They are sometimes calledsimply timber frame structures but they are nottrue skeleton frames. In wallframes, timberwall-panels (stud panels) consisting ofelements of small cross-section placed at closecentres are used to support traditional formsof roof and floor structure (Figs 6.48 to 6.51).The overall size of the structure is usuallysmall: there are rarely more than two storeysand floor spans are in the range 3.5 m to 5 mwith roof spans of around 6 m to 12 m. The

Fig. 6.49 Timber wallframe - balloon frame type. In theballoon frame the studs which form the walls are continu-ous through two floors and the floor joists are attacheddirectly to them. This arrangement provides better struc-tural integrity than the platform frame and produces aweathertight enclosure more quickly. It cannot be prefabri-cated, however, and requires the use of very long timberelements.

system is used principally for domesticbuildings.

The basic planning requirements for timberwallframe buildings are the same as for allloadbearing-wall structures. The wall plan,which must be more-or-less the same on everystorey, must be capable of supporting all areasof roof and floor and, as both the floor androof structures are one-way-spanning systems,this requires an arrangement of primaryloadbearing walls which are parallel to eachother and spaced at roughly equal distancesapart. Atypical arrangement is shown in222

Head binder.

Head plate

Floor joists

Boarding

-Sole plate

Studs

Boarding

Sole plate

Base wall

Header

Header

Head plate

Head binder

Floor joists:

Cill plate

Head binder

Head plate

Floor joists

Cavity barrier

Ledger

Cavity barrier

Floor joists

Sole plate

Base plate

Cill plate

Plywoodsubfloor

Studs

Plywoodsubfloor

Fig. 6.50 Timber wallframe - modified frame type. In themodified frame the wall elements are only one storey highbut the floor joists are attached directly to them. Themodified frame is a compromise between the platform andballoon arrangements which combines some of the advan-tages of both.

Fig. 6.51 Timber wallframe - the independent frametype. In the independent frame the wall elements are onestorey high but the floor joists are carried on a steel angleattached to the inner face of the wall. This arrangementlends itself to prefabrication. The fact that the floor joistsdo not penetrate the outer wall improves the thermalefficiency of the wall.

Fig. 6.52. Note that, as in the case of masonrystructures, spine- and cross-wall forms areused. In spine-wall arrangements the roofnormally spans the full width of the buildingand only the floors are supported on inter-mediate walls. Cross-walls are normally usedin conjunction with purlin roof construction.

The wall-to-floor and wall-to-roof junctionsin this form of structure are non-rigid andlateral stability is provided by the wall, floorand roof panels themselves, which serve asdiaphragm bracing in the vertical and horizon-tal planes. Care must be taken in the planning

of the buildings to ensure that vertical-planebracing is correctly positioned. Bracing wallsmust be provided which are as symmetricallydisposed as possible on plan.

It is also necessary that both the wall andthe floor panels should have sufficient strengthto resist in-plane racking7 caused by wind load.This is done by sheathing with plywood orsome other form of composite board

7 The tendency of the original rectangular shape tobecome a parallelogram due to shearing action. 223

Timber structures

Head binder

Head plate

Floor joists

Cavity barrier

Head plate

Sole plate

Floor joists

Sole plate

Cill plate

Base wall

Tonguedand groovedboarding

Studs

Tonguedand groovedboarding

Head binder

Head plate

Floor joists

Metalconnectingplate-

Steel angle letinto face of studs

Floor battens

Damp proofmembraneto concrete raft

Cill plate

Edge of concrete

Sole plate

Plywoodsubfloor

Cavitybarrier

Head plate

Sole plate

Plywoodsubfloor

Studs

studs are positioned at the same spacing asthe floor joists so as to minimise eccentricityin the transfer of load. The advantages of theplatform frame are that long timber elementsare not required and that the system lendsitself to prefabrication. This is due to therelative independence of the floors from thewalls and to the fact that none of the panels isvery large and difficult to transport. The factthat studs and floor joists are aligned is alsoan advantage because it eases the construc-tional planning by allowing a grid system to beused. The principal disadvantages of theplatform frame are that the building does notbecome a weathertight shell until the construc-tion is at a fairly advanced stage, and that thestructure is slightly less rigid than alternativeforms.

In balloon frames (Fig. 6.49) the wallframesare two storeys in height (this is normally thefull height of the building) and the floor joistsare attached individually to wall studs by lapjoints. The advantage of this system is that itallows the roof to be completed at an earlystage in the construction process, before thefloors are in position, to provide a weathertightshell. It also provides a structure which is morerigid than the platform frame. Its disadvan-tages are that it requires very long elements forthe wall stud, that it does not lend itself toprefabrication and that the joists and studs arenot aligned.

Two other types of wallframe are the'modified' frame and the 'independent' frame(Figs 6.50 and 6.51). Both have wall panelswhich are one storey high. In the modifiedframe the ground- and first-floor wall panelsare attached directly to each other and thefloor joists are then attached directly to wallstuds, as in the balloon frame. In the indepen-dent frame the same arrangement for wallpanels is used but the floor is supported by asteel angle or other element attached to theedge of the lower panel and does notpenetrate the thickness of the wall. Themodified and independent frames are intendedto combine the good features of the platformand balloon frames. They achieve something ofthe structural continuity of the latter without

Fig. 6.52 Typical plan-form of a small-scale timberwallframe building. The arrangement conforms to theparallel-wall format of all loadbearing-wall forms ofconstruction.

(traditional tongued-and-grooved boardingdoes not provide good racking resistance dueto the possibility of individual boards slippingrelative to each other) or by incorporatingdiagonal bracing (tensioned steel wire ortimber) into the wall and floor panels.

A number of different wall configurations areused in timber loadbearing-wall structures. Thetraditional methods are the 'platform' frame orthe 'balloon' frame. In platform frames (Fig.6.48) the wall and floor panels are more-or-lessindependent entities and the wall panels are ofsingle-storey height. Each wall is built on topof the platform which is formed by the floorbelow it and the floor panels thereforepenetrate through the structural part of thewall to the inner side of the cladding. The wall224


First-floor plan

Ground-floor plan

9.5 m

7.0 m

Sta

irwell

Timber structures

Fig. 6.53 Single-storeytimber skeleton-framestructure in which theprincipal elements areplyweb built-up-beamsections.

requiring very long elements of small cross-section, and they allow some degree of prefab-rication to be possible.

6.7.4 The skeleton frameThe fact that timber possesses tensile,compressive and bending strength makespossible the construction of skeleton frames inthe material (Fig. 6.53). A characteristic of thistype of structure, however, is that individualelements are subjected to fairly large amountsof internal force, due to the concentration ofthe load into a small volume of structure, and

while this can be accommodated easily withstrong materials, such as steel or reinforcedconcrete, it can be problematical with timber.

Where a skeleton frame is constructed intimber the whole of the structure will normallybe rather bulky, especially if floor loads arecarried. The structure must normally, therefore,be treated as one of the special architecturalfeatures of the building which it supports.

The floor- and roof-deck systems which areused are usually also of timber and of the one-way-spanning type. The frames are thereforebest planned on a rectangular grid of beams 225


Fig. 6.54 Beam grid for single-storey timber skeletonframe. Closely spaced joists (built-up-beam sections) arecarried on primary beams of relatively short span.

Fig. 6.55 Beam grid for single-storey timber skeletonframe. Strong primary structural elements (one-off trusses/portal frameworks) carry a secondary structure of purlins.

and columns. In floor structures the floor joistsshould span parallel to the long side of therectangle and the main beams parallel to theshort sides.

The most successful timber skeleton framesare single-storey structures carrying low levelsof imposed load (Fig. 6.53). Fairly long spansare possible with this type of arrangement,especially if built-up-beams or trusses are usedfor the principal elements, but the sizes of theelements are likely to be large compared toequivalents in steel.

Two strategies may be adopted for theplanning of single-storey frames (Figs 6.54 and6.55). In Fig. 6.54 the main structural elementsspan in the long direction in a rectangularcolumn grid and are spaced close together tominimise the load carried by individualelements, and to allow them to carry thecladding directly without the use of asecondary structure. The geometry of this typeof arrangement is similar to that of a steelframe with lightweight joists (Figs 3.25 and3.27). Columns are positioned at every third orfourth beam and intermediate members are

carried on 'primary' beams. The latter are moreheavily loaded than the main beams but canusually be of the same depth due to theirsmaller span. The exact configuration which isadopted in a particular case depends on thetype of beam being used.

The second type of arrangement is one inwhich the primary beams are placed at a widerspacing and a secondary structure, on whichthe cladding is mounted, is provided to spanbetween these (Fig. 6.55). The column gridcoincides with the primary beams, which mustbe much stronger elements than those in thesystem described previously. Proprietarybeams (at least in the simply supported form)are rarely suitable for this type of arrangementand large trusses are normally used. Built-up-beam-type cross-sections can be used asprimary elements if a 'strong' structural config-uration, such as a portal frame (Figs 6.3 and6.53) is adopted. This type of structure usuallyhas a plan grid which is similar to those whichare used in single-storey steel frames (Fig.3.30) but in the case of timber the primaryelements are usually placed closer together226

10-20m

3-5 m

Timber built-up-beam(e.g. Plyweb beam,Trussed rafter,Laminated beam)

Bolted truss orlaminated beam

15-30m

3-5 m

Purlins of sawn orlaminated timber

Primary member(bolted truss,built-up or laminatedportal frame)

(3.5 m to 6 m). Element sizes for two versionsof this type of frame are given in Table 6.14.

Where timber is used for multi-storey skele-ton-frame structures and must carry floorloading, the spans must be kept small -around 2 m to 3 m if simple sawn-timberelements are used for floor beams. Longerspans are possible with built-up-beamelements, such as laminated timber or plywebbeams, but the depths of these will be verylarge compared to equivalent steel orreinforced concrete structures (Fig. 6.1).

In multi-storey frames very simple grids,with columns at all grid points, are adoptedwhere solid timber elements are used, so as tominimise the loads which individual beamscarry (Fig. 6.56). Spaced beams are frequentlyadopted to increase the strength of individualelements and this allows longer spans to beachieved. Where built-up-beams are used,more complex primary/secondary-beamarrangements are possible. In all cases thebeam depths must be checked at an earlystage in the design as these are likely to becritical.

