Reinforcement of Timber Structures

Reinforcement of Timber Structures A state-of-the-art report

Editors Annette M. Harte, Philipp Dietsch

Acknowledgement

The members of COST Action FP1101 would like to thank the COST Office in Brussels for financially supporting this publication. The financial support of Rotafix UK, Rothoblaas and SPAX is also gratefully acknowledged. No permission to reproduce or utilise the contents of this book by any means is necessary, other than in the case of images, diagrams or other material from other copyright holders. In such cases, permission of the copyright holders is required. This book may be cited as: Reinforcement of Timber Structures – a state-of-the-art report. Neither the COST Office nor any person acting on its behalf is responsible for the use which might be made of the information contained in this publication. The COST Office is not responsible for the external websites referred to in this publication.

Copyright Shaker Verlag 2015 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publishers.

Printed in Germany ISBN: XXX-xxx-xx ISSN: YYY-yyy

Contents Preface

v

Introduction

1

Annette M. Harte, Philipp Dietsch

Part I: Reinforcement Applications

3

1

Reinforcement of timber beams Steffen Franke, Bettina Franke, Annette M. Harte

5

2

Reversible timber-to-timber strengthening interventions on wooden floors Alessandra Gubana

25

3

Reinforcement of timber columns and shear walls Wen-Shao Chang

39

4

Analysis and strengthening of carpentry joints Jorge M. Branco, Thierry Descamps

55

5

Reinforcement of dowel type connections Laurent Bléron, Damien Lathuillière, Thierry Descamps, Jean-François Bocquet

77

6

Seismic Strengthening of Timber Structures Maria Adelaide Parisi, Maurizio Piazza

89

Part II: Reinforcement Methods

111

7

Adhesives for on-site bonding: characteristics, testing and prospects Benedetto Pizzo, Dave Smedley

113

8

Glued-in rods 133 René Steiger, Erik Serrano, Mislav Stepinac, Vlatka Rajčić, Caoimhe O’Neill, Daniel McPolin, Robert Widmann

9

Reinforcement with self-tapping screws Philipp Dietsch, Reinhard Brandner

161

10 FRP reinforcement of timber structures 183 Kay-Uwe Schober, Annette M. Harte, Robert Kliger, Robert Jockwer, Qingfeng Xu, Jian-Fei Chen 11 Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings Tanja Marzi

209

iii

Outlook

231

Philipp Dietsch, Annette M. Harte

List of contributors

232

List of reviewers

234

iv

Preface This report is a publication developed within the European network COST Action FP1101 “Assessment, reinforcement and monitoring of timber structures”. COST Action FP1101 is a research network established under the aegis of the COST domain “Forests, their Products and Services”. The main objective of COST Action FP1101 is to increase the acceptance of timber in the design of new structures and in the repair of existing structures by developing and disseminating methods to assess, reinforce and monitor them. The Action is structured into three working groups: Working Group 1: Assessment of timber structures, Working Group 2: Reinforcement of timber structures, Working Group 3: Monitoring of timber structures. This report has been prepared by members of Working Group 2. It is a distillation of the work presented and discussed by members at a series of workshops over the course of the Action. These workshops considered the latest research in reinforcement technologies for both modern and historical timber structures, current and proposed design methods and also case studies that illustrate the current state-of-practise for on-site implementation of these technologies. Gratitude is expressed to the COST Office for funding these workshops and the publication of this report. The contribution of Rotafix UK, Rothoblass and SPAX to the technical activities of COST Action FP1101 and their financial support towards the publication of this report is also gratefully acknowledged. Finally, special thanks are due to the authors, reviewers and all working group members who contributed to the preparation of this report.

Annette M. Harte, Philipp Dietsch, Editors

v

vi

Reinforcement of Timber Structures

Introduction Harte, A.M., Dietsch, P.

The maintenance and rehabilitation of existing buildings is an area of increasing importance not only for economical but also for environmental, historical and social reasons. In Europe, as over 80% of buildings are over 50 years old, currently about 50% of all construction is connected to the repair, maintenance, and improvement of existing buildings. The need for structural reinforcement of timber buildings may become necessary for many reasons including change of use, deterioration due to a lack of monitoring and maintenance, accidental damage, changes in building regulations, the requirement to increase seismic resistance or due to the expiration of the planned lifetime. Depending on the situation, the function of the reinforcement may be to restore a weakened structure to its original load bearing capacity (repair) or to increase the load bearing capacity of an intact structure (upgrade). A large range of reinforcement methods is currently available and further technologies are under development. Reinforcement approaches include the addition of metallic, wood-based or FRP reinforcing plates attached to the timber member, the use of mechanical fasteners such as self-tapping screws and glued-in rods, replacement of decayed material using prostheses, adhesive/resin repairs, and more recently nanotechnology. Many of these methodologies are however, not harmonised for their use in-situ nor adapted depending on whether the structure is part of the regular building stock or belongs to cultural heritage. Special consideration is required for buildings that are classified as part of the cultural built heritage. For these buildings, the reinforcement strategy adopted should aim to minimise the removal of original material, conserve the original function of the structure and be reversible. In the past, due to a lack of knowledge on how to select and implement appropriate reinforcement methods, some of the interventions have caused further damage to the structures. For all building interventions, it is important to minimise the disturbance not only to the building and but also to its occupants. The main aim of this report is to summarise the current and emerging methods that are available to repair or enhance the structural performance of timber structures and to provide guidance to the use of these methods. The report is organised in two main parts. In Part I, the different structural elements and subsystems that make up our buildings are considered. These include beams, floors, columns, shear walls and connections. For each of these, the types of failure present in the elements are described and the appropriate reinforcement strategies for each case are presented, including consideration of cultural heritage issues. The reinforcement of buildings to increase their resistance to seismic actions is also included in this part. The focus of Part II is on reinforcement materials and methods. These include adhesive systems, mechanical fasteners such as glued-in rods and self-tapping screws, fibre reinforced polymer (FRP) laminates and bars, and emerging nano-structured materials. 1

Reinforcement of Timber Structures The properties of these materials, their methods of application and relevant design rules are described. The report provides details not only of the latest research findings related to the reinforcement of timber structures, but most importantly how these methods can be best used in practice. Many examples are given of the implementation of the various reinforcement methods. Because of this, the report will be of interest not only to the research community, relevant standardisation bodies, and policy makers but also to practitioners, such as architects, structural engineers and builders, representatives of the timber construction and building industry and product developers in the sector of reinforcement technologies. In conclusion, the reinforcement of existing timber structures is important from economic, environmental, historical and social perspectives. However, until now knowledge on the range of reinforcement techniques available, on the criteria for selection of the best approach for a given situation and on the correct procedures for implementing these methods on site has been defragmented. It is hoped that this report will address these shortcomings thus constituting distinct progress in the field of reinforcement of timber structures at a European level.

2


PART I: REINFORCEMENT APPLICATIONS

3


4

Reinforcement of timber beams

1 Reinforcement of timber beams Steffen Franke1, Bettina Franke2, Annette M. Harte3

Summary High performing, such as highly-loaded, and large span timber beams are often used for sports and industry halls, public buildings or bridges, and provide an aesthetically pleasing and environmentally friendly structural solution. Reinforcement of beams may be required to extend the life of the structure due to deterioration or damage to the material or due to a change of use. The main aim of this chapter is to summarise the current and emerging methods that are available to repair or enhance the structural performance of timber beams. An overview of the main materials, cross sections and geometries used for timber beam structures is presented. Furthermore, their general failure modes are described and typical reinforcement methods and corresponding retrofitting techniques are given. For each of the failure modes, the methods and their advantages are summarized. The reinforcement methods include wood to wood replacements, use of mechanical fasteners such as screws and rods, and methods which add additional strengthening materials.

1

Typology of timber beams

Timber beams can mainly be classified according to the span, the geometry and the material used, as summarized in Tab. 1. The focus here is on high-performance, long-span structures. Tab. 2 gives an overview of typical timber beams in relation to the sizes of the cross section and the span ratio. In Europe, glulam members or block glued glulam members are the main construction elements used for large open span spaces, stadiums or bridges in which the primary structure is timber. The typical layered cross section of glulam reaches from 100 to 250 mm in width and up to 2500 mm in depth but also in bigger dimensions as block glued glulam. Box or composite beams are alternatives providing a lower self-weight. Tab. 1 Classification of timber beams Material Solid wood Glulam, Block glued glulam Laminated veneer lumber Plywood (OSB, LSL) Cross laminated timber

Cross section Solid cross section Box-Beam I-Beam T-Beam C-Beam

Geometry Straight Beam Curved Beam Tapered Beam Truss

1)

Professor for Timber Structures, Bern University of Applied Sciences, Biel/Bienne, Switzerland

2)

Research Associate, Bern University of Applied Sciences, Biel/Bienne, Switzerland

3)

Senior Lecturer, College of Engineering and Informatics, National University of Ireland, Galway

5

Reinforcement of Timber Structures Tab. 2 Overview of timber beam forms Timber beam form Straight beams

Cross section depth h

10 m ≤ l ≤ 40 m h ≈ l / 17

depth h

12 m ≤ l ≤ 25 m h ≈ l / 15

span l

Tapered beams

span l

Curved beams

depth h span l

depth h

Trusses

Span, Depth ratio

15 m ≤ l ≤ 35 m h ≈ l / 17

20 m ≤ l ≤ 85 m h ≈ l / 10

span l

2

Failure modes

2.1

General

Structures have to adopt, and transfer external loads to the ground and also deal with internal loads. This leads to stresses and deformations in the structure which must not exceed design strength and deformation limits. In designing new structures, a full cross section with minor damage and correct material grades are assumed. However, in existing timber structures the cross section and/or the properties of the material of the members can be reduced due to mechanical and biological damage. Both types of damage influence the load carrying capacity and serviceability of single members or the complete construction. Within the assessment of timber structures, damage or failure has to be detected and assessed for the resistance and serviceability of the timber structure. The net cross sections observed must be compared to the designed cross sections. The failure analysis on timber structures in Germany carried out by Blass & Frese [1, 2] gives a good overview of the distribution of main types of failure classified according to the construction, use and region. Most assessment reports state that the timber structures have been built using glulam beams of quality GL28h (see Tab. 3). Their shape, however, is more varied with the most common being, by order: straight (154/426), tapered straight (124/426), pitched cambered (90/426) and curved (47/426). 80 % of the failure cases could be detected in bending members, followed by 8 % in compression members. Furthermore in 75 % of the failure cases cracks could be detected. Typical reasons and types of failure are summarized in Fig. 1 and Fig. 2 . Tab. 3 Most frequent characteristics of the timber structures assessed, from [1] and [2] Characteristic Material Quality (or equivalent) Load type

6

Main result Glulam GL28h Bending

Corresponding no. of assessments 594 80% 68 72% 470 80%


Biological reasons 36%

Crack in the grain direction 75%

StructuralPhysical reasons 38%

Other reasons 12%

Mechanical reasons 9%

Chemical reasons 5%

Shear failure 8%

Fig. 1 Reasons for damage [1], [2]

Tension failure 6%

Other failure 11%

Fig. 2 Types of failure [1], [2]

For high performing and long span timber members the typical failure modes are described in detail in the following sections.

