Timber Engineering Step-2

Timber Engineering STEP 2 Design - Details and structural systems

Edited by H.J. Blass P. Aune B.S. Choo R. Gorlacher D.R. Griffiths B.O. Hilson P. Racher G. Steck

First Edition, Centrum Hout, The Netherlands

First published in 1995 by Centrum Hout Postbus 1350 1300 BJ Almere The Netherlands Printed in The Netherlands by Salland De Lange, Deventer. ISBN 90-5645-002-6 The STEPIEurofortech volume I and I1 will be distributed through the National Representative Organisations (NRO). Copies can be purchased from the NRO, each of whom is responsible for the distribution of the lecture material in their own country only. Eurofortech, International Office is reponsible for the distribution in those European countries where no NRO is identified and for all other countries. All text and pictorial slides will be available and can be purchased from Centrum Hout only. This title is available in paperback edition also in German and French. The paperback edition is sold subject to the condition that it shall not, by way of trade or otherwise, be lent, resold, hired out, or otherwise circulated without the publisher's prior consent in any form of binding or cover other than that in which it is published and without a similar condition including this condition being imposed on the subsequent purchaser. All rights reserved. No part of this book may be reprinted or reproduced, or utilized in any form or by any electronic, mechanical or other means, now known or hereafter invented, including photocopying and recording, or in any information storage and retrieval system, without permission in writing from the publisher.

Contents Foreword Preface Acknowledgements Authors National Representative Organisations Contract implementation D

Design - Details Truss connections with dowel-type fasteners Connector joints in trusses Restoring timber structures - Inspection and evaluation Restoring timber structures - Repair and strengthening Column to beam and beam to beam connections Special connections using steel plates Transportation and erection Frame corners Bracing - Structural detailing Detailing of timber structures in seismic areas Computer aided design and manufacturing

E

Design - Structural systems

El E2 E3 E4 E5 E6 E7 E8 E9 El0 Ell El2 El3 El4 El5 El6 El7 El8 El9 E20 E21 E22

History of timber structures Conceptual design Purlins Straight and tapered glulam beams Curved and pitched cambered glulam beams Trusses made from laminated veneer lumber Trusses made from glulam and solid timber Scaffolding and falsework Plane frames and arches Timber frame houses - Structural Timber frame houses - Serviceability Timber frame houses - Fire resistance Timber concrete composite structures Bracing of structural systems Beam and post structures - Principles Beam and post structures - Design Timber bridges Timber piles Roundwood structures Spatial frames and domes Timber shell roof structures Fatigue design for timber and wood-based materials

STEPIEUROFORTECH

- an initiative under

the EU Comett Programme

E23 E24 E25 E26 E27 E28

Retaining walls, sound and road barriers Structures for transmission systems Concrete forming Timber structures in aggressive environments Learning from failures I Learning from failures I1

STEPIEUROFORTECH

- an initiative under the EU Comett Programme

Foreword The European Confederation of Woodworking Industries, Cei-bois, is the spokesbody and representative organisation of the woodworking sector in Europe. With a production value of 78,000 MECU in 1994 (EU 12 countries) and a workforce of 1,7 million, woodworking is an important manufacturing sector in Europe. With the accession of Austria, Finland and Sweden to the European Union in 1995, this importance has increased even more. Nevertheless and still all too often, the wood sector is looked upon as being traditional, using a raw material with a lot of aesthetic qualities but about which little is known regarding physical and mechanical characteristics. This has certainly been the case in the past with regard to the use of wood for structural construction elements. Whereas semi-probabilistic methods have been used for quite some time to calculate the safety of steel and concrete structures, these were never really applied to wood and calculations are often made on the basis of "traditional" methods with safety coefficients. Yet a lot of research has been performed over the last years aiming at a better characterisation of the material wood. This has contributed to the development of Eurocode 5. Eurocode 5 is a calculation standard for designerslspecifiers with sufficient knowledge of wood. As such, it only provides calculation rules but no background information. The merit of the STEP publication "Timber Engineering" is that it translates theoretical Eurocode-rules into concrete solutions. The fact that this work has been performed at a European level is a sign of its comprehensiveness and quality. It will therefore largely open up the "designer market". In the past, wood and calculation methods for wood were insufficiently known. This being "unknown thus unloved" made designers quickly switch to other materials. We are convinced that this handbook and the evolution towards innovation and quality on the part of wooden element manufacturers, will contribute to a wider and more justified use of wood as a construction material. We therefore wish to congratulate the authors and collaborators, STEP and Eurofortech, for their efforts and hope that in the various countries the necessary attention will be given to this initiative. A.P. Mesquita Cei-bois President

STEPIEUROFORTECH


Preface The second volume of the STEP book series completes both the EurofortechISTEP programme and the production of a unique education and training package for the structural use of timber. Based on the first volume, Timber Engineering - STEP 2 covers in two sections the design of details and structural components. Section D, 'Design - Details' covers joints such as connections with dowel-type fasteners and connector joints, frame comers or the detailing of bracing structures. Aspects such as rehabilitation, transport and erection, design for structures in seismic regions or computer aided design are also taken into account. Section E, 'Design - Structural systems', demonstrates the use of timber and woodbased material in structures. After general topics such as history of timber structures and conceptual design, different structural components like beams, trusses, frames and arches are addressed. Timber frame houses, beam and post structures and timber bridges represent typical uses for timber in structures. Finally, special aspects such as timber shell roof structures, fatigue design and learning from failures are presented to provide a comprehensive view. To assist lecturers and trainers in their use of the training materials, a set of supporting illustration materials has also been developed. The complete package should now provide trainers at all levels of professional education with a tool to instruct future generations of engineers and architects about the use and potential application of wood and wood products as a legitimate, attractive and competitive material. For practising professionals this technical resource should further increase their confidence in the use of this versatile and aesthetically appealing material obtained from a natural and renewable resource. As contractor, EUROFORTECH is pleased to have facilitated the STEP programme. During its three years it has completed an enormous task in bringing together extensive and diverse European expertise. EUROFORTECH would like to acknowledge the commitment and contributions of all the participants who have enabled this project to be realised. The drive and commitment of the STEP Review Group, supported by the STEP secretariat at Centrum Hout, has played a key role in the realisation of the programme. The result will provide the wood industry with a significant tool to assert itself as a genuine material in the market place. It is gratifying to EUROFORTECH and all participants and contributors to the STEPIEurofortech Programme that the successful implementation of the STEP project has been a major factor in the award to Centrum Hout of the U NIVERSITY E NTERPRISE EUROPEAN AWARD 1994 - Section E NTERPRISE TRAINING PROJECTS by 'SERVICIO EUROPA FUNDACION UNIVERSIDAD EMPRESA'.

Jos Evertsen Chairman Eurofortech

STEP/EUROFORTECH - an initiative under the EU Comett Programme

Hans Joachim Blass Co-ordinator STEP

Acknowledgements Authors Timber Engineering - STEP 2 E. Aasheim, The Norwegian Institute of Wood Technology, P.O. Box 113, Blindern, 0314 Oslo, Norway

M.P. Ansell, University of Bath School of Materials Science, Claverton Down, Bath BA2 7AY, United Kingdom P. Aune, University of Trondheim, The Norwegian Institute of Technology, Dept. of Structural Engineering, Rich. Birkelands vei la, 7034 Trondheim, Norway J.P. Biger, Bureau Veritas, Civil and Environmental Engineering, 17 bis Place des Reflets, 92400 Paris la Defense Cedex 44, France G. Bignotti, Holzbau Ag-S .p.a., P.O. Box 224, 39042 BrixenIBressanone, Italy G. Bonamini, UniversitA degli Studi di Firenze, Inst. di Assest e Techn. Forestale, Via S. Bonaventura 13, 50145 Firenze, Italy

H. Briininghoff, Gesamthochschule Wuppertal, Pauluskirchstrasse 7, 42285 Wuppertal, Germany A. Ceccotti, Universith degli Studi di Firenze, Dipartimento di Ingegneria Civile, Via di S. Marta 3, 50139 Firenze, Italy J. Chilton, University of Nottingham, School of Architectures, University Park, Nottingham NG7 2RD, United Kingdom J. Fischer, Lignum - Schweizerische Arbeitsgemeinschaft f i r das Holz, Falkenstrasse 26, 8008 Ziirich, Switzerland

H. Hartl, Zivilingenieur fiir Bauwesen, Kenyongasse 9, 1070 Wien, Austria B. Johannesson, Chalmers University of Technology, Dept, of Structural Engineering, Steel and timber structures, 41296 Goteborg, Sweden G. Johansson, Chalmers University of Technology, Dept. of Structural Engineering, Steel and timber structures, 41296 Goteborg, Sweden A. Kevarinrnaki, Helsinki University of Technology, Laboratory of Structural Engineering and Building Physics, Rakentajanaukio 4a, 02151 Espoo, Finland M.H. Kessel, FachhochschuleHildesheim/Holzminden, Fachbereich Bauingenieurwesen, Labor fiir Holztechnik, Hohnsen 1, 31134 Hildesheim, Germany J-W.G. van de Kuilen, Delft University of Technology, Faculty of Civil Engineering, Timber structures, P.O. Box 5048, 2600 GA Delft, Netherlands

STEPIEUROFORTECH - an initiative under the EU Comett Programme

H.J. Larsen, Danish Building Research Institute, Dr. Neergaarsvej 15, 2970 Hsrsholm, Denmark A.J.M. Leijten, Delft University of Technology, Faculty of Civil Engineering, Timber Structures, P.O. Box 5048, 2600 GA Delft, Netherlands. C.J. Mettem, Timber Research and Development Ass., Stocking Lane, Hughenden Valley, High Wycombe, Buckinghamshire HP14 4ND, United Kingdom L. Mortensen, University of Aalborg, Dept. of Building Technology and Structural Engineering, Sohngaardsholmsvej 57, 9000 Aalborg, Denmark J. Natterer, Ibois-Civil Engineering department, GCH2 Ecublens, 1015 Lausanne, Switzerland

N. Nebgen, Ingenieurburo fir Holzbau, Dieselstrasse 12, 72770 Reutlingen, Germany P. Racher, C.U.S.T.-GCnie Civil, P.O. Box 206, 63174 Aubibre Cedex, France B. Roald, The Norwegian Institute of Wood Technology, P.O. Box 113, Blindern, 0314 Oslo, Norway J .L. Sandoz, Ibois-Civil Engineering department, GCH2 Ecublens, 1015 Lausanne, Switzerland C. Short, Colin Short Associates Consulting Engineers, Brookfield, Glen Road, Delgany, Greystones, County Wicklow, Ireland K.H. Solli, The Norwegian Institute of Wood Technology, P.O. Box 113, Blindern, 0314 Oslo, Norway G. Steck, Fachhochschule Munchen, Fachbereich 02, Karlstrasse 6 , 80333 Miinchen, Germany K. Schwaner, Arbeitsgemeinschaft Holz e.V., Postfach 300141,40401 Diisseldorf, Germany S. Thelandersson, Lund University, Division of Structural Engineering, P.O. Box 118, 221 00 Lund, Sweden

P. Touliatos, National Technical University of Athens, 15 Dinokratous Str, 10675 Athens, Greece

L. Uzielli, Universith degli Studi di Firenze, Dipartimento di Ingegneria Civile, Via di S. Marta 3, 50139 Firenze, Italy G. Vidon, Socotec, Dept. Parois-Isolation, Les Quadrants, 3 Avenue du Centre, 78182 Saint Quentin en Yvelines Cedex, France S. Winter, Ingenieurbiiro, Postfach 40, 36331 Lauterbach, Germany


National Representative Organisations (NRO's) Austria

NRO: PROHOLZ - Holzinformation ~sterreich,Wien Supporting organisations: Fachverband der Sageindustrie, Wien; Fachverband der holzverarbeitenden Industrie, Wien; Bundesinnung der Zimrnermeister, Wien

Belgium NRO: Opleidingscentrum Hout, Brussels

Denmark NRO: "Traebranchens Oplysningsrid", The Danish Timber Information Council, Lyngby Supporting organisations: Danish Building Research Institute, Herrsholm; Technical University of Denmark, Lyngby; University of Aalborg

Finland NRO: VTT Building Technology, Espoo Supporting organisations: Finnish Timber Council; Finnish Forest Industries Federation; The Quality Control Association for Timber Structures (PLY); Finnforest Oy Kerto Division; Technical Research Centre of Finland (VTT)

France NRO: Union Nationale Fran~aisede Charpente, Menuiserie et Parquets, Paris Supporting organisations: Arbocentre, Orleans; Centre Experimental du Batiment et des Travaux Publics, Saint Remy Les Chevreuse; Bureau Veritas, Courbevoie; Centre Technique du Bois et de 1'Ameublement, Paris; C. U.S .T .-GCnie Civil Blaise Pascal University, Aubikre; SOCOTEC, Saint-Quentin en Yvelines

Germany NRO: Arbeitsgemeinschaft Holz, Dusseldorf Supporting organisations: Deutsche Gesellschaft fiir Holzforschung, Munchen; Fachhochschule Hildesheim/Holzminden; Universitat Munchen; Gesarnthochschule Wuppertal; Technische Universitat Munchen; Universitat Karlsruhe; Fachhochschule Munchen

Ireland NRO: Colin Short Associates, Greystones Supporting organisations: Ove Arup & Partners; University of Limerick; Forbairt; Irish Timber Trades Association; Coillte Teoranta; Cork Regional Technical College; Malcolm Jacob - Consulting Engineer; Wood Marketing Federation; Dublin Institute of Technology; University of Ulster; University College Dublin; Trinity College Dublin; Michael Quaide - Consulting Engineer; Irish Timber Council

Italy NRO: Universith degli Studi di Firenze Dipartimento di Ingegneria Civile, Firenze Supporting organisations: Universiti degli Studi di Firenze - Istituto di Assestamento e Tecnologia Forestale; Holzbau AG-S.p.a., Brixen; Centro di Ricerca e Servizi per il Legno e per il Pioppo S.R.L., Sustinente


Netherlands NRO: Centrum Hout, Almere Supporting organisations: Delft University of Technology, Delft; Stichting WESP, Woerden; Stichting Opleidings- en Ontwikkelingsfonds voor de Timmerfabrieken, Bussum; TNO Building and Construction Research, Rijswijk

Norway NRO: Thte Norwegian Institute of Wood Technology (NTI), Oslo Supporting organisations: University of Trondheim; The Norwegian Institute of Wood Technology, Oslo

Portugal NRO: Laboratbrio Nacional de Engenharia Civil, Lisboa

Sweden NRO: Trainformation, Stockholm Supporting organisations: Chalmers University of Technology; Lund University; Swedish National Testing and Research Institute

Switzerland NRO: Lignum - Schweizerische Arbeitsgemeinschaft fiir das Holz, Ziirich Supporting organisations: ETH Eidgenossische Technische Hochschule, Ziirich; EPF Ecole Polytechnique FedCrale, Lausanne; SIA Schweizerischer Ingenieur- und Architekten-Verein, Ziirich

United Kingdom NRO: TRADA, High Wycombe, Buckinghamshire Supporting organisations: Timber Research and Development Association; GangNail Systems Ltd.; Brighton University; University of Nottingham; University of Surrey; Meyer International; SCOTFI; Institute of Wood Science; MiTek Industries Ltd.; Simpson Strongtie International Incorporated; James Donaldson & Son; Donaldson Timber Engineering

Contract implementation Centrum Hout, STEPtEurofortech Secretariat, Westeinde 8, 1334 BK Almere, The Netherlands Commission of the European Communities Taskforce, Human Resources, Education, Training and Youth, COMETT Programme, Contract No 92/1/6960 Eurofortech, International Office, Roebuck Castle, Belfield, Dublin 4, Ireland


Truss connections with dowel-type fasteners STEP lecture D 1 H. Hart1 Zivilingenieur fiir Bauwesen A. Leijten Delft University of Technology

Objectives To present information about the different kinds of dowel-type fasteners used in truss systems, to demonstrate a method for calculating multiple shear joints and to give examples of the calculations involved in their design.

Summary After a demonstration of a method for designing multiple shear timber-to-timber dowelled joints attention is paid to deciding what kind of timber and fasteners are to be used in practice. Examples of the calculation of typical joints and drawings of truss systems showing joints in detail are the main part of the lecture.

Introduction Truss girders are a well established form for timber structures. They are mainly used to build roofs of sports halls, industrial buildings and agricultural buildings and come in many different forms. The advantages of these systems are their low dead weight and their low material requirements although the height of trusses at midpoint might be seen as a disadvantage compared to simple beams. The various members of trusses are either connected end to end with overlapping joints of several thicknesses or butt joints with steel plates. Multiple shear joints allow minimisation of the connection area, because the forces carried by the fastener can be distributed over several shear planes. Because of their high loadcarrying capacity dowel-type fasteners are often used for joints in truss systems, either with timber to timber or timber to steel connections.

Multiple shear joints

Figure I

A typical built-up joint of a truss.

In Timber Engineering - STEP 1 only single shear and double shear joints are treated where only two or three elements are connected. However, in a number of cases a joint consists of more elements. An example is the joint in timber trusses with elements which are not situated in one plane as is the case for trusses with punched metal plates. The top and bottom cords are continuous and run from one support to the other. These structural elements might consist of STEPIEUROFORTECH - an initiative under the EU Comett Programme

Dl11

more than one timber member. The diagonals and verticals frequently fit in between. At the junction of these members a built-up joint is created. An example is given in Figure 1. For simplicity the timber members are connected with one fastener. The number of timber elements connected by the fastener is now five. The number of shear planes is four. The problem arises in how to determine the load-carrying capacity of each shear plane. However, a straightforward application of the conventional design equations is not possible without any modification. In this respect guidance is given by EC5: "In multiple shear joints the total load-carrying capacity should be determined by calculating the sum of the load-carrying capacity for each shear plane as if it were a part of a three member joint". This sentence has two parts. In the first part the phrase 'total load-carrying capacity of the joint' is used. However, since the load which has to be transmitted by every shear plane in general is different, as are the grain directions of the members, the total load-carrying capacity is not always of interest. Of importance is the load-carrying capacity of each shear plane. The second part of the sentence gives an idea of how to analyze this multiple shear joint. A method aimed at determining the load-carrying capacity of each shear plane is now presented. For the analyses the example of Figure 1 is taken as reference. The forces in the members are indicated. It is essential to determine the direction of the force to be transmitted at every shear plane. In Figure 2 the dowel type fastener which in this case connects the members is enlarged and the forces introduced by the separate members are shown. As the loads are symmetrical, only half of the fastener is drawn. At the shear plane the fastener is cut into parts in order to show the direction of the shear force. These shear plane forces can easily be found by considering the equilibrium equations of the individual parts. In reality there are no loads but embedding stresses which act along the axis of the fastener. The basic assumption of the analysis given is that the fastener deforms and fails in the direction of the shear plane force. This direction is important because the embedding strength of the members is dependent on the load to grain angle.

Figure 2

The equilibrium of forces behveen each shear plane.

A procedure to determine the load-carrying capacity of each shear plane is now given. As the joint considered is symmetrical only two shear planes have to be examined. First consider the shear plane between member 1 and member 2 and remove all other members. Mirror member 1 along the symmetry-axis of member 2. In Figure 3b the final situation is shown. The joint members now become 1, 2 and again 1. In this way a conventional double shear joint is STEPIEUROFORTECH - an initiative under the EU Comett Programme

created. The load-carrying capacity of this shear plane is now calculated according to Johansen's model (see STEP lecture C3). The embedding strength substituted in these design equations is modified corresponding to the shear plane load to grain direction of the members. The governing failure mode is noted. Now the next shear plane between member 2 and 3 is considered. Isolate member 2 and 3 and mirror member 2 in the same way, see Figure 3c. The loadcarrying capacity as well as the governing failure mode are noted. To consider all possible failure modes of this shear plane one additional analysis is necessary, considering the combination 3-2-3, Figure 3d should finally be made. So for all shear planes which do not connect an outside member two analyses should be made.

Figure 3

Combination of members for the determination of the load-carrying capacity.

The last, but not least important, consideration is the compatibility of the failure modes of the subsequent shear planes. For the shear plane which connects the outside member in principle all failure modes have to be analyzed. It is physically possible that the fastener end which sticks out of the side member may be inclined. However, for other shear planes this is not possible as at this end another shear plane exists and the dowel must be continuous through the joint. It is evident that the governing failure mode of the next nearby shear plane should be compatible with this failure mode.

Example The joint of Figure 1 is considered. The angle between outside members and diagonals is 45". Members 1 and 5 Member 3 Members 2 and 4 Steel 8.8 yield stress Fastener diameter Density of the timber members

b b b

= = =

fu,k

=

d p,

= =

45 mm 75 mm 35 mm 640 ~ / m m ~ 10mm 380kg/m3

Table 1 shows the results. For the shear plane between members 1 and 2 the load-carrying capacity is 2970 N while for the next shear plane the capacity is 2130 N.


Member Angle between force Member Angle between force Member Angle between force number and grain direction number and grain direction number and grain direction

Mode number

Capacity [N]

Capacity [Nl

Capacity [N]

Ia Ib I1

12600 2970 4540 4070

4270 10500

21000 2130

4x40

6448

3650

6020

111

The load carrying capacities of incompatible failure modes are struck out.

Table 1

Load-carving capacities of the shear planes and compatibility of failure modes.

Material selection The direct connection of members, timber to timber carpenter style, is not always possible, because uneconomical member cross-sections would be needed. The section of the member has to be chosen to satisfy connection conditions and not the action effects in the member. To ensure maximum utilisation of the member cross-sections elements made of other materials, e.g. steel, are used to effect load transfer. The use of steel plates offers the advantage that the connection area can be kept small thus avoiding fixing moments and maintaining the assumptions made in the design method for trusses. For truss systems glued laminated timber is becoming more and more significant on account of its better material properties, its higher resistance to deformation and the possibility of producing bigger cross-sections. With respect to fastener selection a distinction must be made between permanent and temporary structures and whether fastener slip is of any importance. Dowels display a relatively high stiffness and can therefore be used in many kinds of connections. In contrast, the low stiffness of bolted connections is reason enough to restrict their use to cases where high slip can cause no damage to a structure. Bolts should be tightened in such way that the members fit closely and they should be easily accessible for retightening from time to time if necessary when the timber has reached equilibrium moisture content. For moveable structures, falsework or concrete formwork, which should be easily dismantled, bolts are preferred.

Design examples Truss illustrated in Figure 4 The figure illustrates a typical triangular truss which is quite often used. The span of the truss is short enough that the member forces can be transmitted by nails. The internal forces have been calculated by using a computer program. The significant load case is determined considering the self weight and the snow load. It is assumed that the truss system is used in an area which has a heavy snow load for a prolonged period of time, for example the alpine regions in Austria. Therefore the load-duration class is regarded as medium term.


EC5: Part 1-1: 6.3.1.2a EC5: Part 1-1: 6.2.1.1

EC5: Part 1-1: 6 . 3 . 1 . 2 ~

Figure 4

Triangular truss.

Nails:

3,l x 70 - smooth shank and not predrilled y, = 191 f,, = 600 ~/mrn'

Timber:

C27 p, = 370 kg/m3 y, =1,3 service class 1 load duration class medium term k,,, = 098

Characteristic and design embedding strength value for nails up to 8 mm without predrilled holes: = 21'6 N/mm2 f , = 13,3 N/mm2 p =l,O Characteristic and design value for yield moment for common steel wire nails with a minimum tensile strength of 600 ~ / m m ~ : , = 3410 Nmm M y , = 3100 Nmm Nails in double shear: The design value per nail is the minimum of the following values per shear plane.

EC5: EC5: EC5: EC5:

Part Part Part Part

1-1: 6.2.lg 1-1: 6.2.1h 1-1: 6.2.1j 1-1: 6.2.lk

R, R, R, R,

= 907 N = 495 N

=473N = 556 N

2 . 21 = 42 nails 31 . 70

Figure 5

Heel joint.


Connection U , to 0,:

Figure 5

F, = 39300 N nre, = 393001946

Figure 6

= 41,5 say 42 nails

Lower chord node.

Connection V, to U :

Figure 6

F, nreq

Connection D to U,:

= 4560 N = 45601946

= 4,8 say 5 nails

Figure 6 F, = 14600 N nreq

= 146001946

= 15,4say 16 nails

1450 N 2 . 2=4nails31.70

Figure 7

Apex joint.

Connection V, to 0 :

Figure 7 F, = 1960 N nreq

= 19601946

= 2,l say 4 nails


2,4=8nails31 .70 2 . 15 = 30 nails 31 . 7 0 2 . 15=30nails31 . 7 0

I

!

Y

Y

I

I

I-: 5 . 2,4

Y

I

I

Figure 8

Joint in U,:

I

*

A

*

Lower chord joint.

Figure 8 F, = 26200 N nreq = 262001946

= 27,7 say 30 nails

Truss illustrated in Figure 9 The figure illustrates another triangular truss which is a little bit different from the example described before. The internal forces have been calculated by using a computer program and they are of such magnitude that for the connections in the joints dowels have to be used. The significant load case is also determined considering the self weight and the snow load for a load-duration class regarded as medium term.

Figure 9

Triangular truss.

Dowels:

12) 20 mm Fe 360 ,y = 191

Timber:

C24 p, = 350 kg/m3 ,y =1,3 service class 1 load duration class medium term k,,, = 098


-

&~J/:>/>A "

Figure 10

Lb<,."Lb"

v-"

10

,,",."

,,. 8

.6 dowels . 020 . rnm

20

Heel joint.

Connection U , to 0, :

Figure 10 F, = 290000 N

Characteristic and design value for yield moment for round steel bolts: EC5: Part 1-1: 6.5.1.2e

Characteristic embedding strength value for bolts up to 30 mm parallel to the grain: EC5: Part 1-1: 6.5.1.2a EC5: Part 1-1: 6.5.1.2~

Characteristic embedding strength value for bolts up to 30 mm at an angle to the grain of 18 degrees: EC5: Part 1-1: 6.5.1.2b

Design values of the embedding strength: EC5: Part 1-1: 6.2.11 EC5: Part 1-1: 6.2.lm

f , = 14,l N/mm2 f,,,,,,= 13,3 ~ / m n z ~ p = 0,94

Dowels in double shear: The design value per dowel is the minimum of the following values per shear plane. EC5: Part 1-1: 6.2.lg EC5: Part 1-1: 6.2.lh EC5: Part 1-1: 6.2.lj EC5: Part 1-1: 6.2.lk

D 118


"beq = 290000/(13300

Figure 11

- 2 - 2 ) = 5,4 say 6 dowels

Lower chord node.

Figure 11 Connection D, to U,: The sum of forces in D, and D, has to be connected. F, = 86900 N Characteristic and design value for yield moment for round steel bolts: EC5: Part 1-116.5.1.2e

= 384 Nm My,, = 349 Nm

,

Characteristic embedding strength value for bolts up to 30 mm parallel to the grain: EC5: Part 1-1: 6.5.1.2a EC5: Part 1-1: 6.5.1.2~

f,,,,k = 23,O N/mm2 k,,

=1,65

Characteristic embedding strength value for bolts up to 30 mm at an angle to the grain of 1 8": EC5: Part 1-1: 6.5.1.2b

fh,18,k

= fh,l,k

= 21,6 N/mm2

Characteristic embedding strength value for bolts up to 30 mm at an angle to the grain of 34":


fh,l,d

,

p

= 13,3 N/mm2 = 11,7 N/mm2 = 0,88

Dowels in double shear: The design value per dowel is the minimum of the following values per shear plane. STEPIEUROFORTECH - an initiative under the EU Comett Programme

Dl19

EC5: Part 1-1: 6.2.lg EC5: Part 1-1: 6.2.lh EC5: Part 1-1: 6.2.lj EC5: Part 1-1: 6.2.lk

n,,, = 86900/(11700 . 2) = 3,7 say 4 dowels

Figure 12

Apex joint.

Connection D, to 0,: F, = 67400 N

Figure 12

Characteristic and design value for yield moment for round steel bolts: EC5: Part 1-1: 6.5.1.2e

M y , = 384 Nm M y = 349 Nm Characteristic embedding strength value for bolts up to 30 mm parallel to the grain:

EC5:Partl-1:6.5.1.2a EC5: Part 1-1: 6.5.1.2~

fh,O,k

k,,

=fh,~,k

= 23,O ~ / m m ~

= 1,65

Characteristic embedding strength value for bolts up to 30 mm at an angle to the grain of 34":


D1/10

fhuc = 14,l ~ / r n r n ~ f,,,,,,= 11,7 N/mm2

p

= 0,83


Dowels in double shear: The design value per dowel is the minimum of the following values per shear plane. EC5: Part EC5: Part EC5: Part EC5: Part

1-1: 6.2.lg 1-1: 6.2.lh 1-1: 6.2.lj

R, R, R,

1-1: 6.2.lk

R,

= 28300 N

=11700N = 13400 N = 14700 N

.

"Lq = 67400/(11700 2) = 2,9 say 4 dowels

The further added examples should give an impression of how to carry out joints with dowel type fasteners used in different truss systems. Figure 13 shows a trapezoidal truss and Figure 14 the construction of a typical joint in this truss. Wide-spanning trusses, see Figures 15 and 16, often require joint constructions using steel plates to transmit the internal forces as illustrated in Figure 17. Figures 19 and 20 give details of joints of the parallel truss in Figure 18.

Figure 13

Trapezoidal truss.

Figure 14

Example of a joint in the truss shown in Figure 13.

Figure 15

Timber truss with steel-to-timber connections.

I'

'11


Dl111

Figure 16

Details of truss shown in Figure 15.

I-,

2 . 88 nails 6,O. 80

'

L.-*

I

1

Figure 17

Upper chord node with nailed-on steel plates and central bolt.

Figure 18

Parallel chord timber truss.


12 dowels 012 mm

Figure 19

Lower chord node of truss shown in Figure 13.

2 . 10 = 20 nails 9 0 . 310 ( p )

Figure 20

Lower chord node of truss shown in Figure 18.


J I

I

X-

Connector joints in trusses STEP lecture D2 H. Hart1 Zivilingenieur fiir Bauwesen

Objectives To present information about the different kinds of connector used in truss systems and to show examples of joints.

Summary The lecture presents general information about connectors, which are used in truss systems.

Introduction Apart from the fixing of individual beams, mechanical timber connectors have the function within trusses to transfer ring forces from one component to another, Forces are thus introduced into joints by compression and shear, depending on the stiffness of the connector and the embedding strength of the wood. Under these actions elastic and plastic deformations occur in the wood. The dowels used nowadays have been developed from carpenter's dowels. They were mainly produced from dry hardwood with a rectangular section. With the evolution of drilling and milling machines even round hardwood dowels could be placed in timber. Since the twenties a lot of different special dowel types have been developed and their use has been proved with tests. Today only rectangular hardwood dowels, rectangular and T-type metal dowels and special type dowels, made of cast-iron or aluminium, are of any importance. The principles of calculation are covered by lectures C9 and C10.

Joints in trusses in general Eccentricity should be avoided, if possible, when joining the members in a truss systems otherwise considerable additional stresses are caused, see Figure 1. Whenever it is not possible to fulfil this aim the influence on the load-bearing capacity of the connection has to be verified. Tests carried out have shown that in such cases the reduction of the load-bearing capacity of joints can be as high as 30 %. The magnitude of the additional moment can be calculated as follows:

M M

= D, sin a, e = D, sin =IM,I+IM,I

Figure I

a, e

Effect of eccentric connections.

STEPIEUROFORTECH


Examples of joints Rectangular dowels are not often used for joints within trusses but an important example for the use of rectangular hardwood or metal dowels is the combination of a front side recess with a bracket, see Figure 2. This construction is necessary if a fully loaded diagonal member is joined to a bottom chord member and either the compression force cannot be transmitted on account of the load introduction length being too short or because the tie-beam must not be weakened by the recess. This constructional solution offers the advantage that the beams need not to be changed but the effect of any moments resulting from the eccentric connection has to be taken into account in the calculation.

Figure 2

Use of T-shaped dowels in combination with a hardwood bracket (a).

Figures 3 to 25 illustrate the construction of joints in trusses with connectors. When designing the truss joints shown below the following points have to be considered: -

-

-

design loads, load-carrying capacity of connectors and bolts, minimum cross-sections of timber, end distances in the direction of grain, distance from the centre of the connection to the outer surface of the upper chord, with forces acting at an angle to the grain, minimum spacing.

Example: Double-sided ring connector joint, type A1 according to EN912. Calculation according to STEP lecture C9

F, k,,,, t1 t,

d,

= 194 kN (tension) = = 60 mm = 90 ?nnz = 126 mm

b, h, h,

= 160 mm = 160 mm

=15mm

timber strength class C27 according to EN338 p, = 370 kg/m3 spacings and distances a, = 250 mm

a3,t

= 250 ~ n m


Figure 3

C9: Equation (12)

Splice joint with ring connectors.

a

= 0"

n

= 2 x 4 connectors

"bf

-

= 2 x 4 connectors

minimum spacings a3,r,nzin

1,5 d,.

= 1 , s . 126

= 189 mm

minimum thicknesses: -t~,rnin t2,rnin

2,25 h, 3,75 he

= 2,25 . 15 = 3,75 . 15

= 33,8 mm

= 56,2 mm

C9: Equation (16)

Rc,k,l

Rc,k,l Rc,k,2 Rc.k,2

-

-

35kpk,3krdc'35 = 3 5 1 , 0 6 0 , 9 9 2 . 1 , 0 . 1 2 6 1 s 5 51,9 kN 31,s k, k, h, d,. 62,9 kN

= 31,s . 1,051 . 1,O . 15 . 126


Example: Double-sided toothed-plate connector heel joint, type C10 according to EN912. Calculation according to STEP lecture C10

Figure 4

0 0 k,,, tI t, d,

Truss heel joint with toothed-plates.

= 30,O kN (compression) U = 29,l kN (tension) = 41,5 kN (compression) U , = 42,8 kN (tension) = 0,8 = 50 mm b, =120mm = 80 mm b, = 140 mm = 80 mm h, = 27 mm

timber strength class C27 according to EN338 p, = 370 kg/m3 bolt dh

= 20 mm

fU,,

spacings and distances a3,c = 170 mm = 170 mm = 0" (the middle member is supported at the bottom) a = 2 connectors n minimum spacings %,c,rnin

1,2 d,

= 1,2. 80

= 96 mm


minimum thicknesses: t~,min t2,min

C10: Equation (9)

ECS: part 1-1: 6.2.1 (2)

EC5: Part 1-1: 6.5.1.2 (1)

--

= 1,1 . 27 = 1,9 . 27

1,l hc 1,9 h,

-

=29,7 mm = 51,3 mm

Load-carrying capacity of the connector -30 k, k , kt d,',' Rc,k

Load-carrying capacity of the bolt Design values of material properties: Embedding strength (yM = 1,3) side member: fh,o,d = fh,~,d

=

0,082 (1 - 0,01 - 20)

middle member: k, = 1,35 + 0,015

. 20

=

370

08 1,3

2=

149 ~ / r n r n ~

1,70

EC5: Part 1-1: 6.2.1 (1) ECS: part 1-1: 6.2.1 (3) ECS: Part 1-1: 6.5.1.2 (2)

Yield moment (yM = 1,l) 203 1 Myd = 0'8 360 . - . 6 1,l

=

349 Nm


EC5: Part 1-1: 6.2.1 (1)

Design load-carrying capacities per shear plane per bolt

0,5 Rb,a,d= min

. 14,9 . 80 . 20 . 0,96 .

<

1,1 (2 1.3+ 0,96) 50 10 20

b

10-

=

11,4 kN

14,9 11j4 .'349000 20 . 50'

I

r

-0,96

]

=11,4k~

design load-carrying capacity of the joint

Figure 5

Bottom chord joint with toothed-plates and nails; (U,, V,, D, U,).

Figure 6

Truss apex joint with toothed-plates; (VZ, D, 02).( a ) Toothed-plate connectors type ClO according to EN 912, ( b ) contact joint.


m

Figure I0

Bottom chord joint with ring connectors and nails; (U,, D,, V,, D,, U,).

Figure 11

Apex joint with ring connectors; (D,, 0,). (a) Contact joint.

Figure 12

Ring connector heel joint; (01,UI).


Figure 13

Bottom chord joint with ring connectors and contact; (U,, D,, V,, D , U3).

Figure 14

Apex joint with ring connectors; (D,, 0,). ( a ) Contact joint.

Figure 1.5

Toothed-plate connector heel joint; (O,, U , ).


Figure 16

Bottom chord joint with toothed-plate connectors and contact; (U,, D,,V,, 0 9 U3). i

Figure 17

Apex joint with bolts; (D,, 0,). (a) Contact joint.

Figure 18

Ring connector heel joint; (O,, U , ).


Figure 19

Bottom chord joint with ring connectors and contact; (U2, D , V2,D,, U,).

Figure 20

Apex joint with ring connectors; (D,. 0,). (a) Contact joint.

Figure 21

Ring connector heel joint; ( 0 1 , Ul ).


k

100 / 100 1 l M l t

Figure 22

Bottom chord joint with ring connectors and nails; (U,, D,,

Figure 23

Apex joint with ring connectors; (D,, 0,). (a) Contact joint.

Figure 24

Parallel truss; bottom chord detail see Figure 25.

460i80 1100 /801601

Figure 25

Bottom chord joint with ring connectors; (U,, D,, D,,U, j.


Restoring timber structures Inspection and evaluation STEP lecture D3 G. Bonamini Universita degli Studi di Firenze

Objective To present the various objectives and methods for the inspection of existing timber structures, as a basis for the evaluation of their safety, serviceability and prevention of future degradation.

Summary Existing structures, whether old or new, must periodically undergo a thoughtful inspection and evaluation of their safety and serviceability. For timber structures the accomplishment of this difficult task implies the involvement of different experts: wood technologists, engineers, architects and possibly other experts cooperating to check the condition of each structural element, evaluate the serviceability of the whole structure and prevent future degrade. This lecture describes the wood technologist's various methodological approaches.

Introduction Conservation and restoration of existing timber structures is a relatively new idea, the implementation of which requires a multi-disciplinary approach. In the pool of experts (architects, engineers, restorers, etc.), a basic role is played by the wood technologist; in fact, through a careful examination of each structural member and joint, his duty is to derive accurate information pertaining to the properties, performance and condition of the material, and also to determine realistic boundaries within which the designer shall make his calculations. In general, the inspection of an existing timber structure has the following main objectives:

-

to provide the information needed by the structural engineer in order to assess if the strength and the stiffness of members and of connections are satisfactory for structural safety in the intended use;

-

to point out parts which may need specific reinforcement, substitution, or other types of intervention;

-

to evaluate decay factors which may have affected the structure, (in part or in whole) and which may affect it in the future; and to identify and recommend appropriate, remedial measures to prevent such risks from recurring;

-

to cooperate toward the identification of the structural behaviour of the construction work; in the case of structures having historical artistic or cultural value, to cooperate toward a better knowledge of its history, manufacturing techniques etc.

The means to achieve these objectives (concerning all the load bearing elements) are the following:

-

to assess the timber quality: ascertain the wood species and its main physical and mechanical properties, including defects and anomalies;

-

to detect existing decay or damage suffered in service;

STEPIEUROFORTECH


-

to assess the risk of decay or damage in future;

-

to assess the effective cross section(s), and their strength and stiffness characteristics.

Factors influencing the state of the structure Since wood is a material of biological origin, its properties show a much greater intrinsic variability than other quite homogeneous and almost isotropic structural materials, such as steel and concrete. Therefore, no reliable methods exist today which can accurately predict the strength of an individual structural timber member. Grading methods and testing procedures have been developed for the determination of characteristic strength values of newly sawn timber, which are based on the statistical distribution of a group (or grade) of timber members in which the actual strength of an individual member cannot be exactly assessed. Moreover, the uncritical extension of rules and values given in structural design codes to old timber members may be inappropriate, for example because:

-

it is sometimes impossible to visually inspect all faces over the whole length of the members in-situ;

-

wood can be affected by decayldamage, which in general is excluded from new timber by grading rules;

-

old beams are often of large cross-section, which include the pith in a more or less central location.

Gross mistakes in the assessment of old timber members can be avoided by identifying and separating, through different surveying methodologies for each case, the partial contribution of every possible source of variability, and then by operating a final synthesis combining together those results. As a minimum, the following sources of variability shall be considered by the surveyor for evaluation.

Original timber quality A basic characterisation of timber quality can be achieved through the assessment of the following parameters: wood species; provenance (if available); type of timber (round timber, sawn timber, beams, joists, planks, boards, crosssection with pith included or not, quarter sawn, flat sawn, etc.); sapwood and heartwood distribution, density (average value at 12% moisture content), ring width, slope of grain, location and extension of natural defects (knots, fissures, checks, ring shakes, resin pockets, compression or tension wood). The influence of each of these parameters shall be assessed based on procedures similar to those developed for the grading of newly sawn timber.

Service conditions Biological degradation: the wood technologist shall detect biological degradation, indicating with precision its origin (fungi, insects, bacteria, marine organisms), causes, effects and development. It should be pointed out that ordinarily it is necessary to proceed to a refined identification of the wood destroying organisms, because of the broad variability among different species within the same group (life cycle, kind of decay, influences on timber structural properties, prevention strategies). For example, it is essential to distinguish between the discoloration due to stain fungi (neglectable effect on wood strength) from that caused by incipient attack of decay fungi (responsible for brown, white or soft rot, i.e. the complete destruction of the wood molecular structure); or between STEPJEUROFORTECH - an initiative under the EU Comett Programme

the exit holes of Anobium and Lyctus beetles, in order to avoid expensive and dangerous mistakes. Age: small clear wood small specimens extracted from old sound timber members and tested to failure (bending, compression, impact, etc.) show no significant differences in strength and stiffness values from comparable new material (Jessome 1965, Kuipers 1986, Cristelli F. 1986, Ehlbeck and Gorlacher 1987, Rug and Seeman 1991). In the absence of decay, it is in fact (up to now) impossible to discriminate an "age effect" on wood from test data. This does not mean that wood did not undergo any chemical modification in the course of centuries: the amount of crystalline cellulose, for example, seems to decrease with time (Borgin et al. 1975), but simply that the range of strength values obtained from this material fits almost perfectly into that obtained from new wood specimens of the same species and quality. Hence, the surveyor shall not include timber age (no matter how defined) among his assessment parameters. Loading conditions: they shall be considered mainly by the engineer according to the structural design codes. Nevertheless, the surveyor should be aware, for example, of the emphasis to give in his report to large deflections in timber bending members, trying to estimate if these are only due to elastic deformation or partly to hygro-mechanical phenomena (creep). Moisture content: wood is hygroscopic and its moisture content affects almost every property considered in the design of timber structures: strength, stiffness, durability, dimensional changes (shrinkage and swelling), shape stability, etc. It should be remembered that all design values are referred to timber at 12% moisture content and modification factors for different actual moisture content values are given. Surveyors shall put maximum emphasis on a careful assessment of timber members in respect of their past, actual and future relationships with surrounding water (air humidity, rainwater or pipe leakages, condensation, capillary water present in wet walls, etc.), reporting in detail every localised or general situation in which a timber moisture content exceeding 18 20% has been measured or is likely to be attained, with subsequent risks of biological degradation due to fungal attack (see below). ,

Temperature: the effects of temperature in timber structures are very often overestimated by engineers and architects experienced with other materials such as steel, which show a significant thermal expansion. As a general rule, effects on timber strength can be ignored for temperatures under 60 "C. On the other hand, changes in temperature result in changes of wood moisture content, with subsequent dimensional changes due to shrinkage or swelling, which are much greater than thermal expansion or contraction (see below). Radiations: the most common case is that of timber irradiated by the sunlight. Ultraviolet rays can modify the structure of wood only in the very superficial layers (one mm depth or so), causing discoloration (greying) and a sort of carbonization in the worst cases. The underlying wood mass is protected and therefore not affected. Other radiation types such as gamma-ray, x-ray, microwaves and similar, can modify or even destroy the intimate wood structure, but this may happen at a dose rate far beyond the thresholds normally encountered in civil engineering.


Maintenance and repair history It is a matter of fact that the worst damage in old timber structures often occurs as a result of improper restoring interventions. For surveying purposes the following aspects shall be checked:

-

Modifications in the use of the building or of the structure.

-

Unfavourable micro-climate alterations: particular care shall be drawn to insufficient ventilation of all the surfaces of the timber members, to the "sealing" of beam ends into the walls, to the possible formation of condensation as a consequence of roof waterproofing (asphalt coatings, vapour barriers and similar) or air conditioning units installation.

-

Insufficient routine maintenance: any structure must undergo periodic maintenance work, and timber structures are no exception. Actually, many timber structures have been "forgotten" for decades. A slow accumulation of moisture, dirt, decay and mechanical damage may have cumulated into a serious general degradation, whereas timely and qualified maintenance can easily preserve the original structural integrity for centuries.

-

Improper repair and restoration: special attention shall be drawn to the compatibility of timber and other materials often used for repair. Under "compatibility" both physical-mechanical characteristics and static-dynamic behaviour of the wood-other material combination shall be ranked. The commonly used reinforcinglrepairing techniques often do not take into account to the necessary degree the basically different hygro-thermomechanical behaviour of steel, concrete, epoxy-resins and other structural materials in respect to wood and its anisotropy. Another common source of problems may be the beloved but strongly censurable plugging of fissures with wooden wedges (or even their sealing with epoxy resin): these longitudinal fissures derive from the natural shrinkage of drying wood and their width shall be free to continuously vary according to the climatic changes and related variations in the wood equilibrium moisture content (EMC). Their plugging with a stiff material will prevent their physiological movements, hence inducing stresses in the material and the opening of new fissures. It is strongly recommended to avoid plugging or - if absolutely necessary for some reason - to execute it with a soft and yielding material.

Inspection levels The restoration of timber structures is very often part of a more general intervention on buildings, where dimensions and historical importance make it advisable to split the surveying work into two separated steps:

-

a preliminary inspection aiming to give a general idea of the structure's condition and of the actual need for further detailed investigations;

-

a detailed inspection, leading to the assessment of each individual timber member in the structure.

First level inspection (general evaluation) Objectives: preliminary assessment of whole structures without assessment of individual timber members; recommendations regarding intervention priorities (priority classes). Minimum feasibility conditions: the surveyor shall be provided with the plans of the relevant parts of the building; the structure must be accessible and the faces STEPIEUROFORTECH

- an initiative

under the EU Comett Programme

of the timber members should be cleaned and well illuminated; for the best results, coatings should be removed from wood surfaces. General visual inspection: wood species shall be identified to the degree of accuracy permitted by the visual inspection: a skilled wood technologist should be able to discriminate - in normal cases - at least softwoods from hardwoods. In some cases, certain groups of wood species can be accurately identified by visual inspection thanks to their peculiar anatomical features, such as Oak (gen. Quercus), Chestnut (gen. Castanea), Elm (gen. Ulmus), Fir and Spruce (gen. Abies and gen. Picea), Pine (gen. Pinus), Larch (gen. Larix). Average timber quality shall be visually assessed, taking into consideration mainly knots, fissures, slope of grain, apparent decay andlor damage. Expected biological risk evaluation: through a critical examination of the environmental conditions, the surveyor will evaluate the risk of biological attack and place the whole structure under assessment into one of the service classes foreseen by EC5. Timber structures exposed (or having been exposed in the past) to high biological risk conditions should be rated under a high priority class even if their actual conditions seem to be good. Apparent decayldamage evaluation: the surveyor shall identify and report any sign of wood decay, even at incipient stages, giving details about the available technical means to stop its worsening. Grading of the structure according to priority classes: on the basis of the above mentioned operations, the structure will be allocated (as a whole) into one of the three following priority classes: Class "green":

low priority of intervention; as far as timber is concerned, the structure is in good condition; biological risk is low; eventual decay is in incipient stages, no longer active and restricted to uncritical locations; only routine maintenance is required.

Class "yellow":

medium priority of intervention; timber members suffer some decay; damage andlor biological risk is high; further detailed investigations and restoring interventions are required in the short term, even if the structure is not under immediate danger of collapse.

Class "red":

high priority of intervention; timber is affected by decay in critical locations; the structure is in immediate danger of partial or total collapse; further detailed investigations are required, but their execution will be possible only after provisional reinforcinglrepairing interventions in order to comply with safety criteria.

Deliverables: the surveyor should deliver at least a thematic map, in which priority classes are represented on the building plan by appropriate colours: green, yellow, red and grey (this fourth colour is required to indicate those structures intended for visual assessment, but excluded from it because they were not accessible for any reason) and a technical report on the general condition of the timber structures under inspection, including recommendations about eventual need of further detailed analysis or other types of intervention.


Second level inspection (detailed evaluation) Once priorities have been well established, a detailed inspection might be required for those structures showing problems to such a degree that more information on residual sections and/or the effective performance of timber members is needed. Objectives: to assess each individual timber member and each structural joint of a given structure; to provide the engineers with strength and stiffness values consistent with actual member's condition; to assess the effective residual sections of each member affected by decay. Minimum feasibility conditions: the same as for first level inspection, plus access with eyes and hands over the whole length of each member. Detailed visual inspection: if required, wood species shall be identified through microscopical analysis of fragments extracted from the wood member under assessment; wood defects shall be evaluated in any cross-section or length considered as relevant by the surveyor; the same for wood decay evaluation. Timber moisture content measurements: one or more measurements of wood moisture content will be performed with an electrical resistance moisture meter provided with insulated electrodes, in order to determine an average moisture content value and moisture content gradient, localise wet spots, condensation formation, etc. Residual cross sections assessment: through both visual assessment and basic instrumental techniques such as probing test, splintering test and sounding (beating with a hammer) test, the surveyor should be able to plot on the 'paper the effective member cross section(s) determining: pith location; heartwood and sapwood distribution; checks, ring shakes and shrinkage fissure's development; location and extension of decayed wood zones and/or cavities; knots and resin pockets location and extension. Joints assessment: timber-to-timber joints shall be assessed mainly through a careful visual control of decayed areas or disconnections; timber-to-other material joints shall be examined for compatibility aspects (as already described above). The assessment should be made on each joint or at least on a significant sample, upon agreement with the pool of experts. Evaluation of previous restorations (historical and/or recent): this task should be undertaken preferably together with other specialists, in order to assess the actual performance of restored structural parts, the compatibility between timber and other materials used, the expected serviceability and durability of the intervention and any recommendations for improvement. Hypothesis on future decay and/or damage development, including recommendations to prevent them (preservative treatments, measures against fire and against moisture uptake, etc.) Deliverables: the wood technologist shall deliver at least: a detailed map of wood decay, and a set of drawings indicating:

-

the graphical reconstruction of one or more relevant residual cross sections for each timber member;


-

the graphical reconstruction of actual joints;

a technical report including at least a detailed description of: -

the inspection results;

-

the timber structural quality (for each member);

-

recommendations for the prevention of future decay;

-

provisions to obtain serviceable and durable timber structures complying as far as possible with EC5 and/or other building regulations.

In-situ testing techniques A careful visual assessment, coordinated with basic test procedures such as probing, splintering and sounding and supported by a microscopic wood species identification, make it possible - for a skilled surveyor - to obtain results to the degree of accuracy normally needed by engineers and architects. Additional information can be in some cases obtained through non-destructive or slightly-destructive instrumental techniques. Nevertheless, it should be clear that measurements on timber members are always affected to a significant degree by uncertainty due mainly to:

-

variability of wood properties within the timber member;

-

oversimplified fundamental hypothesis of non-destructive methods in respect of the actual behaviour of wood material;

-

difficulties in practical measurement execution in-situ (and often also in the laboratory).

Serious surveying work will then have resort to non-destructive testing preferably as a subsidiary way to strengthen (or weaken) the hypothesis arising from the visual assessment. Commonly used instrumental techniques are the following: Static loading: measuring the modulus of elasticity of a member by static bending techniques is the foundation of machine stress grading of newly sawn timber. Also for in-situ timber members (and, with appropriate modifications, also for structural systems, for example floors), it is possible to compute an average modulus of elasticity by using well known fundamental equations and to infer strength values through the relationship existing between these two quantities. In some cases a clever approach may be the adoption of a sort of "unloading test", i.e. a procedure where the modulus of elasticity is derived by measuring the elastic spring-back of timber members after the removal of existing permanent loads. Endoscopy: this technique is sometimes used to observe hidden faces or even internal cavities of the wood member, hence shall be considered as an extension of the visual assessment. In connection with drilling techniques (see below) it can be used for the assessment of partially decayed cross sections. Thermography: often proposed by beginners in wood structure surveying, it is seldom useful, because of the very low thermal conductivity of wood. X-ray, gamma-ray, computer tomography, nuclear magnetic resonance: these STEPIEUROFORTECH

- an initiative


D3/7

techniques make possible a very accurate internal scanning of whole timber members. Today their use is restricted to a few laboratories and costs are too high for an extensive utilisation. Nevertheless, portable units intended for decay detection in standing trees and timber beams are currently in promising development. Vibrations: free and forced vibrations (transverse and longitudinal) have received considerable attention for non-destructive testing of in-situ timber members. The aim is the same as for static loading tests: to compute a modulus of elasticity (in this case from selected vibration frequencies and related amplitudes) and from it to infer strength values. Acoustic emission (AE): not extensively used on structural timber, it relies upon the application of stress to a member and the analysis of the stress waves generated by it. In connection with static loading techniques, it could be used to locate highly defective cross sections (DAE: defects acoustic emission, Bonamini and Togni 1994), under the hypothesis that failure in structural timber members may occur by stress concentration in small areas (for example, around a knot) far below the elastic limit of the whole section, causing a localised AE (Bonamini and Togni 1994). Stress waves (ultrasound, impact): speed of stress waves through the timber member is used either to detect decay (transverse propagation) or to compute a dynamic modulus of elasticity (longitudinal propagation). If accurate data about wood density are available, and coupling problems between wood and transducer's steel are carefully overcome, stress waves techniques can provide useful indications. Both conditions are often quite difficult to meet on site. Pilodyn: a hardened steel pin is driven into the wood by a spring-loaded device. Depth of pin penetration (after one or more blows, according to the instrument model) is used as a measure of degree of surface degradation. Results are affected by test location within the member, wood anisotropy, wood density, percentage of spring- and latewood, operator's skill. Drilling resistance: specially built electronic controlled drilling machines rely upon the relationship existing between wood density and the rate of penetration of the bit. These machines can bore up to 400 mm depth, automatically plotting on a chart a pattern from which density variations can be easily detected. Useful for internal decay assessment and for ring width assessment, these instruments give information related to the restricted area under test, not immediately extensible to large zones. Displacement transducers and strain gauges: these devices are in some cases used during loading tests in order to evaluate local strains, slip in joints, etc. Hardness: hardness tests on the faces of wood members can provide some information on local surface conditions of the material. Strength values for the whole member can be derived, through repeated tests on different points of the same member, only by very rough approximation. Screw withdrawal: the force needed for extracting a screw shows good relationship with wood strength at that point; hence this technique can provide useful information on surface or deep deterioration of members. The same limitations apply as for Pilodyn, drilling and hardness devices about STEPIEUROFORTECH - an initiative under the EU Comett Programme

the possibility of inferring strength for large members,

Conclusions Timber structures assessment is a complex task. Automatic devices capable of deriving from one or two measurements a complete set of strength and stiffness properties are not available, hence the visual assessment should be the first and the last step of the inspection work. Systematic approach, clear ideas about wood microscopical structure and timber macroscopical behaviour and last but not least a good deal of patience and scientific humility will help in reaching valuable results with inexpensive means.

References Bonamini, G. and Togni, M. (1994). DAE: defects acoustic emission. Un metodo non distruttivo per la localizzazione di sezioni altamente difettose su travi lignee in opera. Technical Report December 1994. Istituto di Assestamento e Tecnologia Forestale, Florence, Italy. Borgin, K., Parameswaran, N. and Liese, W. (1975). The effects of aging on the ultrastructure of wood. Wood Sci. Technol. 9(2):87-98. Cristelli, F. (1986). Caratteristiche fisico-meccaniche di legni antichi variamente degradati ed influenza dell'impregnazione con una resina sintetica. Dissertation No. 112. Universith degli Studi di Firenze, Facolth di Scienze Agrarie e Forestali, Corso di Laurea in Scienze Forestali, Florence, Italy. Ehlbeck, J, and Gorlacher, R. (1987). Erste Ergebnisse von Festigkeitsuntersuchungen an altem Konstruktionsholz. In: Erhalten historisch bedeutsamer Bauwerke, Sonderforschungsbereich 315, Universitat Karlsruhe, Jahrbuch 1987, Ernst & Sohn, Berlin, p. 235-247. Jessome, A.P. (1965). Strength tests on specimens from wood trussess in service for 97 years. Report of the Forest Products Laboratory, Dept. of Forestry, Ottawa, Canada. Kuipers, J. (1986). Effect of age and/or load on timber strength. In: International Council for Building Research Studies and Documentation, Working Commission W18 - Timber Structures, XIX Meeting September 1986, Florence, Italy. Rug, W. and Seeman, A. (1991). Strength of old timber. Building Research and Information, 19(1):31-37.


Restoring timber structures Repair and strengthening STEP lecture D4 L. Uzielli Universith degli Studi di Firenze

Objective To provide an outline of the follow-up procedures to inspection work, including a critical analysis of some commonly used techniques for in situ repairs andtor strengthening of timber structures, members and joints.

Summary This lecture describes the wood technologist's approach to the identification of the aims of the work to be undertaken, and of external constraints, in the repair and strengthening of existing timber structures, after an in situ inspection has been performed. It also discusses, briefly, some techniques that are often recommended. This lecture does not include structural design, which is covered by other lectures.

Introduction Following the inspection of the existing timber structure (see STEP lecture D3), decisions have to be made concerning any follow-up work that is necessary. Several aspects have to be considered, in close cooperation with experts, such as structural engineers, architects, wood technologists, restorers, historians, owner or administrators in charge of the building, and other concerned parties such as building authorities or officers in charge of conservation of the cultural heritage. Consideration of the following three aspects is recommended:

-

identification of objectives, requirements and constraints;

-

structural interventions;

-

maintenance and conservation measures.

Identification of objectives, requirements and constraints Restoration works on existing timber structures, which are often old or ancient and of some cultural importance, will relate to many different needs, which may often be conflicting. A clear identification of objectives and needs to be satisfied by the restoration work must be made before the technical aspects are considered, so that the experts may work on the basis of clearly stated priorities and constraints. Amongst others, the following alternative or complementary objectives of a repairtstrengthening intervention may be listed:

-

conservation of the appearance of the structure;

-

conservation of the original materials, for artistic, historical or cultural reasons;

-

restoration of the original loadbearing capacity;

-

increase or modification of the loadbearing capacity, stiffness, or serviceability;

-

compliance with safety, fire, seismic or other regulations.


On the other hand the constraints need to be considered, such as the following:

-

desires of the involved parties; economics;

-

environmental situation;

-

prevention of future decay and/or other damage;

-

available time, expertise, materials, workmanship and technologies; compatibility between existing timber and materials available for repair and/or reinforcement

For clarity and simplicity only a few of the preceding considerations are detailed. However, it should be appreciated that all the issues are likely to be interactive, and any action taken to solve one of them may significantly affect several others.

Conservation of the original materials and structural concept Because of artistic, historical or cultural reasons, the conservation of old timber structures is becoming more and more important and desirable, and often emphasis is placed more on conservation than on economic aspects. On the other hand (e.g. in many residential buildings) the lowest possible cost in order to obtain durability, serviceability and compliance with building regulations is frequently the main requirement: the cost of strengthening or repair becomes then the guiding parameter in choosing between various criteria for performing work on existing timber structures. Philosophy of restoration and rehabilitation is not an objective of this lecture, but should be considered when establishing the rationale of decisions to be taken (Tampone, 1992; Bertolini, 1992).

Specific roles of the wood technologist Some aspects of restoration work are peculiar to timber structures, and usually require the specific expertise of a wood technologist. His specific contribution may include the following subjects:

-

-

identification of wood species; evaluation of durability of the timber members in their present conditions (according to wood species, heartwood distribution, environmental conditions, etc.), and any need for preservative treatment in order to ensure the desired durability; assessment of the actual structural performance of the timber members.

The better knowledge provided by the wood scientist makes it possible nowadays to conserve structural members that until a few years ago would have simply been removed and substituted. Also, when the original structure does no longer meet the minimum safety requirements, a good knowledge of structural timber often makes it possible to take action so that old members may still contribute to the global loadbearing capacity. Further aspects that the wood expert should especially take care of are:

-

the compatibility between the wood of the member and the new materials used;

STEPIEUROFORTECH

- an initiative


-

the influences of the repair on the serviceability, reliability and durability of both the repaired members and the whole structure.

Conservation of the appearance of the structure The preservation of the appearance of the structure is related to the amount of degradation of timber members. Insects, fungi and fire normally affect the external wood layers, and repair works would often require their removal. When this is not possible or desirable, hardness and conservation of deteriorated wood may be improved through impregnation with appropriate resins.

Restoration of the original loadbearing capacity Timber members affected by heavy decay or damage can be strengthened through repair works. It should, however, be noted that many old timber structures are oversized, considering the structural requirements. After a detailed inspection and careful assessment it is therefore possible that (in spite of past damage suffered by the members) residual cross sections are still sufficient to provide a loadbearing capacity complying with present and anticipated service conditions. In such cases there should be no obstacle to report: "no repair work needed", allowing work to concentrate on the prevention of further decay.

Strengthening The loadbearing capacity of a timber structure needs in certain cases to be improved through appropriate structural consolidation, in order to comply with increased performance requirements (e.g. modifications in the use of the structure).

Compliance with safety, fire, seismic or other regulations It is a matter of fact that in some European countries building regulations have been written by experts with insufficient knowledge and experience of timber structures. The lack of specific regulations may need to be countered by reference to research papers pointing out the good seismic and fire behaviour of timber structures. A substantial improvement is likely to occur after the implementation in these countries of EC5 Part 1-2: "General rules Structural fire design", and EC 8: "De,sign provisions for earthquake resistance of structures".

Working conditions The available time, expertise, materials, workmanship and technologies should be clearly identified before the working plans are completed, in order to assure that the desired quality of the whole work may be obtained. Even more than for other kinds of repair works, timber structures require that no shortcuts be seeked, specially at the initial stages; also, the time required for collecting and analysing information should be allowed for, in order to avoid mistakes that would make successive remedies much more expensive or even impossible.

Compatibility Compatibility between different materials should always be a major concern; this is specially true for materials used for repairing and/or strengthening timber structures. Whenever glued or rigid connections are made, special care is needed in order to avoid stresses (and possible failures) caused by differential deformations related to the following:

-

differential shrinkage between wood and non-hygroscopic materials;

-

differential shrinkage between different anatomical directions of wood;


D4/3

-

different thermal expansion coefficients between wood and metals;

-

different displacements between structural parts or members with different stiffnessess.

Other compatibility problems may arise from chemical factors, such as corrosion of steel fasteners caused by wood extractives under high moisture conditions. Finally, condensation of moisture (and consequent decay problems) is often a result of poor compatibility situations, such as:

-

wood placed in close contact with metals or stone featuring a different heating or cooling rate, under insufficient conditions of ventilation;

-

possible condensation of moisture within timber pieces, at the interface between wood and resin.

Structural repairs Broad classification criteria In general, no two identical situations exist in old timber structures: therefore restoration works and repairs need to be chosen, designed and implemented case by case. Also, as already noted, problems and solutions are deeply interconnected and may be not faced separately. Nevertheless it is useful to lay down the following broad classification criteria, in order to help clarify meanings, scopes and limits of the numerous technical solutions which up to now have been, or may in the future be proposed for structural repair work. Repairs may basically deal with one or more of the following levels of the structure:

-

individual structural timber members; structural units;

-

whole structures;

-

connecting joints;

-

external constraints or connections.

After the repair works the original timbers:

-

may fulfil the same structural functions they were originally assigned (untouched, if recognised as needing no structural repair; after repair or strengthening, if needed);

-

may still fulfil a structural function, although in conjunction with newly added members;

-

may be left in place just for aesthetic reasons, or for conservation of the material's historical authenticity, the structural functions being totally fulfilled by other loadbearing members, such as substitution timbers, steel or concrete.

A note about strengthening Once the strength properties of timber have been lost because of decay or fracture, the original strength of the wood material can not be recovered (even to a partial extent) by means of impregnation with any kind of resin or other process. On the contrary, a timber member, sub-assembly or structure may be STEP/EUROFORTECH - an initiative under the EU Comett Programme

repaired, reinforced and/or stiffened, as needed.

Repair by gluing new parts Decayed or badly damaged segments (often beam ends) may be replaced by newly added parts (wood, glulam, epoxy, etc.), usually connected by glued rods made of steel or fibreglass. Structural design should mainly rely on adhesion of rods parallel to the grain of wood, since moisture variations and differential shrinkages may impair strength of gluelines perpendicular to the grain (Ceccotti, Mannucci and Uzielli, 1990). Appearance and authenticity of original materials are lost.

Figure 1

Replacement of inefficient segments of original members with epoxy. ( a ) Beam end replacement, (b) original timber, ( c ) steel or fibreglass rods.

Repair by means of traditional joints Decayed or badly damaged segments may be replaced by new parts made of solid wood, connected by means of traditional joining or repairing techniques. However, original strength may seldom be fully recovered. Only traditional methods and materials are used. Appearance and authenticity of original materials are lost.

Figure 2

Replacement of inefficient segments of original members by means of traditional joints; left, top to bottom : splice joint covered by bolted wooden plates, nailed spliced bevelled joint, bolted end joint, with steel channel; right, top to bottom: splice joint covered by bolted steel plates, shear reinforcement with nails or steel clamps, splice joint with internal steel plate.

Enhancement of cross-section The cross-section of a member may be enhanced by adding (gluing, nailing, bolting, etc.) reinforcing parts (wood boards, steel beams, etc.). Original appearance and aesthetics are usually lost.

Note that if the same reinforcement parts are simultaneously connected by more STEPIEUROFORTECH - an initiative under the EU Comett Programme

D4/5

than one technique (e.g. nails and glue, or bolts and glue), the strength of the two connections will not add up, because the stiffer connection such as the glue will take most of the load, and only after its failure the more ductile one will gain significance; it is therefore inappropriate to add nails or bolts "just in case" the glue failed, if they are not properly designed to take the whole anticipated load. However nails or bolts may be of great use in providing pressure and keeping members in position while the glue cures. Glued-in plates One or more steel plates (almost as deep as the beam) may be glued into grooves vertically cut in situ along the whole length of beam (see Figure 3). Plates, which are hidden and are protected from fire and corrosion, take up almost all the load. Special equipment is needed for cutting the grooves (Tampone, 1989). Geometrical feasibility, buckling, support conditions, etc. need consideration.

Figure 3

Glued-in plates, along the whole length of a white fir beam (a) (from Tampone, 1989, modified), (b) threaded steel rods, (c) steel plate, 10 mm thick, (d) epoxy resin.

Glued-in rods Steel or fibreglass rods may be glued into grooves cut along the tension edge of the beam. Failure is thus no more caused by strength-reducing defects located at tension edge, and occurs at compression edge, whose strength is not increased; failure behaviour also becomes more ductile (see Figure 4). Limit-state reliability is therefore increased, rather than strength or stiffness. (Ceccotti and Marradi, 1993). Tie-rods Steel cables or rods, equipped with spacers or other devices, may be used in order to contribute to strength and stiffness of individual members or trusses; by means of turnbuckles, the tension may be adjusted either to pre-stress beams or just to control excessive deflections (Marradi, Messina and Paolini, 1989). Periodic adjustments or insertion of elastic components may be needed to compensate for creep and for shrinkagelswelling caused by moisture variations (Ceccotti and Marradi, 1993). The structural conception of trusses may turn out significantly modified. Examples are shown in Figure 5 and 6.


Figure 4

Effects of steel or fibreglass rods glued with epoxy resin into grooves cut along the tension edge of the beam on its load-deformation behaviour. (a) reinforced cross-section, (b) timber cross section without reinforcement. Top diagram: beams with large defects, bottom diagram: beams with minor defects.

Figure 5

Tie-rods on individual members (from Tampone, Franci and Campa, 1989): (a) base plate for strut, (b) steel scar$ (c) cylindrical hinge, (d) clamp, (e) turnbuckle, (f)strut, ( g ) steel cable, left: detail in perspective view.

STEPfEUROFORTECH - an initiative under the EU Comett Programme

D417

Figure 6

Tie-rods on two types of trusses (from Marradi, Messina, and Paolini, 1989): the internal elements marked * have been added together with the tie-rods, significantly modifying the structural conception.

Modification of support conditions Supports of decayed parts (e.g. of beam ends) are moved to locations where timber is sound; this may also result in reducing the span of the beams. Often used for non-visible beams supported by thick walls, where brackets may be effectively fixed (see Figure 7).

Figure 7

Modification of support conditions: the decayed beam end ( a ) has been unloaded by moving the support towards the sound beam part ( b ) which rests on a wooden sleeper, which in its turn rests on an I-beam bracket ( c ) through a neoprene saddle, intended to prevent moisture condensation caused by temperature differences.

Additional loadbearing members Loads are partially or totally carried by additional members (steel or concrete beams, columns, etc.; see e.g. Bertolini, 1992). Structural conception is altered. Appearance and authenticity are partly lost.


Replacing timber members Some or all of the structural members may be replaced by new timber members, adopting the original techniques as far as possible. Great care is needed in considering technical compatibility; e.g. using insufficiently seasoned timber may lead to non anticipated unbalances or stresses. The correctness of replacing original parts in historically significant structures is questionable, since authenticity of materials is lost, whereas aesthetics and authenticity of conception might be conserved. The exceptional case of a six-stories timber-framed house (Knochenhaueramtshaus, Hildesheim, Germany), originally constructed in 1529, entirely destroyed in wartime and reconstructed according to historical techniques, reconciling historical claims with the design codes of the present day and vice versa, is reported by Kessel, Speich and Hinkes (1989).

Re-establishing or improving geometry and/or stability Excessive structural deformability or instability may be prevented or reduced by adding, repairing or re-establishing bracings (generally tension members are made of steel, compression members of timber). Authenticity is conserved, with possibly minor aesthetical alterations. Significant improvements are obtained in static and seismic behaviour, even though the original structural conception might result altered. Great care is needed in designing connections between parts with different anticipated displacements. Similar means as above may be used to re-establish correct geometry (e.g. to reduce or eliminate deformations, members "out of plumb", etc.). Great care and accuracy needed in designing and implementing the work, choosing and manufacturing anchorages, etc. Structural conception might result altered.

Timber-concrete and timber-panel composite structures An effective kind of work, already implemented in a number of cases in different countries, is the timber-concrete composite technique (see STEP lecture E13): a reinforced concrete slab is connected to the beams by means of shear connectors (various types exist, such as glued, screwed-in, fitting in grooves, etc.), so that it is not just supported by beams, but cooperates as well in resisting bending moments (slab is in compression and beams are in tension). Bending strength and stiffness greatly increase, and seismic behaviour improves. Beams need to be in good conditions, in order to contribute in the mixed structure. Timber-panel composite technique is similar as above, except that structural wood-based panels, connected by means of nails or similar fasteners, are used in place of the concrete slab. This system, not yet widely used in rehabilitation of old structures, is much lighter than concrete; on the other hand, it provides a smaller structural improvement.