Another feature which is likely to be criticalin skeleton-frame design is the geometry of the

Fig. 6.56 An example of a beam layout which is suitablefor a multi-storey skeleton frame in timber. Primaryelements are spaced beams (double elements) and areconfined to short spans.

beam-to-column connections and somethought must normally be devoted to thiswhen the design is at a preliminary stage.Simple lap joints with bolted connections are 227

Table 6.14 Span range and dimensions of principal elements intimber trusses in single-storey skeleton frameworks

Span Truss Roof Approximate member size (mm)spacing pitch

(m) (m) Rafter Ceiling tie

10 2 20° 1 X 50 X 150S 1 X 50 X 100S

30° 1 X 50 X 125S 1 X 50 X 75S

15 3 20° 1 X 75 X 200S 1 X 75 X 150S

30° 1 X 75 X 150S 1 X 75 X 150S

20 4 20° 2 X 75 X 225S 1 X 75 X 225S

30° 2 X 75 X 175S 1 X 75 X 150S

30 6 20° 4 X 75 X 200S 3 X 75 X 200S

30° 3 X 75 X 125S 2 X 75 X 200S

40 6 20° 1 X150X 300L Steel tie rod

30° 1 X 150 X 250L Steel tie rod

50 6 20° 1 X 150 X 330L Steel tie rod

30° 1 X 150 x 300L Steel tie rod

60 6 20° 1 X 150 X 400L Steel tie rod

30° 1 X 150 X 330L Steel tie rod

Timber structures

S = sawn timberL = laminated timber

Squarecross-sectioncolumn

Doubled sawn-timberprimary beam

Sawn-timber joists

2m

4m

vertical-plane diaphragms. As with skeletonframes in other materials, the braced panelsshould be positioned as symmetrically aspossible on plan and the disposition of thevertical-plane bracing is therefore a factorwhich affects the internal planning of thebuilding.

6.7.5 Shells and other surface formsSurface structures are those in which theelements which form the skin of the structurehave an important, sometimes dominant,structural role. Stressed-skin panels (Fig 6.40)are in this category; folded forms and curvedshell structures are other examples and theseare described briefly here.

Folded forms are arrangements of flat struc-tural panels; there are two basic types, the'prismatic' type, which consists of flat-sidedcorrugations whose depth is constant acrossthe structure (Fig. 6.58) and more complextypes in which the depth of the corrugations isvaried. The latter are normally based on asystem of triangular panels. In both cases thepanels which form the structures usuallyconsist of a double skin of plywood sheetsseparated by solid timber stringers. Nailed andglued joints are used to assemble the panels,which are then bolted together to form thecomplete structure. The structures are highlyefficient, in the sense that they achieve a highstrength and stiffness for the amount of mater-ial which they contain, and they are thereforesuitable for long-span structures (up to 40 m)or for situations where self-weight must beminimised. They are sophisticated forms,however, which are difficult to design andconstruct and they are not in common use.

Timber shells, which are curved stressed-skin structures of the form-active type, are alsohighly efficient structures which achieve highstrength and stiffness for a given weight ofmaterial. Like folded forms they are difficult toconstruct and are not in common use. Unlikefolded forms, timber shells are usually solidand constructed in layers. Strips of thin board-ing are laid across a formwork or mould of therequired shape and several layers, which arenailed and glued together, are used. These are

normally used and, to avoid eccentricity in thetransfer of load between elements, a 'spaced'arrangement can be adopted for either thebeams or the columns (Fig. 6.57). In frameswith solid timber beams the strength of thebeams is usually a critical factor and so spacedbeams are employed, with single-elementcolumns. Where the horizontal elements aretrusses or built-up-beams a spaced columncan be used. Where tie-beams are required inthe direction at right angles to that of theprincipal members, these will frequently haveto be placed at a different level from the princi-pal elements, so as to allow sufficient spacefor the two sets of lap joints. This increases thetotal depth of the structural zone and can be acritical factor in the design.

Beam-to-column joints in timber frames arerarely rigid and vertical-plane bracing is there-fore required in skeleton frames. This can beprovided either in the form of diagonal bracing(usually of the tensioned-wire type) or bysuitably positioned wall panels which act as228


Fig. 6.57 Typical joint arrangements for timber skeletonframes. One of the features of this type of construction isthe space required for joints.

Timber structures

Fig. 6.58 Timberfolded-plate struc-ture. Forms such asthis, which areconstructed frompanels consisting oftwo skins of plywoodseparated by sawn-timber stringers,have good structuralproperties and allowlong spans to beachieved.

229


Fig. 6.59 Timber hyperbolic paraboloid shells are usedhere to form a roof structure of distinctive shape.

Fig. 6.60 Basic forms of the timber braced dome andlamella vault.

Fig. 6.61 Thebraced dome config-uration allows largeinteriors to beconstructed with avery efficient use ofstructural material.

230

Timber structures

placed in different directions so as to producea shell which has similar structural propertiesin all directions. The shell is normallysupported on edge beams of laminated timberwhich are integral with the shell itself. For easeof construction and analysis shapes whichhave a regular geometry are usually adopted.The hyperbolic paraboloid (Fig. 6.59) is particu-larly suitable but other forms, such as theconoid and the elliptical paraboloid, have alsobeen used.

6.7.6 Lattice domes and vaultsIn this form of construction triangulatedarrangements of timber elements are built upinto domed or vaulted shapes (Figs 6.60 to6.62). The resulting structures are veryefficient due to the form-active or semi-form-

Fig. 6.62 The lamella vault is a system which allowslarge interiors to be created. The degree of standardisationwhich is present can make this an economical form ofconstruction.

active overall geometry and are thereforesuitable for large enclosures (span range15 m to 200 m). In the lamella system eachelement is twice the length of the side of adiamond, and at each joint, one elementpasses continuously through, with adjacentintersecting elements connected to its mid-point. Triangulation of the geometry iseffected with purlins. The particular advan-tage of the lattice vault or dome is the highdegree of standardisation of the componentswhich is possible.

231

Selected bibliography

Addis, W.B., The Art of the Structural Engineer, Engel, H., Structure Systems, Stuttgart, 1967London, 1994 Engel, H., Structural Principles, Englewood Cliffs,

Ambrose, J., Building Structures, New York, 1988 NJ, 1984Balcombe, G., Mitchell's History of Building, Friedman, D, Historical Building Construction,

L o n d o n , 1 9 8 5 L o n d o n , 1 9 9 5Baird, J. A., Timber Specifier's Guide, Oxford, 1990 Gans, D. (Ed.), Bridging the Gap-. Rethinking theBaird, J. A. and Ozelton, E. C, Timber Designer's Relationship of Architect and Engineer, London,

Manual, 2nd edition, London, 1984 1991Benedetti, C. and Bacigulupi, V., Legno Gordon, J.E., The New Science of Strong Materials,

Architettura (with English text), Rome, 1991 Harmondsworth, 1968Benjamin, B. S., Structures for Architects, 2nd Gordon, J. E., Structures, Harmondsworth, 1978

edition, New York, 1984 Gorst, T., The Buildings Around Us, London, 1995.Bill, M., Robert Maillart, London, 1969 Hart, F., Henn, W. and Sontag, H., Multi-storeyBlanc, A., McEvoy, M. and Plank, R., Architecture Buildings in Steel, London, 1978

and Construction in Steel, London, 1993 Hartoonian, G., The Ontology of Construction,BS 449: Part 2: 1969: The Use of Structural Steel in Cambridge, 1994

Building, British Standards Institution, Milton Hayward, A. and Weare, F., Steel Detailer'sKeynes Manual, Oxford, 1989

BS 6399: Part 1: 1984, Design Loading for Buildings, Hodgkinson, A. (Ed.) AJ Handbook of BuildingCode of Practice for Dead and Imposed Loads, Structure, London, 1974British Standards Institution, Milton Keynes Jacobus, J., lames Stirling: Buildings and Projects,

Bull, J. W., The Practical Design of Structural New York, 1975Elements in Timber, 2nd edition, Aldershot, Jodidio, P., Contemporary European Architects,1994 Volume 3, Cologne, 1995

Curtin, W. G., Shaw, G., Beck, J. K. and Lambot, I. (Ed.), Norman Foster-Foster Associates-.Parkinson, G.I., Structural Masonry Designer's Buildings and Projects-. Volume 2 1971-78,Manual, London, 1982 London, 1989

Curtin, W. G., Shaw, G., Beck, J. K. and Bray, W. Lange, Basilica at Pompeii, Leipzig, 1885A., Structural Masonry Detailing, London, 1984 Macdonald, A. J., Structure and Architecture,

Curtis, W., Modern Architecture since 1900, Oxford, 1994London, 1987 Mainstone, R.J., Developments in Structural Form,

Desch, H. E..Timber (6th edition, revised London, 1975Dinwoodie, J. M.), London 1981 Mainstone, R.J., Haaia Sophia, London, 1988

Dowling, P. J., Knowles, P. and Owens, G. W., Mark, R, (Ed.), Architectural Technology up to theStructural Steel Design, London, 1988 Scientific Revolution, Cambridge, MA, 1993

Elliott, C. D., Technics and Architecture-. The Mettem, C. J., Structural Timber Design andDevelopment of Materials and Systems for Buildings, Technology, London, 1986London, 1992 Noever, P. (Ed.), The End of Architecture, Munich,

Elliott, K. S., Multi-storey Precast Concrete Frame 1993Structures, Oxford, 1996 Orton, A., The Way We Build Now, London, 1988 233



Pippard, A.J.S. and Baker, J., The Analysis ofEngineering Structures (4th edition), London,1984.

Reynolds, C. E. and Steedman, J. C, ReinforcedConcrete Designer's Manual, 10th edition,London,1988

Rice, P., AM Engineer Imagines, London, 1994Schodek, D. L., Structure in Sculpture,

Connecticut, 1993

Sunless, J. and Bedding, B., Timber inConstruction, London, 1985

Thornton, C, Tomasetti, R., Tuchman, J. andloseph, L., Exposed Structure in Building Design,New York, 1993

Wilkinson, C, Supersheds, Oxford, 1991Yeomans, D., The Trussed Roof: its History and

Development, Aldershot, 1992

234

The relationship between structural form andstructural efficiency

Appendix 1

Structural efficiency is defined here in terms ofthe weight of structural material which is usedto provide a given load-carrying capacity. Theefficiency will be taken to be high if thestrength-to-weight ratio of the structure is high.

The efficiency of a structure is affected by anumber of factors connected with its form andgeneral configuration. A very significant factoris the relationship between the form of a struc-ture and the type of internal force which occursfor a given pattern of loading. Only this aspectwill be discussed here.1

Elements in architectural structures arenormally subjected to axial internal force, tobending-type internal force or to a combin-ation of these. The distinction is an importantone so far as efficiency is concerned becauseaxial internal force can be resisted moreefficiently than bending-type internal force.The principal reason for this is that the distri-bution of stress which occurs within axiallyloaded elements is almost constant (Fig. Al.l)and this uniform level of stress allows all ofthe material in the element to be stressed toits limit - that is to a level which is slightlylower than the failure stress of the material. Anefficient use of material therefore resultsbecause all of the material present providesfull value for its weight. With bending stress,which varies in intensity in all cross-sections(Fig. Al.l) from a minimum at the neutral axisto a maximum at the extreme fibres, only themost highly stressed parts of each cross-section are used efficiently. Most of the mater-ial present is under-stressed and thereforeinefficiently used.