2.2

Cracks

The most common type of failure, Fig. 2, was observed as the appearance of cracks in grain direction. The variation of the surrounding climate at a timber beam changes the moisture content and lead to shrinkage or swelling of the cross section. Non uniform distributions of the moisture content over the cross section and/or restraint deformations lead to internal stresses and, if the material strength is exceeded, to cracks in the cross section which can significantly reduce the capacity, Fig. 3. For the determination of the influence of cracks in timber beams on the load carrying capacity or stiffness no comprehensive methods are known. Methods and guidelines for this evaluation are currently under development at the Bern University of Applied Sciences. The amount and distribution of cracks depends on several factors, such as timber and glue quality, defects, loading situation or beam shape. Regarding the distribution of cracks in the timber beams, a summary of their characteristics can be found in Tab. 4Note: Failure under tension stress perpendicular to the grain in glulam members has to be distinguished from delamination failure within the glue lines as shown in Fig. 5. Special techniques can be used for the classification of delaminations as shown in [3].

a)

b)

c)

Fig. 3 Glulam cross sections; a) sketch of undamaged cross section with 100% capacity, b) sketch of cross section with some possible damage and unknown capacity, c) real example of cross section with internal cracks

7


Fig. 4 Cracks in grain direction at a glulam member

Fig. 5 Delamination at a glulam member

Tab. 4 Characteristics of cracks and their distributions [1], [2] Location; Quantity

Length; Depth ratio*

At the singularity; Single

1-10 m; mostly 1.0

35%

Normal climate changes

Randomly; Numerous

0.1-1 m; 0.1 to 0.4

33%

Element quality (Glue line or finger joints) Overloading (Shear or bending stresses)

At the defect; Depending on its extent Various; 1 m; Single to numerous mostly 1.0

Crack cause Stress concentration (Restrained shrinkage, notches, transverse forces ...)

Cases

17% 15%

* Ratio of depth of crack to width of beam

2.3

Bending failure

Bending results in longitudinal tension and compression stresses distributed over the depth of the cross section. The tension stress leads to a brittle failure due to the rupture of the wood fibres, as shown in Fig. 6. Longitudinal compression stress results in elastic and plastic deformations which can be described as ductile and leads to the so-called kink bands.

Fig. 6 Principal sketch for bending failure

8

Fig. 7 Tension failure under bending

Reinforcement of timber beams Due to natural defects, such as knots, the tension strength can be reduced compared to compression strength. Therefore, bending failure is mainly described by brittle failure of timber beams within the tension zone, as shown in Fig. 7. Bending failure is classified as critical and can lead to a single failure of the structural element or the complete construction.

2.4

Compression failure

Failure under longitudinal compression stress occurs mainly in timber beams or columns. Failure under compression stress perpendicular to the grain can also be described as a ductile failure with plastic deformations and occurs mainly at supports or at loading points where high loads have to be transferred, as shown in Fig. 8 and Fig. 9. In both cases, these plastic deformations can further lead to eccentricities and load redistributions within the complete structure and therefore overstress parts of the structure. The overall stability will also be influenced.

Fig. 8 Principal sketch for compression failure at support

2.5

Fig. 9 Compression failure at loading point

Tension failure

Fig. 10 Principal sketch for tension failure perpendicular to the grain at a notch

Fig. 11 Tension failure perpendicular to the grain at a notch

9


a)

b)

Fig. 12 Tension failure parallel to the grain of experim. tension tests: a) short-fibred, b) long-fibred, [4] Tension stress has to be considered in the parallel to the grain and perpendicular to the grain directions. When the tensile capacity of the timber is exceeded, brittle failure occurs. Examples of tensile failure parallel to the grain and perpendicular to the grain are shown in Fig. 10, Fig. 11, and Fig. 12. However, due to the low tension strength perpendicular to the grain of solid wood and glulam members, which is almost zero due to natural defects, failure under tension stress perpendicular to the grain occurs more often. Therefore, wood products are mostly optimized to increase the tension strength perpendicular to the grain, but it still has carefully to be considered in the design process. Tension stress perpendicular to the grain occurs in curved, tapered and end-notched glulam members as well as in members with holes, additional connected structural elements or equipment, and at connections loaded perpendicular to the grain.

2.6

Shear failure

In most cases, bending stress and deflection limits govern the design of the members. But for short beams the shear stress can be more important. In general for beams, the shear stress reaches the maximum value close to the supports (Fig. 13). Additionally, end-notched beams and beams with holes can lead to shear stress concentrations. Failure due to shear stress is characterised by a sliding of the fibres and thus cracking parallel to the grain and is considered as a brittle failure. The cracks are mainly closed and therefore hard to detect if they are not at the end of the beam, as shown in Fig. 14.

Fig. 13 Principal sketch for shear failure

10

Fig. 14 Shear failure at holes

Reinforcement of timber beams 2.7

Insects and fungi

Decay due to fungi is possible for timber beams with a moisture content close to/or over the fibre saturation point, see 5. The fibre saturation point varies from wood to wood species and shows a range from 26 M% - 32 M% (in mass percentage). The different fungi and their typical appearance and hazard are summarized in [5]. Generally, decay due to insects can occur within a range of wood moisture content above 6 M%, but can be neglected in construction of service class 1 or 2, where technical dried wood members like solid wood, glulam or wood products are used, [6]. The classification and identification of insects is described in detail in [5].

Fig. 15 Risk of insect and fungal decay in relation of the moisture content

3

Retrofitting techniques

3.1

General

The following sections illustrate possible retrofitting techniques for timber beams. Detailed descriptions of the different techniques and their design can be found in other chapters of this report. In the case of damage or decay, the timber beam or parts of the beams have to be replaced as described in Section 3.3. Retrofit measures to improve the performance of timber beams in bending, shear, and in tension and compression perpendicular to the grain are described.

Fig. 16 Risk of insect and fungal decay in relation of the moisture content

11


3.2

Repair of shrinkage cracks or delaminations

Repairs of shrinkage cracks or partial delamination of glue lines in glued laminated timber may be carried out; however, it is not advisable to repair cracks in solid timber members. In many instances, repair of cracks is carried out in conjunction with other reinforcement interventions. For glued laminated timber, it is generally considered necessary to repair shrinkage cracks in regions of high shear stresses and high tensile stresses perpendicular to the grain. The first purpose of the repair is to restore the load carrying capacity of the glulam member. The visible cracks or delamination in glulam members have always to be assessed by an expert before planning the repair process regarding the load carrying capacity of the whole structure. For crack openings smaller than 10 mm wide and with low to medium fibrosity/splintering the repair can be done by injection of adhesives, [7]. A number of such products with technical approvals for these applications are available. The current regulations according to the requirements in the standards and possible technologies are summarized in [7] for the European market. The methodology of repairing cracks and delamination is also described in [7]. To ensure adequate carrying capacity, the preparation of the bonding surfaces is important to ensure the required quality of the applied technology and to avoid defects. In general, the repair procedure involves cutting out the cracks using a saw, router or grinder to make a clean slot. The slot is then cleaned, optionally brushed with a primer and preparation of filling and ventilation holes as well as supporting bracing system before filling with a suitable adhesive. Fig. 16 shows three different adhesives and various technology for repairing cracks or delamination. For combination of repair methods, the barrier effect of repaired bonded joints against the ingress of water and water vapour has to be considered during the maintenance planning.

3.3

Replacement of damaged or decayed parts

Timber that has decayed due to fungal or insect attack is porous, brittle and has very poor strength properties [8]. This decay often occurs in localised parts of the beam, such as at the ends where the timber is in direct contact with a masonry supporting wall as seen in Fig. 17. In these cases, the condition of the rest of the beam is generally good. Other types of accidental damage, such as fire damage, may cause a reduction in the member cross-section resulting in inadequate strength and stiffness. Decayed and damaged material should be removed and the member upgraded to restore the load-bearing capacity of the member. Repair methods include replacing the damaged section with a timber or engineered wood prosthesis connected to the original beam by means of a scarf joint with wooden pegs and/or adhesive (Fig. 18), by means of bonded-in rods or plates (Fig. 19) or replacing the damaged section with a prosthesis built up from timber boards using screws as shown in Fig. 20 [10], [11]. The most common retrofit method employed involves replacing the damaged timber with a timber prosthesis which is bonded to the sound timber in the original beam using steel or fibre reinforced polymer (FRP) rods or plates. Using this approach, the scale of the intervention is limited and the load bearing function is preserved. The implementation of this type of repair involves a number of different steps [12]. Initially, the beams are propped. The damaged part of the beam is then removed by cutting either vertically or at an angle of 45˚ to the vertical, as seen in Fig. 19 and Fig. 21. Holes or grooves to take the connecting rods/plates are drilled in the beam and the prosthesis and are partially

12


Fig. 17 Beam end decay [9]

Fig. 18 Prosthesis with scarf connection

Fig. 19 Beam end repair using bonded in rods or plates

Beam Prosthesis

Screws

Fig. 20 Beam end repair using bonded in rods or plates

Fig. 21 GFRP rods inserted in ends of propped beams [13]

13


filled with adhesive. The reinforcing elements are inserted into the beam and prosthesis and the adhesive is topped up in holes/grooves if necessary. Additional props are introduced to support the prosthesis. When grooves are used, a timber strip is normally inserted to improve the appearance of the repaired member and to provide fire protection. When the adhesive has fully cured, the supporting props are removed. The prosthesis should be of the same species as the timber to be repaired, or be compatible in terms of its mechanical properties by using, for instance, engineered wood products. The moisture content of the prosthesis should be the same as that of the beam being repaired. The adhesive used is usually a thixotropic epoxy resin and the type used should be specially formulated to bond with the timber and the reinforcement. The design of the repair is based on the requirement that the reinforcement should provide the same load bearing capacity as the section with sound timber.