Maintenance and conservation measures Maintenance work should always carried out with a view to the continued conservation of the structure: no work may be considered as "the final one", needing no further care or maintenance. The action of potential deterioration agents (both biotic and abiotic) should be anticipated and prevented.


Moisture, in its various forms and origins, including the effect of alterations that may have been made on the environment (vapour barriers, waterproofing, sources of condensation, increased or decreased ventilation, closed windows or other openings, sealing of beam end supports, etc.) should always be considered a major threat to the conservation of timber structures, including those which have undergone recent or earlier repair works. Special care should be taken to ensure the proper execution of recommended repair or prevention works. For instance an inaccurately performed preservative treatment (e.g, one leaving unprotected surfaces), or the inappropriate succession of works (e.g. drilling or notching timber after it has been treated, resulting in untreated wood surfaces to be exposed to the attack of insects) will hardly improve the durability of the structure.

References Bertolini, C. (1992). Problemi di recupero: metodologie di indagine, tecnologie di intervento. L'Edilizia, (12), VI, 763-778. Ceccotti, A,, Mannucci, M. and Uzielli, L. (1990). Effetti del riassorbimento di umidith sul comportamento ad estrazione di barre di acciaio ancorate nel legno mediante resina epossidica. In: G. Tampone (Ed.) Atti del I1 Congresso Nazionale "I1 restauro del legno", Firenze, Nardini, Vol. 2. 155-169. Ceccotti, A. and Marradi, P. (1993). Nuove tecnologie negli interventi di recupero delle antiche capriate di legno: materiali e metodi. In: Proceedings of the International Symposium "Ancient Buildings Restoration - Handbooks and new technologies", Naples, 29-30 October 1993, 514-536. Kessel, M.H., Speich, M. and Hinkes, F.J. (1989). The Reconstruction of an Eight-floor Timber Frame House at Hildesheim (FRG). In: G. Tampone (Ed.) Atti del I1 Congresso Nazionale "I1 restauro del legno", Firenze, Nardini, Vol. 1, 217-222. Marradi, P., Messina, C. and Paolini, L. (1989). Recupero di strutture in legno mediante armature parzialmente presollecitate. In: G. Tampone (Ed.) Atti del I1 Congresso Nazionale "I1 restauro del legno", Firenze, Nardini, Vol. 1, 241-248. Tampone, G. (1989). Restauro strutturale con lamine metalliche dei solai lignei della sede del Genio Civile di Firenze. In: G. Tampone (Ed.) Atti del I1 Congresso Nazionale "I1 restauro del legno", Firenze, Nardini, Vol. 1, 263-281. Tampone, G. and Campa, L. (1989). Restauro strutturale con legno lamellare di un solaio e di una volta a carena lignei dell'Accademia di Belle Arti a Firenze. In: G. Tampone (Ed.) Atti del I1 Congresso Nazionale "I1 restauro del legno", Firenze, Nardini, Vol. 1, 283-297. Tampone, G., Franci, F. and Campa, L. (1989). Rinforzo di puntoni e consolidamento di una capriata del teatro di Sarteano mediante centine metalliche. In: G. Tampone (Ed.) Atti del I1 Congresso Nazionale "11 restauro del legno", Firenze, Nardini, Vol. 1, 299-305. Tampone, G. (1992). Tecnologia del restauro delle strutture di legno. L'Edilizia, (12), VI, 729-739.

STEPIEUROFORTECH

- an initiative


Column to beam and beam to beam connections STEP lecture D5 G.Vidon Socotec

Objective To describe the main types of connection used, to point out the design problems to note and to present examples of the design of various types of connection.

Summary The main types of column to beam connections, hinged or fixed, as well as beam to beam connections are described. The principles of design of these connections to resist shear, axial force and bending moment are described. The design principles concerning problems specific to wood such as compressive or tensile strength perpendicular to the grain or dimensional changes are also covered. Actual examples are presented.

Introduction The design of a connection must allow the function selected to be carried out (i.e. hinged or fixed in one or more directions). The connection should be designed to resist all the internal forces and moments about all three principal axes:

-

shear forces in y- and z-direction,

-

axial loads in x-direction,

-

torsional moments about the x-axis,

-

bending moments about the y- and z-axis.

The internal forces or moments in the connection are either balanced by mechanical fasteners such as nails, dowels or bolts or by direct wood to wood contact. Glued connections are not dealt with in this lecture. In general the main internal forces and moments are those which occur in the plane of the frame made up of a column and a beam. The other internal forces and moments essentially arise from the additional forces applied out of the plane of the frame, especially the bracing forces. In any case, in a column to beam connection, the torsional moment must be counterbalanced by some means to prevent the rotation of the beam around the x-axis.

Figure 1

Beam axes.

Figure 2 shows the main types of connection. Connection type 4 in Figure 2 is either free to rotate about the y- and z-axis or continuously fixed in three axes. In connection type 5 the secondary beam is usually simply supported by the main beam. allowing rotation about the y- and z-axis of the secondary beam. STEPIEUROFORTECH - an initiative under the EU Comett Programme

D5/1

Because of the eccentricity of the secondary beam support the main beam is often subject to a torsional moment (see Figure 3).

Figure 2

Main types of connection. ( I ) Corner connection, column to beam or frame corner, ( 2 ) connection between continuous beam and column, (3) column to beam connection in a multi-storey structure, (4) beam to beam connection: two beams in line, ( 5 ) beam to beam connection between a secondary beam and a main beam.

Figure 3

Eccentricity of secondary beam support.

Connection examples Column to beam connection type 1

Figure 4

Joint with steel bars with bolts or screws and with or without connectors. The bearing plate (e.g. of plywood) between the column and the beam centres the load and prevents overloading of the corner of the column.


Generally these connections allow free rotation about the y-axis. The support reaction of the beam is transmitted either by direct contact or by mechanical fasteners. If the column is wide, neoprene sheets may be provided in order to keep the load centred on the column. Both the beam and the column must be torsionally restrained.

-

Figure 5

Joint with nailed gusset plates of steel or plywood. The joint can take up some moment and thereby contribute to the lateral stability of the structure.

Figure 6

Joint with a vertical bolt screwed into a round steel bar with a threaded hole. The hole in the column is plugged after assembly. The bolt should be retightened, especially for deep beams.

Figure 7

Joint with glued-in steel bolts.


Figure 8

Dowelled steel plate in slots in beam and column.

Figure 9

Forked beam bearing on twin column.

Column to beam connection type 2 This type of connection is often used in column to beam structures at right angles between continuous columns and continuous beams supporting floors.

Figure 10

Spaced double beam connected by dowels or bolts to the column.

The connection in Figure 10 needs a large connection area to allow the necessary number of fasteners to be placed.

Figure 11

Double beam connected to indented columns using steel angles to increase bearing surface.


Figure 12

Central beam resting on spacing blocks of double column. Right: the bearing sur$ace is increased by using a U-shaped steel support.

In Figure 11 and 12, the bolts are placed in oval holes to allow for dimensional changes in the timber and are only used to position the beams. To avoid large compressive stresses on top of the beams in Figure 12, a clearance is provided between the upper face of the beam and the spacing block of the double column.

Column to beam connection type 3 This type of joint essentially concerns the connection of a cross member to a continuous column. Except for traditional jointing by means of mortise and tenon, these connections are made using metal fittings.

Figure 13

Nailed plywood or steel gusset plates. The joint is effective and easy to make. It may be necessary to protect the gusset plate against fire.

Figure 14

Gusset plates in slots with nails (plywood gussets) or dowels (plywood or steel gussets). The fire properties are very good.


D515

.

. --

-

..

.-

.. .- --

-------

q b z =,----s - - - + ;---[g --: :i~.jji n

I

iI

1 1

d

L=\,

1

i I

,

1

h

$,k==$ z = : ~ =$ = r z l I I I

I-----I 1

I

1

,

; I

p

'

I

I

1-z

I 1

>:. '

??;,

LX-.

-

1

I L-->

1

i 1;

'- - - _ _ _ _ _ _ - -

I

'

:s>ji; I

1

I

I

1

I

I

)

. I

I I

,

1

Figure 15

Beam between two columns. Part of the load can be taken by a support block.

Figure 16

Beam shoe or bracket. The column will be loaded eccentrically.

Figure 17

Traditional connection by mortise and tenon with hardwood dowel. The support reaction is transmitted by direct contact; the hardwood dowel only keeps the beam in place. This connection is attractive but needs to be carried out by specialised craftsmen and is only suitable for low loads.


Beam to beam connection type 4 Purlins are often designed with cantilever connections or with continuity over the supports. Both solutions are advantageous compared with simple beams on two supports. The necessary timber volume is decreased and the stiffness of the structure increased. Cantilever connections (Figure 18 to 21) are very economic in labour and time.

I 1 I I II I

I

Figure 18

The simplest joint: the short beam is hanging in the cantilever. Where the forces are not too large, the beam can be supported directly on the cantilever.

Figure 19

For large forces it may be necessary to reinforce the joint with steel channels.

Figure 20

Joint with a special steel shoe. In the version shown, only shear forces can be transferred.

Figure 21

Joint with doweled steel plate in slots. The dowels are placed close to the unloaded edges to prevent splitting.


D5/7

Figure 22

Continuous connection with wooden battens on a support.

Figure 23

Continuous connection with crossed purlins on a support.

Beam to beam connection type 5 This type of joint essentially concerns the connection of a cross member to a continuous main beam. Except for traditional jointing by means of mortise and tenon, these connections are made using metal fittings.

Figure 24

Beam shoe.


Figure 25

Beam from one side only supported on a steel bracket. In this case it is necessary to design the main beam for the torsional moment.

Common problems to be avoided Due to slipping and rotation of the connection, the forces which must be counterbalanced by the fasteners act in the tangential direction of the rotation circles through the bolt lines and may not be parallel to the grain (see Figure 26). The fastener force components perpendicular to the grain, which are resisted by the stiff metal side plates or channels, cause tensile and compressive stresses parallel to the grain. The side plates or channels are very rigid and tend to split the wood. In order to avoid this problem, the stiffness of the metal fittings should be decreased at the joint.

Figure 26

(b) (c) Splitting failure due to stiff steel channels used to connect the beam ends.

The net area indicated in Figure 26c has to be large enough to resist the force F in order to avoid a local tensile failure parallel to the grain. For very deep glulam beams the dimensional changes in the wood due to moisture content changes may cause splitting of the timber, if free shrinkage is prevented (see Figure 27). The column in Figure 27, less thick than the beam and better ventilated, dries out more quickly and its shrinkage is hindered by the circle of bolts which attach it to the beam. Since the beam hardly shrinks in grain direction, a split occurs in the centre area of the circle of bolts.


Figure 27

Splitting failure due to differential shrinkage and different moisture variation in beam and column.

For secondary beam to main beam connections it is necessary to place the steel connector of the secondary beam as high as possible to limit the tensile stresses perpendicular to the grain. In practice, if the fitting covers 70% to 80% of the main beam depth, the tensile stresses perpendicular to the grain are not governing the design. However, it will be necessary to check that the fitting does not prevent dimensional changes of the main beam which would cause cracks due to resisting the shrinkage effects.

Figure 28

Connection of secondary beams to a main beam by mechanically connected hangers.


Special connections using steel plates STEP lecture D6 E. Aasheim The Norwegian Institute of Wood Technology

Objective To describe different ways of connecting and supporting timber elements in arches and frames by using steel plates.

Summary The lecture describes the different principles to be followed when designing hinges and supports based on steel plates. The design of the details is discussed, and examples of actual hinges in existing timber structures are presented.

Introduction As pointed out in STEP lecture C1, the serviceability and the durability of a timber structure depend mainly on the design of the joints between the elements. This statement includes the hinges and supports, which in large structures are often connections between two glulam parts.

Basic considerations The selection and design of connections are controlled not only by the loadcarrying and durability conditions, but include other considerations such as aesthetics, the cost-efficiency, the fabrication and the erection. A basic requirement is that all steel details shall be well adjusted to the glulam parts, to avoid time-consuming and costly work at the building site.

It is very important to design the connections in such a way that shrinkage and swelling of the timber parts are possible without creating problems. The moisture content in the glulam during production is normally very well controlled, but the equilibrium moisture content in timber will vary during the year. If free movement due to shrinkage is not allowed, the result may be splitting of the timber caused by tension perpendicular to the grain. The design should avoid the possibility of water being trapped in the joint area, and if necessary drainage holes or slots should be introduced. It is essential to protect end grain from water, because the water absorption parallel to the grain is much larger than the absorption perpendicular to the grain. In many cases a moisture barrier is recommended or gaps may be introduced. If exposed to the weather, or other severe conditions, corrosion of the steel parts may be resisted by rust proofing or by using corrosive-resistant metals. The designer should also consider the compatibility of the metal with the timber treatment. For example, as pointed out in STEP lecture C1, caution should be taken with the installation of steel components into timber treated with preservatives containing copper.

Beam-column and beam-beam connections These connections are covered by STEP lecture D5. STEP/EUROFORTECH - an initiative under the EU Comett Programme

Hinges for frames and arches For hinges in the apex of frames and arches the details shown in Figure 1 may be used.

Figure 1

Hinges for frames and arches based on steel plates.

The details in Figure 1 are applicable in Service Classes 1 and 2 according to EC5. The hinge in Figure l a may be used for frames and arches with slopes of 40 degrees or more. A bolt with a diameter of at least 20 mm should be used with nails as indicated in the figure. In the detail shown in Figure l b the bolts may be reinforced with single-sided toothed-plate connectors, if heavy lateral tension forces occur.

Figure 2

True hinge for frames and arches based on steel plates.

The hinge shown in Figure 2 is a true hinge, which may be used in Service Class 3. STEPIEUROFORTECH - an initiative under the EU Comett Programme

Column supports For simply supported columns, the details shown in Figure 3 may be used.

Figure 3

Column supports based on steel plates, (a) for vertical and horizontal forces, (0) for vertical and horizontal forces and moment about the strong axis of the glulanz member.

The supports in Figure 3 are applicable in Service Classes 1 and 2 according to EC5. In Figure 3a the compression force is transferred directly through contact pressure and not via the bolt. The connectors indicated are "single-sided toothedplate connectors" as described in STEP lecture C10, for example "Bulldog" connectors.

Supports for columns, frames and arches For supporting light frames and arches, or for pin-ended columns, the details shown in Figure 4 may be used.

Figure 4

Column supports or supports for light frames and arches based on steel plates.

The supports in Figure 4 are applicable in Service Classes 1 and 2 according to EC5. The horizontal and vertical forces are transferred through contact pressure between timber and steel. Horizontal forces acting outwards and lifting forces are transferred through the bolt. The bolt may be reinforced with single-sided toothed-plate connectors. STEPIEUROFORTECH - an initiative under the EU Comett Programme

D613

The need for constructional tolerances when casting concrete bases is essential. A moisture barrier is necessary to avoid moisture transfer into end grain.

Supports for frames and arches For simply supported frames and arches the details shown in Figure 5 may used.

Figure 5

Supports for frames and arches based on steel plates.

The support in Figure 5a is applicable in Service Classes 1 and 2 according to EC5. Compression forces are not transferred through the bolts, and single-sided toothed-plate connectors may be added. The support in Figure 5b is a true hinge anchorage, which may be used for outdoor exposure (Service Class 3).

Design of hinges and supports All the steel plate based hinges and supports described in this lecture must be designed separately according to EC5. General joint design is discussed in STEP lecture C1. Welding should be checked to accord with EC3. The resistance to corrosion should meet the protection specifications in EC5 Table 2.4.3.

Examples An example on an apex hinge is shown in Figure 6.

Figure 6

Hinge example.

STEPIEUROFORTECH

- an initiative


The detail in Figure 6 is used in a warehouse building in Norway. The slotted-in "connector" is made of steel plates with thickness 8 mm. The outer steel plates (8 mm x 200 mm x 850 mm) are connected to the glulam members with four 20 mm bolts on each side of the hinge. The support shown in principle in Figure 5b is used in many structures in Europe and other parts of the world. Figure 7a shows an example from a Norwegian structure.

Figure 7

Support examples.

The detail shown in Figure 7b is used to support the arched trusses in H&on's Hall Lillehammer, which was the main ice-hockey arena during the 1994 Winter Olympics in Norway. Slotted in steel plates are welded to the upper part of the steel support, and the glulam parts are connected to the steel plates using steel dowels.

Design example Glulam frame with a cross section b x h = 180 x 900 mm at the support. Strength class GL32 according to prEN 1194.

Figure 8

Support design example.

V, = 250 kN,

HA = 300 kN, medium term loads, Service Class 2, y, = 1,3


Vd = HA sin a - VA cos a

=

300

. sin 70"

- 250

HA cos a

=

250

. sin 70"

+

Nd

=

VA sin a

Figure 9

+

300

. cos 70" . cos

=

196 kN

70" = 338 kN

Static belzavior.

Characteristic material properties according to prEN 1194: fv.g,k

= 3,5 N/rnm2

fc,0,g,k

= 29 N/mm2

fc,90,g,k

= 6,O N/mrn2

Design strength values:

fc,0,d

- 'mod -

&~,k

YM

-

038 . 29 193

=

17,8

~

1

~

Shear stress:

Verification of failure condition:

The axial force is transferred through the steel foot plate:

Verification of failure condition: Oc,0,d

I.6,,0,d


~

2

The shear force is transferred through the steel side plate:

Verification of failure condition:

The moment V, h, should be transferred through the steel dowels, see STEP lecture C 1 . The steel parts should be checked according to EC3.

Concluding summary -

The serviceability and the durability of a timber structure depend mainly on the design of joints and supports.

-

Design the joints and supports in such a way that shrinkage and swelling of the glulam parts are possible without creating problems.

-

The details must be designed separately according to EC5, see STEP lecture C1.

STEPIEUROFORTECH - an iniliative under the EU Comett Programme

Transportation and erection STEP lecture D7 G. Bignotti Holzbau AG-Spa Brixen

Objectives To illustrate the modes of transport for timber construction elements and the most frequent methods of erection for various timber structures.

Summary Transport and erection influence the design and the fabrication of timber structures. The lecture describes the means of transport for various types of structural elements and the erection modes for the different structural configurations. Some suggestions are provided for designers of timber structures so as to optimise costs, quality and safety on-site.

Introduction The final phases in the provision of a timber structure are transport and erection. These may appear insignificant in the realisation of a project but they require the same attention as the preceding phases, in that together they can influence not only the design but also the budgeting and the management of the project.

Figure I

Transport and handling of glularn arches.

Transportation Timber constructions can be built using solid wood, glued laminated timber, plywood or other wood based panels. Solid wood and panel elements will normally have dimensions that can be easily transported. Conversely, glulam structural elements can be manufactured in very large sizes and in a variety of shapes. As a consequence transport solutions are correspondingly diverse. In general the transportation of glulam members does not present loading capacity problems (since their density is only approximately 500 kg/m3) but often the length and the shape of the elements present difficulties. In these cases the ability to transport the structure must be verified at an early design stage both in terms of equipment and route. This is important in estimating costs and in making a rational choice between different types of structural system.


D711

In general glulam beams are long, with a deep, narrow cross section. They can be manufactured up to approximately 45 m long. The length and the overall height determine the best transport solution. The width of the beam, including also the room for any pre-fixed steel hardware (shoes), determlnes the number of units that can be loaded without exceeding the width of the truck (2,5 m). When the overall height of the element is relevant, it is important to gay attention to the limits imposed on the free passage of the vehicle to the site due to the dimensions of viaducts and bridges (clear height of 4,O to 4,8 m).

Figure 2

Transport of a 40 m long glulam beam.

Elements with an overall height exceeding 4,O m are not transportable by road or rail. In these cases structures have to be designed so that the beams can be manufactured in two pieces. They can then be end jointed at the site with steel or timber bolted connections.

Loading The operation is usually carried out using a gantry-crane for the structural elements and a fork-lift for the accessories. The loading operations must be carried out in accordance with safety regulations and the goods must be well secured on the truck in order to avoid problems during transportation.

Means of transport

Figure 3

Three-axle truck with roof trusses.

For domestic deliveries transport is mostly done by road haulage with a few exceptions. Rail transport, for example, is mostly used in a situation where large quantities of standard beams need to be delivered to stock warehouses (maximum length allowed in one wagon, 18 m). For export deliveries it might STEP/EUROFORTECH - an initiative under the EU Comett Programme

be convenient to ship the timber by sea freight in container vessels (permitted length, 11,90 m) or in traditional vessels where the beams are loaded into the hold or are packed on the deck (no restrictions).

Exceptional transportation When the beams' dimensions exceed the dimensions of the truck body the load is considered exceptional and a special transport solution is required. Where possible, a regular truck should be used by letting the beams hang out over the back of the truck but not further than 3 m or 3110th~of the length of the truck. When this solution is not possible, special trucks have to be used.

Exceptional transports with special vehicles

Figure 4

Different forms of special vehicles.

Transportation of very long or high beams requires special vehicles that can be classified in four groups:

-

Tractors with low-bed telescopic semi-trailer (Figure 4a): these are suitable for short and very cambered beams.

-

Tractors with telescopic semi-trailer (Figure 4b): these are suitable for beams up to 3,O m high and up to 24 m long. The semi-trailer telescopic lattice can be extended up to 20 m.

-

Tractors with low-bed trailer (Figure 4c): these are suitable for beams with an overall height up to 4,O m. The trailer is in two parts connected by a non-loadbearing lattice. The beam's large overall height can be slung very low in the trailer which permits the beams to travel only 400 mm from the road level. It must be pointed out that in these cases the road's path has to be carefully checked.


D7/3

-

Tractors with cradle trailer (Figure 4d): these are suitable for any sort of beams with a height up to 3,5 m. The beams are supported by the front centre plate on the tractor and by the back centre plate on the trailer. Centre plates permit articulation between load and conveyance.

Limitations Road regulations prescribe some limitations to exceptional transportation. When the length to be conveyed exceeds 25 m or the width exceeds 3 m, a police escort is required. Regulations can be different in the individual countries. In Germany, when the overall height exceeds 4,O m, transportation has to be done during the night. In Italy, the number of units of loads exceeding the legal limit of 43 to 44 tonnes must not exceed three pieces per convoy. In general a special pass must be obtained from the transport authorities whenever timber elements exceed legal dimensions.

On site Unloading The lifting system available on the site is normally used for unloading. Because transport methods do not include a lift (except for a few trucks) it is necessary to use the site jib crane or a truck crane. Beams are slung with special bands to avoid damage. The bands are slung and tightened around the beams in order to avoid slippage while moving. Handling should always be carried out with the utmost care to avoid possible damage to timber. In the case of trusses, wherever possible the points of lifting should be at the eaves joints, with the truss in the vertical plane, apex uppermost.

Storage

Figure 5

Storage of glulam beams.

After unloading at the site, the timber and its accessories must be stored until assembly takes place. During this period the members and the other materials must not be damaged and they must not represent a hazard to people. Stored members have to be protected from moisture coming from the soil and from rain and snow. For these reasons the elements must be stored to avoid contact with the ground and should be protected by a waterproof cover. If timber has to be stored for a long time in wet conditions, in order to avoid vapour condensation STEPIEUROFORTECH - an initiative under the EU Comett Programme

under the covers, it is important to provide adequate ventilation. This avoids the wood getting stained and mouldy especially when it has not had a preservative treatment. Where large size structural elements have to be stored in the upright position, it is necessary to stabilise them to prevent buckling.

Machining Timber elements (beams, columns, purlins, etc.) normally arrive on the site machined and finished where the beams are surfaced on all sides and their ends are trimmed to a precise length. It is preferable to avoid any activity on site that could have been carried out in the factory at a lower costs and with higher quality. Accurate shop drawings must specify for each timber element how they have to be machined; in the factory rather than on the site. They should also be drilled, slotted and grooved to accommodate connecting hardware. To carry out all this pre-assembly work, it is necessary to develop very accurate shop drawings showing all the details and connections. When possible, a sensible reduction in the erection costs can be achieved by pre-assembling steel connecting parts on the beams. Unless otherwise specified an anti-fungi and antiinsect coating, available in different shades, is brushed on. Nothing should be improvised on the site. Only when site conditions do not permit pre-fabrication of the structural elements, should they be trimmed and machined on-site. The work in this case must be carried out in accordance with highest standards. The construction will be, as a consequence, more expensive.

Site inspection Before erection it is appropriate to check if the site is ready to receive the timber and if foundations and supporting elements are installed properly according to the drawings. It is advisable to check if all materials are on site before assembly begins.

Erection Erection represents, in the construction of a timber structure, the moment in which all the previous stages are verified. A well executed assembly is a primary condition to achieve a good building but it is not a sufficient condition. It is necessary that all the previous phases (design, engineering, fabrication, machining and transportation) are correctly executed, to be followed by skilled erection of the structure. Workmen have to be skilled, with experience in timber construction and with complete and efficient tools and machines. The site must be ready for receiving the timber structural elements in order to allow workmen to operate easily and with safety. Erection of a timber structure is usually carried out by placing first the main beams and then the secondary elements. Due to the fact that these frequently have a role in bracing the structure, provisional bracing must be provided. As soon as possible bracing wood rafters and counter braces have to be connected to the main beams. Assembly procedure depends on the type of structural system. It is useful to consider some typical situations:

Domestic roofs There are two fundamental types of timber roof for housing and similar types of buildings. The first type is characterised by steep pitches and consist of solid wood trussed rafters positioned at close centres. In most cases the structure is hidden by the ceiling. The speed with which trussed rafter roofs can be erected, along with their low cost make this solution used in the majority of domestic STEPIEUROFORTECH - an initiative under the EU Comett Programme

D7/5

roofs. Normally the greatest stresses which truss joints will undergo are those caused by handling. It is important to carefully plan the handling of trusses, taking into account weight, size, access, lift height and whether manual or mechanical handling is required. Temporary raking braces are used during the erection until the last trussed rafter is erected. After that permanent diagonal braces are fixed on both sides of the roof, longitudinal members are fixed making sure that the ceiling ties are accurately spaced at the correct centres. The second type consists of main beams and purlins. These structures have modest dimensions but, due to the high level of finish required, they need to be assembled accurately. Frequently it is preferabIe to hide the steel hardware. This involves more sophisticated connections systems in which the beams have to be machined to insert the steel parts, or indented on the sides to support purlins. As a consequence, machining and assembly are more expensive, but the results are excellent. Scaffolding and jig cranes are normally required on-site.

Beams They are used normally for simple and linear structures. Purlins are frequently supported by standard steel shoes nailed onto the beams. They can be preassembled in the factory or nailed on-site. If the ground under the structure permits it, it can be convenient to use a mobile rise tower.

Three hinged portals

Figure 6

Erection of three hinged portals.

This system is used very frequently because it permits coverage of large areas with a simple and economic structure. Each half portal can be made of a single curved piece or of some straight pieces jointed together. In the case of curved beam portals, it can be highly convenient (for economy and for workers' safety) to pre-assemble on the ground pairs of arches complete with purlins and bracing. They can then be lifted from the centre of the construction allowing them to rotate through the base hinges. The two semi-portals must be lifted beyond their final position so that they czn be dropped down to meet at the top hinge with no interference. A workman operating from a rise tower or from a platform crane fixes the top hinge. This erection system requires the use of two cranes, but it is fast and does not require a central scaffolding to support the semi-portals until they become load bearing. STEPIEUROFORTECH - an initiative under the EU Comett Programme

With highly skilled workmen it is possible to pre-assemble four or five semiportals on the ground and to lift them up simultaneously. Two auxiliary beams are necessary to hang the portals from the crane (see Figure 7).

Figure 7

Erection of groups of pre-assembled portals.

Trusses

Figure 8

A pair of coupled trusses ready to be shifted on rails.

By assembling size limited structural elements, it is possible to build long span trusses (up to 100 m) that solve transportation or production problems. As a result assembly becomes rather heavy because the various members must be jointed and connected on site. Fastening is often carried out using dowelled and bolted connections. Dowels must be hammered and for this reason the truss has to be laid down on one side. The use of bolts requires the truss to be supported clear of the ground. Normally the assembly area available on-site is limited and movement of the trusses by crane is expensive. So, often it is convenient to place rails running along the two trusses support lines. In this way the trusses can be placed in the position nearest to the assembly area and then they can be shifted using trolleys to their final positions. Obviously, secondary beams and bracing must be assembled between trusses before shifting, for reasons of STEPIEUROFORTECH - an initiative under the EU Comett Programme

D7/7

stability

Radial and geodetic domes There are two types of domes: radial and geodetic. When their diameter is under 80 meters they can be erected as a series of radial arches connected at the top. When they are of larger dimensions a geodetic design system is normally used. Their general behaviour is that of a "shell" structure configuration. Assembling a radial dome involves building a scaffolding tower at its centre as support for the steel compression ring to which the top part of each arch must be connected. In geodetic domes a large number of provisional supports is required.

Figure 9

Erection of a radial dome.

Safety Erection of timber structures requires in the majority of cases that great attention is paid to the prevention of accidents at work. From the first project stage it is necessary to consider how the various elements will be assembled. The structure must be designed taking into account systems and connections that reduce risks for workmen. Erection drawings and instructions must be prepared for important constructions. They have to describe unloading, storage and erection of the timber elements. Instructions must be given clearly and it is preferable when they are illustrated by sketches. When possible, preference has to be given to erection systems which permit pre-assembly on the ground. When it is necessary to operate at height, necessary safety measures must be imposed. To protect workmen from the danger of falling from the erected structure holding nets or pre-anchored wire lines should be available. Wire lines can be fixed through special posts pre-anchored to main girders before lifting. Workmen who have to walk on the erected beams, must tie their safety belt to the wire lines. Other facilities that can be used to reach the roof level are fixed or mobile scaffolding and crane-platforms. In addition to safety belts, workers must wear personal protection aids as hard-hats, gloves and shoes with steel toe and anti-punching sole. These safety means together with good training of the workmen and with a correct erection design, permit competitive assembly without neglecting safety requirements.


Figure 10

Use of safety nets.

Figure 11

Pre-anchored wire lines on a dome structure.

Concluding remarks -

Transportation and erection of timber structures, particularly long span or large structural elements, should be considered at an early stage as they can affect the design and cost of the project.

-

Transportation of exceptional loads should conform to the requirements of the relevant highways authorities.

-

Machining and all pre-assembly preparation should be carried out in the factory where possible.

-

Erection of timber structures will have to reflect site conditions and available lifting equipment.

-

Good erection practice to ensure safety on site is essential


Frame corners STEP lecture ~8 P. Racher CUST Civil Engineering Blaise Pascal University

Objectives To describe the types of frame corners commonly used depending on the structural form and the jointing technique. To present design optirnisation of the momentresisting knee joints allowed by EC5.

Summary Three different types of frame corners are presented:

-

tapered members with moment-resisting joints,

-

V-shaped columns or knee brace members,

-

knee-joints using large finger joints or glued-in bolts.

The applications of these frame corner systems are described and some details are given to ensure the structural performance of the frames. In the second part, a design example of a frame corner presents the possible choices for optimising the dowelled moment-resisting joints commonly used in Europe.

Introduction Industrial and recreational buildings are often built using a frame with tapered members, or curved frames, as the main structural system. They are simple systems which permit large areas to be covered with spans ranging from 15 to 50 rn and with spacings of 5 to 10 m.

Figure I

Frames with tapered members (a) and curved frames (b): ( I ) knee joints, ( 2 ) secondary structural roofing elements.

The most critical action effects (forces and moments) are found in the frame comer which has to be designed first. This preliminary design defines the largest crosssection for curved frames, or the knee-joint and the largest cross-section for the tapered members. Though curved frames are mechanically more efficient, their use requires:

-

special attention to transportation or functional requirements,

-

the design of secondary elements (see Figure lb) depending on the shape of the building and the roofing materials.

Because of these disadvantages, the choice of a frame with tapered members is generally preferred. In this case, the designer is faced with the transfer of the moment and forces between the rafter and the column. STEPJEUROFORTECH- an initiative under the EU Comett Programme

D8/1

Frame corner systems Mechanical fastening techniques The most common frames are built, for spans up to 50 m, using tapered members: a single rafter and paired columns (Figure 2a). They are three-hinged to avoid overload due to globally imposed deformations. At the frame corner the internal forces are transferred by using mechanical fasteners placed in a circular or rectangular pattern. To accommodate the fasteners, the mean depth of the members in the area of the joints varies between Ll20 and Ll30, where L is the span of the frame. The main disadvantage of this moment-resisting joint is the possibility of splitting induced by moisture content variations or time effects on the joint behaviour. When using dowels, the splits could be reduced by installing dowels reaching ductile failure modes I1 or I11 (see STEP lecture C3). In case of toothed-plate connectors, the extent of the splits is limited by staggered location of the connectors in the inner pattern relatively to the outer pattern (Figure 2b). For both types of fastener, reinforcement glued at the end of the members reduces the risk of splitting (Figure 2c).

Figure 2

Frames with mechanically jointed members: ( a ) typical frame corner, (b) toothed-plate connectors, (c) reinforcement of members: 1 glued-in rod or 2 glued-on plywood, (d) special arrangement for dowelled joint.

To reduce the stresses near the end-grain of the members, the mechanical behaviour could also be controlled by positioning a stiff rod at the rotation centre. The bending moment is transmitted by a partial pattern of fasteners located along the line of thrust in the area of the joint (Figure 2d). Furthermore, the current trend is to develop knee joints with higher strengths and stiffnesses and with ductile behaviour. In European countries, this has led to studies involving side reinforcement of the members in the joint area using glued plywood or densified veneer wood (Leijten et al., 1994) or fibre glass. These reinforcements prevent overloading of the end corner of the members and ensure plastic behaviour at failure. Another way, used in Japan and Australia, is to design single in-plane members (rafter and column) and to install internal steel plates with dowel fasteners or external plywood gussets (Figure 3). The first arrangement is a good solution for appearance and fire resistance. In both cases, the equilibrium of forces and moment is achieved in the centre section of the plate and the designer has to pay attention to the internal strength of the plate. STEPIEUROFORTECH - an initiative under the EU Comett Programme

Figure 3

Frames with mechanically jointed in-plane members: (a) joint with internal steel plate, (b) nailed plywood gusset

V-shaped column Another concept is to change the loading path using a V-shaped column fixed to the rafter (Figure 4). Depending on the stability criteria and horizontal action effects, this type of frame could be two-hinged with a continuous pitched and curved beam with spans up to 30 rn. For greater spans, they should be three-hinged. The depths of the beam vary between W30 and W40 for h,, and W40 to Ll60 for h,. To transfer the bending moment, the tension and compression members should make an angle a within the range 10 to 20".

Figure 4

V-shaped frame geometry.

With the internal timber member in compression, either a vertical internal timber member or an external steel rod in tension may be used. For such a frame, the design must investigate all combined action effects to take into account possible reversal of loading in the column members.

Glued jointing techniques This technique is carried out using large finger joints or glued-in bolts (Figure 5). Mainly developed in Scandinavian countries, these glued joints lead to high local force transfer and very stiff connections. The main restriction in use results from the brittle behaviour of these joints. So, their production and installation require extended quality control and must fulfil specific requirements as specified for large finger joints in prEN387, "Glued laminated timber - Production requirements for large finger joints". Members connected by large finger joints is the more common glued joint for frames. The joint profile is cut along the depth of the member. In general the knee joint is produced by installing a comer block and cutting two large finger joints. This enables the angle between the forces and the grain direction to be reduced and hence the joint strength is increased. Furthermore, special attention should also be paid to reversal of loading as the strength of large finger joints is greater for negative moment (M,,) STEPIEUROFORTECH - an initiative under the EU Comett Programme

D813

than for positive values (M,,). The absolute value of the ratio M,,/ M,, varies from 0,l to 0,2 for a roof slope of 0 to 30" (Reyer et al., 1991, Heimeshoff, 1976).

Figure 5

Frame comers using large fingerjoint (a), or glued-in bolts (b).

As on-site installation is not allowed for large finger joints, the use of such nondemountable joints is also limited by transportation conditions. To overcome this disadvantage and to limit the risk associated with a single component brittle joint, another solution is to used glued-in bolts (see STEP Lecture C14). Figure 6 shows an installation with inclined glued-in bolts to reduce the influence of possible splitting on the joint strength (Turkowskij, 1991). In this example, forces and moment are transferred by separate compression and tension components and low load reversal is permitted.

Figure 6

Frame corner with inclined glued-in bolts.

Design of frame corners Especially for dowelled moment-resisting joints, EC5 rules and linked standards allow new possibilities for frame optirnisation. Aimed at increasing the use of timber, the first stage is to design the most efficient joints (mechanically and economically).This section presents a design example showing the possible choice for designers and the timber industries.

A three-hinged frame (Figure 7) is designed with glued-laminated tapered members. Related to the short-term load duration class, the critical load combination gives the forces at the knee-joint in the column and the rafter: vC,,= 80 lo3 N N~,,= 120 . lo3 N rafter: v,, = 90 -lo3N N ~ , =, 112 -10' N and

M,, = 260-10' Nm


Figure 7

Geometry of the frame (a) and layout of the frame corner (b).

Current design (frame A) Based on a preliminary design, the geometry and the properties of the selected materials are: Glued-laminated members: strength class GL24 (prEN 1194, annex A) p, = 364 kg/m3 f,, = 2,52 N/mmz t, = 65 mm t, =115mm Dowels: steel grade Fe240

f

= 240 N/mmz

d

=24mm.

k, h

= 0,9 =1150mm

A circular pattern is chosen and the radii are:

With a minimum spacing of 6 d between dowels, the number of dowels per circle is:

The design joint forces induced by the forces and the bending moment are defined in Table 1 (see STEP Lecture C16). Rafter

Column

Table 1

Design values of the forces on the most critical fasteners and the timber.

The mechanical properties of the fasteners have the following values: embedding strength:

fh,oC =

90,082 (1 - 0,01.24) 364 = 193

15,7 ~ / m m ~

yield moment: The coefficient k,, is equal to 1,71. The calculation of the design load-carrying capacity of the critical dowels (see STEP lectures C6 and C16) is given in Table 2. STEPIEUROFORTECH - an initiative under the EU Comett Programme

D8/5

Column

Table 2

Rafter

Design resistance per shear plane for the critical dowels.

For the chosen patterns, the load-carrying condition is checked on the column:

Rip,, = 2 . 119 .lo3 = 23,7 . lo3 N

> Fd,, = 23,l

. lo3 N

and on the rafter: RjPd= 2 .12,8 .lo3 = 25,6

> Fd,c = 23,3

. lo3 N

. lo3 N

In the joint area, the strength of the timber is verified for the calculated force F , , :

At the serviceability limit states, the rotational rigidity of the joint is: Kser,r,d

= Kser (n, r12 +

nz r:)= 9,75 . 10" Nmmlrd

Design with improved steel grade (frame B ) Choosing dowels of smaller diameter and greater tensile strength, EC5 allows the plastic behaviour of the moment-resisting joint to be improved. In this example, the design option is to select dowels of steel grade Fe430 (f,,, = 430 N/mmz) with a diameter of 24 mm. An optimum design is obtained for h = 1080 mm and 45 (30 + 15) dowels installed on two circular patterns with the radii r, = 470 mm and r, = 390 mm. Table 3 gives the results of the design calculations.

Table 3

Results of the design calculations.

For the calculation of the deflection in service, the design has to take into account the joint rigidity:

K,,

= Kse, (n, r:

+

n, r:)= 9,90

. 10"

Nmmlrd


Design with improved glued laminated timber and steel grade (frame C) The other way to optimise the timber structures is to select glued laminated timber of a greater strength class. It should be mentioned that this choice implies greater control requirements during the fabrication process. For this design, the selected components are: Glued-laminated members: strength class GL30 (prEN1194, annex A) pk = 407 kg/m3 , = 2,18 N/mmz Dowels: steel grade Fe430

f , = 430 N/mm2

The possible designs are: C1 h = 1080 mm and 41 (30 + 11) dowels d = 16 mm, C2 h = 980 mm and 70 (38 + 32) dowels d = 12 mm. At the serviceability limit states, the rotational rigidity of the joint is: C1 Kser,r,d= 109

. lo9 Nmmlrd

C2 K,,,r,, = 118

. lo9 Nmmlrd

Design summary Glued laminated timber

dowels

Design

class

h ( mm)

steel grade

A

GL24

1150

Fe240

34 d

24 mm

B

GL24

1080

Fe430

45d

16mm

C1

GL30

1080

Fe430

41 d

16 mm

C2

GL30

980

Fe430

70d

12mm

Table 4

number and diameter

Design possibilities of the moment-resisting joint.

In comparison with design A, the best cost-efficiency could be reached with the design B or C2 depending on the cost of each operation during the fabrication and erection processes.

References Heimeshoff, B. (1976).Berechnung vonRahmenecken rnit Keilzinkenverbindungen. Arbeitsgemeinschaft Holz, Holzbau Statik Aktuell, Germany, Folge 1. Leijten, A.J.M., Cruz H.M., Rodd P.D., Werner H. and Virdi K.S. (1994). Physical and mechanical properties of densified veneer wood (DVW) for structural applications. Final Report of FOREST project N0MA2B-CT91-0033,Delft, Netherlands. Reyer, E., Schmidt, M. and Schmidt, P. (1991). Development and calculation of kinked timber joints elastically prestressed perpendicular to the grain. Proceedings of the International Timber Engineering Conference, Vol. 3, p. 3.85-3.96, London, UK. Turkowskij, S. (1991). Prefabricated joints of timber structures on inclined glued-in bars. Proceedings of the International Timber Engineering Conference, Vol. 3, p. 3.143-3.148, London, UK.


Bracing - Structural detailing STEP lecture D9 Objectives To provide advice on the use and behaviour of bracing. The lecture includes H. Briininghoff Gesamthochschule Wuppertal examples of bracing as well as the necessary detailing and joints.

Summary The behaviour of bracing is discussed and simplified design rules are provided to assist design engineers. Some advice on structural modelling and detailing of the joints is provided.

Introduction It is essential that principal structural members which are liable to buckle are connected to walls, columns, beams or bracing structures which are able to resist the forces involved, in order to ensure local and overall structural stability. The bracing elements prevent large lateral displacements which will otherwise occur perpendicular to the principal axislplane of the structural element. At the same time, they can be used as structural elements for resisting external forces such as wind loading. If main structural elements are perfectly straight and the external loads are applied only in their principal planes, i.e. without any eccentricity, then deflections will occur only in the principal plane without inducing any reactions onto the bracing. However, in practice, it is necessary to allow for the lack of straightness due to imperfections in the production process and which can occur during the erection process. Bucking forces will also occur if wind or other external loads occur in the plane of the bracing, resulting in lateral deflections.

Uses of bracing Structural bracing can be used to resist external forces which do not arise from the behaviour of the structure but are applied onto the structure and have to be transmitted to the foundations. Examples are wind loads or horizontal loads e.g. lateral shocks, crane braking forces and seismic loads. The second type of forces are internal forces which result from deviations of the main structural element from its intended position. These forces can be balanced within the structure which is correctly detailed and therefore do not have to be transmitted to the foundations. Examples are

-

forces due to the lateral displacements of frames and columns,

-

forces arising from supporting beams and the compression chords of trusses which are liable to buckling,

-

forces at the intermediate supports of compression members,

-

forces at the lateral supports of tension chord nodes in trussed members.

Examples are given in Figures 1 to 7.


A

t

Figure I

Transfer of wind loads from gable columns through roof and wall bracing.

Figure 2

Transfer of external loads through bracing into the foundation. ( a ) braking force from crane, ( b ) bracing to provide torsional resistance at beam ends.

Figure 3

Transfer of P-A forces from inclined columns through roof and wall bracing.

Figure 4

Support reactions for inclined structures and vertical actions.


Figure 5

Lateral intermediate support to decrease column lengths.

Figure 6

Lateral bracing of beams.

Figure 7

Lateral support of tension chords using knee-bracing against purlins.

1 In a structure, the main load bearing elements have to transmit the vertical forces e.g. self-weight and snow loads on roofs whilst the bracing elements have to resist horizontal wind loads and buckling forces from the main element. The designer usually considers these actions separately and the design of the main elements and bracing is carried out in two steps. In reality, however, structures are threedimensional systems such as the simple truss, shown in Figure 8, with upper and lower purlins. In this case two diagonal bracing elements, which together with the purlins acting as chords form a truss system in the plane of the roof, to ensure lateral stability of the simple truss. There are eight support reactions. One of these STEPIEUROFORTECH - an initiative under the EU Cornett Programme

D9/3

is needed to prevent the four structural elements at eaves level from forming a rectangular mechanism, thus leaving seven support reactions, with six possible equilibrium conditions in space. Thus the system is statically indeterminate to one degree.

Figure 8

Support reactions of three-dimensional braced system.

For design purposes, the design may be simplified by neglecting one of the support reactions in the x-direction (see Figure 8) and assigning the horizontal bracing forces to the two supports in the y-direction. This simplification is only possible if the displacements of the supports (assumed as fixed in the y-direction) are nearly the same; otherwise the "omitted" support will have to resist forces in the xdirection resulting from the rotation of the structure in plan. This approach is also only valid if the deformations of the chords (purlins) are small enough to be negligible but would only cause displacements in the y-direction. Due to the extensions in the internal members of the truss and movements in the joints, shear deformation as shown in Figure 9 will occur. Hence the vertical element at the support will remain vertical without causing any horizontal reaction at the top of the truss.

Figure 9

Deflection of a bracing structure.

In sloping roofs, the span of the bracing system is assumed to be equal to the length of the roof plane and is a plane structure. The three-dimensional load carrying behaviour leads to inclined forces (see Figure 10) which should be taken into STEPIEUROFORTECH

- an initiative


account. This is a practical calculation model. For roofs with large slopes, it is sensible not to consider the transfer of shear forces along the apex of the roof resulting in two cantilever beams, as shown in Figure 11.

Figure 10

Vertical forces at the top of the structure caused by horizontal actions.

Figure 1 I

Support reactions of cantilevered system.

If the structural system and the loading are symmetrical and the deformations of the chords are not considered, the horizontal displacement of the apex is the same for both bracings. Thus there would be no forces along the apex but there would be two horizontal support reactions in the x-direction which have to be considered. It is also advisable to support the bracing system at the apex and/or at intermediate locations. These support forces should be transferred to structural strong points, say at the eaves - see Figure 12. In such cases, where the stiffness of the support system is sufficiently high, small deformations of the bracing system would occur, resulting in smaller lateral forces. In general this is a sensible approach, however, with nailed plate trusses which are usually narrowly spaced, it is impossible to provide such a bracing systems with sufficient depth and hence stiffness.

Figure 12

Support of bracing truss at apex.


Bracing details Types of bracing Bracing systems are mostly formed by adding diagonal members to the main structural elements (e.g. beams and purlins) to form trusses. Prefabricated trusses which are placed in between the members to be braced can also be used. In this lecture only trusses are considered although beams, shear walls or single members can also be used for bracing purposes. In most cases, the main structural elements also form the chords of the bracing system. In the case of trusses, the compression cord which needs to be braced should also be part of the bracing system. Where beams have to be braced, the bracing system should be placed in the compression zone. For roof structures with purlins, the purlins can be used as part of the bracing system. This is achieved by adding diagonal members. The characteristics of the different possible forms are described below.

Crossed diagonals resisting only tensile forces usually made of steel e.g. steel rods with turn buckles for tightening purposes; -

ease of erection;

-

the purlins will be subjected to the additional stresses arising from being part of the bracing system.

Figure 13

Crossed diagonals. ( a ) Braced beam, ( b ) purlin.

W- trusses -

no additional forces in the purlins;

-

all diagonals have to transmit both tensile and compressive forces due to reversal of forces, thus timber members are recommended;

-

where the diagonal members are fixed to the purlins, the buckling lengths of the diagonals can be reduced.

Figure 14

W-truss. ( a ) Braced beam.

STEPIEUROFORTECH


N-trusses this system is only sensible if the external loading in one direction produces relatively higher forces;

-

the vertical internal members, if in compression, have shorter buckling lengths. (h)

Figure 15

(11)

N-truss. ( a ) Braced beam, ( b ) purlin.

K-trusses the forces in the internal members are reduced by up to 50%; -

buckling lengths are relatively shorter;

-

offers larger openings when used as a vertical wall bracing system, see Figure 17;

-

the diagonals support the purlins at midpoint, thus reducing the buckling length in the plane of bracing.

Figure 16

K-truss. ( a ) Braced beam, ( b ) purlin.

Figure 17

Wall bracing.

Trussed beam ease of erection e.g. screwing or nailing steel members onto the purlins; the use of a parabola shape for the tension chord is optimal as it results in a constant tensile forces; STEPIEUROFORTECH - an initiative under the EU Comett Programme

D9/7

-

only applicable where the loading is in one direction;

-

non-uniform external loading must be considered.

Figure 18

Trussed beam. (a) Braced beam, (b) purlin.

Connections Connections can be detailed in a number of ways and the examples shown in F w e s 19 to 21 have worked well in practice. Timber diagonals can be connected with steel plates (Figure 19) and nails or dowels. Slotted steel plates should be predrilled together with the timber since the required spacings are smaller and the connections are stiffer and more effective when compared with non-predrilled nailed connections. For small forces, thin steel plates on one side are sufficient. The connection area required for non-predrilled nailed connections is four times larger than that required for predrilled connections with slotted steel plates. The diagonals have also to be designed for the bending moments resulting from the eccentricity of the steel plates.

Figure 19

Timber diagonals connected by slotted steel plate. (a) Braced beam, ( b ) slotted steel plate, (c) timber diagonals, (d) predrilled nailed connection.

The diagonals can be connected through specially designed steel connections such as those shown in Figure 20. This kind of joint can be assembled easily.

Figure 20

Connection of steel diagonals with wedge formations and flat bars. ( a ) Steel angle, (b) braced beam, (c) connectors, (d)purlin, (e) turnbuckle.


Steel rods which penetrate the chords can be connected at the other side as shown in Figure 21. Special steel connections which allow for a wide range of diagonal angles are available.

Figure 21

Connection of steel diagonals on the back side of the chords. ( a ) Braced beam, (6) nailed steel fitting, ( c ) steel bar, ( d ) purlin, ( e ) joint using timber block and connectors.

Example In STEP lecture B15, the forces on the bracing members due to the buckling of the bending members shown in Figure 22 were determined as qd,hr= 5,04 kN/m

Figure 22

Bracing structure.

Assuming a uniformly distributed wind load of w, = 3,96 kN/m

and two bracing trusses, the following forces are obtained.


The existing purlins need to be designed for the forces V, and V,. The additional timber diagonals are designed for forces Dl and D,.The connection details are as shown in Figure 19 and the following points should be considered:

-

the lowest purlin should be designed for the bending moments due to the vertical actions and a compressive force Vl = 45 kN. This compressive force is to be transferred to the main truss;

-

the purlin subjected to a compressive force V2 = 22,5 kN is to be designed as a column;

-

the purlin V, is to be designed for a compressive force of 22,5 kN to allow for load reversal;

-

the top rafters of the main truss which form the chords of the bracing truss should be designed for the force U, = 37,5 kN which is compressive when the loading acts in the opposite direction;

-

the vertical component of the force Fvfor a roof slope of a is Fv= 2U2 sina

-

the diagonal members should be designed for a compressive force D,= 43,9 kN;

-

it is recommended that all members are connected with the same number of nails, rather than designing the joint in accordance with the shear force diagram for the truss. This is because designing the joints for the actual shear forces would result in identical joint displacements when the bracing is loaded which, however, is not realistic. By using the same number of nails per joint, and thus having the same joint stiffness for the entire truss, realistic differential jofnt slips are achieved;

-

the requirement to limit the deflections to 11700 for lateral buckling loads or 11500 for the combination of wind loads and buckling loads is not part of the serviceability limit state design. These limits are given because they were used in the assumptions for the development of the EC5 equations. These limits can be exceeded if a more exact calculation for the buckling loads is carried out in which case, the actions should be increased by y, and 5%values for material properties modified by ,y and k,, should be used. The slip modulus K, should be taken as 213 K,,,;

-

it is not necessary to limit the deflections when the span to depth (l/h) ratio of the bracing is less than 6. However, accurate design and good detailing is particularly important. The deflections should be calculated in accordance with appendix C of EC5. The elastic deformations of the internal members as well as the deformations of the joints have to be taken into account.


Bracings are in general three dimensional structures although for design they are often considered as plane trusses.

References Briininghoff, H. (1988). Verbande und Abstiitzungen, Grundlagen und Regelnachweise In: Informationsdienst Holz, Bericht der Entwicklungsgemeinschaft Holzbau in der Deutschen Gesellschaft fiir Holzforschung.


Detailing of timber structures in seismic areas STEP lecture D10

Objective

A. Ceccotti University of Florence P. Touliatos University of Athens

To give guidelines for the design and evaluation of structural details for timber construction in seismic zones.

Summary Structural detailing is a very important issue for earthquake resistant buildings. This is particularly true in the case of timber structures where the conception of the structural behaviour as a whole and of single joints play fundamental roles. In this lecture particular attention is therefore paid to the structural form and to the ductility and dissipation properties of joints that can be reached only if details are properly conceived. The intention is not to present an exhaustive list of possible cases but to give the key for a better understanding of real needs. Examples are taken also from ancient constructions in order to enable modern designers to learn from past experience about earthquake-safe constructions, when calculation codes were not yet available.

Introduction In the design process of construction in seismic zones great attention is always paid to the calculation of the load-bearing elements: codes dedicate a large space to it. Nevertheless designers must be aware that the diligent application of the calculation rules will not be enough for the success of the construction. The reality is that constructional arrangements and details are also important. In other words calculation per se is not sufficient without good detailing. Eventually the opposite may be true; i.e. for constructions of small dimensions, regularly arranged in plan and in height, with some minimum dimensioning of element sections, few guidelines about bracing and connections had been in general sufficient, as shown by experience, to resist earthquakes. Therefore in many seismic codes around the world a list of minimum dimensions and requirements with structural examples are given for small structures which do not need to be calculated (Ceccotti, 1989). In the present version of the European Seismic Design Code (Eurocode 8) a similar possibility has been not considered because of the diversity of construction techniques through different countries of the European Union. Anyway in Eurocode 8 in addition to calculation criteria a lot of detailing rules are given in order to:

-

assure compliance with some very important hypotheses which form the basis of the calculation methods and to give at the same time advice against the most dangerous mistakes;

-

assure the attainment of the required ductility level and, consequently, of the relevant "behaviour factor" used for the evaluation of the inertia forces. For example in part 1.3, chapter 5 - "Timber Structures" - of Eurocode 8 some detailing rules are given in order to assure good ductility behaviour of the mechanical joints.

Actually, in order to avoid limiting the building activity with timber, Eurocode 8 is a performance-based code and in principle any joint may be acceptable if it fulfils some ductility performance test requirements (see STEP lecture C17), but in most current cases a few detailing rules are sufficient to avoid tests. STEP/EUROFORTECH - an initiative under the EU Comett Programme

D10/1

Nevertheless the designer must also be aware that even if the part of the code dedicated to detailing was to be full of structural detailing examples a satisfactory outcome is not assured by the uncritical application of these rules. The most imgortant thing is to understand the real meaning of such detailing rules and behave accordingly. A code could never contain and solve respectively all kinds of details and problems. Consequently, in this lecture the principal idea is to focus on the most important points to be controlled for the best performance of a generic timber structure in a seismic zone and leave the solution of particular cases to the designer.

Structural continuity Basically, the earthquake action may be considered to be a horizontal action that, in contrast to vertical actions, involves the entire structure and not only the small part of it just underneath the load (Figure 1). Obviously the same thing can be said for wind, but if hurricanes are excluded, fortunately not present in Europe, the action of the earthquake, according to the modern point of view of the Eurocodes, can be more important than wind especially for heavy structures (Ceccotti & Larsen, 1988).

Figure I

Different structural effects of vertical and horizontal loading.

This means that the continuity of the link between different members at all positions is particularly important and the effectiveness in both tension and compression. All the components of the shear walls and diaphragm systems (see STEP lecture E10) must be adequately fastened together so that the structure acts as an effective unit. In Figure 2, for example, the main positions where such features are necessary are focused, and a possible solution for realising continuity is presented. At the floor level the presence of a continuous girder all around the floor should be considered in order to collect the tension forces that will arise when the floor is loaded laterally and it is considered rigid in plan acting as a diaphragm (Figure 2b); and the necessary continuity at the corners will be assured through the diaphragm panelling putting edge nails at closer spacing (Figure 2a). Also in the height the load bearing vertical elements should be continuously connected in order to guarantee the transmission of the vertical efforts (Figure 2c). STEPIEUROFORTECH - an initiative under the EU Comett Programme

Figure 2

Details assuring structural continuity under horizontal actions. ( a ) corner reinforcement; (b) tension girder continuity; (c) continuity of tension studs; (d) prevention of uplifting from foundation and sliding of foundation.

In some cases, at floor level, the liaison between two corresponding - upper and lower - shear walls is merely obtained by panel sheeting nailing through the header beam of the floor. That is not suitable when tension efforts are important as in the case of an earthquake (e.g. a good solution is shown in Figure 2c and Figure 3b). Particular attention should be paid to the connections between the timber structure and the foundations both in order to prevent uplift and sliding (Figure 2d). Openings weaken both diaphragms and shear walls, therefore openings must be reinforced around them in order to maintain as much as possible the same in-plane rigidity. In Figure 3 an example of a real application is given. Great attention shall be paid to tension perpendicular to grain. For that reason Eurocode 8 prescribes, with reference to Figure 4d, that be has to be more than 213 h, where h is the depth of the member, so that splitting due to tension perpendicular to the grain is less likely; and when using a strap this should surround the timber piece (Figure 4e). Connections must obviously be able to work in both directions because the action of the earthquake is bi-directional. For that reason simply contact joints without any STEPIEUROFORTECH - an initiative under the EU Comett Programme

Dl013

possibility of reacting against opposite actions are not suitable (in Figure 5 some possible provisions are given). Foundations shall be tied to each other as much is possible in order to minimise the effects of differential ground movements. In particular the foundations of houses should be interconnected to act as a whole especially if the nature of the ground is soft in order to realise a rigid foundation functioning as a "raft" when the soil is moving.

Figure 3

Detailing examl~lesfor a timber framed house. ( a ) prevention of uplift from foundations; ( b ) continuity of tension members; ( c ) stzffening of openings in shear walls by framing with additional studs, lintels and corner hangers; ( d ) stiffening of openings in diaphragms by framing with doubling of trimmer and header joists; ( e ) stiffening of diaphragm floors (blocking); (f) prevention of sliding of foundations.

Figure 4

Detailing against tension perpendicular to grain. (a),(c):poor; (b),(d)and (e): good.


Figure 5

Possible provisions against the loosening of support in old and modern con~tructions.

Figure 6

Schematic exarizples of distrib~rtioaof bracing and stiffening. (a) and (e): poor; ( b ) and (f): fair; ( c ) and (dl: good.

Building regularity Regularity in plan and height is very important in order to assure good behaviour under earthquakes. The reason is that torsional effects induced by irregularity are not easily determined by the calculations, especially the static one. On the other hand, unfortunately, the attempt to control torsional effect by sophisticated dynamic calculations is often no more than an academic exercise: it would be much better to realise the regular building, also without real axes of symmetry but at least with seismic resistant parts regularly distributed and, better, also homogenously distributed (Figure 6, rearranged from Dowrick, 1977). In that way torsional effects are quite limited, strength properties are more uniformly distributed and calculation results more reliable. In the presence of large unbalanced openings in order to reduce the tendency of the building to twist under lateral forces, the best solution is to try to approximate the rigidity afforded by the shear wall at the opposite end by means of additional bracing, or by increasing panel thickness, with edge nailing STEP/EUROFORTECH - an initiative under the EU Comett Programme

Dl015

at closer spacing, or by affixing panels to both sides of the framing around the openings. Internal partitions have usually a positive effect as they contribute an increase in the dissipation of energy by hysteresis and friction.

Structural compatibility The problem arises in the case of connection between parts with different rigidities, e.g. for the liaison between the load bearing timber structure and a chimney or an external wall - often only decorative - of masonry (or even glass).

Figure 7

Timber framed sport-hall with special joint allowing independent movements between structural and non-structural parts. In the enlarged detail (p) are visible: (a) frame stiffening beam; ( b ) heavy duty springs; ( c ) bracing system joint; (d) main frame steel rod with end spring; ( e ) building facade steel column; ( f ) independent movements between main frame and facade frame; (g) main frame glulam beam; ( h ) facade wooden frame and windows.

The designer has two possibilities. One is to realise an external wall rigid, selfsupporting and independent of the flexible timber structure, for example in Figure 7 the glass facade wall is independent of the main structure (Touliatos, 1991). The other possibility is to connect the external wall part, i.e. masonry, to the timber part so strictly that the two structures will act together as a composite structure. Actually in most cases the connections between external masonry walls and the internal timber structure are not accordingly conceived so that they just increase the horizontal action on the masonry wall because they add to the inertia forces acting on the wall per se, the pushing-action due to the larger movement of the timber structure. If different material parts are connected, two simplified alternative design approaches are possible according to two different limit situations. On one hand it STEPIEUROFORTECH - an initiative under the EU Comett Programme

can be considered that the masonry weight should be carried by the timber structure (i.e. light masonry with mass but no rigidity, that means the masonry is considered fissured); on the other hand it may be assumed that the masonry, more rigid than timber, will collect the totality of lateral actions (i.e. heavy masonry). In other words the timber structure will still carry the vertical loads but will lean on the masonry structure with regard to the horizontal forces (Figure 8).

Figure 8

Example of mixed stone-timber building (Greek Islands, 1500 B.C.). Timber framing supports vertical loads but only heavy masonry can resist lateral loads. Note that floors because of their workmanship cannot be considered rigid in plane.

When a strong earthquake causes the masonry to collapse, dissipating a lot of energy, if detailing is well conceived, the timber structure has still the possibility to stand up after the shock and the masonry will be easily repaired, as shown in the case of Figure 9.

Ductility and dissipation of energy In old buildings (e.g. timber framed buildings with brick infill, like in Figure 9), dissipation of energy was obtained by friction between timber and masonry, and by hysteresis due to compression perpendicular to grain (Touliatos, 1993). In modem structures, in order to reduce inertia forces, unless a decision is taken to increase the natural period of vibration as in Figure 7d, the most usual way is to dissipate energy by hysteresis in mechanical joints (see STEP lecture C17). Eurocode 8 prescribes that when designing with reduced inertia forces (q > 1) ductility and energy dissipation properties shall be demonstrated by tests. For some particular cases tests may be avoided if certain details are followed. These details are mainly based on past experience. STEP/EUROFORTECH - an initiative under the EU Comett Programme

D 1017

Figure 9

Example of mixed masonry-timber structure building in Greek Islands (1800 A.D.). ( a ) diagonal stiffening timber rod acting perpendicular to grain at the corner; ( b ) anchoring detail of the wooden frame to the masonry wall; ( c ) masonry wall bearing the wooden frame of the upper floors; ( d ) secondary load bearing system of wooden columns, just behind the main load bearing system of masonry walls, able to support provisionally the entire structure after a severe earthquake (see the lower detail, right hand side); ( e ) wooden curved one piece component stiffening the timber wall frame; ( f ) special joint easy to replace when damaged after an earthquake (see the enlarged detail, below on the left hand side); ( g ) wooden curved one piece component stiffening the roof; ( h ) tie beam to avoid thrust on walls; ( i ) brick infill. Also in this case floors cannot be considered rigid in plane due to their workmanship.

STEPIEUROFORTECH


Nailed shear panel systems have given excellent ductile behaviour, much better than any diagonal bracing system. For that reason in the connection of sheeting to the timber framing, in Eurocode 8 it is stated that for proper ductility it is sufficient that the sheeting material is wood-based and the thickness t, of the sheeting material is at least 4 d, where d, the nail diameter, does not exceed 3,l mm. This is valid provided that the wood-based materials meet one of the following conditions: particleboard-panels with a density of at least 650 kg/m3, plywood-sheeting at least 9 mm thick, or particleboard and fibreboard sheeting at least 13 mm thick. Smooth nails and staples are still suitable in diaphragms when connecting sheeting materials to the timber frame. A point side penetration t, of 4-6 times the sheeting thickness is appropriate. However in the general case smooth nails are not recommended without additional provisions against withdrawal, e.g. by coating or clenching.

Figure 10

Basic detailing requirements in Eurocode 8.

Besides Eurocode 8 considers that dowelled and nailed timber-to-timber and steelto-timber joints, when the minimum thickness of the connected members is at least 8 d and the dowel diameter does not exceed 12 mm, are sufficiently ductile. The reason why this applies is that for the best performance under cyclic load a mode 111 failure of the joint is desirable (that means thick timbers and slender dowels). Now, referring to the diagrams in STEP lecture C3 (see Figures 11 and 12) it is easy to recognise that the fields corresponding to the desirable behaviour are characterised by values of t, / , / v d and t, / d-d ratios bigger than about 3,5. Therefore with reference to the usual values of timber embedding strength and fastener steel yielding strength it is possible to state that the minimum thickness of the connected timbers -8 d- is very much on the safe side, in ductility terms. If a designer wants to propose different fastener arrangements or different materials, it is allowed, provided it can be demonstrated by tests that the EC8 performance requirements under cyclic loading of the joint are fulfilled (see STEP lecture C17). Obviously before proposing new arrangements it is important to have clearly in STEPIEUROFORTECH - an initiative under the EU Comett Programme

D 1019

mind that the basic idea is to try to obtain a mode I11 failure in order to couple the dissipation due to the embedding of the timber with the dissipation due to the plasticity of the fastener's steel. Note: it is evident that some detailing rules given by Eurocode 8 are thought to achieve the necessary ductility level. But what about the case when the designer chooses to design the structure without making reference to ductile and dissipative behaviour, i.e. with q = l ? In principle it is not essential, for example, to use slender dowels, but the use of slender dowels will for certain give to the structure a reserve of ductility that is always very welcome, without any extra cost. So the suggestion of the authors is to follow as much as possible the detailing for ductile and dissipative behaviour even for structures calculated as non dissipative.

Concluding summary Structural detailing is very important in timber structures in seismic areas in order to guarantee the flow of the efforts through the entire resistant structure from foundations and ground to roofing and vice versa (structural form and continuity); and to ensure a sufficient resistance reserve (ductility and dissipation of energy). Codes and experience can give guidelines for proper detailing but the designer has the challenge to find for each case the most appropriate solution.

References Ceccotti, A. & Larsen, H.J. (1988). Background Document for specific rules for timber structures in Eurocode 8. Report EUR 12226 for the Commission of the European Communities, Bruxelles, Belgium. Ceccotti, A,, editor, (1989). Structural Behaviour of Timber Constructions in Seismic Zones. Proc. of the relevant CEC DG I11 - Univ. of Florence Workshop, Florence, Italy. Derrick, D.J. (1977). Earthquake resistant design. John Wiley & Sons pub., London, U.K. Touliatos, P. (1991). Some problems of the timber framed construction in seismic zones. Proc. of 1991 International Timber Engineering Conference, London, U.K. Touliatos, P. (1993). Seismic disaster prevention in the history of structures in Greece. National Technical Univ. of Athens, Greece.


Computer aided design and manufacturing STEP lecture D 11 M.H. Kessel Labor für Holztechnik LHT Hildesheim

Objectives To provide information about the state of today's CADJCAM technology, i.e. computer-aided design and manufacturing systems, in the field of timber construction.

Summary The lecture Covers the specific requirements for CAD systems used in timber construction and on the direct control of wood working machines. CAD systems are no longer used as pure drawing devices, but are capable of simulating complete building designs and thus producing consistent building models and planning data. An example taken from the timber frame construction industry is used to illustrate the level of development.

Introduction In aircraft, automobile and machine construction the use of CADICAM systems is already a Part of everyday work to simulate a product and its construction on a computer. The aim is to ensure that the geometry, properties and behaviour when simulated deviate only marginally from the real product. Currently the microprocessors of workstations are so powerful that almost any degree of similarity can be achieved. The similarity is only limited by the amount of work necessary for the data input, which is of Course subject to cost-effectiveness and directly dependent on the number of pieces andlor components that are to be manufactured. This number is much higher in the series production of the manufacturing industries mentioned above than in timber construction, in which usually only unique products (Figure 1) are manufactured. However, there are also good reasons for timber engineers to portray and simulate the planned building in a full three-dimensional form on a computer:

-

By means of a complete 3D-simulation of all main components a consistent model of the building is produced. The consistency, for example of fitting-accuracy, and assembly can be controlled visually on the screen in the planning phase.

-

The complete 3D-simulation automatically generates complete parts and production lists.

-

By means of the complete 3D-simulation all the geometric and working data of each component can be obtained and the trimming installation or the wood working centre can be numerically controlled.

-

The complete 3D-simulation enables the generation of photo realistic representations by means of shading and rendering. Downstream image processing programs may be used to project buildings into existing surroundings. Use of these techniques can offer decisive advantages in acquisition and sale.

For the simulation of timber constructions powerful 3D-CAD-systems are already available (Kessel, Gnutzmann, 1992). Due to the three-dimensional STEPIEUROFORTECH - an initiative under the EU Comett Programme

D1 111

complexity of the timber constructions they have become more and more popular in practice. In the field of computer-aided manufacturing (CAM) ever more powerful computerised numerical control (CNC) woodworking machines (e.g. trimming machines) are being developed. They allow for flexible, rational and very high quality manufacturing.

Figure 1

Simulation of a gymnasium.

When combining CAD and CAM, geometric and material information about each component is transferred in the form of Parameters from the CAD simulation to the woodworking machine. Each working process on a component consists of several working steps of the CNC machine. The number of steps depends on the complexity of the working process.

Timber and CAD By the use of CAD it is possible to design a consistent building model with any desired dimensional accuracy of its components. This accuracy is only limited by the accuracy of the numeric representation in the Computer. The dimensions of the building model are usually transferred to the manufacturing System by means of shop drawings or records that are used for tracing dimensions of individual components or for direct machine control. When the components manufactured in this way are installed on the project site the building geometry planned should be precisely realised because of the consistency of the CAD model. But the STEPIEUROFORTECH - an initiative under the EU Comett Programme

accuracy realised depends on the accuracy of the production and the dimensional accuracy of the construction material used. The latter is especially important in timber construction. EC5: Part 1-1: 3.2.3

The target sizes needed for the calculation of the effective cross-section and the geometric properties of the supporting structure and thus the basis for the CAD building model refers to timber with a moisture content of 20%, corresponding to service class 2. Therefore the dimensions for structural timber of softwood also refer to a moisture content of 20%. Thus it can be assumed that when used in service class 1 with an equilibrium moisture content of 12% the cross-section will be reduced by several millimetres due to shrinkage. Usually this is taken into account by referring all dimensions in the shop drawing to the upper edge of the component. In order to meet today's quality demands in building construction only structural timber should be used, the moisture content of which should be equal to the equilibrium moisture content of the corresponding service class. Only when the quality of planning, production and building material is consistent can the building geometry be satisfactorily realised.

3D-CAD-Simulation CAD may no longer be understood as a tool for electronic drawing alone. The data generated by CAD contain much more information than a conventionally produced drawing and, above all, are much more reliable. This is illustrated in the roof construction in Figure 2.

Figure 2

First floor walls und purlins of a house in wood frame construction.