Fig. Al.l (a) Elements which carry purely axial load aresubjected to axial stress the intensity of which is constantacross all cross-sectional planes,(b) Pure bending-type load (i.e. load which is normal tothe axis of the element) causes bending stress to occur onall cross-sectional planes. The magnitude of this varieswithin each cross-section from a maximum compressivestress at one extremity to a maximum tensile stress at theother.

1 For a more wide-ranging discussion of structuralefficiency see Macdonald, Structure and Architecture, op. cit.

The type of internal force which occurs in anelement depends on the relationship betweenthe direction of its principal axis (its longitu-dinal axis) and the direction of the load which 235

Axial load

(a)

(b)

Axial stress

Bending-type load

Compressive

Tensile

Bending stress


is applied to it (Fig. A1 .2). If an element is. straight, axial internal force occurs if the load is

applied parallel to the longitudinal axis of theelement. Bending-type internal force occurs if itis applied at right angles to the longitudinalaxis and combined axial and bending-typeinternal forces occur if it is applied in a direc-tion which is inclined to the longitudinal axis.The axial-only and bending-only cases can beregarded as special cases of the more generalcombined case, but they are in fact the mostcommonly found types of loading arrangementin architectural structures.

If an element is not straight then it willalmost inevitably be subjected to a combin-ation of axial and bending internal forces whena load is applied but there are importantexceptions to this as is illustrated in Fig. A1.3.Here the structural element consists of a flex-ible cable, supported at its ends, and fromwhich various loads are suspended. Becausethe cable has no rigidity it is incapable ofcarrying any other type of internal force butaxial tension; it is therefore forced into a shapewhich allows it to resist the load with an inter-nal force which is pure axial tension. Theshape traced by the longitudinal axis is uniqueto the load pattern and is called the form-activeshape for that load.

As is seen in Fig. A1 .3 the shape which thecable adopts is dependent on the pattern ofload which is applied; the form-active shape isstraight-sided when the loads are concentratedat individual points and curved if the load isdistributed along it. If a cable is allowedsimply to sag under its own weight, which is adistributed load acting along its entire length,it adopts a curve known as a catenary.

An interesting feature of the form-activeshape for any load pattern is that if a rigidelement is constructed whose longitudinal axisis the mirror image of the form-active shapetaken up by the cable, then it too will besubjected only to axial internal forces when thesame load is applied, despite the fact that,being rigid, it could also carry bending-typeinternal force. In the mirror-image form all theaxial internal forces are compressive(Fig. A1.4).

Fig. A1.2 Basic relationships between loads and struc-tural elements.(a) Load coincident with the principal axis; axial internalforce only.(b) Load perpendicular to the principal axis; bending-typeinternal force.(c) Load inclined to the principal axis; combined axial andbending-type internal force.236

(a)

(b)

(c)

The cable structure and its rigid 'mirror-image' counterpart are simple examples of awhole class of structural elements which carryaxial internal forces only, due to their longitu-dinal axes being shaped to the form-activeshapes for the loads which are applied tothem. These are called form-active elements.

If, in a real structure, a flexible material suchas steel wire or cable is used to make anelement, it will automatically take up the form-active shape when the load is applied. Flexiblematerial is in fact incapable of becominganything other than a form-active element. Ifthe material is rigid, however, and a form-active element is required, then it must bemade to conform to the form-active shape ofthe load which is to be applied to it or, in thecase of a compressive element, to the mirrorimage of the form-active shape. If not, theinternal force will not be pure axial force andsome bending will occur.

Figure AI.5 contains a mixture of form-activeand non-form-active shapes. Two load patternsare shown: a uniformly distributed load acrossthe whole of the element and two concentratedloads applied at equal distances across them.For each load, elements (a) carry purebending-type internal forces; no axial force canoccur in these because there is no componentof either load which is parallel to the axis ofthe element. The elements in (b) have shapeswhich conform exactly to the form-activeshapes of the loads. They are therefore form-active elements which carry axial internalforces only; in both cases the forces arecompressive. The elements (c) do not conformto the form-active shapes of the loads and willnot therefore carry pure axial internal force.Neither will they be subjected to pure bending;they will carry a combination of bending andaxial internal force.

So far as the shape of their longitudinal axesare concerned, structural elements can thus beclassified into three categories: form-activeelements, non-form-active elements and semi-form-active elements. Form-active elementsare those which conform to the form-activeshape of the loads which are applied to themand they contain axial internal forces only.

Fig. A1.3 Tensile form-active shapes. Because it has norigidity a cable must take up a shape - the form-active shapewhich allows it to resist the load with a purely tensileinternal force. Different load arrangements produce differ-ent form-active shapes.

Fig. A1.4 Compressive form-active shapes.

Fig. A1 .5 Examples of the relationship between elementshape, load pattern and element type. The latter is deter-mined by the relationship between the shape of theelement and the form-active shape for the load patternwhich it carries.(a) Non-form-active (bending stress only).(b) Form-active (axial stress only).(c) Semi-form-active (combined bending and axial stress). 237

Appendix 1

(a) (b) (c)

Non-form-active elements are those whose active element when subjected to the uniformlylongitudinal axis does not conform to the form- distributed load (Al ,5c).active shape of the loads and are such that no Because they are not subjected to bendingaxial component of internal force occurs. These stress, elements with form-active shapes arecontain bending-type internal force only. Semi- potentially the most efficient types of struc-form-active elements are elements whose ture. Non-form-active elements are potentiallyshapes are such that they contain a combina- the least efficient because bending stress istion of bending and axial internal forces. The the principal type to which they are subjected,cranked-beam shape in Fig. Al .5 is a fully-form- The efficiency of semi-form-active elementsactive element when subjected to the two depends on the extent to which they differconcentrated loads (A1.5b) but a semi-form- from the form-active shape.


238

Approximate methods for allocating sizes tostructural elements

Appendix 2

A2.1 Introduction

The methods which are outlined here for deter-mining the sizes of structural elements are notthose of rigorous final structural design calcu-lations. Rather they are quick methods whichallow the sizes of structural elements to bedetermined approximately. They are prelim-inary planning tools which are sufficientlyaccurate, in most cases, to allow the feasibilityof proposed structural arrangements to beassessed. They are applicable to themainstream forms of structure.

One of the main objectives of structuraldesign is to produce structures which are safeand serviceable at reasonable cost. Thisrequires that the sizes of the cross-sections ofstructural elements be sufficiently, but onlysufficiently, large to carry safely the loadswhich are applied to them. The principalmechanism by which this aspect of structuraldesign is controlled is through calculations.

A2.2 Structural analysis

The principal factor which determines the sizerequired for a structural element is the amountof load which it carries and the element-sizingpart of structural calculations must normallytherefore be preceded by an assessment ofthis. The process by which it is done is knownas the analysis of the structure. Even approxi-mate element-sizing calculations require thatsome form of rudimentary structural analysisbe carried out.

Structural analysis can be subdivided intothe three distinct processes of load assessment,preliminary analysis and final analysis. Load

assessment involves the prediction of themaximum load which will occur on the struc-ture in its lifetime. In the case of architecturalstructures there are three principal types ofload: these are dead load - the permanent gravi-tational load caused by the weight of thebuilding and its fixtures; gravitational imposed load- variable load caused by the weights of theoccupants of the building, furniture and othermoveable items; and wind loading - non-gravita-tional imposed load caused by the action ofwind pressure. In the case of most structuralelements the most unfavourable load groupingis the combination of dead and imposed gravi-tational load. The approximate element-sizingcalculations which are presented here willtherefore be based principally on this loadcombination. The figures which are specified inthe current British Standard for dead andimposed gravitational loads are given in TablesA2.1 and A2.2.

The preliminary analysis of a structure is aprocess in which the three-dimensional objectwhich is the structure is broken down into itsconstituent elements so that the forces whichare imposed on the elements under the actionof the maximum load condition can be deter-mined. It involves the tracing of the path whichis taken through the structure by the load fromthe floor and roof surfaces, to which it isapplied initially, to the foundations, where it isultimately resisted. Some structural elements,such as floor slabs, are acted on directly by theloads. Others, such as columns, receive loadfrom the parts of the structure which theysupport. The end product of the preliminaryanalysis is a set of individual structuraldiagrams (one for each element) on which theforces which act on each element are marked. 239


The final part of the analysis of a structureconsists of the evaluation of the internal forcesin the individual elements - the bendingmoments, shear forces and axial forces whichact on their cross-sections. It is the magni-tudes of these which determine the shapes andsizes which are required for the cross-sections.The analytical technique by which they arecalculated is the device of the 'imaginary cut'through the structure, which is used inconjunction with the concept of equilibrium

and the 'free-body-diagram' to evaluate theinternal forces (see Macdonald, Structure andArchitecture, Chapter 2 and Appendix 1, for anexplanation of these terms).