3.4

Flexural reinforcement

In order to increase the flexural strength and stiffness of beams, reinforcing elements are added that act compositely with the existing member. A large variety of reinforcement configurations is available. The reinforcing elements can be in the form of rods, plates or other structural shapes which are connected to the beam using mechanical fixings or structural adhesives. These reinforcing elements can be placed inside or outside of the member and may be slack or prestressed. The reinforcement material can be a metal, fibre reinforced polymer (FRP) or engineered wood product. Fig. 22 and Fig. 23 show some possible configurations for external and internal reinforcement. Apart from the structural requirements, the configuration selected for a particular application may depend on other factors: the presence of decorative ceilings or painting on beams may require that access for reinforcement is restricted to the top or sides of the beam; fire protection requirements may exclude the use of externally bonded plates on exposed surfaces. As timber beams generally fail in tension in a brittle fashion, positioning of the reinforcement on the tensile face of the beams is very effective for increasing bending strength. With increasing percentage of tensile reinforcement, the neutral axis moves towards the bottom of the beam. As a result the compressive strain in the timber increases relative to the tensile strain and compressive yielding may occur before the timber eventually fails in tension. The load-deflection response for a timber beam reinforced with carbon fibre reinforced polymer (CFRP) plates, which were inserted from the top, is shown in Fig. 24, [14]. The unreinforced beam A has a brittle response. For the reinforced beams, two of the beams display significant ductility in their response before failure. Kliger et al. [15] investigated the influence of the distribution of the reinforcement between the tensile and compressive faces of the beam on the bending strength and stiffness. They concluded that for maximum strength, 75% of the reinforcement should be on the bottom face and 25% on the top. To achieve maximum ductility, all of the reinforcement should be placed on the bottom. The maximum stiffness enhancement was achieved when the reinforcement was equally distributed between the top and bottom faces. However, for low percentages of reinforcement the stiffness gain by distributing the reinforcement between the two faces may not justify the additional work involved.

14


Fig. 22 External reinforcement arrangements

Fig. 23 Internal reinforcement arrangements

Fig. 24 Load-deflection response of unreinforced (A) and reinforced (D1, D2, D3) beams, [14] Steel and other metals have been used for reinforcing timber for many years. Mark [16] bonded aluminium sheets to the top and bottom faces of timber beams and reported an increase in the flexural strength and stiffness. Dziuba [17] tested timber beams reinforced with steel rods on the tension face and noted that compressive yielding occurred prior to failure in tension. DeLuca and Murano [18] reinforced spruce beams with 0.82% steel bars and recorded mean increases of 48% in peak load, and 26% in stiffness. Nielsen and Ellegaard [19] investigated the use of punched metal plate connectors as flexural reinforcement for timber but with limited success. Fibre reinforced polymer in the form of pultruded rods or plates have been the subject of a considerable amount of research for the reinforcement of timber and have been used in practise to reinforce solid timber and glulam structures. Several fibre types are available including carbon (CFRP), aramid (AFRP), glass (GFRP), basalt (BFRP) and steel (SFRP). CFRP [14], [15], [20]-[22] and GFRP [23]-[26] have been widely used as externally bonded plates or internally as near surface mounted reinforcement bonded into grooves cut into the beams. CFRP materials have high strength and stiffness properties and, depending on the properties of the unreinforced beam and the percentage of reinforcement used, strength and stiffness increases of over 100% can be achieved. For lower grade timber, less expensive GFRP materials are generally sufficient to provide the required strengthening but the stiffness increase can be limited. Steel fibre reinforced polymer bars have been found to provide a significant increase in capacity and ductility but insignificant improvement in stiffness [27].

15

Reinforcement of Timber Structures The use of FRP materials has a number of advantages over steel due to their light weight, their corrosion resistance, and their ease of handling on site. It should be noted that the routing of grooves to house reinforcement may cause a weakening of the beam as a discontinuity is introduced in the wood fibres in the vicinity of the grooves. Prestressed steel or FRP plates bonded on the tension face with epoxy resin [28]-[32] can provide further increases in strength. A pre-camber is introduced in the beam due to the eccentric prestress, which can be offset against the deflection to the external loads. However, this technique is currently not used in practise due to the difficulty in installation and insufficient knowledge of the long-term performance of the prestressed members. As the flexural capacity of the beam is enhanced, the shear capacity may be exceeded. In these cases, a combination of both flexural and shear strengthening may be required.

3.5

Reinforcement in tension perpendicular to the grain

Failure in tension perpendicular to the grain in timber beams can arise in notched beams, around holes and in curved, tapered or pitched cambered beams. Reinforcement of beams in these situations can be achieved using internal or external reinforcement. Types of internal reinforcement include selftapping screws, bonded-in or drilled-in threaded steel rods or bonded-in FRP rods or tubes. External reinforcement is achieved by mechanically fixing and/or gluing on sheets of wood-based panels, such as plywood, or FRP sheets or nail plates. For the case of notched end beams, the stress concentration at the corner of the notch leads to crack initiation and rapid crack propagation results in a sudden brittle failure of the beam as shown in Fig. 12. The high tensile stresses perpendicular to the grain are accompanied by high shear stresses. Different reinforcement methods are illustrated in Fig. 25. The reinforcement can be deployed perpendicular to or at 45˚ to the beam axis. Due to the presence of high shear stresses, the performance of notched beams reinforced at 45˚ is superior. This has been validated by a number of experimental investigations [34]-[37]. Reinforced notches have enhanced load-bearing capacity but also display less brittle failure modes than is the case for unreinforced notches. For screws or glued-in rods, the requirements for minimum edge distances and spacing must be satisfied while keeping the reinforcement as close as possible to the notch corner. Externally bonded sheets of FRP or plywood are placed on both sides of the beam and extend over the full height. Screws or nails are normally used to provide the required bonding pressure while the adhesive is curing. Irrespective of the type of reinforcement used, the usual design approach is to assume that the tensile forces perpendicular to the grain are carried entirely by the reinforcement. For beams with a rectangular notch at the support, the tensile force for which the reinforcement is designed is given by [34] 2 3 Ft ,90,d  1,3 Vd 31     2 1      

(1)

where Vd is the design shear force and α is the ratio of the reduced beam height at the notch to the total beam height. For the case of beams with a round or rectangular hole,

16


Ft ,90,d 

Vd hd 4h

 hd2  Md 3  2  0,008  hr  h 

(2)

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 25 Typical reinforcement arrangements for notches. (a) & (b) self-tapping screws, (c) & (d) glued-in rods, (e) & (f) EWP or FRP side plates

Fig. 26 Reinforcement of openings

Fig. 27 Reinforcement of pitched camber beams where Vd and Md are the design values of the shear force and bending moment at the section, respectively, h is the beam height, hd is the hole height and hr is the distance from the edge of the hole to the top or bottom of the beam [34]. Typical reinforcing configurations for beams with holes are shown in Fig. 26. For curved, pitched tapered or tapered beams, the tensile stress perpendicular to the grain occurs in the apex region, which is highlighted in grey in Fig. 27, [39]-[41]. Reinforcement of this region can be achieved through the use of screws, glued-in rods or side plates, as shown in Fig. 27. The design tensile stress, σt,90,d, may be calculated as

 t ,90,d  k p

6M ap ,d 2 b  hap

(3)

17

Reinforcement of Timber Structures where Map,d is the design moment at the apex, hap is the depth of the beam at the apex and kp is a function of the taper angle, the radius and the depth at the apex [42]. The load to be carried by discrete connectors, such as screws or glued in rods, is the total of all tensile stresses on an area equal to the connector spacing by the beam width. The load capacity of the connectors is determined by the withdrawal capacity and the tensile strength. The reduced section of the beam due to drilling of holes for the radial reinforcement must be considered in the design. As the portion of the holes below the neutral axis cannot be considered effective in tension, the section modulus in bending is reduced [41].

3.6

Shear reinforcement

The methods available to strengthen beams in shear are the same as those described for reinforcement against tension perpendicular to the grain. These include internal reinforcement in the form of screws and bonded-in rods of steel or FRP and external reinforcement in the form of side plates. Akbiyik et al. [43] investigated the shear reinforcement of timber stringers with horizontal splits using hex bolts, lag screws, and plywood and GFRP side plates. The bolts were epoxy bonded in vertical holes drilled from the top. The lag screws were installed vertically and at 45˚. The plywood and GFRP side plates were attached to the sides of the beams using screws. All repair types were effective with an average increase in the residual shear capacity of 62%. None of the repaired specimens recovered the original undamaged stiffness. The extent of the existing damage had a big influence of the effectiveness of the repair. For the more highly damaged stringers, the use of GFRP side plates was the most efficient method. Several investigators examined FRP shear reinforcement of beams [37], [43]-[46]. Radford [44] reported an increase in stiffness of over 270% when using epoxy bonded side plates of GFRP with the fibres oriented at +/-45˚ to the beam axis. The use of vertical GFRP shear spikes produced a stiffness enhancement of over 160%. Inserting the shear spikes at a spacing equal to the beam depth was found to be the most effective. Gentry [46] also used a combination of FRP flexural plate and FRP shear pins to reinforce glulams. Svecova and Eden [47] used GFRP bars to reinforce beams from a bridge. This resulted in a significant increase in strength and decrease in variability. Widmann et al. [37] investigated the shear reinforcement of delaminated glulam beams. Glulam beams with delaminated middle lamellae were loaded to failure and then reinforced with self-tapping screws or epoxy bonded CFRP side sheets oriented at 45˚ to the beam axis. Both approaches showed a significant increase of the shear strength. The ultimate shear strength could not be determined as different failure modes were found. Trautz and Koi [48] described a series of tests performed on glued-laminated beams reinforced with screws using different arrangements to carry tensile and compressive forces. The shear stiffness of beams reinforced with screws arranged in a nested pattern with screws carrying loads in tension and compression was superior to that achieved by reinforcing with diagonal tension screws only. Dietsch et al. [49] describe design approaches for the shear reinforcement of timber beams in the unfractured and fractured states. The types of reinforcement considered are self-tapping screws or threaded rods deployed at an angle to the beam axis. The models account for the enhancement in shear performance resulting from compression induced perpendicular to the grain by the reinforcement. Comparison with results of experimental tests provided the validation of the shear stiffness and strength predictions. For unfractured beams, an increase in capacity of 20% is achievable

18

Reinforcement of timber beams when reinforced with threaded rods. For fractured beams, the maximum increase in bending stress compared to the intact beam is 33%.