The useful and defect-free development of the detail solution shown in Figure 3 requires an experienced carpenter with a very good three-dimensional imagination for both conventional and 2D working procedures. Even when the carpenter meets these demands the final verification regarding defects in the STEPIEUROFORTECH - an initiative under the EU Comett Programme

D1 113

design can only be carried out in the building process itself. This is the essential difference to the 3D-CAD-simulation of a construction where, by means of immediate visual control, defects can be detected very early in the design process. Due to 3D-visualisation and extensive edit-functions there is a lot of room for new developments and unconventional solutions. New solutions can be found on the screen and their practicality can be verified at very low cost.

Figure 3

Intersection of the main roof with tower (left) and side roof (right).

The three-dimensional building model is the basis for all subsequent working steps in the design and manufacturing processes. Lay-out plans (Figure 4), projections, sectional views and shop drawings (Figure 5) are automatically generated from the 3D-model. Dimensioning shop drawings is also done automatically for Standard components. Therefore the amount of work necessary for postprocessing is usually very small. However, the main point is that because of the consistency of the building model the design is very reliable since, for example, transcription errors are avoided.

114

Figure 4

Rafter und purlin layout.


Shop drawings. Top: rafter position 110, bottom: r-aper position 11 I .

Figure 5

Some time ago CAD was only used as a tool for processing graphic data. Nowadays a complete 3D-CAD-system is able to archive and process nongraphic data. Furthermore, the existing geometric data can be combined with subject-specific data and processed in various ways. An example of this is the fully automatic generation of parts and production lists (Figure 6). These lists are generated by the System without any additional data input on the basis of the current machining state of the components. Subsequently they can be used as an important tool for inventory control, manufacturing, building and calculation. ARCHITECT CUSTOMER COMMENT

:

......,.,,....

: :

SCHMIDT rafters + purlins

No.

Constr. group

Constr. element

PURLINS PURLINS PURLINS PURLINS PURLINS PURLINS PURLINS PURLINS PURLINS PURLINS RAFTERS RAFTERS RAFTERS RAFTERS RAFTERS RAFTERS RAFTERS RAFTERS

CENTRE PURLIN RIDGE PURLIN CENTRE PURLIN CENTRE PURLIN RIDGE PURLIN CENTRE PURLIN EAVES PURLIN EAVES PURLIN EAVES PURLIN EAVES PURLIN JACK RAFTER JACK RAFTER COLLAR BEAM BAR BAR RAFTER-ZA-3 RAFTER-ZA- I RAFTER-3

RAFTERS RAFTERS RAFTERS RAFTERS RAFTERS RAFTERS

JACK RAFTER TRIMMER-2 TRIMMER-3 TRIMMER- 1 TRIMMER-2 JACK RAFTER-I

RAFTERS RAFTERS

RAFTER TRIMMER

No.TL

.-

---------180 181

0 0

TOTAL

Figure 6

PROJECT NAME PROJECT NUMBER DATE Pcs.

1 2

SCHMIDT :

........,,,..

18-MAY-94

Mat. b.

h,d

1

I (tot)m

V(mA3)

C24 80 C24 80

180 180

665 600

0,67 1,20

,010 ,017

452,57

7,515

82

Productiort List.


Computer-aided manufacturing The use of CNC-woodworking machines (e.g. trimming machines) has substantially increased over the last years. In Figure 7 a NC-record for a trimming machine is shown. Up to now these data are often generated directly on the machine by taking the data from a shop drawing and then transcribing them into a NC-program. On the one hand transcription errors can occur in this process and on the other hand the component geometry is generated twice; once in the CAD system and again on the machine.

Figure 7

Machine control ( 2 NC-records).

Company philosophy and interfaces Using Computers in planning, design and manufacturing offers a large number of advantages. However, benefit can only be derived from them if primarily isolated components such as CAD-simulation were to be integrated into company internal or inter-company philosophies. Then data for machine control could be generated in the CAD-system resulting in substantial time-saving and minimised error rate. The main problems with computer-aided concepts as a whole are the necessary interfaces. In the following the present state and possible developments of the three main interfaces are discussed below:

Man-computer interface Data input is still very time-consuming since only primitive devices such as keyboard, mouse or tray are available for the interface between man and computer. The improvement of communication between man and computer, i.e. the acceleration of the data transfer, becomes more and more important for further CADICAM developments since the computer internal data transfer has been accelerated in such a way that earlier processing bottlenecks, such as the time necessary for producing pictures and calculating the suppression of covered lines, have been eliminated. Important developments concerning this interface are not expected in the near future. By means of an even more User friendly design of the User surface of CAD-systems only minor improvements can be achieved. A substantial improvement could be achieved if expert knowledge were to be stored in the operating surface. This would have to include the restriction that this knowledge could only be applied on special geometries and designs. An important acceleration can be expected through developments in Sensor technology. In this STEPIEUROFORTECH - an initiative under the EU Comett Programme

field it might be possible some day to transfer human knowledge into the Computer by means of Sensors without using manual input devices.

Inte$ace to external programmes In order to achieve a higher degree of efficiency and reliability it is necessary that generated data can be further processed without having to repeat the corresponding data input. This requires the ability to exchange data between CAD-systems, calculation programmes and CNC-machines. The currently existing interfaces are only individual solutions for communication between specific programmes. Except for a few exceptions (e.g. STEP-2D and DXF) there are no interfaces that have become European standard and that meet all the demands of timber construction. Therefore, a general interface for timber construction is needed (see Deutsche Gesellschaft für Holzforschung 1994).

Interface CAD-machine CNC-woodworking machines still possess machine specific interfaces which enable the machine to memorise processing records (Figure 7). These highly differing records often describe the geometry of a component incompletely, because only data which can be processed by the machine can be memorised. Therefore a general, machine-independent interface is being developed which contains the complete geometric and specific description of the component. Due to the often very high complexity of timber components it is not expected to achieve a complete automation of the data generation for woodworking machines in the near future. The automation will at first be restricted to standard components. Therefore a CAD-system should allow the possibility of identifying components or complete construction forms by hand. This can be realised by the identification modules shown in Figure 8.

Figure 8

Identification modules.


By means of these modules it is possible graphically to choose the type of process to be used for a component, designed and simulated by a CAD system, from a menu. The choice can be made on three different levels. The menu contains components for which all analysis algorithms are stored parametrically. On the upper level the complete working process can be chosen for a comglete component if its form is similar to the form of a component in the menu. At the middle level the process for realising individual details can be chosen in the Same way. At the lowest level the machine unit which should work on the component can be selected together with the particular operation that is required. However, the lowest level should only be available directly on the machine since the User must possess machine-specific know-how and furthermore, no machinespecific data should be transferred via the interface CAD-machine.

Staff The use of computer-aided design and manufacturing systems requires staff that is motivated, trained in computer technology and possesses a good three-dimensional imagination. CADICAM does not lower the demands on the qualification of the staff but raises it. Since usually only young employees possess the qualifications required, they should be given the chance to gradually become familiar with the complex CAD/CAM system of the company until they finally master it. However, it must be taken into account that it is not enough to be able to master the technology alone, but also its influences on the design (fitting details) and manufacturing and building processes. So far only a little experience has been gained in this field. CADICAM is not a tool that can be used efficiently if the staff are trained only once on how to handle the system. These systems are still being further developed and in relatively short periods better and extended versions and possibilities will be available. Therefore continuous training Courses should be attended and the staff must be motivated, able to detect new possibilities themselves and to use them in the company-specific conditions.

Concluding remarks

-

The complete 3D-simulation of a timber construction on a computer provides a consistent model which allows for the visual control of consistency and assembly of all main components already in the planning phase.

-

In order to meet today's quality demands in building construction only structural timber should be used, the moisture content of which should be equal to the equilibrium moisture content of the corresponding service class.

-

Investments in CADICAM-technology are only useful, if the technology will be applied by well trained, motivated and creative staff.

-

Only when the quality of planning, manufacturing, building material and staff is harmonised i.e. consistent, can the building construction be satisfactorily realised.

-

After 10 years of experience in CADICAM-technology in timber construction it can be assumed that future developments will continue to be as fase as in the past, particularly in the fields of CAM, the connection of CAD and CAM and knowledge-based systems.


References Kessel, M.H. and Gnutzmann, D. (1992). Graphische Datenverarbeitung irn Holzbau. Karlsruhe: Bruderverlag. Deutsche Gesellschaft für Holzforschung DGM (1994). Projektschnittstelle Holzbau Version 1.04. München.


History of timber structures STEP lecture ~1 J.C. Chilton University of Nottingham

Objectives To describe the historical development of the use of timber in building and engineering structures and to develop an awareness of the significance of timber as a traditional structural material.

Summary The lecture traces the history and development of timber building and engineering structures, both small- and large-scale, from their humble beginnings in 'primitive' shelters through to modern long span structures. Some examples of timber structures from throughout the world are reviewed to illustrate this development and present the potential of timber as a structural material.

Introduction Timber has been available as a constructional material to most societies since the human race first started to build crude shelters at the beginning of civilisation. A diversity of tree species exist and most climatic zones have at least one that has adapted to the prevailing conditions within that area. Thus timber is generally available in most inhabited regions of the world. The history and development of timber structures is an extensive topic. Timber has been used in the construction of buildings, bridges, machinery, war engines, civil engineering works and boats etc. since mankind first learnt to fashion tools. Here it will only be possible to give some examples, generally limited to buildings, bridges and works of civil engineering, to illustrate this development. These examples will not be restricted to the European experience and, where possible, they will be drawn from other continents to demonstrate the adaptability of timber as a structural material, to stimulate study of alternative building forms and to show the engineering and architectural potential of timber structures.

"Primitive" structures

Figure I

'Primitive' structures ( a ) withy frame, (b) teepee, (c) yurt.

Probably the earliest shelters constructed by mankind were made from a framework of tree branches covered in leaves or animal hides; one of the simplest, the "withy frame" being a structure formed from saplings inserted in the ground in the form of a rough circle, bent over, woven and tied together with ropes of natural fibres to produce a lightweight dome. The "teepee" of the native North Americans is also an example of a simple structure where timber poles are used to construct a conical framework to provide a hide-covered shelter. Mongol nomadic desert people use a further example, the yurt, which consists of a folding timber lattice or trellis that is expanded and anchored to form a circular wall which then supports a set of thin timber radial roof arches. The skeletal framework is then clad with thick felt STEPIEUROFORTECH - an initiative under the EU Comett Programme

Elf1

material to provide shelter from sand-laden desert winds. These simple timber structures are illustrated in Figure 1.

Traditional Chinese building system A very sophisticated modular timber building system was developed over 1000 years ago in China and was well established by the time of the Sung Dynasty (960 - 1279) when, in 1103, the Ying-tsao Fa-shih was published. This was a detailed manual of building procedure and practice that, although revised by subsequent dynasties, survived in modified form until the founding of the Chinese Republic in 1912 (Needham, 1971). The basic structure of a traditional Chinese building consists of a grid of timber columns founded on large stone bases which then support floor beams and the heavy roof construction (see Figure 2). Under the modular building system, depending on the importance and physical scale of a building, the structural elements were selected from a standard set of components of prescribed dimensions. The roof structures of the traditional standard Chinese building system appear somewhat alien to the modern western observer as they do not employ any form of roof truss. Main roof beams span between columns and these in turn support a series of simply-supported beams, stacked one upon another, each slightly shorter than the one below. Purlins span between, and perpendicularly to, the ends of the beams in adjacent stacks and short individual rafters span from purlin to purlin to form the classic curved profile of the Chinese roof. Complex timber bracketing is used to support the ends of the lowest beams at each column, the size and complexity of the brackets also being described in detail for each building type in the standard system. Over many centuries the main development in this system was in the complexity and aesthetic appearance of the bracketing whilst the other modular components remained relatively unchanged.

Figure 2

( a ) Chinese building system and (b) bracket system.

Historic Japanese structures In Japan there is also a long tradition of using timber structures and many historic buildings were constructed following the example of the Chinese building system but in a slightly less elaborate form. The largest ancient timber building in the world is the Todaiji temple at Nara in Japan (Figure 3a), which is 57 m wide by 50 m deep and 47 m high and houses the Daibutsu or Great Statue of the Buddha. The current building, which dates from 1708, is only about two-thirds the size of the original which was destroyed by fire. Nearby, there is a pair of elegant timber multi-tiered pagodas, at the Yakushiji Temple, Nara. One of these, the East Tower is original but the other was rebuilt as a copy, in 1981, to replace that destroyed by fire in the 16th century. In the former capital of Kyoto there are many fine STEP/EUROFORTECH - an initiative under the EU Comett Programme

examples of timber temples and palaces of great antiquity. Several cities in Japan (e.g. Osaka, Kumamoto) had dramatic timber-framed castles with tiered tiled roofs, however, they were mostly destroyed or severely damaged by fire during times of war and are mainly now reinforced concrete frames cloaked with timber framed roofs and details. In Japan, the quality of traditional timber joinery is high and many complex and inventive joints were used to connect structural elements without using metal fastenings (Sumiyoshi and Matsui, 1991).

Figure 3

( a ) Todaiji Temple, Nara, Japan - the largest ancient timber building in the world and ( b ) detail of its extensive cantilever bracketing.

Norwegian stave churches

Figure 4

Left: structural system of Norwegian stave church. ( a ) Nave purlin, ( b ) nave collar beam, ( c ) nave scissor brace. ( d ) nave rafter, ( e ) nave roof bearing beam, (f) upper nave wall plate, (g) lower nave wall plate, ( h ) scissor brace quadrant bracket, ( i ) applied bracket, ( j ) nuve wall, ( k ) nave stave splay, (1) port hole, ( m ) nave bresummer, (n) nave arcading, ( 0 ) aisle rafter, ( p ) aisle strut, ( g ) quadrant bracket, ( r ) upper aisle wall plate, ( S ) lower aisle wall plate, ( t ) aisle wall, ( U ) aisle, ( v ) aisle sill beam, ( W ) raft beam, (X) nave intermediate stave, ( y ) nave corner stave. Right: Stave church at Hopperstad, Norway (Photograph by B.R. Lindstad, NIKU).


Up to the 14th century, between 800 and 1000 stave churches are estimated to have been built in Norway. The name derives from the substantial timber columns, or staves (from 0,3 to 0,4 m in diameter) which are a dominant structural element in their construction. That 29 of the churches still survive, some having done so for more than 800 years, is a tribute to the builders' skill in the selection of durable materials, a stable structural system and construction details appropriate for the climate. (Sack and Aune, 1989). An example of one of the surviving buildings is shown in Figure 4.

Medieval timber frames Timber-framed construction was commonly used in Europe during medieval times for houses, barns etc. In the UK, three methods of construction dominated; cruck frame, box frame and aisled frame (see Figure 5 to 7) with regional variations in the popularity of each form. For cruck frame construction, a series of transverse, laterally-stable, "A" frames were erected at regular intervals along the building on a continuous timber sole plate. In its simplest form, each cruck frame was produced by longitudinally splitting the trunk and main branch of a tree (or an appropriately curved trunk) of sufficient length to span from the sole plate to the roof ridge. The two halves of the trunk then formed a matching pair of slightly curved members (known as cruck "blades") that closely followed the cross-section of the building and could be jointed at the ridge (and connected with a collar beam if required) to make structurally-stable, symmetrical frame. Generally, all joints were made using hardwood dowels. Purlins and wall plates connected the frames longitudinally and lateral stability was achieved in this direction by knee braces inserted between the purlins and cmck frames to resist wind forces. Rafters were added to support tiles or thatch. Where longer clear spans were required, the construction was modified to produce a "base" cruck. In this form, the two cmck blades did not meet at the roof ridge but were connected by a braced cross beam to create a rudimentary timber portal frame. A "crown post" roof (as described in the section on roof trusses below) was then added to the centre section, supported by the cross beam.

Figure 5

Cruck frame. (a) Ridge purlin, ( 6 ) purlin, (c) wind brace, (d) tie beam, (e) wall plate, (f) cruck spur, ( g ) cruck blade, ( h ) wall post, (i) sole plate.


Figure 6

Box frame. (a) Tie beam, (b) wall plate, (c) wall post, (d) bressummer, (e) storey plate, (f) sole plate or sill, (g) summer, (h) girder.

Figure 7

Aisled frame. ( a ) Collar, (b) collar plate, (c) rafter, (d) crown post ( e ) tie beam, (f) arch brace, (g) roof plate, (h) shore, (i) aisle tie, ( j ) aisle post, ( k ) aisle plate, (l) wall post, (m) sole plate, (n) post plate.

Box framing was basically post and beam construction with the addition of diagonal bracing within the wall planes to resist lateral forces. A common feature of this construction was that the upper-storey floor cantilevered a short distance out from the walls of the ground floor. In densely populated medieval cities of Europe this provided additional accommodation over the narrow streets and the continuity of the floor beam cantilever improved the structural efficiency of the beams. In some box-framed buildings only limited diagonal bracing is provided with additional stability deriving from the infilling panels of wattle and daub, or brickwork etc. However, in others the diagonal bracing becomes an architectural feature, in addition to its structural function, with bracing patterns varying from region to region. As the name suggests, aisled construction, used predominantly for storage barns, has a building cross-section consisting of a main central "nave" and two side aisles, analogous to that of many churches. The central space was defined by two rows of timber columns, the aisle posts, which were usually connected longitudinally at their STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E115

heads with a continuous timber plate and knee bracing to form two long frames. Tie beams spanned transversely between adjacent columns in the two frames and these were also laterally stiffened with corner bracing. Wall posts were connected with additional framing members to form the side aisles. Usually, diagonal shores were installed, between the aisle posts and the wall sole plate or the post plate, to provide additional lateral stability against wind forces. Above the aisle frame a simple "collar" or "crown post" structure completed the roof. In some ways the crosssection of an aisled barn was a representation, in timber, of that of a typical Gothic cathedral.

Log-houses, balloon and platform frames In highly forested regions of Eastern Europe a different house building technique developed using the almost unlimited supply of logs (predominantly round) in which they are usually used laid horizontally one upon another to form walls. Structural stability is provided by notching the logs at the corner intersections so that the wall planes interlock. Evidence of the use of this notching technique dates from the Stone Age. However, the earliest evidence of notched log-houses has been found at Buchau, Bodensee, dating from the period 100-800 BC and the earliest in Scandinavian countries from around 1000 AD (Haulgid, 1980). Log-house construction was widespread in regions of Slavonic influence and was not restricted to small building. In Russia, for example there were many instances of tall church buildings constructed using the same technique. In Western regions a more sophisticated variation developed using trimmed logs and dove-tail notches at the corners. A less common alternative utilised halved logs set vertically in the ground to form palisade-like walls. This technique was still used in Europe up to the l lth. century, for example in the churches at Greensted, in England and Lund in Denmark. Of course, the notched log-house was taken to the New World by settlers from Europe. The log cabin was widely used in the forested areas of the USA and Canada where rapid colonisation required the construction of a large quantity of residential accommodation using relatively unskilled labour in a short time.

Figure 8

Farm log building in Nonvay (Photograph by J.N. Christensen, NIKU).

In North America, as prosperity increased, wood was still used as the prime construction material but using finished timber, produced by mechanised sawmills, in framed construction. The most common methods of timber-framed construction are platform frame, where storey-height wall panels a set on the intermediate floor platforms and balloon frame where two-storey external wall panels are erected first then the intermediate floor is suspended from them. STEP/EUROFORTECH - an initiative under the EU Comett Programme

Development of timber roof structures Ancient Greek roof construction was based on a beam and post system. Longitudinal ridge beams were supported either on props from massive horizontal main beams up to 13 m long or on an internal colonnade. Depending on the roof span rafters then spanned from the ridge beams to the perimeter walls, producing a low pitch. The Romans developed triangulated trusses, with spans up to 30 m, for the roofs of their basilicas and these greatly influenced the form of medieval Italian and later European roof structures. The Romanesque roof trusses of St. Paul's Outside the Wall, in Rome, which exceeded 24 m span, were repaired in 816 and finally destroyed in a fire in 1823. In Northern Europe the Romanesque truss was built with a steeper pitch, with more internal web framing to improve their structural efficiency. In the more domestic scale medieval buildings other roof structures were developed. Crown post roofs had braced props supported on transverse collar beams to maintain the longitudinal ridge timbers. Early timber trusses were often built without much understanding of the structural action involved and often resembled tied arches more than the modern idea of a truss, for example trusses in the roof of Wren's, Sheldonian Theatre, in Oxford (Mark, 1993). The early development of roof trusses is shown in Figure 9 to 12. The need to support purlins spanning between principal rafters of basic triangular trusses led to the development and widespread use of king post and queen post trusses (Figure 9 and 10). In both cases including "post" in the name of the truss is misleading as it implies that the significant elements are in compression whereas both the central vertical king post and the two vertical queen posts are in fact in tension under downward roof loading. Therefore care must be taken to design the joints at each end of the posts to transfer the tensile force which is present, usually by using metal straps.

Figure 9

Developnzent of timber roof construction: king post truss.

Figure 10

Deveeopment of timber roof construction: queen post truss.

Another form of roof construction used mainly for open roofs of halls and churches is that of the hammer beam, Figure 11, one of the best known examples being the STEPIEUROFORTECH - an initiative under the EU Comett Programme

E117

20,7 m span roof of Westminster Hall, London, completed in 1394 by Hugh Herland (TRADA, 1985; Mark, 1993).

Figure 11

Developnzent of timber roof construction: ( a ) hammer beam.

The parabolic arch is a very efficient structural form when subjected to uniform vertical loading provided that the supports are fixed in position. To overcome the problems associated with the imposition of lateral and point loads andtor movement of supports, timber bowstring girders emerged for use in spans from about 25 to 55 m. These consist of a laminated tied arch with web bracing. A related form is the Belfast truss (see Figure 13), developed in the 19th. century, which has a curved upper chord, a straight lower chord and a dense "criss-cross" pattern of web bracing. This truss form had the distinct advantage that spans up to about 36 m could be fabricated from short lengths of timber (TRADA, 1985).

Figure 12

Development of timber roof construction: Wren's Sheldonian theatre root

Figure 13

Belfast roof truss for an aircrap hangar at Duxford, England.

Massive timbers were used in roof and spire construction in the great Gothic cathedrals of Europe. For instance at Notre Dame in Paris the timber spire erected over the crossing in the early 13th century survived, though damaged, until its STEP/EUROFORTECH - an initiative under the EU Comett Programme

removal in the early 19th century and its replacement, in 1860, by the present spire designed by Viollet-le-Duc (Mark, 1993). Also, after the Norman tower over the main crossing of Ely Cathedral, in the East of England, collapsed in 1322, it was replaced by a timber structure consisting of an octagonal shell base supporting an octagonal lantern. This was completed in 1334. The total weight of the timber structure was approximately 400 tonnes, including the main lantern posts 19,2 m long, 508 X 813 mm in section, each weighing about 10 tonnes.

Timber bridges Early timber bridges constructed by the Romans were simple beam structures of hewn tree trunks spanning between timber piled piers. One of the earliest recorded is the Pons Sublicius built during the time of Ancus Marcius (640-616 BC) which survived, with regular repair, until the time of Constantine (306-337 AD) over 900 years later. The bridge known as Caesar's Bridge, across the Rhine, is believed to have been built under the direction of Vitruvius (the Emperor's Chief of Artillery) and, in a later drawing by Palladio, is shown to have longitudinal beams resting on cross-beams supported by inclined piles. An interesting joint was used to connect the piles and cross-beams so that the addition of load to the bridge deck caused the joint to become tighter. In 104 AD Trajan's Bridge, consisting of 20 piers up to 45 m high joined by semi-circular timber arches of 52 m span, was raised across the Danube River (ASCE, 1976). In 1570, Andrea Palladio, published an illustration of a 30 m span, timber-trussed bridge over the Cismone River, in north-east Italy, constructed around 1550 AD (see Figure 14). The joint details show an appreciation of, and are appropriate for, the forces that are generated by the pedestrian loads on the bridge which are supported on the bottom chord of the truss (Mark, 1993; ASCE 1976).

Figure 14

Palladio's design for a 30 m span trussed bridge.

At Queen's College, University of Cambridge, a timber bridge, known as the "mathematical" bridge, designed by William Etheridge, was erected in 1750, to replace an earlier design. (The current bridge has been rebuilt, twice since, to the design of 1750). Although the bridge soffit appears to be an arch, it is in fact formed from straight timbers placed tangentially to the curve. These timbers extend to the level of the parapet handrail and are bolted together at all intersections thus combining to produce a truss (Labrum, 1994). In France during the 18th century timber bridges were developed which had level decks with low parapets supported on flat profile arches formed from planks clamped together to produce a laminated section. Examples such as the Pont Louis, over the Isar, near Fresingen, had spans up to 45 m (ASCE, 1976). Grubenmann also constructed some renowned covered bridges, with spans of up to 120 m, in the 18th. century (Stiissi, 1961). For instance, the complex trusslarch at Schaffhausen over the Rhine, built in 1754, and an elegant 60 m span arch bridge at Wettingen. His 30 m span bridge at Kubel combines arch and suspension principles with the STEPIEUROFORTECH - an initiative under the EU Comett Programme

E119

bridge deck effectively tying the arch and the roof structure acting as a strut between the tension elements (see Figure 15). In all of these structures complex scarf joints are used to form long tension members from shorter pieces of timber.

Figure 15

Grubenrtlann's bridge at Kubel.

During the dramatic period of railway expansion in the 19th century many bridges and viaducts were constructed from timber. The renowned engineer Isambard Kingdom Brunel used timber for viaducts and bridges on the South Devon Railway and South Wales Railway. As these structures were quicker and cheaper to construct initially they therefore appealed to the under-funded railway companies. For example, for the Bourne Viaduct of 1842 he used a combination of a "king post" truss and "queen post" truss to span 20 In and for the Landore and Usk Viaducts he erected timber, trussed arches of up to 30 m span to carry the trackbed. In Cornwall he used a timber modular system for a series of bridges and viaducts. For instance, the St. Pinnock Viaduct had, at the top of each masonry pier, three frames each of four Baltic Pine struts, arranged in a fan shape to support the beams of the track-bed. Initially, the general public were troubled by the slenderness of the bridge components but eventually they came to have full confidence in the skill of the engineer. Brunel had accepted that constructing timber bridges and viaducts would result in a continuing maintenance commitment and had had the timber treated with preservative (kyanised, using a solution of mercuric chloride). Although the timber structures remained serviceable they were replaced by embankments or masonry arches as funds permitted, due to the high cost of upkeep (Vaughan, 1991 ; Otter, 1994). In the USA timber has been used extensively in bridge construction as native forest trees of great length and large cross-section were readily available to settlers. The shortage and high cost of other construction materials meant that timber bridge and truss design developed more quickly than in Europe. Early North American timber bridges were of simple beam and pile construction. For example the "Great Bridge" over the Charles River erected in 1660, Samuel Sewall's York River Bridge, in Maine, 82,3 m long and having 13 piers, constructed in 1761 and the Charles River Bridge at Boston, 458 m long and having 75 piers, built in 1786 also by Sewall. In 1785 Enoch Hale erected what was probably the first bridge in the USA requiring more than simple span beams, the bridge over the Connecticut River at Bellow Falls, Vermont. From 1792, Timothy Palmer constructed innovative trussed arch bridges and, in 1796, patented the idea of fully-enclosing timber bridges with roof and walls to improve their expected life from the accepted 10-12 years to up to 40 years. His works included the "Permanent" Bridge over the Schuylkill Rfver at Philadelphia (1804-6), which had spans of 45,7 l 59,5 l 45,7 m acting as a continuous truss over the piers, and a bridge over the Delaware River at Easton (1806-7) which was covered and remained serviceable until its replacement in 1895 when the applied vehicular loads had become excessive. Lewis Wernwag produced a total of 29 bridges, the most notable being the "Colossus" over the Schuylkill River in Philadelphia (1812). Fletcher and Snow in (ASCE, 1976) said that "This bold design, scientific and architecturally beautiful, STEP/EUROFORTECH - an initiative under the EU Comett Programme

probably was never surpassed in America". It was a fully-covered, trussed-arch, timber structure, of 103,7 m clear span and 6,l m rise, incorporating 5 parallel trusses having arch-rib lower chords 1,07 m deep by 0,33 m consisting of 7 layers of timber clamped together. In 1813-14 he also built a timber truss bridge over the Delaware River at Newhope. This bridge had timber parallel top and bottom chords, timber verticals but "X" iron diagonals in each panel, a truss form that anticipated the Pratt Truss. Theodore Burr patented his bridge truss system (parallel chord truss combined with an arch) in 1817. Burr's truss system was used for many timber bridges throughout the USA, a notable example being the Waterford Bridge (1804) over the Hudson River in New York State. The four spans were between 47 and 54,9 m and the covered bridge survived for 105 years until destroyed by fire in 1909. In 1820 Ithiel Town patented his lattice web truss and, in 1835, the modified form with double webs and secondary chord members. The Town truss was eminently suitable for timber construction as it could be made from standard-sized timbers of relatively small cross-section and length which were assembled using readily available bolts and rods rather than special joint assemblies. If suitably detailed and covered to protect the structural components from decay 60 m spans for railway bridges were expected to survive for at least 50 and up to 100 years. The cost of spruce timber bridges of this type and of medium span, in 1890, was approximately half that of an equivalent iron bridge. The truss patented in 1840 by William Howe had timber web bracing members that ran diagonally between the top and bottom chords over two panels and vertical metal rods connecting adjacent top and bottom chord panel points. This was modified later to have timber "X" diagonals in each panel. The Pratt Truss, patented in 1844, had the timber and metal web elements interchanged, with metal "X" diagonals in each panel and timber verticals. Compression elements were, therefore, shorter than in the Howe Truss and the joints between web and chord members were easier to fabricate.

Figure I6

Development of timber bridge trusses. ( a ) Androscoggim bridge Brunswick, 1804, ( b ) Town lattice truss, 1819, ( c ) Col. Longs brace truss, 1830, ( d ) Haupt lattice truss, 1839, ( e ) Howe truss, 1840,(f) Whipple bowstring truss, 1841, ( g ) Pratt truss, 1846, ( h ) Whipple double-intersection truss, 1846, ( i ) Bollman truss, 1852, ( j ) Fink truss, 1857.

STEP/EUROFORTECH - an initiative under the EU Cornett Programme

Ell1 1

Laminated timber structures Although laminated timber was in common use in other countries and had also been used for railway bridge construction in the UK (in particular by Brunel), one of the first major uses in a building was the train-shed roof of Kings Cross Station, in London. The roof was originally constructed using laminated timber arches of 30,5 m span. These were formed from boards 38 mm thick bolted together to make arch ribs 600 mm in depth. Subsequently the laminated arches were replaced by a metal structure due to deterioration of the timber, caused by steam from the locomotives standing at the station platforms. In 1866, arches of a similar form were used for the roof of the German Gymnasium in London (TRADA, 1985). Early laminated timber structures used mechanical methods, bolts, dowels etc., to connect the individual laminations. However, the development of synthetic resin glues, such as urea-formaldehyde, phenol-formaldehyde and resorcinol enabled joints to be made that were as strong as the timber being joined.

Long span structures (hangars, hypar shells, domes etc.) Some of the largest clear span timber arch buildings in the world are the non-rigid airship (blimp) hangars constructed for the US Navy, in 1942. A total of 16 hangars (the majority of which still survive) were built from treated Douglas Fir at nine different locations in the USA to house marine patrol craft. To conserve valuable steel and other metal resources the hangars (305 m long, 91 m wide and a maximum of 52 m high) were constructed using timber arch trusses 5,6 m deep, spaced at 6,l m centres. The arches were fabricated from various sections, typically double 100 X 300 mm chords and 100 X 200 mm to 150 X 200 mm web bracing, jointed using split-ring and shear-plate connectors. Trussed wood purlins were used to span between the arches and support rafters at 2 m centres which in turn carried tongue-and-groove boarding 50 mm thick (ASCE, 1989). Since the late 1950s timber has been used in the construction of shell structures that fully utilise their three-dimensional form to resist applied loads. An elegant example was the 46,6 m long by 28,3 In wide and 85 mm thick, turtle-backed, five-layer, teak shell constructed for the assembly hall of the Rangoon College of Engineering in 1956. Timber is particularly suited to the construction of shells based on hyperbolic paraboloid geometry as this double curved surface can be generated using straight elements. This was demonstrated by the erection of a 35 m square roof at Wilton Royal Carpet Company in 1957 (TRADA, 1985). Timber has also been used both to generate the outer surface of masonry domes (e.g. St. Marks, Venice, St. Pauls Cathedral, London etc.) and in more recent history to generate long span geodesic, lamella and ribbed domes up to 162 m plan diameter.

Medieval timber machinery and engineering works Timber was also used extensively in Europe for agricultural, industrial and military machinery and often as temporary works in the construction of major engineering projects. War engines, such as catapults, were constructed, since the time of the Assyrians and Greeks, for hurling missiles at the walls of cities under siege. In medieval times large timber frames were built to support battering rams used to break down fortifications. Water wheels were used for pumping water for irrigation and for driving corn mills and other machinery (some examples of water wheels at Hama in Syria up to 20 m in diameter). Timber post mills utilising wind power were constructed, from medieval times, for grinding corn in areas where water STEPIEUROFORTECH - an initiative under the EU Comett Programme

power was scarce. This required the construction of a stable timber frame base for the vertical post about which the mill structure rotates and large wooden rotating sails designed to resist high bending effects from the wind. Because of its durability in the corrosive marine environment timber was used extensively in the construction of docks, jetties and harbours. Also, until the advent of reinforced concrete and steel as structural materials timber was the only material available for piled foundations to buildings and engineering works on poor ground. Timber was (and still is) used extensively for temporary works in building and engineering, for instance centring to support masonry arches whilst under construction, formwork for casting reinforced concrete etc.

Concluding summary Representative examples of the development and use of timber in building and engineering structures, through history, have been presented. The significance of timber as a structural material for architectural and engineering structures has been demonstrated. It is hoped that this brief review will stimulate a desire to investigate further the history of timber structures as a way of informing contemporary design philosophy.

References American Society of Civil Engineers (1976) American Wooden Bridges, ASCE, New York American Society of Civil Engineers (1989) Classic Wood Structures, prepared by the Task Committee on Classic Wood Structures, Structures Division, ASCE, New York. Harris R. (1993) Discovering Timber-Framed Buildings, 3rd edition, Shire Publications Ltd., Princes Risborough, Aylesbury, UK. Hauglid R. (1980) Laftekunst - laftehusets opprinnelse og eldeste historie, Dreyers Forlag, Oslo Norway (with Summary in English). Labrum E. A. (1994) ed. Civil Engineering Heritage - Eastern and Central England, Thomas Telford Ltd. London. Mark R. (ed) (1993) Architectural Technolcgy

-

up to the Scientific Revolution, MIT Press.

Needham J. (1971) Science and Civilisation in China, Volume 4, Part 3, Cambridge University Press, Cambridge, U.K. Otter R.A. (1994) ed. Civil Engineering Heritage - Southern England, Thomas Telford Ltd. London. Stiissi F. (1961) Der Baumeister Johann Ulrich Grubermann und seine Zeit, Industrielle Organisation. Sumiyoshi T. and Matsui G. (1991) Wood Joints in Classical japanese architecture, translated by Ferenc Kovacs, Kajima Institute Publishing CO Ltd, Japan. TRADA (1985) Timber in Construction, Batsford, London. Vaughan A. (1991) Isambard Kingdom Brune1,John Murray, London.


Conceptual design STEP lecture E2 J . Natterer, J. L . Sandoz Swiss Federal Institute of Technology

Objective To introduce the basis of conceptual design, taking into account the more important quality criteria for timber in design.

Summary A summary of quality criteria for the design of timber structures is presented. The summary focuses on a discussion of general design considerations which have developed through many years of research and practice.

Introduction Many years of practice and research have resulted in the consideration of a number of problems, some of which are certainly applicable to other building materials but all of which are applicable to timber. Long term fundamental research and research work of shorter duration oriented toward practical applications have led to a variety of structural wood based products and more efficient connection details for instance. Despite the increasing role given to timber as an ecological building material, the fact that timber is avoided by many engineers is perhaps not so surprising since timber requires more judgement than other building materials because of its anisotropic behaviour and variability of properties. This is unfortunate because timber has so many good qualities. Given its properties, timber as a building material, requires particular conceptual work in the same way concrete and steel do in their own domains. The following paragraphs are intended to highlight some of these timber specific design concepts.

Roles of the architect and the engineer The architect is concerned with the totality of the building whereas the engineer generally focuses on structural optimisation. The key to a successful design is the determination of the critical criteria from which the final decisions are made. Structural optimisation should not be over-emphasised at the expense of other important criteria. Architectural aspects such as building function, space, circulation and lighting are certainly as important during the conceptual stages of a project as the design, fabrication and erection of the structure. For the architect and engineer, the basic design considerations include the determination of the areas of responsibility for the design of a timber construction. During the preliminary discussions, a design inventory which includes the development of space requirements and schemes for the building function is established. Their suitability within the environment surrounding the building must be considered. The deciding criteria should be put on paper and cleared with the owner and the building authorities. Some of the criteria are listed below:

- Restrictions

- fire protection - Site - acoustic protection - surrounding right of way - storey number - volume to be built

land access - aspect - topography - weather exposure

-


E211

- Restrictions

-

- Building Usage

- size - Loading - lighting - fire protection - in-plan organisation - functions distribution - space division

- Technical Installations

surface to be built building line materials ecology

- Site

- environment - geology

- dead load -

wind load snow load utility load crane load impact load

- ventilation - heating - lighting - acoustics - water supply - liquid waste evacuation

This check-list forms the basis that influences the decisions of the engineer. The overall aim is to achieve an optimisation for safety, cost, functionality, durability and aesthetic appeal. Elements of the architectural check-list which result from discussions between the architect and the engineer are outlined below.

Restrictions Designers have to work within constraints. Legal restrictions set down by the authorities have become multi-faceted and although they result in increasingly heavy bureaucratic limitations, they provide an improved safeguard for the requirements of the jndividual and the community and they have a stabilising effect on the appearance of the built environment and surrounding landscape. Restrictions set down by the planning authorities have the same effects. Renewed emphasis on environmental considerations has become a necessary and welcome challenge in recent years.

Site The size and the shape of the site have an important influence on architectural features of a building, such as the height of the frontage and the angle of the roof. At the same time the requirements of the authorities concerned with the protection of landscape and monuments, the topography and adaptation to the existing built environment need to be considered very carefully. Site conditions can affect the orientation of the building which in turn affects the use of natural light. The engineer must work out the conditions of the site regarding climatic influences such as snow, wind and earthquake loads. In addition the ground conditions and accessibility of the site are criteria for the foundations and the determination of the size of the building.

Building usage The requirements for a building's usage determine its height, the number of storeys and the possible points of support. Special functional requirements should be taken into account from the very beginning of the design process. Certain building functions require natural lighting which will determine the roof shapes to a great extent. The current fire-safety requirements are also dependent on the intended use of the building.


Loading During the early stages of design, loads which are likely to affect the structure are evaluated. The engineer should be conscious of less common loads such as those due to various installations, cranes and storage of goods as well as the more common loads such as wind, snow, earthquake and traffic. The loading may well affect the structural requirements and allowable spans.

Preliminary design and planning The development of the shape of a building can be affected to a great extent by the surrounding townscape or landscape in order that it is consistent within an overall design framework. The profile image of a sports hall, that of the diving tower for a swimming pool or that of a garbage disposal plant, together with its lighting and ventilation, can become determining factors in the profiles of the surrounding built spaces and the shape of the structures.

Development of the building shape During the preliminary design phase, the architect develops the basic character of the building. During this phase, which usually ends with the submission of plans, the engineer must suggest, in the first place, an achievable structure. If he recognises the architectural image behind the preliminary design, then he will be capable, in co-operation with the architect, of developing structures that not only correspond to static criteria but also to functional and site-related requirements. The engineer should, at the same time, develop structural alternatives that ensure both the accomplishment of the design-concept and the consideration of the various technical and economic requirements. On the other hand, the architect must recognise the behaviour of the structural systems. The architect should be capable of transforming the static systems given to him into a building shape. Examples of building shapes are shown in Figure l . Only when the shape of the static systems has been adapted to the shape of the building, can further consideration be given to the choice of materials and connections. In the initial stages of a project it is particularly important to consider the viability of a great number of structural alternatives for various design criteria, by both architect and engineer, in order to ensure the quality of the end-product.

Figure l

Building shapes. ( a ) end elevation, (b) side elevation.

Development of the structural form The decision to allow the structure of a building to be visible on the inside andfor on the outside is important from both a creative and an economic point of view. In this regard timber is superior to other building materials due to aesthetic and insulation considerations. It is possible to achieve attractive architecture and fulfil structural demands with an exposed structural system. This can be achieved STEPIEUROFORTECH - an initiative under the EU Comett Programme

E213

by a suitable arrangement of main and secondary structural elements that can be linear, plane, curved or doubly curved. The development of the inner space must be in tune with the structural demands from which appropriate sections follow. The structural shapes for the building cross-section can be developed along the length or width of the building plan or equally well along the diagonal of the building plan. The main support systems define the visual appearance of the structural system as shown in Figure 2. Their design in combination with the wind and stability bracing give the building its characteristic shape. The overall behaviour of the structure as well as the effect of the bracing can thus be expressed from the outside of the building.

Figure 2

Some complex structural forms. ( a ) section, ( b ) elevation.

Load path systems The number of main supporting structures as well as their spacing and distribution are determined by the development of the ground plan. The design of the frame depends on the building usage as well as partitions and lighting. Locally unstable soil requires special foundations which in turn result in longitudinal structural support whereas circular or polygonal foundations lead to radial or spatial arrangements of the main supports. Roofs which are both functional and economical can be achieved by means of branching systems. Secondary support systems shape both the roof and inner space of the building. The secondary supports are determined by the position and the type of the main supports, as well as by the number, shape and type of connections of independent support units. Well selected secondary systems support small spans and, at the same time, fulfil stability requirements. A great number of economical and interestingly shaped arrangements of bracing can be obtained. Examples of typical primary and secondary load path systems are shown in plan in Figure 3.


XXX!

p--+--+---:, l

(d) A---*--+--d Figure 3

Load path systems. Primary systems: (a) linear-parallel, (b) lineardiagonal, ( c ) radial, (d) diagonal; secondary systems: ( E )linear-parallel, (B, diagonal, ( g ) tangential, ( h ) linear-diagonal.

Modelling of support systems for stability The shape and size of the main and secondary supports are developed largely according to their effective height, since the effective height is a cubic function in the moment of inertia of the section rigidity of the supports. The building function fixes the height between ceiling and floor below. The construction height affects the distance between ceiling and floor above. Low construction heights permit the use of joists andtor beams for small spans and require closer supports. Larger construction heights can take advantage of planar truss systems which are more economical for longer spans. Spatial combinations of planar truss systems eliminate the need for stabilisation of the individual systems. Structural support systems can be optimised by adjustment of the construction height as well as by the alteration of supports and by the integration of the bracing system with the main support system as shown in Figure 4.


Figure 4

Structural forms. Continuous systems: ( a ) standard, ( b ) with knee-bracing, framework, ( d ) truss, ( e ) cable-stayed; braced systerns: kneebracing, ( g ) split-column, ( h ) propped beam, (i) strut frame, ( k ) kneebraced strut frame.

(C)

v)

Shell structures are economical for roof systems of large spans since they require neither system nor member bracing. In addition, they respond well to the demand for minimum structure weight and maximum inner space.

Material selection The choice of material and sections for main, secondary and peripheral structural systems should be decided upon as early as possible in order to determine the support system requirements. In timber construction, a variety of sections is available: logs, square timber in single or assembled sections, as well as beams, planks and glued-laminated sections. A well thought out section geometry can not only provide stability but can also provide fire resistance (except in the case of small cross-sections) and insulation for sound and heat. The qualities of the surfaces offered by timber allow the architect a range of choices: natural, roughsawn, planed or polished. In the area of manufactured timber products, materials such as veneered sheets, particle boards and numerous types of sandwich boards are available to the user. Timber-concrete composite systems are used in timber construction in order to reinforce existing structures and to reduce material and production costs of new structures. Such composite elements are very often used for beams, floorsystems, panels, diaphragms and shear walls. They provide good acoustical properties, they are fire-resistant and flame-retardant. Bridges are often constructed with timber joists and a concrete deck which is used to distribute concentrated wheel loads and to protect the joists from chemically aggressive rain waters. There is, in addition to a variety of timber sections, a choice in the types of timber to be used for instance spruce, pine, fir, larch, birch and tropical timbers. STEPIEUROFORTECH - an initiative under the EU Comett Programme

The choice of a species depends on the expected strength of the required material, its availability and cost, and occasionally its visual effect. A combination of light and dark timber species may be used to better underline a detail or a structural system.

Fasteners In the process of detail optimisation, the expense on both material and labour regarding the fasteners should be in proportion to the aesthetic and technical demands. The type of fasteners and the technology for their production should be discussed as early as possible with the architect. The finished aspect of a timber construction is dictated to a great extent by the fasteners used and their cost. For instance, these can be timber-timber connections, dowelled joints or spliced steel fasteners with or without visible steel components. The choice depends on the materials used for the members such as roundwood, sawn timber or glued-laminated timber. Steel components which are exposed require constructional timber protection against weather andlor fire.

Representation of the structure Isometrics and perspectives An important criterion for the quality of a timber construction is the effectiveness of the enclosed space. The representation of the internal space by means of an isometric or perspective drawing as shown in Figure 5 is recommended as early as the preliminary design stage since it illustrates the distinctive character of a construction even to untrained readers of a drawing. In addition, this type of drawing aids in the establishment and the optimisation of the production technology for timber fabricators.

Figure 5

Perspective of an ice arena.

Construction model A construction model is an important means for the representation of the inner space. It allows the quality of the design to be represented in an understandable fashion to decision makers such as building committees, municipalities and especially, to the prospective owners. On a construction model, details become visible. Discussions between architect and engineer are facilitated. Detail variations of the support structure and of the elements of construction become obvious and are sufficient for the aesthetic appraisal at scales of l to 20 or 1 to 50.

Computer modelling As timber engineering deals with a relatively easy to use material, there is an increased freedom for structural design and detailing. The contribution of CAD is therefore that much more valuable. An additional reason for the use of CAD is the large investment in labour costs required to determine the geometry as well as the detailing and working drawings. By using CAD in timber design, even STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E217

very complex space connection details can be shown easily to fabricators. It can be an advantage for both the architect and the engineer to work from the very beginning with spatial representations. This requires three-dimensional CAD programs that are not only drafting programs but also a means of alding thea building engineer. Design alternatives which include structural analysis as well as the simulation of inner space with the representation of surfaces and colours are all possible. Interactive data processing in constructional timber engineering results in a simplification of the recording and modelling of geometric structures for structural analysis as well as for the development of drawings. This accelerates the optimisation process considerably. The aim in the development of CAD is to define the building and the structure in spatial terms. Based on the support of the data obtained, the computer is able to carry out the planning of the construction and to transfer the total package in order for workshop drawings to be produced and to provide instructions for automated fabrication machinery.

Conclusion This chapter is intended to provide a general outline of requirements for the design optimisation of timber structures. The co-operation between the architect and the engineer is stressed from the beginning of a project study until the final presentation to the authorities and the potential buyers. It is only by increasing the engineering component in timber structures that constructional savings can be made making timber competitive with other building materials.

References Gotz, K.-H.; Hoor, D.; Mohler, K.; Natterer, J. Holzbau Atlas, Institut fur Internationale Architektur-Dokumentation GmbH, Munchen, 1978. Gotz, K.-H.; Hoor, D.; Mohler, K.; Natterer, J. Construire en Bois, Presses Polytechniques et Universitaires Romandes, Lausanne, 1987. Gotz, K.-H.; Hoor, D.; Mohler, K.; Natterer, J. Timber Design and Construction Sourcebook, McGraw-Hill, Inc., 1989. Natterer, J.; Herzog, T.; Volz, M. Holzbau Atlas Zwei, Institut fiir Internationale ArchitekturDokumentation GmbH, Miinchen, 1991. Natterer, J.; Herzog, T.; Volz, M. Construire en Bois 2, Presses Polytechniques et Universitaires Romandes, Lausanne, 1994.


Purlins STEP lecture E3 C. Short Consulting Engineer

Objective To show the design and detailing of purlins in compliance with EC5.

Summary Purlins from round pole, solid or glued laminated timber, and built up sections of I- or Box beams are discussed. The basis of design is developed, the ultimate and serviceability limit states are identified, design values of actions and resulting load cases are developed. Methods of verification or design conditions are addressed for members with biaxial actions in relation to solid timber. Durability, structural detailing and control in relation to purlin members are discussed.

Introduction A purlin is a horizontal member in a roof supported on the principals and supporting the common rafters.

Figure 1

Section shapes used as purlins. (a) Round solid timber or natural timber poles, (b) solid timber sections, sometimes used in groups fixed together to act as a single unit, (c) glued lanzinated sections, ( d ) Laminated Veneer Lumber, (e) I-beams, (f)box beams, ( g ) trussed rafters with punched metal plate connectors, sometimes fixed together and used in groups to form a multiple truss, (h) truss with glued or nailed plywood gussets, bolted, or toothed plate connections.

Purlins can be manufactured from a variety of materials and in many configurations. The most commonly used shapes are those shown in Figure 1. Purlins may be constructed as simply supported beams as shown in Figure 2a, or as continuous beams as shown in Figure 2b. If the timber element is continuous over three or more spans the arrangement will reduce the design moment and the effective deflection for a given section due to the continuity over the support. The limit for this structural arrangement is the maximum continuous length which the timber can be purchased or transported to the construction site.

STEPIEUROFORTECH - an initiative under the EU Cornett Programme

If)

Figure 2

1

l

Typical purlin forms.

When the building length exceeds the maximum timber length one of several structural forms may be adopted to simulate the properties of the continuous beam by connecting together successive timber lengths to achieve continuity. Two methods are shown in Figure 2c and 2d. In the first case the bolted splice would occur at a point of contra-flexure and in the second case the continuity is maintained by the moment of resistance in the plated or bolted joint close to the support. Trussed purlins, braced spans and reinforcement with tension steel are shown in Figure 2e, 2f and 2g, respectively.

Design In all but the simply supported span shown in Figure 2a critical points for the design would contain combined stresses from bending, shear, tension andtor compression forces. Care should be taken to evaluate the effect of the geometrical orientation of the purlin and the interaction of the force vectors which can combine to generate not only combined stresses but in some cases biaxial combined stress coexisting at the same point on the purlin. The structural design method is similar to that for a simple or continuous beam except that the purlin is generally subject to biaxial bending and torsional effects. Refer to STEP lecture B3 for basic beam bending considerations and STEP lecture B4 for shear and torsion verification. The normal loading arrangement on the purlin comes from the rafter and consists of the self weight of the rafter, the permanent load from the roof materials, transient load from snow or wind loads, and imposed load on the surface of the roof. The imposed load is generally applied to the horizontal projected plan area of the roof whereas the snow, weight of the roof materials and the self weight of STEPIEUROFORTECH - an initiative under the EU Cornett Programme

the rafter are applied to the true area and true length respectively. These loads have a line of action which is vertical. The wind loading is applied at right angles to the true surface area of the roof as a pressure, and can act in an upwards (suction) or downwards (positive pressure) direction depending on the prevailing wind direction and the geometric properties of the building being designed. The loads supported by the purlin should be determined from the structural action of the supported rafters.

Example Continuous purlin over 6 spans of 6 m length (see Figure 2d) roof slope a = 10" purlin spacing e = 1,15 m

Timber strength class C24 Service class 1, load duration class short-term kmod = 099 f = 16,6N/mm2

Characteristic values of actions Permanent actions (self-weight) 0,20 kN/m2 (roof area) Roofing 0,06 k ~ / m(roof ~ area) Insulation 0,lO kN/m2 (roof area) Ceiling 0,10 kN/m2 (roof area) Purlins and bracing 0,46 kN/m2 (roof area) G, = 0,47 kN/m2 (horizontal area) Variable actions (snow and wind): Q, = 0,75 kN/m2 (horizontal area) Wind action is suction only and not governing.

Design values of actions

Partial safety factors y, = 1,35 Ye =1,5

qy,d

permanent actions variable actions

+ yQQ,) = 1,15 sin lO0cos10" (1,35 0,47 + 1,5 0,75) = 0,35 kN/m

= e sin 10" cos loo( y, G,

.

-

Bay moments: d = qz.d IZ d = 77 qy,d IZ STEP/EUROFORTECH - an initiative under the EU Comett Programme

end span

0,078

5,50

0,98

second span

0,034

2,40

0,43

third span

0,043

3,03

0,54

The position of the connections between the timber members is chosen in a way that the bay moments govern the design. Although the bending moments at the supports are larger, they are not governing since two cross-sections are present at the supports. The sum of the moments in the two adjacent bays always exceeds the moment at the support for uniformly distributed loading and constant bay lengths. From the condition that the moment at the connection position equals the value of the maximum bay moment, the following positions result (see Figure 3): Inner bays: l,, = 0,10 1 End bay: l,, = 0,17 1

end span

Figure 3

second span

Simplqied moment distribution and connection forces at support.

Forces in the connections (inner bays): The forces to be transferred in the connections are determined assuming a constant moment in the single cross-section over the length l, from the support and disregarding the influence of the uniform loading on the cantilevered part. This assumption is slightly conservative.

K, K,

M , ,l = M+ /lc

=0,43q,L =5,06kN = 0,43 q, 1 = 0,90 kN

Ultimate limit state design

End bay: C24, b

X

h = 120 x 160 mm


Inner bays: C24, b X h = 60 X 160 mm The calculation corresponds to that for the end bay.

Connection in the second bay choose 6 ringed shank nails, d = 4 mm, L = 110 mm

Lateral load-carrying capacity

p t1 t,

d

p,

=l,O = 60 mm = 50 mm =4mm = 350 kg/m3 = 0,082 pk d '0,3 = 0,082 . 350 4-Ox3= 18,9 N/mm2

EC5: Equation 6.2.lf is governing. R,,, = 873 N

Axial load-carrying capacity Assume: = 6,12 ~ / m m ~ = 50 - 1 0 . ~pk2 f,, =600.10-~p: = 7 3 , 5 ~ / m m ~

Interaction

If it is appropriate to limit deformations, the deflections may be calculated as for a continuous beam using the bending stiffness of the single cross-sections. STEP/EUROFORTECH

- an initiative


E315

Straight and tapered glulam beams STEP lecture E4 H.J. Larsen Danish Building Research Institute

Objectives To give examples of the use of straight glulam members, and of joints between beams and other structures. Further, to describe the strength and stiffness properties for inhomogeneous glulam members.

Summary The lecture begins with examples of plane structures made from straight glulam members. For single span beams with constant depth guidelines are given on recommended spanldepth ratios. Some typical details are shown. The advantages of tapered beams are described and some details of the design are given, e.g. determination of the maximum bending stresses and deflections. Lastly, the determination of characteristic values for glulam with different lamination qualities in the outer and inner laminations is described, together with an example of the calculations.

Figure I

Examples of plane structures made from straight members. (a) Truss made of two beams and with a tension tie of steel (or glulam). (b) Large truss with top and bottom chord of glulam. The lattice can be made of glulam, solid timber or steel (tension members). (c) Trussed beam with top chord and posts in glulam. The tension tie can be made of steel or glulam. (d) Frame with glulam columns and (steel) tension tie. (e) Frame made of 2 X 2 straight members. The corner can be made with gusset plates of plywood or steel or with large finger joints.

Introduction Straight glulam members with constant cross-section are one of the most common structural elements. Beams can in principle be produced in any size. In practice, the size is limited by the capacity of the production equipment and transportation problems. Volumes above 10 m3 require special consideration, but beams up to 40 m' with lengths up to 50 m have been produced. The minimum width is about 60 mm. Beams with widths less than about 70 mm are often made by cleaving wider beams; they therefore have a tendency to warp STEPIEUROFORTECH

-

an initiative under the EU Comett Programme

E411

and the surface quality may be slightly inferior. If this is not acceptable, the buyer will have to specify that the members shall be produced directly to the specified width. For widths over about 200 mm, each lamination is often made from two boards placed side by side. This gives a more costly production, and it may often be cheaper to use twin members. If a lateral interaction is requlred, it can be ensured by mechanical connections or gluing. However, gluing of this type is not covered by European standards. The minimum depth is 100-135 mm corresponding to three laminations. Straight members can be used alone as beams or columns, but can also be used to build up more complicated 2- or 3-dimensional structures. Examples of plane structures are given in Figure 1. This paper deals only with simple beams.

Single span beams with constant depth Straight beams with constant cross-section are commonly used in roofs and floors, and as wall plates. The minimum depth is about 1/20 to 1/17, where l is the span. To reduce problems with lateral instability the width should not be less than h/7, where h is the depth. Most glulam factories produce "standard beams" according to a national size standard. "Standard beams" normally have a shorter delivery time and are cheaper than individually produced beams. Free spanning beams are normally cambered corresponding to at least the final deflection from the dead load and the quasi permanent part of the variable loads. Normally only deflections LL,, from bending are taken into consideration. The deflection from shear U, can be found from U Junl = UJu,,,

15 ( ~ 1 ) ~

- 1 ~(wL)'

for a uniformly distributed load for a single midspan force

For l/h 2 15 the contribution from shear is less than about 8%.

Continuous beams with constant depth Beams over several spans are often made as continuous beams or as cantilever beams, see Figure 2. They have a more favourable moment distribution, especially where there is more than one load case, and continuous beams are stiffer than corresponding simply supported beams. The minimum depth can therefore be reduced to about 1/25 - 1/20. Generally, continuous beams are not given any camber.

Figure 2

( a ) Continuous beam. ( 6 ) and ( c ) Cantilever beams. For cantilevered beams the joint pattern slzown in ( b ) should be chosen. For tlze pattern ( c ) tlzere is an unnecessary risk of progressive collapse.

Cantilever beams are very common in structures made from structural timber because of problems in getting long lengths. In glulam structures continuous beams are normally more advantageous. In cantilever systems the joints should STEPIEUROFORTECH - an initiative under the EU Comett Programme

be located in such a way that the bending capacity is utilised both in sagging and hogging (often corresponding to different load cases) and such that failure fn one beam will not spread to the whole structure. To avoid tension perpendicular to the grain failures in the notched beam ends, the short beams should preferably be hanging from the cantilevered beams.

Single tapered beams Single tapered beams, see Figure 3a, are very common both alone and as part of roof panels. The volume is slightly higher than for a corresponding beam with constant depth, but they are often advantageous because secondary beams, roof panels, etc. can be placed directly on them, and eaves boards are avoided. The slope is normally between 1/40 and 1/10 (a c: 5"). The mid-depth should not be less than about 1/20 and the minimum depth 1/30. The beams can be made with camber.

Figure 3

( a ) Beam wit11 single taper. (b) Beam with double taper. The beam widths are b.

The strength has to be controlled at two cross-sections: section A, where the shear stresses are maximum, and section B, where the bending stresses are maximum. For simply supported beams with a uniformly distributed load, q, the distance X from the support to the critical section is X

= 1 hn,in/(hnt,n+ ht7,<,x)

(1)

and the maximum bending stress is

on,, = 0.75 q, l2/@ h,,,, hn1ux) For other increased found by along the

(2)

types of load, or if the characteristic value of the bending strength is by the factor k, for depths less than 600 mm, the maximum stress is calculating the stresses at different cross-sections at regular intervals span.

It should be verified that

STEPfEUROFORTECH - afi initiative under the EU Cornett Programme

The factors k,,),, and k, take into account the influence of taper on the stress distribution and on the strength, see STEP lecture B8. The maximum bending deflection U,, from a uniformly distributed load can be calculated from the corresponding deflection U , for a beam with a constant depth of (hnzin + hnlm)/2as where k, is given in Figure 4.

Figure 4

Factor k, to determine the maximum bending deflection of a beam with single taper (top) and double taper (bottom).

Double tapered beams Beams with double taper have the same advantages as single taper beams. In addition, the material economy is better because the depth follows the moments. Normally the maximum depth should not be less than about 1/20. The strength verification is in principle as for tapered beams. In addition, the strength of the apex cross-section should be verified in accordance with EC5, see STEP lecture B8. For a uniformly loaded, symmetrical, simply supported beam, the bending strength of the apex cross-section will never be critical. The most unfavourable bending situation will in this case be found at section B with a distance X from the support of and the maximum bending stress - for which (3) applies - is

o,,, = 0775 q,

hmin(2h,,,, - h,,,))

(6)

For other types of load the bending strength of the tapered beam parts as well as the apex cross-section have to be verified. The maxfmum design tensile stress perpendicular to the grain can be found from the apex moment M cl,,,,, as


For a uniform load the midspan deflection can be found from Figure 4 as for a beam with single taper; in this case the deflection is smaller than for a corresponding beam with a constant depth (h,,, + h,,,)/2.

Details Some typical details are shown below (numbering as shown in Figure 5).

Figure 5

Joint types. Joint types numbered l and 2 are shown in this lecture. The others are shown in STEP lectures D5 and 0 8 .

Figure 6

Joint type I , beam to wall. ( a ) Steel strap from the foundation nailed to the beam. ( b ) ( c ) Anchoring with a steel shoe or two steel angles. In the latter case, the angles as well as the bolts have to be designed for the eccentricity moments. 1 = asphalt felt.


E415

l

I

L.!

l l

l l

L.!

Figure 7

Joint type 2, column to foundation. (a) (b) Simply supported colunzns with nailed, bolted or screwed steel bar or column shoe. ( c ) Simply supported column with steel plates in a slot. It is important to place an effective vapour barrier between wood and concrete. (d) (e) Built-in columns with channels or glued in bolts cast into recesses in the foundation. The column end can be free or packed up to increase the load-carrying capacity in compression as well as bending. I = asphalt felt.

Characteristic values The characteristic values for glulam shall be determined in accordance with European Standard EN 1194 "Glued laminated timber - Strength classes and determination of characteristic properties". This standard also gives the characteristic values for a range of glulam qualities produced by most glulam manufacturers. For special lay-ups the characteristic values are normally calculated and published by the manufacturer. .In both cases glulam is designed as a homogenous material. The designer may prefer a special lay-up, e.g. to utilise local materials. The calculations needed in this case are briefly described below for the symmetrical non-homogenous glulam cross-section shown in Figure 8.


Figure 8

( a ) Cross-section of a glulam beam made with two different lamination qualities. ( b ) Stress distribution in bending. (c) Stress distribution under axial load.

Regarded as solid timber, the inner laminations have the properties Ei, ft., etc. (note that some subscripts have been omitted for the sake of simplicity. For example, denotes the characteristic tensile strength parallel to the grain for the inner laminations for which the complete notation is f,,,,,).The outer laminations - of a better quality - have the properties E,, f,,,,, f,, etc. The corresponding properties for the laminations regarded as blocks of glulam are denoted E,,i, fm,,,i and f,,,, etc. and are found from EN 1 194, e.g.: (1,25 - Ei/60000) Ei with Ei in N/mm2 E,,i

= max

1,05 Ei

The effective stiffnesses by = (E,,,(h"hf)

@A),

in bending and (EA),f under axial load are given

+

h i3 ) b/12

= ( 1 - ( l - E g a,()i l)(h,lh)3)~,,0bh3/12 ~ = k, E bh 3/12

( 11)

= (1-(1-E,,,/E,,,>(h,/h))E,,"bh = k, E,,,bh

(12)

This corresponds to a formal modulus of elasticity in bending for the glulam regarded as a homogenous material of E,

= k, E,,"

(13)

For axial load a slightly lower modulus of elasticity - k, E,, - is found, but EC5 permits the use of E, in all cases. The stress distribution in bending is shown in Figure 8b. The strength will always be determined by the stress in the outermost fibres. The characteristic bending capacity is thus Mk

= 2 (El)cffnt,g,,l(Eg,oh) = ' 1 fm,g,obh2/6

corresponding to a formal bending strength of fm,,

= ' 1 fm.K.0

The stress distribution under axial load is shown in Figure 8c. The strength will normally be determined by the stress in the outermost fibres, and the formal STEPIEUROFORTECH


E417

axial tensile strength is then

A.,

=

A,,

= k/. f,,g,iEg,oJE,,i

f,,,g,,o (16) and correspondingly for the compressive strength. In rare cases the stress in the inner laminations may be critical. In this case, the formal tensile strength is kA

(17)

Example As an example, the properties of a glulam member with h,lh = 213 and with laminations of strength class C27 and C16 according to European Standard EN 338 "Structural timber - strength classes" are calculated. The laminations have the following properties:

0,806 19,7 = 15,9 f,,,

= 15,9 ~ / m m ~

= min

0,806 - 14,7 . 12600J8930 =16,7


Curved and pitched cambered glulam beams STEP lecture E5 H.J. Larsen Danish Building Research Institute

Objectives To give examples of the use of curved and pitched cambered glulam beams and to describe some special design considerations for these types of structures, especially the influence of moisture variations on the deformations and the internal stresses. Also, to describe possibilities of reinforcing the beams to reduce the influence of tensile stresses perpendicular to the grain.

Summary Curved beams and pitched cambered beams offer many architecturally interesting possibilities. Often the design of these elements is governed by tensile stresses perpendicular to the grain. These may be caused by external loads, but moisture variations may also have a major effect: they give rise to internal stresses and increase the deformations from external loads. To reduce the risk of failure perpendicular to the grain, curved and pitched cambered beams are often reinforced. Two reinforcement systems are described: glued-in steel rods and glass fibres glued to the surfaces. The design of the reinforcement is described together with a design example.

Curved beams Curved beams, i.e. members where the actions are predominantly carried by bending, as opposed to arches, where most of the actions are carried by axial forces, are used to achieve special architectural expression. Examples of the use of single curvature beams are shown in Figure 1.

Figure l

Example of a structure with single curvature beams.

In the structure shown in Figure l a the main actions (dead load and snow load) will reduce the radius of curvature, i.e. give rise to compressive stresses perpendicular to the grain; the tensile stresses perpendicular to the grain (from wind) will normally be small and not cause any serious problems. In the structure shown in Figure l b the main action on the cantilever will cause relatively large stresses perpendicular to the grain, and these stresses may be critical for the load-carrying capacity. It is possible to create architecturally interesting structures by using beams with multiple curvatures in a single plane; an example is shown in Figure 2. However, production of such members is difficult, they are expensive, and it is impossible to avoid relatively large stresses perpendicular to the grain. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E511

Figure 2

Example of a structure with reversed curvature beams.

Except for the problems related to the stresses perpendicular to the grain, the design of curved beams poses no problems. The strength verification is as for corresponding straight or tapered beams with a small axial force.

Pitched cambered beams Pitched cambered beams, see Figure 3, are used for roof slopes between about 1/10 and 113 ( a between 5" and 20"). The minimum depth h,,, should not be less than 1/30, where 1 is the span. The apex depth h,,,,, is normally between 1/15 and 1/10.

Figure 3

Pitched cambered beam comprising two tapered parts with the centre portion (the shaded apex zone) having a curved soffit. ( a ) Reinforcement bar.

The strength of the tapered parts is verified as described in STEP lectures B8 and E4. The strength of the apex zone is verified as described in STEP lecture B8. The strength of the beams is very often governed by tensile stresses perpendicular to the grain in the apex zone, and insufficient consideration of these stresses - or failure to consider them at all - has led to a considerable number of failures. Stresses perpendicular to the grain may result not only from external loads but also from moisture variations. For structures in Service Class 1 dried to an appropriate moisture content before erection the effect of moisture variation may normally be assumed to be covered by the assignment of characteristic values and the normal safety factors. In some cases, however, it may be necessary to estimate the influence of moisture variations. Assume as a rough estimate, see Figure 4, that the moisture content in the outer sixth on each side is decreased by 3% corresponding to a free strain of about 3 0,002 = 0,006. With .F9, = 300 N/mm2, the internal tensile stresses will be Acs = 2 300 0,006/3 = 1,2 N/mm 2, i.e. of the same order of magnitude as the short-term tensile strength, and there is a risk of splitting. In practice, creep will reduce the stresses, but they should not be neglected. STEPIEUROFORTECH - an initiative under the EU Comett Programme

Figure 4

The influence of a decrease in the moisture content in the outer sixth. (6) Free strain, ( c ) resulting internal stresses.

Deformations The vertical deflections of curved and pitched cambered beams are normally of no importance. The horizontal deflections at the supports may be rather large, however, and it is necessary to ensure that they can take place without giving rise to unforseen horizontal forces on the supporting structures (walls and columns) or damage to secondary elements. Not only external forces but also moisture variations contribute to the deformations. The main effect of an increase in the moisture content will be an increase in the depth of the beam from h to h(l + E), where E is the strain corresponding to the increase in the moisture content. Since the moisture influence in the fibre direction is marginal, the result, see Figure 5a, is that the angle d q is reduced to dq' and the radius of curvature is increased from r to r':

These changes mean that the chord length is increased and the camber reduced, see Figure 5b, by AV

=

Figure 5

r ~ ( -l coscp) + a&q/2

(3)

Influence on the curvature of a beam of an increase in the moisture content. ( c ) Geometry after moisture increase, (d) undeformed state. e is the total midpoint deflection.

STEPlEUROFORTECH - an initiative under the EU Comett Programme

E513

Reinforcement As shown in Figure 3 structures with tensile stresses perpendicular to the grain can be reinforced by gluing in steel rods or by gluing plywood or a fibre material to the surfaces.

Reinforcing materials Glued-in steels rods are treated in detail in STEP lecture C14. They can be smooth or threaded; in the latter case the calculations should be based on the tensile stress area. The effective beam width is reduced by the diameter of the hole or the outer diameter of the rod, whichever is the larger, and the width or the depth may have to be increased correspondingly. Various fibre materials - glass, aramid and carbon - offer interesting possibilities. For the time being, only glass fibres are of interest: they are cheap, easy to glue on the wood with polyester or epoxy adhesives and the reinforcement adds pleasantly to the appearance of the glulam: Glass fibres are delivered in sheets with unidirectional fibres, as mats made of about 50 mm long chopped fibres with random orientations or as woven sheets. In the last mentioned case the fibres are gathered in rather thick bundles (roving) and the surface becomes rough and less attractive than in the other two cases. The unidirectional fibres are the most effective. Typical values per unit width for a reinforcement of 100 ,g/m2 loaded in the fibre direction are (tf,,,),, = 60 N h m and (tE,),, = 3000 N/mm. The effective glass thickness (the sum of the layers on both sides) is denoted t, L,, is the tensile strength of the glass and E, the modulus of elasticity. If the load acts at an angle cp with the fibre direction the stiffness is reduced according to

The strength reduction is less. For a mat with random fibres the stiffness is 318 of the stiffness of a corresponding unidirectional reinforcement with the same weight, and the strength is halved.