In the context of approximate calculationsfor preliminary element sizing a much abbrevi-ated version of structural analysis can becarried out. Often, the requirement is to deter-mine approximately the sizes of the criticalelements only (e.g. the most heavily loadedelement in a timber roof truss or the most240

Table A2.1 Weights of building materials (based on BS 648 [1 kg = 10 N (approx)])

Asphalt Piaster

Roofing 2 layers, 19 mm thick 42 kg/m2 Two coats gypsum, 13 mm thick 22 kg/m2

Damp-proofing, 19 mm thick 41 kg/m2

Road and footpaths, 19 mm thick 44 kg/m2 Plastics sheetingCorrugated 4.5 kg/m2

Bitumen roofing felts

Mineral surfaced bitumen per lay 3.5 kg/m2 Plywoodper mm thick 0.7 kg/m2

Blockwork

Solid per 25 mm thick, stone aggregate 55 kg/m2 Reinforced concrete 2400 kg/m3

Aerated per 25 mm thick 15 kg/m2

Rendering

Board Cement:sand (1:3) 13 mm thick 30 kg/m2

Blockboard per 25 mm thick 12.5 kg/m2

ScreedingBrickwork Cement:sand (1:3) 13 mm thick 30 kg/m2

Clay, solid per 25 mm thick, medium density 55 kg/m2

Concrete, solid per 25 mm thick 59 kg/m2 Slate tiles(depending upon thickness and source) 24-78 kg/m2

Cast stone 2250 kg/m3

Steel

Concrete Solid (mild) 7850 kg/m3

Natural aggregates 2400 kg/m3 Corrugated roofing sheets per mm thick 10 kg/m2

Lightweight aggregates (structural) 1760 kg/m3

+240or- l60 Tarmacadam25 mm thick 60 kg/m2

Flagstones

Concrete, 50 mm thick 120 kg/m2 Terrazzo25 mm thick 54 kg/m2

Glass fibreSlab, per 25 mm thick 2.0-5.0 kg/m2 Tiling, roof

Clay 70 kg/m2

Gypsum panels and partitions

Building panels 75 mm thick 44 kg/m2 TimberSoftwood 590 kg/m3

Lead Hardwood 1250 kg/m3

Sheet, 2.5 mm thick 30 kg/m2

Water 1000 kg/m3

Linoleum

3 mm thick 6 kg/m2 WoodwoolSlabs, 25 mm thick 15 kg/m2

Table A2.2 Minimum imposed floor loads (based on BS 6399 Part 1 1996)

Type of activity/occupancyfor part of the building orstructure

A Domestic andresidential activities(Also see category C)

B Offices and workareas not coveredelsewhere

Examples of specific use

All usages within self-contained dwelling unitsCommunal areas (including kitchens) inblocks of flats with limited use (See note 1)(For communal areas in other blocks of flats,see C3 and below)

Bedrooms and dormitories except those inhotels and motels

Bedrooms in hotels and motelsHospital wardsToilet areas

Billiard rooms

Communal kitchens except in flats covered bynote 1

Balconies Single dwellingunits andcommunal areas inblocks of flats withlimited use (Seenote 1)

Guest houses,residential clubsand communalareas in blocks offlats except ascovered by note 1

Hotels and motels

Operating theatres, X-ray rooms, utility rooms

Work rooms (light industrial) without storage

Office for general use

Banking halls

Kitchens, laundries, laboratories

Rooms with mainframe computers or similarequipment

Machinery halls, circulation spaces therein

Projection rooms

Factories, workshops and similar buildings(general industrial)

Foundries

Catwalks

Balconies

Fly galleries

Ladders

Uniformly distributed loadkN/m2

1.5

1.5

2.0

2.0

3.0

1.5

Same as rooms towhich they give accessbut with a minimum of3.0

Same as rooms towhich they giveaccess but with aminimum of 4.0

2.0

2.5

2.5

3.0

3.0

3.5

4.0

5.0

5.0

20.0

—

Same as rooms to whichthey give access butwith a minimum of 4.0

4.5 kN/m run distributeduniformly over width

—

Concentrated loadkN

1.4

1.8

1.8

2.7

4.5

1.4

1.5/m runconcentrated at theouter edge


4.5

1.8

2.7

2.7

4.5

4.5

4.5

To be determined forspecific use

4.5

To be determined forspecific use

1.0 at 1 m centres

1.5/m run concentratedat the outer edge

—

1. 5 rung load

Appendix 2

241


242.

Table A2.2 Selected gravitational imposed loads [continued)

Type of activity/occupancy

for part of the building or

structure

C Areas where peoplemay congregate

Cl Areas with tables

C2 Areas with fixedseats

C3 Areas withoutobstacles for movingpeople


Public, institutional and communal dining roomsand lounges, cafes and restaurants (See note 2)

Reading rooms with no book storage

Classrooms

Assembly areas with fixed seating(See note 3)

Places of worship

Corridors, hallways,aisles, stairs, landingsetc. in institutional typebuildings (not subject tocrowds or wheeledvehicles), hostels, guesthouses, residential clubs,and communal areas inblocks of flats notcovered by note 1. (Forcommunal areas in blocksof flats covered by note 1,see A)

Corridors, hallways,aisles, stairs, landing, etc.in all other buildingsincluding hotels andmotels and institutionalbuildings

Corridors,hallways, aislesetc. (foot trafficonly)

Stairs andlandings (foottraffic only)

Corridors, hall-ways, aisles etc.(foot traffic only)

Corridors, hall-ways, aisles etc.,subject towheeled vehicles,trolleys etc.

Stairs and land-ings (foot trafficonly)

Industrial walkways (light duty)

Industrial walkways (geneal duty)

Industrial walkways (heavy duty)

Museum floors and art galleries for exhibitionpurposes

Balconies (except as specified in A)

Fly galleries

Uniformly distributed

load

kH/m2

2.0

2.5

3.0

4.0

3.0

3.0

3.0

4.0

5.0

4.0

3.0

5.0

7.5

4.0

Same as rooms towhich they giveaccess but with aminimum of 4.0

4.5 kN/m rundistributed uniformlyover width

Concentrated load

kN

2.7

4.5

2.7

3.6

2.7

4.5

4.0

4.5

4.5

4.0

4.5

4.5

4.5

4.5


Table A2.2 Selected gravitational imposed loads (continued)

Type of activity/occupancy

for part of the building or

structure

C4 Areas withpossible physicalactivities (See clause9)

C5 Areas susceptibleto overcrowding (See

clause 9)

D Shopping areas

E Warehousing andstorage areas. Areas

subject toaccumulation ofgoods. Areas forequipment and plant.

F

G


Dance halls and studios, gymnasia, stages

Drill halls and drill rooms

Assembly areas without fixed seating, concerthalls, bars, places of worship and grandstands

Stages in public assembly areas

Shop floors for the sale and display ofmerchandise

General areas for static equipment not speci-fied elsewhere (institutional and public buildings)

Reading rooms with book storage, e.g. libraries

General storage other than those specified

File rooms, filing and storage space (offices)

Stack rooms (books)

Paper storage for printing plants andstationery stores

Dense mobile stacking (books) on mobiletrolleys, in public and institutional buildings

Dense mobile stacking (books) on mobiletrucks, in warehouses

Cold storage

Plant rooms, boiler rooms, fan rooms, etc.,including weight of machinery

Ladders

Parking for cars, light vans, etc. not exceeding2500 kg gross mass, including garages,driveways and ramps

Vehicles exceeding 2500 kg. Driveways, ramps,repair workshops, footpaths with vehicleaccess, and car parking

Uniformly distributed

load

kN/m2

5.0

5.0

5.0

7.5

4.0

2.0

4.0

2.4 for each metre ofstorage height

5.0

2.4 for each metre ofstorage height butwith a minimum of 6.5

4.0 for each metre ofstorage height




7.5

—

2.5

Concentrated load

kN

3.6

9.0

3.6

4.5

3.6

1.8

4.5

7.0

4.5

7.0

9.0

7.0

7.0

9.0

4.5

1.5 rung load

9.0

To be determined for specific use

Appendix 2

243

NOTE 1. Communal areas in blocks of flats with limited use refers to blocks of flats not more than three storeys in heightand with not more than four self-contained dwelling units per floor accessible from one staircase.NOTE 2. Where these same areas may be subjected to loads due to physical activities or overcrowding, e.g. a hotel diningroom used as a dance floor, imposed loads should be based on occupancy C4 or C5 as appropriate. Reference should

also be made to clause 9 (BS 3699 Part 1: 1996)NOTE 3. Fixed seating is seating where its removal and the use of the snare for other purposes is improbable

Table A2.3 Bending moment and deflection formulae for beams (after A) Handbook of BuildingStructure)

The following tabulates formulae and values of moments (M), reactions (R), shear force (S) and deflection (8) in beams for a number ofcommon loading and support conditions. It covers cantilevers, free support beams, fixed-end beams, and propped cantilevers


244

245

Appendix 2


246

247

Appendix 2


248

heavily loaded beam in a floor) in order thatthe feasibility of a proposal can be tested orthe depth of the structural zone required for afloor estimated. In the case of elements loadedin bending, Table A2.3 can be used to estimatethe maximum bending moment involved.

The loading on critical elements can beestimated from a judgement of the total areaof roof or floor which they support. These areasare then simply multiplied by the intensity ofgravitational load to give the load on the struc-tural element concerned.

Basic versions of element-sizing calculationsare normally based on either the permissiblestress or the load factor method of design. In thepermissible stress method the failure stress ofthe material is divided by the factor of safety togive a permissible stress and the calculationsare used to determine the sizes of cross-sections required to ensure that the actualstress in the structure, when the peak load isapplied, is never greater than the permissiblestress.

In the load factor method the peak loadvalue is multiplied by the factor of safety togive a factored design load which is then used,in conjunction with the actual failure stress ofthe material (normally taken to be the yieldstress), to determine safe values for the sizesof cross-sections (i.e. values which ensure thatthe structure will have a margin of strengthover that which is required to resist the peakvalues of the applied loads).

In some present-day final design calcula-tions it is normal to split the factor of safetyinto two or more 'partial factors of safety'which are applied separately to load andmaterial strength values to take account of thevarying degrees of precision with which thesecan be known. The relative advantages anddisadvantages of the different approaches tothe incorporation of the factor of safety intostructural calculations will not be discussedhere. All of the approximate sizing calculationspresented here are based on the permissiblestress method, which is the simplest to applyin practice.

A2.3.2 Elements subjected to axial tensionElements which are subjected to axial tensionare normally constructed either of steel or oftimber. The axial tensile stress in the elementis normally considered to be uniformly distri-buted across the cross-section and is calcu-lated from the equation,

fat = P/A (A2..1)

where: fat = axial stressP = applied axial forceA = area of cross-section

A2.3 Element-sizing calculations

A2.3.1 IntroductionStructural elements must be of adequatestrength and must not undergo excessivedeflection under the action of load. These aredistinct requirements and either one of themmay be the critical factor which determines thesize of cross-section which must be adopted.The calculations which are outlined here arealmost exclusively strength calculations. In themajority of cases these are sufficient for thepurpose of determining approximately the sizeof cross-section required for a structuralelement. The basic principles and elementaryforms of element-sizing calculations areoutlined in this section. The modified versionsof these which are used for the principal struc-tural materials, and which allow for thepeculiarities of different materials, arediscussed in Sections A2.4 to A2.7

Element-sizing calculations must have afactor of safety incorporated into them to allowfor the uncertainties which are inevitablypresent in the design and constructionprocesses. The agencies which give rise tothese inaccuracies include the imprecisionwith which loads and material strengths can beknown, the inaccuracies which are present inthe calculations themselves and the discrepan-cies which occur between the sizes andstrengths which are specified for structuralelements and those which are actually realisedin the finished structure. 249

Appendix 2


If the size of cross-section does not vary alongthe length of an element the magnitude of thestress is the same at all locations.