3.7

Reinforcement in compression perpendicular to the grain

Crocetti et al. [50] undertook experimental investigations of the compressive strength perpendicular to the grain of glulam beams reinforced internally with glued-in steel rods and glued-in hardwood dowels and externally with screwed-on steel side plates. The beams reinforced with glued-in rods all failed in buckling and resulted in a significant increase in both strength and stiffness over the unreinforced beams. The beams reinforced with side plates also produced a significant enhancement in the compression strength. Blass and Bejtka [34] proposed a design model for the compressive capacity, R90,d, of a beam support reinforced using self-tapping screws that accounts for buckling and screw withdrawal. This may be expressed as  nRd  kc,90  lef ,1  b  f c,90,d   R90,d  min   b  lef ,2  f c,90,d   

(4)

where n is the number of screws, Rd is the lesser of the withdrawal capacity and the buckling capacity of the screw, fc,90,d is the design value of the compressive strength perpendicular to the grain, b is the beam width, lef,1 and lef,2 are the lengths as defined in Fig. 28, and kc,90 is a load distribution coefficient in the range [1;1.75] In existing structures, the insertion of screws or glued-in rods at locations of concentrated loading may be difficult to achieve.

Fig. 28 Reinforcement for compression perpendicular to the grain

3.8

General remarks

In practical terms, the choice of reinforcement method for existing timber beams will be based not only on the ability of the reinforcement to provide adequate strengthening of the structure but will be constrained by other factors such as aesthetics, need for reversibility, access for repair, and available expertise. As the reinforcing elements generally have different stiffness, thermal expansion and moisture absorption properties than the timber element, factors such as constrained shrinkage and swelling due to thermal or moisture changes must be considered and if necessary additional thermal or moisture induced stresses should be accounted for in design. Where the reinforcement results in a significant change in beam stiffness, it is important to consider the consequences for the overall behaviour and load distribution of the entire structure.

19


4

Case studies

4.1

Clyne Castle, Wales – Replacement of decayed parts [13]

Clyne Castle is a Grade II listed building near Swansea in Wales, which was originally built in 1791 but which has had numerous annexes added over the intervening years resulting in a complex roof structure. Failure of the roof valley drainage system resulted in prolonged exposure of the roof structure to moisture leading to wet rot in the span beams, hip rafters and ceiling joists. An upgrade procedure was required that did not interfere with the ornate suspended ceiling. The solution that was adopted was to replace the decayed timber in-situ with a prosthesis made from laminated veneer lumber (LVL). Due to the restriction on access, the connection between the prosthesis and the hip rafters was effected using slots routed in the sides of the prosthesis and the rafter into which six 16 mm diameter high tensile steel rods were bonded using a two-part epoxy structural adhesive. Fig. 29 shows the prosthesis in place with the rods inserted in the slots prior to topping up with adhesive. The span beams were repaired using an LVL prosthesis that was attached to the beams using six 20 mm diameter steel rods. The rods were bonded into holes drilled into the end grain of the beams and into matching side slots in the prosthesis, as shown in Fig. 30. Due to the lighter loading carried by the ceiling joists, it was sufficient to use a C24 softwood timber prosthesis that was bonded to the joists using two 12 mm GFRP rods. Fig. 21 shows the rods inserted in the joists before adding the prosthesis.

Fig. 29 Hip rafter repair [13]

Fig. 30 Span beam repair [13]

Fig. 31 Long-term monitoring of Sins Bridge, [52]

20


4.2

Sins Bridge, Switzerland – Flexural reinforcement [51], [52]

Sins Bridge is a historic two-span timber arch bridge over Reuss River at Sins in Switzerland. It was originally built in 1807 and the eastern side was rebuilt after being blown up during the 1852 Civil War. It comprises two equal spans of 30.8 m and was designed for horse-drawn carriages. In 1992, it was upgraded to carry 20 tonne trucks. This involved the installation of a new 200 mm thick transversely prestressed timber deck and the strengthening of two transverse cross-beams with CFRP laminates. The cross-beams comprised two solid oak beams placed one on top of the other. The 1 mm thick CFRP laminates were bonded to the top and bottom surfaces using epoxy resin (Fig. 31). Two types of CFRP were used: One was a high modulus material (E = 305 GPa, tensile strength = 2300 MPa) and the other was a high strength material (E = 152 GPa, tensile strength = 2600 MPa). The reinforced cross-beams and a number of unreinforced crossbeams were instrumented with electrical resistance strain gauges and Demec gauges in order to monitor their long-term performance (Fig. 31). The deflection of the reinforced cross-beams was 20 - 50% lower than the unreinforced beams. Fourteen years after the original installation, the performance of the reinforced beams continued to be satisfactory [51].

5

Conclusion

Due to the impact of different aspects like moisture changes, fungi and insect attacks, timber elements can be damaged and result in lower capacity and larger deformations. Furthermore, high stresses exceeding the strength limits can also lead to different failure cases, like bending, compression, tension or shear failure, where cracks mostly appear. An analysis of several assessment reports has shown that most damaged structural timber elements show cracks in the grain direction. In the case of damage or decay, the timber beam or parts of the beams have to be replaced or reinforced in bending, shear, and in tension or compression perpendicular to the grain to recover the performance. The choice of reinforcement method for existing timber beam structures will be based not only on the ability of the reinforcement to provide adequate strengthening of the structure but will be constrained by other factors such as aesthetics, need for reversibility, access for repair, and available expertise. As the reinforcing elements generally have different stiffness, thermal expansion and moisture absorption properties than the timber element, factors such as constrained shrinkage and swelling due to thermal or moisture changes must be considered and if necessary additional thermal or moisture induced stresses should be accounted for in design. Where the reinforcement results in a significant change in beam stiffness, it is important to consider the consequences for the overall behaviour and load distribution of the entire structure. Many factors have to be considered selecting the best method and sometimes lack of knowledge exists. Ongoing research needs to be done to further improve the retrofitting and replacing methods for reliable and efficient repair.

21


6

Acknowledgement

Parts of the report and research work are within the COST Action FP 1101 – Assessment, Monitoring and Reinforcement of timber structures. The Swiss State Secretariat for Education, Research and Innovation (SERI) proudly supports the research work.

7

References

[1]

Blass, H. J., Frese, M., “Failure analysis on timber structures in Germany”, In: Kohler, J., Fink, G., Toratti, T., Assessment of failures and malfunctions, Shaker Verlag, Germany, 2011. Blass, H. J., Frese, M., Failure analysis on timber structures (only in German: Schadensanalsyse von Hallentragwerken aus Holz), KIT Scientific publishing, Karlsruhe, 2010. Franke, B., Scharmacher, F. Müller, A. “Assessment of the glue-line quality in glued laminated timber structures”, Advanced Materials Research. Vol. 778, 2013, pp. 424-431. Dietsch, P., Gamper, A., Merk, M., Schopbach, H. „Versagensmechanismen im Holzbau“, DVD, TU Munich, Germany, 2012. Zabel, R. A.,Morrell, J. J., Wood Microbiology: Decay and Its Prevention, Academic Press Limited, London, UK, 1992. Aicher S., Radovic, B., Folland, G. Probability of decay by house longhorn beetle on glued laminated timber (only in German: Befallswahrscheinlichkeit durch Hausbock bei Brettschichtholz), IRB-Verlag, Germany, ISBN 978-3-8167-5977-5, 2001. Lehringer, C., Salzgeber, D., “Repair of cracks and delaminations in glued laminated timber”, in Franke, B., Franke, S. Cost Workshop – Highly performing timber structures: Reliability, Assessment, Monitoring and Strengthening, Switzerland, 2014. Wheeler, A.S., Hutchinson, A.R., “Resin repairs to timber structures”, International Journal of Adhesion & Adhesives, Vol. 18, 1998, pp. 1-13, 1998. Tampone, G., “Mechanical failures of the timber structural systems”, In: Proceedings of ICOMOS IWC - XVI International Symposium, Florence, Venice and Vicenza, 2007. Pizzo, B., Schober, K.U., “On site interventions on decayed beam ends”, In: Core Document of COST Action E34, Bonding of timber, University of Natural Resources and Applied Life Sciences, Vienna, 2008. Menichelli, C., Adami, A., Balletti, C., Bertolini Cestari, C., Bettiol, G., Biglione, G et al., Le strutture lignee dell’arsenale di Venezia. Sudi et restauri., In Proceedings of XXV Convegno Internazionale Scienza et Beni Culturali – Conservare e restaure il legno, Bressanone, Italy, 2009. (In Italian) Smedley, D., Cruz, H., Paula, R., “Quality control on site”, In: Core Document of COST Action E34, Bonding of timber, University of Natural Resources and Applied Life Sciences, Vienna, 2008. Rotafix Ltd., “Clyne castle timber repairs. A case study.” Rotafix Limited, Swansea, UK Nowak , T.P., Jasienko, J., Czepizak, D., “Experimental tests and numerical analysis of historic bent timber elements reinforced with CFRP strips”, Construction and Building Materials Vol. 40, 2013, pp. 197–206 Kliger, R., Johansson, M., Crocetti, R., “Strengthening timber with CFRP or steel plates – short and long-term performance”, In: Proceedings of World Conference on Timber Engineering, Miyazaki, Japan, 2008. Mark, R. “Wood-aluminum beams within and beyond the elastic range. Part 1: Rectangular sections”, Forest Products Journal, Vol. 11, No 10, 1961, pp. 477-484. Dziuba, T., “The ultimate strength of wooden beams with tension reinforcement”, Holzforschung and Holzverwertung, Vol. 37, No. 6, 1985, pp. 115-119.