Stress distribution The behaviour perpendicular to the grain is linear elastic to failure and the load perpendicular to the grain is distributed according to the stiffnesses. When estimating the load taken by the wood it is proposed to use the short- term mean stiffness values for both wood and reinforcement, i.e. for the wood E,,,,,,,,,,. Since the tensile strength perpendicular to the grain for wood is approximately proportional to the stiffness, there is no reason for taking "safe" values e.g. the 95-percentile, for the stiffness of the wood. According to EN 1194 "Timber structures - Glued laminated timber - Strength classes and determination of characteristic values", E, ,,,, = E,,,,,, 430, where E,,,, is the stiffness parallel to the grain. Normally, and on the safe side, the reinforcement is designed to take the total tensile force perpendicular to the grain, i.e. corresponding to the situation where the wood has failed in tension perpendicular to the grain. Where a more detailed calculation of the reinforcement is thought appropriate it is proposed that a low estimate be used for the stiffness of the wood, i.e. the 5percentile reduced by creep. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Tests on glulam beams with glass fibre reinforcement - see Larsen and Enquist, (1993) and Traberg and Larsen, (1993) - show that if the reinforcement is designed to take the full load the influence of volume and stress distribution may be disregarded, i.e. k , , ( ~ , j ~ ) ~=, ~ 1. The explanation may be that the reinforcement arrests small initial cracks in weak zones which in an unreinforced beam, would have led to brittle, catastrophic failure.

In some cases the adherence strength of the reinforcement to the wood may limit the strength of the reinforcement. According to Gustafsson and Enquist (1993) the adherence strength per unit width can be determined as

where Gf is the fracture energy of adherence, which may be taken as Gf = 0,35 ~mrn/mrn~. For tensile failure in steel rods or in the glass fibre a partial safety factor for yM = 1,l is proposed; for the adherence strength yM = 1,3 is proposed. Strengthening a n d repair There are many examples of failures in curved or pitched cambered beams but few total collapses, and experience and tests have shown that it is rather easy to repair the beams by filling the crack with glue, jacking the structure up and reinforcing it with a reinforcement capable of taking the full load perpendicular to the grain in the original beam.

Design example Structure Pitched cambered beams spaced at 4,8 nz and with a width b = 0,165 m are made of GL 28 (see STEP lecture AS). They have the following geometry, see Figure 3: r, = 13,50 m

l = 16,OO m

h,, = 0,60 m

h,,

= 1,50 m

a = 15"

Strength verification A dead load of 0,5 kN/m2 and a snow load (short-term) of 0,s kN/m2 give the following design load:

The shear strength at the support and the bending strength in the straight (tapered) parts and at the apex are sufficient. The maxiumum tensile stress perpendicular to the grain is calculated according to STEP lecture B8, Equations (14) to (18): Y

=

k5

= = = = =

k6

k7

'5 Or,90,d

ri, + h,J2 = 13,50 + 1,5012 = 14,25 m 0,2 tana = 0,0536 0,25 - 1,5 tana + 2,6 tan 2a = 0,0347 2,l tana - 4 tana = 0,2755 k, + k, (haJr) + k, (hair)' = 0,603 k, 6 MOpe,J(bhy2) = 0,0603 . 6 288/(0,165 - 1,5 ) I O - =~ 0,286 N/mm2

STEPIEUROFORTECH

.

- an initiative


h,Jr = 0,105 B8-(16) BS-(17) B8-(18) B8-(15) B8-(14) E515

The volume of the apex zone is V = 1,l m3, k,, = 1,7 andf,,,, = 0,312 N/mm2.

= 0,9

0,45/1,3

The tensile strength - see STEP lecture B8, Equation (21) - is not sufficient since 1,7 (0,01/1,1)0~2 = 0,664 . 0,312 = 0,207 ~ / m m ~ . 01,90,d > kdis( v ~ v ) ~ )=, ~

Reinforcement It is proposed that the beam be strengthened with a reinforcement that can take the full tensile force perpendicular to the grain, i.e. per unit length:

- E,,,,,,/30 The modulus of elasticity perpendicular to the grain is E,,,,,,, = 12000130 = 400 ~ / m m giving ~, a stiffness per unit length of 400 - 165 1 0-3= 66,O kN/mm.

Reinforcement with steel rods The reinforcement consists of 14 mm threaded rods (tensile stress area 118 mm 2) per 500 mm. Assuming a yield stress of 235 N/mm2 the design strength per unit length is 118 235/(1,1 500) = 50,4 > 47,2 N/mm. With E = 210 k ~ / m mthe ~ stiffness per unit length is 118 210/500 = 49,6 kN/mm, and the stresses in the wood perpendicular to the grain are reduced by the factor 66,0/(66,0 + 49,6) = 0,57, i.e. to 0,57 0,286 = 0,164 N/mm2, which is acceptable.

Reinforcement with glass fibres The reinforcement is made with 200 g/m2 unidirectional glass fibre sheets on each side with properties as given above. The average angle between the force and fibre direction is less than 5" and its influence on strength and stiffness is marginal. The stiffness per unit length is (tE,), = 4 design strength is sufficient since

(

4

6011,l

3000 =

10.~= 12,O kN/mm and the

218 Nlmm

FR.,*= min 165400

= 100 Nlmm

The stiffness per unit length is 4 - 3000 . 10-3= 12,O kN/mm, i.e. the stress in the wood perpendicular to the grain is reduced by the factor 66,0/(66,0 + 12,O) = 0,85:

References Larsen, H.J. and Enquist, B. (1993). Fibre Reinforcement of Glulam, Report No. 2, Small reinforced beams. Div. of Struct. Mech., Lund Institute of Technology, Report TVSM-7083, Sweden. Traberg, S. and Larsen, H.J. (1993). Fibre Reinforcement of Glulam, Report No. 3, Medium-sized and large reinforced beams. Div. of Struct. Mech., Lund Institute of Technology, Report TVSM7083, Sweden. Gustafsson, P.J, and Enquist, B. (1993). Fibre Reinforcement of Glulam, Report No. 7, Adherence of reinforcement to wood. Div, of Struct. Mech., Lund Institute of Technology, Report TVSM7083. Sweden. STEPIEUROFORTECH

-


Trusses made from laminated veneer lumber STEP lecture E6 A. Kevarinmaki Helsinki University of Technology

Objectives To describe the design of long span LVL-trusses with Multiple Nail Connectors and present typical examples of their use.

Summary The lecture begins with a description of the use and application of Multiple Nail Connectors (MNCs) in LVL-trusses. It presents the analysis of MNC LVL-trusses, the strength verification of members and the capacities and the strength verification of MNC joints. Some examples of LVL-trusses used in structures are presented.

Introduction The laminated veneer lumber truss is a long span roof header. LVL is a praduct similar to plywood except that veneers are parallel and longer lengths - nowadays up to 23 m - are available (see STEP lecture A9). Members of the LVL-truss are all constructed of two separate LVL-elements which are jointed together with Multiple Nail Connectors (MNCs), a jointing method of high load transfer capacity especially developed for this use. All members are jointed together with these MNCs to obtain centred node joints.

Multiple Nail Connector A Multiple Nail Connector (MNC) is a joint element used for jointing members of glulam, solid timber or LVL-truss. It is a steel plate which has nails welded on both sides perpendicular to the plate. Its function is based on the nail joint between the steel plate and the timber element. Flat cone nails shown in Figure 1 have been chosen to give an economical MNC-type, based on tests performed by Kalliomaki et al. (1986). The nails are rectangular in section, 3 X 4 mm2, and 50 mm long. The longer sides of the rectangular cross section of the nails are profiled to increase anchorage strength. The nails have a 4 mm long cone in the base for increasing the effective length of the nails to improve strength and rigidity of the joint. The nails in the corner plates of the truss are directed so that the longer sides of the nails are parallel to the truss member.

Nail spacing is usually 40 mm parallel to the grain (and thus the direction of the force) and 12 mm perpendicular to the grain in a staggered arrangement when the nail-type of Figure 1 has been used (see Figure 2). These nail spacings have been chosen according to calculations supported by joint tests (Kalliomiiki et al., 1986), such that failure occurs by exceeding the yield moment of the nail. Thus the failure of the joint is always ductile, when the number of nails in the force direction is no more than six. This leads to a maximum length of the nail group of 220 mm. The peeling off of Kerto-LVL has been observed in tests when longer nail areas with these nails and nail spacings have been used in MNCs. MNCs are produced by welding the nails with a peg welding machine to the steel plate which is cut to shape before welding. It can be cut either mechanically or with a cuttlng torch. To achieve a full strength weld the steel must be cleaned, normally by sand blasting. Generally the plate thickness is 10 mm and the steel grade Fe510D. The MNCs can be corrosion protected if it is needed; a zinc galvanization of 275 g/m2 or a hot-dip zinc coating of 400 g/m2 has been used. STEPIEUROFORTECH


E611

Figure I

The nail dimensions of a Multiple Nail Connector.

Figure 2

Normal nail spacings in MNCs with the nail-type of Figure l .

The use of MNCs in Kerto-LVL-trusses The members of Kerto-LVL-trusses are composed of two partial elements fixed together with MNCs. These members are also jointed with the same MNCs. In the truss nodes the members are assembled 1 to 2 mm apart (see Figure 3) so that the axial force of compression members is transferred through the MNC instead of by direct contact pressure between LVL members. Thus the MNC transfers forces according to the truss theory and no stresses perpendicular to grain are induced. The truss nodes are centred when MNCs are used. The shape of the joint element follows the shape of the truss corner and each transferred force in the members has a corresponding nail group on both sides of the steel plate. The top and bottom chords of the truss are usually continuous if they are under 23 m long except at the ridge joint of the top chords. Due to these continuous chords the MNCs need to transfer only the difference of axial forces in two successive internal members and, in the case of loaded chords, the support reactions of the chords. The ridge joint differs from the other joints in that the top chord members are tightly together. Thus the horizontal proportion of the chord force is transferred by timber contact. In manufacturing of the MNC Kerto-LVL-trusses, the timber elements of one side are spread out on an assembly jig after cutting. Then the MNCs are pressed into the elements using ribbed pressure plates. The plates are compressed one by one. The second set of side elements is assembled over the MNC nail points and pressed into position. The MNC nail pattern with the minimum nail spacings requires a applied ~ over the effective area of the steel compression pressure of about 7 ~ / m m plate during assembly. STEPJEUROFORTECH - an initiative under the EU Comett Programme

The steel joint plates of Kerto-LVL-trusses can easily be fire protected if required by one of two methods. Either a groove to fit the steel plate is milled into the members or, more commonly, the gap of the steel plate thickness between the members is filled with a timber strip. No reduction of the design value of the connector capacity is required for either kind of protected Qoint for standard fire resistance not greater than R60 (STEP lecture C19). The charrfng rate of Kerto-LVL is 0,8 mmhmin on the under side of Kerto-LVL beams based on information given by the manufacturer and type approval in Germany. Taking this charring rate value, the minimum protection thickness af is 12 mm in fire class R30 and 36 mm in R60 (see STEP lecture C19). For fire protection designed joints the distance between the MNC plate and member edge and the distance from the nail point to the LVL surface should be at least ay

Figure 3

Truss joint assembled with a Multiple Nail Connector.

Load carrying capacity of MNCs The nail capacity of MNCs may be calculated by Johansen's equations for steel-totimber joints (see STEP lecture C3). The shear strength of the welded nails in MNCs is calculated by the equations for thick steel plates (equations 6 . 2 . 2 ~and d from EC5: Part: 1-1). A typical yield mechanism and a free-body diagram of a MNC nail is shown in Figure 4. The cone shaped thickening in the base of the nail is presumed to be so rigid, that the plastic hinge is formed at the top of the cone as shown in the figure. When the nails are shorter or more rigid the failure mode number 2 with only one plastic hinge located in the base of the nail is also possible (see STEP lecture C3). When a nail has a cone shaped thickening so that the distance from the first plastic hinge to the interface between the steel and the timber is I,,,, the shear force capacity of the nail increases by a factor l,,, dh,,,. When the axial force effect is 10% as in EC5 (STEP C3), the equations of EC5 may be derived in the following form:

1 3 1

R, = min

MY,d

fh,d(t,- ken> d fh,d

d

('1

-

'con)*


(1)

Figure 4

EC5: Part 1-1: 6.3.1.2

Function of nails in the MNC, and afree-body diagram of a nail. ( a ) Plastic hinges.

The embedding strength f,,, for MNC nails may be calculated from the nail size d and the LVL density p, by the equation number 6.3.1.2a of EC5. The equation is for timber, but it is also suitable also for Kerto-LVL (Koponen et al., 1992). The characteristic density of Kerto-LVL p, is 480 kg/m3 (see STEP lecture A9). The design value of the embedding strength h,,is calculated using the actual value of the modification factor k,,,,given for plywood in EC5 (see STEP lecture A9) and the partial coefficient ,y = 1,3. The equations for yield moments M,,,, of round and square nails are presented in EC5 (Part 1-1: 6 . 3 . 1 . 2 ~and d). In these equations the characteristic strength of the nail steel f,,, has been assumed to be 0,8f,,,, where the tensile strength f,,,is 600 N/mm2.The yield moment calculated using the equations in EC5 corresponds with plastic theory when the nail size, d, is 8 mm, that is also the maximum nail size accepted by EC5. With smaller nails the yield moment of EC5 is higher than the value calculated value. For example with nails where d = 3,5 m m the calculated M,,, is exceeded by 36%. Higher yield moments for the special shape, or profile nails, or for high strength nail steel, may be utilized only if the nail yield moment M,:, has been tested according to prEN409 (see STEP lecture C4).

EC5: Part 1-1: 6.3.1.4(2)

The minimum spacings and distances of MNCs nails may be designed according to EC5 using the rules given for nailed steel-to-timber joints with the density value p, = 480 kg/m3.So in MNCs the minimum nail spacing in the main direction parallel to the grain and force is 10,5 d and the spacing perpendicular to the main direction is 4,9 d. The distance from the first nail line to the end of the Kerto-LVL member should be at least 20 d in tension and 15 d in compression. The edge distance between the truss member and the nail group should be at least 7 d. However, the peeling off of Kerto-LVL has been observed in tension tests when long nail areas are used in MNCs (see Figure 5). If the minimum values are used for the nail spacings, a nail group length over 220 m m in the loading direction should not be used in tension joints unless it has been confirmed by testing. Higher tension loads may be transferred without the risk of brittle failure by greater nail spacings or by two separate nail groups with an un-nailed part between the groups in the MNC plate. According to EC5 any influence in number of fasteners on the load-carrying capacity of nailed connections may be neglected (see STEP lecture C15). The tests of MNCs have shown that the number of nails has no significance on the mechanical behaviour of a single nail unless there are so many that the peeling off of Kerto-LVL has been observed. A big reduction for more than six MNC nails in line with the 'Load direction has been required: the load carrying capacity of the extra nails being reduced by 60%, i.e. for n nail rows the effective number n,, is: STEP/EUROFORTECH - an initiative under the EU Comett Programme

with 6 < n 5 15.

Figure 5

Peeling off failure mode of a MNC test joint with Kerto-LVL.

The following concerns load-carrying capacity of the MNC nail shown in Figure 1. The calculated yield moment M,,, parallel to the longer side of the nail (4 mm) is 7850 Nmm, when the increase over the theoretical value is taken into account, as in EC5 for a nail size d = 3,5 mm. The embedding strength f,,,of Kerto-LVL is 28,3 N/mm2 calculated according to EC5 when the nail size d is 3 mm perpendicular to the force. The length of cone l,,,, is 4 mm in the base, and the length of nail t , is 50 mm. According to Equation (1) the characteristic load-carrying capacity R, is 2,11 kN and the design value R, is 1,53 kN in the medium-term load-duration class with service classes 1 and 2. MNC tests with these nails and with the nail spacings of Figure 2 have been carried out. According to the tests over a series of different width and length joints, the average load-carrying capacity R, was 2,7 kN per nail and the characteristic value R, calculated from test results is 2,4 kN. This comparison shows that EC5 is suitable for estimation of the load-carrying capacity of the MNCs but it is somewhat conservative with these types of special nail. In this case the number of nails could be reduced by 15% if design is based on test results.

Analysis of MNC LVL trusses The analysis model of a MNC Kerto-LVL truss is simple, because the connections are designed without eccentricity, so that all systemlines of members coincide with the member centre line and in the node they meet each other always at the same point at the centre of the chord. Fictitious beam elements may be needed only to model the eccentric supports (see EC5: Part 1-1: 5.4.1 . l ) . The joints in compression have the same rigidity as in tension because any timber-to-timber contact between external and compression members is neglected. The use of a simplified analysis (EC5: Part 1-1: 5.4.1.3) is also normally possible. The axial forces in the members are determined assuming that every node is pinjointed. The bending moments in the chords are determined as if the member was a beam with a simple support at each node. The effect of deflection at the nodes and partial fixity at the joints may be taken into account by a reduction of 10% in the node bending moment. The reduced node moments should be used to calculate the span bending moments. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E615

Trusses should be analysed as framed structures in general analysis, where the deformations of the members and joints, the influence of support eccentricities and stiffness of the supporting structure are taken into account in the determination of the member forces and moments (EC5: 5.4.1.2). The analysis should be carried out using the appropriate values of modulus of elasticity given for Kerto-LVL in STEP lecture A9 and the joint slip defined below. The fictitious beam elements should be assumed to be as stiff as the adjacent elements. Joints may generally be assumed to be rotationally pinned, but the moment capacity and the rotational stiffness of MNC joints are taken into account in the chord splices and with other eccentric joints by modelling the nail group of a MNC with elastic theory. The joint slips of MNCs may be estimated by the slip modulus of a single MNC nail. The instantaneous slip modulus K,,, per shear plane per fastener under servlce load is given in EC5: 4.2. The nail slip modulus K,,, for Kerto-LVL with p, = 480 kg/m3 is 420~!.~ in N/mm. The final joint slip is calculated by kdefvalues given for calculation of creep deformation of plywood (STEP lecture A9). For MNCs the instantaneous slip modulus K, per nail for the ultimate limit state design should be taken as: (3)

KU = 2K,,??I 3 The final slip modulus Kutf,,is given by Ku,fin

=

Ku

'

(1

+

)

Strength verification of members The cross section of an internal LVL element is designed for the member force F42 using the design strength of Kerto-LVL given in STEP lecture A9. The external members are designed according to EC5 at the most critical point for the design load combination of the moment MJ2, normal force Nd/2 and shear force Vd2. It is dangerous to design the cross-section using the sum dimensions of parallel compression members where lateral buckling is possible. The standard thicknesses of Kerto-LVL are shown in STEP lecture A9, and the heights of members may be freely determined so that the cross-section will be fully utilized. The same values as with glued laminated timber may be used in the design of compression members of Kerto-LVL assuming the straightness limits: P, = 0,l. For internal members in compression, the effective length for in-plane strength verification is usually taken as the distance between the centroids of the MNC nail group areas. Also the influence of the rotational stiffness in these semi-rigid connections may be taken into account in a more sophisticated analysis (see STEP lecture B7). The effective length for chords in compression should generally be taken as the distance between two adjacent points of contraflexure (EC5: Part 1 - 1: 5.4.1.4). When a simplified analysis has been carried out, the effective lengths of the top chords may be simply assumed to be a certain factor times bay lengths, but then the calculated axial forces should be increased by 10% for the strength verification of members in compression and also in the connection design. The baylength-factors are given in EC5: Part 1 - 1: 5.4.1.4(3). The buckling lengths of the internal members for the lateral (out-of-plane) strength verification can be assumed to correspond to the distance between the braces that are often at the outer edges of the top and bottom chord (STEP lecture B7). The separate design of the parallel compression members is a conservative method in the case of lateral buckling, because the MNCs connect the ends of members to each other. If lateral buckling is the critical factor and no additional brace is STEPIEUROFORTECH - an initiative under the EU Cornett Programme

wanted, the theory of mechanically jointed columns may be utilized, but then usually additional MNCs are needed so that the shafts are connected at the ends and at the third points (EC.5: Part 1-1: Annex C).

Joint strength verification EC5: Part 1-1: 2.3.3.2

The cross-section of steel plates are designed according to the structural steel codes for the maximum loads of internal members using the material safety coefficient y, = 1,l. The width of the steel plate parallel to chord is designed for the force component of the chord direction in the joint node. When the plate has been cut so that each member part of the steel plate has a constant width, the centre lines of each part coincide with the centre line of the truss members. These centre lines meet in the node point at the centre of the chord. Thus the capacity of the steel plate in the joint line between the chord and the internal members will not be a critical factor. The number of nails for each nail group of internal members is designed to resist half of the member force FA2 using the design strength of the nail (n > FJ2R,). The truss member force is divided by two because the MNC has the same nail groups on both sides. The numbers of nails for the chord is calculated from the difference in forces of all internal members meeting at the node. In the strength verification of chord splices or other eccentric joints where a loading combination of N,, V, and M, is acting on the MNC plate at the ce~ltroidof the nail group area, the following condition ( 5 ) should be satisfied for every nail:

where

R, R , , and R , , n X, and

y,

is the design load-carrying capacity of round or square nails, are the design values of rectangular nails in X- and y-axis directions, is the number of nails and are the nail coordinates from the centroid of the nail group area when x-axis is parallel to the normal force N, and y-axis is parallel to the shear force V,.

Use of MNC Kerto-LVL-trusses The MNC Kerto-LVL-truss is an economical alternative when long spans (18 to 50 m), high loads (> 10 kN/m) or fire resisting requirements (R30, R60) are needed. The MNC joints have such a high load-carrying capacity and rigidity that they are not the critical factors in the dimensioning of members or in the deflection of rafters. Unnecessary wastage of Kerto-LVL material with its high strength and stiffness values is avoided because the height of every member may be optimised to the force in the element. The most common Kerto-LVL truss type has been made from 75 mm thicknesses of Kerto-LVL with members jointed together with MNCs using nails and nail spacings as shown in Figures 1 and 2. Thus the total width of the rafter is 160 mm STEPIEUROFORTECH - an initiative under the EU Comett Programme

E617

when a steel plate thickness of 10 mm is used. Typical Kerto-LVL-trusses have a span of 20 to 40 m with truss spacings of 4,8 to 7,2 m. Normally, long rafters over 23 m have been made from two parts jointed together on the building site by chord splices as shown in Figure 6. Two examples of site joints between Kerto-LVLtrusses and columns are shown in Figure 7.

welded splice

bolted splice

Figure 6

The building site splices of MNC Kerto-LVL-trusses.

Figure 7

Examples of joints between MNC Kerto-LVL trusses and columns.

References Kalliomaki, l,,Hirsi, H and Kanerva, P. (1986). Kerto-Laminated Veneer Lumber Truss. Report 14. Helsinki University of Technology Laboratory of Structural Engineering and Building Physics. Espoo, Finland. Koponen, S. and Kanerva, P (1992). Summary of European Kerto-LVL tests with mechanical fasteners. Report 29, Helsinki University of Technology. Laboratory of Structural Engineering and Building physics. Espoo, Finland.


Trusses made from glulam and solid timber STEP lecture E7 G. Steck Fachhochschule Miinchen

Objectives To describe the design considerations and analysis of trusses and to present two design examples.

Summary As a common construction method, the general design of trusses will be briefly explained. The numerous considerations in design and the working steps for the computation of the structure will be briefly presented. By means of two examples, the design and computation of trusses consisting of glulam or solid timber will be detailed using the rules in section 5.4.1 of EC5.

Introduction In timber structures, trusses are a wide-spread construction form taking full the potential advantage of the wood and the benefits of mechanical connections. Trusses find applications in wide spans, frame constructions, and as roof and wall bracing and are for example used in hangar constructions. The top chord of the truss usually is adapted to the shape of the roof and the profile of the bottom chord is governed by clearance, ceiling and construction height requirements. For simply supported trusses the top chord is stressed in compression and also in bending in the case of load distribution between the nodes. The internal members (diagonal and vertical) are arranged in such a way that a triangulated web is generated. In order to reduce the number of joints and with it the production costs, widely triangulated systems are preferred. In addition to the common triangular, trapezoidal and parallel chord shapes, a variety of special forms are possible for trusses. A selection of such trusses is recorded in STEP lecture B12, which also deals with the possibilities and constraints of three-dimensional systems. A number of design considerations forms the basis for the selection of the appropriate system, the material of the web members, and the fasteners in the truss joints. Then, the computation of the structure, the design of the joints and the internal members follows. The accurate transformation of the computational results into complete and detailed construction and working drawings represents the last but very important stage of work in order to achieve correct fabrication and erection of the truss.

Design considerations Depending on the use of the building, the following items have first to be clarified:

-

ground plan elevation protection from heat and moisture fire protection illumination openings roof drainage installation.


In order to select an adequate type of truss the following data are required:

-

-

minimum clearance slope of the roof height of the eaves and of the ridge span truss spacing roof loads.

Trusses spanning up to 30 m can be made of solid timber members. Members consisting of glulam, laminated veneer lumber (see STEP lecture E6) or parallam (see STEP lecture A9) may be used for wide-spanning systems or structures, whose wooden surfaces usually are to be free of cracks. Glulam is also preferred for truss members when large cross-sections are required. Cross-sections of spaced glulam members may be used for heavily loaded trusses of long span. The member forces can typically be transmitted by dowels and steel plates. Trusses make it possible to adapt the construction height, and also the crosssectional area and the quality of material, to the bending moment. Unlike thin webbed beams, installations and components can easily be run transverse to the plane of the truss. The selection of the fasteners (see also STEP lectures D1 and D2) is governed by the magnitude of the forces and the extent of the connecting areas, however it is also important to consider if steel plates and fasteners need to be hidden for reasons of aesthetics or fire protection. Joints, which can be industrially produced using machinery to a large extent, should be preferred to mainly hand made joints with higher labour costs. The design appraisal should also consider transportation and erection of the trusses (see STEP lecture D7), as these factors influence the arrangement and position of the chord connections.

Computation of the structure Part 1-1: 5.4.1

The system drawing, containing axial dimensions, numbers of members and nodes, joints and conditions of support, is required for the preparation of the structural EC5: calculations. According to EC5 the computation is classified into a simplified analysis and into a general analysis (see STEP lecture B12).

Simplified analysis determination of the axial member forces on the basis of a model where all nodes are pinned -

dimensioning of members and joints

-

calculation of the deflection

-

taking into account the elastic straining of the members and the displacement stiffness of the joints, or

-

taking into account only the elastic straining of the members. In this case it is expedient to halve the deflection limit values recommended in EC5.

General analysis initial calculation of the truss as a frame structure on the basis of estimated cross-sections of the members and, eventually, on estimated displacement stiffnesses STEPIEUROFORTECH - an initiative under the EU Comett Programme

-

dimensioning of the members and joints

-

improved computation of the internal forces and moments leading to improved cross-sectional areas and numbers of fasteners

-

computation of the deflection.

Example of a truss consisting of solid timber Provided that either the truss construction is hidden or, if exposed, flawless surfaces are not required, trusses consisting of solid timber are quite common for intermediate spans up to 30 m.

@

200 X 260 C24

@ 200 X 200 C24 @ 2 60 X 140 C24

Figure I

~ c ipart :

I - I : 5.4.1.3(1)

detail ,,B,,

4 @ type Cl l , d, = 80 mm

Kingpost truss made from solid timber.

A simplified analysis for the kingpost truss shown in Figure l is not permitted, because the condition H > 10 hChr,,,, is not fulfilled since H = 1200 mm and 10 h,,ord ,,,, = 2600 mm. 2 @ tjpe C10, d, = 80 mm o 200

Figure 2

X

260 C24

Detail "A".

l

\4~rlpeCII,dc=80mrn

I-9

Figure 3

Detail "B".


Material Members: solid timber C24 Characteristic values: f , , = 24 ~ / m m ~ A,,,, fc,90,k = 5,3 N/mm2 E,,,, p, = 350 kg/m3 E, ,,,,,

= 21 N/mm2 fr,o,k 2 = 7400 N/mm E,,,,, = 11000 ~ / m m ' E,,,,,,,

= 14 ~ / m m ' = 2 37013 N/mm2 = 370 N/mm2

Steel element in Detail B : Fe360 Fasteners: connector types C10 and C11, 80 mm diameter

Design value of action Permanent point load F, = 25,O kN km,, = 0,6 solid timber k,,, = 0,6 solid timber, service class 1

Calculation of internal forces The internal forces can be calculated by using a computer program for frames or as follows by solving the equation for the geometrical condition for the statically indeterminate force X in the member 2. U,,,

- u I x= 62x + 63Plsin a

+ 6,,,

/ sin

a + 6,, 1 sin a + 6,

(1)

is the deflection of beam 1 (span l = 6,O m) due to the point load F, is the deflection of beam 1 (span l = 6,O m) due to the point load X U~~ is the elastic deformation of member 2 parallel to grain direction and the 62,x deformation of beam 1 perpendicular to grain direction due to the force X 6 , is the elastic deformation of member 3 due to the force D = X / (2 sin a) 6,(,,, is the slip of the connector joint 4(5) is the vertical slip caused by tolerance of the hole diameter in Detail B 6, U

The instantaneous deformations are:

X 3231 = 3,50 2 sina . 7400 - 2 . 60 - 140 A = 8,oi - 10-5 X 2 sina 2 K , 2 mm (assumed)

-

6 4 ( 5 ) 9

6, STEP lecture C10: Eq.(6)

=

.

10-5 X

with K, = 2 K,s,, / 3 = 2 . 0,45 80 350 / 3 = 8400 N/mm, as 5-percentile values for K, are not available. Substitution in equation ( 1 ) gives X = 22600 N N , , = 22,6 kN M,,, = 3,6 kNm STEPIEUROFORTECH - an initiative under the EU Cornett Programme

= 22,6 1 (2 sin a ) = 30,4 kN N I , = 22,6 / (2 tan a ) = 28,3 kN ,d

When K,,, E,,,,,,, and E,,,, are used instead of Ku,E,, and E,,,,, the results are: X =18700N N , = 18,7 kN M,,, = 9,45 kNm D3,d = 25,2 kN N , = 23,4 kN The comparison of the results shows the governing influence of the bending stiffness of the beam.

Verifications Diagonal strut D, (member 3)

STEP lecture B6

Beam (member 1)

The instantaneous deflection of the beam midpoint at serviceability limit state due to the point load F, = F, l 1,35 = 18,5 kN is: u

~= u, * ,-~u I X~=

7 ~m m ~= 1 1857

and the final deflection is: ~ , , , ~ , , = 7( 1 +0,6) = l 1 m m = 11545 < 11200 Joints: see STEP lecture C10.

Example of a truss made from glulam Trapezoidal truss in the roof of a hall, service class 1 Members: glulam GL32h Joints: steel plate with t = 8 m m in slots & dowels of 16 m m diameter Wind load will be neglected. The top chord is laterally restrained at points 3,8 m apart.

STEPIEUROFORTECH


the EU Cornett Programme

1 Figure 4

22,5 m

Truss system.

Characteristic values of actions Permanent actions: Roof covering + purlins + bracing Truss dead load (0,15 + (22,5 - 15) / 200) 5,O top chord 2,O + 0,5 = Gto1,,, bottom chord = Gott,k

Load split

Snow load

2,O kN/m 1 ,O kN/m

=

= 2,5 kN/m = 0,5 kN/m

Qt,,k

=

g,ol,,, = 1,35 2,5 ghotSd = 1,35 .0,5 q , , , = 1,35 - 2,5 + 1,5 - 5,O qhott,d q,, = 1,35 . 2,5 + 1,5 5,O 1 2 qhott,d

= =

5,O kN/m

Load cases 1) g 2)g+S 3) g + s/2

3,38kN/m 0,68kN/m = 10,90 kN/m = 0,68kN/m = 7,12kN/m = 0,68 kN/m

Design values of action for verification of the

Characteristic values of actions

Serviceability limit states

Gk

Qk

top chord

23

5,O

bottom chord

OS

Ultimate limit states for load case

1,0Gk

It0 Q,

1)

2)

3)

2,s

50

3,38

10,90

left :7,12 right:3,38

0,68

0,68

0,68

0,5

Characteristic and design values of actions in kN/m.

Table I

Material prEN 1194: 1995

prEN 1194: all members in GL32h fm,g,k = 32,O N/inm2 .L,O,g,k = Nhm2 f,,90,,k E~,~,mean = 13300 N/mm2 E90,g.mrtm Eo,s,05 = lo600 (E,o,g,05 Pg,k = 430 kg/m3

x,o,s,k

3095

t Steel plates Characteristic yield strength f, Characteristic tensile strength f,

= 22,O N/mm2

ft.90,gjk

= 6,O ~ / m m ~ = 440 ~ / m m ~ = 350 ~ / m m ~ )

fv,g.k

= 0,45 N/mm2 = 2,9 ~ / m m '

= 830 N h m 2

(Ggo5

= 660 N/mm2)

= 8 mm Fe 360 according EN 10 025 = 235 N/mm2 = 360 ~ / m m ~


Modulus of elasticity E = 210 000 N/mm2 Dowel 0 16 mm Fe 360 according EN 10 025; flush with timber surface

Simplified analysis Conditions there are no re-entrant angles in the external profile; some part of the bearing width lies vertically below the support node; the truss height H = 3,O m > 10 h,,,, thus h,,,, c 300 mm; H = 3,O m = 0,133 1 = 15% of l.

430

2,25

430

1

'I

Figure 5

Structural system for the simplified analysis. Serviceability limit states design

at top chord

F,,, =

5,6

at bottom chord F,,, =

2,3

Table 2

11,3

Ultimate limit states design for load cases

75

24,5

left: 16,O right: 7,6

3,1

3,1

3,1

Design values of point loads in the nodes.

Calculation of the internal forces by using a computer program for frames. Verification for top chord: Load case 2 (g + S) governs: km,, = 0,9 N, = 284 kN (compression)

chosen

U 200 X 200 mm GL32h A = 40 -10 mmz W, = 1,33 106mm3


Az =

3810 0,289 - 200

=

66

thus: kcSz= 0,718

Verification for bottom chord: Load case 2 (g + S) is governing: km, = 0,9

N, = 270 kN (tension) 200 X 160 mm GL32h using A = 32 103mm2 A, = 3 2 . 103 - 10. 160 - 2. 16 (200-10) = 24,3 103mm2

.

Design load-carrying capacity of a dowel:

R, = 32,5 kN for a = 0"

Figure 6

Dowelled joint.

Calculation of the required number of dowels per joint: n , = N, l R, for the joint of the internal member to the steel plate. For joints between the steel plate and the external member (chord) the resultant of the point load (in the node) and the change of force in the chord should be taken into consideration. Furthermore, the angle between the resultant force and grain direction of the chord will not be zero. Calculation of the deflection at the serviceability limit: The influence of the slip occurring in the truss joints can be assessed by using an effective cross-section STEPJEUROFORTECH - an initiative under the EU Comett Programme

with

K,,, = 14300 N/mm per dowel is the cross-section of member i A, is the length of member i li are the numbers of dowels at the member ends nu,i3 ne,i E ,,

= 13300 N./mm2

u~,,,~, = 24,6 mm = l / 926 u ~ = ,14,8~ mm ~ ~ ~ U, = 40 mm (chosen)

< 1 I300

+ 0) = 24,6 mm = 14,8 (1 + 0,6) = 23,7 mm uIdin unetdi, = 23,7 + 24,6 - 40 = 8,3 mm = l / 2711 < l / 200 u~~~ = 24,6 (1

General analysis First order linear analysis of the truss as a framed structure can be applied by using a computer program for plane frames.

Z

Figure 7

System.

Initial analysis of internal forces and moments for the ultimate limit states design with: bottom chord continuous along the total length; top chord continuous up to the ridge; ultimate limit states design for load case g + S; = 10600 N/mm2. modulus of elasticity of the member is Final analysis of internal forces and moments for the ultimate limit states design with: conditions according to analysis above; the effective cross-section A* based on K, = 2 K,ye, / 3; K,, = 14300 N/mm;

-

displacements in node 5, 7, '19, 20 due to change of force in the chord are disregarded.

The results of the analyses are shown in Table 3. The normal forces are independent of joint stiffnesses unllke bending moments of the chords.


Mem- Internal ber No force1 moment

Simplified analysis

General analysis disregarding the influence of slip

General analysis including the influence of slip

Comparison of internal forces and moments in the load case g

Table 3

+ S.

Initial analysis of serviceability limit states design with: bottom chord continuous along the total length; top chord continuous up to the ridge; internal members pinned; load cases g and S; E,,g,,e,, = 13300 ~/mrn'.

Deformation of the system in the load case s (initial analysis).

Figure 8

Final analysis of serviceability limit states design with: conditions according analysis above; the effective cross-section A* based on = 14300 N/mm; A,*

-

=

Ai / ( l + Eo>b'3meun li

(-+- 1 nu,i

',ser

1 ne,i

>> ;

K.ser

displacements in node 5, 7, 19, 20 due to change of force in the chord are disregarded.

Figure 9

Deformation of the system in the load case s (final analysis).


Results of the initial analysis (disregarding the influence of slip occurring in the joints):

Results of the final analysis (showing the influence of slip): U,,,, = 24,l mm < 1 /300 u ~ = 14,5 , ~mm~ ~ ~ = 24,l (1 + 0) = 24,l mm U,, U,$, = 14,5 (1 + 0,6) = 23,2 mm U, = 40 m m (chosen) unttfin= 23,2 + 24,l - 40 = 7,3 mm = l / 3082 < l / 200 An example of the truss joints is given in Figure 10.

Figure I0

Node 20.

Simplified analysis including slip

Table 4

General analysis disregarding slip

General analysis including slip

Comparison of the deflections of the bottom chord midpoint. Deflection values in mm; the chosen precamber value U , is 40 mm.

Concluding summary concerning the glulam truss example -

The computation of the internal truss forces based on the simplified analysis leads to reliable results (see Table 3).

-

The internal moments in the loaded top chord depend significantly on the displacements of the nodes.

-

For the calculation of the deflection slip should be taken into account (see Table 4). Otherwise it is expedient to halve the deflection limits recommended in EC5.


Scaffolding and falsework STEP lecture E8 P. Aune University of Trondheim The Norwegian Institute of Technology

Objectives To describe scaffolding systems and their load bearing elements, and to explain the main factors which have an influence on the safety and stability of the C~n~tr~~ti~n.

Summary The lecture gives a short introduction to the subject by mentioning the purpose, history and development of such temporary structures. Requirements, structure systems, loads and particular details in scaffolding are briefly covered. A recently developed nailplate based falsework is described. The combined use of steel and timber in a heavy structure by a bridge construction completes the lecture.

Introduction In many cases a supporting framework of scaffolding is needed during the construction (or dismantling) of a structure. For workers, platforms may be used during the construction, repair, painting or cleaning of buildings. Even though scaffolding and falsework are temporary structures they have to be designed and treated seriously and according to the same principles as if they were permanent structures. Collapse involves risks for human life or serious injuries and also has economic consequences. Many countries have regulations or codes, in particular, to ensure the safety of human life. In Asia bamboo is widely used for scaffolding. In Europe scaffolding and falsework in former days with few exceptions were made of timber and constructed on the site. Some of the old structures were huge and impressive and often based on complex structural systems. It required certainly good design, material knowledge and workmanship of high quality. Figure 1 shows the falsework system by the Svinesund road bridge, an arch bridge at the border between Sweden and the south-eastem part of Norway (built 1938-42).

Figure l

Falsework at Svinesund bridge (Aune, 1994).


In recent years, however, a great number of flexible elements and systems have been marketed. The elements may be made of steel (often steel tubes) and aluminium profiles, light weight, easy adjustments, and simple connections. The fact that such systems may be used over and over again represents of course great advantages. Even so a lot of scaffolding and falsework is still made of timber. EC5 has no specific rules covering such constructions, and this lecture deals with general principles in design and refers to the national specifications and practice in some European countries.

Scaffolding Norway is one of the countries where there are regulations concerning scaffolding. The regulations are worked out by the National Labour Commission and are quite detailed. They have to be regarded as any specification. The main purpose which is strongly emphasised, is consideration of the safety of workers and people in general. This implies requirements for particular details, as well as for a sufficient load bearing capacity of the structure as a whole. The timber used should be of a structural grade and unpainted. If second hand the material must be undamaged. Similar to the assumptions in EC5 regarding the qualifications of the designer and other persons involved, scaffolding work should be carried out by skilled people. It is the employer's responsibility to provide a final control/inspection before the scaffolding is used. It is important to provide signs giving details of the owner, the designer and the design loads.

Loads Experience over recent decades has provided the background for the present classification of scaffolding according to load values in six classes. The given uniformly distributed loads over the whole platform area vary between 0,75 and 6,00 kN/m2 (CEN, 1988). In addition there should be two concentrated loads, one distributed over a certain area and the other the weight of one person. The position of these loads shall be chosen to represent the most unfavourable conditions. Also load actions due to wind forces according to the National load specifications, have to be included.

Load bearing system and bracings In Figure 2 a typical scaffolding system is shown (CEN, 1988). Main components are identified by numbers in the caption of Figure 2. The loads carried by the platform may be transferred to the ledger and further to the transom and finally to the standard, see Figure 3a. By the system shown in Figure 3b the loads are transferred directly to the transom and then to the standard. The joints must be designed accordingly. The purpose of the plane brace is to keep the system in a rectangular shape in the horizontal plane. The longitudinal and the lateral braces provide stability in the respective vertical planes and directions. The tie members fix the scaffolding to the wall.

Detailing The guard rails shall be designed to resist a required concentrated load (vertical or horizontal) at the ultimate limit state situation. There is a similar requirement (smaller load) to meet the servfceability situation.


For some scaffolding a toeboard might be required at the platform. The standards should preferably be in one length, but if necessary they might be jointed by butt joints and long gusset plates along two or four sides. Access to the platforms should be safe, comfortable and suitable preferably a stairway (with a guardrail). For long stairways a landing is required.

Figure 2

Scaffolding structure (Identification according to CEN, 1988). ( 1 ) Platform, ( 2 ) guardrail, (3) transom, (4) ledger, (5) standard, (6) plane brace, (7) longitudinal brace, (8) lateral brace, (9) tie member, (10) toe boards, (11) bridging ledger, (12) base plate, (13) wall support.

Figure 3

Different load-bearing systems (Arbeidstilsynet, 1989).

Falsework Despite the wide-spread availability of different flexible falsework systems, the use of timber may be beneficial in some cases. Figure 4 gives an example of a rather heavy structure used for bridge construction (Holzbau-Taschenbuch, Band 1, 1986). There are longitudinal steel beams (or lattice girders), the towers, however, are timber structures. It should be noticed that controlled lowering of heavy falsework is definitely necessary. This structure has to be designed according to accepted principles with particular emphasis on buckling, lateral torsional buckling and overall stability. In many details stresses perpendicular to the grain may be critical.


Figure 4

Heavy falsework used for bridge construction. (a) Steel beam, ( b ) lattice girder, (c) timber tower, (d) spindle for lowering.

Figure 5 shows a section of the falsework used in the construction of parts of the Condeep platform, Troll, in the North Sea. The bottom part of the platform consists of several cylindrical cells, each covered by a spherical dome with a diameter of 32 m. The structure is formed by 144 curved nailplate girders arranged radially and supported by a compression ring at the center of the dome. The complete structure was prefabricated and lifted into position by a number of cranes. The handling caused an entirely different load situation compared to the weight of the concrete (of a thickness for the regular cells varying between 0,5 and 0,7 m, and up to 4,5 m in special cases). This rather complex and newly developed nailplate structure was designed and detailed using computer programs (Ringstad, 1993).

Figure 5

Section of a nailplate based domelike falsework.

Concluding remarks -

Even when scaffolding and falsework structures are temvorarv constructions, they should be designed according to the principles given in EC5.

-

Accidents related to temporary structures are often caused by disregarding the overall stability and the necessity of horizontal and vertical bracings.

-

Both safe overall erection and disassembly procedures as well as relevant details are essential to ensure safety on site.


References Arbeidstilsynet (1989), Trestillas - timber scaffolding (in Norwegian) Oslo, Norway Aune, P (1994), Timber Structures. Examples (in Norwegian) Tapir Publisher, Trondheim. CEN (1988), Service and working scaffolds made of prefabricated elements. Materials dimensions design loads and safety requirements. Harmonization Dokument HD1000, Brussels. Holzbau - Taschenbuch Band 1 (1986). 8. Auflage (in German), Verlag Ernst und Sohn, Berlin, p. 537. Ringstad, H. (1993), Trekonstruksjoner som forskaling p i TROLL-plattformen (in Norwegian), Byggeindustrien no. 1, 1993 p. 60 - 61.


Plane frames and arches STEP lecture E9 L. Mortensen Aalborg University

Objective To give an introduction to the design and use of plane frames and arches.

Summary The lecture gives an introduction to the field of application of frames and arches and to some aspects of their design. The influence of the structural system and of the geometry of the structure on the moment distribution will be discussed. An example demonstrating the design of a three-hinged curved frame concludes the lecture.

Introduction Plane frames and arches are widely used structures in timber engineering. They are used for sports halls, assembly halls, hangars, churches, halls in industry and farming and also for domestic housing. Besides covering a wide range of applications the frames and arches also cover a wide range of spans, i.e. from a few metres up to about 50 metres for frames and more than 100 metres for arches. The structures will normally have solid cross-sections but especially in the case of greater spans trussed structures may be chosen. The structures may be fabricated from glued laminated timber or LVL or wood based material combined with glued laminated timber, LVL or solid timber in box and Isections.

Structural systems Frames and arches are normally designed as two-hinged or three-hinged structures. Structures with fixed supports are very seldom used because they are more difficult to assemble and because they transmit moments to the foundations. The most commonly used design is the statically determinate three-hinged structure with hinges at apex and supports. Compared with the two-hinged structure it may require more timber because of a less efficient distribution of the internal moments and, furthermore, it will be less rigid. But these disadvantages of the three-hinged structure may easily be eclipsed by the fact that the distribution of internal forces in the static determinate structure is independent of possible displacements of the supports and deformations due to possible changes in the moisture content. Furthermore, the connections at the joints will be simpler and less costly since they are not subjected to bending moments.

Frame designs Examples of glued laminated timber frames are shown in Figure 1. The frames may either be curved, Figures l a and Ib, or they may be designed with sharp knees, Figures l c and Id. The three-hinged curved glued laminated timber frame, Figures l a and b, is a very common solution. This is not only because it is a very economical structure but also because of its aesthetic qualities which contribute to a pleasant and often graceful interior. The roofing may follow the curved shape of the structure but most often a post and rafter assembly will be employed as indicated In Figure la. STEPIEUROFORTECH

- an initiative


E911

Figure 1

Examples of glued laminated timber frames.

Frames with inclined legs as indicated in Figure l b will often be advantageous, e.g. in silo buildings. By fitting the inclination of the legs to the angle of friction of the material to be stored the pressure on the walls will be reduced and the structure will therefore be more efficient. The inclined leg will reduce the transport height of the frame. Curved frames are normally manufactured with an inner radius of 3 to 5 m at the curved knee. The bending of the laminates will induce bending stresses and in order to limit these stresses the laminate thickness shall be chosen in accordance with prEN 386 "Glued laminated timber. Performance requirements and minimum production requirements". Since the maximum bending moment will normally occur at the knee the largest cross-section is required here, and the rafter is therefore tapered from the curved knee to the apex, and often the straight part of the leg or part of it is tapered towards the base. The span of the curved frame ranges from approximately 10 m to approximately 50 m but larger spans are possible. In frames with sharp knees the knee joint may typically be established by employing one or two large finger joints, see Figure Ic. The introduction of a knee segment and two finger joints at the knee reduces the angle between the fibres of the joining pieces of timber, and this will increase the strength when compared with the joint with only one finger joint. Since the finger joint is located in the section with maximum internal moment the performance of the structure is very dependent on the quality of the joint. In order to ensure a reliable and durable finger joint the production shall follow prEN387 "Glued laminated timber - Production requirements for large finger joints". As an alternative to the finger joint the joint may be produced by gluing plywood gussets on both sides of the members or by employing steel or plywood gusset plates with mechanical fasteners. If either the leg or the rafter is chosen as a double member the gussets are omitted, Figure Id. Designs with glued-in bolts, steel plates or steel rods represent the latest developments for this joint.

STEPIEUROFORTECH - ar? initiative under the EU Cornett Programme

The span of the frame with sharp corners ranges from approximately 10 m to approximately 35 m but larger spans are possible. When deciding on the geometry of the frame it is important to realize that the dimensions and thereby the economy are effected by the magnitude of the maximum bending moment in the structure which develops at the knee. The closer the centre line of the structure follows the thrust line of the load the smaller the moment that will develop at the knee resulting in a more economical structure. The extent to which the geometry of the frame can be fitted to follow the thrust line will depend on the functional requirements for the building. It is clear that the introduction of the curved frame improves the possibility of fitting the geometry to the thrust line and thereby reduces the maximum bending moment. If the leg of the frame is inclined it will further reduce the maximum bending moment because of the reduction in eccentricity e, with respect to the knee as shown in Figure 2. Curved frames are therefore better suited for wide spans than the frames with sharp knees where larger moments will develop and therefore result in larger cross-sections and the consumption of more timber. For both structures the bending moment at the knee will be reduced if the eaves height is reduced ( h in Figure 2).

Figure 2

Axis of structures and thrust lines for uniformly distributed load. ( a ) frame axis, ( b ) thrust line.

Arch designs Arches are more appropriate for larger spans than frames because the internal moments are relatively small compared to the moment that would develop in frames. This is because the geometry is fitted more closely to the thrust line of the applied load. Ideally the geometry should follow the thrust line exactly in order to avoid internal moments but since different load combinations will produce different thrust lines it will not be possible to avoid internal moments altogether. Normally the parabolic or circular shape will give good approximations and is therefore often chosen.

Figure 3

Three-hinged arch.


The arches are normally connected to the foundation via hinges. At the apex a hinged connection is also most often introduced and a statically determinate three-hinged structure is formed (see Figure 3). Alternatively, it may be chosen to establish a more rigid two-hinged arch by making all the joints between adjacent arch pieces moment resistant. The height of arches ( h in Figure 3) is normally chosen in the interval 0,13 to 0,20 times the span but greater heights are used when required. The arches are often built up of a constant rectangular, glued laminated timber cross-section but other materials may be chosen, and it may also be decided to vary the cross-sectional depth for structural or architectural purposes.

Design of frames and arches The design procedure is essentially by trial-and-error. Normally the geometry of the structure is decided on the basis of functional and architectural considerations. The cross-sectional dimensions are then estimated and the design will show if they are adequate. A revision of the originally estimated values may be necessary.

Design example. Preliminary design of a three-hinged curved frame Figure 4 shows a three-hinged frame.

Figure 4

Symmetrical three-hinged frame.

The following dimensions are given: l = 25 m, h = 4 m, h,,,= 7,2 m, a = 16", rin = 3 m, lamination thickness t = 22 mm, width of cross-section b = 160 mm, inclination of leg a,,, = 14". The cross-sectional dimensions required will depend on the specific loading conditions, frame geometry and frame spacing and also on the strength class of the glued laminated timber. The guidelines for estimating cross-sectional depths given below should therefore be considered as rough estimates. depth of curved section h, h, depth at supports h, depth at apex

- 0,03 1, where I is the span - 0,5 h,

- 0,9 h,


In this example the following cross-sectional depths are chosen: curved section supports apex

h, = 748 mm (34 laminations of 22 mm) h, = 500 mm h, = 300 mm

The frame geometry is hereby defined. In this example only dead load and snow load are considered but it should be noted that this is only one of many cases that must be checked. It is assumed that the design load from the combination of dead load and snow load produces a line load of q, = 9,00 kN/m on the left half and q, = 10,13 kN/m on the right half of the frame. This is an approximation with regard to the loading in the corner because the rafter assembly will distribute the load here as point loads to the post and the rafter. The load duration class is short term since the snow load is short term. The structure is assumed to be assigned to service class 1. The strength class GL36 is chosen according to prEN 1194:1993 "Timber structures - Glued laminated timber - Strength classes and determination of characteristic values".

EC5: Part 1-1: 2.2.3.2, 2.3.3.2 and 3.1.7

The design strength properties are found as: ,,R, 'l =24,9N/mm2 h , , , ,=2175N/mm2 h,,, = 2,40 ~ / m r n ~

fn

f,,90,,d

= 4,36 ~/rnrn'

In the following three cross-sections will be examined. First section D at the frame corner (see Figure 4) where the maximum bending moment occurs. Then section F (see Figure 4) where the maximum bending stress in the straight part of the rafter occurs. Both cross-sections are examined for the combined action from the bending moment and the axial compressive force acting at these sections. Finally, section A at the support (see Figure 4) is examined for the shear force acting here.

Calculation of internal forces EC5: Part 1-1: 5.4.4

EC5: Part 1-1: 5.2.1

The geometrical and structural imperfections and induced deflections shall be taken into account when determining the internal forces. This may be done by carrying out a second order analysis on an imperfectly shaped structure. In this example the influence of imperfections and deflections will be taken into account by employing the column equations. As a first step the internal forces are therefore found on the basis of the perfect and undeformed structure. They are found from hand calculations or more conveniently obtained from computer programs.

Frame corner In this example the maximum internal moment occurs at the left-hand corner. (Section D in Figure 4). At section D the Internal forces are found as:


As mentioned above the increase of the moment due to imperfections and deflections will be taken into account by considering the frame corner as a column with the effective length lef according to STEP lecture B7, Equation (16).

h S

I I, K,

= 3830 mm (length of column between A and D) = 11820 mm (length of rafter between D and C) = 160 7483/12 = 5,58 109 mm4 = 160 6423/12 = 3 3 3 109mm4

.

-

.

m

With respect to the lateral stability of the frame it is assumed that it is supported laterally at the support and by purlins spaced at 2 m. The purlins closest to the frame corner are assumed to be located where the assembly rafter is connected to the rafter at (X, y) = (9,66, 4,43) and the effective length with respect to deflection out of the frame plane is therefore calculated as l,,., = 5300 mm (see Figure 4). The stresses at the corner section, D, of the undeformed structure are calculated as: o,,, = 1,23 N/mm2 and o,, = 1,11 14,6 N/mm2. EC5: part 1-1: 5.2.4

Because of the curvature the bending stress is multiplied by the factor k, = 1,11.

EC5: part 1-1: 5.2.4

Because of the strength reduction caused by bending of the laminates the bending strength is reduced by the factor k, = 0,90 (see the column equations below).

EC5: part 1-1: 5.2.1

The strength of the frame corner is sufficient if the stresses satisfy the column equations.

k , , and k , , to be employed in the equations are found as:

By inserting in the column equations the following is found:

The conditions are satisfied.

Rafter Similarly to the frame corner the rafter is examined for the combined action from the axial compressive force and the bending moment. The rafter on the STEP/EUROFORTECH - an initiative under the EU Cornett Programme

right-hand half of the frame will be subjected to the heaviest load and will therefore be considered. The rafter is considered as a single tapered beam (a = 2,6"). The section of the rafter where the maximum bending stress occurs will be examined. Because of the taper of the beam the maximum bending stress will not occur where the bending moment is maximum but at the section located at a distance X, from the apex where

(Larsen and Riberholt, 1994). is the x-coordinate to the section of zero moment in the rafter. It is found as = 6,57 m. The corresponding cross-sectional depth is found as h, = 605 mm and the cross-sectional depth at the apex is h, = 300 mm.

X ,

X,

Therefore

The corresponding cross-sectional depth is h, = 401 mm. The internal moment at F is calculated as M, = 48,5 kNm. The bending stresses in the outmost fibres at this section are calculated on bending stress as G, = 11,4 N/mm2 (10,l) and G,,~,, = 11,2 N/mm2 (9,93). The figures indicated in brackets are the stresses at the section with maximum bending moment (X= 3,29 m).

,

EC5: Part 1-1: 5.2.3

The axial compressive force at F is found as N, = 105 kN and the corresponding stress is calculated as G,,~,,, = 1,64 N/mm2.

EC5: Part 1-1: 5.2.3

Due to the combined action from the axial force and the moment the stress at the tapered edge will exceed that at the bottom edge and the tapered edge will therefore be considered. At the tapered edge the bending strength shall be reduced. The reduction will, however, be negligible in the case of compressive stresses parallel to the tapered edge. It is found that f,,,, = 24,7 ~ / m m ~ . The strength of the rafter is sufficient if the stresses satisfy the column equations.

EC5: Part 1-1: 5.2.1

k , , and k , , to be employed in the equations are found as: k , , = 0,46 and k , , = 0,97.

When calculating k , , the depth of the rafter cross-section was taken as h = 642 mm which is the depth at 0,65 s (see Figure 4) and the effective length of the rafter was calculated according to Equation (17). lef was found as l,, = 16000 mm. The column equations may now be employed:


And it is seen that the conditions are satisfied. Having considered the frame corner and the straight part of the rafter it still remains to consider the sections at support and apex. Both cross-sections are subjected to an axial compressive force and a shear force. Here only the investigation at the support will be demonstrated.

Support It is assumed that the leg has been cut vertically and the cross-section at the support thereby reduced to 160 X 500 mm2. The axial force (113 kN at A and 119 kN at B) can easily be absorbed. The shear force is greatest at A. Because of the inclination of 14" of the leg it is found that V , = 72,7 kN, and therefore

Conclusion As mentioned earlier the design is a trial-and-error procedure and the results of the first trial above indicate that it will be possible to reduce the estimated crosssectional dimensions of the frame and still satisfy the strength requirements. A second trial based on reduced cross-sectional dimensions will not be demonstrated here. It should be remembered that requirements for the frame stiffness in the serviceability limit state might be determining for the cross-sectional dimensions.

References Larsen, H.J. and Riberholt, H. (1994). Trxkonstruktioner. Beregning. SBI-anvisning 135. 3. edition. Statens Byggeforskningsinstitut, 1994. ISBN 87-563-0849-3. Riberholt, H. (1991). Trakonstruktioner. Eksempler. Afdelingen for Bzrende konstruktioner. Danmarks Tekniske Hfijskole. ISBN 87-7740-057-7. Riberholt, H. (1991). Trakonstruktioner. Udforming. Afdelingen for Barende konstruktioner. Danmarks Tekniske Hfijskole. ISBN 87-7740-090-9.


Timber frame houses - Structural Objective

STEP lecture E10 B. Roald

To describe the structural behaviour of timber frame house construction.

The Norwegian Institute of wood Technology

Summary This lecture describes, in principle, how timber frame houses are constructed and designed. It is based mainly on the Scandinavian tradition. It contains a brief description of the main calculation methods concerning timber frame walls, floors and roofs, and also a brief description of other materials which form part of the timber frame house.

Introduction Timber frame housing is the traditional type of construction in North America and Scandinavia, where more than 90% of all homes are built this way. Timber frame housing has also increased its market shares in a number of other European countries, particularly in the United Kingdom. Fast and dry erection, good thermal insulation and suitability for all types of architectural designs are important factors for making wood frame construction a strong alternative in house building.

Construction principles The most commonly used method of building timber frame houses is the platform method shown in Figure 1. Platform construction means that the walls are placed directly on a sub-floor which acts as a working platform during the erection of the house. Originally an American method, platform construction is now becoming the normal way of building in most countries, because it combines better safety during erection with the possibility of using either prefabricated panels or on-site assembly.

Figure I

Platform construction (NB1 Handbook 38 Timber Frame Houses). ( a ) Stud, (b) bottom rail, ( c ) perimeter beam, ( d ) sole plate, (e) subfloor, @floor joist.


E1011

Timber frame construction uses timber members and components to form a structural frame which transmits all vertical and horizontal loads to the foundations. The exterior cladding is non load bearing; it is used to weatherproof the building and provide the desired external appearance. Conventional timber frame walls are constructed of vertical timber studs, breather membrane and sheating on the outside, vapour barrier and lining on the inside and with the cavity filled with mineral wool. Timber joists make the floor. Roof trusses or rafters are spaced at 0,6 m or 1,2 m, depending on snow loads andlor tradition. Roof structures correspond with the position of wall studs in well designed houses, in order to avoid unnecessarily large head sills and the possibility of visible deformations.

Walls Load bearing walls

-

structural requirements

Load bearing walls support roofs andtor floors. The studs are designed as columns (STEP lecture B6) with buckling length the same as the length of the stud. One assumes that the studs are prevented from buckling in the wall plane by the sheathing. The slenderness and buckling capacity are only calculated for buckling perpendicular to the wall plane.

Figure 2

Timber frarne wall (NB1 Building Details A 523.251). ( a ) Lintel, ( b ) top rail, (c) stud, ( d ) nogging, (e) bottom rail.

Load bearing walls in low-rise houses in Scandinavia are designed for vertical loads only, and not in combination with horizontal wind load. The argument for this is that the probability for the occurrence of both wind and snow-load at their maximum values at the same time is very small. Also, the wall sheathing gives extra safety against buckling perpendicular to the wall plane. This is not considered when designing the studs as simple columns. The wind load on external walls of normal height is never decisive for the stud size in a timber frame house. For houses with especially high walls the studs must be checked for wind load too, that is combined compression and bending (STEP lecture B6). In addition to buckling, compression perpendicular to grain on the bottom rail must be checked. Lintels over openings in load bearing walls are designed as simply supported STEPIEUROFORTECH - an initiative under the EU Comett Programme

beams across one span. In reality, the top rail, the wall sheating and the roof construction's load distributing ability, all contribute to increased load capacity. As long as the spacing of the roof rafters is no more than 0,6 m, the lintels may be calculated with a uniformly distributed load. This is much simpler than designing with point loads from each rafter because the exact position of the rafters does not need to be known.

In addition to the bending moment, compression perpendicular to grain at the support of the lintel must be checked. Calculation of deflection may be omitted, as the top rail, sheathing etc. provide sufficient stiffness.

External walls External walls are constructed from vertical studs, normally at 400 mm or 600 mm spacing, nailed with simple butt joints to top and bottom rails. Strength graded timber must be used. In Northern Europe, it is the requirement of thermal insulation which in practice governs the size of the studs, along with the need for satisfactory butt-jointing of sheathing and plasterboard, not the structural design and needs for load bearing capacity. The normal sizes are 48 mm X 98 mm or 36 mm X 148 mm. Tables are given for load bearing capacity and stud sizes (NB1 Building Details A 523.251, TRADA (1989)). Where openings occur in load bearing walls, they must be spanned by suitably designed lintels, and the load on the lintels should be transmitted to the foundations by cripple studs. The cripple studs should be of the same crosssection and at least the same number as those removed from the opening. Wind bracing is usually provided by a wood based sheet material or plasterboard, normally nailed to the external face of the frame. All four walls of a house must have at least one layer of sheet material nailed along all four sides of the sheet and along each stud.

Figure 3

Wind bracing with sheet material (NB1 Building Details A523.251). ( a ) Panel sheathing, ( 6 ) nail spacing approx. c/c 100 mm along the edges of the panels, (c) nail spacing approximately c/c 200 mm along the middle of the panels.

Internal walls Internal load bearing and non load bearing walls may be constructed simply by using a stud frame lined on both sides with a sheathing material. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E 1013

Separating walls Separating walls are commonly constructed from two separate stud frame walls with a cavity between. Depending on the requirements for fire resistance and sound insulation, the cavity is filled with insulating material and one or more layers of adequate lining is applied.

Figure 4

Separating wall (NBI Building Details A 525.301). (a) studs 48 mm x 73 mm, ( b ) mineral wool, ( c ) lining.

Floors Floors are made of joists or I-beams with spacing 400 mm or 600 mm, with or without thermal insulation. On top flooring is applied and underneath ceiling lining.

Figure 5

Floor construction (NB1 Building Details A 522.351).

Intermediate floors shall be calculated with an imposed floor load of 1,5 k ~ / m ~ for one-family dwellings. In normal dwellings, it is not the need for load bearing capacity which is decisive for the design of wood joist floors, but the need for stiffness of the floor. A floor designed for a uniformly distributed load can give floor joists with such a span that deflection and vibrations will be unacceptable for normal use. It is very complicated to decide what the acceptable stiffness for a timber joist floor is, as the acceptable level of vibrations and deflections is very individual. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Vibration may be experienced as jingling in cupboards, light furniture which moves and the feeling of vibrations by the inhabitants themselves. The Norwegian Building Research Institute (NBI) uses a maximum deflection of 0,9 mm under a point load of 1,O kN as a design criteria for timber joist floors. The deflection depends on the flooring and ceiling linings load distributing and stiffening ability. An accurate calculation is complicated and special data programs have been made to calculate joist span tables (NB1 1990), table 1 and 2. Due to the consideration of vibrations, the capacity of thin-webbed I-beams and box beams can not be fully utilised. The point load criteria should not be used uncritically especially for long spans (> approximately 5 m). In the serviceability limit states part of EC5 an investigation method for the vibration of residential floors is presented, see STEP lecture E l l .

Roofs Pitched roofs may be constructed using prefabricated or site-constructed trussed rafters or with traditional rafters and purlins.

Figure 6

Roof constructions (NB1 Handbook 38 Timber Frame Houses). ( a ) Trussed rafters, ( b ) rafter rooj ( c ) purlin roof.

Accurate calculation of trussed rafters is quite complicated and is usually done with specialist computer programs (see also STEP lecture B12). The manufacturers of prefabricated trussed rafters design these by the use of data programs and deliver them according to the purchasers specifications of load, roof pitch, span etc. For site-constructed trussed rafters there exist design sheets (NB1 Building Details, TRADA Design Sheets etc.).

EC5: Part 1-1: 5.4.6

Traditional rafters with supports at each end are designed as simply supported beams across one span. When calculating the bending moment, any overhang should be ignored. The load bearing capacity of the overhang should be designed separately as a cantilever beam. The load sharing factor k,, may be used for rafters and purlins with a spacing no more than 0,6 m. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E 1015

When designing purlins, the fact that the load is applied at a slope to the crosssection must be taken into account. If the roof has panel sheathing providing diaphragm action, the load component perpendicular to the purlin height may be reduced slightly. For steep roofs (pitches greater than about 40") the combination of snow load and wind load may give the ultimate design load. For low rise timber frame houses, it is assumed that sufficient safety is obtained when the load bearing capacity for snow load and wind load is checked separately. Design sheets with tables for rafters and purlins exist (NB1 Building Details, TRADA Design Sheets etc.).

0ther components

Figure 7

Timber frame wall, floor and roof - components (NB1 Building Details A 523.255, A 522.355 and A 525.100). Wall: (a) Interfial lining, (b) vapor barrier, ( c ) furring strip, ( d ) mineral wool 100 mm (e) mineral wool 50 mm, (f) wind barrier, (g) batten, ( h ) ventilated space, (i) external cladding. Floor: (a) Subfloor, (b) thermal insulation, ( c ) wind barrier, ( d ) ceiling, floor joist, (e) perimeter beam, (f) floor joist, ( g ) flooring. Roof: (a) Roofing, (b) sheathing, (c) ventilation space, (d) wind barrier, (e) insulation, (fl vapour barrier, (g) ceiling lining.


Cladding External cladding can be chosen from a wide range of materials, including brickwork, tiling, timber boarding or cement render. External cladding in Scandinavia usually consists of timber boarding, either vertical or horizontal. This is supported on counter or horizontal battens fixed to the studwork to provide a drained and ventilated cavity. This is to ensure that any water or vapour that has penetrated the cladding can evaporate or be drained away.

Wind barrier The wind barrier must be strong enough to resist site and wind damage, be waterproof but allow the escape of any water vapour. Building paper or sheet materials are used as an external wind barrier. The material must to some extent be able to withstand the effect of moisture - also in the form of water. The most common products are:

- Asphalt impregnated breather paper, - 12 mm bitumen-impregnated porous fibreboards with windproof coating, - 9 mm plaster board with impregnated cardboard,

Thermal insulation Thermal insulation requirements of the building regulations are achieved in timber frame construction by placing insulation materials in the walls, ceilings and floor cavities. Most countries have experienced a strengthening of building code requirements concerning thermal insulation performance, and it may often be economical to insulate better than the minimum code requirements. Insulation thickness in walls is now 100 mm minimum as a standard for wood frame housing in most countries, while 150 mm and more is the norm in northern areas. Mineral wool is the most commonly used insulation material, and used in the form of bats or rolls. Standard thicknesses are 50, 75, 100, 150 and 200 mm.

Vapour barrier The importance of an airtight vapour barrier increases with better insulated low energy houses. Polyethylene sheeting has proved to be the most versatile material for vapour barriers, because it can overlap the critical joints between building components and is easy to install. Polyethylene sheets are often exposed to damage during construction, and thicker qualities (0,15-0,20 mm) are now commonly used. The vapour barrier is placed on the warm side of the insulated wall and ceiling, directly behind the internal lining, to minimize the passage of air with high vapour content into the structure. Air with a high vapour content will condense when it is cooled inside the structure.

Lining Internal lining materials may be wood panel, plywood, fibre board, particle board or gypsum plaster board.


References NB1 Handbook 38 Trehus (1987). The Norwegian Building Research Institute, Oslo Norway. NB1 (1988). Building Details A522.355 Trebjelkelag. Varmeisolasjon og tettesjikt. The Norwegian Building Research Institute, Oslo, Norway. NB1 (1987). Building Details A523.251 Bindingsverk av tre. Dimensjonering og utf0relse for smbhus. The Norwegian Building Research Institute, Oslo, Norway. NB1 (1981). Building Details A523.255 Timber Frame walls. Thermal insulation, wind and vapour barriers. The Norwegian Building Research Institute, Oslo, Norway. NB1 (1993). Building Details A524.301 Skillevegger av tre mellom boliger i rekkehus. The Norwegian Building Research Institute, Oslo, Norway. NB1 (1987). Building Details A525.100 Tretak. Varmeisolering, tetting og lufting. The Norwegian Building Research Institute, Oslo, Norway. TRADA (1989). Timber Frame Housing Structural Recommendations TBL 52. Timber Research and Development Association, High Wycombe, England.

STEPIEUROFORTECH


Timber frame houses - Serviceability STEP lecture El l S. Thelandersson Lund University

Objective To describe the principles of design and detailing of timber frame houses with particular reference to serviceability requirements.

Summary This lecture deals with the design and detailing of timber frame houses with particular reference to serviceability requirements. Important structural serviceability aspects which have to be considered in timber frame houses are vibrations in floors and load and moisture induced movements in timber and wood based materials. Non-structural aspects of design, such as sound and heat insulation, the provisions of installations and protection of timber against excessive moisture exposure during construction and use, are discussed. The necessity of an integrated approach in design, where structural and non-structural problems are considered simultaneously to achieve rational and economic solutions, is pointed out.

Introduction Timber frame building systems consist basically of wall and floor units composed of timber studs or joists, covered with different types of sheathing materials and with insulation material in between. For floor units, light-weight elements may be used as an alternative to solid timber joists as shown in Figure 1. Such systems are rational since most necessary functions can be integrated in the same structural unit, such as:

-

transfer of vertical loads stabilisation against horizontal loads physical separation

-

-

heat insulation sound insulation fire separation

In contrast, a skeletal structural system is normally used in steel buildings. The only function of the structural skeleton is to carry the loads. All the other functions must be provided by other structural components such as infill walls. The wall and floor units in timber frame building systems can be highly standardised, but still easy to modify if needed. The units can be assembled on site using simple equipment or they can be prefabricated either in a temporary shop at the building site or in a permanent factory. The technology is well known especially in Northern Europe, where a majority of single family houses are built with timber frame. In some parts of the world, timber frame building systems are widely used also in buildings up to four or five storeys. In North America, for instance, timber frame is the dominating system used in multiresidential construction. In the design and construction of timber frame building systems, special care must be taken to ensure good serviceability taking into account the special properties of timber. Structural timber systems have very low weight compared to other types of building systems. This is advantageous with respect to ease of transport, erection and construction, requirements on the foundation and for earthquake resistance. However, light-weight structures are sensitive with respect STEPIEUROFORTECH - an initiative under the EU Comett Programme

E l 111

to floor vibrations, deflections, sound transmission and fire resistance. Another special feature in timber construction is the fact that wood is a hygroscopic material which absorbs and desorbs moisture in response to climate changes in the environment. Changes in moisture content are associated with shrinkage or swelling, and the simultaneous influence of loading will enhance these movements. If wood is permanently exposed to high moisture levels, there is a considerable risk of growth of fungi and rot which may jeopardise the durability and serviceability of the building or its components. It is extremely important that such hazards are prevented by appropriate design of the building envelope and by using dry timber during construction. For the same reason the timber must be protected from extensive exposure to rain and snow during the construction period.