The size of cross-section which is requiredfor a particular tensile element is calculatedfrom the following variation of the aboveequation:

Areq = P/fatp (A2.2)

where:Areq = area of cross-section requiredP = applied axial loadf a t p = maximum permissible axial

tensile stress

The area of cross-section which is required inpractice is normally slightly larger than thatgiven by (A2.2) due to complications with theend connection.1 In particular, the need toallow for the effect of stress concentrationsaround fastening elements, such as bolts andscrews, and for any eccentricity which may bepresent in the connection (such as might occurwhere a steel angle element is connectedthrough one leg only) normally require that aslightly larger cross-section be adopted than isgiven by equation (A2.2). For the purpose ofapproximate element sizing the adoption ofthe very crude device of simply increasing thesize of the cross-section calculated fromequation (A2.2) by 15% will normally give asatisfactory result.

A2.3.3 Elements subjected to bending

A2.3.3.1 Calculation of bending stressBending stress occurs in an element if theexternal loads cause bending moment to acton its cross-sections. The magnitude of thebending stress varies within each cross-sectionfrom peak values in tension and compressionin the extreme fibres on opposite sides of thecross-section, to a minimum stress in the

1 The making of satisfactory tensile connections is one ofthe classic problems of structural engineering. SeeGordon, |. E., Structures, Harmondsworth, 1978,Section 2 for a discussion of this.250

Fig. A2.1 Distribution of bending stress. Where bending-type load is present the resulting bending stress on eachcross-section varies from maximum tension on one side tomaximum compression on the other.

centre (at the centroid) where the stresschanges from compression to tension (Fig.A2.1). It will normally also vary between cross-sections due to variation in the bendingmoment along the length of the element.

The magnitude of bending stress at anypoint in an element depends on the followingfour factors: the bending moment at the cross-section in which the point is situated; the sizeof the cross-section; the shape of the cross-section; and the location of the point withinthe cross-section. The relationship betweenthese parameters is,

fby = My/I (A2.3)

where fby = bending stress at a distance yfrom the neutral axis of the cross-section (the axis through thecentroid)

M = bending moment at the cross-section

I = the second moment of area of thecross-section about the axisthrough its centroid; this dependson both the size and the shape ofthe cross-section.

This relationship allows the bending stress atany location in any element cross-section to becalculated from the bending moment at thatcross-section. It is equivalent to the axialstress formula fa = P/A.

Neutral axis

High compressive stress

High tensile stress

Equation (A2.3) is called the elastic bendingformula. It is only valid if the peak stress iswithin in the elastic range. It is one of themost important relationships in the theory ofstructures and it is used, in a variety of forms,in the design calculations of all structuralelements which are subjected to bending-typeloads.

For the purpose of calculating the maximumbending stress, which occurs at the extremefibres of the cross-section, equation (A2.3) isfrequently written in the form,

f b y m a x = M/Z (A2.4)

where: Z = I/ymax

Z is called the modulus of the cross-section.(It is often referred to as the 'section modulus';sometimes the term 'elastic modulus' is usedand this is unfortunate because it leads toconfusion with the term modulus of elasticity.)If the cross-section of an element is notsymmetrical about the axis through itscentroid the maximum stresses in tension andcompression are different. Where this occurstwo section moduli are quoted for the cross-section, one for each value of ymax.

A2.3.3.2 Calculation of shear stressShear stress acts on the cross-sectionalplanes of bending elements due to thepresence of shear force. The distribution ofshear stress within a cross-section is notuniform (the pattern of distribution dependson the shape of the cross-section) butnormally only the average value of shearstress is calculated.

average shear stress = shear force/area ofcross-section whichresists shear

v = V/Av (A2.5)

In the case of a rectangular cross-section thearea which resists shear is the total area of thecross-section. For I- and box-sections the areaof the web only is used.

A2.3.3.3 Approximate sizing of bending-typeelementsBending-type elements are subjected to bothbending and shear stress and the size of crosssection which is adopted must be such thatneither is excessively large. Normally, thebending strength criterion is used for initialselection of the cross-section size and thechosen section is then checked to ensure thatit will be satisfactory in respect of shear.

Appendix 2

Fig. A2.2 Stresses in compression elements.(a) If the element is straight and perfectly aligned with theload the stress in each cross-section is axial.(b) If the element has a slight curvature the eccentricitywhich is present gives rise to bending stress and the totalstress is a combination of this and axial stress. 251

Bending stress dueto eccentricityAxial stress due tocompressive load

Combinedstress

Maximumstress

Axialstress


Initial selection of section size can be basedon the following version of the elastic bendingformula:

Zreq = M/fbp (A2.6)

where: Zreq = required modulus of sectionM = maximum applied bending

momentfbp = permissible bending stress

A2.3.4 Sizing of elements subjected to axialcompressionCompression elements are the most problem-atic to size due to the need to allow for thephenomenon of buckling. Compressive forcesare inherently unstable because if any eccen-tricity is present in a compressive system theaction of the forces causes the amount ofeccentricity to increase. For example, in Fig.A2.2 the internal forces in a simple compres-sive element are exposed by use of the deviceof the imaginary cut. It can be seen from (a)that the internal force is one of pure axialcompression if the element is perfectlystraight; an axial stress is produced which isevenly distributed across the cross-section andthe system is in a state of equilibrium. It is,however, potentially unstable.

If the element is given slight curvature, as in(b), which might occur due to the presence of asmall lateral disturbance, a couple is producedby the misalignment of the internal axialforces. This causes a bending moment to acton each cross-section. The bending strainwhich the curvature generates producesbending stress which resists the bendingmoment and which tends to restore theelement to its original straight condition.Unlike in a beam, however, the bendingmoment and the bending stress which occurwhen curvature is introduced into a compres-sive element are not directly related. Thebending moment is dependent solely on themagnitude of the applied compressive forceand the amount of eccentricity which ispresent. The bending stress is determined bythe bending strain (dependent on the amountof curvature which has developed) and on the

Fig. A2.3 Stability of a 'perfect' strut (i.e. a compressionelement which is straight initially and perfectly alignedwith the load). The condition with respect to stabilitydepends on the level of applied axial load. If this is small,as in the first diagram, the strut is stable and will return tothe original condition following a disturbance. If the axialload is high the strut is unstable and will buckle ifdisturbed. The third diagram depicts the situation whenthe 'critical' level of load, which occurs at the transitionbetween these two load ranges, is applied. When the criti-cal load is applied the strut remains in the displacedposition following a disturbance.252

Axiallyloaded

strut

Stable state

Unstable state

Critical state

P large

P small

P critical Strut remainsin disturbedposition

Bucklingfailure

Strut returnsto originalstate

Disturbingforce

removed

properties of the element itself (specifically, onthe geometric properties of its cross-sectionand on the modulus of elasticity of the mater-ial). For a particular amount of curvature (i.e.of bending strain), the size of the restoringcouple generated by the bending stress isalways the same, but the size of the disturbingcouple caused by the eccentricity depends onthe magnitude of the compressive load.Compressive elements are therefore potentiallyunstable internally depending on the relation-ship between these couples. The bending typefailure which results from this type of instabil-ity is known as buckling.

The critical factor in determining the suscep-tibility of a particular element to buckling, isthe magnitude of the applied compressive load(Fig. A2.3). If this is small the disturbing couplewill also be small, even if the eccentricity islarge, and the restoring couple will increase ata faster rate than the disturbing couple if somesideways-acting external agency destroys theoriginal straight alignment of the element. Thesystem is stable because the restoring couplewill always be able to return the element to itsoriginal straight condition when the externalagency is removed. If the compressive load islarge, however, this will produce a disturbingcouple which is greater than the restoringcouple at all levels of curvature and theelement will be unstable because any externalagency which introduces a small amount ofcurvature will precipitate a progressive increasein curvature until buckling failure occurs. For aparticular compressive element the magnitudeof the compressive load at which instabilitydevelops is known as the critical load [Pcr].

The analysis of buckling is one of the classicproblems of structural design and manymathematicians and engineers have investi-gated methods for predicting the critical loadsof structural components. Perhaps the bestknown of these is due to the eighteenth-century Swiss mathematician Leonhard Eulerand, although the formula which Euler derivedfor the calculation of critical loads is notsuitable for most practical designs, it is never-theless described here because the study of itprovides a good introduction to the factors on

which the stability of compressive elementsdepends.

Eider's analysis of buckling

Euler's analysis, which is described in detail inPippard and Baker, The Analysis of Engineering

Structures (4th Edition, Edward Arnold, London,1984), is a theoretical investigation of a perfectstrut, that is of a compressive element which isperfectly straight initially and in which noeccentricity is present, either in the elementitself or in the application of the load. It yieldsthe following formula for the critical load of aperfect strut with hinged end connections,

Pcr = π2EI/L2 (A2.7)

where Pcr = the Euler critical loadE = the modulus of elasticity of the

material/ = the second moment of area of

the cross-section of the strutL = the length of the strut

The Euler critical load for an ideal strut isequivalent to the buckling strength of a realstrut. In the case of the ideal strut, a curvaturemust be deliberately introduced to causebuckling failure and the compressive loadconcerned must be greater than the criticalload. Eccentricity is always present in a realstrut, however, and so if a compressive loadgreater than the critical load is applied to itthe strut will automatically fail by buckling.

It will be seen that Euler identified theslendemess of a compressive element as themost significant factor which determines thecritical load, with slenderness being defined, inthe most basic version of the formula as L2/I. Themore slender the element, i.e. the higher theslenderness ratio, the smaller is the critical load.

It is significant that the quantity by which the'thickness' of the element is judged in the deter-mination of its slenderness is the secondmoment of area of the cross-section (/). This is,of course, a measure of the bending perform-ance of the cross-section and its use in thiscontext of compressive stability is not surprisingbecause buckling is a bending phenomenon.

Appendix 2

253


254

(A2.8)

Fig. A2.4 If the conditions of restraint are the same in allplanes a strut will buckle about its weakest axis. Thus, theelement with the rectangular cross-section is less stable incompression than that with the square cross-section of thesame total area.

/ = r2 A

where: r = radius of gyrationA = area of cross-section

The effect of second moment of area on thebuckling characteristics of an element can beseen by considering the behaviour of elementswith different shapes of cross-section. If theconditions of lateral restraint are the same inall directions and an increasing amount of axialload is applied to an element, it will alwaysbuckle about the axis through the centroid ofits cross-section about which the bendingstrength is least. This is the axis which givesthe smallest second moment of area. In Fig.A2.4, for example, the strut with the rectangularcross-section will fail by compressive instabilityat a lower load than that with the square cross-section, even though their total cross-sectionalareas are the same, because the secondmoment of area of the rectangular cross-section, about one axis through its centroid, isvery small. The fact that the second momentsof area of its cross-section about other axes arelarger than those of the strut with the squarecross-section does not affect its compressivestrength since it buckles about its weakest axis.