[2] [3] [4] [5] [6]

[7]

[8] [9] [10]

[11]

[12]

[13] [14]

[15]

[16] [17]

22

Reinforcement of timber beams [18] DeLuca, V., Murano, C., “A comparison of un-reinforced and reinforced timber glulam with steel bars”, European Journal of Technology and Advanced Engineering Research, Vol. 2, 2011, pp. 45-54. [19] Nielsen, J., Ellegaard, P., “Moment capacity of timber reinforced with punched metal plate fasteners”, In: Proceedings of 1st RILEM Symposium on Timber Engineering, Stockholm, Sweden, 1999. [20] Schober K.U., Rautenstrauch K., “Post-strengthening of timber structures with CFRPs”, Materials and Structures, Vol. 40, 2006, pp. 27–35. [21] Borri, A, Corradi, M, Grazini, A., “A method for flexural reinforcement of old wood beams with CFRP materials”, Composites Part B: Engineering, Vol. 36, 2005, pp.143-153. [22] Li, Y.F., Xie, Y.M., Tsai, M.J., “Enhancement of the flexural performance of retrofitted wood beams using CFRP composite sheets”, Construction and Building Materials, Vol. 23, 2009, pp. 411-422. [23] Hernandez, R., Davalos, J.F., Sonti, S.S., Kim, Y., Moody, R.C., “Strength and stiffness of reinforced yellow-poplar glued-laminated beams”, Research Paper FPL-RP-554, Department of Agriculture, Forest Service, Forest Products Laboratory, Madison, WI, US, 1997. [24] Gentile, C., Svecova, D., Rizkalla, S.H.,, “Timber beams strengthened with GFRP bars: development and applications”. Journal of Composites for Construction, Vol. 6, No. 1, 2002, pp. 11-20. [25] Raftery, G.M., Harte, A.M., “Low-grade glued laminated timber reinforced with FRP plate”, Composites: Part B, Vol. 42, 2011, pp. 724-735. [26] Raftery, G., Whelan, C., Harte, A., “Bonded-in GFRP rods for the repair of glued laminated timber”, In: Proceeding of World Conference on Timber Engineering (WCTE12), Auckland, New Zealand, 2012. [27] Borri, A., Corradi, M., “Strengthening of timber beams with high strength steel cords”, Composites: Part B. vol. 42, 2011, pp. 1480-1491. [28] Peterson, J., “Wood beams prestressed with bonded tension elements”, Journal of the Structural Division, ASCE, Vol. 91, No. 1, 1965, pp. 103-109. [29] Triantafillou, T.C., Deskovic, N., “Prestressed FRP sheets as external reinforcement of wood members”, Journal of Structural Engineering, Vol. 118, No. 5, 1992, pp. 1270-1283. [30] Lehmann, M., Properzi, M., Pichelin, F., Triboulet, P., “Pre-stressed FRP for the in-situ strengthening of timber structures”, In: Proceedings of World Conference on Timber Engineering (WCTE 2006), Portland, OR, USA, 2006. [31] Brunner, M., Schnueriger, M., “FRP-prestressed timber”, in: Proceedings of 8th International Symposium on Fiber Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-8), Patras, Greece, 2007. [32] Brady, J.F., Harte, A.M., “Prestressed FRP flexural strengthening of softwood glue-laminated timber beams”, In: Proceedings of World Conference on Timber Engineering (WCTE 2008), Miyazaki, 2008. [33] McConnell, E., McPolin, D, Taylor, S., “Post-tensioning of glulam timber with steel tendons”, Construction and Building Materials, Vol. 73, 2012, pp.426-433. [34] Blass, H.J., Bejtka, I., “Reinforcements perpendicular to grain using self-tapping screws”, In: Proceedings of 8th World Conference on Timber Engineering, Lahti, Finland, 2004. [35] Amy, K., Svecova, D, “Strengthening of dapped timber beams using glass fibre reinforced polymer bars”, Canadian Journal of Civil Engineering, Vol. 31, No. 6, 2004, pp. 943-955. [36] Jockwer, R., Frangi, A., Serrano, E., Steiger, R., “Enhanced Design Approach for Reinforced Notched Beams”, In: Proceedings CIB – W18 Meeting 46, Vancouver, Canada, 2013. [37] Widmann, R., Jockwer, R., Frei, R., Haeni, R., “Comparison of different techniques for the strengthening of glulam members”, In: Proceedings of COST Action FP1004 Early Stage Researchers Conference, Zagreb, Croatia, 2012.

23

Reinforcement of Timber Structures [38] Blumer, H., Spannungsberechnungen an anisotropen Kreisbogenscheiben und Sattelträgern konstanter Dicke, Lehrstuhl für Ingenieurholzbau und Baukonstruktionen, Universität Karlsruhe, 1972/1979. [39] Jonsson, J., “Load carrying capacity of curved glulam beams reinforced with self-tapping screws”, Holz als Roh- und Werkstoff, Vol.63, 2005, pp. 342-346. [40] Kasal, B., Heiduschke, A., “Radial reinforcement of curved glue laminated wood beams with composite materials”, Forest Products Journal, Vol. 54, No. 1, 2004, pp. 74-79. [41] American Institute of Timber Construction, Standard for radially reinforcing curved glued laminated timber members to resist radial tension, AITC 404, Centenniel, CO, US, 2005. [42] CEN, EN 1995-1-1: Eurocode 5: Design of timber structures – Part 1-1: General – Common rules and rules for buildings, 2004, CEN, Brussels, Belgium. [43] Akbiyik, A., Lamanna, A.J., Hale, W.M., “Feasibility investigation of the shear repair of timber stringers with horizontal splits”, Construction and Building Materials, Vol. 21, No. 5, 2007, pp. 991-1000. [44] Radford, D.W., Van Goethem, D., Gutkowski, R.M., Peterson, M.L., “Composite Repair of timber structures”, Construction and Building Materials, Vol. 16, No. 7, 2002, pp. 417-425. [45] Hallstrom, S., Grenestedt, J.L., “Failure analysis of laminated timber beams reinforced with glass fibre composites”, Wood Science and Technology, Vol. 31, 1997, pp. 17-34. [46] Gentry, T.R., “Performance of glued-laminated timbers with FRP shear and flexural reinforcement”, Journal of Composites for Construction, Vol. 51, No. 5, 2011, pp. 861-870. [47] Svecova, D., Eden, R.J., “Flexural and shear strengthening of timber beams using glass fibre reinforced polymer bars-an experimental investigation”, Canadian Journal of Civil Engineering, Vol. 31, No. 1, pp. 45-55. [48] Trautz, M., Koj, C., “Self-tapping screws as reinforcement for timber structures”, In: Proceedings of International Association for Shell and Spatial Structures Symposium, Valentia, Spain, 2009. [49] Dietsch, P., Kreuzinger, H., Winter, S., “Design of shear reinforcement for timber beams”, In: Proceedings CIB – W18 Meeting 46, Vancouver, Canada, 2013. [50] Crocetti, R., Gustafsson, P.J., Ed, D., Hasselqvist, F., “Compression strength perpendicular to grain – full scale testing of glulam beams with and without reinforcement”, In: Proceedings of COST Action FP1004 Early Stage Researchers Conference, Zagreb, Croatia, 2012. [51] Meier, U., “Restoration of a historic wooden bridge with CFRP”, FRP International, 2(1) pp 12 1993. [52] Czaderski, C., Meier. U., “Long-Term Behaviour of CFRP Strips for Post-Strengthening.” In: Proceedings of 2nd International fib Congress. Naples, Italy, June 5-8, 2006. Vol. 2. pp. 110112.

24

Reversible timber-to-timber strengthening interventions on wooden floors

2 Reversible timber-to-timber strengthening interventions on wooden floors Alessandra Gubana1

Summary The goal of the State-of-the-Art-Report on High Reversible Strengthening Interventions on Timber Floors is to give an almost complete overview of the new techniques developed and tested to achieve in-plane and out-of-plane stiffness upgrading by means of less invasive and reversible interventions. In recent years a growing sensibility towards the preservation and maintenance of heritage buildings has led researchers to test different dry retrofitting systems. The chapter focuses on strengthening interventions based on the use of timber or timber based elements: the most adopted or most promising techniques are so briefly described. While the problem of upgrading the bending stiffness has a solid analytical background in the concrete-to-timber composite section theory, the problem of upgrading the in-plane stiffness still requires attention. The importance of the correct evaluation of the in-plane mechanical properties of floor timber diaphragms is nowadays clearly assessed in order to determine the building structural response under lateral seismic loads. In some codes, simplified analytical procedures are proposed to determine the in-plane stiffness, but generally with regard to new timber building floor typologies. The tests described in literature are generally referred to different setups, test rigs, boundary conditions, aspect ratio of the floor samples and also the recorded parameters are not always the same ones: it is so difficult to compare the experimental data as to achieve a general stiffness evaluation approach.

1.

Introduction

Existing timber floors often require strengthening. The main problem is their low stiffness, which results in high bending deformations and vibrations under service loads. Permanent deflection due to creep can also reach critical values. Moreover, in earthquake prone areas, if seismic resistance has to be assured in existing masonry buildings, floor diaphragm behaviour must be achieved. The importance of an effective diaphragm action in the floors of multi-storey masonry buildings is wellknown in earthquake engineering. Thanks to the diaphragm action, in fact, the floors can transfer the forces due to earthquakes to the lateral load resisting systems [1]. One of most widely used and effective techniques for strengthening floors is based on the connection of a new 40-50 mm height concrete topping on the existing timber joists. Different types of metallic fasteners, notched shear keys or slotted in perforated plates generally assure the effective collaboration between the two different materials [2,3,4,5,6,7,8,9].

1)

Associate Professor, University of Udine, Udine, Italy

25

Reinforcement of Timber Structures The new composite section ensures a significant floor stiffness upgrade, while the concrete topping connected with the vertical walls is able to give an effective three dimensional behaviour of the masonry buildings and therefore it markedly improves the lateral load resistance. The use of concrete also allows load distribution to take place and provides acoustic and fire insulation. On the other hand, a thin concrete slab adds undesirable additional weight on the floor, and consequently the seismic actions and the foundation loads increase. This technique, while simple and very efficient, is now often considered not sufficiently reversible: in particular in Italy it is frequently not authorized by the Cultural Heritage Offices, therefore it may be not allowed in buildings of historical value. The increased sensitivity towards the conservation of cultural heritage leads to the adoption of restoration techniques which can guarantee as much as possible the maintenance of the building authenticity and integrity, the conservation of the materials and of the original function of the structure, the reversibility of the intervention and its compatibility with the existing parts of the buildings [10,11,12,13]. Various solutions have been studied in recent years with the aim of developing less invasive and reversible techniques, based for example on the use of steel profiles or mortar topping [14]. The paper focuses particularly on strengthening interventions based on the use of timber or timber based elements, dry connected to the existing floors, with the aim of increasing the bending or in-plane stiffness.

2.

Timber-to-Timber Bending Stiffness Upgrade Interventions

Bending stiffness upgrade can be achieved by connecting timber planks or Cross Laminated Timber panels to existing floor joists. In this way, it is possible to rely on a composite T section beam. In timber-to-timber composite sections, the use of traditional materials and dry assembly methods are in agreement with the restoration issues of compatibility, reversibility or recoverability of the intervention. For all composite sections the mechanical characteristics of the connection are the main factor which influences the structural response. The design of timber-to-timber composite sections requires consideration of partial composite behaviour, due to the deformable shear connection between web and flange. Analysis can follow the Eurocode 5 [15] approximate ‘gamma method’, where an effective flexural stiffness (EI)ef for the composite section is calculated as a function of the stiffness of the shear connection, taking into account the slip between the flange and the joist. Also the ‘shear analogy method’, where the composite beam is divided into two virtual components coupled with stiff bars can be used to determine internal forces [16,17,18]. Simplified equations to evaluate the increment, with respect to complete connection, of the camber, of the central vertical displacement and of the bending stress due to maximum slip at interface can be found in [14]. For example the increment of the central sag, for ordinary values of the ratio L/H (18÷25), where L is the length of the beam and H is the height of the cross section, can be estimated with:

w  10

(1)

where w is the increment of the vertical deflection and  is the maximum slip at the interface. Several connection systems have been proposed for concrete-to-timber systems [19] and some of them have been tested for timber-to-timber solutions. New connection systems have also been proposed and different elements for upper flanges have been tested, with good performance in any case.