Figure l

Floor system with light-weight I-joists as primary load-bearing elements. (a) Gypsum plasterboard, (b) floor board, (c) light-weight I-joist, (d) insulation, (e) resilient channel.

Many of the key issues in the design of timber frame building systems are related to serviceability, and floor systems especially are in many ways critical for the quality of the building.

Design to control floor vibrations The most common means to improve the vibrational performance of floor structures is to increase the stiffness in the principal load bearing direction. This can be done by increasing the joist dimension, and also by gluing the floor board to the joist, thereby providing a composite action between the joist and the floor board. The most efficient measure, however, is to increase the bending stiffness perpendicular to the principal bearing direction. This may be achieved by using a flooring material with higher stiffness or by strutting between the joists. Two commonly used methods of strutting are solid bridging and herringbone strutting as shown in Figure 2. The effectiveness of these strutting methods is limited, however, due to shrinkage and compression perpendicular to the grain in the joists. The best result is obtained with solid bridging, if the nogging pieces are glued to the board panels at the top and the bottom of the joists. In this way, the transversal load transfer capability is more or less independent of moisture movements in the wood. STEPIEUROFORTECH

- an initiative

under the EU Cornett Programme

Figure 2

Strutting between jibor joists will increase the transverse st~rnessof a floor. ( a ) Solid bridging, ( b ) herringbone strutting.

To avoid disturbing vibrations between different rooms or flats, continuous beams should be avoided. It should be noted that, contrary to popular belief, reducing the joist spacing does not always improve the vibration performance.

Sound performance One of the key issues in light weight timber frame design is acoustics. The concerns over acoustics fall into two categories:

-

Airborne noise (e.g. stereo sounds, voices).

-

Structure borne noise (e.g. footsteps, plumbing, elevators).

By careful design and construction the problem with air borne noise can generally be handled to achieve good sound ratings. As an example a wall separating two flats from each other is shown in Figure 3. This type of wall has a sound rating which is comparable with a 240 mm thick concrete wall. The performance may be further improved if the linings on one side of the wall are placed on resilient channels connected to the studs. It is important that the cavity between the two parts of the wall is not broken by structural connections. By similar measures, a good performance with regard to air borne noise in floor structures can be achieved.

Figure 3

Wall with high sound perjormance, designed as two individual wall units separated by a cavity. Each wall unit has plasterboard linings on its sides and is filled with insulation between the wall studs. ( a ) Gypsum plasterboard.

The structure borne noises can be more difficult to address, especially in floor structures. To obtain good sound insulation in wooden floors, the top and the bottom part of the floor must be separated from the primary structure as much as possible. At the bottom, a very effective method is to place the ceiling structure on resilient acoustic channels, see Figure 1. The purpose of the resilient channel STEPIEUROFORTECH - an initiative under the EU Comett Programme

E1 113

is to isolate the gypsum board from the supporting timber structure. Therefore, it is important to ensure that the screws which fasten the gypsum board to the resilient channels do not touch the timber floor joists. It is also advantageous to place insulation material between the beams. From an acoustic point of view, there is no significant difference between solid timber beams and light-weight composite beams, as those shown in Figure l . At the top the floor sheathing may be placed on a thin insulation underlay (floating floor) as shown in Figure 4. To improve the footstep sound performance it is advantageous to increase the mass of the flooring. This can be accomplished by placing gypsum plasterboard on top of the structural floor sheathing and using double plasterboard sheathings in the ceiling underneath the floor. An alternative is to apply a thin (30 - 40 mm) concrete or plaster layer on top of the floor boarding instead of gypsum boards.

Figure 4

Floor system with very high impact sound rating. From top: 2x13 mm gypsum board, 20 mm mineral wool, 13 mm gypsum board, 22 mm chip board, 300 mm light weight 1-joists, 145 mm mineral wool, 25 mm resilient acoustic channel, 2x13 mm gypsum board.

The choice of floor covering is one important factor which influences the impact sound level. A thick carpet is more advantageous than a hard floor covering such as wooden parquet or PVC carpets. Many structure borne noise problems can be avoided through careful suite layouts. In multi-residential buildings it is best to line up bathrooms and kitchens from storey to storey. A bathroom or a kitchen should not be placed above a bedroom in the apartment below. Finally, it is extremely critical for a good sound performance to ensure that no plumbing pipes are touching the gypsum board ceiling or the wall linings. It can be concluded, however, that very good sound performance can be achieved with light-weight timber construction. The floor design shown in Figure 4 has been shown to have an impact sound rating comparable to a 290 mm thick concrete floor.

Moisture movements in structural timber systems. Moisture induced deformations in timber framed systems are inevitable due to environmental changes in temperature and relative humidity. The largest STEPIEUROFORTECH

- an initiative


deformations occur perpendicular to the grain, i.e. in the transverse directions of timber members. The difference between the highest and lowest values in moisture content is 2-4 percentage units in indoor climate and 8-10 percentage units in outdoor climate. The most serious situation occurs when the timber is wet during construction. The subsequent drying will then lead to significant shrinkage. If the timber is installed in green condition, it may shrink 3-4% in the transversal direction, when exposed to indoor climate. This shrinkage corresponds to a dimensional change of 6-8 mm over the depth of a normal floor joist. It is therefore very important to use dry timber in construction, especially for multi-storey buildings. The problem with shrinkage is accentuated when a timber member is loaded in compression perpendicular to the grain, e.g. in a wall-floor connection as shown in Figure 5. The transverse compression in the horizontal members cannot lead to failure in the usual sense, but it will lead to large deformations especially when it is combined with shrinkage in the member. The deformations at each floor level can be of the order 5-10 mm. For single family houses up to two storeys the consequences of these movements are less serious, since the load levels are normally lower and the additive effects are limited. For timber buildings with several storeys, however, the loads are higher and the deformations from a number of connections may add up to very large deformation in the upper part of the building. Also for buildings with large horizontal dimensions, moisture movements over long lengths may give rise to large displacements in critical points of the structural system.

Figure 5

Wall-floor joint in typical platform frame construction, Sunley et a1 (1985).

The moisture and load induced displacements in the structural system may affect the serviceability of the building in various ways. Large movements may cause damage to plumbing, electrical and mechanical systems and differential movements between exterior brick veneer and the structure may lead to cracking. Similar problems may arise for other types of exterior cladding with limited deformability. Normally the major part of the deformations develop during the construction period, but those parts of the movements which occur after the building is finished, may cause damage to linings, paint and wallpaper. In particular, differential settlements of the supports leads to slopes and level STEPIEUROFORTECH - an initiative under the EU Comett Programme

E1 115

differences on floors. For instance if a floor is supported on a concrete or a brick wall on one side and on a timber frame wall on the other side significant slopes can be created. For buildings with several storeys a special "shrinkage analysis" of the structural system should be made to avoid such detrimental effects. The settlements in floor-wall connections are most significant in platform framing type systems as shown in Figure 6a. In this case the wall units are constructed in storey heights and the intermediate floor is placed on top of these panels to form a platform and a stabilisation for erection of subsequent wall panels. A way to limit the settlements is to use so called balloon framing, i.e. using wall units two storeys in height with the intermediate floor suspended between the walls, see Figure 6b. In this case the vertical load is transferred through the connection via stresses parallel to the grain, and the settlement due to shrinkage is also minimised in this type of connection. In spite of its disadvantages with regard to settlements, the platform framing technique is almost always used in practice. The reason is that this method is considered to be far more rational and efficient in the construction process. Another way to reduce settlements in timber frame systems is to block the vertical movements in the floor-wall connections by using short vertical pieces of timber which can transfer the vertical loads by stresses parallel to the grain. The disadvantage with this solution is that it is time consuming and it can also be difficult to combine with good solutions to obtain efficient heat insulation in wall-floor connections at the exterior walls.

Figure 6

( a ) Pla@orm frame construction. ( b ) Balloon frame construction.

The building envelope The exterior wall plays a vital part in creating the desired indoor climate in the building. The exterior wall should therefore:

-

provide thermal insulation,

-

prevent air leakage,

-

prevent damage from internal and external moisture exposure.


Thermal insulation in a timber framed wall is easily provided by fixing insulation material between the studs, see Figure 5. If high insulation capacity is required an extra insulation layer may be placed outside the timber frame as shown in Figure 7. In the latter case a special design of the wall-floor connections has to be made to minimise the effect of thermal bridging. To prevent air leakage and to protect the wall from moisture diffusion from the inside, a vapour barrier is normally provided on the inside face of the studs, i.e. on the "warm" face of the insulation. Behind the exterior cladding a breather membrane is provided, which protects the wall against wind and prevents air circulation which can impair the thermal insulation performance of the wall. The exterior surface of the breather membrane should be water repellent to prevent water from penetrating the exterior cladding, which could damage the membrane, the heat insulation or the timber frame. It is important, however, that the breather membrane is permeable enough to allow moisture to ventilate out from the wall construction. To ensure ventilation of moisture behind the exterior cladding, an air cavity should be provided as shown in Figure 7.

Figure 7

Exterior wall and wall-floor connection with high thermal insulation peiformance, Trainformation (1992). (a) Horizontal batten, ( b ) bottom plate, ( c ) breather membrane, ( d ) rim joist, (e) top plate, (fl internal sheathing, (g) vapour barrier, ( h ) floor joist.

Figure 7 shows an exterior wall with a wall-floor connection designed to fulfil high requirements for thermal insulation, which may be relevant in colder climates. In timber framed building systems the exterior walls are normally also load bearing. It is quite evident that an optimal design of such a wall and its detailing is obtained only if all relevant requirements are taken into account simultaneously. Optimising the design from the structural point of view only, would lead to a design which would be inappropriate in view of other functions, such as thermal insulation and moisture protection.

Concluding remarks Special care must be taken in design and construction to ensure good serviceability in timber frame systems. The most important aspects are performance with regard to acoustics, vibration and springiness, and avoidance of adverse effects of moisture during construction and use. Most of these problems can be solved in a satisfactory way with proper design and workmanship. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E1 117

References Sunley J. and Bedding B., Editors (1985). Timber in Construction, BT Batsford Ltd. LondoflRADA, United Kingdom. TrainformationITratek (1992). Manual for Timber Construction (Trabyggnadshandbok, in Swedish), Stockholm, Sweden.


Timber frame houses Fire resistance STEP lecture E12 H. Hart1 Zivilingenieur fiir Bauwesen

Objective To provide information on the fire behaviour of timber frame houses.

Summary A survey upon design and calculation procedures to determine the fire resistance of timber frame houses is given. The fire behaviour of single components made of wood andtor wood-based materials as well as non-combustible panels and insulating layers is discussed in detail.

Introduction The fire resistance of a timber frame house depends on two specific functions. Single components such as beams, columns, walls, floors, stairs and roofs have to fulfil a load-bearing function and (especially walls and floors) are necessary for compartmentation. Fire compartmentation should prevent fire expansion and keep rooms or escape routes free from flames and hot gases. According to these functions the design procedure for the single component is the following: EC5: Part 1-2: 2.1.

Criterion R Mechanical resistance in the event of fire is required. Structures shall be designed and constructed in such a way that they maintain their load-carrying function during fire exposure. Members shall be designed and constructed in such a way that they maintain their separating function during the relevant fire exposure, i.e.

Criterion E No integrity failure due to cracks, holes or other openings large enough to permit fire penetration by hot gases or flame.

Criterion I No insulation failure leading to temperatures of the non-exposed surface exceeding admissible limits. The members shall comply with criteria R, E, and I as follows:

-

R (load carrying): RE (I) (load carrying and separating): EI (separating):

beams, columns, walls. floors and walls. partition walls.

The criteria are time dependent and vary with the use of the structure. The time period is defined in minutes as 15, 20, 30, 45, 60, 90, 120, 180, 240 or 360 minutes.

Behaviour of the entire structure The classification of a single component requires that each connected element fulfils the same criterion as the component. This means that for example a beam STEPJEUROFORTECH - an initiative under the EU Comett Programme

E1211

only fulfils it's load-carrying function for a certain time if all the appropriate single components (stiffening, bracket, bracing ...) do not fail within the required time period. The interplay of the abilities of each element forms the fire resistance of the entire structure.

Figure l

Example entire structure R30 (Kordina and Meyer-Ottens, 1994).

In the following some calculation methods will be presented for single components. In most cases classification according to the rules given will prove conservative and testing of the element is indispensable.

Behaviour of the single component Members The fire resistance of unprotected timber members is discussed in lecture B17. In some cases a protection by cladding might be necessary. But from the aesthetical point of view and from a designers point of view it is better to reach a fire resistance time by an unprotected member. For calculating the fire resistance of a protected timber member the kind of protection gives information about the fact from how many sides charring has to be considered (1-4 sided).

Figure 2

EC5: Part 1-2: 3.6

Member unprotected, totally covered and partly covered.

Charring need not to be considered for surfaces of members covered by fire protective cladding when tPr 2 tLreq where 'l>r

tJXq

is the failure time of protective board or other protective material, i.e. the duration of effective protection against direct fire exposure. is the required fire resistance time under standard fire exposure.

When surfaces of members are covered by fire protective cladding or are aligned with other structural members having a failure time smaller than the required fire resistance time t,, charring of the member starts at the failure time l,,. of the cladding.


Panels, insulating layers Panels are produced from wood andlor wood-based materials with binders. Panels are non combustible under the following conditions:

-

-

use of binders such as gypsum, cement, etc. use of fire retardants.

The main panels are: all veneers only board (AV0 board), particle board, fibreboard, gypsum plasterboard. Due to the fact that some panels have a charring rate lower than solid wood they increase the time of fire resistance of the component. In most cases they are used in combination with walls and floors to satisfy separating functions inside the house (e.g. escape routes, stairways). For panels used as fire protective cladding (Figure 3) a failure time has to be determined. Failure time in this case means that the increase of temperature on the opposite side of the fire exposure has not yet reached the maximum of 180 K for combustible panels and 500 K for non combustible panels (during fire tests at this time black areas could be seen on the opposite side). According to the different behaviour of the materials in the event of fire it is necessary initially to assess the failure times on the basis of tests.

EC5: Part 1-2: Figure 3.1

Figure 3

Example for panels used as fire protective cladding.

The failure time for fire protective cladding of wood and wood-based panels may be found from the equation EC5: Part 1-2: 3.7

where

0, t, tr

is the charring rate according to EC5: Table 3.1. is the thickness of wood or wood-based panel cladding. In the case of two or more layers of board t,, is the sum of thickness of each layer. = 4 min has been introduced to prevent falling-off of panels and/or premature penetration of fire into empty spaces (holes in the panel).

If the cladding consists also of more than one material, the time of fire resistance for the panel is in general the sum of each layer's fire resistance. A fire protective cladding should always be fixed to the member itself and in case of multiple layer cladding not one to the other. Load-carrying panels need not to be analyzed, if their residual thickness is at least 60% of the thickness required for normal temperature design. According to EC5 Part 1-2 insulating layers have a melting point > 1000 "C and are non-combustible. The most common products are wood-wool panels and mineral fibre board. Especially for wall- and floor construction insulating layers are of great importance to reach the separating criterion. It is indispensable that the insulating layer are fixed carefully and tight (compressed).


E 12/3

For non-combustible insulating material with a thickness of more than 20 mm and a density of more than 30 kg/m3 which remain coherent up to 1000 "C failure time may be taken as EC5: Part 1-2: C5

t,, = 0,07 (ti, - 20) min where t,, is the thickness of insulation material in mm and p,, is the density of the insulting material in kg/m3.

Bracing Bracing should not fail within the required fire resistance period of the member. Bracing should have a residual cross section of at least 60% of the cross sectional area required for normal temperature design, otherwise the frame must be analyzed as an unbraced frame. This rule should also be applied for walls and floors (panels).

Walls For the sake of classification in case of fire, wall constructions are divided into the following groups:

-

Non-load-bearing walls are flat components that are, also in case of fire, stressed only by their dead load. In structural design they do not perform any stiffening function.

-

Load-bearing walls are stressed by compression. If walls are required to maintain the buckling stiffness of load-bearing walls the above-mentioned rules (bracing) should be applied.

-

Separating walls are located between apartments, used for stairways, escape routes or fire walls. Their main function is to avoid fire transfer from one room to another. For separating constructions fire exposure from only one side has to be considered.

-

Non-separating (non - load-bearing) walls. They do not need any fire resistance.

A wall only performs the separating function if openings (especially doors and windows) have the same fire resistance (e.g. separating function). In practice only outer and not inner walls meet this requirement. Experience has shown that outer walls with a length of > 1 m perform the separating function. The construction of a wall follows the principles shown in Figure 4.

Figure 4

Increasing stud cross section according to thinner panels (Kordina, MeyerOttens, 1994).

Figure 4 shows that the thickness of a wall is in general depending on the thickness of the panels (load-bearing wall) or insulating layers (separating wall). Using no insulating layer or reduction of the panel thickness implies a larger cross section of the studs to reach the same fire resistance. STEP/EUROFORTECH - an initiative under the EU Cornett Programme

Figure 5

Exarrzples for load-bearing walls R30. (a) Wood-based panel, (b) gypsum plaster board.

Figure 6

Example for separating wall REI 30. (a) Wood-based panel, ( b ) insulating layer.

Another problem for wall constructions are installation. Especially sockets reduce the necessary wall thickness and therefore it is not allowed that they face each other. This smaller wall thickness might diminish the fire resistance and has to be made up for by insulating layers.

Figure 7

Socket protected by insulating layer. (a) Insulating layer, (b) socket.

For separating constructions the main verifications are as follows: EC5: Part 1-2: 2.3

I

The increase of temperature on the unexposed side is limited to 140 K. This criterion is satisfied if the residual layers have a remaining failure time of 15 min beyond the required fire resistance time (see Figure 2, 3).

I1

The maximum temperature rise at any point is limited to 180 K, and no fire penetration through panel joints occurs. This criterion is satisfied if the panel joints in the outer layer of the non-exposed side are not directly exposed to fire, with a safety margin of 5 min (see Figure 2, 3).

111

To allow for building services in walls it should be verified that - after removing 45 mm of the layer on one side at a time - the reduced construction has a residual failure time of 5 min beyond the require fire resistance time.


E1215

It should also be ensured, that the panels remain fixed to the timber frame on the unexposed side. This requirement is fulfilled when criterion I1 is observed.

Floors In general the same principles (cladding and insulating layers) as for wall design could be applied. Composition flooring (or floor covering) on the upper side represents an effective protection against fire. Charring and breaking through (of loads) can be prevented by composition flooring. The under side is in most cases the unfavourable side of the floor construction (concerning fire resistance) and must be protected (insulating layer). According to this charring has to be considered from the under side.

EC5: Part 1-2: Figure C4

Figure 8

Example for floor construction REI30. (a) Flooring, (b) panels, (c) insulating layer.

In addition to the above-mentioned facts (fire protective cladding), failure times for floors exposed to fire from below should, in the vicinity of panel joints, be taken as EC5: Part 1-2: C1

where 5 is a reduction coefficient accounting for increased charring at joints (see Figure 4) and h, is the height of the panel.


Figure 9

Reduction coefficient accounting for failure time at joints for floors exposed to fire from below.


Roofs The function of the roof structure is to be resistant against fire spread or heat radiation as shown in Figure 9. Only in some special cases roof structures take over stiffening or separating functions (e.g. attic flats).

(a) - -:I. c ,-,----

EC5: Part 1-2: Figure C.4

Figure 10

(b)

Possibilities of fire load on roofs. (a) fire spread through sparks, ( b ) heat radiation.

Stairs Calculation of fire resistance follows the principles discussed for member design. In the event of fire stairs can be necessary (escape routes) or not. As concerning timber structures stairs used as escape routes have to keep their load-bearing function upright during the fire period time. Stairs used for separating functions are carried out very seldom.

Joints, grooves Joints are dealt with in lecture C19. Especially connections between walls and floors for rooms with separating function must be carried out in a way that makes them resistant against premature fire (or hot gases) penetration. Otherwise the whole construction fails before the required time. Also grooves between cladding have to be smoothed with a filler.


Figure 11

Example wall - floor connection. (a) Side of fire exposure, (b) fire penetration.

The connection on the left side would (after a certain fire period) not be as resistant against fire penetration as the connection on the right side. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E 1217

Connections to adjoining floors and walls should be detailed such that EC5: ~art:l-2:Chap.C4.3

-

the fixing is not affected by failure of panels.

-

gaps at interfaces will not give way to fire penetration into the void between panels and frame.

-

failure of panels of one construction will not give way to fire penetration into the void of an adjoining construction.

If floors or walls are connected to massive constructions different deformations and expansions may occur. In this case the interface should be sealed with noncombustible material.

Minimum dimensions In most countries minimum dimensions are given to guarantee (without verification) a certain time of fire resistance. Unprotected timber frame members should have minimum dimensions of 38 mm. For walls individual panels and sheets a minimum thickness t,,,, related to the span l,, in mm of the panel is given by: EC5: Part 1-2: C6.7

Wood based panels in single-layer constructions should have a characteristic density of at least 350 kg/m3.

Example Non load-bearing construction for inner or outer walls. The test result showed failure on the opposite surface after 65 minutes, with fire exposure from the inside.

Figure 12

(a) Non-load-bearing wall. (a) outside, (b) inside, (c) nails 2,1 X 40, ( d ) woodbased panel, (e) mineral fibre board, (f) sawn timber 60 by 97 mm, (g) sawn timber 25 by 46 mm, (h) nails 3,4 X 80, (i) wood panels with groove and tongue joints.

According to EC5 Part 1-2 the main verification for separating constructions (EZ) are as follows: a)

Fire exposure from one side.

b)

Residual layers should have a remaining failure time of 15 minutes beyond required fire resistance time (increase of temperature limited to 140 K).


C)

Safety margin in the outer layer of 5 minutes (maximum temperature rise at any point limited to 180 K).

d)

Residual cross section of the timber frame member are at least 60% of the cross section necessary for normal temperature design is required.

Verification for El 40: a) b)

Fire exposure is only from one side (see above). Failure times.

Wood based panel Mineral fibre board Wood panel

8,4 minutes. 20,4 minutes. 20,8 minutes.

.

Sum of all layers = 8,4 + 20,4 2 + 20,8 = 70,O min r 40 + 15 = 65 minutes. c)

Safety margin in the outer layer of 5 minutes.

Sum of all layers except outer layer = 8,4 minutes. d)

+ 20,4

2 = 49,2 min > 40

+ 5 = 45

Residual cross section 60%.

Charring for the timber frame member Charring rate for solid timber Charring depth Remaining cross section (height)

31,6 minutes. 0,8 mmlminutes. 25,3 mm. 71,7 mm = 74% > 60%.

Conclusion: This wall meets all the requirements (a) to (d) necessary for separating function E1 40. Although the construction reached separating function for 65 minutes in the fire test it would not perform E1 60 in calculation.

References Arbeitsgemeinschaft Holz e.V. Diisseldorf (1971). Brandverhalten von Holzkonstruktionen. Informationsdienst Holz Hartl, H. (1990). Informationsheft "Brandverhalten von Holzkonstruktionen" Kordina, K., Meyer-Ottens, C. (1994) Holz Brandschutz Handbuch. DIN 4102 Teil 4 (1994). Brandverhalten von Baustoffen und Bauteilen.


Timber-concrete composite structures STEP lecture E13 A. Ceccotti University of Florence

Objectives To describe composite timber-concrete load-bearing structures and to discuss the parameters affecting their design.

Summary This lecture deals with the composite timber-concrete technique with special reference to its performance in buildings. In particular this lecture considers structural elements where the cross-section is built from both materials linked by mechanical fasteners. Therefore the most important features of this technique and its hygro-thermo-mechanical behaviour are illustrated from a designer's point of view. A simple design example is included.

Introduction Why is it that in some countries timber-concrete composite structures are popular? Because the coupling of a concrete layer on the compression side and of a timber section on the tension side, allows the best properties of these two materials to be utilised. In fact concrete is used only in compression, where it gives its best performance in terms of strength and stiffness, and timber is used in tension, so that concrete in tension, which is only dead weight, is eliminated. Therefore it is possible to have a structurally efficient section, rigid and light at the same time. In this way the load carrying capacity of a traditional timber floor can be doubled and its out-of-plane rigidity improved three or four times. Figure 1 presents the self weight for a given service load-span combination for different types of floors.

Figure I

Floor self weight g versus span l for a service load q of 2,5 KN/m2, in the case of (a) an all-timber, ( b ) timber-concrete and (c) all-concrete section from Natterer (1993).

Moreover, the spring effect, so annoying to the user when jumping or just walking on the floor, is reduced; and vibrational damping is, in general, closer to 2% than to l%, which means that serviceability verifications for vibration are more easily satisfied. Further, the in-plane rigidity becomes so large that it may be considered infinite. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E13/1

In other words a floor, for example, is so rigid that it is able to keep its shape and consequently the shape of the entire building. This is very important for example for the survival of the structure under an earthquake. In addition this permits seismic calculation procedures based on this assumption. This is not the case for a masonry structure with simple timber floors without any in-plane rigidity as in old European constructions, see STEP lecture D10 (Figure 9). Of course it is necessary that both the timber beams and the concrete slab are well connected to the masonry wall (see Figure 2).

Figure 2

Example of earthquake-resistant design with an existing timber floor in a middle-European masonry building. (a) main beam; (b) secondary beam; (c) brick tiles; (d) concrete slab; (e) steel mesh; Cf) steel fasteners epoxy glued in to timber; ( g ) steel stirrups connecting concrete layer with masonry; (h) all-around reinforced concrete girder.

Sound insulation is also improved. On one hand, for air-transmitted noises, it is improved with respect to an all-timber floor due to the increased mass of the floor, and on the other hand, for impact noises, sound insulation is improved with respect to an all-concrete floor due to the higher damping. Regarding fire the upper concrete slab constitutes an efficient barrier against fire propagation that increases the fire resistence in comparison with an all-timber floor. In addition on the bottom timber joists are more fire-resistant than corresponding joists made of steel or pre-fabricated prestressed concrete. Finally, the cost of a composite timber-concrete floor is competitive when compared with an all-concrete floor. Actually it is not only the cost/square metre of the product per se, as may be found in a common building prices list, that has to taken into account, because there are other factors that contribute to saving money on the rest of the structure and on the building site too (e.g. more rapid execution, less concrete formwork and less stabilisation needs because timber elements can partly provide these features themselves, reduced foundations because of less structural weight, etc.). For all the above reasons composite timber concrete structures are popular in some countries, both in the case of renovation works of old timber structures (i.e. timber STEPIEUROFORTECH - an initiative under the EU Comett Programme

floors in old masonry structures) and in the case of new floors in new masonry houses too. This technique is also used for bridges (mostly in America and in the Pacific region). This technique has also been used, although less frequently, for composite walls where the concrete layer gives the racking strength and the timber ribs contribute to the flexural stiffness out-of-plane and provide the necessary stiffening against buckling. In principle, however, it could appear very difficult to obtain a good marriage between these two materials, because their mechanical properties and their hygrothermal behaviour are so different. But on the other hand no collapses or difficult serviceability problems have been reported in twenty years of use (RILEM, 1992). In the following pages the reasons why this marriage can work and the main design criteria to be followed to ensure good results will be given.

Types of connection In Figure 3 a table of the most commonly used connection systems is shown.

Figure 3

Examples of different timber-concrete connection systems. ( a l ) nails; (a2) reinforced concrete steel bars, glued; (a3/4) screws; (b1/2) connectors, split rings and toothed plates, respectively; (b3) steel tubes; (b4) steel punched metal plates; ( c l ) round holes in timber and fasteners preventing uplifting; (c2) square indentation and fasteners; (c3) cup indentation and prestressed steel bars; (c4) nailed timber planks deck and steel shear plates slotted through the deeper planks; ( d l ) steel lattice glued to timber; (d2) steel plate glued to timber.

The stiffness of the connection system may be assumed as a sort of classification index. For example, elements connected by nails, screws or dowel shaped fasteners STEP/EUROFORTECH - an initiative under the EU Comett Programme

E 1313

(A) are less rigid than elements connected by surface connectors (B) and even less rigid than elements when some notches have been cut into the wood itself (C). The stiffest connections are those where a bond between concrete and wood is obtained (D). For cases A,B and C the behaviour is like that of composite structures with a semi-rigid behaviour where the main point is that cross sections do not keep their planarity, see Figure 4. Only in sections with a D system of connection will plane sections remain plane. Roughly speaking it is possible to say that the values of the effective bending stiffness EIef may be about 50%, with A types, and up to loo%, with D types, of the bending stiffness of the correspondent sections rigidly connected. Design calculations in D case may be made easily, since there is no slip, by just "transforming" the concrete section into an equivalent timber section having the same centre of gravity but with an increased width EJE, times the real width, or by just using the method described in STEP lecture B10 for glued composite sections. In the other cases, when the semi-rigid behaviour is obtained, the method, used for mechanically jointed beams and columns (see STEP lecture B11) may be used.

Mechanical performance On the concrete side basic parameters are of course the characteristic strength f,, and the average stiffness E,, and creep coefficient @ of concrete layer. For these characteristics reference should be made to Eurocode 2. The parameters needed on the timber side are the strength f,,,,and stiffness E,,,,,, and the creep behaviour coefficient, kdeY But the most important point is the knowledge of the mechanical characteristics of the joints, mainly stiffness, i.e. the slip modulus value K,,, per fastener. This is because the rigidity of the joint determines the stress distribution along the composite structural element. In general it should be stated that strength and stiffness properties must be evaluated by tests, in particular by tests made according to EN 26891 "Timber Structures -[Testing of] joints made with mechanical fasteners - General principles for the determination of strength and deformation characteristics", that means short term tests on specimens that reproduce the real arrangement in the structural element, for example as shown in Figure 5. In order to avoid an influence of the number of fasteners on the test results (see STEP lecture C15) the number of fasteners in the specimen should not exceed two. In this way, no matter what connection system is used, the designer will be able to use the correct slip modulus in the calculations. For K,,, reference is made to the initial slip modulus K,,, = 0,4 F,,Jv,,, (see STEP lecture Cl). It is useful here to remind the reader that, when performing global analysis i.e. calculating internal actions and the consequent stress distribution at an ultimate limit state, mean values of material stiffnesses and slip modulus of connections have to be used. Actually this is because in EC5 only the mean values of slip moduli are given and the characteristic values are not available. Therefore only the mean values of the modulus of elasticity may be used: in fact if using at the same time the characteristic value of the modulus of elasticity and the mean value of the modulus of slip the calculated values of the resultant stresses would be on the unsafe side.


Figure 4

Basic behaviour of a timber-concrete beam with semi-rigid connections. ( a ) cross sections do not keep their planarity; (b) concrete layer is under compression and bending, timber beam is under tension and bending and fasteners are under shear; ( c ) the strain distribution has the same slope because the section parts keep the same curvature; stress diagram is the result of compression-bending and tension-bending stresses.

Figure 5

A possible test arrangement for determining load-slip behaviour of a connection system.

Another reason for using mean values is that Eurocode 2 only specifies a nominal value for the modulus of elasticity of the concrete, which is assumed to be an average value. Of course when performing the cross-section verification i.e. calculating design strengths, reference has to be made to 5-percentile characteristic resistances. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E1315

An elastic analysis is allowed also at the ultimate states. That means that not only timber is considered linearly elastic but also joints and concrete that have, in fact, a very well pronounced plastic behaviour. This is allowed, and is on the safe side, by considering a "nominal" secant modulus of elasticity for concrete (see Eurocodes 2 and 4) and an "equivalent" secant slip modulus for joints. For global analysis i.e. for the calculation of the global internal actions, normal force and moment, on the concrete layer and the timber beam, the concrete part is considered uncracked. This means that the entire second moment of area "I," must be considered. On the other hand, for the cross-section verification the concrete will be considered to have no tensile strength. For the concrete compression side the verification is made assuming as upper limit of resistant strength the limit value of the plasticization stress in compression (according to Eurocode 2). On the tension side, if it is the case of a cracked section, reinforcing bars shall be provided. For the slip modulus of joints reference is made to the K, value where K, = 0,6 F,,Jv,,, according to EN 26891. In general K, is not so different from 213 K,s,,). For long term behaviour, reference has to be made to creep performances of concrete, timber and joints utilising their relevant creep coefficient as conventional reduction factors. For the concrete to timber connection just the coefficient for timber joints will be used, considering that for the most part joint deformation will occur in the timber. Of course, it is evident that the method of reducing the elastic moduli is only a conventional system for taking into account time deformation but the real value of elastic moduli has no reduction in time. This method tends to overestimate the real deformation at the final stage as it can be calculated, for example, utilising a step by step method. Therefore it is on the safe side from a designer's point of view. Constraint forces would arise when the two parts are subjected to dimensional variations: e.g. when concrete shrinks. In this particular case the length reduction of the concrete slab will favour fasteners because it tends to reduce their deformation. On the other hand this would increase the deflection of the beam, although this can be counteracted by giving the beam a camber. But, finally, most of the shrinkage happens when the entire structure is still propped and the usual cracks in the concrete layer will reduce the importance of the phenomenon dramatically. Most interesting is the case of temperature variations in the concrete (the concrete layer is more affected by temperature variations than by environmental humidity variations) and moisture variations in the timber. The level of stress that can be calculated by a simple elastic calculation utilising the conventionally reduced elastic moduli, will only have a significant influence for very stiff connections and for longer elements.

Recommendations for design and construction Do not use wet timber. If it is unavoidable, use timber without pith or be sure that fissures will not affect fastener lines. Leave the propping in place for more than the time allowed for all-concrete elements. Use corrosion-protected fasteners: zinc-coated steel or stainless steel. Reinforce the concrete especially if thick concrete sections are being designed, in order to avoid loss of stiffness due to large cracks on the concrete tension side.


When casting try to protect the timber from moisture, i.e. using plastic layers or using concrete with additives in order to reduce the waterlcement ratio (that also allows smaller concrete shrinkage). This is not crucial for timber but for appearance underneath. Pay attention to timbers that do not allow the concrete to harden (e.g. in the case of larch, due to sugar extractive). With increasing spans prefer softer connections in order to minimise eventual constraint actions, and, when possible, the author prefers a structure where the concrete layer is mainly important for reducing deflections rather than for reducing the stress values in the timber.

Design example - data A composite timber-concrete simply supported floor has a span l = 4,00 m, beams spacing 0,50 m and semi-rigid connections like type A2 in Figure 3. Dimensions are shown in Figure 6.

Figure 6

Example of a composite beam with semirigid connections under bending actions: reference values ( S , fasteners spacing).

Action side Characteristic values of permanent and variable loads, per beam: = 0,70 kN/m permanent load g, variable load q, = 4,15 kN/m (medium term) I design load combination (only permanent) M , = 1,35 g, PI8 = 1,9 kNm Vd,l = 1,35 g, 112 = 1,9 kN I1 design load combination (permanent + medium term) M,, = (1,35 g, + 1,5 qk)P/8 = 14,3 kNm = (1,35 g, + 1,5 qk)1/2 = 14,3 kN

v,,


Note: because the ratio 1,35 g J(1,35 g, + 1,5 q,) = 0,13 is much less than the ratio kmo,,,,,r~km,d,m,d,,,rm (0,6/0,8 = 0,75) it is evident that only the ZZ load combination is decisive. For this reason, in the following, reference will be made only to the medium term load combination.

Material side Top slab:

concrete strength class C25130, according to ENV 206 "Concrete performance, production, placing and compliance criteria". Effective flange width b, is calculated according to EC4 as 2118 but not greater than the beam spacing, 500 mm. Beam: solid timber, strength class C22, according to prEN 338 "Structural timber. Strength classes". Fasteners: steel bars for reinforced concrete B500B, according to EN 10080, d = 10 mm, sef = 0,75 S,, + 0,25 S,, = 120 mm. Service Class 1.

Material properties The characteristic strength values as well as the modulus of elasticity are taken from EC2 and prEN 338, respectively. As already said, for the modulus of elasticity and the slip modulus of the connection, the mean value is used in the design although an ultimate limit state is considered. Concrete

fLk,cuhe = 30 N/mm2 f,, = 2,6 N/mm2 , = 2,25 (permanent load) @,,, = 1,35 (medium term load)

Timber

f

= 22 N/mm2

pO,k = 340 kg/m3

,

= 13 N/mm2

fh,",,

= 25,l N/mm2

E,,

= 30000 N/mm2

= 2,4 N/mm2

,

E,,,,,,

= 10000 N/mm2

EC5: part 1-1: 3.1.7 EC5: part 1-1: 4.1

Service class 1:

km,, = 0,s (medium term load combination) kdef = 0,6 (for permanent load) kdCf = 0,25 (for medium term load)

Fasteners

f = 500 N/mm2 M = 0,s 500 103/ 6 = 66700 Nmm K,ser = 09125 d E ~ , r n e u n = 12500 N/mm K, = 2 K,,, / 3 = 8330 N/mm according to test results valid when deformation in the concrete may be considered negligible, i.e. the case when f,,,,,, 2 30 N/mm2 and the height of the fastener penetrating the concrete is bigger than 3d.

Design strength values are calculated according to EC2, EC4 and EC5 respectively. Concrete

0,83 0,85f ckycube

fcd =

fcwd

=

14,l Nlmm2

1,5 -

0985 195

=

1,47 N

/

~

~

~

Timber


Connections concrete side (localised compression)

where is is is is

the dowel diameter, the characteristic cylinder strength of the concrete Cf, = 1,5 L), the partial factor of the material (y, = 1,25), the nominal value of the modulus of elasticity of the concrete = 30000 N/mm2

d

f,, y, E,, E,,

shear failure of the fastener

wheref, is the specified tensile strength of the steel of the dowel OC, = 500 ~/mm,). timber side

R,

=

1,5d-4

=

6,49 kN

Computation Ultimate state Initial stage The effective bending stiffness of the cross-section is calculated according to EC5, Annex B. The writer suggests a small change in the EC5 formula for the a, calculation: i.e. change the expression (h, + h2)/2 into (a, + a,). In this way it is possible to take into account also cross-sections that have a gap between part 1 (concrete) and part 2 (timber). Values of cross-section: = 20000 mm2 A, I, =2,66.106mm4 yl = 0,16 a, = 91,2 mm EId = 2,12 . 1 012 Nmm2

A, I, y2 a2

= 30000 mm2 = 100 106 mm4 = 1,00 = 28,8 mm

Design normal stresses: Concrete side 2 G , , , , = 2,93 N/mm 2 G,,,, = 4,06 N/mm STEPIEUROFORTECH - an initiative under the EU Cornett Programme

G,,, G,,

= 2,93 + 4,06 = 4,06 - 2,93

= 6,99 < 14,l N/mm2 = 1 , l 3 < 1,47 N/mm2

concrete side,

the last formula states that concrete is not cracked; therefore calculation stops here (see "Mechanical performances"); timber side, 2 G,,,, = 1,95 N/mm 2 G,,,,, = 6,77 N/mm

Design shear stress in the beam: on the safe side and for the sake of simplicity the shear force V, is considered to be totally carried by the timber beam, T,,,,, = 0,72 N/mm2 < 1,47 ~ / m m ~ Design fastener load:

Final stage With time due to the higher creep of concrete with respect to timber, action forces tend to migrate from concrete to timber. In other words stresses decrease in the concrete and increase in the timber. At final stage an effective modulus of elasticity can be used which is calculated as an average value of kW coefficients weighted according to the loads (i.e. in this case 15% for permanent and 85% for mediumterm load). The new fictitious moduli of elasticity (and consequently slip modulus) to be taken into account are therefore the following:

Values of cross-section A, = 20000 mm2 I, = 2,66 106mm4 y, = 0,26 a, = 94 mm EIef = 1,53 10" Nmm2

A, I, y2 a,

= 30000 mm2

=100.106mm4 =1,00 = 26 mm

Design normal stresses Concrete side G,,,,, = 2,83 N/mm2 G ,,,, = 2,29 N/mm2 STEPIEUROFORTECH - an initiative under the EU Comett Programme

G,, G,,

= 2,83 + 2,29 = 2,83 - 2,29

= 5,12 < 14,l ~ / m m ~ = 0,54 N/mm2 (Compression)

Timber side = 1,88 N/mmz Gm,2,, = 7,25 N/mm2

G,,,,,

Serviceability limit state

-

Deflections

Initial stage Values of cross-section AI = 20000 mm2 A, E0,rnean = 10000 ~ / m m ' K[,,, Y1 = 0,22 a2 = 36,7 mm a1 = 83,3 mm (Ear = 2,4 10" Nmm2

Final stage permanent load ECrn = 30000/(1+2,25) E0,rnean = 10000/(1 + 0,60) K,Yer = 12500/(1 + 0,60) YI = 0,36 a2 = 31,7 mm a, = 1,36 . 10" Nmm2 (El)ef U1 jin

E,,

= 30000 ~/mrn'

= 12500 N/mm

= 9230 N/mm2 = 6250 N/mm2 = 7810 N/mm = 88,3 mm

= 1,7 mm

medium term load =30000/(1+1,35) E0,mean = 10000/(1+0,25) KT,r = 12500/(1 +0,25) K = 0,35 a2 = 32,3 mm a, = 1,76 10" Nmm2 (EI),f 'Crn

= 30000 mm2

=12760N/mm2 = 8000 N/mm2 = 10000 N/mm = 87,7 mm

Although in the final stage the effective moduli of elasticity decrease considerably, the stiffness of the composite cross-section decreases to a much lesser extent.


E13111

Concluding summary Timber-concrete composite load-bearing structures are very useful for the production of stiff (in- and out-of-plane) and resistant floors (and walls). Following few design rules it is possible to realise suitable structural elements which are easy to calculate and with a suitable long term performance.

References RILEM TC 111 CST (1992). "Behaviour of timber-concrete composite load-bearing structures". Proc. of ACMAR-Ravenna International Symposium, Dept. of Civil Engineering, University of Florence, Italy. Natterer, J. (1993). "Constructions en bois 11", Notes du cours IBOIS - EPFL, Lausanne.


Bracing of structural systems STEP lecture E14 S. Winter Beratender Ingenieur der ARGE HOLZ

Objectives To develop an understanding of the minimum requirements for bracing of structural systems and to give some examples of bracing elemenets in practical use.

Summary The lecture begins with the minimum requirements for bracing structural systems. It presents the calculation of the load distribution in unsymmetrical statically indeterminate systems and shows the different elements used for bracing. The bracing of a roof structure is detailed as an example.

Introduction Structural systems must be designed to carry lateral loads caused by wind andlor earthquakes. Together with lateral bracing forces the load-carrying capacity of structures to resist wind-loads is discussed in detail and the problems of resistance against earthquake loads are shown in a qualitative manner.

Bracing criteria Lateral load resisting systems consist of vertical elements or a combination of vertical and horizontal elements. The minimum criteria are shown in Figure 1:

-

Together with a horizontal diaphragm three vertical elements are required. The elements, or their line of action, must not intersect at one point. Nor must they all be parallel to one another.

-

Without a horizontal diaphragm four vertical elements are required, which only two of these elements are allowed to cross at one point.

More parts of the structural system could be added to this "basic box", which when braced as described allows the lateral loads to be transferred partly or completely to the foundations through the elements. The bracing shown in the Figure 1 gives balanced situations (e.g. wind loads along the narrow side of the building). If the horizontal loads are not centred, torsional situations occur which lead to additional forces and deformations.

Figure I

Bracing of a structural system - minimum requirements.


-

Bracing geometrical location Bracing elements of a structural system should be located in such a way, that symmetrical systems are created. Otherwise the centre of resistance will not coincide with the centre of pressure of the loads and additional torsional loads have to be taken into account. This may be ignored for buildings up to two storeys in timber frame construction, if vertical shear wall units are located in a minimum of four circulating walls (Briininghoff et. al., 1989). Support loads to continuous horizontal diaphragms may be calculated assuming a number of non-continous twohinged beams. The bracing elements should be of the same stiffness for each direction of lateral loads, otherwise an additional excentricity has to be taken into account. The calculation method given in Figure 2b is only valid for an equal stiffness per unit length of all wall diaphragms. For a more detailed calculation with regard to displacements due to bending, shear and slip in joints, see Steinmetz (1992).

Figure 2a

Calculation of forces in vertical elements, statically determinate system.

Figure 2b

Vertical elements in a statically indeterminate system.

The forces in the system shown in Figure 2b may be calculated according to Equations (1) to (8).

STEPIEUROFORTECH

- an initiative


Wind direction W,

bxi Hxi = wx

C 'xi

HY .f

=

wx

ey ',i bxi 2

+

E bxiS;

+

by, sxi

Wx ey 'xi byi 2 2 C bxisyi + C byi sxi

Wind direction W,

Hxi

Wy ex

=

C bxiS; byj H .= WY yf

C byi

'yi 'xi 2

+

byi S,

Wy ex'xi +

CbxiS;

+

byi xbyi sxi 2

Vertical elements Frames Frames have rigid moment-resisting connections which resist vertical and lateral loads. They are commonly used in single-storey industrial buildings. To produce moment-resisting connections fasteners (e.g. dowels) or finger jointed corners are used. For horizontal loading timber frames are normally used only to resist wind loads. For higher loads, e.g. horizontal braking-forces from cranes, the cross-section requirements would rapidly increase mainly due to the need to accommodate more fasteners. The restricted clear width of the building due to the large corner crosssections is another disadvantage. All these reasons lead to uneconomic solutions when higher loads are present.

Cantilevers In industrial buildings with cranes, steel or concrete cantilevers are used. If lower load-values are expected, roundwood (see STEP lecture E19) and glulam (Heimeshoff, 1983) cantilevers could be economic, e.g. for agricultural buildings or small sized halls. Deformations and stresses should be calculated according to EC5 except in the design of the base connection. Rigid base connections for clamped columns are given by Heimeshoff (1983) and to avoid fungal attack caused by high moisture content very careful detailing of the base connection is necessary.

Diagonal bracing Diagonal members may be designed to act in tension andlor compression, e.g. STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E 1413

timber diagonals in frameworks. When acting only in tension, crossed steel rods with threaded ends are used. Diagonal bracings are often located within the wall, but sometimes they are located outside the cladding and are used for architectural reasons.

Vertical shear wall units Vertical shear wall units consist of timber studs and plates and wood-based or gypsum-based panels. Shear forces are transferred through a large number of connections (nails or screws) around the panel edges (see B13). To prevent uplift vertical shear wall units must be fixed to the foundation. The compression-studs have to be designed for vertical loads and the additional vertical forces caused by horizontal loads. Because of their good strength and stiffness, fibreboards, Oriented Strand Board and plywood are the most common materials to produce shear wall units. In timber framing construction they act as multi-functional elements. They also fullfill functions such as fire-resistance, separation and noise insulation and, structurally, resist vertical and horizontal face loadings.

Figure 3

Vertical bracing elements, ( a ) frame, ( b ) cantilever, (c) diagonal bracing, ( d ) shear wall unit

Bracing elements for horizontal forces All buildings require horizontal bracing at each floor and at roof level using horizontal diaphragms or horizontal diagonal bracing. At roof level the lateral load resisting system is normally located within the plane of the roof. In floors woodbased panels are carrying vertical loads and shear forces. In the perimeter edge beams and attached wall plates carry the flexural forces of the horizontal diaphragm. The floor acts as a horizontal beam carrying lateral loads to the vertical elements (see STEP lecture B13).

Figure 4

Combination of different types of bracing to horizontal forces in a industrial building.

Diagonals at floor or roof levels can form parts of trusses carrying the horizontal loads to the adjacent vertical elements. Therefore all connections and the tension and compression members have to be detailed very carefully. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

At roof level steel rods or lightweight thin gauge steel diagonals acting together with purlins, or small horizontal trusses can be used. Trussed rafters always have to be laterally braced by additional trusses at roof level. Using trusses and glulam beams the bracing not only acts as a lateral load resisting system but is also used to avoid lateral buckling (see STEP lecture B15).

-

Design Details

Design procedure -

Establish the building geometry and the lateral load resisting permanent walls and parts of the structure. In buildings in earthquake regions the system must be designed to resist seismic and wind loads. The ductility of the bracing system will influence the value of the equivalent static lateral force. Nonductile structures or joints may be designed for higher lateral loads but give uneconomic solutions in earthquake design.

-

Specify lateral loads according to EC1 andlor EC8 taking care to relate the load to the practical structural system, for example checking if studs are continuous over storey height when considering horizontal wind load distribution.

-

Calculate the resulting forces in bracing elements. For symmetrical buildings simple rules could be used. For buildings with uneven distribution of bracing elements or differing stiffnesses of the bracing elements the additional torsion forces may be calculated as given in Figure 2b. The best torsional resistance is provided by widely spaced vertical elements.

-

Calculate the load carrying capacities, S, 5 R,

-

Consider joint design for the bracing elements. The flow of load from application point to foundation should be traced. Design these joints:

-

Connections of load carrying members such as studs or rafters to horizontal elements, e.g. horizontal diaphragms, with special care taken in dealing with uplift forces caused by suction.

-

Joints within the horizontal bracings, e.g. connections of prefabricated floor or wall elements, joints within horizontal trusses, connections of purlins and frames etc.

-

Connections of horizontal elements to vertical elements e.g. horizontal diaphragms to shear walls or chords of horizontal trusses to shear walls.

-

Connections within vertical elements, e.g. connection of prefabriated shear walls or nailing of wood-based panels.

-

Connections of vertical elements to foundation, e.g. nail-on plate at the end of shear walls or anchoring systems of steel rods to the bottom of studs.

When detailing joints it is important to avoid stressing timber perpendicular to grain, because timber is weak in this direction and sudden failure could occur. -

Calculation of displacements. The allowable displacements depend on the type of building. The calculation is more important for structures with high vertical loads. High values of deformation lead to high values of additional forces and will lead to cracks in claddings which should be avoided.


E 1415

-

Calculation of lateral load resisting systems in place during the erection of the building to secure stability of the structure in all stages of fabrication. During the erection of prefabricated timber framing buildings only a temporary bracing is necessary, e.g. by adjustable, oblique steel studs. In this case a special calculation is normally not necessary. For frames and spaced frames etc. special calculations may be required. Here it is normal practice to start erection with the vertical parts of the structure connected directly to bracing elements, e.g. Figure 4, section A.

Bracing elements

-

Light steel diagonals are mostly used for roof constructions acting within the roof level. They are also used at walls and floors in small buildings. The end connection is provided by nails (EC5: Part 1-1: 6.2.2(1)). The design value of the lightweight thin gauge steel diagonal is normally governed by the design load-carrying capacity of the nails. The maximum design loadcarrying capacity should be taken as wlax R, = A,,,

f,,,

The number of nails can be increased by using additional pieces of timber beside the rafters at the end of the diagonals (see Figure 8). Lightweight thin gauge steel diagonals must be tensioned during erection but have the advantages of low cost and an easy and quick installation. However erection in practice is often incorrect and disadvantages are the low load carrying capacity and the thermal elongation. If a lightweight thin gauge steel diagonal is erected during a cold winter period, due to thermal elongation of steel great deformations along the diagonal could occur in the summer. Therefore the load carrying capacity of the diagonal could only be activated by large deflections of the entire structure.

-

Steel rods with threaded ends are effective and aesthetic bracing elements. They are normally custom designed and can transfer large loads. Many alternative systems are possible, some details are shown in Figure 5. It is an advantage that the simple tensioning system of the steel rods can be used to adjust the structure during erection. To reach a good ductility of the bracing, all connections have to be overdesigned to ensure tension yielding of the rod is the failure mode. If steel rod bracing is designed carefully it can be used to enhance architectural appearance.

Figure 5

-

Different details of diagonal bracing, ( a ) timber diagonal with nailed steel plate, (b) steel rod with simple end connection, (c), ( d ) steel rod with castiron connectors.

Timber or glulam diagonals are custom designed according to their loads and buckling lengths. They carry tension and compression forces. Connections take the form of nailed steel-plates, nailed wood-based panels, nails, bolts,


connectors etc. (see Figure 5). Timber diagonals are often more economic than steel rods but since adjustment of the structure during erection is not possible, precise workmanship is necessary. Canying tension and compression forces stiff constructions with small deflections are produced. To create good ductility well detailed nail plate connections or nails and bolts with small diameters should be used. Diaphragms made from wood-based panels and timber plates and studs are normally used in single- and multi-storey houses. Using the multi-functions of the panels this kind of bracing is often the most economic and also gives an excellent ductile behaviour, much better than any diagonal bracing system. Wall- and floor-elements could be prefabricated, connections of prefabricated elements have to be calculated, especially the connections of horizontal and vertical diaphragms. Board edges not fixed to a stud or a beam have to be secured with additional timber joists if no special calculation and detailing is used. The characteristic load-carrying capacity of the wood-based or gypsum-sheathed panels are given in European Standards, National Application Documents or European Technical Approvals. These are normally determined by test since calculation would involve embedment values which are only available for timber, plywood and hardboard in EC5. Tension and compression studs of shear walls have to be calculated considering also vertical loads. Holding-down details such as those given in Figure 6 must also be calculated.

Figure 6

Typical holding-down detail along shear wall edge.

Example Bracing of a roof construction with lightweight thin gauge steel diagonals.

Figure 7

Roof system.

STEP/EUROFORTECH

- an initiative


Material properties:

softwood C24 according to EN 338 steel Fe 360 annular ringed shank nails 4,O X 60 mm, Service class 2 kmod = 0 , 9 a = 45" P = arctan 7 64 = 39,4" 9,30 W

H

= - =

2

'l

5,25 kN

HL = Fw tanp cosa

=

=

F, = 6,79 kN Z = -

F,

cos p

3,05 kN

=

V

Calculation of bracing member

Lightweight thin gauge metal steel diagonal 40 A,,, = 40 . 2 - (2 . 5 2) = 60 mm 2

X

2,O

Calculation of nails

Annular ringed shanked nails 4,O X 60 mm

where t, = 58 mm (penetration depth), d = 4,O mm (nail diameter). Rd = fh,,

=

1'1 \ / 2 W d

.

0,082 p, d4',3= 0,082 350

R, = 1,l 42

n

=

6790 873

=

6015

. 13,l

7,77 say n

=

4,O

=

. 4,04.3 =

18,9 Nlmm2

873 N

8 nails


Angle bracket to transfer V and H,,

V = 3,05 kN ; H,,= 5,25 kN Figure 8

Base connection of lightweight thin gauge diagonal: ( I ) rafter 1001'180 mm, ( 2 ) additional timber 120/160 mm, (3)purlin 100/140 mm, (4)lightweight thin gauge steel diagonal 40 X 2,0 mm, (5) angle 105 X 105 X 90 mm .

The additional timber (2) is necessary to keep nail distances when nailing lightweight metal gauge diagonal. The angles are each nailed with 8 nails on the side and with 10 nails on the bottom. According to technical reports from the producers of the angles the design load-carrying capacity may be taken as V,

= 9,94 kN

and

Combined load v v, + H,, / HI,,, 3,05 / 9,94 + $25 1 7,89

H,,, = 7,89 kN I 1,O = 0,98 c 1,0


The flow of forces must be followed throughout the structure, in order that all connections be accurately calculated.

-

Minimum bracing requirements must be taken inro account.

-

Location of bracing elements with respect to the stiffness of the bracing should be symetric to load location to avoid additional torsion of the structure.

-

In earthquake regions ductility of bracing has to be confirmed. Design of the bracing structure should be calculated according to wind and earthquake loads.

-

Selection of bracing structures should consider load-canying, ductility, economic and erection aspects (e.g. prefabrication).

-

Lightweight thin gauge steel diagonals have to be tensioned to guarantee their immediate take up of bracing loads.

References Briininghoff, H et.al. (1989). Holzbauwerke: Eine ausfuhrliche Erlauterung zu DIN 1052, Teil 1 bis 3. Deutsches Institut fur Normung e.V., Deutsche Gesellschaft fur Holzforschung e.V., 1. Auflage. Beuth Verlag, Berlin, Koln, Germany. Steinmetz (1992). Die Aussteifung von Holzhhsern am Beispiel des Holzrahmenbaues. Holzbau Statik Aktuell, Ausgabe Juli 199211. Arbeitsgemeinschaft Holz e.V., Diisseldorf, Germany. Heimeshoff, B. (1983). Einspannung von Stiitzen aus Brettschichtholz durch Verguf. in Betonfundamenten. Holzbau Statik Aktuell, Ausgabe Juli 1983/7. Arbeitsgemeinschaft Holz e.V., Diisseldorf, Germany. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E1419

Beam and post structures - Principles STEP lecture E15 K. Schwaner Arbeitsgemeinschaft Holz

Objective To set out the principles governing the design of beam and post structures.

Introduction A continuous cycle of development spanning many centuries precedes the timber-frame construction methods of today. The first attempts to construct shelters were made in prehistoric times. Vertical posts were dug into the ground, then, horizontal planks were placed side by side and joined to form a frame which bore the brunt of the weight. Dividing walls were filled out with wattle and partly coated with clay. In this way, man erected such constructions for the first time, where the load-bearing frame was separate from the non-load-bearing dividing walls. In the Middle Ages, the classic half-timbered building was developed from above mentioned method of construction. Examples which can still be seen today pay tribute to the very high level of the builders craftsmanship (Figure l). Towards the end of the 18th century the half-timbered buildings became less significant in comparison with the solid masonry construction methods. This was not least due to the fact that the bourgeoisie, who were gaining more and more influence in society, also developed a new set of values where construction methods were concerned, including, for example, a preference for more "solid" and more "durable" buildings. Timber-frame construction still played a significant role when no particular importance was attached to representative architecture and where buildings had to be erected as cheaply as possible.

Figure 1

The town hall in Frankenberg, Germany

Since the mid 20th century there has been a fundamental change in the attitude towards timber constmctions: timber is gaining importance once again as more and more people learn to value its high quality and the numerous things it can be used for. However, due to rising wage costs, the half-timbered constructions built by craftsmen became uneconomic and instead different methods of construction were developed, such as timber-frame construction and modern beam and post structure methods. STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E15/1

Modular construction A timber-frame structure is a three-dimensional load-bearing construction, erected within a grid, consisting of linear elements which are supplemented by a curtain wall or panels. There is a clear division between the load-carrying construction and the f a ~ a d eelements (Figure 3).

Figure 2

The skeleton is erected.

Figure 3

The house is completed.

All beam and post structures, whether made from timber, concrete or steel, are based on a basic module. This module ensures that the structure of each beam and post building is standardised and structurally sound. A multiple of the module produces the grid. The module and grid determine the position of the load-bearing columns, the location of the main and secondary beams and often also the way in which the non-load-bearing finishing elements are arranged. An example of this type of elementary unit is the traditional Japanese home which, in the ground plan and cross and longitudinal section, is built on the basic unit of the Tatami mat (0,91 X 1 3 2 m). The planning of beam and post structures begins with the choice of a ground plan module. Its size, and thereby that of the grid, depends on a wide range of factors and can be given a new and different specification for each building.


Module and grid are governed by

-

human dimensions: all types of scales drawn up by man such as, for example, Leonardo da Vinci's theory of proportion (Figure 3), Le Corbusier's Modulor, the Japaneses floor mat, the span of a hand, the length of a stride or the height of a step

Figure 4

Leonardo da Vinci's proportion study of 1490. Pen-and-ink drawing, Venetian Academy.

-

architectural design: configuration of the building parts, order and arrangement of the inner rooms, size and scale of the rooms, structuring of the fa~ades,specific requirements regarding what the building is to be used for

-

structural and material considerations: economic beam spans, standardized joints, customary formats of finishing elements, insulating panels, panelling, standard sizes of windows, doors, stairs, spans of load-bearing floor elements

-

site conditions: geographic position, orientation, size of plot, tree population, vicinity of existing buildings, specifications in the master plan.

There is no one single definitive module size: for example, there is the European module (DIN 8000) = 100 mm; the brick size (DIN 41721) = 115 + 10 mm; the stair module = 90 mm (half height of step); multi-module = 600 mm. These produce the customary grid scales of 1,20 by 1,20 m, 1,25 by 1,25 m, 3,60 by 3,60 m or 4,80 by 4,80 m. In addition, the distinction is made between the band grid and the axial grid (Figure 5). The band grid defines the clearance between the parts of the building in a load-bearing construction, the axle grid always refers to the centre lines of the columns or of the load-bearing system. The number of different possible modules alone illustrates that there is not one universal grid scale. The specification of the respective grid is part of the design process and therefore also constitutes a design criterion. The selected grid is reflected in every part of the building. It can, for example, clearly be seen in the format of the windows, the distances of the load-bearing construction or in the STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E1513

division of the exterior cladding. The effects on the formation of the skeleton are as follows:

-

small grid (e.g. 1,20 by 1,20 m): The structural system is identical to the finishing system; every column is load-bearing; small cross-sections; simple e.g. nailed joints.

-

largegrid(e.g.3,60by3,60m): There is a clear division between primary and secondary system. In this case there are larger cross-sections manufactured, for example, from glued laminated timber; large, open spans and less joints.

fa) Figure 5

(bJ

Axis ( a ) and band (b) grid.

The economy of the building measures undertaken is also dependent on the choice of grid and load-bearing system. As with other timber engineering structures, the number and form of the joints will also greatly influence the overall cost of a beam and post structure. The most common systems are based on a right-angled ground plan grid. In principle, however, other grid systems (radial grid, triangular grid, etc.).

Load-bearing systems and joints Almost all beam and post structures can be traced back to a few principal loadbearing systems. Each of these construction forms distinguishes itself from the other by the order and position of the load-bearing elements (columns, main and secondary beams) and by the joints of these building elements. Columns and main suporting beams have single or multiple sections. The secondary supporting beams are flush or stacked (Figure 6). The most important configurations are described below.

Figure 6

Load-bearing systems in beam and post structures.

System 1: Single section main supporting beam and single section column, single-storey Wfth this system, the continuous or single span main supporting beam rests directly on the column. The secondary supporting beams can be stacked on the main supporting beams or hung level between them. In the case of a small grid scale, load-bearing boards or wood-based panels can be laid directly on the main STEPIEUROFORTECH - an initiative under the EU Cornett Programme

supporting beams. Cantilevers in both directions are possible. The load transfer between main supporting beam and column occurs through bearing column ends must be accuratly cut for a good fit. There are simple geometric ratios for the finishing and fagade connections. When conneting to main and secondary supporting beams, the different construction heights and other details should be taken into consideration. Possible connections for supporting beam to column: These joints are mainly for allocation and eventually against uplift loads: tenon joints timber or wood-based material exterior or interior butt plate straps exterior or interior steel plates glued-in bolts end-grain joints with Type A connectors according to EN 912 steel angle brackets T-shaped integral fasteners (Figure 8).

Figure 7

Single section main supporting beam and single section column, singlestorey. (1) Column, ( 2 ) secondary beam, (3) main beam.

Figure 8

T-shaped integral fastener.

System 2: Multiple section main supporting beam and single section multistorey column This load-bearing system consists of single section multistorey columns and multiple section main supporting beams. The secondary beams (ceiling beams) normally rest on the main supporting beams to enable the use of the continuity effect for an economic structural design. They can be placed parallel on both sides of the column in the grid axes STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E 1515

or be built in as single span beams between the column faces. This system is suitable for two- or multi-storey buildings, in particular for split level floors, since in this case the beams can be fixed to the columns at any chosen level. Cantilevers in both directions are then also possible. Possible conections for main supporting beam to column:

-

dowels or bolts toothed-plate connectors, ring and shear-plate connectors wooden cleat supports steel angle brackets column with scarf joint, part of the bearing force as compression mixed constructions with these types of joints

The secondary beams are normally laid over the main supporting beam. This stacking method of construction results in relatively high floors. In addition, coupled with the multiple section method of construction, it also creates voids and thereby complex connections from the inner walls and the f a ~ a d e . The principle of construction is simple, the construction itself and the loadbearing functions are easily recognisable. Characteristic of this type of construction are the protruding ends of the beams, which can normally clearly be seen, with the fastener end distances which often prove necessary. It should, however, be noted that it is almost impossible to avoid penetrating the faqade of the building. The great demands of building physics which are made on modern outer walls (wind tightness, heat loss in joints, thickness of insulation) are particularly difficult to fulfil1 using this method of construction. In order to avoid having to treat the timber with preservative, no timber parts within the construction must be directly exposed to the weather. As a result, ends of the beams protruding through the f a ~ a d eshould be avoided.

Figure 9

Multiple-piece main supporting beam and single section multistorey column. (1) Column, ( 2 ) secondary beam, (3) main beam.

System 3: Single section main supporting beam and multiple section multistorey column This method of construction is often chosen for architectural reasons. The continous main supporting beams are fixed onto the multiple section columns by mechanical timber fasteners. However, due to the slenderness of the columns, additional lining rails or battens often have to be built in as well. Sometimes the main supporting beam can lie directly on top of this batten. Cantilevers in both directions are possible here, too.


The small size of the column cross-sections is problematic where fire design is concerned. The cross-sections often have to be enlarged in order to obtain the required fire resistance. Due to the complexity of this principle of construction, restrictions similar to those of the above-mentioned system apply, regarding finishing and faqade element joints.

Figure 10

Single section main supporting beam and multiple-section multistorey column. ( l )Column, ( 2 ) secondary beam, (3) main beam.

System 4: Single section main supporting beam and single section multistorey column The load-bearing skeleton consists of single section multistorey columns. The main supporting beams are then joined to this column as single span beams at any level. Here, the secondary support beams can be laid flush with the main supporting beams or stacked.The advantage of the former construction is that, because the joints are the same level in both directions of the axis, very simple geometries are created for connecting internal and external walls. If all main supporting beams are lying in one particular alignment and all secondary beams are lying at right-angles to these, an aligned system is produced. A non-aligned system is created when the main supporting beams are arranged at the same height in both directions, but the direction of the secondary beams is changed to a chequered pattern. Thus all main supporting beams are proportionately loaded - the connections are symmetrical and the number of different construction elements and joints can be greatly reduced. This standardisation makes up considerably for the disadvantage of the large number of fasteners and the fact that more timber is used. The system is particularly suitable for prefabrication of the load-bearing elements and the complete finishing components. It enables construction elements to be produced with precision and independent of weather conditions and to be assembled rapidly. Moreover, the faqade of the building can be placed around the building without being affected by wall openings. Only in this way, for example, can a beam and post structure fulfil1 the high standards expected from a low-energy building. Balconies and canopies are aligned against the f a ~ a d eusing independent loadbearing systems. Possible connections for main supporting beam to column:

-

dowelled joints with centre member of welded steel T-shaped integral fasteners cold-formed steel fasteners, e.g. joist hangers wooden cleat supports and steel angle brackets contact pressure on specially formed cross-sections of the columns proprietary fasteners, e.g. Janebo hook plates (Figure 12), BSB-system (Figure 13)


E 1517

Figure l 1

Single section main supporting beam (level) and single section multistorey column. ( I ) Column, ( 3 ) main beam.

Figure 12

Proprietary fastener: Janebo-hook-plates.

Figure 13

Proprietary fastener: BSB-system.

Most of these joints can also be used as a connection at the same height as the main and secondary beams. In addition there are Z-profiles, end-grain joints with ring and shear-plate connectors and supports on load-bearing battens on the main supporting beam. If the timber-frame elements remain visible, the joints must then also fulfil the aesthetic requirements. The choice of load-bearing system can also be seen as a statement of architectural creativity. A distinction is made between different systems: those which arc fully integrated into the wall, partially integrated or non-integrated. In the case of an integrated load-bearing system the skeleton itself is not visible and many different types of materials and surface conditions are possible for both the internal and external panels. In a partly integrated load-bearing system the wall surfaces are arranged flush against the columns in accordance with the grid. Despite structural texture, the whole wall surface is preserved. In a nonSTEPIEUROFORTECH - an initiative under the EU Comett Programme

inteerated load-bearing system the beams protmde and become the main ---. characteristic of the interior design. -

V

Figure 14

Delivery of complete beam and post structure components to the building site.

Figure l 5

Assembly of columns and main supporting beam.

Figure 16

Assembly, alignment and fixing of the first structure in one axis.


Figure 17

Aligning the rest of the structure.

Figure 18

Placing of the secondary beams.

Figure 19

Completed assembly of the groundfZoor.

Bracing In a beam and post structure the functional load-transfer construction is entirely separate from the faqade and external wall elements. Only the linear elements (beams and columns) ensure the stability of the building, i.e. the transfer of all STEPIEUROFORTECH - an initiative under the EU Comett Programme

vertical and horizontal loads into the ground. In particular, the spatial stability and the bearing and transfer of bracing actions and wind loads must be calculated. Bracing actions are horizontal forces which, for example, are caused by wind, column inclinations and deviations of straightness of the beams due to imperfections in the processes of manufacture and assembly. Beam and post structures are braced in both horizontal and vertical directions. In certain cases, if enough vertical bracing elements are available, horizontal bracing constructions are not necessary. Bracing in a horizontal direction is carried out using diaphragms in floor and roof levels. Bracing in a vertical direction is carried out using shear walls, bracing constructions, portal frames, restraint columns or rigid staircases (Figure 20). The diaphragm can be constructed using panelling with wood-based material, aligning bracing struts or by laying floorboards with the tongue and groove in a diagonal direction. It should be noted that a new series of floor systems are currently being developed in which the diaphragm effect is already an integral part. Glulam panels, boxes with webs and panels of wood-based material, laminated veneer lumber, stacks of vertically arranged solid timber planks or boards as well as timber-concrete-composite systems are particularly worth mentioning.

1

Figure 20

2

3

4

Vertical bracing elements. ( I ) Shear walls (timber, brickwork, concrete), ( 2 ) crossed diagonals, (3) portal frames, (4) knee-braced frame.

Vertical linear or planar bracing elements can be aligned to bear the horizontal forces from the floor and roof level diaphragm. The planar elements are shear walls of concrete, brick or timber frame walls with wood based panels. If a structural calculation is carried out to transfer the bracing actions into the joints, brickwork between the columns can then be used for bracing. Bracing can also be carried out by aligning steel or timber bracing struts or aligning crosswise diagonals, or by building portal frames or restraint columns. If the floor and roof level diaphragm is braced to a sufficient extent, a rigid staircase or maintenance shaft can then act as a complete vertical bracing element.

Figure 21

Gunz house, Keltern, Germany.


Figure 22

Isometry of the Gunz house.