Because the quantity second moment ofarea must always be calculated with respect toa particular axis through the cross-section ofan element, the critical load which is calcu-lated by the Euler formula applies to bucklingin a particular plane. This is the plane which isnormal to the axis which is used to calculate /(Fig. A2.4). For a given element cross-section, anumber of different critical loads can often becalculated from the Euler formula dependingon the axis which is chosen for the / value.Each relates to a particular plane of buckling. Ifthe conditions of lateral restraint are the samefor all possible planes of buckling, the truecritical load of the element is the value whichis calculated from the smallest value of / of thecross-section. If the conditions of lateralrestraint are not the same for all bucklingplanes the plane for which the ratio of L/l isgreatest determines the critical load.

The property of the cross-section of acompressive element which is normally usedto gauge the critical load is not in fact thesecond moment of area but a related quantitycalled the radius of gyration. This is defined bythe equation,

Weak axis

Strong axis

Axis from which/ calculated.

Plane to which assessment ofbuckling based on / relates

The radius of gyration is therefore given by,

(A2.9)

The substitution of I = r2 A is normally made inthe Euler buckling formula which thenbecomes,

Pcr = π2E[r/L]2

This can be rearranged into the form,

Pcr/A = π2E[r/L]2

fcr = π2E[r/L]2 (A2.10)

The introduction of 2A into the formulainstead of I therefore allows the critical load tobe expressed in terms of a critical averagestress. This is a more convenient form for usein design.

Design of compression elements to resist bucklingThe relationship between critical average stressand slenderness ratio, which is expressed inequation (A2.10), is shown in the form of asketch graph in Fig. A2.5. This can be used asthe basis of a design method for compressionelements.

Fig. A2.5 The relationship between critical load andslenderness ratio. The more slender the element the loweris the critical load. The graph shows the relationship aspredicted by the Euler formula (ideal strut) and gives anindication of the type of relationship which occurs inpractice and which can be derived empirically. The discrep-ancy is due to the invalidity in practice of one of theassumptions on which the Euler formula is based, namelythat all of the material is stressed within the elastic range.

The design of a compression element is amatter of determining a cross-section whosesize and shape are such that its bucklingstrength is greater than the compressive loadwhich will be applied to it. The followingprocedure can be used in conjunction withversions of Fig. A2.5 for particular materials toachieve this.

1 The compressive force to be carried is deter-mined from the analysis of the structure andthe effective length of the strut judged fromits actual length and proposed end condi-tions (see below for an explanation of effec-tive length and the importance ofconsidering end conditions).

2 A trial size and shape of cross-section areselected.

3 From the properties of the trial cross-section and the effective length of the strutthe slenderness ratio is calculated and theEuler formula (or graph) is used to calculatethe magnitude of the average compressivestress at the critical load value. This is thevalue of the compressive stress which mustnot be exceeded if buckling failure is to beavoided.

4 The magnitude of the average compressivestress which will actually occur in the strutis calculated from the applied compressiveforce and the cross-sectional area of thetrial section.

5 If the relationship between the actual stressand the critical stress is not satisfactory theproperties of the cross-section are amendedand the above sequence repeated. In theinterests of safety and economy it is desir-able to achieve a cross-section which resultsin the actual stress being slightly smallerthan the critical stress but not excessivelyso.

The cyclic process is continued until a satisfac-tory cross-section is achieved.

Slenderness ratioThe critical load of a compressive element isaffected by its length and the fact that thelength term L appears in the lower part of the 255

Appendix 2

Ideal strut

Slenderness ratio

Criticalload

Real strut


Euler formula indicates that, as would beexpected, the longer an element is the smalleris the load which it can safely carry incompression. For a particular element thelength which must be used in the formula isthe distance between the points at which it isrestrained against lateral movement, which isfrequently different from the total length of theelement. Another factor which is significant isthat, in real structure, the conditions of lateralrestraint are frequently different for differentplanes of buckling (Fig. A2.6). In such cases thevalue which is used for length in the bucklingformula must be compatible with the radius ofgyration which is used as both values mustapply to the same plane of buckling.

A number of different slenderness ratios cannormally be calculated for a particular elementdepending on the conditions of lateralrestraint which are provided and on the shapeof the cross-section. Each refers to a particularplane of buckling - the plane which is normalto the axis from which the radius of gyrationwas calculated. The highest slenderness ratiois the one which determines the bucklingstrength.

The concept of effective lengthThe characteristics of its end conditions affectthe critical load of a structural element. Anelement which has its ends fully restrainedagainst rotation will carry a higher compressiveload before buckling than one whose ends arehinged, because the buckled shape of thefixed-ended element is more complex thanthat of one with hinged ends and a greaterload is required to force the element into thisshape. An element with one end fixed and theother end hinged has a simpler buckled shapethan the element which is fixed at both endsand its buckling strength is therefore inter-mediated between the other two cases.

The Euler buckling formula applies only tocompressive elements which are hinged atboth ends but it is possible to use it forelements with different end conditions byemploying the concept of effective length. Theeffective length of an element which does nothave hinged end conditions at both ends is the

Fig. A2.6 The slenderness ratio of an element dependson the conditions of lateral restraint. Two examples areshown here.(a) Assuming that adequate restraint is provided at rooflevel and by the intermediate floor, the slenderness ratioof the column is based on the storey height for the assess-ment of stability in the plane of the cross-section of thebuilding.

(b) The lengths on which the slenderness ratios of thecolumns in this skeleton frame are based depend on theplane of buckling under consideration. In the plane of thewall it is the distance between the points at which lateralmovement is restrained by the cross-bracing. In the cross-sectional plane of the building it is the height of thecolumn.256

(a)

Totallength

Possible modeof buckling Length used

in assessmentof stability

(b)

Length which determinesstability in plane normal to wall

Length whichdeterminesstability inplane of wall

same as the actual length of the hinge-endedelement which has the same buckling strengthand which is identical to it in every otherrespect. Thus the effective length of anelement with hinged ends is the same as itsactual length. That of an element with fixedends is approximately 0.5 times its actuallength and that of an element with one endfixed and the other completely free is twice itsactual length (Fig. A2.7).

Limitations of the Euler formulaBecause the assumptions on which it is basedare not strictly valid, the Euler formula doesnot predict accurately the critical load of realelements. In the Euler analysis it is assumedthat the critical load value is achieved whilethe stress in the element is within the elasticrange of the material and that instability is dueto the inability of the bending stress to resistthe bending moment which is caused by theeccentricity which is present. It is not neces-sary for the stress in the material to pass theelastic limit for this to be possible. The elasticlimit of the material would of course eventuallybe exceeded when the element failed but isnot exceeded in the initial stages of the failure.The element does not fail therefore due toyielding of the most heavily stressed part, butbecause the disturbing couple is greater thanthe restoring couple while the stress is withinthe elastic range. The system is unstable, inother words. The phenomenon is known aselastic buckling and it occurs in real structuresonly if they are very slender.

In most real structures the failure mecha-nism is slightly different because when a realelement is subjected to an increasing amountof compressive load the maximum stress in thecross-section, which of course is the sum ofthe axial stress and the compressive bendingstress, usually becomes greater than the yieldstress of the material before the Euler criticalload is reached (an indication of the relation-ship between 'ideal' and 'real' behaviour isgiven in Fig. A2.5). This causes a suddenincrease in the lateral deflection which initi-ates buckling when the load is less than theEuler critical value. Most real elements there-

Fig. A2.7 The concept of effective length. The effectivelength of a compression element depends on the endconditions. Effective lengths for different end conditionsare illustrated here. (Note that these are theoretical valuesThe values which are used in practice are different fordifferent materials and will be found in design codes.)

fore fail at load levels which are smaller thanthe Euler critical load and the less slender theelement the greater is the discrepancy betweenits true failure load and the failure load whichis predicted by the Euler formula. The extent ofthe discrepancy also depends on the type ofstructure of which the element forms part. Thetype of material which is used is a particularlyinfluential factor - the behaviour of a masonrypier, for example, is significantly different fromthat of a steel column.

In practice, therefore, while the design ofcompressive elements is based on procedureswhich are similar to the one which has beenoutlined in connection with the Euler formula,the precise details of the procedures are differ-ent for different structural materials. Thesequence of operations is broadly the same asthe one given above; a cyclic process is used toarrive at a suitable size and shape for the 257

Appendix 2

Both endshinged

Both endsfixed

0.5LL

0.7L

Points of contraflexure

One end hinged,one end fixed

One end fixed,one end free

2L


element's cross-section. The permissible stressvalues which are used in practice are not basedon the Euler formula, however. For somematerials they are derived from equationswhich are similar to, but more sophisticatedthan, the Euler formula while in others theyare based almost entirely on experimentaldata. The practising designer is not normallyconcerned with the derivation of permissiblelevels of stress, however, as recommendedvalues can normally be found in codes ofpractice. They are usually presented in theform of tables or graphs giving permissiblevalues of average compressive stress for differ-ent values of slenderness ratio.

The procedures which are used to assessslenderness ratio and effective length tend alsoto be different for different materials.

A2.4 Steel structures

A2.4.1 IntroductionSteel is used principally in skeleton-frame-typestructures so the principal types of element arecolumns, beams and triangulated girders. Thestructures are normally hinge-jointed whichmakes approximate analysis straightforward.Permissible stresses for steel are given in TableA2.4.2

A2.4.2 Beams3

The most common types of these are floorbeams in multi-storey frames, for which theI-section is suitable, or secondary elements, such as purlins or cladding rails, for whichsmaller sections such as the channel arenormally used.

A good approximation to the size requiredfor a beam is given by the equation:

Zreq = M / f p b (A2.11)

2 These should only be used for preliminary sizing ofelements to test the feasibility of a proposed structure.They should not be used for final element-sizing calcu-lations.258

Zreq is the required modulus of section andprovides the basis for selecting a suitable sizeof cross-section from the available rangesspecified in Table 3.2. If a suitable section sizecannot be found from Table 3.2 another type ofsection will be required (refer to manufactur-ers' tables). It is normal to select the lightestcross-section which will provide the requiredsection modulus.

M is the maximum applied bending momentdetermined from the structural analysis.This isdetermined either from first principles, usingthe 'imaginary cut' technique, or from therelevant formula in Table A2.3. The loadpattern is determined from the area of floor orroof which the beam supports or by estimatingthe point loads which will be applied to it fromother skeleton elements which it supports.

fpb is the appropriate value of permissiblebending stress selected from Table A2.4.3

The above procedure will give a reasonablyaccurate prediction of the size required for anelement which is subjected to bendingmoment. A more accurate prediction isobtained if the shear stress and deflection arealso checked and the section size adjusted ifnecessary.