26

Reversible timber-to-timber strengthening interventions on wooden floors 2.1

Timber flanges connected to main beams by hardwood dowels

The reinforcing technique consists of placing a new plank above each floor joist, connected by means of dowels (Fig. 1) made of hardwood pins [20]. With proper design of the connection of the beams to the load bearing walls and additional planks in the upper surface, the system can work also as an effective diaphragm (Fig. 1b) [21]. A composite timber-to-timber T beam is thus obtained with deformable connection between the flange and the web. These may or may not be separated by the existing floor board (Fig. 1d). The presence of the floor boards influences the dowel collapse mechanisms and consequently the beam failure: the best performance was obtained when boarding was present.

Fig. 1 Stiffening intervention made with new planks connected to existing beams by means of dry hardwood pins (from [20])

2.2

Timber flanges connected to main beams by self-tapping screws

Among the different types of possible connectors available, the use of inclined self-tapping screws seems to give promising results, based on experimental testing of timber-to-timber connections [22,23,24]. In [25,26] a strengthening intervention on an ancient floor of a castle is described. The existing beams were coupled with glulam planks strength class GL 24h, 80 mm in thickness. A T-beam compound section was so obtained, with deformable connections between the flange and web. The web and flange were separated by a new boarding, with a thickness of 30 mm, which replaced the existing decayed one (Fig. 2). The connection system consisted of self-tapping double thread screws, steel grade 10.9, X-crossed at an angle α = 45˚. A final load test on a portion of the reconstructed floor showed good performance and good agreement with the design data.

27


Fig. 2 Diagrams of a) the transversal section, b) longitudinal section of the composite structure, and c) detail of the geometry of the connection system. (i) Beam interspace; (ls) screw length; (s) spacing of the fasteners (from[26]) The slip modulus of the connection was determined according to the calculation method proposed in [22,23,24], on the basis of the results from a series of push-out tests for different screw inclinations and joint configurations.

2.3

Timber-to-CLT Composite Section

Cross laminated timber panels (XLam or CLT), which has been used to build walls and floor slabs in new timber buildings for the last 15 years, can be efficiently adopted also in restoration interventions to obtain a diaphragm effect under seismic actions, by connecting them to the existing timber beams and to the perimeter walls. Cross laminated timber panels consist of a sequence of layers made of boards, glued one to the other. Experimental tests showed they are stiff enough in their plane to assure a diaphragm effect and to resist the common lateral forces due to earthquakes. In this way the superposition of XLam panels connected to the existing timber joists can substitute for concrete topping as a less invasive, lighter and more reversible solution [27]. Structural XLam panels are generally made of a series of layers and can be over 120 mm thick. Special panels of 60 mm thickness have been specially produced for use in strengthening interventions, as in restoration it is important not to vary the existing floor levels too much. A series of experimental tests has been conducted to investigate the behaviour of the composite section beams. Ten specimens made by XLam panels and glulam beams, six connected together by means of predrilled steel dowels (T1A÷T6A) [27] and four by self-tapping screws (T1B÷T4B) [28], were tested under bending loads up to collapse. The screws were inserted with a 90 degree angle with respect to the beam axis (Fig. 3). All the beams were loaded over 2 points at a distance corresponding to 1/3 of the beam length. The increasing of the load was checked by displacement control. The data acquisition was automatically carried out. The instrumentation consisted of 10 potential transducers of 1/1000 mm precision, in order to monitor the vertical deflection in the middle of the beam, the rotation and the slip at interface near the supports (Fig. 4).

28

Reversible timber-to-timber strengthening interventions on wooden floors

Fig.3 Predrilled steel dowel or screw spacing along the beam (from [27,28])

Fig.4 Load configuration and monitoring device set up (mm)(from [27,28]) As for all types of composite section, the behaviour depends on the connection stiffness and it ranges between the two limit states of composite section without connection (EJ)0 and composite section with infinite stiffness connection (EJ)∞. The diagrams (Fig. 5) show a notable increase in stiffness. The use of self-tapping screws as connectors leads to a greater initial stiffness compared with predrilled steel dowels, as they do not need an initial slip to come in contact with the surrounding timber [27]. Moreover, there is also a friction effect, due to screws, which gives an initial stiffness equal to the case of infinite stiffness connection up to 10% of the collapse load. XLam panels can simultaneously provide diaphragm action as they have significant shear stiffness in their plane.

Fig. 5 Load-Deflection diagrams of XLam-to-Timber composite section beams

29

Reinforcement of Timber Structures 2.4

General Remarks

All these interventions have to face the problem of the thickness of the cross section after the interventions, as in restoration it is important not to vary too much the existing floor levels, otherwise problems can arise with the window and the door thresholds, with the level of the stairs or with eventual wall decorations. From this point of view the use of CLT panels of only 6 cm height, or eventually less, can be an interesting solution. The efficiency coefficient of the composite section η [7] is a significant parameter to evaluate the capacity of the connection in order to limit slip between the two parts of the composite section. The expression of the coupling coefficient η is defined by the following equation: (2)

where (EJ) actual bending stiffness, (EJ)0 bending stiffness of the composite section without connection, (EJ)∞ bending stiffness in case of complete connection. The results presented in literature show that the typical values of concrete-to-timber cross sections ( [29] or more recently up to 0,9) can be reached also by means of timber-to-timber techniques: the experimental maximum value obtained by composite CLT to timber beam section was around =50%, even if numerical analyses show that with different connector spacing it can increase [28], while the use of inclined self-tapping screws can upgrade the coupling coefficient values (=0,74 [26]). If the existing beams have significant cambers due to permanent loads and creep, some problems can arise during the execution of the intervention, as the bottom of the new element and the upper part of the existing joist do not match. In these situations compensation board can be needed or forced shoring can be eventually tried.

3.

In-plane stiffness upgrade by using timber or timber based elements.

An effective diaphragm action in the floors of multi-storey masonry buildings is particularly important because the forces due to earthquakes can be transferred to the lateral load resisting walls. One of the first documents to propose the in-plane stiffness upgrade of timber floors by means of timber elements was produced by the local authority of Friuli Venezia Giulia Region (Italy) after 1976 earthquake [30], in fact, at the time, Italian national codes did not provide guidance on strengthening interventions design on existing masonry buildings damaged by earthquakes. The suggested technique consisted of superposing a second layer of floorboards over the existing one, but laid in the orthogonal direction (Fig. 6). This multiple layers of floorboards was a technique commonly diffused in ancient buildings in the earthquake prone areas of the region. The importance of the connection of the floor with the perimeter walls was underlined by the drawings of possible details of the joining of the floorboards with walls (Fig. 7) and of the joists with walls (Fig. 8).

30

Reversible timber-to-timber strengthening interventions on wooden floors New layer of floorboards

Existing layer of floorboards

Fig. 6 New layer of floorboards over the existing one (from [30])

New layer of floorboards

Steel plate

Nails

Exisisting layer of floorboards

almost

Masonry wall to stiffening layer connection

Fig. 7 Detail of the connection of the floorboards with the wall (from [30]) Nails spacing Masonry wall Stiffening layer of floorboards Existing layer of floorboards Masonry wall

Fig. 8 Detail of the suggested connection of the joist with masonry wall (from [30]) A series of full scale tests to measure experimental in-plane stiffness is described in [31] and [32]. In these tests, two samples of unstrengthened timber floor elements, having a common typology of mono-directional floors (FMSB and FM samples), were subjected to five types of strengthening interventions. The interventions involved the use of diagonal punched metal strips (FMSD), of a single layer of diagonal planks placed at 45° with respect to the original floorboards with common or tongue and groove boards (FM SP(A) and FM SP(B)), of a double diagonal layer placed at ± 45° made with tongue and groove boards (FM± 45°DP(A)), of one diagonal plank equivalent to the punched steel strip in term of axial stiffness (FMWD(D)) and finally of a double diagonal made with thick planks (FMWD(E)). As it can be seen in Fig. 9, the monotonic tests on reinforced floors showed that the best result was achieved by the use of a double board layers. The stiffness increase is significant with respect to the traditional existing floors.

31


Fig. 9 Force versus top displacement behaviour for the unreinforced and reinforced specimens (from [32])

Fig. 10 Specimens used in the full-scale tests (from[33]) In [33], five full scale timber floors were tested to analyse the in-plane behaviour: one was made with timber beams and floorboards only (Fig. 10 S-specimen), the second was strengthened by a second layer placed orthogonal to the first board (Fig. 10 S-specimen) and the other three by two (Fig. 10 CLT2-specimen) or three CLT panels (Fig. 10 CLT3.1 and CLT3.2). In Fig. 11 the test setup is described. Test results show that a significant increase in the in-plane stiffness, up to 5-10 times than that of the unstrengthened specimen, can be achieved, so the efficacy of the intervention was experimentally demonstrated. Numerical modelling showed that the main influence on the in-plane

32

Reversible timber-to-timber strengthening interventions on wooden floors behaviour was due to the screw connection in the direction perpendicular to the axis of the floor beams.

Fig. 11 Setup of the floor test (from [33]) In [34] a full-scale experimental program consisting of tests on four as-built diaphragms and four retrofitted diaphragms in both principal loading directions is presented. As-built configurations were typical of those found in historic unreinforced masonry buildings in North America and Australasia, whereas retrofitted diaphragms consisted of plywood panel overlays with stapled sheet metal blocking systems (SMBS). The nonlinear and low stiffness behaviour of the as-built diaphragms was confirmed in each principal loading direction. The plywood overlay and SMBS dramatically improved the asbuilt diaphragm shear strength and shear stiffness and were shown to perform satisfactorily from a serviceability perspective.