Conclusion The changing face of the building industry, with new emphasis being put on construction methods which save costs and protect resources, has led to increased opportunities to make use of building materials and types of construction which on the one hand can meet the individual needs of the customer and on the other are technically advanced, uncomplicated and which can be assembled without difficulty by the local manufacturer. Modern beam and post structures in particular fall into this category (Figures 21,22). Construction methods developed from past models, new connection techniques and materials ensure that buildings constructed using the beam and post structure method are every bit as stable and durable as buildings constructed using other methods. In addition, they offer such advantages as adaptability, ease of building alteration and extensions, pleasant aesthetic design, exemplary construction physics, ease with which components can be prefabricated, practicability on different technology levels (small workshops, rationalized medium-sized firm, industrial production), easy to transport and assemble. Due to the fact that the customer plays an important contributory role in the construction process, particularly where finishing elements are concerned, it is also possible to reach a marked reduction in building costs. The beam and post method of construction provides a variety of creative design possibilities for a wide range of different buildings and structures. It is employed for housing and office buildings as well as for public buildings - kindergartens, schools, senior citizens' homes, sport and leisure halls. Religious buildings can also be constructed in suitably aesthetic architectural designs using the beam and post structure method. STEPJEUROFORTECH - an initiative under the EU Comett Programme

Beam and post structures - Design STEP lecture E16 N. Nebgen Ingenieurbiiro fur Holzbau

Objectives To show the design of a small office building in Reutlingen, built in 1994. To present the calculation and detailing of the most important elements of the building, such as roof and floor beams, stringers and columns. To demonstrate the design of the connection between the stringer and column.

Introduction The load bearing structure of the office building consists of single section beams, stringers and columns. The beams, stringers and the columns in the roof are designed with a level upper surface hence all beams are single-span girders. The beams are connected with T-shaped steel-plate castings to the stringers. Fastener joints are rod dowels in the beams and grooved nails in the stringers. The stringers and the columns are connected by slotted steel-plates and dowels. The horizontal force is transferred through the roof area and some of the walls, which are formed as wood-based diaphragms. The calculations allow for an additional storey to be added later. In the top of the columns screw sockets are included, to which the columns of the second storey can be fixed in the case of extension.

Design example 4,52 $

Figure 1

4,52 P

,

r

4,52

a

r

4,52

Ground plan and cross section.


E1611

Secondary beams (pos. 1 in Figure l ) Single storey case; roof loading Permanent actions Roof felting Thermal insulation Timber boarding Self weight of the beams

0,05 kN/m2 0,12 kN/m2 0,18 kN/m2 0,20 kN/m2 G, = 0,55 kN/m2

Variable actions Snow Imposed load

Q, = 0,75 kN/m2 2,OO kN/m2

Wind pull with uplifting function not governing Actions per beam spacing 1,13 m on centres g, = 0,55.1,13 = 0,62 kN/m q , = 0,75.1,13 = 0,85 kN/m Two storey case; floor beam Permanent actions Finishes Particle board flooring Concrete slabs, thickness 50 mm Boarding Self weight of the beams

Variable actions Imposed load Light partition walls

Actions per beam

0,lO kN/m2 0,20 kN/m2 1,lO kN/m2 0,18 k ~ / m ~ 0,20 kN/m2 G, = 1,78 kN/m2

2,OO kN/m2 0,75 kN/m2 Q, = 2,75 k ~ / m ~ spacing = 1,13 m g , = 1,78. 1,13 =2,01 kN/m q , = 2,75 . 1,13 = 3,11 kN/m

Thus the flooring condition governs the beam design. System

Internal forces and moments A , , = B , , = V,,, = 0,5 . 2,01 . 4,52 A , , = B,,, = V,,, = 0,5 3,11 . 4,52 = 0,125 - 2,Ol . 4,52' M&k M,j k = 0,125 3,11 . 4,52'

= 4,54 kN = 7,03 kN = 5,13 kNm = 7,94 kNm

Design value of the internal forces for ultimate limit state for serviceability limit state STEPIEUROFORTECH - an initiative under the EU Cornett Programme

chosen GL24 b X h = 120 X 260 mm service class 1 load duration class: medium A =31,2.103mm2 W =1,35.10~mrn~ E = 10800 N/mmz km,, = 0,s kdej;g = 0,6 (permanent) Bending

Shear

Deflection

Main beam (pos. 2 in Figure 1) Permanent actions Self weight of the stringer Floor load 1,78 . 1,13 I 2 G,

= 0,40 kN/m = 1,Ol kN/m =1,41kN/m

Reaction from floor beams

N,

= 4,54 kN

Variable actions from floor 2,75 . 1 ,l3 1 2 Reaction from floor beams

Q,

= 1,55 kN/m

N,

= 7,03 kN

System N

N

N


I = 176 . 106 mm4 km, = 0,25 (medium)

Internal forces and moments A,, = B,,, = V,,, = 0,5 . 1,41 . 4,52 + 1,5 . 4,54 A,, = B,, = V , , = 0,5 . 1,55 . 4,52 + 1,5 7,03 = 0,125 1,41 . 4,522 + 0,5 - 4,54 . 4,52 M,jk = 0,125 . 1,55 . 4,522 + 0,5 . 7,03 . 4,52 M,j k

= 10,O kN

= 14,l kN = 13,9 kNm

= 19,9 kNm

Design value of the internal forces for ultimate limit state for serviceability limit state

chosen GL24 b X h = 160 X 360 mm service class 1 load duration class: medium A = 576 . 103mm2 W = 3,46 . 106mm3 E = 10800 N/mm2 km,, = 0,8 kdeJ, = 0,6 (permanent)

I = 622 . 106mm4 kdeJ,=0,25 (medium)

Bending

Shear

Deflection


Column, (pos. 3 in Figure l )

Permanent actions N , , = 0,55 . 4,522 = N3,g= 1,78 . 4,522 N I , + N 2 , = 11,2 + 36,4

= 11,2 kN = 36,4 kN = 47,6 kN

the self weight of the column may be disregarded. Variable actions N I ,= 0,75 4,522 NzSq= N3,q= 2,75 . 4,522

= 15,3 kN = 56,2 kN

the eccentricity of the load N2 will be disregarded. Load combinations Value of combinations

snow imposed load

Design value of internal forces and moments " ~ ~ ~ , i ~ o , i Q k , i z~G,jGkj "+ " If+

Buckling in the middle of the column Combination N, = 1,35 Combination N, = 1,35 Combination N, = 1,35

l (permanent load + imposed load + snow) 47,6 + 1,5 56,2 + 1,5 0,7 15,3 2 (permanent load + snow + imposed load) 47,6 + 1,5 15,3 + 1,5 0,7 56,2 3 (permanent load + imposed load) . 47,6 + 1,5 . 56,2

.

Combination 3 is governing chosen GL24 b X h = 160 X 160 mm service class 1 load duration class: medium STEPIEUROFORTECH - an initiative under the EU Cornett Programme

= 165 kN

short

= 146 kN

short

= 149 kN

medium

Check axial load at base (no buckling): Combination 3 (permanent load + imposed load ) = 1,35 47,6 + 1,5 2 56,2

.

N,

= &,o,d

. 10' 193 . 103 - 0,8 26,O

233 25,6

=

= 233 kN

medium

0,57 < 1

The columns loaded by wind bracing loads will not be treated here.

Joint stringer-column

Plan view of the upper connection

Elevation showing upper connection

Elevation showing lower connection Column base


The following calculations show the design of the main beam to column connection. V, = 34,6 kN (support reaction of the stringer)

Maximum force on dowel in the column: = 34,6 / 3 = 11,5 kN Moment M = 34,6 0,19 = 6,57 kNm

D,

assumed: hinge in the column axial force on dowel in the stringer: D, = 34,6 / 8 = 4,32 kN

r,

=

AD,

=

=

6,57 4

0,126 m

0,126 0,126~+ 4

. 0,056~

=

10,9 kN 0

Resultant force on dowel in main beam: a =arctan(4/12) = 18,4" AD, = 10,9 sin 18,4" = 3,44 kN AD, = 10,9 COS 18,4" = 10,3 kN D, = 3,44 + 4,32 = 7,76 kN = 10,3 kN D, = p = arctan (7,76 1 10,3) = 37,0°

D

-4 =

=

12,9 W

e

0

8,

Design value of load-carrying capacity km,, = 0,s ,y = 1,3 for timber STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E1617

= 1,l for steel used in joints

,y

=410kg/m 3 = 1,35 + 0,015 d

pk

k,,

= 1,35 + 0,015 16 = 1,59

Embedding strength

EC5: Part 1-1: 6.2.1 (3)

Yield moment (y,

ECS: Part 1-1: 6.5.1.2 (2)

MYg

=

0,8

= 1,l):

16~ 1 = 179 . 360 . 6

10' Nmm

1,l

Design value of the load-carrying capacity p = oo

R,

=

R,,,,

min

l

=2

. 11,O

1,l .17,4 - 7 0 - 1 6

= 22,0 kN

14

2

+

4 . 179000 17,4 - 16 702

-

I

1 10-3

=

12,6 k~

> 12,9 kN

Concluding summary The calculations are limited to those essential to the design. Only the worst beam, stringer and column cases are covered. Wind and horizontal load effects are ignored and would not normally effect the design of a low rise structure of this type.


Timber bridges STEP lecture E17 J. Fischer LIGNUM, Ziirich

Objective To describe the use of timber in modern bridges.

Summary A short history of timber bridges in Europe is given as a background to design. Structural forms and details of the principal structural elements are considered and examples shown. Timber is thus shown to be effective, economic and durable for bridge construction and to allow innovative and aesthetic solutions.

History The oldest known timber bridges go back to 600 years BC. The limited information available on these bridges shows the builders to have had excellent knowledge of timber properties and applications to structural forms. Whilst masonry bridges have survived for many centuries, these early timber bridges were mainly destroyed by war, natural disasters or fire. One of the oldest timber bridges in existence in Europe is the Kapellbriicke in Luzern. It was built in 1333 and over the centuries much of the structure has been rebuilt. Originally, the overall length was 285 m, but in the 19th century this was reduced to 222 m. The bridge is covered and is formed of simply supported beams on interconnected timber piles. In August 1993 a large part of the bridge was destroyed by fire.

Figure I

Kapellbriicke; longitudinal section.

Rebuilding of the bridge following the original form began immediately and the bridge was reopened for pedestrian use in April 1994. The supporting structure mainly consists of oak piles driven into the bed of the river Reuss. Cross-girders again of oak connect the pile caps and support the 26 spans of the main bridge structure; the average span is 7,65 m and the maximum span is about 13,5 m; the total length is now 204 m. Between 1755 and 1758 the master carpenter Hans Ulrich Grubenmann built the well-known Rhinebridge at Schaffhausen. He designed the bridge as a single span of 119 m but was forced by the town authorities to change the design and incorporate an existing central pier into the bridge. Shortly after the completion of the bridge he removed packing members over this pier and was able to demonstrate that his original concept had been possible. STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E17/1

Figure 2

Kapellbriicke after rehabilitation (1994).

Figure 3

Rhinebridge at Schaffhausen ( l755-1758) by Hans Ulrich Grubenmann.

Developments in long spanning masonry and reinforced concrete bridges were made possible by the skilled use of timber falsework which itself had to span large distances and carry heavy loads. At the turn of the century Richard Coray (1869 - 1946) constructed the falsework for some of the most important bridges in Switzerland. His structures were examples of daring and imagination:

-

the falsework of the Wiesner viaduct (190611907) cantilevered out over the 88 m deep gorge,

-

for the Sitterbriicke trusses spanned onto a scaffolding tower 96 m high,

-

for the 100 m span Langwieser viaduct an unconventional roundwood fan was developed.

Once the main structure had been built, these temporary bridges were dismantled or demolished. The development of timber bridges in the United States was in two main areas. Firstly in the use of the through truss, often covered to protect the structural timber, used in increasing lengths to carry roadways over rivers and secondly in the use of timber trestles. The latter use shows no technical advances in the use of timber but is noted for the volume of wood used in bridging and which allowed the railways to be developed quickly and cheaply through very difficult terrain without the need of highly skilled labour whilst using the raw material of the forest, through which the railway passed.


Figure 4

Langwieser viaduct 60 m above the valley.

Figure 5

Suspended bridge close to Brail (1910) used for material transports to a tunnelling building site; span 168 m, 60 m above the river Inn.

The early professional bridge builders in America were Palmer (1751-1821), Wernwag (1770-1843) and Burr (died 1822). They all patented different truss forms, mainly combining the truss with an arch and using mechanical lamination to achieve large spans. Wernwag's Collosus Bridge over the Schullylkill River at Philadelphia built around 1805 had a span of 111 m with a shallow arch forming the bottom chord of a superimposed truss. The first modern truss design not relying on an arch component was developed by Ithiel Town in 1920. His trusses had a single or double web lattice and were very simple to fabricate and build. Their use was wide spread and they soon became the most common form of American covered bridge. Other truss systems were developed, but in the 1840's Howe and Pratt both introduced trusses incorporating iron tie members. Howe's patent used diagonal STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E 1713

bracing in timber with vertical ties whereas Pratt reversed the process using vertical timbers and diagonal cross bracings. The precompression stresses available by tightening the end nuts on the iron rods had the advantage of stiffening the joints between the large section timbers improving the performance of the bridges. However the advent of iron into these trusses brought about the demise in popularity of the timber truss bridge as more elements of the truss were replaced with iron and later steel. Covered timber trusses are still around today especially in forestry areas. In many areas timber trusses are still seen as a competitive solution in bridge design. The British engineer I.K. Brunel (1806-1859) was a great believer in the structural use of timber and incorporated many timber structures in his London to Bristol Railway. However, in timber bridging he is best known for his railway viaducts built mainly in South West England and the Welsh valleys. On the main line route through Cornwall there were 43 viaducts with an aggregate span of 8 km built between 1850 and 1859. These were slender graceful structures often built on sweeping curves to bridge the deep valleys of the area at heights of up to 50 m. The viaducts followed a number of standard designs mainly incorporating fan like supports. The main beams used 300 mm by 300 mm section of yellow pine. They were often mechanically laminated using Brunel's special "joggle" or shear key to achieve greater spans. A special feature of the designs were that any stmctural member could be replaced within about an hour without disruption to service. The timbers were expected to last 30 years but as labour costs for maintenance increased the bridges were replaced and most had gone by 1940. In developing these bridges Brunel was noted for his analytical work and stmctural testing of timber elements. During recent years a real renaissance of timber bridges has taken place. One reason is the increasing interest in using such a durable and ecologically sound building material as timber. Apart from that, technological developments have contributed to new and more efficient jointing techniques and the use of woodbased materials. Timber bridges are thus shown to be effective, economic and durable and to allow innovative and aesthetic solutions.

Structural background Planning and construction Road bridges carry the traffic loads of main or secondary roads, footbridges and cycle bridges are mainly used by pedestrians and cyclists and only occasionally by maintenance vehicles or ambulances. Due to their lower loadings, footbridges and cycle bridges are mostly lighter than road bridges and are often seen to be more suited to the use of a low modulus material like timber. With regard to appearance and also to structure, open and covered bridges are distinguished as two main types of bridge. The choice of the best structural form of a bridge depends on several parameters:

-

topography and landscape, span, loading, clearance and clear width, soil conditions, architectural features.

STEPIEUROFORTECH

- an initiative under the EU Cornett Programme

Figure 6

Open bridge (left) and covered bridge (right).

Another principal distinction concerns the position of the carriageway relative to the main structure: the carriageway may be on top or at the base of the main structure. For the structure itself, a large number of forms are possible but most bridges consist of one of the basic forms shown in Figure 7 or a combination thereof.

Figure 7

Structural forms of bridges. ( a ) Beams on two or more supports, (b) trussed systems, ( c ) kings and queens post trusses, ( d ) strut frame systems, ( e ) frame systems, (f3 arch bridges, (g) suspended and cable-stayed systems, (h) chainlike structure.

Durability The reason for decay in timber bridges is nearly always poor detailing for durability and neglected maintenance. All aspects of timber protection and maintenance should therefore be considered even during the planning phase of the bridge. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E1715

Wood as a natural product is part of a life cycle of growing and decomposition. One method to break this cycle and hence preserve the timber is to keep the wood material constantly dry. This is a very effective method to preserve the load carrying capacity and the functioning of the bridge during its planned life time. This goal can be achieved using a timber protection plan relating to the elements shown in Table 1. Measure

Element

Objective

Conceptional design

Prevention or decrease of Roof or covering of the intense weather exposure main structure

Choice of material

Prevention of damage through adequate choice of materials

Use of either naturally durable or preservatively treated timbers; low moisture content during erection

Design of details

Prevent unfavourable consequences of shrinkage and swelling due to water contact

Covering of horizontal surfaces, of joints and of end grain; enable quick drying out of wet parts

Preservative treatment

Prevention of fungal or insect attack

Pressure treatment using chemical solutions

Surface treatment

Prevent weathering of Several layers of surfaces; achieve pigmented coating dimensional stability and avoid cracks; limited protection against fungi or insect attack

Table I

Timber protection plan.

Depending on the actual structure, single elements of the timber protection plan can be omitted. The strict application of the other measures then becomes more important.

Figure 8

Effective protectiorz against the influence of the weather: (a) traditional rooj ( b ) easy to replace inclined and vertical cladding, ( c ) waterproof deck with lateral cladding.

Detailing for durability Detailing for durability is by far the most effective way to protect structural timber elements. All old covered timber bridges still in use today have survived because of careful and consistent detailing for durability but often the covering is STEPIEUROFORTECH - an initiative under the EU Comett Programme

considered sacrificial in its protection of the more important structure below and can be quickly and easily replaced. When designing modern timber bridges a roof is often omitted for architectural reasons. In these cases the structural protection of the main structure has to be achieved in different ways. Examples are a deck which acts as a roof, or a covering of the main structure with lateral cladding. Preservative treatment Pressure treating sawn and glued laminated timber using CCB or CCF salts can protect the timber members against fungal attack as well as against surface decay due to the influence of the weather. It is important that preservative treatment is supported by a careful detailing of the timber structure. Protection of fasteners Apart from the timber members the metal fasteners and fittings have to be protected against corrosion. Especially from the use of salt to keep the roads free from ice and snow in the winter and the use of salts for treating the timber which leads to accelerated corrosion of steel parts. Because fasteners in timber connections are not accessible after erection of the structure, they have to be protected permanently. Hot-dip galvanising and eventually additional layers of protective coating lead to a prolonged lifetime of fasteners and metal fittings. In more aggressive environments stainless steel is preferred. Maintenance A regular check of the whole bridge should be made at intervals not exceeding three to five years. The extent of regular maintenance is small if detailing for durability of the structure has been consistently applied. The regular checks should cover the following points: -

-

-

weathering of surface coating of timber members, cracking of the timber, delamination of glued laminated timber parts, mechanical damage, fungal growth, growth of plants, collection of dirt.

Deck The deck made of planks or plates may be designed as an open or closed construction. The deck may be used for two main functions:

-

Load-carrying The deck distributes the concentrated wheel loads and transfers them to the main structure. If the deck is designed as a diaphragm, it can be used to brace the main beams and to transfer horizontal wind or brake loads.

-

Protection The deck protects the main structure from moisture and mechanical damage from traffic. An effective and durable protection of the timber structure is achieved with closed decks.

Since the 1940s timber decks have been used in USA and Canada. Naillaminated joists were used which were also connected with to main structure by nailing. Later glued laminated beams connected through shear connectors were also used. Nowadays timber decks exist in a great variety of designs utilising STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E 1717

round timbers, sawn timber, glued laminated timber or wood-based panels and sometimes as composite beams together with a concrete compression zone and wearing course. Prestressed timber decks consist of laminated joists transversely prestressed using high strength steel. Although the stiffness of timber is low perpendicular to the grain, prestressing allows a plate action in the deck. In order to avoid damaging the timber due to the high bearing stresses perpendicular to the grain under the end blocks of the prestressing steel, edge members of hardwood or steel channels are used. Prestressed timber deck plates may be either glued laminated or not glued. The shear forces between joists in non glued plates are transferred by friction made possible through prestressing. This method is very simple and economic. The joists are often treated with creosote before assembly. Gluing enables timber plate decks to have larger dimensions. Here, prestressing avoids cracking of the deck plate due to bending stresses perpendicular to the grain.

Figure 9

Unglued prestressed timber plate deck. ( a ) corrosion protected steel prestressing bar, ( b ) direction of span, ( c ) spruce/fir joist, ( d ) oak edge member.

Due to shrinkage and creep the prestressing force decreases with time. Measurements in timber decks in service, however, show that the necessary force can be maintained at an adequate level for a long time. If necessary, early in the structural lifetime, restressing can be done. Prestressing decreases moisture induced variations of timber dimensions and provides dimensionally stable members. The prestressed plate therefore is suited to serve as base for durable wearing surfaces like asphalt. Prestressed timber decks are increasingly used to replace existing decks as well as to provide a simple and effective deck structure in new timber bridges.

STEPJEUROFORTECH - an initiative under the EU Comett Programme

Figure 10

Straight and curved ( r = 20 m ) prestressed timber decks.

Modern examples of timber bridges DorfZi-Briicke Eggiwil, Switzerland (1984) Two-lane road bridge for normal highway loading consisting of double axle loads of 270 kN each and including an allowance for impact together with a uniformly distributed load of 4,2 kN/m2. The bridge is a direct replacement for the original timber structure built in 1885. It is a covered bridge designed to follow the appearance of the traditional bridges of this region. However, it represents a milestone in the development of modern bridges. Behind the traditional covered roof structure the structural framework incorporates many developments made possible by research work and testing. For the first time a 200 mm deep continuous deck structure spanning over the cross girders made from glulam has been adopted. The deck is transversely prestressed with threaded steel rods to provide plate action instead of the low stiffness of timber perpendicular to the grain. The wearing surface is isolated from the deck by protective membranes which ensure the durability of the deck and the supporting cross girders. For the first time the cross girders have been connected to the structural arches by hangers made of glued laminated beech. A high quality timber was necessitated by the high loads of up to 500 kN transferred by the hangers which are at 4,5 m centres. The beech is preserved by creosote. The hangers are connected to the arch using slotted steel plates and dowel bolts. The cross girders are built as double beams encompassing the hanger. The limited clearance, the long span and the high load has necessitated these members being built also of glued laminated beech. The glued laminated timber arches span 30,6 m carrying the 6,5 m wide carriageway and 1,5 m footpath. The clearance under the bracing members linking the arches is 5,O m.The vertfcal loads are transferred from the deck through the cross girders and the hangers into the arches. The horizontal wind loads are transferred both by the stiff plate deck and a wind girder system under the roof supported at the ends of the bridge by portal frames. The longitudinal braking forces are transferred to the foundations through the deck. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E 17/9

Erection For transportation reasons, the arch was manufactured in two halves. As with the hangers, the connection between the arch halves was made with slotted steel plates and high strength dowel bolts. The deck unit was manufactured in three parts 2,2 m wide by 30,6 m long which were glued together on site and then prestressed. Materials used Glued laminated timber (sprucelfir) Glued laminated timber (beech) Sawn timber (sprucelfir) Steel sections, plates, bolts High strength dowel bolts

112 m3 36 m 3 42 m3 13000 kg 2300

Figure 11

View of the Diiifli-Briicke, Eggiwil.

Figure 12

Erection.


Figure 13

Cross girder with hanger. ( a ) Spruce glulam arch, ( b ) beech glulam hanger, (c) spruce glulam tension tie, ( d ) beech glulam cross girder, ( e ) steel fitting for cross girder support.

Langlaujbriicke Pradella, Scuol, Switzerland (1990) This bridge is used in the summer as a footbridge and in the winter as part of a cross-country ski trail when it must carry the load of special tracked vehicles. The characteristic loadings are a uniformly distributed imposed load 4,O kN/m2, concentrated imposed load 30 kN and snow load 3,4 k ~ / m ' .The total length is 85,O m broken into three spans of 21,5, 42,O and 21,5 m. The outer spans are continuous with a suspended centre span. Four main beams at 1 m centres carry a 4,5 m wide carriageway. The beams are 200 mm wide and vary in depth between 1287 mm and 1650 mm. Cross girders (90 mm by 140 mm) link the beams at 1 m intervals and support the deck which is made from 39 mm thick laminated veneer lumber erected in 22 m lengths. This deck distributes well the concentrated load and acts as a diaphragm in resisting the horizontal forces. In order to protect the main timber structure, the deck acts as a roof. The asphalt wearing surface is isolated from the deck by a polyurethane foam layer. A vertical cladding system is used to protect the exposed faces of the beams. This cladding together with the handrails is fabricated from durable larch which renders unnecessary any preservative treatment. The lamella boarding of the cladding is used as an architectural feature achieving a ribbon-like appearance. Erection Temporary towers were erected at the hinge positions in the main beams and then the main beam units were positioned using mobile cranes. The cross girders together with the hand rail supports were prefabricated as U-shaped sections. The erection of the main beams and cross girder units was completed in four days. Finally the cladding was fixed from a travelling scaffold. Materials used Glued laminated timber Sawn timber (sprucelfir) Laminated veneer lumber Larch cladding

98 m3 22 nz3 16 m3 306 m2


Figure 14

Figure 15

View of the Langlaufbrucke Pradella.

Detail of main structure.


Bicycle and pedestrian bridge over the Simme at Wimmis, Switzerland (1989) As a focal point of a pedestrian and cycle way system this covered, transparent looking slender bridge crosses the river Simme 25 m above water level. A standard characteristic load of 4,O kN/m2, uniformly distributed, together with a snow loading on the roof of 1,8 kN/m2 was applied. Additionally a requirement for a concentrated imposed load of 35 kN to allow access for emergency vehicles was stipulated. The design was chosen in competition with steel and concrete. The position of the two concrete piers was determined by the topography and the need to provide a clear span over the river. This resulted in spans of 27,O , 54,O and 27,O m. The columns are clamped in the foundations and are connected to the main structure through concrete diaphragms at the top of the column (1,2 m by 5,2 m). Two continuous trusses at 4,4 m centres and 2,94 m deep represent the main structure. Top and bottom chords (200 mm by 700 mm) and the diagonals (240 mm by 360 mm) are manufactured of glued laminated timber and are reinforced with lateral layers of laminated veneer lumber (2 X 63 mm by 400 mm) glued to the glulam members. Depending on the forces, the vertical members are either of sawn or glued laminated timber. The width of the carriageway is 3,5 m and the clear height 2,5 m. The top and bottom chords of the two main beams are connected by cross girders 6,75 m at centres. Crossed steel diagonals together with the cross girders under the bridge deck form the bracing truss for the transfer of horizontal loads. The piers and the end supports of the bridge form the supports of this horizontal truss. Careful detailing for durability made a chemical preservation of the timber members unnecessary. The roof cantilevers 1,5 m on each side over the main trusses and a lateral cladding of the railings made from larch as a more durable timber species additionally protects the structure. The connections between the timber members are formed using nailed on steel plates and a central bolt which transfers the load between the plates and provides a perfect hinge. Additionally, specially made steel shoes were used which allow loads up to 700 kN to be transferred. Erection Although the river bed provided enough room for the preparatory erection work, this area was endangered by flooding. The deck structure of the bridge consisting of main truss bottom chords, cross girders, longitudinal secondary beams and steel cross ties was prefabricated on the ground in units 36 m long. These parts with a weight of 13000 kg were lifted into place using mobile cranes. The ends of the two parts covering the 27 m spans cantilevered into the middle span and were supported by temporary columns built up of three logs each. After that, the middle part was connected to the two end parts using steel plates and ringed shank nails. The timber deck structure then served as a working platform for the erection of the remaining bridge parts. Materials used Glued laminated timber Sawn timber Laminated veneer lumber

95 m3 51 m3 18 m3


Larch deck Larch cladding Steel sections, plates, bolts

27 m3 1150 m2 7000 kg

Figure 17

View of the bridge over the Simme at Wimmis.

Figure 18

Bridge cross-section in the middle of the span.

Figure 19

Internal view.


Figure 20

The transparent looking bridge rests on concrete piers.

Flurwegbriicke over the river Aabach, Uster, Switzerland (1986) This bridge is used as a pedestrian and bicycle bridge and is situated in a nature park. The characteristic loadings are a uniformly distributed imposed load of 4,O kN/m2 and a concentrated imposed load of 10 kN. Two main beams of glued laminated timber (200 mm by 1200 mm) carry a 3,4 m wide carriageway on the bottom chord of the glulam beams. For architectural reasons the main beams have a camber of 200 mm. The span is 14,2 m. U-shaped units of cross girders and vertical tension members made of glulam connect, and at the same time brace, the main beams at 2,4 m intervals and support the deck which is made from longitudinal beams (120 by 180 mm) and diagonally placed deck planks 70 mm thick. To avoid slipping on the deck surface, it is covered with a layer of sand in epoxy. To protect the main timber structure, structural detailing as well as chemical preservation of the glulam was used. The white fir laminations of the glued laminated timber were pressure treated before gluing. Additionally, the glulam members were coated with several layers of surface coating. On the inner side, the main beams are covered with a cladding which acts as a protection against mechanical damage and the weather. This cladding consists of pressure treated spruce planks and can easily be replaced. The water is led away from the bridge through gaps between the deck planks. Erection The whole bridge was completely assembled in the workshop. The transport of the 10000 kg structure to the building site was arranged during the night. Materials used Glued laminated timber (white fir) Sawn timber (white fir) Larch deck planks Spruce cladding

12 m3 3,6 m3 3,6 m3 61 m2


Figure 21

U-shaped bridge cross-section of the Flurwegbrucke over the river Aabach at Uster.

Figure 22

A mobile crane is used to erect the bridge.

Figure 23

Bridge resting on supports.


Drostobel-Briicke, Klosters, SWitzerland (1992) The new forest road bridge replaced a derelict concrete structure from the 1920s. The characteristic loadings are a uniformly distributed imposed load of 5,O kN/rn2 and four concentrated imposed loads of 75,O kN multiplied by an impact factor of 1,8. The length of the bridge is 18,O m and the width of the carriageway 3,6 m. The structural system is a strut frame with a continuous beam which carries the deck. The strut frame consists of diaphragms rather than beam elements. The deck is built up of closely spaced glulam beams connected with two layers of laminated veneer lumber 39 mm thick. The laminated veneer lumber distributes well the concentrated load and acts as a diaphragm in resisting the horizontal wind and bracing forces. The bridge gives an unconventional architectural image. Erection The whole bridge was completely assembled in the workshop and quickly erected using a mobile crane. Materials used Glued laminated timber (spruce) Sawn timber Laminated veneer lumber Sawn timber cladding

23 m3 2 m3 5 m3 45 m"

Figure 24

View of Drostobel-Briicke, Klosters.

Figure 25

View of the cladded strut frame.

STEPIEUROFORTECH


Footbridge and cycle bridge over the N9 motorway, Ballaigues, Switzerland (1989) The bridge is part of a forest trail for walkers. The loadings are a uniformly distributed life load of 4,O KN/m2, a snow load of 2,8 ~N/rn' and a concentrated load of 10 kN. For the traffic on the N9 it was necessary to provide a clearance of 4 3 0 m. The span of the bridge is approximately 24 m and the width 2,50 m. The local contours dictated a sloping deck and environmental considerations suggested the use of round wood and a cable-stayed structure with an inclined tower. The soil on the high side was retained by a wall of close centred concrete piles which were then used to provide support for the tower with an anchorage system for the back tie. The lower side approach ramps consisted of simply supported beams spanning 5,O m. The main bridge is also made from single beams of 4,35 m and 5,10 m supported by cross girders which are in turn supported by the inclined cables. The structural system comprises pressure impregnated round wood made from Spruce and Fir. Slots were machined into the timber to control the points at which cracks due to shrinkage will occur. Compression members are made from a single log and bending members from pairs of logs. Larch was used for the deck and handrails. The cables and bracing members are made from round reinforcing rods with turnbuckles to adjust the tension. The joints of the ties to the top of the tower, and to the cross members, are made from steel and the forces are transmitted by bearing pressure. All steel parts are hot-dipped galvanised except the parts of the main supports which are made from stainless steel (quality V4A) in order to resist corrosion from road salts in the winter. Erection During construction it was important that the traffic on the main road should not be disrupted. Construction was carried out in stages: erection of the approach ramps, erection of the tower and back tie and construction of the five deck elements of the main span starting from the higher lever. Materials used Roundwood (spruce) Sawn timber (larch) S tee1

Figure 26

View of the foot bridge, Ballaigues.


Figure 27

View of the access ramp from below.

W Figure 28

Drawing of the steel fitting at the pier head.

F o o t a n d cycle bridge over the river Neckar, Remseck, Germany (1988) The bridge is a part of a foot and cycle path system and is designed for a uniformly distributed load of 5 k ~ / m ' and 4 kN/m2, respectively, according to DIN 1072. The main feature of this innovative construction spanning 80 m is its lightness and transparency. The main structure comprises three trusses used to form an equilateral triangular system with a side length of 7,56 m. The truss member connections are made using dowelled steel-to-timber joints. The covering to the sides is made from safety glass so that the user has the impression that he is invited to pass through and while crossing he is in immediate contact with his surroundings. The covering improves the long-term durability of the timber. Erection The whole bridge was constructed on one river bank and then one end was lifted onto a barge and the span floated across the river to reach its permanent supports. Materials used Glued laminated timber (spruce) Sawn timber deck planks

296 m3 10 m3


Steel

Figure 29

View of the foot and cycle bridge over the river Neckar.

Figure 30

View into the bridge.

Figure 31

Drawing of a bridge cross-section.

Wennerbrucke over the river Mur, St. GeorgenMurau, Austria (1993) Two-lane road bridge for normal highway loading consisting of double axle loads of 250 kN each together with a uniformly distributed load of 5 kN/m 2 or alternatively, a single axle load of 600 kN.


The main structure consists of four parabolic three-hinged arches (360 mm by 1200 mm) supporting the horizontal deck by evenly spaced columns (360 mm by 360 mm). The clear arch span is 45 m and the arch height 12,s m. The total length is 85 m and thus represents the longest timber road bridge in Europe. Between the end supports of the bridge and the columns supported by the arches V-shaped columns (360 mm by 600 mm) support the deck. Together with the four longitudinal main beams (360 mm by 1000 mm) they form a strut frame and transmit part of the braking forces. The arch tops are connected to the main beams through slotted steel plates. The total width of the bridge is 8,6 m. The deck consists of prefabricated prestressed concrete elements. Between the main beams and the concrete deck an 8 mm neoprene layer is inserted. Glued-in steel rods with a diameter of 30 mm provide the connection between concrete elements and main beams. The horizontal bracing is made of steel tubes and steel ties. Erection For transportation reasons, the longitudinal main beams were manufactured in four parts and the arches in two halves. First, the eight identical parts consisting of an arch halve, one inclined column and the longitudinal beam connecting both were assembled on site. Using a mobile crane, the assembled parts weighing 21000 kg each were positioned and connected. After that the bridge parts connecting the arches with the end supports were positioned. Materials used Glued laminated timber (larch) Steel sections, plates, bolts Prestressed concrete decks

300 m3 69000 kg 180 m3

Figure 32

View of the Wennerbrucke over the river Mur, St. Georgen, Murau.

Figure 33

Bridge structure.


Figure 34

Deck seen from below with three-hinged arches, longitudinal beams, inclined columns, columns on arches and bracing.

Pedestrian and cycle bridge. Main-Donau Canal near Essing, Germany ( l987) The bridge spans the Main-Donau Canal, a main road and two secondary roads. To check the stability of the structure it was necessary to carry out a static analysis and a study of the dynamic effects induced by wind and the normal uses of the bridge. The latter study was undertaken both by analysis and by performing tests in a wind tunnel. There were two conditions to be met. First the bridge had to harmonise with its surroundings and this led to the slender shape of the bridge. Second there had to be sufficient clearance for the ships on the canal. This form was chosen to enable vertical loads to be resisted by tension action with the bridge acting like a chain. This is the first timber bridge to adopt this form of construction over such a large span. The structure consists of nine glulam beams (220 mm X 650 mm) which are grouped together in threes. These tension members run continuously over three V-shaped lattice supports and are shaped like a chain. The calculated maximum force is 4000 kN. The total length of 190 m is made up of spans of 30, 32, 73 and 35 m. At the end supports the bridge is anchored by transmitting the tension into high tensile steel wires and hence to anchor blocks. The forces in the wind braces are transmitted by diagonal tension members to the supports. In the upper part of the deck the planks are arranged diagonally and made watertight using a titanium zinc layer. The components of the joints of the main structure are made from cast iron or stainless steel. Erection The glulam tension members were prefabricated in the factory with lengths between 32 and 43,5 m. On slte they were connected by large finger joints. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Materials used Glued laminated timber (spruce) Sawn timber (larch) Diagonal decking (spruce) Deck wearing surface(ekki)

320 rn3 20 rn3 1330 rn2 665 rn2

Figure 35

View of the foot and cycle bridge at Essing.

Figure 36

Bridge cross-section.

Figure 37

Detail of trussed columns acting as supports.

STEPIEUROFORTECH

- an initiative under the EU Comett

Programme

Road bridge San Nicla, Tschlin, Switzerland (1992/93) The bridge forms part of a bypass and crosses the river Inn. The road is a oneway system but with no load limitations. The load assumptions relate to a truck with two axle loads each of 205 kN and a distributed load of 3,2 kN/m2. The bridge is an arch bridge with a U-shaped cross-section and has no roof. It employs new research information both in the field of grading of lamination using ultra sonic and in connection techniques. The glulam arches have a span of 39 m and variable cross-sections both in width and depth. At the abutments of the arch the cross-section is 800 X 1000 mm and at the crown 650 X 1500 mm. The heavily loaded cross members are vertically laminated and have beech plywood at each end. The spruce glulam and the beech plywood are connected by large finger joints, and consequently, allow the full shear strength to be developed. The different connections between the arch and hanger, arch and crossbeam, arch and support are made from a newly developed anchorage system. The end plates of the arch are fitted with steel lugs. Threaded rods are screwed into the plate and inserted into pre-drilled holes in the glulam and fixed by the injection of epoxy resin. With this system it is possible to transmit tension forces up to 500 kN. An essential requirement was to protect all timber parts from the effects of changes in the environment. The consequence was that the arch ribs, the edge beams and the cross beams were covered on top with copper and on the sides with Larch boarding. The cross beams made from Beech were coated to avoid changes in moisture content. Erection The moment resisting connection between the arch ribs at the crown was made using an 80 mm thick steel plate. The steel lugs meet like fingers and are joined by 20 mm diameter bolts arranged over the full depth of the section. The connections which can transmit tension and compression forces are joined to the end grain by the system described above. Quality assurance The client specified in the preliminary project phase a plan for the use of the bridge and for quality assurance. This plan specified a life of 100 years as there are many examples in Switzerland of timber brides of this age. Also the client engaged the Swiss Federal Institute of Technology in Zurich in order to achieve quality control in areas including production of the components, the grading of laminations, the accuracy of the dimensions of the members and the processes of gluing and prestressing of the deck.

Figure 38

View of the glulam arches of the San Nicla bridge.


Materials used Glulam Laminated veneer lumber Larch boarding Stee1

Figure 39

Cross-section of the main structure.

Figure 40

Cross girder with beech plywood reinforcements at the ends.

Figure 41

Detail of connection between hanger and arch.


E17/25

Pedestrian bridge Werdenberg, Sevelen, Switzerland over the N13 motorway (1989/90) The bridge connects two service stations. There were many conditions imposed by environmental factors, traffic, wind load, use, Local Authority regulations and last, but not least, the risk of corrosion from road salt. Live load is 4,O k ~ / m ' and 10 kN concentrated load. The snow load is 1,2 kN/m2 and the wind load 1,7 kN/m2. The total length including both access towers which are made in reinforced concrete is 103 m. The internal width is 1,8 m and the internal clearance 2,4 m. The clearance for the motorway traffic is 5,5 m. The main glulam beams running over three equal spans of 30 m have hinges in the spans to make the system statically determinate. The cross-section is 200 mm by 1419 mm tapered to larger depths over the supports. The wind bracing is in the plane of the floor and comprises a horizontal I-beam made from 45 to 63 mm thick LVL webs and flanges made from sawn timber. The cross-section of the bridge is formed as a series of box frames, spaced at 2,45 m centres, fixed to the main beams. These box frames carry the roof and are braced against lateral forces. The timber structure is protected against moisture at all exposed points using wood based panels. The traditional type of covered bridge was not considered appropriate in this instance. The chosen architectural form is more like an pavilion-like building or a look-out tower. It has a very light appearance, not solid, and so does not form a visual barrier. Because of the high risk of corrosion the structure was built with few parts made from steel. The welded parts are hotdipped galvanised and in the joints that are exposed to the weather cast iron has been used. All steel parts are additionally painted. Erection The columns and access towers are made from concrete. The wooden bridge was constructed in three parts on the ground beside the road and then lifted by two mobile cranes overnight. These parts were 43 m long and weighed 35000 kg. Materials used Sawn timber (Spruce & Fir) Glulam Roof cladding (Spruce & Fir) Three layer panels (Ash) Roundwood (Robinia) Cement bonded panels

Figure 42

37 m3 53 m3 410 m2 160 m2 72 pieces 160 m*

View of the pedestrian bridge Werdenberg, Sevelen.


Figure 43

Bridge support.

Figure 44

Internal view. . p

Figure 45

Isometric representation of the main structure.


Conclusion In Central Europe the use of timber as a structural material for bridges has increased considerably during the past ten years. Engineers, architects as well as building authorities have again discovered this traditional way of building bridges. The developments of new, efficient fastener techniques and of new wood-based structural materials allow innovative and architecturally pleasing structures. Modern timber bridges are much more than just an alternative for concrete bridges. Numerous structures recently built either to replace existing bridges or as new bridges prove this fact.

References Bener, G. (1939). Gerust- und Seilriesenbauer Richard Coray. Kommissionsverlag Bischofberger Co., Chur, Switzerland.

+

Bernasconi, A. (1993). Sperrholz im Bruckenbau. In: Holzwerkstoffe auf Fumierbasis, Schweizerische Arbeitsgemeinschaft fur Holzforschung / LIGNUM, Zurich, Schweiz, 1993. p. 219-228. Bieler, W. (1993). Fumierschichtholz fur Bruckenbauten. In: Holzwerkstoffe auf Furnierbasis, Schweizerische Arbeitsgerneinschaft fur Holzforschung 1 LIGNUM, Zurich, 1993. p. 213-218. Blaser, W. (1982). Schweizer Holzbriicken. Birkhauser Verlag, Basel. Bruninghoff, H. (1993). The Essing Timber Bridge. Structural Engineering International. Joumal of the International Association for Bridge and Structural Engineering (IABSE), Zurich. Bruninghoff, H., Samuel, S., Rampf, G., Heimeshoff, B. und Sengler, D. (1992). Brucken: Planung Konstruktion - Berechnung. Entwicklungsgemeinschaft Holzbau (EGH) in der Deutschen Gesellschaft fiir Holzforschung, Munchen. Fischer, J., Graf, E., Richter, S., Sell, J. (1995). EMPNLIGNUM-Richtlinie, Holzschutz im Bauwesen. Lignatec 111995. LIGNUM, Zurich. Gehri, E. (1985). Bedeutung von Holzwerkstoffen hoher Festigkeit fur die Quertrager-Ausbildung von Strassenbriicken. Xylorama, Tendenzen in der Holzforschung. Birkhhser Verlag, Basel. Gehri, E. (1985). Holzbriicken fur Feld- und Waldstrassen - Die QS Holzplattenbrucke. ETH, Zurich. Gehri, E. (1989). Brucken und Stege - Gestern und heute. In: Brucken und Stege aus Holz. Schweizerische Arbeitsgemeinschaft fur Holzforschung l LIGNUM, Zurich, Switzerland, p. 9-18. Gehri, E. (1989). Ungedeckte Brucken - Statische und konstruktive Betrachtungen. In: Brucken und Stege aus Holz. Schweizerische Arbeitsgerneinschaft fur Holzforschung / LIGNUM, Ziirich, Switzerland, p. 157-182. Killer, J. (1985). Die Werke der Baumeister Grubenmann. Birkhauser Verlag, Basel. Kropf, F. (1989). Langzeitverhalten von Holz im Aussenbau - Ergebnisse von Uberwachungen an Pilotobjekten. In: Brucken und Stege aus Holz. Schweizerische Arbeitsgerneinschaft fiir Holzforschung l LIGNUM, Zurich, Switzerland, p. 123-131. LIGNUM (1993). Ubergange. Holz Bulletin No 3211993. Schweizerische Arbeitsgemeinschaft fur das Holz - LIGNUM. Zurich. Natterer, J. und Merz, K. (1989). Beschrieb Holztragwerk Briicke uber die Simme zwischen Reutigen und Wimmis. In: Briicken und Stege aus Holz. Schweizerische Arbeitsgemeinschaft fur Holzforschung I LIGNUM, Zurich, Switzerland, p. 97-108. Pizio, S. (1989). Fahrbahnausbildung bei Holzbrucken. In: Brucken und Stege aus Holz. Schweizerische Arbeitsgerneinschaft fur Holzforschung / LIGNUM, Zurich, Switzerland, p. 141-156. Schickhofer, G. (1993). Eine Bogenbrucke uber die Mur. Projektbeschreibung einer Strassenbrucke

STEPIEUROFORTECH


der Bruckenklasse 1 aus Larchenbrettschichtholz. Holzforschung und Holzverwendung Nr. 211993, Graz. Sell, 3. (1989). Spezielle Schutzmassnahmen - praktische Erfahrungen mit einem umfassenden Holzschutzkonzept fur Brucken und Stege. In: Briicken und Stege aus Holz. Schweizerische Arbeitsgemeinschaft fur Holzforschung I LIGNUM, Zurich, Switzerland, p. 115-121. Stadelmann, W. (1990) Holzbriicken der Schweiz - ein Inventar. Verlag Bundner Monatsblatt, Chur. Steurer, A. (1989). Entwurfsgrundlagen und Konstruktionsprinzipien von Holzstegen. In: Brucken und Stege aus Holz. Schweizerische Arbeitsgemeinschaft fur Holzforschung / LIGNUM, Ziirich, Switzerland, p. 19-42. Unger, G. (1993). Neue Dimensionen durch Holzlejmbau. Holz Kurier Nr. 2511993, Wien. Von Biiren, Ch. (1985). Funktion und Form. Birkhauser Verlag, Basel.


Timber piles STEP lecture EI 8 J-W.G. van de Kuilen Delft University of Technology

Objectives To explain the principles for the design and the calculation of foundations with timber piles.

Summary This lecture provides background to the design of foundations with timber piles. An introduction is given on the Dutch grading rules for piles and the characteristic strength value of piles. These values can be used in design calculations. Details are given of foundation types used in the Netherlands.

Introduction Pile foundations are commonly used in areas where the load bearing capacity of the soil is insufficient. In such cases the foundation of the structure can be made of timber piles. The length of the piles varies, but the maximum available length is 23 m which is sufficient for most areas in the Netherlands but also for most expected loads. The piles are driven into the ground, in order to transfer the loads of the structure to a stronger soil layer. The piles are naturally tapered with a diameter varying between 300 mm at their top, which is directly under the structure to 110 to 160 mm at the bottom, which is placed at approximately 1,5 m below the upper level of the load bearing soil layer. Details of two foundation types widely used in the Netherlands are shown in Figure l .

Figure l

(4 (e) Two commonly used foundation methods. (a) Shear block, ( b ) foundation slab, (c) pile head, (d) type traditionally used in Rotterdam, ( e ) type traditionally used in Amsterdam.

In many areas where pile foundations are needed ground water is usually present. In those cases the top of the pile must be placed at least 500 mm below the lowest expected ground water level. The pile is then always saturated, eliminating the possibilities of decay by fungi. Species which are currently allowed to be used in foundations are Spruce, Larch and Douglas Fir. Pine is not allowed due to the large amount of sapwood, which is susceptible to bacterial deterioration, even under the ground water level. To connect the timber pile with the structure above a concrete extension pile is used. An example of such an extension is shown in Figure 2. STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E1811

Figure 2

Concrete extension pile.

This concrete extension pile bridges the gap between the pile and the structure in order to avoid part of the timber pile being placed above the ground water level. These extension piles have a maximum length of about 4 metres and in general a diameter between 230 and 350 mm, although special types with smaller dimensions are also available. The building sequence is as follows. The timber pile is hammered into the ground for most of its length. Before the last metre of the pile is hammered, the concrete extension pile is placed on top of the timber pile and the pile driving is continued until the combination of timber pile and concrete extension is at the required depth. The placement of the concrete slab and the pile driving equipment is shown in Figure 3 and 4.

Figure 3

A concrete extension pile is placed on top of the timber pile which is already hammered into the ground.

The concrete extension pile is connected to the concrete foundation slab or the floor of the structure. In Figure 5 the connection between the concrete extension pile and the foundation slab is shown. The load-carrying capacity of the timber piles in compression as well as the load carrying capacity of the soil have to be determined and a check must be made to ensure that these capacities are sufficient for the design loads. The soil is treated in a similar manner as for other structural material and design rules for the verification of the soil strength should be used. Both ultimate limit states and serviceability limit states must be verified. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Figure 4

Pile driving.

Figure 5

Connection between concrete extension pile and foundation slab.

The compression strength of timber piIes The compressive characteristic strength of piles has been determined from a sample of 95 specimens cut out of 95 piles. The sample consisted of 57 Spruce, 20 Larch and 18 Douglas Fir pieces. Measured growth characteristics in the specimens were the largest knot diameter, the sum of knot diameters and the sum of knot diameters divided by the circumference. Furthermore all specimens were tested in wet condition in accordance with practice. The strength of the Spruce specimens was a little lower than the strength of the Douglas Fir, which in turn had a little lower strength than the Larch specimens. Since the differences were small, and the main species used is Spruce it was decided to derive one characteristic value, only based on the Spruce subsample. A statistical analysis did not show any correlation between the measured growth characteristics and the compression strength. The characteristic compression strength value in wet . only the modification condition was determined to be 16,3 ~ / r n r n ~Hence, factors for load duration need to be applied. This results in the design strength values given in Table 1, based on ,y = 1,3. The design compression strength of the pile is checked at a level of 1,5 metres above the pile tip, which is supposed to be at least the length over which the pile should be driven into the load bearing layer. The strength verification of the soil is done at the pile toe.


E1813

Load duration class:

km,

Permanent

0,60

fc,o.d

(N/mm2)

7,5

Long-term Medium-term Short-term Instantaneous Table l

Design compression strength values of timber piles.

The values in Table 1 are valid for single piles. Depending on the final situation in the structure these values can be increased. A group of piles under one foundation block will act as a unit. If one pile starts to fail in compression the stiffness of that pile reduces and load is redistributed to the other piles under the foundation block, provided the stiffness of this block is high enough. In that case the piles will act as a parallel structural system and a load sharing factor can be applied in the design. According to the Dutch design code NEN 6743 "Geotechnics - Calculation method for bearing capacity of pile foundations Compression piles", the minimum number of piles acting together under a foundation block is three. For that minimum case the load sharing factor was found to be 1.1.

The quality of timber piles The quality of timber piles is regulated in a special standard, the Dutch NEN 5491:1983 "Quality requirements for timber piles". This standard regulates that only Spruce, Larch and Douglas Fir as species are allowed where a batch of spruce may contain both Spruce (Picea abies) and Fir (Abies alba). Since no relationship could be found between growth characteristics and strength the quality requirements are not very strict where growth characteristics are concerned. The major restriction on knots is that the knot area, the sum of knot diameters divided by the circumference, must be smaller than 0,5 and that the maximum permitted diameter of a knot is 1/12 of the circumference or 50 mm. Other restrictions relate to curvature, with a maximum deviation in the middle of the pile of half of the middle diameter, and fissures. Additionally permissible deviations of sizes are given for both circumference and length of the piles.

Ultimate limit state design of the soil In order to make a design for a foundation the composition and load bearing capacity of the soil must be studied. Normally this is done by means of cone penetration tests (CPT) where a cone is pressed into the ground and the resistance is measured. This resistance can be used to derive the load carrying capacity of the soil. With special equipment not only the cone resistance is recorded but also the skin friction of the soil. This friction gives information on the forces which can act on the pile shaft. These forces on the shaft can be either positive or negative. A positive skin friction contributes to the load carrying capacity of the pile, since the vertical loads can now be transferred both through friction and through normal forces at the pile toe. On the other hand if the soil layers decrease in thickness over the years due to settlement, or a sand layer is put on top of the soil, the positive skin friction can become negative. This decreases the load carrying capacity. This latter possibility is sometimes not accounted for, leading to damage to the structure.


If a study of the soil is performed, the results of that study are treated in a similar manner as any other building material. The design code for geotechnics (NEN 6740) is written as any other load and resistance factor design code. The maximum design strength of the soil is calculated as follows:

where: Ffound;rn;rel.,

is the maximum characteristic value of the soil determined from the cone penetration tests,

F'u,,~m,d

is the design maximum value of the soil,

4

is a factor taking into account the number of CPT's which have been performed. In most cases is 6 = 0,8,

Ym;l)

is the material factor which is 1,25 in case the soil strength is determined by means of cone penetration tests.

The Dutch design code takes into account the number of CPT's. If more CPT's have been performed the actual load carrying capacity is better known and thus higher loads may be applied on the foundation, or fewer piles are needed. Based on these tests the load carrying capacity can be determined giving the necessary length of the pole. The result of the CPT can be used to determine the ultimate load carrying capacity of the soil layer in which the pile toe is to be placed. One method to calculate this strength is called the Koppejan - de Beer rule. This rule provides a safe estimate of the load carrying capacity of the soil and can be found in the Dutch standard NEN 6743 Geotechnics - Calculation method for bearing capacity of pile foundations - Compression piles. The maximum design force (N) can now be calculated by multiplying this strength value by the area of the pile toe. The positive skin friction is calculated as: fskin;m;rep =

a qc

where a is a constant for a given soil and q, is the cone resistance in N/mm2 with an upper limit value. The maximum skin friction is calculated as the sum of the skin friction over the circumference of the pile and the length in the soil over which skin friction is acting. In general these design calculations are not done by the structural engineer of the building but by geotechnical engineers who are specialized in the design of foundations and who carry out cone penetration tests. The final scheme of a foundation, including the forces which may act on the pile is shown in Figure 6.

Serviceability limit state design For structures with a high stiffness deformation criteria have to be set. This means that the deflection of tRe top of the pile must be determined. If all piles under the same foundation element do not have the same deflection, damage may occur in the building. Relatively harmless damage occurs when doors can no longer be opened or closed. More harm is done when cracks in masonry STEPtEUROFORTECH

- an initiative


E 18/5

walls occur, which may lead to disintegration of (parts of) the structure. Since a total description of the phenomenon is not appropriate here, reference is made to the Dutch standard NEN 6743 "Calculation method for bearing capacity of pile foundations - Compression piles". The principle of the determination is that the deflection at ground level consist of two parts. The first part is the deflection of the top of the pile due to the elasticity of the pile and the penetration of the pile toe in the soil. For this an average force in the pile is determined based on the ratio between skin friction and normal force at the pile toe. The second part is the displacement of the whole pile due to settlement of the layers under the pile.

Figure 6

Forces acting on the timber pile after the structure is completed. (a) Structure, (h) concrete extension pile, (c) timber pile, (d) ground water level, (e) negative skin friction, (f) weak clay, (g) positive skin friction, ( h ) load bearing sand layer, (i) pile toe resistance.

Design example The structure of Figure 6 is taken as an example. The pile toe is placed 13,s m below surface level. The vertical design load F, is permanent and is 100 kN. The design strength of the timber pile for load duration class permanent is 7,s N/mm2. The pile diameter is 140 mm. A, = 15400 mm2 The load bearing capacity of the pile consists of the summation of the pile toe resistance and positive skin friction. Positive skin friction is acting over 2,l m. The average pile diameter over this 2,l m is 148 mm. There is no negative skin friction to be accounted for. 4 = 0,s. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Verification of the soil strength: The strength value of the soil at the pile toe is determined by the geotechnical engineer:

The design load carrying capacity at the pile toe becomes:

The skin friction can be added to this value: the average cone resistance over 2,l meters is 7 ~ / m m ~ . Using Equation (3) the skin friction is found:

and the design value of the load carrying capacity of the skin resistance is:

The total design load bearing capacity of the soil now becomes:

Verification of pile compression strength: The diameter A, = 15400 mm2. The compression stress is now:

A pile toe diameter of 140 mm is sufficient to carry the design load.


Timber piles are available with a length up to 23 m and are a good alternative to foundations with concrete piles.

-

Three species are allowed for timber piles (Spruce, Douglas Fir and Larch). Care must be given to situations where the ground water level may change for which special concrete extension piles are developed.

-

The characteristic strength value of timber piles has been determined based on 95 compression tests and the verification of the strength according to the EC5 is given.

-

An introduction is given to the quality requirements on timber piles. A Dutch standard contains the requirements on allowable defects and permissible deviations from preferred sizes.

References NEN 549 1:1991 Quality Requirements for Timber - Timber Piles. NEN 6740 Geotechnics - Basic requirements and loads. NEN 6743 Geotechnics - Calculation method for bearing capacity of pile foundations Compression piles.


Roundwood structures STEP lecture E19 S. Winter Beratender Ingenieur der ARGE HOLZ

Objective To give an overview of the types of structure that can be constructed from roundwood.

Summary Different uses of roundwood are shown by means of a few examples and a calculation example for a timber pole construction is given designed in accordance with EC5. Connection methods are also considered within the lecture,

Introduction The construction of buildings with roundwood is one of the oldest construction types known. Roundwood as construction material could be obtained without great effort allowing the material to be used by so-called primitive cultures. The applications cover simple log cabins up to highly developed pole and tower buildings. The famous Norwegian stave churches are excellent examples. The use of roundwood construction is increasing due to the low cost of the basic material and the low construction costs.

Applications Today roundwood construction is mostly used in agricultural buildings. There are several reasons for this:

-

Stables and sheds, storage for machines or barns for corn have no special need of airtightness and particularly in barns a ventilated construction is preferred to make the additional drying of harvest possible.

-

The farms require high volume low cost construction.

-

A lot of farms are close to woodland or own forest areas. They could use their own raw products when they only have to cut the trees, remove the bark by hand or small machines and form the connection areas with a chain-saw.

-

Agricultural organisations sell kit buildings or just instructional information together with calculations and the necessary special connectors for different types of buildings from sheds to barns.

-

The simplicity of construction allows the possibility of self-build.

Roundwood construction is also used for simple halls in other applications, e.g. for boat yards because the costs of the structure can be minimised. In house construction timber poles are not used very often. However, in regions with flooding problems, difficult soils, steep slopes or in cases of special architectural requirements, timber poles are sometimes used to provide an elevated building platform and the poles may be continued to form the structural frame. A new application is the construction of sound barriers using cheap readily available small diameter poles. The walls are anchored between cantilever steel columns or timber poles. This form of construction is also appropriate to small retaining walls. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E1911

For simple wall construction in agricultural buildings, as well as in retaining walls and noise barriers, pole diameters up to 140 mm are used, see Figure 1.

Figure 1

Pole retaining wall.

Sports halls in particular those for equestrian activities are built in roundwood. The roundwood is used for the columns and walls and in few cases additionally for roof trusses. The joints in the trusses are designed as steel-to-timber joints by incorporating steel nodes. Further applications of roundwood construction are in harbour construction, e.g. bank stabilisation or catwalks for boat landing, and in simple short span bridges for agriculture and forestry use. Roundwood is also used for look-out and observation towers particularly in woodland areas.

Material Properties Roundwood gives a very good ecological balance, because sawing is not needed. and the use of primary energy is minimised. It is necessary to remove the bark from the roundwood to minimise the risk of insect attack and for some uses the poles must be peeled to obtain constant diameters and surfaces. No sawing also means the full strength of the wood is preserved. In roundwood construction all European species of softwood are used, e.g. spruce, fir, pine, larch and Douglas Fir. If timber of high durability such as larch or Douglas Fir is used in a carefully detailed construction preservatives may not be necessary. Normally a chemical preservation treatment is used, particularly for structural usage. To improve the quantity and the penetration depth of the impregnation particularly where there is ground contact or at joints predrilled holes are used, see Figure 2. The kind of preservation for load bearing construction is carried out in accordance with "EN 335-1: Durability of wood and wood-based products - definition of hazard classes of biological attack. Part 1: General", "EN 335-2: Durability of wood and wood-based products definition of hazard classes of biological attack. Part 2: Application to solid wood" and "EN 350-2: Durability of wood and wood-based products - natural durability of wood - Part 2: Guide to natural durability and treatability of selected wood species of importance in Europe". For roundwood which is only debarked higher strength values can be used because the material is left in natural form and longitudinal fibres are not cut as in sawn timber. However, currently EC5 gives no special rules for roundwood. Also in the standards "EN 518: Structural timber - Grading. Requirements for visual strength grading standards" and "EN 519: Structural Timber - Grading. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Requirements for machine strength graded timber and grading machines" reference is only made to sawn timber, not to roundwood. Guidance could be taken from EN TC 124.210: Timber Poles for overhead lines - Requirements for strength grading standards" however this standard does not cover load-bearing construction.EN 518 is a standard for national standards. Therefore material properties could be used in accordance with the national grading standards for roundwood. In the following example roundwood class I1 as given in the German national standard DIN 4074-2 is used.

Timber Pole Structures Roundwood could be used in all fields of application of sawn timber if all joints were specially designed. One of the most economic construction forms with roundwood is the use of cantilever timber poles, and this is practised in a lot of countries all over the world. Timber poles are placed in predrilled holes and the void is filled with concrete, or occasionally only with gravel. In this way cantilever poles are created which can resist vertical and horizontal forces. The diameters of such poles are normally between 200 mm and 350 mm. The costs for foundations and bracing of the buildings are mainly reduced. Droge (1986) noted a decrease in costs of 45%. To allow this type of construction a soil is required which allows predrilling and will provide adequate characteristic values of ground pressure. Timber pole structures could be used with non-cohesive soils with a minimum medium density or cohesive soils with a stiff consistency. An example of the construction of a base to a timber pole structure is given in Figure 2.

Figure 2

Example of a base of a cantilevered timber pole (Merk, 1990). (a) Pressure treated timber pole, (b) additional perforation to facilitate penetration of preservative, (c) pressure treated board, (d) pressure treated wall planks, d = 60 mm, (e) pressure treated timbers, d = 100 mm, (f) concrete C12l1.5.

The concrete used to fill the holes should be in accordance with class C12/15 given in ENV 206 and consistency class CO given in IS0 4111. In buildings with spans up to 12 m, utilising lightweight roofing and with standard wind and snow loads, the distance between the poles will normally be 3,00 - 5,00 m. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E 19/3

The example shows how a timber pole structure is designed. The ground pressure caused by vertical loads can be calculated based on a diameter d,, = d, + 0,10 m for the ground area (symbols are given in Figure 3). To design a timber pole according to EC5 the buckling length must be known. For an elastic restraint a buckling length S, = P l (P > 2) has to be calculated. As an alternative a non-linear analysis could be carried out. In reality the rotational stiffness of the foundation depends on the rotation of the pole requiring an iterative calculation based on an estimation of the deflection of the pole head. Rotation of the foundation and with it the buckling length could be calculated using loads with a partial safety factor y = 1 and by using the national codes for engineering soil properties. After calculation of the buckling length the regulations of EC5 may be used. The following example is calculated according to "DIN 18900: Timber Pole Structures - design and construction". Some characteristic values for common soils are given in Table 1. If the soil underlying the building does not correspond with these classifications, the characteristic values of mechanical properties and density have to be determined by testing in accordance with National or European standards.

Example timber pole structure (taken from Niirnberger, 1988) Assumptions Service class 2. In many cases service class may be more appropriate. deflection of pole head W , I h 1 75 characteristic soil properties according to Table 1.

Figure 3

Symbols for timber pole structures.

Figure 4

Geometry of the building, calculation of pole A.


Characteristic soil properties compactness or consistency

kind of soil

Y

'P

l,

c'

hp,

kg ----

cosh

[kN/m3]

["l

[kN/rn2]

medium density high density

18,O

32,5

--

O,9

11,O

--

19,O

35,O

--

0,s

14,s

medium density high density

18,O

35,O

--

O,8

14,5

---

19,O

37,s

--

0,7

19,5

medium density with U > 15 high - density

20,O

32,s

--

0,9

11,0

---

22,O

-10,O

0.8 2,05

14,5 2,O

-3,6

5,0

130

4,0

5,2

sand with U 1 6 nongravel cohesivesoil gravel-sand

cohesive clay

stiff

19,O

35,O 17,s

cohesive silty clay

stiff

19,5

223

Table 1

Characteristic soil properties.

Calculations for pole A Characteristic values of actions vertical load, self-weight: vertical load, snow: vertical load, wind (suction): moment at point E caused by wind: horizontal load caused by wind:

F, F, F, M, q,

= 20,7 kN =25,6kN = -19,9 kN = 14,9 kNm = 2,73 kN/m

Materials Concrete C12115 according to ENV 206, consistency CO according to I S 0 4111, earth moisture. Softwood C24 according to EN 338 E,,,, = 7400 hVmm2 Pole diameter at bottom: d, (Minimum diameter) d, Pole diameter at point E: Hole diameter: d, Depth of hole: t Thickness of concrete: S, Thickness of bottom concrete: 4 Check of minimum criteria: 120mm 11,2~, 1 2 S, = 1,2 . 120 0,14 t = 0,14 . 1500 400 mm < d,

= 260 mm > 150 mm

= 260 mm = 500 mm = 1500 mm = 120 mm = 150 mm

I0,14t 1250 = 144 mm I 1 5 0 mm = 210 mm 2 150 mm I800 mm t,T

Soil: sand, small particle-size distribution ( U 5 6), medium density Density: y = 18,O kN/m3 Angle of internal friction: cp = 32,5" Cohesion: C' = O According to Table 1, see Nurnberger (1988) Rate of plane of rupture: b = 0,9 Volume-coefficient of passive lateral soil pressure: = 11,o

ap,


Reduction factor: v y = vc = 0,35 (Using soil values determined by tests v, should be v, = 0,25)

Geometric properties of the pole dE AE WE I, iE

= = = = =

260 mm d: - ~ 1 4 d; ' ~ 1 3 2 d: d 6 4 dE 1 4 = 26014

= 2602 .n/4 = 2603 -1~132 = 2604 n164 = 65 mm

= 53100 mm2 = 1,73 106mm3 = 224 106mm4

. .

Limiting value of fixed end moment U,

M, hR n k,

= 40 mm (estimated) =M,+uKF = 14,9 + 0,04 . (20,7-19,9 + 25,6) = qw h = 2,73 - 3,30 = ME / HR = 16,0/9,01 = 1,77 m =h, / t = l77011500 =0,217/(n+0,6) =0,217/(1,18+0,6)

= 16,O kNm

= 9,Ol kN = 1770 mm

= 1,18 > 1,0 = 0,122

Limiting value of H, HR,li, = V , k, (y I 3) 9 h

H,,

=

0,35

E (t + (b db I 4)), + vc kl

. 0,122 .

'

3

4

C'

COS^ (t + (b db I 4))2

h

0~)1 +

0

=

11,8 W > 9,01 W

Deflection of pole head U~

a tan a tan a

= U, + U,, = M E l ME,lim

= (a, + a, + a ) 10-3 = [(16,0/20,9)3+ (16,0/20,9)~+ 16,0/20,9] 10-3

.

= 0,00180

U, S

= (0,6 t + h) tan a = (0,6 1500 + 3300) 0,00180 = 7,6 mm = h + dE = 3,30 + 0,26 = 3,56 m

Y

=h/s

Condition: UK UK UK

= U, + U,, = 7,6 + 33,l = 40,3 mm

40 mm (estimated value)

therefore ME = 14,9 + 0,0407 26,4 = 16,O kNm = 16,O kNm new calculation of deflection of pole head is not necessary

.

Pole buckling length ME tan a K, S

= 16,O kNm = [(16,0 / 20,9)3 + (16,O l 20,9)~+ 16,O l 20,9] = ME l tan a = 16,O 1 0,00180 = 8890 kNm = h + dE = 3300 + 260 = 3560 mm


. 10.~= 0,00180

Design values of actions according to EC5 Load case 1: max F,,,

permanent load

.

= 1,35 20,7 = 27,9 kN

Load case 4: max F,,,

permanent load + snow = 1,35 20,7 + 1,50 25,6 = 66,4 kN =0 permanent load + snow + v wind = 1,35 20,7 + 1,SO - 25,6 - 1,50 0,60 19,9 = 48,4 kN = 13,4 kNm permanent load + v snow + wind = 1,35 20,7 + 1,50 0,7 25,6 - 1,SO 19,9 = 25,O kN

Load case 5: max F , ,

permanent load + wind = 1,0 20,7 - 1,50 * 19,9 = - 9,15 kN

Load case 2: max F , , Md

Load case 3: max F,,, M,

.

Characteristic values of material properties Softwood C24 4 = 260 mm fnl,k = 24 ~ / m m ' ',,mean = 1 1000 N/mm2

,

E,,,,

= 21 N/mm2 = 7400 N/mm2

Design values of material properties Service class 2 YM = 1,3 = 0,6 for load case 1 km", fm,d = 1 1,l ~ h r n m ~ = 9,7 ~ / m m ~ = 0,9 for load cases 2 to 4 fm,d = 24 0,9/1,3 = 16,6 N/mm2 fc.,o,d = 21 . 0,9/1,3 = 14,5 N/mm2

,

Column calculation according to EC.5: Part 1-1: 5.2.1

0, = 0,2 for solid wood k

=

0,5 (1

+

P,

A(,

-

0,5)

+

A )

= 2,712

Verification = F,,, / A

Gc.~,d Gm (oc,o,d

/ kc L , , o , ~ )+

(Gm

/ fm.J

=M,/W 5 190


governing load combination: load case 4

Soil pressure d, A, Oc,d

=db+lOOmm

=500+100

= d; n/4 = max /

= 6002 n/4

= 600 mm = 0,283 106 mm2

.

fc,d

calculation according to EC7 or NAD

Pull-out calculation of load case 5 lower radius of frustum of cone = 0,25 m = 0,50 / 2 '-U = d, 12 upper radius of frustum of cone = r, + t tan 6 = 0,25 + 1,50 tan 17,5" = 0,72 m r0 Weight calculation with volume of radius of frustum of cone and cylinder = 18,O . 1,50 . n [(0,722+ 0,72 . 0,25 + 0,252)/3 - 0,502/4] Gsoil = 16,21 kN = 23,O . E [0,502 1,50 - 0,26' (1,50 - 0,15)/4] Gconcrete = 5,13 kN = 0,43 kN Gtirnber = 6,O 0,26' 1,35 . ~ 1 4 = 16,21 + 5,13 + 0,43 = G.soiI + Gconcrere + Grirnber G[,, = 21,77 kN max F,,, = 9,15 kN < 21,77 kN

Joints Connections in roundwood construction are more difficult to manufacture than connections with sawn timber. Often roundwood can only be cut by chain-saws. The roundwood will need to be cut to facilitate splice joints and to ensure good bearing between members in compression joints. If the loads are low or the joint is used purely for location, joist hangers or thin metal plates may be used. Special joist hangers for roundwood construction with different diameters are distributed by special firms. If these metal parts have no load-bearing function or if they only act in tension they could be designed according to EC5. Joist hangers for roundwood construction which are similar to joist hangers for sawn timber are not covered by any European Standard. They have to be designed according to an European Technical Approval or to a National Technical Approval. Figure 5 shows examples of typical joints, Figure 6 shows a joint at a pole head to connect a double beam. If pole heads are connected by girders it should be noted that compression loads will be set up. Curved or pitched cambered beams should not be used in this form of construction in order to avoid lateral movements of the column heads. STEPIEUROFORTECH - an initiative under the EU Comett Programme

Figure 5

Diferent types of roundwood joints.