The average shear stress is given by:

Vav = V/Aav

vav should not be greater than the relevantvalue given in Table A2.3. V is the maximumvalue of the shear force in the beam. This willnormally occur close to the supports. Aav is thearea of the cross-section which resists shear. Inthe case of steel sections this is the area of theweb (the total depth of the section multipliedby the web thickness).

3 No allowance is made in this procedure for secondaryeffects such as the compression instability of thinflanges or webs - see Draycott, Structural Elements DesignManual (Butterworth-Heinemann, Oxford, 1990). It musttherefore be used only for preliminary element sizing.In most architectural structures the parts of elementswhich are likely to buckle due to local compression areadequately restrained laterally. The estimates obtainedfrom the procedure are therefore likely to be reasonablyaccurate in most cases.

Table A2.4 Basic permissible stresses for steel[suitable for preliminary sizing of elements only.Not to be used for final design calculations](BS 449)

Stress type Basic permissible stress {Grade 43 Steel)

Tension 155 N/mmCompression 155 N/mmBending 165 N/mmShear (average) 100 N/mmBearing 190 N/mm

Appendix 2

Shear stress is only likely to be critical in casesof very heavy loading on relatively short spans.

The deflection of the beam can be checkedusing the relevant formula from Table A2.3.The critical requirement is that the maximumdeflection under the action of imposed loadonly should not exceed span/360. It is possiblefor this limit to be exceeded even if the sectionselected is adequately strong, especially if theload is relatively light and the span long.Where this occurs a larger size of section thanthat required to satisfy the strength criterionmust be selected.

A2.4.3 ColumnsUnless there is reason to believe that they willcarry a significant amount of bending momentcolumns should be regarded as axially loadedfor the purpose of approximate sizing. For theapproximate sizing of steel structures a trial-and-error procedure similar to that outlined inSection A2.3.4 is used and consists of:

1 The selection of a trial cross-section. If thecolumn is axially loaded and the conditionsof lateral restraint are the same for allpotential buckling planes, the best sectionshapes are those which have similarbending strength about all their principalaxes. The H-shaped universal column (Table3.3) or square or circular hollow sectionstherefore perform best in this situation.Section shapes in which the bendingstrength about one principal axis is signifi-cantly greater than about the other, such asthe I-shaped universal beam, should be

used if the column is subjected to a combi-nation of bending and axial load or if theconditions of lateral restraint are signifi-cantly different for different planes.

A preliminary estimate of the size ofsection which will be required can beobtained by dividing the axial load by anestimate of the final value of the permis-sible stress. Unless the column is verylightly loaded this will normally be in theupper third of Table A2.5.

The calculation of the slenderness ratio ofthe trial column. As was discussed inSection A2.3.4, a slenderness ratio appliesto a particular plane of potential buckling. Ifthe trial cross-section is not symmetricaland/or the conditions of lateral restraint ofthe column are different in different poten-tial buckling planes the slenderness ratiosfor these planes, will also be different. Thepermissible compressive stress is deter-mined by the highest value of slendernessratio which is calculated.

For a given potential plane of bucklingthe slenderness ratio is calculated from:

slenderness ratio = Le/r

Le is the effective length of the column forthe plane concerned. This is determinedprincipally by the distance between pointsat which the column is restrained againstlateral movement (normally the storeyheight of the building). It is affected by theconditions of restraint at these locations,however, and in particular with whether ornot any restraint is provided againstrotation. This must be assessed and theeffective length adjusted in accordance withTable A2.6. In most steel frameworks theeffective length is between 0.7 and 1.0 of thedistance between lateral restraints.

r is the radius of gyration of the cross-section about the axis which is at rightangles to the potential plane of buckling.Radii of gyration of cross-sections are givenin Steel Section Tables (see Tables 3.2 and3.3). 259


Table A2.5a Permissible stresses for steel compressive elements (after BS 449)

Intermediate values may be obtained by linear interpolation.

NOTE. For material over 40 mm thick, other than rolled I-beams or channels, and for universal columns of thicknesse:exceeding 40 mm, the limiting stress is 140 N/mm2.

3 The permissible compressive stress isobtained from Table A2.5, based on thehighest of the slenderness ratio value calculated in 2.

4 The actual compressive stress which willoccur is calculated from the applied load

and the area of the trial cross-section (givenin the Steel Section Tables) and comparedwith the permissible value. The actual valueshould be equal to or slightly less than thepermissible stress. If it is not, a new trialsection is selected and the above sequencerepeated.260

Appendix 2

A2.4.4 Triangulated elementsTriangulated girders and trusses can have avariety of overall forms. The most commonlyused are the pitched truss and the parallelchord truss. Pitched trusses normally have apitch angle of between 30° and 45°. This deter-mines their overall depth. The depth of parallel

chord trusses in relation to their span can varywidely. Principal structural elements whichcarry substantial areas of floor or roof normallyhave span/depth ratios in the range 12 to 14.Lightweight triangulated steel joists are muchless deep with span/depth ratios in the range20 to 30. 261

Table A2.5b Permissible stresses for steel compressive elements (continued)

Intermediate values may be obtained by linear interpolation.

Table A2.6 Effective lengths of steel compressive elements for different conditions of end restraint(after BS 449)

Conditions of end restraint Effective length of element (L = distance between restraints)

Effectively held in position and restrained in direction at both ends 0.7L

Effectively held in position at both ends and restrained in 0.85Ldirection at one end

Effectively held in position at both ends but not restrained in Ldirection

Effectively held in position and restrained in direction at one end I.5Land at the other partially restrained in direction but not held inposition

Effectively held in position and restrained in direction at one 2.0Lend but not held in position or restrained in direction at theother end


The internal geometry of triangulated girdersis arranged so that very small internal angles(less then 30°) are avoided. The ideal arrange-ment is one of equilateral triangles.

The individual sub-elements of triangulatedstructures can be regarded as carrying either pureaxial tension or pure axial compression. Themagnitudes of the internal forces on any particu-lar sub-element can be determined by using themethod of sections. This is a version of the 'imaginarycut' technique. Once the magnitudes of the axialforces in the sub-elements are known the size isdetermined by using the techniques outlined inSections A2.4.2 or A2.4.3.

Axial internal forces are best resisted by cross-section shapes which are symmetrical, such ascircular or square hollow sections. Angle andchannel sections are also commonly used.

The feasibility of a particular structuralproposal can be checked by sizing only themost heavily loaded of the sub-elements. Inparallel-chord arrangements these are thehorizontal sub-elements at mid-span and theinclined or vertical sub-elements close to thesupports.

A2.4.5 Elements subjected to combinedstressThe basic rule for elements which aresubjected simultaneously to more than one

type of internal force is that the followingequation must be satisfied:

f a / fp a + fb/fpb<1 (A2.12)

where: fa = actual axial stressf pa = permissible axial stressfb - actual bending stressfpb = permissible bending stress

The equation must be satisfied at all locationsin the element. If the internal forces vary alongthe length of the element it may be necessaryto check that the equation is satisfied at morethan one location.

A2.5 Reinforced concrete structures

In reinforced concrete structures the elementsare subjected to primary load actions whichare either of the bending type (beams andslabs) or of the axial compression type(columns and walls). In the case of bending-type elements there are two considerationswhich affect the overall sizes of the cross-sections which must be adopted - the provi-sion of adequate bending strength and theprevention of excessive deflection. In the caseof compressive elements, the prevention of262

buckling-type failure is an important consider-ation. This is dependent on slenderness. Oftenreinforced concrete walls or columns arerelatively short, however, and the overalldimensions which are adopted can be deter-mined by the need to prevent the level ofcompressive stress from becoming excessiverather than the need to maintain the slender-ness ratio low enough to maintain adequateresistance to buckling.

The strength of a reinforced concreteelement is determined principally by theoverall size of its cross-section, by the strengthof the constituent concrete and by the quantityand location within the cross-section of thereinforcement which is positioned within it.The strength of concrete can vary widely(typically between 25 N/mm2 and 70 N/mm2)depending on the mix proportions andwater/cement ratio which are specified. It istherefore possible, by manipulating thestrength of the concrete and the amount ofreinforcement which is present, to produce arange of strengths within a given overall size ofcross-section. For this reason the overalldimensions of elements tend to be determinedby other criteria than the need to provideadequate strength. These are the need to limitdeflection in the case of bending-typeelements and the need to avoid excessiveslenderness in the case of axially loadedelements. The 'rules of thumb' given in Table4.1 provide reasonably realistic approximationsto the sizes which will be required forreinforced concrete elements.

A2.6 Masonry structures

As with reinforced concrete, the basic strengthof masonry can vary over a fairly wide rangedepending on the strengths of the constituentbricks or blocks and the mix proportions of themortar which is used. For this reason theoverall thickness of walls and columns tends tobe determined by considerations of limitingthe slenderness ratio or constructional con-venience rather than by the maximum load

which is carried. Basic masonry elements maytherefore be sized approximately from the datagiven in Tables 5.1 and 5.3.

A2.7 Timber structures

A2.7.1 IntroductionRigorous structural design calculations fortimber are complex because the carryingcapacity of timber elements can be affected bya large number of factors. Included in these arethe duration of the load, the moisture contentof the timber, the extent to which load sharingbetween elements is possible, the overalldimensions of the element, the direction of theload in relation to that of the grain, and anumber of other effects. These are taken intoaccount in rigorous element-sizing calculationsby adjusting the value which is used for thepermissible stress to suit the individualcircumstances of the structure. The permissiblestress which is used in a particular case is thebasic allowable stress for the species involvedmultiplied by one or more stress-modificationfactors. The exact procedure which is used toevaluate the permissible stress in a particularcase depends on the type of element underconsideration (i.e. on whether it is solid orlaminated, etc.) and on the nature of the inter-nal force which it will carry. These procedureswill not be explained in detail here.

The fact that so many factors can affect thestrength of a timber structure means thatsimplified approximate sizing calculations givea slightly less reliable indication of the finalsizes which will be required than is the casewith equivalent methods for other materials.Particular care is required when the limits ofthe spans in which a particular component isnormally used are approached. This isespecially the case with bolted trusses inwhich the viability of a proposed arrangementis likely to be determined by the feasibility ofthe joints. Where the limits of normal practiceare approached, the feasibility of a proposalcan only be tested by carrying out rigoroussizing calculations.

263

Appendix 2

A2.7.2 Tension elementsThe area of cross-section required can bedetermined from:

Areq = P/fapp (A2.13)

where:Areq = area of cross-section required. If

the end connection is of thebolted type the cross-sectionselected should be increased by20%

P = applied axial loadfap - permissible axial stress. This

will depend on the species oftimber. Typical values are:softwood 4 N/mm2

laminated timber 6 N/mm2

hardwood 9 N/mm2

A2.7.3 Bending elements

Solid rectangular sections

The section modulus required is found from:

Zreq = M/fpb (A2.14)

where:Zreq = section modulus required (bd2/6

for a rectangular cross-section)M = maximum applied bending

momentfpb = permissible bending stress.