Fig. 12 Setup of the floor test for parallel to joist loading (a) and perp. to joist loading (from [34]) The tests were performed for loading parallel to joists and perpendicular to joists (Fig.12). Testing in

33

Reinforcement of Timber Structures both principal loading directions confirmed the orthotropic nature of timber diaphragms. While the shear strength remained consistent for as-built diaphragms, the shear stiffness in the direction perpendicular-to-joists was up to 32% less than the corresponding value in the orthogonal configuration. For retrofitted diaphragms, the difference in shear stiffness increased to 60%, and the shear strength in the direction perpendicular-to-joists was almost 50% of the shear strength parallelto-joists. In [35] ten floor specimens were tested, including five floors representing an as-built configuration and five floors retrofitted through the addition of a plywood layer on the top of the floor boards. Square-edge straight pine flooring boards (25 × 150 mm) were nailed to the joists with two standard nails (3.15 mm diameter) at each joist intersection. The flooring board layout is shown in Fig. 13. Timber floor boards were staggered meaning that the floor was composed of a combination of 2 m and 1 m long boards. For five specimens, plywood panel overlays were screwed to the original structure.

Fig. 13 Flooring board layout (from [35]) In Fig. 14 the load-displacement curve related to the specimen R-1 underlines that the effect of the plywood panel overlays results in a significant increase in both diaphragm strength and stiffness; the maximum load reached during the test is equal to 150 kN, equal to three times that for the as-built specimen. The curve shows a strong nonlinearity after the first load cycle, proving that the diaphragm behaviour is strongly affected by the connection between the timber elements (nails and screws). The failure occurred principally in the screws between panels and boards (short screws) and between panels and joists (long screws).

Fig. 14 Load versus displacement curves for as-built (AB-1) and retrofitted specimens R-1 (from[35])

34

Reversible timber-to-timber strengthening interventions on wooden floors 3.1

In-plane stiffening of wooden floors by means of nailed plates connecting the planks.

The in-plane stiffness of wooden floors can be increased by using punched metal plates (gang nails) to connect the boards longitudinally, so as to prevent the slip due to shear forces (Fig. 15). Different fasteners may be used as nail plates embedded into the wooden board by using special devices or knuckle nail plates, applied by hammering in the “knuckle nail” [36]. The technique requires the use of an appropriate device capable of ensuring that the nail plate achieves complete penetration of the nails in the wooden boards, such as an hydraulic jack arranged on a special steel frame. The hammer application technique is simpler, but it may be less effective. A perimeter L-shaped steel profile is used to increase the floor resistance: the steel profiles perpendicular to the horizontal action constitute the two chords needed to resist the bending moment, the stiffened wooden floor supports shear forces and the steel profiles parallel to horizontal action transfer the in-plane forces to the shear walls.

Fig. 15 Schematic view of the floor (a) and gang nails connecting the boards (b) (from [36])

4.

Conclusions

The recent literature on the problem of bending and in-plane floor stiffness upgrading shows that there is an increasing sensibility towards the problem of reversible and less invasive techniques of restoration. Several papers on these problems have been presented with an evident increase in number in the last two to three years. The design rules for upgrading the bending stiffness by means of timber-to-timber dry techniques are based on the composite section theories developed for concrete-to-steel and concrete-to-timber composite sections. As new types of connectors and new types of upper flanges have been introduced, experimental tests have been performed to check the applicability of the codified formulations or to validate the new proposed ones. As in all types of composite sections, one of the main parameters governing the structural answer of the beam is the connection load-slip relationship. The importance of the correct evaluation of the in-plane mechanical properties of floor timber diaphragms to determine the building structural response under lateral loads due to earthquakes is addressed. In some codes, simplified analytical procedures are proposed to determine the in-plane stiffness, but generally with regard to new timber buildings floor typologies. In [37] design details are given to obtain a rigid floor, while in [38] formulas are presented to evaluate the different components

35

Reinforcement of Timber Structures of the diaphragm horizontal deflection, which are due to nail slip, to flexural deformation and to shear deformation in the sheathing. More experimental tests are necessary to check the reliability of the analytical models proposed, to develop new ones or to complete an exhaustive experimental database. Moreover, the tests described in literature papers are generally referred to different setups, test rigs, boundary conditions, aspect ratio of the floor samples and recorded parameters and so it is difficult to compare the experimental results as to achieve a general stiffness evaluation approach.

5.

Acknowledgements

The present paper is a result of the activities of COST Action FP1101 Assessment, Reinforcement and Monitoring of Timber Structures.

6.

References

[1]

Tomaževič M., Earthquake-resistant Design of Masonry Buildings, Imperial College Press, 1999. Turrini G., Piazza M., “Una tecnica di recupero dei solai in legno”, Recuperare, No.5, 1983, pp. 396-407. Ronca P., Gelfi P., Giuriani E., “Behavior of a wood–concrete composite beam under cyclic and long term loads”, In: Proceedings of the International Conference on Structural Repair and Maintenance of Historic Buildings, Seville, Spain, 1991. Natterer, J., Hamm, J., Favre P., “Composite wood–concrete floors for multi-story buildings”, In: Proceedings of the 4th International Wood Engineering Conference, New Orleans, Omnipress, Madison, Wisconsin, USA, 1996. Gelfi, P., Giuriani, E., Marini A., “Stud shear connection design for composite concrete slab and wood beams”, Journal of Structural Engineering, 2002, 128(12), pp.1544-1550. Giuriani, E., “L’organizzazione degli impalcati per gli edifici storici”, L’edilizia 134:30-43, 2004. Gutkowski, R., Brown, K., Shigidi,A., Natterer J., “Laboratory tests of composite wood– concrete beams”, Construction and Building Materials, Vol. 22. No.6, 2008, pp.1059-1066. Bathon, L., Graf, M., “A continuous wood-concrete-composite system”, In: Proceedings of World Conference of Timber Engineering, Whistler, BC, 2000. Clouston, P., Bathon, L., Schreyer, A., “Shear and Bending Performance of a Novel Wood– Concrete Composite System”, Journal of Structural Engineering, Vol. 131, No. 9, 2005, pp. 1404–1412. Venice Charter, Second International Congress of Architects and Technicians of Historical Monuments, Venice, May 25-31, 1964. Krakow Charter, International Conference on Conservation Krakow 2000, Krakow, 2000. ISCARSAH-ICOMOS, Principles for the analysis, conservation and structural restoration of architectural heritage, 2003. UNI Ente Nazionale Italiano di Unificazione, UNI 11138 - Cultural Heritage - Wooden artefacts - Criteria for the preliminary evaluation, the design and the execution of works, Milano (Italy), 2004. Giuriani E., Consolidamento degli edifici storici, UTET, Torino (Italy), 2012. EN 1995:2004, Eurocode 5 Design of Timber Structures, 2004. Kreuzinger, H., “Platten, Scheiben und Schalen: Ein Berechnungsmodell für gängige Statikprogramme“, Bauen mit holz, Vol. 1, 1999, pp. 34-39. Deutsches Institut für Normung: DIN 1052; Entwurf, Berechnung und Bemessung Deutsches Institut für Normung: DIN 1052; Entwurf, Berechnung und Bemessung, 2004.

[2] [3]

[4]

[5] [6] [7] [8] [9]

[10] [11] [12] [13]

[14] [15] [16] [17]

36

Reversible timber-to-timber strengthening interventions on wooden floors [18] Kuhlmann, U., Michelfelder, B., “Optimised design of grooves in timber-concrete composite slabs”, In: Proceedings of the 10th World Conference on Timber Engineering, Portland, Oregan, USA, 2006. [19] Yeoh, D., Fragiacomo, M., De Franceschi, M., Heng Boon K., “State of the art on timberconcrete composite structures: Literature Review”, Journal of Structural Engineering, Vo. 137, No. 10, 2011, pp.1085-1095. [20] Valluzzi, M.R., Garbin, E., Modena, C., “Flexural Strengthening of timber beams by traditional and innovative techniques”, Journal of Building Appraisal, Vol 3, No. 2, 2007, pp.125-143. [21] Modena, C., Valluzzi, M.R., Garbin, E., da Porto, F., “A strengthening technique for timber floors using traditional materials”, In: Proceedings of the Fourth International Conference on Structural Analysis of Historical Constructions SAHC 04, Padova, Italy, 2004. [22] Bejtka, I., Blaß, H.J., “Screws with continuous threads in timber connections”, In: International RILEM Symposium on Joints in Timber Structures, Stuttgart, Germany, ed. S. Aicher and H.W. Reinhardt, pp.193-201, 2001. [23] Bejtka, I., Blaß, H.J., “Joints with inclined screws”, In: Proceedings of Meeting 35 of the International Council for Building Research Studies and Documentation, CIB, Working Commission W18 – Timber Structures, Kyoto, Japan, CIB Paper No.35-7-4, 2002. [24] Tomasi, R., Crosatti, A., Piazza, M., “Theoretical and experimental analysis of timber-to-timber joints connected with inclined screws”, Construction and Building Materials, Vol. 24, 2919, pp. 1560-1571. [25] Angeli, A., Piazza, M., Riggio, M., Tomasi, R., “Refurbishment of traditional timber floors by means of wood-wood composite structures assembled with inclined screw connectors”, In: Proceedings of 11th World Conference on Timber Engineering WCTE 2010, ed. A.Ceccotti and J.W. Van de Kuilen, Riva del Garda, Italy, 2010. [26] Riggio, M., Tomasi, R., Piazza, M., “Refurbishment of a traditional timber floor with a reversible technique: importance of the investigation campaign for design and control of the intervention”, International Journal of Architectural Heritage, Vol. 8, 2013, pp.74-93. [27] Gubana, A., “Experimental tests on Timber-to-Cross Lam composite section beams”, In: Proceedings of 11th World Conference on Timber Engineering WCTE 2010, ed. A.Ceccotti and J.W. Van de Kuilen, Riva del Garda, Italy, 2010. [28] De Cillia, L., “Experimental and numerical analysis of Timber to XLam composite section beam” (in Italian), Master of Civil Engineering Thesis, University of Udine, Gubana A. Supervisor, 2013. [29] Piazza, M., Tomasi, R., Modena, R., Strutture in legno, Hoepli, Milano (Italy), 2005. [30] Regione Autonoma Friuli-Venezia Giulia - Segreteria Generale Straordinaria: Legge Regionale 20 giugno 1977, n. 30 - Recupero statico e funzionale degli edifici. Documento tecnico n.2 DT2: Raccomandazioni per la riparazione strutturale degli edifici in muratura. Gruppo Disciplinare Centrale, Maggio, 1980. [31] Valluzzi, M.R., Garbin, E., Dalla Benetta, M., Modena, M., “Experimental Assessment and modeling of in-plane behaviour of Timber Floors”, In: Proceedings of the VI International Conference on Structural Analysis of Historic Construction, SAHC 08, Bath, UK, ed. D. D’Ayala and E. Fodde, 2008. [32] Valluzzi, M.R., Garbin, E., Dalla Benetta, M., Modena, M., “In-plane Strengthening of Timber floors for the seismic improvement of masonry buildings”, In: Proceedings of 11th World Conference on Timber Engineering WCTE 2010, ed. A. Ceccotti and J.W. Van de Kuilen, Riva del Garda, Italy, 2010. [33] Branco, J.M., Kekeliak, M., Lourenço, P.B., “In Plane Stiffness of traditional Timber Floors Strengthened with CLT”, Materials and Joints in Timber Structures, ed. S. Aicher et al., RILEM Book series 9, pp.725-737, 2014. [34] Wilson, A., Quenneville, P.J.H., Ingham, J.M., “In-Plane Orthotropic Behavior of Timber Floor Diaphragms in Unreinforced Masonry Buildings”, Journal of Structural Engineering, Vol. 140, No. 1, 2013, 04013038:1-11. [35] Brignola, A., Pampanin, S., Podestà, S., “Experimental Evaluation of the In-Plane Stiffness of