Figure 6

Timber Pole Construction: joint at pole head.

If the restraint of poles is not strong enough or if an additional bracing to reduce buckling length is necessary, additional diagonals could be used, see Figure 7.

Figure 7

Timber Pole Construction: Connection of diagonal.

Concluding Summary Roundwood construction enables cheap buildings particularly suited to agricultural needs but also available for many other fields of application. Using timber pole construction can reduce costs up to 45%. Roundwood structures are particularly good in their low use of primary energy and they also provide an excellent ecological balance. Most European softwood can be used bot poles STEPEUROFORTECH


E 1919

should be protected by wood preservatives where they are buried in the ground or in concrete to form foundations. Special rules for roundwood construction have yet to be provided in any European Standard.

References Drijge, G. (1986). Holzmastenbauart. Kap. 20 aus Holzbau Taschenbuch, 8. Auflage, Band l: Grundlagen, Entwurf und Konstruktion. Verlag Ernst & Sohn, Berlin, Deutschland. Nurnberger, W. (1988). Informationsdienst Holz - Zweckbauten fur die Landwirtschaft. Hrsgb.: Entwicklungsgemeinschaft Holzbau (EGH) in der Deutschen Gesellschaft fiir Holzforschung, Munchen, und Centrale Marketinggesellschaft der Deutschen Agrarwirtschaft (CMA), Bonn, Deutschland. Merk (1990). Werkzeichnung der Merk Holzbau GmbH & CO, Aichach, fur das Bauvorhaben Freilandhalle Obernach, FuBpunkt - Ausbildung, Aichach.

STEPIEUROFORTECH

- an initiative


Spatial frames and domes STEP lecture E20 J.C. Chilton University of Nottingham

Objectives To develop a basic understanding of the structural principles and behaviour of threedimensional structural forms, such as spatial frames and domes, and to describe the benefits to be gained from their use. To discuss the major design considerations and construction details for these types of structure where timber is the primary structural material.

Summary The lecture begins by introducing the concepts of three-dimensional and two-way spanning structures and their potential advantage over planar structures such as beams, arches and trusses. For space frames this is demonstrated by examining the load sharing behaviour of two beams spanning orthogonally connected at their midpoints. From this, the influence of the aspect ratio of the two spans on the efficiency of a two-way spanning structure is demonstrated and the influence of the number and location of supports is also addressed. Some single and double layer grid types are described. For domes, the highly-efficient, load-carrying capacity which derives from their overall, three-dimensional, structural form is explained. Examples of different dome geometries such as Schwedler, geodesic, lamella and rib are described and, where appropriate, the way in which these geometries relate to basic polyhedral forms. Subsequently, some design and construction considerations, particularly those associated with joints, supports and movement, are outlined for both spatial frames and domes, and reference is made to the relevant sections of STEP for guidance on detailed element and joint design. Finally, a few examples of timber space frame and dome structures are briefly described to illustrate the potential of solid and glued-laminated timber for this type of construction.

Introduction Simple beams and trusses are planar structures that span in one direction and must sustain every load that may be applied to them (including any heavy point loads) transmitting these to the two end supports. However, by appropriately connecting these planar systems with elements in the orthogonal direction, an efficient, threedimensional, load-sharing system may be produced, that ensures that a load applied anywhere in the structure is resisted by all of its component elements. If composed of beams, such a structure is classified as a single-layer grid or grillage, whereas, if formed from trusses, it is referred to as a double-layer grid, space frame or space truss. Similarly, a three-dimensional dome distributes loads more efficiently than a single arch. The behaviour of these structures is discussed in more detail below. The benefit of using domes for efficient, three-dimensional structures has been known since antiquity, e.g the 43 m span, masonry dome of the Pantheon in Rome (118-128 AD). There are also many historical examples of timber domes, e.g. the outer dome of St. Paul's Cathedral in London (1705-1708) and St. Mark's in Venice (13th. century modifications) although these are not pure timber domes but work in combination with masonry domes. However, space frames are a much more recent innovation, the pioneer of this structural type being generally recognised to be Alexander Graham Bell, the inventor of the telephone, who constructed many STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E20/ 1

experimental space frame structures in the 1900s. However, space frames were not widely used until the 1950s with the development of modular systems by, for example, Mengeringhausen, Buckrninster Fuller, Wachsmann, du Chateau, and Fentiman, and the wider availability of the electronic computer which enabled the analysis of these complex, three-dimensional structures. Despite the wide acceptance of space frame structures today the great majority are constructed from steel or aluminium with only limited numbers being fabricated using timber members.

Geometry and stability in 3 dimensions To form a stable pin-jointed structure in two dimensional space a fully triangulated structure must be formed. In three-dimensional, pin-jointed structures it is a necessary but not sufficient condition for static stability that, n23j-6 where,

(1)

n = number of bars in the structure j = number of joints in the structure 6 = the minimum number of support reactions.

A study of the stability of the Platonic polyhedra (tetrahedron, cube, octahedron, dodecahedron and icosahedron) helps in the understanding of the stability of threedimensional structures. The tetrahedron is the minimum stable, three-dimensional, pin-jointed, bar structure. It has 4 joints or nodes connected by 6 bars or members and, given the necessary support conditions, it complies with Equation 1 above having only axial forces in the bars when loads are applied at the nodes (i.e. j = 4, n = 6 and 3j - 6 = ( 3 4) - 6 = 6). The cube has 8 joints and 12 bars, thus, n = 12 but 3j - 6 = ( 3 8) - 6 = 18 and the pin-jointed cube is unstable unless additional bars are introduced between the nodes or further support reactions are inserted. In the case of the octahedron n = 12, j = 6 and 3j - 6 = (3 . 6 ) - 6 = 12 thus it is a stable pin-jointed bar structure. Following similar reasoning, the pin-jointed dodecahedron is found to be unstable as a bar structure but the icosahedron is stable. For this reason, most double-layer space grid geometries are based on linked tetrahedral or half-octahedral modules. As the vertices of the Platonic polyhedra occur on the surface of a circumscribed sphere, dome geometries are frequently based on triangulated subdivision of the faces of these polyhedra, particularly of the icosahedron.

Two-way spanning systems As mentioned above, load-sharing systems may be produced, that ensure that a load applied anywhere in the structure is resisted by all of its elements. This principle may be illustrated by considering an orthogonal grid system of two horizontal beams, of span l, and l,, connected together at their midpoints, where a vertical point load F is applied (as shown in Figure 1). Assuming that the beams have the same material and cross-sectional properties (i.e modulus of elasticity (E) and the second moment of area (Z) are the same for both) and that each beam will carry a portion of the applied load F ( F , by beam 1 and F, by beam 2 ) the midspan deflection (U)of each beam can be calculated. For beam 1 the midspan deflection

and for beam 2 the midspan deflection


The beams are connected together at their midpoint, and, from consideration of compatibility, their deflections must be equal (U,= U,). Thus

Figure l

Two-way orthogonal beam grid system (plan andperspective views). (a) Beam 2, (b) beam l , ( c ) load F.

Using the above equation and the fact that F, + F, = F, the proportion of the total load F carried by each of the two beams can be found for different span (or aspect) ratios, as shown in Table 1. This table demonstrates that, when the beams are of the same length, equal load is carried by each but that, in all other cases, the greater proportion of the load is carried by the shorter span. When the ratio of spans exceeds 2,O most of the load is carried by the shorter beam which indicates that the benefit of two-way spanning grids is usually greatest if the structure is supported in approximately square structural bays. Alternatively, it is possible to increase the stiffness of the longer beams (usually by increasing I) to even out the load distribution in situations where one span is longer than the other. Span ratio (1,/1,)

1,O

1,5

2,O

3,O

Beam 1

(F,)

0,50 F

0,77 F

0,89 F

0,96 F

Beam 2

(F21

0,50 F

0,23 F

O,11 F

0,04 F

(E and I constant, l, longer span and 1, shorter span)

Table l

Proportion of load carried by each beam for dzfferent span ratios.

Advantages of using space frames There are many benefits to be gained from the use of space grid structures some of which are outlined below:

-

all elements contribute to the load carrying capacity,

-

loads are distributed more evenly to the supports,

-

there is a wide choice of support location (discussed in more detail below),

-

deflections are reduced when compared with plane structures of similar weight (alternatively a lighter three-dimensional structure results in similar deflections),

-

the high redundancy of space grids means that, in general, failure of one or


E2013

a limited number of elements, for instance, the buckling of a compression member, does not necessarily lead to overall collapse of the structure and consequently they have good resistance to damage caused by fire, explosion or seismic activity,

-

as they are usually modular, factory-fabricated, with accurate components they are easily transportable and simple to assemble on site,

-

mechanical and electrical services and air-handling ducts are easy to install within the structural depth.

Disadvantages of using space frames There are also disadvantages to using space grids and some are given below:

-

the cost, which can be high when compared with alternative structural systems, particularly when space frames are used over short spans,

-

the number and complexity of joints can lead to longer erection times on site depending on the joint type and grid module chosen,

-

when fire protection is required it is more expensive due to the high number and relatively large surface area of the space frame elements,

-

visually, the lightweight structure can appear to be very dense when viewed from certain directions.

Grid configurations Regular shaped grids are usually adopted for both the top and bottom layers of space grids to limit the number of different member lengths in the structure. There are only three regular polygons, the equilateral triangle, square and hexagon, that completely fill a plane with a regular tiling, thus, these are the most commonly used geometries. In square, two-way grids the grid lines have members orientated in two perpendicular directions, usually either parallel to the edges of the grid or set on the diagonal, at 45" to the edges. However, plane grids of triangles and hexagons produce three-way grids with members orientated in three directions. By combining the regular polygons or by using them in combination with other polygonal shapes (e.g. triangles with squares, triangles with hexagons, squares with octagons) more complex grid geometries may be produced. In space grid structures, where two plane grids are separated by web members to form a double-layer grid, it is not necessary for the top and bottom grids to have the same configuration. Nevertheless, cost implications and ease of web member connection limit the number of common forms of double-layer grids. Some common configurations are shown in Figure 2 and are as follow:

-

square on square - where the top grid is directly above the bottom grid and the web members connect the layers in the plane of the grid lines (Figure 2a),

-

square on square offset - where the bottom grid is offset by half a grid square relative to the upper grid and web members connect the intersection points in the top and bottom grids (Figure 2b),

-

square on diagonal square - where the lower grid is set at 45O to, and is usually larger than, the top grid; again with web members connecting the intersection points on the top and bottom grids (Figure 2c),

-

triangle on hexagon - where the upper grid is triangular and the lower, more open, grid is hexagonal; again with web members connecting the intersection points on the top and bottom grids (Figure 2d).


Figure 2

Typical common double-layer grids. (a) Square on square, (b) square on square offset, (c) square on diagonal square, (d) triangle on hexagon.

Members in the lower layer of a double-layer grid are generally in tension, i.e. not vulnerable to buckling, and may, therefore, be longer than the upper compression members, permitting a more open grid geometry in that layer. Choice of grid configuration and depth will also affect the economy of the space frame due to the cost of node joints when grid spacing is small and the larger cross sections required for compression elements to avoid buckling when grid spacing is large.

Support locations The choice of the most favourable support locations will depend on the plan form of the structure and architectural considerations but the positions chosen will have a significant influence on structural efficiency. Either top or bottom node joints can be supported. For example, considering a space frame square in plan, providing supports for each edge node in either the top or bottom layer is most economical whilst having supports only at the corners greatly increases the maximum forces in the space frame members and the vertical deflections are also much bigger. Placing a few intermediate supports along each edge considerably improves the space grid performance compared with the corner supported condition whilst keeping the number of columns to a reasonable minimum. Single columns located at the middle of each side also produce an efficient support system, as the corners of the space frame are cantilevered and counter-balance the central area, consequently deflections and member forces in the middle are reduced (Makowski, 1981). To reduce deflections for each of the above support conditions, the supports can be brought in slightly from the edges of the space frame. This produces a cantilever around the whole structure allowing the opportunity to have column free elevations, if desired. Alternatively, both deflections and member forces in the space grid can be reduced by use of 'tree' supports instead of discrete columns.

Domes Domes are particularly suited to covering circular or polygonal plans. Constructed from a continuous material, such as reinforced concrete, the dome is a doublecurved synclastic shell. With suitable detailing of the connections between cladding and supporting structure a timber shell may also be formed, however, this lecture is confined to spatial dome structures where the cladding is assumed not to contribute significantly structurally, apart from resisting lateral buckling of compression elements. Timber shells are considered in detail in STEP lecture E21.

Dome geometry and stability Geometrically, the surface of a dome is usually determined by the rotation of a planar arch profile about a central vertical axis. In timber domes, the threeSTEP/EUROFORTECH - an initiative under the EU Comett Programme

E2015

dimensional curved surface is generally supported on some form of ribbed structure generated from curved or straight members (or both in combination) depending on the chosen structural geometry. Several different ways in which this geometry may be formed in practice are shown in Figure 3 and described below: -

radial curved ribs running continuously from a central compression ring to a perimeter tension ring and connected by a series of meridional rings and bracing members (e.g. Schwedler dome, Figure 3a),

-

triangulated lattice dome (Figure 3c),

-

geodesic geometry, as described by Buckminster Fuller, derived by projecting subdivisions of the faces of Platonic polyhedra (described above) onto a spherical surface (Figure 3d).

More organic curved forms can be generated using grid shells but these are beyond the remit of this lecture.

Figure 3

Typical timber dome geometries (a) radial rib dome, (b) Schwedler dome, (c) lattice and (d) geodesic dome.

If carefully selected for the expected load configurations, a two-dimensional arch form will be subject primarily to axial compression with relatively small in-plane bending moments. However, as with all compression elements, the arch is liable to out-of-plane buckling and will normally require lateral restraints at appropriate intervals. In a dome structure with radial ribs, the provision of suitably placed meridional rings and bracing elements between the ribs generates a threedimensional form with high buckling resistance. Similarly, in geodesic domes, the curved surface is generally triangulated with members in axial tension or compression connected at nodes to form a stable structure.

Analysis The analysis of spatial frames and domes is facilitated by the use of electronic computers running finite element software (usually based on the stiffness method) to evaluate member forces and deformations of the structure for critical load cases. EC5: Part 1-1: 2.3.1 & 2.3.2 This software can also be used to appraise the effect of initial imperfections in the structural geometry and second order effects, as required by EC5. The rise in popularity of these structural forms is undoubtedly due to developments in computer analysis.

Element design In small spatial structures, members may be solid timber sections but for larger structures and curved elements glulam members are normally used. As it is quite difficult and expensive to produce rigid joints between timber members, most joints in spatial timber structures are designed (or assumed in the design process) to be STEPIEUROFORTECH - an initiative under the EU Cornett Programme

pin-jointed. The individual members in most spatial frames and domes are, therefore, considered to be in axial compression or tension with only a small degree of local bending, usually associated with cladding loads. For members subject to these actions, suitable cross-sections are solid circular sections. Once the overall stability of the structure has been assessed and assured, the individual members can be designed for the combined effects of axial force and bending moment (see STEP lectures covering structural components).

Joint design The connections between the individual elements of spatial frames and domes are normally made using steel components which must be detailed to transfer axial forces whilst minimising eccentricities which induce secondary bending in the members. Guidance on the design of joints in timber structures is given in the STEP lectures covering joints. Normally, in timber space frames, some form of metal insert is provided at each end of the timber sections and it is this that is connected to the node joint. These may be designed for a particular application or a proprietary jointing system may be used.

Construction details Three-dimensional structures are particularly sensitive to dimensional inaccuracies in the constituent elements. For instance, in long span space frame roofs, one way of providing a fall for rainwater run-off is to generate a camber in the structure by using members of a slightly shorter length in the bottom grid. Therefore, it is essential that the overall length of members is accurately controlled during manufacture or that a means of adjustment is incorporated in the joint details. As noted above, pin-jointed, three-dimensional structures require a minimum of six support restraints. The location and direction of these will obviously depend on the situation of the supports within the plan of the structure and the ability of the supporting structure to provide the necessary reactions to the applied actions. Typically, long span space frame structures will have at least three vertical restraints and three horizontal restraints. For instance, if one corner of a square plan structure is restrained vertically and in both horizontal directions the space frame will need at least two additional vertical restraints to prevent rotation about a horizontal axis and one more horizontal restraint to prevent rotation about a vertical axis. Ribbed domes will usually require a central compression ring to facilitate connection of the radiating ribs, as it is difficult to devise a joint which permits all of the ribs to be joined easily at the centre of the dome. A tension ring or radial buttresses are required to resist any outward thrust from the ribs at the perimeter.

Examples of timber spatial structures To illustrate the construction details discussed above examples of some timber spatial structures are included. The sophisticated proprietary Mero steel ball node joint is shown connecting members of glulam timber in a roof structure at Mittelstadt (Figure 4a) and a bridge in Munich (Figure 4b). As a contrast a system using roundwood poles, developed by Huybers (1987) at Delft University, is also shown (Figures 5a and 5b). In this case, the joints are formed from components fabricated from steel plate and subsequently galvanised for corrosion protection. Individual components are inserted into the roundwood poles and fixed by hollow steel dowels, retained in position by wire lacing. Then the metal connectors are bolted together at the nodes. The structure shown is the roof of a 10,8 by 16,2 m span agricultural building at Lelystad in the Netherlands (Huybers et al., 1987). STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E2017

Figure 4

Joint details of a roof at Mittelstadt (left) and a bridge in Munich (right).

Figure 5

Agricultural building at Lelystad (left) and joint detail (right).

Recent examples of large span timber domes have been constructed at Northern Michigan University in Marquette (timber decked, 162,5 m span) and the Asemia Dome in Japan (140,7 m span, covered with a Teflon coated fabric membrane).


Timber spatial frames and domes provide three-dimensional, highly-efficient, load-distributing, lightweight structures for medium to long spans.

-

In the case of spatial frames and for some dome configurations, short, timber elements are connected by metal connectors to generate large structures.

-

The ratio of spans in two directions and the location of supports greatly influence the efficiency of spatial frames and domes.

-

Most timber members within these structures can be designed as pure tension or compression elements. Bending normally only being produced by local cladding loads or secondary effects due to joint eccentricities.

-

Connections between elements may be made with either proprietary node joints or purpose designed metal fasteners.

References Makowski, Z.S. (1981). Analysis, Design and Construction of Double-Layer Grids, Applied Science Publishers, England. Borrego, J. (1968). Space Grid Structures, MIT Press, Cambridge, Massachusetts, USA. Huybers P,, Ende, G. van der, Lanser J., and Reijken, S.Th. van der (1987). Opbouw en beproeving van een bouwkonstruktie uit larix rondhout te Lelystad, Report C4-86-03, Technische Universiteit Delft, Netherlands.

STEPIEUROFORTECH

- an initiative


Timber shell roof structures STEP lecture E21 C.J. Mettem TRADA Technology Ltd

Objectives To give an appreciation of the opportunities to create exciting building forms in in timber, covering large clear spans without internal supports, using as a basis the structural principles of shells.

Summary Shells are thin and stiff curved membranes which act alone to provide a complete roof, including the structure, as well as the form of the roof surface. After an introduction further explaining this concept, the lecture states the applications and advantages of timber shells. It outlines how they are constructed. The historical development of structural shells is very briefly discussed. Basic theoretical principles are given, including membrane stress theory. A numerical example reinforces this. Orthotropicity, important for timber construction, is then touched upon. In the main, the lecture concentrates upon shell forms where membrane action predominates, but shells with bending, and with ribs, are also mentioned. Constructional aspects related to the standards which support EC5 are considered. The lecture concludes with two illustrated case studies.

Introduction In the context of this lecture, a shell may be considered to be a thin and stiff curved membrane, acting alone to provide a complete roof, including the structure, as well as the form of the roof surface. The thickness of a shell roof is small compared with the other dimensions of the surface. The geometry of the form is closely prescribed, since its strength and stiffness is enhanced by its curvature. Shell forms may be of single or of double curvature, and a considerable variety of shapes is possible. In general, forms of double curvature are likely to be stiffer, for a given material thickness, than single curved types. In all instances, bending moments perpendicular to the shell surface are intentionally minimised in at least one planwise direction, through the choice of the geometrical form. In many cases, both of the orthogonal in-plane directions are associated with a curvature that leads to minimal out-of-plane moments. Normal and shear forces acting within the membrane of the shell are collected at stiffening edge beams, and are thus transmitted to the foundations of the structure. In certain cases, a small number of principal tension members may also be required to complete the stability of the shell roof.

Applications Timber shell roofs should be considered whenever an elegant, architecturally appealing solution is sought for the two-way spanning of considerable clear areas. They are particularly suitable for structures in which large or frequent gatherings of people occur. The form of the structure is very evident, from both without and within the building. Extreme lightness of structure is achieved by the technique, but to be successful, this must be accomplished through careful design, by meticulous attention to detail, and with first-class workmanship. Shell roofs should not therefore be entertained for the more mundane building purposes. Nor are they likely to be architecturally advantageous for situations STEPJEUROFORTECH - an initiative under the EU Comett Programme

E2111

where internal load-bearing walls are essential, since these destroy the sense of "openness" of the internal space. To construct a timber shell, a series of thicknesses of material are laid at an angle to one another, and are nail-glued together. This forms a membrane, which in its entirety is less orthotropic than would be the case for a unidirectional layer of timber. Traditionally, timber boards laid side-by-side have been used to form each individual layer of the shell. In some instances, strips of plywood have been used as an alternative. Nowadays, other wood-based panels or structural composites such as Laminated Veneer Lumber (LVL), as described in STEP lecture A9, might offer advantages. Edge beams are usually of glued laminated timber, although solid structural timber is sometimes used, on small shell components. Here again, LVL might also now be considered for such a function.

Historical development Early historical developments Early historical structural developments tended to favour the use of timber as a material for beams, which are clearly unidimensional elements. Indeed, there is a semantic connection between the words "beam" and "timber". At first, stone, too, was most conveniently used in this way, as monolithic columns, for example. However, as stone vaulting succeeded stone arches, two-way spanning structural forms were born. Classical, Gothic and Renaissance vaults and domes share this important characteristic with shells. The cellular, vaulted construction of the Pantheon dome in Rome, is witness to the genius of Roman engineering. Its construction is like that of a modern gridshell. It held the world record for a free-spanning dome, at 43,2 m, until the birth of modern reinforced concrete techniques. With these, comparable domes can now be built with a structural mass of some 1/50th of the unreinforced concrete or stone types. Steps leading directly to the development of modern shell roofs resulted from the convergence of two paths, nineteenth century applied mathematical progress, and the use of reinforced concrete, followed by timber for two-way spanning forms.

The Surjiace Tension Analogy The theories used for the determination of membrane forces in shells and other types of structural membrane, such as tent and pneumatic structures have their roots in a very fundamental early nineteenth century theory, relating to surface tension. This was expressed by de Laplace (1806), and had considerable importance in natural science, as well as in engineering. It was discovered that a fundamental law of capillarity is that a liquid film in equilibrium assumes a form which gives it the minimum possible area under the conditions to which it is subjected. A soap-film taken upon a plane wire ring will be exposed to equal atmospheric pressure on both sides, and obviously has its minimised area in the form of a plane. The wire ring may then be bent, so that it no longer lies in a plane, and the film will become curved. However, the film will remain the smallest possible surface which can be drawn continuously across the boundaries of the ring. A quadrangular wire ring thus arranged, and warped into the outline of the edge beams of a hyperbolic paraboloid shell, for example, will take up a soap film of exactly the shape of the structural shell surface, showing this form to be a surface of minimum area.


Hence it can clearly be seen that the basis of membrane theory is that all forces are assumed to lie within the shell surface, and no bending moments are deemed to exist. Adaptation of the theory in 1826 by Lam6 and Clapeyron, and its perfection by Airy and Love (1927), cannot be described here, but may be pursued by those mindful of the importance of engineering history.

Developments of modern theory and practice Theoretical methods of analysis for engineering shells, which include components such as boilers and other pressure vessels, as well as roof structures, have attracted many famous investigators, such as Finsterwalder (1932), Lungdren (1949) and Timoshenko (1959). A comprehensive and classic textbook, written in English, is that by Flugge (1960). The same author published extensively in German. This textbook also contains a comprehensive bibliography, which is recommended for those wishing to pursue the development of modern theory. Major advances in concrete shells began in approximately 1928, when Bauersfeld and Dischinger (1928) published papers on the theory and erection of these types. A dome shell for the planetarium was constructed in Jena, at this time, and this building was followed by other notable shells, including those in Frankfurt, Leipzig and Zurich, the latter being a hyperbolic paraboloid. The use of timber as shuttering for in-situ cast concrete shells may well have given rise to the inspiration to use timber alone. Early examples of timber shells are stated by Tottenham (1961) to have occurred in Russia, not long after the first concrete types mentioned above. These are said to have had spans in the order of 48 m, and to have used boarded membranes with mechanical fasteners. Little is now known about the fate of these structures. However, it is well known and documented that timber shell roof construction was developed and promoted extensively by the Timber Development Association (TDA) in the 1950's. This work led to a number of actual building projects, as well as laboratory prototypes. Examples include the roof built in 1957 for the Royal Wilton Carpet factory, which was formed from four 17 m span square plan hyperbolic paraboloids, and a multiple conoid roof for a railway station in Manchester, mentioned again later in this lecture. Following the principles explained above, the theory of structural shells developed very rapidly in the decades just after the mid-twentieth century. A large number of shells in timber, using forms including hyperbolic paraboloids, cylindrical and conoid shells, dome shells and elliptical paraboloids were built in the UK and other parts of Europe. All were multi-layer boarded, with naiVstaple gluing and glulam edge beams. To these theories can now be added the advantages of cheap, high powered computational facilities. Considerable experience has also been gained in the better detailing of structures to avoid biodeterioration, and in environmentally acceptable methods of ensuring durability (STEP lectures A14 and A15). The release of the engineer and architect from the constraint of having to use only those forms whose behaviour can be understood and analyzed through methods of discrete, algebraic stress analysis, has led to many exciting and very free shell roof forms, a number of good examples of these being illustrated in Holzbau Atlas (Natterer, Herzog and Volz, 1991).


Basic Theoretical Principles Shell roofs Shells constructed from cross-layered timber elements, in the manner described above, will always have a radius-to-thickness ratio approximately in the order of r/t = 120 and hence will qualify for the structural theories applicable to thin shells. The following assumptions are generally associated with fundamental theories: -

The shell material is homogenous, isotropic and linearly elastic;

-

The system behaves according to small deflection theory, thus deflections under load are sufficiently small that changes in the geometry do not alter the static equilibrium;

-

Only static loading and static structural response are considered.

The first assumption, regarding isotropicity, has been modified by certain investigators, who have shown how to handle organised orthotropic systems, such as those found in timber structures. This will be further discussed later. Tests, both in full-scale and on models, tend to support the assumptions of homogeneity and elasticity, at the levels of normal design. Small deflection theory has also been checked by timber researchers, using models and other tests. In addition to requiring that actions (loadings) are static, many of the theoretical derivations for thin shells deal only with uniform cases of loading. Lack of symmetry, or concentrated load effects, require special consideration.

Geometry Conventions must be adopted in order to define every co-ordinate in a structural shell in an unequivocal manner. As an introduction to these conventions of shell geometry, consider the typical cylindrical shell element of Figure 1. It will be noted that the principal axes X, y and z are denoted using a convention similar to that adopted by EC5 (EC5: Part 1-1: 5.1.6) for beams. The thickness is denoted by t. At t/2 there is a middle surface, which bisects the thickness of the shell. In the case of a cylindrical shell, this middle surface is straight longitudinally, and circular transversely. The middle surface has a radius r in the transverse plane. For more complex shell elements, such as bicurved surfaces, these general conventions are retained, so that the geometry is always defined by reference to the middle surface. normal axis

,longitudinal axis

cenrroidal surface middle and centroidal surface

radius of middle surface

Figure I

Cylindrical shell element.


Centroidal surfaces must also be defined. Figure 1 makes an important point, which is that in a curved shell element such as this, middle surfaces and centroidal surfaces are not necessarily coincident. In the case of an element from a cylindrical shell, the middle and centroidal surfaces are coincident in the sections (y = const) shown at the edge of the figure, but not in the sections (X = const) shown as the front face. General shell elements of arbitrary curvature will have centroidal and middle surfaces which are separate in both planes.

Stress Resultants Q,

Figure 2

The loads and stress resultants acting on a shell element.

Figure 2 shows the loads and stress resultants which may act upon a bicurved shell element of arbitrary shape. The loads p , p, and p, may be considered as uniform, and unitary, for the purposes of presenting the essential theory. The cross-hatched front edge (in Figure 2) is part of the cross section (X = const) through the shell. Since this is an element of calculus, this section is considered to have an area ds, t. The dimension ds, is deemed to approach zero, leading to a force divided by the length of section which reaches a finite limit. Such a quotient is termed a "stress resultant", having units such as kN/m for example. For analysis, the stress resultants are resolved into three components, also shown in Figure 2.

-

The normal forces N, and N,, considered positive if tensile and negative if compressive;

-

The shear forces N, and NW N, is considered positive if it points in the direction of increasing y on the side of the element where + N, would point towards increasing X. N,, is the reverse;

-

The transverse forces Q, and Q,.

The three types of stress resultant which were defined as forces, and described above, give rise to corresponding normal and shear stresses, as illustrated in Figure 3. Because of the factors (rx + z)/r, and (r, + z)/r,, the moments are not zero, even when the stresses are uniformly distributed across the thickness of the shell. Adjustment for the separation of the middle and centroidal surfaces are usually made by means of shape factors, which depend upon the geometry of the shell form. For thin shells however, such factors are seldom far from unity. The full development of the relationships between the projected membrane forces, the loads and the shape of the middle surface of the shell requires rather STEPIEUROFORTECH - an initiative under the EU Comett Programme

E2 115

lengthy explanation, and the reader is recommended to study the texts for a fuller understanding of this, especially Fliigge (op. cit. 1960). In principle, an important first step is to relate the normal forces and the normal stresses, by means of integrals of the latter. For example, the total normal force on the element ds, t, Figure 2, is given by the expression: -t

Figure 3

The stresses acting on a shell element of curvatures r, r,; cross hatched area is an element of di8erential magnitude.

Corresponding relationships exist between all six in-plane stress resultants and their respective stresses. It should be noted that the integrals show that the equality of the shear stress X, = X, does not necessarily imply equality of the shearing forces, although if t and z are small compared with r, and r,, the difference may be negligible. When the stresses are not distributed uniformly across the thickness, some of them are in reality associated with moments with respect to the centre of the section, M, and M,, and there also exist two twisting, or warping moments, M, and M,,. Since it is the intention of this lecture to present only the essentials of the more elementary membrane theory, these within-thickness moment effects are not further detailed. Furthermore, in the design of timber shells, (Keresztesy, 1966) analysts have generally tended to check by means of experimental stress analysis, for example using accurate, scaled structural models, in order to establish the domain of the simpler membrane theory, and thus avoid dealing with moment complexities. STEPEUROFORTECH - an initiative under the EU Cornett Programme

Hence, in the simplest of cases, it is possible to consider in-plane forces only, Figure 4. These are considered to be distributed uniformly over the thickness. Responses of the structure to the actions which it experiences may be taken to include the stress resultants which are used in shell theory to denote the force per unit length on the middle surface. Thus the stress resultant N, acting on the transverse section of the shell element in Figure 4, when divided by shell thickness t, yields an actual design stress, which would correspond to EC5 symbols such as G,,,,for a beam element (EC5: Part 1-1: 5.1.2).

Figure 4

Membrane stress resultants.

Membrane Stress Theory In many thin-shell analyses, the membrane theory has provided a reasonable basis for the design, except at the boundaries, where the shell is supported or stiffened. Fliigge (op. cit. 1960) describes the adaptation of membrane theory as a "spectacular simplification". He goes on to explain that: "the inadequacies of the membrane theory can be discovered by a critical inspection of the membrane solutions, without any need for first solving the bending problem - a task which ojlen enough is out of reach of the practical engineer and even of the research worker". Local restraints usually exist at boundaries; edge members need to be added to receive the reactions indicated by membrane theory, and these edge members in turn introduce their own mass and rigidities. Bending effects at such boundaries are usually evaluated by approximations, to make allowance for the effect of the edge loads and displacements on both the normal stress resultant and the introduced shell moment. Following membrane theory it can be shown by examining the equilibrium of the differential moments that the transverse forces Q, and Q, cancel out. Also by equilibrium, the two shearing forces equal one another:

Thus of the ten stress resultants belonging to the general theory, only three remain: N,,N, and N, = N,,. Three equations of force equilibrium exist, which may be used to solve these stress resultants. It should be noted that although the normal forces are indicated as positive, or tensile, in Figure 4, there are types and regions of shell where these are compressive. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E2 117

Shells of Arbitrary Shape Shells described by analysts as being 'of arbitrary shape' include all those types with bicurved surfaces, such as hyperbolic and elliptic paraboloids, often built in timber. Hyperbolic paraboloids have been especially popular in timber because certain of the generator lines giving rise to this shape can be produced as a 'ruled surface'. The development of the differential equations for the membrane stress, however, is quite complex. The principles are demonstrated on the basis of a form with simpler equations in the following. Direct stresses in shells of arbitrary shape can be computed using a differential equation attributed to A. Pucher (1934) who discovered the usefulness of Airy's stress function for the purpose. The following is a brief overview, succeeded by a simple worked example. The reader is referred to the classical literature for further study.

z

Figure 5

Shell of arbitrary shape in rectilinear co-ordinates.

Figure 5 shows a shell of arbitrary shape, defined geometrically in rectilinear coordinates. Vertical co-ordinates of the middle surface, the z plane, are expressed as a function of X and y. On the middle surface, the co-ordinates x = y = const do not meet at right angles to one another. Hence the horizontally projected plane shown at the top of the figure is constructed to reflect the co-ordinates orthogonally. To develop the differential equations, the skewed stress resultants are also projected onto this horizontal plane as shown in Figure 6. The projected membrane stress resultants are obtained by differentiation, giving the following: cos X Nx = Nx cos e -

Ny = Ny

cos 0 cos X

where tan 8 = dddy and tan

X = dddx (Figure 5).


Figure 6

Shell element with membrane stress resultants projected onto horizontal plane.

To assist with the solution of these basic equations, expressions are written which make use of an auxiliary variable, the stress function 0.This is the step based upon Airy's principle:

Hence the differential equation for Qi becomes:

in which

and where P,, Pyand g a r e loads per unit area of the horizontal projection. If it

-

is necessary only to consider vertical loads, then p, = p, = 0, and q = -pZ, another considerable simplification.

Example Consider a curved paraboloid roof shell, which is triangular in plan (see Figure 7). Shells of this type have been built in timber to form part of a series of interconnected roofs over a building (Natterer op. cit. 1991). The reason for this choice as a worked example is that such triangular paraboloid shells can be shown to perform closely in accordance with simple membrane theory, and that the particular form of the Pucher equation from the stress resultants is also simple. Each individual shell unit can be considered to be cut from a paraboloid surface which is developed from a square planform, that is with only one rise h, = h, = const. Such a paraboloid of revolution has the equation


Figure 7

Triangular paraboloid shell.

In relation to the general differential equation for the stress resultants of a membrane shell, one of the terms on the left hand side vanishes, and the coefficients of the other two become constant. Furthermore, restricting consideration to q = -pz, (vertical loading only), the differential equation simplifies to: t

This is known as the plane-harmonic equation, and its solution is stated in mathematical texts as follows:

where a is an arbitrary distance from the co-ordinates origin. Applying the equations shown above for the projected membrane stress resultants, together with the simplified differential equation one obtains:

Numerical results for a shell of this type with a = h and plotted in terms of N/ph are shown in Figure 8. The reader will find that if typical values for timber structures are entered into the parameters obtained from this diagram, N/ph, and if normal stresses are estimated using a typical thickness, say 57 mm, then the apparent membrane stresses are very low. This is usual for such structures, and is in agreement with tests, such as those reported by Keresztesy (op.cit. 1966).

Orthotropicity As is evident from the basic theoretical principles explained above, the membrane stress resultants in a shell may be either tensile or compressive, depending upon the form of the shell and the region of the membrane under consideration. It has been shown that the loads and stress resultants acting on a shell element must come into equilibrium, in accordance with the fundamental STEPfEUROFORTECH - an initiative under the EU Cornett Programme

theory. The membrane analogy suggests that an ideal shell material would be isotropic, whereas of course the individual boards of timber or wood-based material forming a timber shell roof are significantly orthotropic in their behaviour. Textbooks such as that by Bodig and Jayne (1982) give a clear introduction to the equations for the orthotropic elasticity of wood and wood composites.

Figure 8

Numerical values (N/ph) of the stress resultants of the paraboloid shell example.

Orthotropic composite systems have long been recognised in timber technology, one of the best known and longest established being plywood. Special plywoods are also known, with certain veneer types arranged such that the longitudinal axis (grain direction) is neither at O0 nor at 90°, but is positioned at some intermediate value, aimed at establishing an overall composite sheet with nearisotropic properties. These have been used in aircraft manufacture. The individual boards in a shell roof have to be laid up in a workshop, or even possibly on site, rather than in a plywood press. However they can be arranged in a similar fashion to plywood, thus providing the degree of orthotropicity required. With particular types of shell, such as hyperbolic paraboloids for example, the form generator directions may also be taken into account in planning the layup, so that each layer of boards may for example require to be either twisted or bent, but not both. Those boards forming the individual layers which receive most tensile stress are able to resist such forces without difficulty, timber being a material of high tensile strength and stiffness to mass ratio, particularly when defects such as sloping grain are eliminated by strength grading. However, unless the individual layers are thoroughly connected to one another, there would be a tendency for the pieces which are in compression to buckle about their less-stiff axis, causing a distorted and possibly unserviceable roof form. There would also be problems if layers were insufficiently joined at the edges of the shell, where strong shear connections to the edge beams may be required. Unconnected layers could inadvertently introduce twists into such beams. For these reasons, it is recognised practice substantially to join the individual layers to one another, at least with mechanical fasteners, and in most instances with both fasteners and adhesive. It STEPIEUROFORTECH

- an initiative


E21111

should be noted that in this second case, the adhesive should be regarded as the primary joining medium, and the fasteners only as a means of ensuring glue pressure, since the difference in stiffness of fasteners and adhesives renders the two systems non-additive. The following explains briefly how the effective modulus of elasticity E, at a specified angle 0 from any convenient axis can be computed, for a layered shell composite, thus enabling it to be considered as a thickness-homogenous material, obeying Hooke's Law. This procedure considerably simplifies the relationship between stresses and strains in the membrane. It depends, of course, upon the layers being properly connected, as described above. Rectilinear orthotropicity is assumed, that is to say the effect of growth ring orientation is ignored. Hence ER = E , and two of the Poisson's ratios are taken to be similar, namely v , = v., This is a simplification of practical significance, since the method of sawing to define growth ring orientation cannot usually be specified, in any case. Then, for each single, orthotropic, straight grained layer, denoting E, and E, as the moduli of elasticity parallel and perpendicular to the grain respectively, the effective modulus of elasticity at an arbitrary angle 8 to the grain is given by: W

Suitable approximate relationships are:

Then, denoting the thickness of one layer and the number of layers by t and n respectively, the effective modulus of elasticity at angle 8 of the complete shell is given by:

Where experimental stress analysis has been undertaken, either on models or on a full-sized structure, it will be necessary to relate the measured strains to equivalent membrane stresses following the above principles. This can be carried out by a simple averaging process as follows. If o, = E, E, is the stress in the top fibre, and o, = E, E, is the stress in the bottom fibre, at an arbitrary measurement co-ordinate, where E, and E, are the measured strains, then the equivalent membrane stress is given by:

Shells involving bending As explained in the introduction, shell forms are provided with curvature, in order to obtain increased stiffness, within a relatively thin membrane. In general, forms having double curvature are stiffer than single-curved types. In barrel vaulted and conoid shell forms ideal, moment-free behaviour cannot possibly be attained, since there is substantial longitudinal beam action. This is illustrated STEPIEUROFORTECH - an initiative under the EU Comett Programme

qualitatively for a barrel shell of the proportions that might be built in timber in Figure 9. Similar bending effects in timber conoids have been noted (Tottenham, op. cit. 1961). In such shells, the actions may be considered to create a response of compression, tension and shear, all in the plane of the middle surface.

In long barrel shells, beam action is significant.

Figure 9

Long barrel shells Long barrel shells may be discussed as a typical example of such forms, in which these types of bending action are important. Barrel shells have transverse sections which are segments of cylinders, or which are occasionally of other closed curved profiles, such as ellipses. They may be considered to transfer load by a combination of longitudinal beam action, and transverse arch action. Short barrels have been constructed from materials other than timber, to build aircrafthangers for example, but in these types the structural action is predominately arch-like. In long barrels on the other hand, defined arbitrarily by a ratio in the order of r/L I0,6 any remaining arch action occurs only near the crown, and the bending is very significant. For such long barrels, in-plane stresses are approximated well by

where is M, is I is V is b

the the the the

bending moment about centroidal axis. moment of inertia of shell cross section. total shear at cross section. total cross-sectional thickness of shell measured horizontally.

Figure 10 compares the bending stresses calculated by this approximation, with those given by more comprehensive barrel shell theories. It can be seen that the correspondence, for r/L = long = 0,2 is good, and that as the proportion of the shell becomes shorter, the maximum tensile stress is progressively underestimated by the simpler theory. Barrel shells have been of considerable economic importance in concrete construction, and a number of publications are available explaining their design STEPEUROFORTECH - an initiative under the EU Comett Programme

E21/13

more fully, and providing tables of coefficients for the stress resultants. These may provide a first approximation for proposed solutions in timber, bearing in mind the necessary adjustments for orthotropicity. Explanations of edge beam design procedures, and means for the satisfactory design of multiple roof forms, can also be found in such literature.

Figure I0

The bending theory approximation for barrel shells. (d) Beam method, (e) shallow-shell theory.

Conoids Conoid shells in timber were introduced in the early 1960's. Following the first example, which was a railway station in Manchester, and which is currently under consideration for listed building status and for major refurbishment, a number of similar types were built elsewhere in Britain, and also in the Netherlands, Natterer (op. cit. 1991). In the town of Yeovil, England, for example, a market was constructed using six pairs of conoids, each varying in length from 5 to 7 metres, in span from 12 to 18 metres, and in spanlrise ratios ranging from approximately 5 to 1, to 10 to 1 (Booth, 1966). Substantial research including load testing of 1:2.5 scale models, and prototype testing of laminated chord trusses, was undertaken, prior to these projects.

Stiffened Shells Diverse forms of timber shell are possible, in which the membrane is stiffened by means of ribs, arches or gridwork arrangements. Some discussion of the theories for gridwork shells may be found in the classic texts such as Fliigge (op. cit. 1960). In general, the approach is via anisotropic elasticity principles. Naturally however, with such unique forms of structure, each case must be considered on its own merits, and if favoured, then analysed as a special project. Specialist software is now available to ease the path of such investigations. For an exhibition hall in Lausanne, Switzerland, (Natterer and MacIntyre, 1993), a timber shell named as a 'polydome' was chosen. This had a 25 m span form based upon a shallow slice taken from the crown of a circular dome. It was supported at four abutment positions, which were tied through the foundations. The gridwork was formed from intersecting laminated boarding, which was assembled at ground level on site, and subsequently lifted onto simple falsework for completion of the membrane. Site assembly considerations often lead to a choice of forms other than the purer membrane types, which were described initially. The difficulties of erecting shell components on an inaccessible site, over piers in the middle of the river Thames, STEP/EUROFORTECH - an initiative under the EU Comett Programme

and the consequent necessity to break the form into prefabricated sections, was one of the reasons for the choice of stiffened timber shells in the Thames Barrier project, described in the first case study below.

Architecture There is no doubt that timber shell roofs offer the possibility of creating exciting building forms that can cover large spans, without internal supports, and with openings or glazed areas in the perimeter walls. Equally, there are the general advantages of timber structures, such as high strength and stiffness to weight ratios, ease of fabrication, and energy saving benefits both in production and in use (STEP lecture A16). Armed with a knowledge of EC5, which is a most modern design code, containing the best of information derived from tremendous international cooperation, it should be a realisable challenge for the structural engineer to participate in teamwork with the architect and others to revive the best features of these elegant and slender structural forms.

Choice of form The choice of shell form will depend upon a number of factors. Some of the more important are as follows: -

the general plan shape of the area to be covered;

-

the number and position of supports that can be accepted;

-

the position and extent of any roof lighting which is required;

-

for certain shell types or combinations, the acceptable position of ties.

Table l gives geometrical definitions for three bicurved shell forms commonly used for roofs.

Shell form

Geometrical Equation

Elliptic Paraboloid z = -x+2 - y

4

Hyperbolic Paraboloid, edges bisecting directions of generators

Hyperbolic Paraboloid, edges parallel to generators

Table 1

z=-

2

h,

x2-yZ

2c

z =2= 2 h c ab

Description

Intersections with planes, = const and y = const, both parabolae, of two different sizes

X

May be stretched in X direction to provide rectangular plan, with

Intersections with vertical planes, x = const and y = const are straight lines, the generators

Geometrical definitions of three bicurved shell forms commonly used for roofs.


A range of shell types which are known to have been used for timber construction is shown in Figure 11, whilst Table 2 shows the approximate range of economically feasible dimensions for these forms. Individual shell-form components can also be combined into multi-shell roofs. A range of possible types is illustrated in Figure 12.

Figure l 1

Range of shell types commonly used in timber construction. (a) Hyperbolic paraboloids, (6) elliptical paraboloids, (c) long barrel vault (left) and conoid (right).

Figure 12

Examples of possible types of multi-shell roofs.


Shell form

Width (span)

Hyperbolic paraboloid Elliptical paraboloid Barrel-vault shell Conoid Table 2

Length

m

m

7,2 - 24 7,2 - 30 7,2 - 20 12 - 27

7,2 - 24 7,2 - 30 12 - 36 7,2 - 12

Approximate range of economically feasible dimensions for the forms shown in Figure 11.

Construction and specifications Timber shell components normally consist of two fundamental parts. These are the shell membrane itself, and the edge beams. In some types, stiffening ribs may also be present. The membrane is constructed from two or more layers of boarding, plywood or other composite. These layers are fastened together, either in the workshop or on site, using a combination of adhesive and mechanical fasteners. Figure 13 illustrates this taking place. Glued laminated timber (glulam), other structural composites, or in the case of small shells, solid timber may be used for the edge beams.

Figure 13

Fastening of the membrane layers.

Stages and details of construction to consider are as follows.

Solid timber This is most commonly used for the cross-layered boarding of the membrane. The solid timber should be visually or machine strength graded (see STEP lecture A6). Visual grading should be carried out according to standards which fulfil the minimum requirements of EN 518 "Structural timber - Grading Requirements for visual strength grading standards". Machine strength grading must meet the requirements given in EN 519 "Structural timber - Grading Requirements for machine strength graded timber and grading machines". Boards of a thickness of only approximately 20 mm may be required, and not all visual grading rules which conform to EN 518 or grading machines which comply with EN 519 may be capable of interpretation/operation to produce such a small thickness. Enquiries of suppliers should therefore be made at an early stage in STEPiEUROFORTECH - an initiative under the EU Comett Programme

E21117

the design. The specialist timber engineering fabricator may undertake this grading himself.

Strength classes A strength class system has been established in EN 338 (see STEP lecture A7). Softwood boarding for the membrane, and if solid strength graded timber is used, the softwood for the edge beams, will be selected from one of these strength classes. This will enable the characteristic values of the material properties to be determined by the designer.

Glued laminated timber The production of glued laminated timber (glulam) and its strength classification system is described in STEP lecture A8. European standards are either available already, or shortly to be published, covering a number of essential aspects. These include classification and performance of the glulam adhesives; finger jointing standards for laminations and for complete glulam members; and production requirements including delamination and glue line shear tests for quality control purposes. The five strength classes for glulam, defined in EN 1194, will provide a means of determining the characteristic values of strength, stiffness and density for this material.

Additional aspects Shrinkage and distortion effects within the elemental constituents of the shell (membrane and edge beams) are likely to give rise to unserviceablility and even possibly to instability unless materials are correctly dried prior to stress grading, and other production processes are correctly carried out. Small deflection theory was mentioned in Basic Theoretical Principles. The performance of shell structures may be sensitive to departures by the actual constructed form from the initial, theoretical shape which was assumed in the design. It is important that throughout both the workshop prefabrication stages (if any, see below), and the construction work on site, recognition is given to this factor. Suitably strict quality control measures are thus called for, at all stages in the work processes. The significance of moisture in timber is thoroughly explained in STEP lecture A4. At least the outer layers of the shell boarding, and preferably all layers, should be formed from boards which are end-jointed in accordance with EN 385 "Finger jointed structural timber". Serious consideration should be given to the specification of preservative treatment for both the solid timber, and also for the glulam elements. This is a topic giving rise to a number of aspects and requiring cross-references to other European standards. It is well covered however in STEP lecture A15. Both glulam and shell membrane elements may be treated by vacuum, or vacuum pressure methods, after manufacture, but enquiries should be made as to the size of the facility. An alternative may be to consider the choice of a species which is sufficiently durable for the performance required under the relevant service class and exposure conditions. Cases have occurred where the choice has been to treat the shell boarding timbers, but to use a naturally durable hardwood for the glulam, and all such possibilities should be carefully investigated at the initial design stage.

Manufacture Where possible, it is recommended that manufacture in a workshop should be considered, with the shell components broken down into prefabricated portions STEP/EUROFORTECH - an initiative under the EU Comett Programme

of a size capable of being delivered to site and erected. Sometimes, it may be possible to arrange a temporary covered area which is sufficiently well controlled for manufacture adjacent to the site. Obviously, such decisions should be contemplated at the design stage, since they influence important details. Cross-layered membrane boarding is laid up on formers, which may be constructed from items such as solid timber, plywood, or tubular scaffolding. The costs of providing this formwork are not insignificant, and these must also be considered at the design planning stage. The formwork must be constructed in a sufficiently stiff, rigid manner to ensure that the correct shell shape is created, and that tight contact is obtained when the successive layers of boarding are nail-glued together. The need to achieve a reasonably balanced construction has been mentioned under Orthotropicity. Suitably chosen angles between the successive layers of boarding will ensure a membrane which is approximately isotropic through the total thickness. Not only is this desirable to comply with the structural theory, but it also ensures that the shell does not warp when changes in moisture content occur through seasonal variations. The use of thin, and relatively narrow boards in a succession of layers is also advisable, since the boards, unlike the infinitely thin theoretical generator lines, have a distinct cross-section. Even with straight line generator forms such as hypars therefore, they require to be twisted during the lay up process, in order to achieve the desired shape. For these several reasons, thicknesses of boards usually range from 12 mm to 30 mm, and widths are generally less than 100 mm. Edges may either be loosely tongued and grooved, or plain.

Examples of Timber Shell Roofs Case Study l The Thames Barrier Shell Roofs To safeguard London from the possibility of flooding, a moveable tidal barrier was constructed in the early 1980's (Johnson, 1980). The Thames Barrier is located on the river at Woolwich Reach, below the City of London. It takes the form of a series of rising sector gates, which are normally swung down, beneath the water, and retained in semi-circular cills on the river bed. Ships are able to pass between the piers above these cills. When there is a danger of surge tides flowing up the river, the 200 m span steel sector gates can be raised to check the level of the water. The project incorporates seven timber shell roofs, Figure 14, which form a major architectural feature seen everyday by many thousands, both from the river and also from the air when approaching London Heathrow. The roofs cover the flood-gate operating machinery, and are positioned on each of the concrete piers. Five of the roofs measure 19 m in height, and are 11 m wide and 24 m long. Two smaller roofs of a similar form also exist. The shells are of an archstiffened type, rather than being pure membrane shells. This was a structural decision resulting from both geometrical and constructional considerations, as explained below. The roof surfaces are of a complex double curvature, and the shape is generated by the rotation of the plan of each pier about two centres, which are defined on the elevations. The shells are of triple boarded construction, and are clad on the exterior with stainless steel. When viewing the STEPJEUROFORTECH - an initiative under the EU Cornett Programme

E21119

structures from the river, it is possible to see the laminated timber arches, and also part of the undersides, which are unclad, and obviously of timber construction. A requirement of the design brief was that prefabrication should be maximised, in order to speed the site work, since expensive crane barging, critical shipping closures and tide timetables were involved in the operation. Furthermore, it was envisaged that certain parts of the structures might be damaged by shipping collisions. In this event, components were to be demonstrable and replaceable. This proved a wise decision, since on one occasion it was necessary to put the plan into effect. Such considerations, together with the inappropriateness of the required geometry for any of the pure shell types which are described in the literature, led to the choice of the arched-ribbed method of shell consideration.

Figure 14

The timber shell roofs of the Thames Barrier.

Each roof contains four main glulam arches, which are made from a hardwood timber which is classified as 'very durable', and which has been proven by experience to be suitable for such rigorous exterior exposure conditions. Twistlaminated edge beams, having double curvature, are used to frame the shells. The roof membrane is formed from three layers of European redwood boarding. The first layer, 22 mm thick, was applied at an angle of 45" to the arches. The second layer, 32 mm thick, was laid horizontally, and the third (outer) layer, also 22 nzm thick, was crossed at an opposing 45". These softwood boards had been pressure treated with preservative, and a number of details were incorporated in the design, in order to ensure the free circulation of air in certain areas where decay might otherwise occur. As a means of confirmation and support for the several stages of structural analysis and design, which included numerous computer runs, a structural model was built and tested, Figure 15. This 1/5 scale model was an exact replica of the real structure, with everything scaled down in minute detail, including the use of the same types of timber, scaled-down fasteners, and replicas of all other important aspects. The model was load tested using dead-weight and pulley systems to represent both the permanent and the variable actions. A series of serviceability load combinations were first examined, and these were followed by heavier loadings representing several important, potentially dangerous, ultimate limit state conditions, including icing.


The shell roofs have now been in service on this important and very exposed site for approximately fifteen years. They have been functioning entirely correctly, and in one case a particular component of a shell was put to the ultimate test, requiring partial replacementJrepair, after being hit by a ship. Needless to say, no disproportionate collapse occurred.

Architects Greater London Council. Consulting Engineers Rendel, Palmer & Tritton. Specialist Timber Consultants Timber Research and Development Association.

Figure 15

Structural model of a rooJ:

Case Study 2 Nursery School, Stainz, Austria A small, but very imaginative and attractive project, using timber shell roofs, was nominated for the 1994 Awards Scheme of 'GLULAM', the European trade organisation for glued laminated timber. This nursery school is of square-cruciform plan shape. It consists of an atrium area 8 m square, to which is linked, in each corner, four 'classic' hyperbolic paraboloid shell-roofed classrooms, see Figure 16. The atrium is framed from horizontal perimeter beams in glulam, which are supported off braced, spacedcolumn posts of similar material. This post and beam system supports a glazed roof of pyramidal hipped form. The hypars themselves, which are clad in copper roofing, are formed with the normal cross-boarded construction. They use several layers, some of which run parallel to the edge beams, whilst others lie at 45" to these. Glulam is also used for the edge beams. Each roof is cross-tied at the lower points, using unobtrusive round steel linking bars. Areas of high-level lighting are incorporated into the timber-framed walls, which have sufficient stiffness to provide the necessary edge support and overturning resistance to the shells themselves. This high-level window lighting reflects onto several staged areas, which are reached by means of wooden staircases, with further space beneath. A warm, restful interior is created by means of large areas of softwood, in natural colour, together with light which is reflected from the plain wall surfaces. STEP/EUROFORTECH - an initiative under the EU Cornett Programme

E21121

The whole is brightened by means of the nursery furniture and play objects, which are finished in primary colours. The exterior, of white walls, timberwork with natural pine finishes, and attractive patinated roofing, blends perfectly with the beautiful Austrian countryside.

Figure 16

The nursery school, Stainz.

Architect Engineers

Helmut Hafner. Walter Worle and Dr. Pius Worle.

References Bodig, J. and Jayne, B.A. (1982). Mechanics of wood and wood composites. Van Nostrand Reinhold, New York. Booth, L.G. (1966). The model testing, design and construction of a timber conoid shell roof for Yeovil Cattle Market. International Conference on Space Structures, University of Surrey. Dischinger, F. (1928). Schalen und Rippenkuppeln, in Emperger, F. Handbuch fur Eisenbeton, 4th ed., Vol. 6, Berlin. Finsterwalder, U. (1932). Die Theorie der kreiszylindrischen Schalengewolbe System ZeissDywidag. International Association for Bridge and Structural Engineering Publications, Zurich. 1 (1932) p. 127-152. Flugge, W. (1960). Stresses in shells. Springer-Verlag, Berlin. Johnson, V. (1980). The Thames Barrier Roofs and other Timber Structures. J. Inst. Wood Sci. Vo1.8, December. Keresztesy, L.O. (1966). The determination of the membrane forces, bending moments and displacements in an inter-connected, prefabricated, laminated timber diamond-type shell. International Conference on Space Structures, University of Surrey. De Laplace, P.S. (1806). MCchanique CCleste, BKX, suppl. Thkorie de l'action capillaire. Love, A.E.H. (1927). A treatise on the mathematical theory of elasticity, Fourth edition. Cambridge Universiy Press. Lungdren, H. (1949). Cylindrical shells. The Danish Technical Press. The Institution of Danish Engineers, Copenhagen. Natterer, J., Herzog, T. and Volz, M. (1991). Holzbau Atlas. Rudolf Miiller, Koln. Natterer, J. and Maclntyre, J. (1993). Polydome: A timber shell, Switzerland. Structures STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Worldwide, Structural Engineering International. 2/93. Pucher, A. (1934). Uber den Spannungszustand in doppelt gekriimmten Flachen. Beton Eisen. 33 (19). Berlin. Timoshenko, S. and Woinowsky-Krieger, S. (1959). Theory of plates and shells, Second edition. McGraw-Hill, New York. Tottenham, H. (1961). The design of timber shells. Proceedings of the First International Conference on Timber Structures, Southampton.


Fatigue design for timber and wood-based materials STEP lecture E22 M.P. Ansell University of Bath

Objectives The lecture explains the nature of constant amplitude and complex fatigue loads, examines methods for establishing the fatigue characteristics of wood, describes life prediction techniques for the fatigue design of timber and wood-based materials and outlines property changes which occur in fatigue.

Summary The introduction explains basic fatigue terminology and describes fatigue loading configurations. The lives of wood-based materials subjected to fatigue at constant amplitude are represented by G-N (stress versus number of cycles to failure) or E-N (strain versus number of cycles to failure) curves which vary with the R ratio, which is the ratio of the minimum cyclic stress to the maximum cyclic stress. Constant life diagrams may be constructed from such curves and relationships such as the Goodman equation describe the form of the constant life lines. The prediction of fatigue life under complex loading conditions, for example a wind or vehicle load, is achieved by a cycle counting technique which separates the load spectrum into cycles of mean stress and stress amplitude. This information is combined with constant life information and a fatigue life prediction is achieved by performing a simple Miner's Rule summation. The fatigue design approach is being applied routinely in the development of wood composite wind turbine blades. The lecture concludes by describing the microstructural damage caused by fatigue stresses in wood.

Introduction Wooden structures are frequently subjected to dynamic loads, for example vehicle loads acting on factory floors and bridges and wind loads acting on timber roofs. These fatigue loads often cause sub-critical, microstructural damage which under more extreme conditions may lead to fatigue failure. Fatigue failure of materials occurs following the application of cyclic stress (or strain) with a peak value which is less than the static strength (or strain at failure) of the material. Laboratory fatigue tests which last for less than 104 cycles are known as low cycle fatigue tests whereas those which last longer are termed high cycle tests. The fatigue loading mode, eg. tension, compression, shear, torsion or mixed mode, Figure 1, affects the fatigue life and the failure mechanism. Metal alloys are susceptible to tensile fatigue, especially at welds and notches which results in fatigue crack propagation. Wood is much less sensitive to tensile fatigue for propagation modes across the grain and fatigue damage occurs by microstructural damage events throughout the volume of the wood, unless joints or defects initiate localised damage. Hence it is not appropriate to apply fracture mechanics to fatigue in a tensile crack opening mode across the grain, although the Paris law can be applied to crack opening modes along the grain. The Paris Law is an empirical law which describes the rate of steady state crack propagation in fatigue where,

STEPIEUROFORTECH


E2211

where a is the crack length, N is the number of cycles, C and m are materials constants and AK is the critical stress intensity range.

Figure I

Fatigue loading confgurations. (a) Axial, (b) bending, (c) combined torsional and bending, (d) torsional.

Fatigue design methods described in this lecture have been applied to solid wood and to unidirectional laminated veneer lumber (LVL) loaded along the grain.

o-N (S-N) curves and the R ratio The fatigue life of engineering materials is traditionally represented in the form stress-life (o-N or S-N) or strain-life (E-N)curves following Wohler's (1867) classic work on iron railway axles in the 1850s and 1860s. The stress (stress-controlled test) or strain (strain-controlled test) can refer to the peak value, the amplitude or the range. For the sake of simplicity o-N tests are referred to from now on. Sinusoidal or sawtooth stresses may be applied to test pieces but it is necessary to identify the loading configuration and whether the load is repeated, reversed or of variable amplitude. The waveform presented in Figure 2 depicts a repeated tensile stress.

Figure 2

Sawtooth stress versus time waveform illustrating characteristic stresses.


The characteristic stresses are defined as follows: The stress range, ACJ -- o m a x o m i n

= 20dt

The mean stress,

orneun

= 0 3 (G,,

+ omin)

The alternating stress,

G,,,

= O S (G,,

- G,,,)

The R ratio,

R = o,do,,

= (o,,,,

- ou~t)J(%ean+

The R ratio is a key to understanding the loading mode in tests involving combinations of tensile and compressive stresses, for example, R = -1 represents reversed loading. Figure 3 illustrates some typical R ratios where omin is always taken to be the most negative stress.

Figure 3

Single stress cycles depicting typical R ratios. (a) static tension, ( b ) reversed loading, (c) static compression.

Other than static tensile and compressive stresses, where R is equal to one, the R ratio identifies the fatigue loading mode as follows:

R = 1 to0 R=Oto-1 R = -1 to +oo R = +W to -1

Tension-tension. Tension-compression, Compression-tension. Compression-compression.

The reader should ponder carefully the mode transitions at R = 0 and &W. At R = -1 the loading is reversed and our,= o,, = -G,,. Wood and wood composites are evaluated at R ratios which most usefully simulate loading in the engineered product. In order to fully evaluate wood composites in fatigue they should be tested at several R ratios which will result in a set of o-N curves, Figure 4, which are usually plotted in a linear stress versus log cycles form. Similar graphs have been published by Tsai and Ansell (1990) and Bonfield and Ansell (1991) for flexural and axial fatigue, respectively, of wood composites. The schematic diagram represents regression lines through a scatter of fatigue failure points (not shown) at five R ratios and static strength values at one quarter of a cycle (log N = -0,6). Bearing in mind that stress is the independent variable and that log cycles to failure is the dependent variable, it is important that the regression analysis should be performed on data in log N-o form. It may then be plotted in the traditional o-log N format. Figure 4 demonstrates that as the R ratio approaches 1 (static stress) the fatigue life increases and that reversed STEPIEUROFORTECH - an initiative under the EU Comett Programme

E2213

loading results in the shortest fatigue lives. The fatigue life of wood decreases as the moisture content increases (Tsai and Ansell, 1990).

I

Nfl

I

I

1

l

Nj3

I

I

I

log N (cycles)

-0,6

Figure 4

l

Nj2

Set of G -log N curves for tension-tension ( R = O,1, 0,3 and 0,5) and tension-compression ( R = -0,5 and -1) cyclic stress configurations.

At this point it is worth mentioning the overriding importance of the difference between the mean static tensile strength of wood and the mean static compressive strength. Whilst the static strength in tension is considerably greater than that in compression the slope of the o-log N curves in tension-tension is considerably greater than in compression-compression. In other words wood is remarkably fatigue resistant in compression-compression until o,, approaches the static compressive strength. Once loading becomes mixed mode (R = -1 to ?c=) the tensile component of loading alters the mechanism of fatigue damage accumulation.

Constant life lines It is useful to be able to predict the fatigue life of wood at any R ratio by interpolation from o-log N data at any R ratio. This is achieved by plotting constant life or Goodman diagrams, such as Figure 5, using the o-N data from Figure 4. The lines depicted radiating from the origin are lines of constant R ratio because the constant ratio ominlomm is equal to the ratio o,,Jo,,,,.

c -c

Figure 5

'J/ /

Set of constant life lines for combinations of alternating and mean stresses resulting in constant lives of Nfl, Nf2 and Nf3 cycles. (a) Safe combination of stresses below constant life line.

STEPIEUROFORTECH



In Figure 4 each vertical line Nfl, Nfl, Nf3, representing chosen constant lives, intersects the o-N curves for the R ratios 0,5, 0,3, 0,1, -0,5 and -1. The intersection points a, b, c, d, and e for a constant life Nf3, for example, can be represented as corresponding points on the constant life diagram, Figure 5, through which the constant life line can be drawn. The values o e l , oe2 and oe3 represent the alternating stress values for which constant lives of Nfl, Nfl, NJ3 are achieved respectively. The area under each constant life line represents combinations of mean and alternating stress which will not result in fatigue failure at cycle totals less than the constant life. The Goodman equation for the constant life line on the tension side of the constant life diagram is represented as, r Oult =
1 -

1

"

Oult, tension

1

(2)

for constant life points which fall on a straight line. Other versions of this equation are possible, including the following: Gerber r

Factored Goodman

The Gerber equation represents a parabolic constant life line with the same limits as the Goodman line. The factored Goodman equation includes a safety factor M > 1 which reduces the area under the Goodman line. Wood species tested to date (Bond et al., 1993) obey Goodman relationships and two equations are required to characterize each side of the constant life diagram, (C-CJC-T and T-CIT-T). Constant life diagrams for wood are also very similar in form to those for fibre-reinforced plastics (Ansell et al., 1993). It is clear that the constant amplitude fatigue life at any R ratio and peak stress can be interpolated from constant life diagrams of the form depicted in Figure 5 . Furthermore it is only necessary to measure the static tensile and compressive strengths and to obtain the o-N curve at R = -1 to obtain the two Goodman relationships for each side of the constant life diagram, which significantly reduces time spent in fatigue testing.

Complex loads and life prediction When wood is subjected to fatigue stresses in the field it is often the case that stresses are complex (variable amplitude) rather than constant amplitude. High amplitude cycles will cause the most fatigue damage. Wood is a linear-elastic material up to a proportional limit, illustrated in somewhat exaggerated form in Figure 6, but it is usually strained to levels well below the elastic limit. However it is essential to appreciate that during fatigue wood dissipates progressively more energy through hysteresis (Hacker and Ansell, 1994) as microstructural damage accumulates, following visco-elastic and plastic deformation.


0,O

0,l

0,2

0,3

0,4

0,s

0,6

0,7 E, %

Figure 6

Stress versus strain hysteresis loop. ( a ) proportional limit, (b) yield stress, 0,2% ofiset, (d) tensile strength, (e) design tensile strength.

(C)

In order to predict fatigue life under complex load a simple summation rule such as that due to Palmgren-Miner is used,

where ni is the number of cycles of stress experienced that result in failure after Ni cycles. The complex spectrum can be split into individual cycles which are sorted into "bins" (data stores), labelled by mean stress (or strain) and range stress (or strain), by the use of a suitable counting technique in order to estimate the proportion of fatigue life used up. Rainflow counting is an appropriate technique for binning cycle counts and this method is illustrated, with reference to hysteresis, in Figure 7. (The term "rainflow" relates to the concept of rain flowing down the series of roofs represented by Figure 7(a) which is used to define loading cycles. This is difficult to visualise and is avoided in this account). The strain-time (or stress-time) history ABCDEDFGA is part of a complex spectrum which corresponds to a stress-strain hysteresis loop containing three sub-loops, two of which are elastic and one of which is elastic-plastic. Loading from A to B involves elastic and plastic deformation and the unloading from B to C is elastic. Subsequent loading from C to D exceeds the strain level experienced at B so this stage involves elastic (from C to B) and plastic (from B to D ) deformation. Unloading from D to E returns the state of strain elastically to the same strain as at C but the stress level has changed due to the plastic deformation experienced. Returning the strain to D after a small amount of inelastic relaxation has occurred closes the hysteresis loop DED. Finally the strain returns to A via the elastic loop FGF. Two open hysteresis loops remain, ADA and DED, where elastic and plastic deformation occurred, and two closed loops BCB and FGF where only elastic deformation occurred. These loops are recorded and binned in terms of their mean and range strain or stress. Each loop is related via the constant life diagram to the o-N curve for the associated mean and range stress using appropriate computer software. The accumulated number of cycles for all loops are summed using the Palmgren-Miner rule and the fatigue life can be predicted in terms of the number of passes through the load-time history. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

(a)

strain -5-4-3-2-1 0 1 2 3 4 5

l time

Figure 7

Diagram demonstrating the Rainflow Counting technique. ( a ) Complex load-time history, (b) stress-strain hysteresis loop.

Fatigue evaluation of wood composites for wind turbine blades Recent fatigue research on wood has been driven by the design of wind turbine blades of up to 45 m diameter constructed from LVL, (Ansell et al., 1991). This work has involved the evaluation of several wood species and the performance of joints (Bonfield et al., 1992). The problems associated with using wood composites in offshore wind farm locations has also been addressed (Bond and Ansell, 1993).

Fatigue damage and property changes in wood Microstructural evidence for fatigue damage in Sitka spruce in four point bending has been observed by Tsai and Ansell (1990). On the compression side of the specimen damage is initiated in the double tracheid wall of the softwood by cell wall buckling, observed by polarised l i g h optical microscopy. As fatigue cycling continues diagonal arrays of buckled zones develop across many adjacent cells until cell buckling occurs and visible compression creases form at the compression surface. The neutral axis of the wood beam moves towards the tensile surface, increasing the load on this face, and the first manifestation of beam failure is jagged tensile fracture. Hacker et al. (1994) have reported on the dynamic mechanical property changes that occur in compression-compression, tension-tension and reversed loading experiments. Changes in the dynamic modulus, hysteresis loop area, underlying creep strain and the fatigue modulus are described.