Typical values for this are,softwood 6 N/mm2


hardwood 15 N/mm2

The deflection is more likely to be critical thanwith other materials and should be checked.Because most timber beams carry distributedloads, the requirement for adequate stiffness issatisfied if the following relationship iscomplied with:

/req = 4.34WL2/E (A2.15)

where:/req = second moment of area of

cross-section [bdV\2 for arectangular cross-section)

W = total applied loadL = spanE = modulus of elasticity of timber.

Typical values for this are:softwood 8000 N/mm2


hardwood 15000 N/mm2

A2.7.4 Compressive elementsTimber columns and other compressiveelements must be sized from the trial-and-error procedure outlined in Section A2.4.3. Thepermissible stress is determined by theslenderness ratio. Typical values of permissiblestress are given in Table A2.7.

Table A2.7 Typical permissible compressivestresses in timber

Slendernessratio fpc (N/mm2)Le/r Softwood Laminated timber Hardwood

0 7 8 165 6.83 7.80 15.60

10 6.66 7.61 15.2220 6.27 7.17 14.3730 5.79 6.62 13.2340 5.15 5.88 11.7650 4.35 4.97 9.9460 3.54 4.05 8.1070 2.86 3.26 6.5380 2.31 2.64 5.2890 1.90 2.17 4.34

100 1.58 1.80 3.60140 0.85 0.97 1.94160 0.66 0.75 1.50200 0.43 0.49 0.97

For rectangular cross-sections, r = 0.288tt = the smaller of the cross-section dimensions.


264

Index

NOTEThe dating convention used in this book is the following:BCE: Before the Christian Era (for BC)CE: of the Christian Era (for AD)

Aggregate, 118-9, 123-5arch, 118, 162Ove Arup Partnership, 10, 31, 58, 64, 71 ,105, 106,

141Aspden, 100

Badge's Mill, Shrewsbury, 50Baker, B. 17Banqueting House, London, 185Basilica Nova, Rome, 150Baths of Caracalla, Rome, 150Bauhaus, Dessau, 104Behnisch, G. 11, 22-3, 59Béton brut, 106Blenheim Palace, Oxfordshire, 160Bolles-Wilson, 44Boots Warehouse, Beeston, 105Botta, M. 15bracing, 4, 79, 82, 93-6, 137, 139, 170-2Brunei, I. K. 19Brunelleschi, F. 155

Calatrava, S. 9, 114Candela, F. 114cement, 118, 122-3Centre Pompidou, Paris, 37, 57, 61, 71, 73Channel Four Headquarters, London, 49concrete, 14-15, 99-146Coop Himmelblau, 22, 27, 35, 47, 59Crystal Palace, London, 54

Dieste, E. 151-2dome 9, 118, 153-8, 162Domino house 102, 112

Economist Building, London, 108effective length, 256-7Eisenmann, P. 112elastic bending formula, 250Villa Emo, Fanzolo, 160Engineering Building, Leicester, 111

Erith, R. 1equilibrium, 4Euler, L. 253Exhibition and Assembly Building, Ulm, 46

Fairbairn, W. 19Falling Water, Pennsylvania, 110Farnsworth house, Illinois, 57Fleetguard Factory, Quimper, 87Florey Building, Oxford, 43, 111, 112form active structure, 33-4, 97, 117-18, 145, 148-

228, 236-7Forth Railway Bridge, 17Foster, N. 3, 12, 29, 31,64, 111-12Fowler, H. 17

Gehry, F. 22, 34, 114-16German Pavilion, Barcelona, 56Giacomo della Porta, 158Goetheanum, Dornach, 14Graves, M. 112Grimshaw, N. & Partners, 57, 87Groninger Museum, Groningen, 27Gropius, W. 104, 107

Hagia Sophia, Istanbul, 153, 155Hawksmoor, N. 160Hejduk, J. 112Hennebique, F. 100Highpoint, London, 105, 141History Faculty Building, Cambridge, 111HongkongBank Headquarters, Hong Kong, 3, 94Hunt, A. 12,42,80,87

Ice Rink, Oxford, 87-8Inmos Microprocessor Factory, Newport, 87

Invalides, Les, Paris, 158

Jones, I. 185

Kowalski, 189265


Lasdun, D. 107-8La Tourette, monastery of, 107Le Corbusier, 29, 102, 106-8, 112Lloyds Headquarters, London, 31,61loadbearing wall structure, 1, 6-8, 34, 37, 40, 10

117, 140-1, 162-78, 185-8, 222-5loading, 4, 239-243Lubetkin, B. 105-8, 141

Maillart, R. 100-1, 114Mansart, J. H- 158masonry, 15, 40, 147-78

cross-wall, 168-9, 171diaphragm wall, 37, 151, 176-8fin wall, 37, 151, 176-8hoop stress, 153, 158slenderness ratio, 169-71spine wall, 166, 168wind loading, 173-7

McLachlan, E. & F. 8Meier, R. 22,46, 112-5Mero space framework, 82, 84Michelangelo Buonarotti, 158Mies van der Rohe, L. 2, 31, 53, 56, 161-2modulus of section, 251Museum of Decorative Arts, Frankfurt, 112-15

National Theatre of Great Britain, London,107-8

Nervi, P. L. 114New City Library, Munster, 44New York Five, 112Nodus space framework, 85non form active, 34Notre-Dame Cathedral of, Chartres, 152Notre-Dame Cathedral of, Paris, 151Notre-Dame-du-Haut, Ronchamp, 109

Olympic Stadium, Munich, 11

Palladio, A. 159-60panel structure, 7-8Pantheon, Rome, 153-4Parthenon, Athens, 159, 183Paxton, J. 54Penguin Pool, London, 105-6Perret, A. 101Petit Maison de Weekend, 107Piano, R. 10, 71Piranese, G. B., 184-5Poleni, G. 158portal framework, 37post-and-beam structure, 5, 34, 37, 158-62

Powlesland, J. 204principal stress, 127

radius of gyration, 254-5Ragnitz Church, Graz, 189Ransome, E. 100reinforced concrete, 40, 42

beams, 130bond stress, 128columns, 130flat slab structure, 42, 101, 135-8hybrid structures, 143-5low-heat cement, 123mix design, 125post tensioning, 121precast, 132-4, 141-4pre-stressing, 120, 131pre-tensioning, 121shells, 145-6shrinkage, 123slabs, 131stairs, 132water/cement ratio, 122workability, 119

Renault Warehouse and Sales Headquarters,Swindon, 31, 64, 87

rigidity, 5Ritz Hotel, London, 51Rogers, R. 71, 87

Ste-Cécile Cathedral of, Albi, 148St loseph's Chapel, Ipswich, 204St Paul's Cathedral, London, 157-8St Paul's Outside the Walls, Rome, 183-4St Peter's Basilica, Rome, 158, 183-4Sainsbury supermarket, London, 57Samuelly, F. & Partners, 44, 111-12Santa Maria del Fiore, Florence, 155-6, 158Villa, Savoye, 29, 103, 110Seagram Building, New York, 55Sears Tower, Chicago, 59Selfridges, London, 52semi form active, 34skeleton frame, 6, 13,34,40,51, 117, 188-9,225-8Skidmore, Owings & Merrill, 59slenderness ratio, 253-6Smithson, A. & P. 108Spacedeck space framework, 83Staatsgalerie, Stuttgart, 3stability, 4steel, 12-14,40,49-98

bolts, 72bundled-tube, 95, 97

266

cable network, 10, 37, 97-8cold-formed section, 69diagrid, 81fire protection, 61, 72-3framed-tube, 95heat-treatment, 62hot-rolled section, 64interstitial-truss, 91space-framework, 37, 79-86staggered-truss, 91trussed-tube, 95, 97tube-in-tube, 95

Steiner, R. 14Stephenson, R. 19Stirling, J. 3strength, 4stress

axial, 249-50bending, 250-1shear, 251

Sullivan, L. 51Szyszkowitz, 189

Taut, B. 53Tecton, 105-6tent, 10Terry, Q. 1timber, 40, 174-231

balloon frame, 222-4fire resistance, 196gluelam, 202grade stress, 194-5grading, 190, 196-8half-timbered building, 185-7independent frame, 223-4joints, 210-14

laminated board, 201-2laminated timber, 202-3lattice dome and vault, 231loadbearing wall, 185-8modified frame, 223-4moisture content, 192-194moisture movement, 192particle board, 201-2platform frame, 222-4plyweb beam, 203plywood, 201-2sawn timber, 198-9seasoning, 193shell, 228-31skeleton frame, 188-9, 225-8stressed skin panel, 215-7truss, 206-9trussed rafter, 180, 204-6wallframe, 185, 187-8, 222-5

Torroja, E. 114

Unistrut space framework, 83Unite d'Habitation, Marseilles, 29, 107

Vanbrugh, J. 160Vitra Design Museum, Basel, 35, 114-16vault, 9, 148-52

Williams, O. 105, 108Willis, Faber & Dumas building, Ipswich, 22, 42, 80,

111-12Winslow House, River Forest, 39World Trade Centre, New York, 95Wotruba, F. 99Wren, C. 157-8Wright, F. L. 37, 39, 109-10

267

Index

ARCHITECTURE

Following on from the author's previous book, Structure and Architecture, this book is

concerned with the preliminary stages of the design of structures for buildings and deals with

the issues involved in structural decision making at a strategic level.

It is intended principally for use by students of architecture and will provide them with the

information required to make sensible choices on the structural aspects of architectural

design. The topics covered in the book include choice of structural material, the

determination of the form and geometry of structures, the integration of the structure with

other aspects of architectural design and the preliminary allocation of dimensions to the

principal structural elements. It is essentially a non-mathematical treatment. Calculation

procedures are reduced to the minimum compatible with achieving a meaningful

approximation to the required member dimension.

Dr Angus Macdonald is currently the Head of the Department of Architecture at the

University of Edinburgh. He specialises in the teaching of structural design and architectural

history. He is the author of several other books including Windloading on Buildings and

Structure and Architecture which was published by Butterworth-Heinemann in 1994.

Cover: Cambridge Law Faculty, Cambridge, England. Sir Norman Foster & Partners, Architects.

Anthony Hunt Associates, Structural Engineers. Photograph: Dennis Gilbert

Architectural Presshttp://www.bh.com

An imprint of Butterworth-Heinemann

arcpress structural design for architecture

Documents