References Ansell, M.P., Hancock, M. and Bonfield, P.W. "Wood composites - the optimum fatigue resistant materials for commercial wind turbine blades". In the Proceedings of the International Timber Engineering Conference, London, 4.194-4.202, September 1991. Ansell, M.P., Bond, I.P. and Bonfield, P.W., "Constant life diagrams for wood composites and polymer matrix composites", Proceedings of the Ninth International Conference on Composite Materials (ICCM-9), Madrid, Spain, ed. Miravete, Antonio, Pub. University of Zaragoza, Woodhead Publishing Ltd., V, 692-699, 1993. Bond, I.P., Ansell, M.P. and Hacker, C.L., "Fatigue testing of wood composites for aerogenerator rotor blades. Part VIII. Statistical treatment of constant life data for design optimization", Proceedings of the 1993 European Community Wind Energy Conference, Liibeck-Travemiinde, Germany, Pub. H.S. Stephens and Associates, 137-140, 1993. Bond, I.P. and Ansell, M.P., "Effect of the marine environment on the fatigue life of wind turbine materials", Wind Engineering, 17, 2, 100-107, 1993. Bonfield, P.W. and Ansell, M.P., "The fatigue properties of wood in tension, compression and shear". J. Mat. Sci., 26, 4765-4773, 1991. Bonfield, P.W., Bond, I.P., Hacker, C.L. and Ansell, M.P., "Fatigue testing of wood composites for aerogenerator rotor blades. Part VII. Alternative wood species and joints", in "Wind Energy Conversion 1992", ed. Clayton, B.R. Pub. Mechanical Engineering Publications Ltd, 243-250, 1992. Hacker, C.L. and Ansell, M.P. "Stress-strain hysteresis loop capture in the assessment of fatigue damage accumulation in wood composites". To be published in the Proceedings of EWEC '94, the European Wind Energy Association Conference, Thessaloniki, Greece, 1994. Tsai, K.T. and Ansell, M.P., "The fatigue properties of wood in flexure". J. Mat. Sci., 22, 865-878, 1990. Wohler, M,, "Wohler's experiments on the strength of metals", Engineering, Aug. 23, 1867.


Retaining walls, sound and road barriers STEP lecture E23 C. Short Consulting Engineer

Objective Design and detailing of retaining walls, sound and road barriers of structural timber in compliance with EC5 Design of Timber Structures.

Summary This lecture covers road division barriers, acoustic attenuation barriers and earth retaining walls; three separate but related subjects requiring the design of essentially vertical elements but with differing environmental and functional requirements. The early history of these structural types are described and the basis of design is developed. The ultimate and serviceability limit states are identified and design values of actions and resulting load cases are developed. Methods of verification for design conditions are considered for members in relation to sawn timber and round pole sections. Durability, structural detailing and control in relation to retaining walls and barriers are discussed.

Retaining walls Timber retaining walls are of two main forms, either pile and board or crib.

Figure I

Simple contiguous pole wall (in plan showing retained material).

Pile and board The pile and board style is perhaps the oldest form of retaining wall and may have started as a simple row of poles or piles driven into the ground to form a contiguous wall of vertical elements. Stability was provided by the depth of penetration into the soil by the mobilisation of passive resistance in the sub soils.

A later development spaced the vertical elements and filled the gap with horizontal elements of smaller cross section. In this arrangement the earth behind the wall was retained by the horizontal elements spanning onto vertical elements which are driven into the ground to mobilise the passive resistance of the sub soils. The pole (or pile) and board style of retaining wall is a common sight today in use as temporary works on construction sites. The round small section elements are replaced by sawn boards which span horizontally onto the vertical pile. In some instances the pile is not driven deep into the ground but braced with a raking or flying shore. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E23/1

Figure 2

Plan of spaced pole and board wall.

Figure 3

Trench showing a braced pile and board wall (note reduced penetration). (a) boards, (b) pile, (c) bracing.

The design of walls retaining soils or granular material would start with the assessment of the active pressures on the wall. In some cases full slope stability must be considered. The designer can then concentrate on the design of the timber horizontal and vertical elements. Durability is the first consideration. The species selected requires consideration of the expected life of the structure and the ability of the timber to be driven through the expected soil type. It is possible for different timber species to be selected for the pile and the board. The pile must resist the impact forces during installation. In some cases steel shoes may be fitted to the toe of the pile and in most cases a steel cap is placed on the head of the pile for protection during the driving process. The durability period would be important for both pile and board.

Figure 4

Pressure diagram and loads for a retaining wall. F, = active force and F, = passive force.


The active soil pressure behind the wall varies with depth. In many cases the simple earth and hydrostatic pressure is increased by additional loads from surcharge due to vehicle loads, soil slopes rising from the top of the wall and building foundations close to the back of the wall. Careful analysis is required for the assessment of the imposed actions on the timber elements. Given the active earth pressure distribution behind the wall the boards can be designed using simple static considerations. Friction between the board and the soil behind the board can induce a vertical force in the boards and this should be checked to ensure adequate bearing capacity on the edges of the board. It should be verified that the stresses induced in the board by bending,. shear, bearing and compressive actions are less than or equal to the strength of the timber modified by the partial material factors for the service class (Class 3 high moisture content) and the load duration classes of each load combination. The piles support the loads transferred by the boards and may be assumed to behave as a simple vertical cantilever.

Design example A retaining wall is required for a retained height H of 3,00 m and the backfill bulk density, y, is 18,O k ~ / m with ~ , an internal angle of friction a of 35". There is a transient surcharge of 5,00 kN/m2 to be allowed for at the top of the wall, and the Rankine coefficient k, is (l-sina)l(l+sina) = 0,27. Strength Class is C24.

Board Earth pressures at the lowest board level are; Backfill = k, y H = 14,6 kN/m2 Surcharge = k, 5 = 1,35 kN/m2

Backfill on board F,

4,28

1,35

5,78

Surcharge on board F,

0,40

1,50

0,60

Force on board

4,68

Table I EC5: Part 1-1: 2.3.3.1 EC5: Part 1-1: 2.3.3.1

6,38

Maximum characteristic and design forces on board with b = 225 mm.

y, for unfavourable effect of permanent action is 1,35 yQfor unfavourable effect of variable action is 1,5 Pile spacing s is 1,30 m; Assume simple supports

Trial section Breadth of board Depth of board Area Second moment of area

b h A I,,

= 225 mm = 50 mm

=11200mm 2 = 2,34 106 mm4


.

EC5: Part 1-1: 3.1.5 EC5: Part 1-1: 3.1.6 EC5: Part 1-1: 3.2.4

EC5: Part 1-1: 3.2.2

EC5: Part 1-1: 2.3.3.2 EC5: Part 1-1: 2.4

EC5: Part 1-1: 5.1.6 EC5: Part 1-1: 2.2.3.2 EC5: Part 1-1: 5.2.2 EC5: Part 1-1: 5.1.7

Section modulus Z, = 93,8 103mm3 Service Class 3. Load duration class is Long-term The factor for moisture content and duration of the load duration km, = 0,55 The solid timber is assumed to be normally prepared before installation by drying to near equilibrium moisture content. Depth factor k,, = 1,25 Load on board F, = 6,38 kN (Table l ) Design shear force V, =3,2kN Design moment M , , = F, S / 8 = 1,04 kNm y M =l,?

Durability The timber would be selected from the more durable species, refer to STEP lectures A14 and A15 or subject to preservation treatment before installation and detailed to insure long life in service. Ultimate limit state Bending only about one axis

M , , = 1,04 kNm 0 m,, = M O , W = 1 1,l N/mm2 fm,d = km", fm,k kh YM = 1297 N/mm2 m,d fm,d = 0,87 < 1 Board stability is provided by the lateral pressure exerted by adjacent boards as a result of frictional transfer of vertical loads to the board edges. T, =1,5V,IA = 0,43 N/mm2 h,, = h , k / YM = 1,06 N/mm2 = 0,40 < 1 7, l f,,, Pile Earth pressure Backfill pressure at base of wall is 14,6 k ~ / m ' ;Surcharge at the base of wall is 1,35 kN/m2 F

r

rF

X (m)

M (k~m)

Backfill on pile

28,5

1,35

38,5

1

38,5

Surcharge on pile

5,28

1,50

7,93

1,5

11,9

Earth force on pile

33,8

Table 2

Forces on pile.

Pile spacing

S

46,4

50,4

= 1,3 m; H = 3,00 m; assume pile is a simple cantilever.

Trial section Breadth of pile Depth of pile Area Second moment of area Section modulus

b h A I W

= 300 mm = 350 mm = 105000 mm2

=1,07.109mm4

.

= 6,13 106 mm3

EC5: Part 1-1: 3.1.5

Service Class 3

EC5: Part 1-1: 3.1.6

Load duration class Long-term

E2314


EC5: Part 1-1: 3.2.4 EC5: Part 1-1: 3.2.2

EC5: Part 1-1: 2.3.3.2

The factor for moisture content and duration of the load km,, = 0,55 Depth factor k,, =1,0 Load on pile F, = 46,4 kN Shear force V, = 46,4 kN Moment MO,, = 50,4 kNm y, =1,3

Ultimate limit state EC5: Part 1-1: 5.1.6

Bending only about one axis

EC5: Part 1-1: 2.2.3.2

EC5: Part 1-1: 5.2.2

M , , = 50,4 kNm 0 =MO, W = 8,23 N/mm 2 fm,d = kmodfm,k kh 1 YM = 10,1 N/mm2 = 0,81 < 1 m,, l fm,d

,,,

Lateral stability is provided by the bearing friction between boards and pile in the Z direction but a check should be applied for the Y direction.

EC5: Part 1-1: 5.1.7

Crib Walls

Figure 5

Typical crib retaining wall.

The crib wall is made from a frame work of simple elements of headers and stretchers forming a collection of cells filled with well graded crushed rock or natural stone materials. The headers and stretchers lock into each other with simple halving joints, providing tensile restraint to the internal forces from the mass of rock material which would otherwise redistribute to its natural angle of repose. Some loading derives from mass of stone bearing down on the embedded timber but the individual timber elements are designed as tensile members as described in lectures B2 and B12 with appropriate adjustments for factors derived for Service Class 3.

Traffic barriers Traffic barriers are typically used to separate lanes of traffic from other vehicles heading in the opposite direction or for separation of traffic from pedestrian STEPIEUROFORTECH - an initiative under the EU Comett Programme

E2315

areas. The structural form of such barriers is a profiled, functionally continuous steel rail set at typical car bumper height on timber or steel vertical elements driven or set into the ground. The height chosen in transportation guidelines (Transport Research Laboratory, 1991) to reduce the chance of the vehicle overturning on impact. The barriers are not designed to stop vehicles instantly as they cross the barrier but slow the vehicle to a stop by absorbing the kinetic energy of the impacting vehicle, or deflecting its path back into the correct direction. The forces involved in stopping a speeding vehicle vary with the starting velocity and the deceleration rate imposed. If the structure is too stiff the deceleration rate is too high and the driver of the vehicle would be damaged, However, if the structure is not stiff enough the rate of slowing would be too low and the car would pass into the area being protected by the barrier.

Figure 6

Traffic barrier.

Based upon test information one design publication (Transport Research Laboratory, 1991) requires the post to be designed such that the following condition is satisfied: "Load applied 600 mm above paving or jinished ground level whichever is higher such that ... is acceptable if bending moment of 9000 Nm is sustained without deflection exceeding 150 mm". The design case relates to accidental damage to the post, and therefore factors which allow for accidental actions and instantaneous stress levels should be used in the verification calculations. The supports may be designed as cantilevers subjected to a design moment of 9 kNm and an appropriate shear force. Both actions are coincident at a point near to the point where the vertical support enters the ground or the foundation concrete.

Land division barriers, fences and acoustic barriers Fencing with a variety of infill rail and panel arrangements are designed in a similar way to the Acoustic Barriers described below. The infill panel weight is generally negligible and the predominant action for consideration is that derived from the wind loading. Acoustic barriers are erected between areas of high noise levels and areas which are to be protected from such noises. They are required on the boundaries of airport taxiways, aircraft test centres, motorways, and noisy industrial sites. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

The barriers are designed to either deflect upwards or absorb inwards the energy waves from the noise source. In each case the timber supports are vertical elements cast into concrete foundations set in the sub soil. The actions to be catered for derive from the weight of the absorbing material in the barrier or the deflecting vanes, and the wind force on the projected area of the barrier. In general the weight of the absorbing material is transmitted directly to the sub soil via a strip foundation and the material stability is maintained by internal ties balancing the tension forces generated.

Figure 7

Typical acoustic barrier.

The design of the vertical cantilever elements in the acoustic barrier is therefore limited to consideration of the self weight of the containing frame, deflecting vanes if present and the wind force on the barrier. National Codes of practice (BS 6399, 1995 and Building Regulations, 1991) are available until the Euronorm dealing with Actions is published. The forces generated are proportional to the square of the design velocity after adjustments are made for roughness of the terrain, height and proportions of the barrier, and the design life of the barrier. The resulting pressure from the wind is expressed in kN/m2 on the projection of the barrier profile onto a vertical plane. From the pressure the force vector can be derived and with it the design bending moment and coincident shear force. The vertical timber member can then be designed using EC5 methodology. The overall stability of the barrier and its foundation should be verified to ensure adequate factors of safety against overturning, and sliding.

References Transport Research Laboratory (1991). Highway Construction Details: Part 3: Detail PTElO Test Procedure for 125 X 90 'Z' section steel and 150 X 150 Timber posts. Her Majesty's Stationery Office, United Kingdom. BS 6399 (1995). Loading for Buildings, Part 2, Code practice for Wind Loading, British Standards Institution, London, United Kingdom. Building Regulations (1991). Part A, Wind Speed Map, Published by Department of the Environment, Dublin, Ireland.

STEPIEUROFORTECH


Structures for transmission systems STEP lecture E24 K.H. Solli The Norwegian Institute of Wood Technology

Objective To introduce and to describe the structural use of timber and glulam in tran~mis~ion masts.

Summary The lecture gives a general background to transmission systems constructed in timber and covers structural forms, preservative treatment, actions, and the different functions of masts and foundations.

Introduction Electric power and signals from telegraphs and telephones have normally been transmitted and distributed by means of overhead transmission lines. The most common supporting structures have been round poles treated with different preservatives such as creosote. During the last decades glued laminated cross sections have been used even for high voltage lines. The reasons for choosing a timber construction for this purpose can be many. A high voltage line hardly fits in to the surrounding countryside, especially if the line passes through areas of natural beauty. If the masts are made of a natural material such as timber many people might find them more acceptable. In addition to timber being a natural material, it also allows the designer great latitude in the configuration of the structure. The possibilities of getting an acceptable design are quite good. Timber constructions are also easy to transport, for example by helicopter. Hollow cross sections of glued laminated timber will in addition weigh only about 60-70% of a solid pole with the same load carrying capacity. Finally, the high electrical resistance of timber with low moisture content offers significant advantages when compared with metal structures.

Structural forms Round poles Round poles of pine have been the most common form for use as transmission lines. They are made of large trees, have often a length of 20 metres or more and a top diameter from 150 mm to 250 mm. The available length of the poles has been limited by three main factors. The length of the trees, the length of the preservative treating cylinders and the length that can be transported (truck or train).

Glued laminated masts Glued laminated masts can be produced with different cross-sections which can be solid or hollow. Figure 1 gives examples of some types of glued laminated cross-sections. Glued laminated cross-sections have the advantage compared with solid round poles that the necessary dimensions and lengths can easily be achieved without the need for extremely large trees. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E24/ 1

Figure l

Examples of different glued laminated cross sections.

Preservative treatment Round poles Poles are treated with either creosote or water-borne preservatives. As for pressure treatment of other wood products the moisture content in the poles must be lower than the fibre saturation point before treatment. Because of health and environmental risks the water-borne preservatives have to be fixed in the poles before handling. There are restrictions on the use of wood treated with creosote and some water-borne preservatives in some countries.

Glued laminated masts Glued laminated timber can be treated in two different ways, and both methods involve certain problems. One method is to treat the laminations before gluing. This method is limited to water-borne preservatives since creosote treated timber cannot be glued. To achieve a good glueline it is necessary to plane the laminations. In this way the outer parts with the best protected areas are removed, often exposing untreated heartwood. In addition glued beams are often planed to give exact dimensions and smooth surfaces, whereby the resistance against micro organisms is weakened further. The other method is pressure treatment of the glued product. This method has a limitation caused by the size of the actual treating cylinder. The penetration of the preservative can be reduced because the laminations in practice will have heartwood zones out to the edge of the cross-section. Because of the drastic swelling during treatment and corresponding shrinkage during the following drying and thereby risk of cracking, water based preservatives are avoided. Creosote, however, is ideal in this connection as the heat from the oil tends to post-cure the gluelines, giving better strength and water resistance to the gluelines.


Actions While the design of transmission structures for telephone lines etc. is based on experience and standardised systems, the masts for high voltage power lines must be designed as a normal structural member. The design must be based on the actual actions and load combinations and also the surrounding conditions which might exert an influence on the construction. The actions are divided into three main groups:

-

vertical actions (Figure 2a),

-

horizontal actions perpendicular to the line direction (Figure 2b),

-

horizontal actions parallel to the line direction (Figure 2c).

Figure 2

The three difSerent types of actions on transmission masts.

Vertical actions

Vertical actions are permanent actions of the different parts of the construction and the lines. In most of Europe actions from the weight of snow and ice on the line might become the major vertical load. The fact that the vertical actions can be applied symmetrically or asymmetrically must be taken into consideration, see Figure 3a, b and c.

Figure 3

Vertical actions. Example of symmetrical and asymmetrical actions.

Horizontal actions perpendicular to the line direction

Horizontal actions perpendicular to the line direction are mainly caused by wind on the mast construction and the line. Especially during the winter when the line is covered with thick ice and snow, the wind actions can be huge.

STEPIEUROFORTECH

- an initiative


Horizontal actions parallel to the line direction Horizontal actions parallel to the line direction are caused by unbalanced tension in the line. This can occur for example when the snow and ice fall off the line in the span on one side of a mast, but not on the other side, or when the line accidentally fails on one side. As for vertical actions the horizontal actions parallel to the line direction can occur in both symmetrical and asymmetrical combinations, see Figure 4.

Figure 4

Horizontal actions parallel to the line direction. Example of ( a ) symmetrical and ( b ) asymmetrical load combinations.

For an asymmetrical load as given in Figure 4b it is important to allow for the torsion effect on the mast construction. Depending on different conditions the components of the three main action types must be combined in different ways. For example if the transmission line goes down a steep hill, the action parallel to the line direction will give a bigger contribution to the resulting vertical action than if the line was crossing a flat area of land.

The function of the masts In a transmission line the masts can be divided into standard masts and masts with special functions. The main function of the standard masts is to carry the dead loads from the cable and the mast itself and the natural loads caused by snow and wind.

Figure 5

( a ) Horizontal actions parallel to the line direction when the line changes direction. ( b ) An example of how mast can be designed to resist lateral actions.


In addition to the standard masts a line needs different types of special masts. One of the most important types of special mast is needed when the line changes direction. Since the actions in the line direction are not parallel but form an angle at this mast, the resulting action will have a different direction compared with the standard masts. While designing these mast it is important to consider the lateral actions and stability. If a cable failure occurs in the line it will introduce an impact load effect. The result of this effect might be the same as the chain reaction made by falling domino pieces. In order to avoid too many masts failing by this domino effect, it is necessary to have specially designed masts placed in the line at frequent intervals. These special masts (see Figure 6) must be designed to accommodate the actual impact loads.

Figure 6

Example of a mast designed to accommodate the impact actions parallel to the line.

Foundations The foundations for timber poles and masts depend on the ground conditions, but they are normally embedded in the ground. This method provides great savings when compared to the expensive foundations required for other types of construction. It is important to design the foundation detail having regard to shear forces, see Figure 7. The static model in the foundation must in each case be determined according to the actual foundation method. F

Figure 7

Example of moment and shear forces in the foundation.

It is also important to remember that some of the load combinations might have actions with an upward direction i.e. trying to lift the timber member from the ground. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E2415

Lateral restraint The main part of a transmission structure is normally one or several vertical members fixed at one end, i.e. the foundation. Since many of the actions have horizontal components, lateral restraint to avoid deflections, instability etc. is needed. This can be achieved by using oblique bracing members of timber or steel, or by guys. The restraint system should also take care of the special deflection, torsion, which might be imposed on a structure by asymmetric actions.

Design process In the design process all parts of the transmission mast must be controlled against the worst combination of actions for each specific part. Since this combination often varies from mast to mast in the same line, each structure normally needs to be designed separately. Important stages in the design process are:

The main structure The main structure must be controlled against bending moment, buckling, shear forces and torsion. The effects of asymmetrical actions will be different on a pole structure compared with a double mast structure with regard to torsion.

The foundation The foundation must be controlled against bending moment and shear forces. If the structure is embedded in the ground the foundation system should also be designed with regard to decay hazard, even if the structure is treated with preservatives. If the foundation structure is performed with connector details, the effects from torsion of the main structure must be regarded. The moisture content is normally very high in this part of the structure.

Mechanical connectors All connectors and joints must be controlled against actual load combinations. High humidity and influence from pollution (acid rain) might cause corrosion of mechanical fasteners. Untreated heartwood is often exposed in the holes for the fasteners, and this may involve decay hazard. All details round the points of line suspension must be designed with caution.

Cross bars The cross bars must be controlled against actual load combinations. Be aware of the possibilities of forced actions caused by different deflections in the main structure, for example effects from asymmetrical loads on a double mast structure.

Bracing The design of the bracing system is normally based on axial stresses. For members which resist compression, buckling must be considered.

Transportation and erection Extraordinary actions caused by transportation and erection of the structure or part of it must be considered.

STEPIEUROFORTECH



Concrete forming STEP lecture E25 H. Hart1 Zivilingenieur fur Bauwesen

Objective To present information about different concrete forming systems and their use in practice.

Summary The lecture gives a brief description of concrete formwork. The essential components of all formwork, that is, the shell, or lining, and the supporting structure are described. The influence of different materials on the concrete surface is explained. Different formwork systems for walls, columns, or columns and floors are dealt with and their use in practice is described. Climbing formwork as a special type of formwork is also mentioned. A short account of planning formwork assignments with the aid of special computer software and the designing of system formwork completes the lecture.

Introduction Essentially, two different types of formwork are used in practice. Conventional formwork consists of a formwork shell, or lining, built with boards, or panels and a supporting structure made of straps, yokes, etc. Formwork systems, on the other hand, normally consist of three-ply panels transverse and cross girders and a supporting construction of floor props or telescopic frames. On account of their well designed components these formwork systems have, in most cases, attained such a high degree of flexibility that they have often ousted conventional formwork.

Formwork components The main components of concrete formwork are the formwork shell, or lining, and the supporting structure, including its bracing and anchoring members. Also of importance is a separating aid, known as formwork oil, which makes for easier removal of the formwork.

Formwork skin The choice of the formwork skin to be used is influenced by the production costs, the working time, the number of reuses, including the decision to use a formwork system or not, the shape keeping resistance, the planned structure of the concrete surface, the labour costs and architectural aspects. In the following the most commonly used materials are described. In principle a distinction can be made between absorptive and non-absorptive surfaces of the formwork shell. Absorptive formwork linings, with their rough surfaces, take water from the concrete and lead to a darker grey colouring of the finished concrete surface. This effect is very important in the case of fairfaced concrete. Wood as a formwork material is cheap, easy to handle and elastic and is of adequate strength. If the surface is improved with plastic resin it also offers the advantage of easy removal, simple cleaning and a long service life. Plain boards are used for fixing and filling for visually unimportant concrete surfaces, for highly sectionalised surfaces with a low frequency of use, or as open formwork. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E2511

On the other hand, in some cases it is desirable to show the texture of the timber on the concrete surface. The disadvantages of boards are the high wastage and the high labour costs. Due to changes in the moisture content deformation occurs, e.g. warping of boards and gaps between individual boards, with resultant ridges on the concrete surface. Dry boards take water from the concrete, which would be needed for the hydration of the cement. An average moisture content of about 14% to 18% is, therefore, recommended for the formwork shell. Rough sawn boards produce a "living" concrete surface, because not only the outlines of the boards, but also their surface texture can be seen. New boards contain a high percentage of tannic acid, which must be neutralised by artificial ageing using lime wash. Rough sawn boards are used about four to five times for concrete surfaces in civil engineering. Planed boards for fairfaced concrete reproduce only the outlines of the board and give a neutral surface. In most cases they are planed on one side only, the reverse side being mechanically aligned to ensure the same thickness of all boards. The frequency of use is two to three times. A special danger with planed boards is the formation of laitance or scum, which is hard to remove from the concrete surface. Profiled boards are produced by treating the surface by sandblasting, brushing and, in some cases, by flaming. The texture of the wood shows up clearly on the concrete surface. Profiled boards are suitable for visually uniform fairfaced concrete. There are prefabricated boards on the market with improved surfaces, ready for immediate use. Profiled boards are economical because they can be used ten to fifteen times. To connect individual boards different types of joints have been developed, see Figure 1. They are useful for keeping the surface of the formwork shell flat and for tightening the formwork. For fairfaced concrete they are obligatory.

Figure 1

Types of joints. ( a ) Butt joint, ( 6 ) half lap joint, ( c ) tongue and groove joint, ( d ) special joint.

Butt joints are the easiest and most frequently used type of joint. In forming these the anisotropic properties of wood have to be borne in mind. If the heartwood side faces the concrete and the moisture content is adequate a good joint can be ensured, because the further swelling of the boards closes the joints. In other cases the danger of sedimentation and formation of holes is high, because the waterlcement mixture seeps through the joints. In the case of half lap joints the boards can be removed without breaking, but ridges in the concrete cannot be avoided. This kind of joint is seldom used. Tongue-and-groove joints are subject to fracturing and material wastage is thus very high. So-called special joints are very effective: they ensure a high degree of tightness, no ridges on the concrete surface and the possibility of removal without breaking of the boards. STEPIEUROFORTECH - an initiative under the EU Comett Programme

In formwork structures wood-based materials are often used for the formwork shell, or lining. Fibreboard panels are produced from felted timber fibres with or without filling and binding material. When of simple quality they are used only once and may be left in the building. Their advantage is a low dead weight and low cost. Fibreboard panels are suitable for curved surfaces and also for fairfaced concrete. Soft types of fibreboard have no static function. They are therefore always used in conjunction with an additional supporting structure, which is built using boards with a spacing of 50 to 300 mm. Normally fibreboards tend to swell. They are used as "lost" formwork linings for ribbed ceilings, etc. Oil-hardened fibreboard panels can be used several times if no demands are made on the smoothness of the finished concrete surface. Three-ply chip- or particleboards are made by compressing small wood chips bonded together with resin. Their surface is either rough or improved. Because of their low stability the supporting structure has to be designed accordingly. There are no special restrictions governing the use of chipboards but their tendency to swell, with the result that the impression of the nail heads shows on the concrete surface, must be reckoned with. Rough chipboard can be used about four times, improved board about twelve times. Three-ply formwork panels consist of three plies of the same thickness, glued together over the whole of their surfaces. The grain of the inner ply is at right angles to that of the outer plies. They are available as standardised formwork panels with a width of 500 mm, in most cases with a mechanically treated surface. Their tendency to swell and shrink is low. Standardised formwork panels can be used for a lot of jobs, up to forty times, large-area systems about ten to twelve times. In the case of plywood a distinction must be made between wood core and veneer plywood. Wood core plywood consists of at least three plies. The middle ply is of wood battens placed side by side at right angles to the grain of the outer plies. The battens need not be glued together. The outer plies consist of one or two veneers. The outer plies are glued to the centre ply, thinner outer plies are also glued together. This type of plywood can be used for many different jobs. Veneer plywood is produced from at least three sheets of veneer with the grain similarly oriented, that is, with the grain of the middle ply at right angles to that of the two outer plies. It is usually improved with a plastic coating. For fairfaced concrete formwork its thickness is 4 to 12 mm, as a selfsupporting system 15 to 30 mm. It can be used ten to twenty times. The frequency of use is not only influenced by the panel itself but also by the kind and thickness of improvement. Webbed board consists of solid timber webs, which are covered on both sides with solid boards of wood or wood-based material, in most cases with plywood, see Figure 2. Their advantage is that a supporting structure is only needed at right angles to the webs.

Figure 2

Composite panel.


Plastic formwork lining is rather expensive. It is available as a liquid or as shaped formwork lining. Good experience has been found with fibreglass reinforced synthetic resin. Metal formwork shells, mainly of steel, are used for precast concrete or for high frequency of use.

Formwork oil To facilitate formwork removal the use of formwork oil is necessary. Formwork oil is available as oil, emulsion, compound or as chemically effective substances. Formwork oil should be applied in films of about 2 to 50 microns. Thicker coatings lead to the formation of patches and influence the concrete surface. Manual application is thus not recommended. The rusting of steel formwork cannot be prevented by formwork oil. For fairfaced concrete it is important to use one type of formwork oil only. General demands on formwork oil are: a uniform surface and a uniform colouring of the concrete, no sedimentation, no formation of patches or pores, easy removal of formwork, ensuring an undamaged surface ready to take plaster without further treatment and, last but not least, environmental acceptability. For rough timber formwork mineral oil or chemically effective substances are preferable to emulsions. Compounds should not be used. For planed timber formwork the same holds good but compounds may be used. Plywood formwork can be treated with mineral oil, chemically effective substances or compounds. For plastic and steel formwork all types except emulsions are suitable.

Supporting structure Despite the development of special I-beams and lattice girders solid timber has retained its significant importance for formwork structures thanks to its easy handling properties and adaptability. Solid timber is used for formwork systems in combination with service girders but mostly as formwork for irregular surfaces, special constructions or bridges. The disadvantage of solid timber is the difficult removal, no variability in length, a low frequency of use and a high material wastage. Special I-beam and lattice girders are very important because of their low price, long service life, low maintenance costs and their high stability. For large-area walls service girders in combination with strong formwork yokes are very useful. For practical reasons service girders are no longer built with a camber, and the flanges are parallel. The length of girders can be up to 20 m. Lattice girders are built in such a way that the braces or struts are glued to the flanges only over small areas, so that secondary stresses have no effect. The hollow spaces between flanges and the braces can be conveniently used for handling the girders. Normally these trusses are supported at the joints. The braces of some systems consist of double members, finger-jointed to the chords. One end of such a girder is always vertical, the other may be inclined. Two girders can be joined end to end to increase or vary the length. The use of solid timber service girders has increased over recent years. With these girders a three-ply, cross-glued web is connected with solid finger-jointed timber flanges. The longitudinal joints are also finger joints. Holes are drilled in the middle of the chords for lifting. Such girders are available with depths from 160 to 360 mm. Figure 3 shows different type of girders. STEP/EUROFORTECH - an initiative under the EU Comett Programme

Figure 3

Different types of timber formwork girders. (a) with three-ply web, ( b ) glulam, ( c ) truss.

Wall formwork Small wall formwork areas, especially if they are irregular or complicated can be built using conventional formwork. Conventional formwork has the lowest material costs but the highest labour costs. If panels are used instead of boards labour costs can be reduced. If a multiple use of similar sized elements is possible the use of large-area formwork is preferable, for economical reasons. In most cases frequency of use is too low. Consequently basic formwork components have been developed, which can be put together horizontally and vertically to give the planned size, see Figure 4. This development has resulted in standardised elements supplied fully assembled. Frame-type formwork systems are a good example of this. They are light enough to be moved by hand, see Figure 5. This type of formwork is a synthesis of conventional and large-area formwork.

Figure 4

Wall formwork. (a) three-ply formwork sheet, ( b ) timber girders, (c) connection with connectors, splice plates and fastening bolts, (d) stacking flange, ( e ) steel waling.


E2515

Figure 5

Frame-type formwork system.

Formwork systems also include such components as corner elements for inside and outside corners, hinged corner elements for variable angles, corner stiffening plates, closing plates, quick action connectors, waling and stabilising components, all of which are well-constructed and supplied by various producers. Parallel with the development of formwork systems formwork clamps of a very high load-carrying capacity have been developed. A formwork clamp consists of the anchor plate, the anchor nut and the anchor bolt, plus a distance piece. During concreting the clamps are prestressed against the distance piece, to compensate for the deformation of the tie bolt and the pressure of the anchor plate on the timber flange. In principle it can be said that the number of formwork system components has decreased, because they can be used for more purposes. When deciding the choice of wall formwork it should be born in mind that for large-area systems a crane is needed, whereas frame formwork can be moved without one.

Components of wall formwork The formwork lining consists of horizontal boards or panels. These are held in place by vertical straps and horizontal yokes but more often with service girders. The prestressed formwork clamps take up the pressure of the concrete. The supporting structure with uprights and struts or brackets and bolts ensure the stability of the whole system. In the case of formwork systems the concrete placing platform is mounted on the formwork and stays there for multiple use. Service girders allow a significantly larger spacing of the waling than sawn timber. Girders are used in wall formwork as vertical members. As horizontal chords double squared timber, in most cases double steel C-sections are used. They have been developed for the various formwork systems into useful chord constructions, permitting any spacing of chords and ties, see Figure 4. Frame formwork systems are industrially produced structures of low dead weight and can therefore be handled by one or two persons. The basic construction of a conventional formwork system is dispensed with, because the frame components are delivered ready for use. Connection of the several components is by means of quick acting clamps. Chords are no longer necessary. Erection time and labour costs are much lower than with conventional formwork, but the price of STEPIEUROFORTECH - an initiative under the EU Comett Programme

the material is much higher. The use of frame-type formwork calls for more planning than with other methods. Today this problem is taken care of by special computer software supplied by the relevant producers. In contrast to large-area systems frame formwork requires no special job preparation. Frame formwork systems are very economical if the frequency of use is low, variable ground plans exist and no crane is available. Frame components can also be put together to form larger systems, see Figure 6. With circular forming panels curved concrete surfaces can be produced. Because of many visible joints frame-type formwork is not very suitable for fairfaced concrete.

Figure 6

F r a m e - ~ p formwork e system.

Column formwork Column formwork always involves high labour and material costs. Normally the columns are concreted before the casting of beams and slabs using standardised prefabricated components, which project beyond the top of the columns. Conventional formwork is only recommended for a small number of columns or for irregular sections. The formwork shell is built with vertical boards, held together as panels by horizontal boards or steel straps.

Figure 7

Column formwork.

Modern column formwork consists of plywood panels fixed to open formwork or service girders, see Figure 7. For the chords squared timber or more often STEP/EUROFORTECH - an initiative under the EU Comett Programme

E2517

standardised components are used, e.g. steel waling with slotted holes. These steel waling can be fixed with flange clamps at any position of the chords. If they are placed in such a way that each of them projects at one end beyond the edge of the formwork different cross-sections of columns can be produced using the same components. The edges are stiffened with corner connectors. For the stabilisation and alignment of the system adjustable braces are used. As for walls frame-type formwork can also be used for column constructions. The panels are connected over the edge with universal fixing bolts, to give variable column cross-sections. Thanks to their quick assembly frame panels for column formwork are economical even for a low frequency of use.

Floor formwork Components of floor formwork The formwork shell consists of boards, three-ply panels or plywood. The horizontal supporting structure is a lattice of squared timbers or service girders. In former times solid timber posts were used for vertical supporting structures. These have now been replaced by adjustable floor props. If large telescopic distances are involved and a sufficiently high frequency of use of the same sized units is possible the formwork shell and horizontal and vertical supporting structures are put together to form mobile floor "formwork tables".

Floor formwork systems In recent years the development by the formwork production firms of floor formwork systems based on well-designed components has led to a decline in the importance of conventional floor formwork. The floor formwork can be built step by step. Two men can set up the whole formwork for large areas, supported by means of free-standing folding props with tripods, see Figure 8.

Figure 8

Floor formwork.

A vertically adjustable prop head can be fixed, which makes it possible to lower the beams a few millimetres so as to remove the formwork shell and the transverse and cross beams. The prop head is constructed in such a way that either one beam is safely held or, by turning the head through ninety degrees, two overlapping beams. In this manner longitudinal extension is possible, without destroying the formwork material. Simple props are placed between the folding props. The girders are of standardised I-sections. Thanks to their low dead-weight, they can be handled by one man. Their load-bearing capacity is STEPIEUROFORTECH - an initiative under the EU Cornett Programme

higher than that of comparable squared timber beams. To ensure immediate reuse of the various components it is necessary to decrease each period of use. Because concrete needs a rather long time to develop adequate load-bearing capacity the floor props of the supporting construction must be left under the floor after removing most of the formwork. These floor props are placed under supporting beams, which are included in the formwork shell. The rest of the formwork can thus be removed and used for the next assignment. As a rule this floor system is very economical for low frequencies of use and for variable ground plans. Another important possibility with such formwork systems is to assemble the basic components - formwork shell, longitudinal and transverse girders and supporting structure - to form complete units for repeated assignments. The top part of the formwork - formwork shell and beams - consists of the same materials as the above described system. It can be adapted to variable ground plans and different floor designs. The supporting structure is built using hotgalvanized basic frames, which can be quickly and safely joined together to form larger units. Vertical adjustment is exact to a few millimetres. The system is fully variable in width and height. The basic units can be combined for any structure. For greater heights the basic frames are placed one upon the other and fixed with connecting bars. Stability is achieved by means of horizontal and diagonal braces or by special supporting frames. On top of the supporting structure telescopic frames with specially constructed heads for connection to the top part of the formwork ensure accurate adjustment. The bases of the frame are either constructed using heavy-duty screw jacks with lock nuts or using telescopic frame legs, with telescopic spindles. For removal of the formwork system the "tables" are loosened with the spindles, lowered by means of transport winches and transported on the wheels provided. If the crane is available the units can be lifted and moved for a further assignment. This saves expensive crane down-times. This type of formwork system can only be used if a crane is available and it is very economical for high frequencies of use. With this system it is also possible to form floor joists and T-beams. The formwork for these can be integrated into the floor formwork and lowered for removal by means of swivels. The systems high flexibility offers the possibility of producing any shape of floor, e.g. mushroom floors, or floors with complicated ground plans.

Special formwork systems For a number of assignments special formwork systems have been developed. One very important type is climbing formwork, which is used for piers, towers, skyscrapers or concrete dams. Climbing formwork allows the technical advantages of wall formwork systems at any height. Large-area formwork units, climbing brackets and suspended platforms are assembled in such a way that the whole structure can be shifted with one movement of the crane. This saves down-times. The work is divided into three phases. With the concreting of the first section special cones are installed, which serve to position the next section. Once the concrete displays adequate strength the formwork is removed far enough from the concrete for it to be easily cleaned. The positioning cones are replaced by climbing cones and at the fixing points the nuts are loosened. The formwork is held together by the bracket and the crane shifts the formwork as a simple unit by means of the climbing cones. With scissor-action spindles the formwork is STEP/EUROFORTECH - an initiative under the EU Comett Programme

E2519

pressed into the former concrete section and fixed. For counteracting wind effects wind struts are used. The combination of vertical waling, scissor-action spindles and waled-to-bracket connectors results in formwork units of different heights, which means that the concreting phase can be varied as required. A modern type of climbing formwork is the self-climbing system. Special climbing shoes are built into the front section. The automatic climber consists of two intermeshing frames, which lock alternately into the climbing shoes. The climbing is effected by means of a hydromechanical system.

Formwork system design with computer software or makeshifts of the formwork producers Many formwork producers have developed computer software to ensure that their formwork systems are used with the highest possible efficiency. These computer programs are very flexible and can be used to solve various problems. After inputting the ground plan of a building, which can also be taken from other CAD-software, the program provides a formwork solution, suitable for the selected building sections and including the waling, stabilising components and the supporting structure. Details of corner or connecting elements of wall formwork are drawn. Floor formwork solutions are adjusted to the room height, joists and floor thickness. Elements available in the contractor's yard or on the building site can also be taken into account. In addition the total amount of formwork for each section, a parts list and the price of all material used is automatically displayed. For the design of formwork systems without the use of a computer the production firms provide makeshifts, that is, job preparation aids for the quick and easy calculation of wall and floor formwork. With the help of neatly arranged tables and diagrams the user is able to design the formwork shell, girders and supporting structure.


Timber structures in aggressive environments STEP lecture E26 H. Hart1 Zivilingenieur fiir Bauwesen

Objective To present information about timber structures under the influence of aggressive environments.

Summary A brief explanation is given of the behaviour of timber and wood-based materials when exposed to chemical actions, high temperatures and the influences of ultraviolet rays and weathering. The material properties and the resistance to these actions are described. The use of timber in practice in aggressive environments is shown with examples of projects carried out.

Introduction The material properties of timber are mainly determined by its structural composition. As a natural product wood shows great deviations in respect of these properties, not only between different species of wood but also in the same species. A certain homogeneity of material properties can be achieved by chipping the wood or cutting into bits and subsequently putting then panels together again to form glued laminated timber or wood-based materials. Glued laminated timber displays the same anisotropic characteristics of solid timber with significant differences between properties in the direction of grain and perpendicular to it but reduces the significant defects. With the use of wood-based materials or glued laminated timber material final properties can be varied by changing the properties of the constituents and their shape and assembly. Of significant importance is the choice of the glue used. Factors to be considered in choosing an adhesive for structural timber include behaviour on exposure to the environment and durability, ease of use during production and the nature of the joint to be made. For more information concerning material properties see lectures A4 and A12.

A number of chemical processes and other similar events may give rise to chemically polluted atmospheres often associated with the presence of hightemperature moisture-laden air. Swimming-pools will also seem an obvious situation for such exposure risks, although the equilibrium moisture content of timber structures can be quite low due to the heated environment. On account of their high resistance to the effects of aggressive environment, timber and wood-based materials are often used for structures where such conditions prevail. This natural resistance of timber is not a fixed value, but the behaviour under different negative influences varies greatly and also changes with the course of time.

Resistance to chemical actions In comparison with other materials, such as steel, timber offers remarkable resistance to the effects of chemical actions. Damage or destruction of timber, starting from its surface as a result of physical and chemical interaction, is called corrosion. It must be realised that aggressive media mainly attack the hemicelluloses and lignin and not the cellulose. For this reason coniferous STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E2611

timber, with its higher percentage of lignin, displays in general a higher corrosion resistance than deciduous timber. Corrosion phenomena are mostly associated with a brown or dark colouring of the timber, which spreads from the marginal zones to the central sections. All these effects depend on the strength of the chemical and the period of exposure. The manner of destruction is similar to the behaviour of timber under the influence of fire. The decrease in strength effects only the first 10 mm to 20 mm depth, while for the rest of the section the natural properties remain unchanged. In designing timber structures it is usual to allow for these effects by increasing the cross-section. Timber is very resistant to acids and also to salt solutions. Alkalies lead more quickly to a destruction of timber, especially if high temperatures occur at the same time. At a pH-value of between 3 and 10 it may be assumed that no damage will occur, however certain types of timber react differently and the time and temperature of influence play an important part. For a short time higher or lower pH-values will cause no damage, but higher temperatures accelerate the speed of destruction. The influence of aggressive gases, for example ammonia and formaldehyde, have no negative effects on either solid timber or glued laminated timber. Sulphur dioxide causes certain damage only in combination with high temperatures and high moisture content. To determine the real characteristics of the resistance to chemical actions it is often necessary to carry out tests. Such tests are normally conducted in two parts, first of all using small pieces of timber exposed to highly concentrated attacks, in order to obtain qualitative statements, and later with full scale tests to quantify these values. This method is of special importance for glued laminated timber and wood-based materials, where the choice of glue or binding material is also significant in increasing the resistance to chemical actions. Glued laminated timber pieces have been treated with highly concentrated sulphuric acid to measure its effect on the timber and to determine the most suitable type of adhesive. From these tests it was found that only certain adhesives of the phenol-formaldehyde and resorcinol-formaldehyde type and their combinations satisfy this particular requirement. With such adhesives no seeping of acid into gluelines was seen, which would have led to the same dark colouring of the timber alongside the gluelines. A special form of chemical action may occur where timber is in immediate contact with metals, for example with mechanical fasteners. Depending on the type of timber and the moisture content, corrosion of the metal surface leads to colouring of timber and also changes its mechanical properties. Iron, when in contact with wood of high tannic acid content, like oak, causes a black colouring. The tensile strength decreases with long contact, the compressive strength is not affected. On small contact areas the attack occurs very slowly and is limited in extent. For trussed rafter roofs in stables or similar buildings, with nail-plate connections, special corrosion protection have to be provided. In such cases the metal parts have to be galvanised or coated with plastics.

A secondary effect of corrosion appears if the decomposition products of corrosion are washed out by water or other fluids. The consequence is dark coloured traces on other parts of the timber, which can be unsightly.

STEPIEUROFORTECH


In most cases where timber is exposed to chemical actions no special conservation measures are necessary, because the natural resistance is adequate. In fact it is better if the planed surfaces of timber beams are not coated with wood preservatives, since the paint forms grooves in which sedimentation of aggressive chemicals is facilitated.

Examples of projects carried out For the reconstruction of a storage battery manufacturing plant (see Figure 1) glued laminated timber was used in order to resist the attacks of sulphuric acid vapours, and to guarantee the durability of the structure for at least twenty years. Tests have determined the amount of weakening to be expected over a period of twenty years. These results formed the basis for the calculation of the crosssection.

Figure I

Timber roof structure of a storage battery manufacturing plant.

As described before, these tests were performed in two steps. The first was to choose the right adhesive with small-scale tests, and the second to observe the behaviour of a girder made with this adhesive under operating conditions in the manufacturing plant over a period of several weeks. The roof of a central storehouse of a cheese production firm~wasalso built using glued laminated timber. Timber being able to resist the expected formation of lactic acid during fermentation better than other materials, for example steel. To solve the corrosion problems occurring during the storage of de-icing salt in storehouses or silos, many such buildings have been erected using glued laminated timber, see Figure 2. Metal components were in general of stainless steel. Because of the formation of condensation water and ammonia vapours, set free during the production of fresh ice, glued laminated timber girders and stainless steel connectors were chosen, for a roof construction over an ice skating stadium, see Figure 3. The roof over a thermal spa, which is exposed to thermal spring vapours, has been built with glued laminated timber, see Figure 4. The timber girders used were produced using water-proof and boil-proof glues. The quality of the metal fasteners was chosen to withstand the thermal spring vapours. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E2613

Figure 2

Silo from glulam for salt storage.

Figure 3

Timber roof structure of an ice skating arena.

Figure 4

Timber roof structure of a heated swimming pool.

Resistance to high temperatures The thermal properties of timber are extremely important. The specific heat of timber is comparatively high, in relation to its density, but of minor importance compared with its thermal conductivity. The thermal conductivity of timber is approximately 0,4% of steel. It varies, approximately in proportion to density, STEP/EUROFORTECH - an initiative under the EU Cornett Programme

being lowest for low density softwood and highest for high density hardwood. The thermal conductivity also increases significantly with moisture content, by perhaps a third for a moisture content increase of 40%. High temperatures affect material properties of timber only if the time of exposure is long enough. The degree of destruction also depends on the level of these temperatures. Temperatures around 100 "C lead to a brown colouring of timber but the weakening of strength can hardly be detected. At temperatures up to 150 "C, the first decrease of strength occurs within a couple of weeks. Higher temperatures accelerate the destruction and above 250 "C unprotected timber, depending on its density and environmental conditions, is rather quickly destroyed. In general, it can be said that the higher the density the greater the resistance to high temperatures. On account of the low thermal conductivity of timber the thermal destruction is transmitted slowly from the outer parts to the central section, so that appropriately designed elements have a sufficient resistance to high temperatures. The simultaneous influence of high temperatures and increased humidity causes significant changes in the material properties of timber; both the strength and the stiffness decrease.

Example of projects carried out Figure 5 shows the roof over an enamel smelting house. Timber has been used successfully to resist the continuous high temperatures inside the building augmented by the changing climatic conditions on the outside.

Figure 5

Timber roof structure of an enamel smelting house.

Resistance to the influences of ultraviolet radiation and rain Resistance to the effects of weathering depends mainly on the structure of the timber. Bulk density, thickness of cell walls and constituent substances are highly significant. When estimating degree of resistance a distinction has to be made between the effects of ultraviolet radiation alone and conditions where timber can also be reached by rain. Timber exposed to UV-radiation, suffers from discolouring, which can be irregular for different radiation intensities. In this way timber yellows and in course of time takes on an intensive brown colour, resulting from little dark decomposition products of the corroding surface. Translucent varnish is not STEPIEUROFORTECH - an initiative under the EU Comett Programme

E2615

suitable for protecting timber from these effects. Instead, paints available in a large variety of colours should be used. With the additional influence of fluids or high humidity these decomposition products are washed out, leading to a grey colouring of the wood. The surface thus created is rough and full of corrugations, which reduce the cross-sectional area. Besides this direct reduction of area or quality reduction of performance, secondary destructive processes caused by fungi are likely, and therefore further destruction of timber can take place, and which is usually of more significance. The resistance of wood-based materials is not only determined by the structure of the timber used; it is also influenced by the adhesive or binding material used. Urea resin based glues are rather quickly destroyed by swelling and shrinking of adjacent timber parts, which decreases the overall strength. For full exposure to weather, water-proof and boil-proof adhesives should be chosen; these can be obtained by using certain adhesives of the phenol-formaldehyde or resorcinolformaldehyde type. Only these types can provide resistance to the weathering of wood-based materials. Besides the proper choice of adhesives improved surface treatments also increase the resistance to the effects of ultraviolet radiation and weathering. Although it is normal to dry timber to a moisture content equivalent to the average relative atmospheric humidity anticipated movement problems are often also encountered: shrinkage with drying and swelling with wetting. Faults such as cracks or fissures appear, in other situations cross-sectional movement may become apparent as warping through the effects of twisted grain. The obvious solution to all these problems is to use only timber with low movement but this is not always realistic. The alternative is to impregnate timber with chemicals which induce stabilisation, although processes of this type are also frequently unrealistic because of the difficulty of achieving deep penetration. One approach is to enclose timber within a protective film in order to stabilise the moisture content. Paint and varnish coatings will act in this way, provided they completely cover the timber and are not damaged in any way. Unfortunately whilst these coatings give good protection against rainfall, they are unable to prevent moisture content changes resulting from slow seasonal fluctuations in atmospheric humidity. As a result the painted timbers shrink or swell with changes in relative humidity, causing the surface coating to fracture. Rain is absorbed into the crack by capillary action and the remaining paint coating restricts evaporation, so that the moisture content increases and thus causes the dampness to accumulate. Therefore in any case the paints to be used should be micro-porous, that water may evaporate. The most widely used conservation system is to ensure that timber remains dry by taking appropriate structural precautions. Thus buildings have to be designed to protect the structure from the penetration of dampness, especially in the form of dripping water, e.g. rain but also as condensation water. For example, walls should be designed to resist penetrating rain, perhaps as cavity constructions, which have to be built with the possibility of adequate ventilation. In some cases where timber girders are exposed directly to the weather it is often more useful to protect them with a cheap cladding, which can be replaced from time to time, than to preserve the whole girder by chemical treatments.


Concluding remarks Timber offers a high resistance against aggressive environments and can be successfully used under such conditions. The maintenance cost for steel or concrete structures in identical environments are much higher than for comparative timber structures. Steel structures need a regular coating with corrosion protection and concrete buildings a permanent observation of the concrete surface in order to detect cracks which can lead to corrosion of the reinforcement bars. Timber, however, needs very little maintenance in most aggressive environments although, unless stainless steel is used, the corrosion protection of the steel fasteners has to be checked in regular intervals.


Learning from failures I STEP lecture E27

Objective

B. Johannesson G. Johansson Chalmers University of Technology

To describe some structural failures caused by snow, wind or landslide.

Summary Structural failures caused by wind, snow and landslides are discussed. The wind and snow failures are mostly due to mistakes in design or manufacture. The behaviour of the houses in the landslide shows that a "stiff" wood frame house can withstand a very severe load condition and yet hold together. Proper design with accurate design loads results in buildings that can withstand severe load combinations very well.

Storm damage General Some damage cases from the heavy storms that occurred in West Sweden, in September 1969, are described. It happened a long time ago, but many of the experiences are still as relevant as they were 25 years ago. The damage was investigated and the results were given in a report (Johansson 1970). Many general observations are valid for all types of structures. Many of the roofs that blew off were just lying loose on top of the houses, they were not fixed to the main structure at all. In several cases the main reason was so obvious that no further extensive investigation was needed. The wind velocities were high, the maximum 10 min mean velocity recorded was 31 m/s and the maximum gust velocity recorded was 37 d s . Compared with earlier and later storms, the storms of 1967 and 1969 passed over areas with big cities and they lasted for a long time. For about 4 hours around noon, the wind velocity was at its maximum value or was very close to it. Many of the damaged houses were quite newly built, most of them were less than ten years old.

Timber roof structure on a concrete slab There were quite a lot of two or three storey houses that lost their roofs. On the upper concrete slab a timber frame system was built with vertical posts 50 X 100 mm and rafters 50 X 100 mm to 50 X 150 mm. The spacing of posts and frames was dependent on the house and rafter dimensions, however the frame spacing was very often 1,2 m. A typical sketch of the structure is shown in Figure l . All of the examined damage had either insufficient or non-existent anchorage of the timber structure to the concrete slab. In many cases the roof structure was attached at the eaves, but none of the posts was fixed in any way. This means that when the storm succeeded in breaking the anchorage at the eaves the whole roof was free to blow off. Then the wind force could also influence the roof plate with pressure from below, see Figure 2.


Figure I

Typical sketch of a timber roof structure on a concrete slab; (a) concrete.

Figure 2

Probable damage sequence.

Figure 3/4

Damage examples. Fixing of post to the roof slab. Note the wrong position of the steel strap (top). Part of the roof is lying on the leeward side of the damaged building (bottom).

Different types of anchorage were used. The most common was thin steel straps, cast in the slab and nailed to the timber structure. In other cases a horizontal 50 X 100 mm timber was bolted to the concrete slab, to which the timber structure was fixed with just a few nails. However, many of the anchorages showed very poor workmanship and could not be expected to resist the wind forces. In Figures 3 to 6 some examples of damage are shown.


Figure 5/6

Part of a damaged roof (top). Fixing of the posts to the secondary member laid on the concrete slab (bottom).

Temporary timber frame building A temporary timber building was totally destroyed during the storm. The building was 30 X 15 m in plan and 5,2 m high at the ridge. The frame system can be seen in Figure 7. All the joints in the structure were bolted. Unfortunately the size of the timber used is unknown but the rafters were approximately 50 X 150 mm. The purlins were fixed to the frame rafters with a supporting wood block. Each block was nailed with three nails to the rafter. The purlin, however, was only nailed with one nail to the block. Because of this the whole roof blew off. The frames were sway deformed with a maximum deflection of 0,5 m. If the roof had not been removed, the frames would have probably been broken instead.

Figure 7

The frame structure and the deformation of the frame; (a) deformed frame.


E2713

Figure 8

View of the damaged building. A blown-off roof section can be seen to the left.

Snow load damage General During the winter of 197611977 heavy snow loads caused several structural failures in Sweden. The damage was of different types, ranging from failures in roof membranes caused by freezing water to total collapse of a building. In an investigation, (Johannesson and Johansson 1979a) almost one hundred different damaged structures were studied. The investigation included steel, timber and concrete structures. The snow depth during the winter, measured by the Swedish Meteorological and Hydrological Institute (SMHI), was in some cases found to considerably exceed the 50 year Mean Recurrence Interval (MRI) ground snow.

Causes of snow damage to timber structures The main reason for the structural failures was the heavy snow loads causing the total load to exceed the load carrying capacity of the structures. Excessive snow loads - that is the amount by which snow loads exceeded the snow loads given by the Swedish Building Code, drifting snow and sliding snow were estimated to have caused about 30% of the failures in the timber structures. Manufacturing faults, including all types of deviations from the prescribed design, both in factory and on site, and also underdesign, that is mistakes made either in calculations or as an inappropriate design, were estimated to be the cause in about 65% of the failures.


Some typical failures Bad gluing of wooden structures had caused failures in glulam beams and plywood webbed I-beams. The glulam beams were of an I-type where the gluing of the flanges to the web was badly done. Due to the thickness of the flanges it was probably impossible to achieve proper pressure along the whole gluelines during manufacture. One of the basic conditions for successful gluing is that the glued surfaces are plane enough to come in close contact and that the applied pressure is sufficiently high. Especially in the case of nail-gluing it is necessary to use planed surfaces. The plywood beams failed due to badly planed flanges resulting in bad glue joints.

Figure 9

Failure in a glued laminated timber structure due to bad gluing. Section (top) and sketch of the failure (bottom); ( a ) glulam, ( h ) tension bar.

Timber connections with nail plates (punched metal plates) are sensitive to misplacement of the nail plates since the plate size is often small. Collapses of roof trusses were found to be due to undersized nail plates at the supports. These small plates were placed so that cracks could develop in the rafter leading to a collapse. When nail plates are used it is necessary not only to determine the minimum size of the plate from strength considerations but also to take into consideration the possibility of crack development. A similar failure occurred in a glued laminated beam with a notch at the support.

Figure 10

Joint with a nail plate that is too small; (a) failure sui$ace, (b) nail plate.


E2715

The design of carports usually means rather limited resistance to withstand horizontal loads. In some observed cases the carports were built next to the neighbour's house to which, due to legal difficulties, they could not be attached. Sliding snow from that house caused horizontal forces in the carports resulting in permanent horizontal sway. Methods to prevent this kind of damage probably have to concentrate on preventing the snow from sliding, since it is difficult to stiffen a carport, due to its function.

Figure l l

Typical placing of a carport next to the neighbour's house.

Figure 12 shows one type of roof truss that is easily "overloaded". The moment distribution is very much dependent on the distance a in the figure. The failure is often a pure bending in section 1 or in the joint 2. In the case shown the bending stress in section 1 was calculated to 66 N/mm2.

Figure 12

Section through a timber store house; (a) round timber pole.

Special failures The primary structure of a cold store house was a three-hinged wooden truss frame. After about 20 years of service the building was moved to another place. A drawing of one half of the truss is shown in Figure 13. The flanges of the truss-beam were made of 50 X 125 mm sections. The contractor cut the frames in the sections A-A (Figure 13) when the building was moved to its new place. In these joints lap splices of 50 X 125 mm were nailed with just a few nails to each part. The length of the lap splices was about one metre. The whole building collapsed due to failure in the joints A-A. This was probably a progressive failure starting from one frame. At failure the snow load was STEPIEUROFORTECH - an initiative under the EU Cornett Programme

estimated at 1,l kN/m 2 based on in-situ measurements. According to the Swedish Building Code the joint in the tension flange should have been nailed with about 130 nails. The estimated number of nails actually used in the joint corresponded to an allowable load of 0,2 kN/m2 which is less than the dead load of 0,3 kN/rn2.

Figure 13

One half of a three-hinged truss frame, spanning 15,5 m, with joints at A-A.

The structure in Figure 14 probably collapsed due to the fact that the deformation in the steel wire tie rod was neglected. The failure occurred between joints B and C. The load on the structure was a combination of wet snow and unevenly ~ distributed on wind forces. The snow load was approximately l k ~ / m the roof. Stresses calculated for this load and wind load at a wind speed of 15 m/s gave a bending stress of 20 ~/rnrn' if the deformation in the steel wire was considered and 4 ~ / m r ifn ~not. Hence it may be concluded that it is vital that the theoretical model in the calculations is correct in order to avoid a disastrous failure. (d) 180 X 180 mm 50 X 230 mm (c) 50 X 180 mm

D

Figure 14

,2 - 50 X 230 mm

a-a

Timber frame structure with a steel tension rod. ( a ) steel cable, ( 6 ) purlins, ( c ) cover plate, ( d ) distance piece, ( e ) splice, If) bolt, ( g ) round timber.

STEPIEUROFORTECH

- an initiative


E2717

The landslide at Tuve General The slide area was about 27 hectares with a maximum length of 750 m. The slide width in the area where most of the houses were situated, was about 200 m. The number of houses in the landslide was 65 and about 100 houses close to the slide area were evacuated. Most of the houses moved about 100 m horizontally and 10 m vertically. The maximum transportation of any of the houses was 180 m. In Figure 15 a map of the slide area is shown. The map also shows the house slide paths. The houses in the slide were of different types. Most of them were timber houses of different types - old houses (50 years or more), new houses with prefabricated elements and new houses with timber frames. Some of the houses had timber floor structures, but most of them had cast concrete floors, constructed of lightweight concrete or normal concrete elements. A few of the houses were built with brick or light weight concrete walls and concrete slabs. The most damaged houses were those built of lightweight concrete blocks. Those walls could not withstand the forces from the landslide.

The terrace houses at Alrnhojdsvagen The main reason for the severe damage to these houses was that the concrete walls in the basement broke down. The walls had very little or no reinforcement and they split at the corners. Then the concrete slab slid off the walls and broke. The basement walls were intact in only two of 29 houses. After the concrete slab was broken, the timber frame superstructure was exposed to loads which exceeded the design loads many times. In most of the houses the "roof triangle" was not damaged. Also the "non load carrying walls" acted as load carrying to a great extent. As a matter of fact, the bathroom on the second floor seemed to be the most secure room in the whole building. Figures 16 and 17 show the frame system, the roof triangle and a sketch of possible damage development. In Figures 18 and 19 some details are shown.

Figure 15

Map of the slide area with the direction the houses slid indicated.

STEPIEUROFORTECH

- an initiative


Figure 16/17 The frame system (top), the roof triangle (middle) and possible damage development (bottom). ( a ) timber, ( b ) concrete.

STEPIEUROFORTECH

-


Figure 18/19/20

Details of damaged houses at Alrnhojdsvagen.

Small-element houses These houses were made of small wooden elements. Each element was built as a wooden box, about 200 mm square, length 2,5 m, filled with sawdust. The roof structure was a timber frame. The floor structure was light weight concrete elements. The connection between the wood elements and the floor structure was too weak. These houses lost their walls to a great extent. When most of the gable wall had fallen, the roof structure also broke. In one case a roof structure held together for four weeks before breaking. Probably the influence of humidity weakened the nailed joints. There was one house in the slide area structure. This house had a timber floor function as a closed square box. The wooden superstructure was nearly intact,

of the same type, except for the floor structure, which made the whole house concrete cellar was destroyed but the just a little deformed.


Figure 21

Roof cantilever. (a) fracture.

Figure 22/23 Photo of a house with roof cantilever. It is the same house on both pictures, the photo with the broken roof was taken four weeks later.

Ordinary element house The house had a timber super structure built on concrete block walls in the basement. There were timber beams in the floor structure and timber trusses in the roof. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

E2711 1

This house was very interesting. Although it had lost approximately 75% of its supporting basement walls, it did not break down, although it showed some deformation and a few broken windows. This is an excellent example of "box-function". A box with 4 stiffened walls (the inner walls act as stiffeners) and an intact top and bottom is rather stiff and safe. This was proved many times in the slide area. Figure 25/26 shows the building, resting on "nearly nothing".

Figure 24

Section of the house.

Figure 25/26 The house rests only on the centre walls and cantilevers in all four directions. Notice the shear deformation in the window openings. STEPIEUROFORTECH - an initiative under the EU Cornett Programme

Summary The reason for the storm damage in most cases can be traced to detailing performed as a matter of routine, overestimation of the strength of the connections, unsuitable design and structural details and negligence on the site. All parties involved in the construction process bear a share of the blame, no one is without blame. Heavy snowloads sometimes combined with wind loads caused a lot of damage. The damage occurred mainly in light-weight structures like timber and steel structures. These are often more sensitive to excessive imposed loads. The most obvious conclusion from this investigation is that accurate design with appropriate loads gives a satisfactory level of safety against failures. Bad design andlor poor workmanship during construction or erection often result in structures where the failure load can easily be lower than the design load. The behaviour of the houses in the landslide at Tuve very clearly showed that a "stiff" wood frame house can withstand a very severe load condition and yet hold together. The most impressive sight, however, was to see how the "roof triangle" in almost all houses kept "together" even if all walls had disappeared and the house had slid more than 100 metres. The overall conclusion is that correct design procedures using accurate design loads result in buildings that can withstand severe load conditions very well. However, this naturally raises the question as to what a design load in different situations should be and what the proper design should be.

References Johansson, G. (1970). Stormskador i vastra Sverige. ("Storm damage in Western Sweden") Byggforskningen. Rapport R33 (in Swedish). Johannesson, B. and Johansson, G. (1979a). Snoskador vintern 1976-1977. ("Snow damage winter 1976-1977"). Byggforskningen. Rapport R15 (in Swedish). Johannesson, B. and Johansson, G. (1979b). Tuveskredet 1977. Undersokning av smihusstommars skadetilighet. ("The Tuve landslide 1977. Investigation of the damaged houses"). Byggforskningen. Rapport R137 (in Swedish). Johansson, G. and Johannesson, B. (1984). Damage due to snow loads. IABSE 12th Congress, Vancouver, BC, finalreport, p. 829-835.

STEPIEUROFORTECH


Learning from failures I1 STEP lecture E28 J.P. Biger Bureau Veritas

Objective To describe cases of disorders and accidents caused by production, design or installation errors.

Summary The pathologies ot traditional framework, lightweight structures and glulam structures are examined, photographs and sketches are shown, and short comments are given.

Traditional structures

Figure I

Roof deformation.

The carpentry supporting the roof in Figure 1 is of a classical type: rafters, purlins, supported by gable walls and a central truss. The ridge beam of the roof shown in Figure 1 has become concave, and rows of tiles have slipped to the bottom of the roof planes. This very common situation is the result of the absence or inadequacy of blocking of rafters at the base of the planes, allowing the purlins placed obliquely to become deformed by transverse bending. The trussed rafter shown in Figure 2 has broken because of the presence of a largesized knot. The timber was obviously not graded properly. The ridge purlin shown in Figure 3 has broken because of the steep slope of the grain of the timber, which causes local transverse stresses, resulting in breakage. The timber was obviously not graded carefully enough. The opening of the assembly of a king post on the tie bar of a traditional truss is shown on Figure 4. It may be seen that the king post is cracked more or less on the plan of the bolt, and that the metal suspender is embedded in the bottom ot the tie bar. The fracture process is in progress. STEPIEUROFORTECH - an initiative under the EU Comett Programme

E2811

Figure 2

Fracture caused by a knot.

Figure 3

Fracture caused by the slope of the grain.

Figure 4

Assembly of king post on tie bar

This case involves a design error in the assembly of the central diagonals, which are blocked on the tie bar with the result that the metal suspender and bolt are overloaded. A good design would tie the diagonals to the king post. STEPIEUROFORTECH



Figure 5

Serious damage by fungus.

The structure shown in Figure 5 is about fifty years old. Infiltrations of water caused by gaps in the waterproofing of the roof created conditions favourable to the growth of fungus. This structure is very seriously damaged at places where infiltrations took place. Proper maintenance of the roofing is needed to avoid, or reduce, the risk of fungus attack on the supporting timber structure.

Lightweight structures

Figure 6

Buckling of trussed rafters.

Cases of buckling of trussed rafters, as shown in Figure 6, are rather common. They are caused by over estimation of the transverse rigidity of the framework, resulting from the idea that the length of transverse buckling for trussed rafters would be equal to the distance between batten lines. This idea is an illusion because the lateral restraint given by the tile battens and the bracing is not sufficiently rigid. The numerous nailed joints which are involved may slip and drag, as shown on Figure 7. The design of continuous bracing bars under the roof planes, with girtrails and torsional blocking of truss bases, as well as the adoption of reasonable design buckling lengths, are essential precautions needed to prevent the risk of lateral buckling of trussed rafters.


Figure 7

Slipping and dragging of batten nailing.

Figure 8

Failure of timber girder.

The broken girder shown in Figure 8 is used as a rafter. The end of its lower flange has broken, and its metal diagonals have torn away. The support of this girder is not properly designed: offsetting the supporting point from the triangulation node causes a secondary moment, which has resulted in the failure. Attention must be paid in the design of supports with respect to the calculation hypothesis in order to avoid secondary moments.

S-

Figure 9

'

-2

W

..L

..

Warping of LVL frame.


The frame shown in Figure 9 consists of two thicknesses of 36 mm LVL, side by side for the crossbeam and forming a coffer for the columns. Warping is due to insufficient lateral and torsional stiffness, caused here by a loose nailing of the LVL sheets together. The specification of adequate nail densities in the design documents is necessary to avoid this instability.

Figure 10

Instability of timber frame walls.

The timber frame house shown in Figure 10 is supported by props. The ground floor walls slope outwards, probably because the connections between floor and walls are absent or inadequate. Design, fabrication and erection should be done by professionals, preferably.

Glulam structures

Figure I I

Ridge deformation.

The ridge of this sports centre (Figure 11) has subsided about 300 mm, following an increase in the curvature of the flanks of the frames that form the main structure. This increase in curvature is not caused by any extra load on the frames, nor by creeping, but results from the transverse shrinkage as the wood dries out. Shrinkage moves the soffit and back of the curved flanks of the arches closer together, while their respective lengths remain approximately the same, which


E2815

means an increase in their curvature, and consequently subsidence of the key point as indicated in Figure 12. Control of drying conditions with regard to the service condition, and anticipation of the possible deformation will prevent this problem.

Figure 12

Increase in curvature and deflection due to shrinkage.

Curved beams are subject to transverse tensile stresses accompanying bending forces. This is a cause of serious cracking. Many cases have been recorded of cracking similar to the case shown in Figure 13. They are generally the result of over-optimism in the transverse tensile strength. Consideration of the influence of curvature on the transverse tensile stress (see STEP lecture B8) as well as the influence of service conditions on shrinkage and surface cracking is needed in the design of curved tapered beams. The use of reinforcements made of glued in bolts or glass fibre is recommended (see STEP lecture E4).

Figure 13

Cracking of tapered curved beam.

Another case of cracking caused by transverse tensile stresses is shown on Figure 14 and 15. These steeply tapered glulam columns are situated outside. Their exposure to climatic changes helps cracking. A good design would need a smooth taper, and inside location, which is, in most cases, an architectural concern.


Figure 14

External uprights with cut-off strips.

Figure 15

Detail view of cracking.

Waterponding may occur on flat roofs. It is responsible for many cases of collapse, the beam in Figure 16 is one example. Beam deflection under water ponding loads increases beam loading. In case of insufficient beam stiffness, this process leads to instability, gluing defects (Figure 17) and finger jointing defects (Figure 18) may reduce the stiffness and the resistance of beams, increasing dangerously their susceptibility to water ponding loads. Special care is needed for beam design in the case of flat roofs, the stability under water ponding loads must be checked. Prevention of gluing and finger jointing defects needs quality control in the manufacturing process.

Figure 16

Beam collapse.

Figure 17

Defective gluing,


Figure 18

Finger jointing defects.

Joints play an important part in glulam pathology. Shrinkage of beams and columns at frame corners often result in cracks as shown in Figure 19. This can be prevented by adequate drying of the glulam prior to assembly, associated with gluing process control. The designer must bear in mind that the risk increases with the depth of beams and columns. Reduction of design stresses may be recommended in particular cases.

Figure 19

Cracked frame corner.

Insufficient rotational capacity of joints may create problems for glulam structures. This is the cause of the collapse shown in Figure 20.


1 .,

Figure 20

Failure of cantilever joint.

The beam had two bays, a cantilever joint was located at a short distance from the central support, see Figure 21.

Figure 21

Two bays beam and cantilever joint (a).

The design of the cantilever joint shoe shown on Figure 22 does not allow the rotational possibility required by the position of the joint, which is different from the zero moment point.

Figure 22

Detail view of cantilever joint.


The canopy of the stadium stand shown in Figure 23 has been almost entirely carried away by an ordinary gust of wind.

Figure 23

Attachment failrire.

Figure 24

Detail view of attachment.

The detail view in Figure 24 shows that the beams were attached to the tops of the column by bolts located near the edge of each beam. This arrangement generates transverse tensile stresses, which have caused cracks to cut the beams at the level of the upper bolt. The design of the joint according to the provisions described in STEP lecture C2 is required to avoid this.


Timber Engineering Step-2

Recommend Documents