37

Reinforcement of Timber Structures Timber Diaphragms”, Earthquake Spectra, Vol. 28, No. 4, 2012, pp. 1687-1709. [36] Gattesco, N., Macorini, L., “High reversibility technique for in plane stiffening of wooden floors”, In: Proceedings of the VI International Conference on Structural Analysis of Historic Construction, SAHC 08, 2-4 July 2008, Bath, UK, ed. D. D’Ayala and E. Fodde, 2008. [37] EN 1998-1:2005, Eurocode 8 Design of structures for earthquake resistance, 2004. [38] NZS 3603, New Zealand Timber Structures Standards, 1993.

38

Reinforcement of timber columns and shear walls

3 Reinforcement of timber columns and shear walls Wen-Shao Chang1

Summary This chapter provides an overview of state-of-the-art repairing and reinforcing techniques on timber columns and shear walls in both research and practice. It covers two levels of intervention, repair and reinforcement of timber elements. The former focuses on damaged elements and the latter focuses on enhancing the mechanical properties of the elements. Although it was found that most of the research foci were on reinforcement of timber connections and flexural members, columns and shear walls play a crucial role in the prevention of structural collapse. With the future development of taller timber structures, these issues will become more and more important.

1. The need to reinforce/repair timber columns and walls A column is a member in a structure that takes vertical load and sometimes bending moment transferred from a beam via connections. It is crucial to the stability of a structure. A timber shear wall is an important structural element that provides lateral stability to the structure and resists the horizontal forces, such as earthquake and wind. They provide substantial in-plane stiffness and only limited out-of-plane stiffness. The reasons to reinforce timber shear walls are: (1) to enhance stiffness and strength; (2) to improve ductility; and (3) to increase energy dissipation capacity. Note that in this chapter; only shear walls made of timber will be discussed. For example, in some half-timber framed structures, stones, bricks (Fig. 1) and wattle and daub (Fig. 2) are often used as in-fill elements, and therefore are outside of the scope of this chapter. There are a number of situations where a column and a shear wall in a building need to be repaired or reinforced. These include biodeterioration, mechanical failure, cracks, and the need for higher strength.

1.1 Biodeterioration Columns, when touching the ground without any measure to isolate them from damp, are prone to elevated moisture content levels which will lead to bio-deterioration due to insects (such as termites) and fungal attacks. This is a common form of decay that can be found where the column touches the ground (Fig. 3). When designing a timber column, one should select the timber carefully as the most common form of deterioration is from attack of the sapwood by insects, while the heartwood remains untouched [1]. The rise of moisture content in a column will lead to fungal defects and attract termites to attack the member, and these are often unseen as shown in Fig. 4. This failure mode in a column not only reduces the mechanical properties of timber but also reduces the effective section area. 1)

PhD, Department of Architecture and Civil Engineering, University of Bath, UK

39


Fig. 1 Half-timber frame with brick infill

Fig. 2 Half-timber frame with wattle and daub infill

1.1 Biodeterioration Columns, when touching the ground without any measure to isolate them from damp, are prone to elevated moisture content levels which will lead to bio-deterioration due to insects (such as termites) and fungal attacks. This is a common form of decay that can be found where the column touches the ground (Fig. 3). When designing a timber column, one should select the timber carefully as the most common form of deterioration is from attack of the sapwood by insects, while the heartwood remains untouched [1]. The rise of moisture content in a column will lead to fungal defects and attract termites to attack the member, and these are often unseen as shown in Fig. 4. This failure mode in a column not only reduces the mechanical properties of timber but also reduces the effective section area.

Fig. 3 Bio-deterioration in a column that has contact with the ground

Fig. 4 Timber strut attacked by termites

1.2 Mechanical failure Compared with beam members, creep is less onerous in a column member. A column normally takes only vertical load; in some circumstances it will take combined compression and bending. The former will lead to buckling of the column, whilst the latter will result in partial yielding or split along the grain as shown in Fig. 5. Slender compression members are susceptible to buckling. When a compression member has (1) insufficient section size; (2) vertical cracks so the effective section is reduced; or (3) low material strength, it is prone to buckle. The buckling of a compression member is

40

Reinforcement of timber columns and shear walls a failure that often occurs without warning. It is therefore important to consider whether compression members within a structure are highly stressed, and if any action needs to be taken to ensure the prevention of the column from buckling.

1.3 Cracks Cracks occurring in timber members often result from the differences between the drying speed in interior layers and outer ones. The drying stresses will be built up if the outer layers are dried to a level that is much lower than the fibre saturation point while the interior is still saturated [2]. Rupture in timber occurs and in consequence cracks occur if the drying stress exceeds the strength perpendicular to the grain as shown in Fig. 6.

1.4 Need for higher strength Recently there has been a trend all over the world to strive for higher timber constructions. Two midrise timber apartments had been completed in London, UK prior to 2011. Another 10-storey timber apartment was completed in 2012 in Melbourne, Australia. Further tall timber buildings are at the planning stage and therefore we will see more and more tall timber buildings in the future. To achieve taller timber buildings, we need timber products with higher strength, in particular those which will be used in the lower parts of the buildings. As such, we also need timber columns and shear walls to have higher strength in order to resist the self-weight built up when the buildings go higher.

2. Repair and reinforcement of timber columns 2.1 Prosthesisation When dealing with historic buildings, engineers and architects need to balance authenticity of the structures after renovation/repair and assurance of the strength of the structure to carry the load needed. To minimise the amount of timber being replaced, prosthesisation has become common practice when the timber members are bio-deteriorated due to termites or insects. It is a method that replaces only the decayed or failed part with a new portion. Timber used for prosthesisation, in particular for the conservation of historic buildings, must be carefully selected so that the nature of the new timber will match that of the old. The moisture content of the timber being used should be close to that of the existing members so that moisture movement can be avoided. Fig. 7 shows an example of a column being prosthesised after it was partially damaged. When a new prosthesis is adopted to replace the damaged portion in a timber member, two methods exist to connect the old and new members: (1) local and traditional carpentry as shown in Fig. 8; and (2) glued-in members for the connection. For both cases, modern adhesives are often used to ensure the continuity of the new column. Although prosthesisation is common practice nowadays in historic building conservation in many countries, there is a lack of research work on this method.

2.2 Screw reinforcement Song et al. [4] carried out a series of tests to study the effect of self-tapping screws to repair timber columns with vertical cracks and compared that with timber columns repaired by Fibre Reinforced Polymer (FRP) pads. The vertical cracks were simulated by making slots in the column with a width of 6 mm and a length of 1500 mm. Vertical load was applied on the columns with pin connections at both ends. The conditions for the specimens are shown in Tab. 1 and the specimen design is shown in Fig. 9. The failure modes of each specimen are shown in Fig. 10.

41


Fig. 5 A column damaged due to an earthquake

Fig. 6 Vertical cracks occur in a column

Fig. 7 A new timber component was used to Fig. 8 Partial replacement repair in Daibei Temple partially replace rotten column with traditional (1550), China [3] carpentry

It was observed from the tests that the maximum loading capacity of Specimen 2 (cracked and unrepaired) was more than 30% lower than that of the intact column (Specimen 1), and this shows that the vertical crack will weaken the column. The experimental results also showed that the self-tapping screws will improve the strength of the cracked specimen to a level close to the intact ones. The additional work of filling the crack in a column does not affect the strength of the cracked column. The strength of the cracked column repaired by FRP pad showed similar results Fig. 9 Specimen design of repairing cracked timber column by using screws and FRP pads [5] those repaired by self-tapping screws.

42

Reinforcement of timber columns and shear walls Tab. 1 Specifications of column and experimental results by Song et al (data source: [5]). Retrofit

Diameter/width1 Spacing (mm) (mm)

Ultimate strength (kN)

̶

̶

̶

̶

846

Y

N

̶

̶

̶

571

200 x 200 x 1800

Y

N

STS

6

250

736

4

200 x 200 x 1800

Y

N

STS

6

250

895

5

200 x 200 x 1800

Y

N

STS

6

250

675

6

200 x 200 x 1800

Y

Y

STS

6

250

812

7

200 x 200 x 1800

Y

Y

FRP

100

200

835

No. Dimensions (mm)

Slotted Filled

1

200 x 200 x 1800

N

2

200 x 200 x 1800

3

Note: 1 diameter for screws and width for FRP sheets

Fig. 10 Failure mode of columns reinforced by different strategies tested by Song et al. [5]

This study shows self-tapping screws to be a good repairing measure; in particular because it is reversible, i.e. the self-tapping screws can be removed in the future once more efficient ways of repairing timber columns have been developed. More work needs to be done on investigating factors, such as the dimensions of the cracks and the spacing of the screws, before this method can be widely implemented.

2.3 Steel member reinforcement In the early stages of reinforcement and repair of timber structures, the focus was mainly on using metallic reinforcement, such as steel bars and plates. However, the focus was also mainly on beam

43

Reinforcement of Timber Structures elements and connections; efforts being devoted to the reinforcement of columns were relatively scarce. Tanaka et al. compared the effect of a column reinforced by steel plate with that of one reinforced by carbon fibre sheets [6]. Buckling tests were carried out, and the parameters considered included (1) slenderness ratio of column, (2) boundary conditions for steel plates in the reinforced column, and (3) reinforcement methods (steel plates and carbon fibre sheets). The sections of the specimens of the experiments are shown in Fig. 11 and the reinforcement arrangements are depicted in Fig. 12.

Fig. 11 Section of the columns reinforced by steel Fig. 12 Different reinforcement arrangements [6] plates and carbon fibre sheets [6] The experimental outcomes showed that steel plates reinforced timber columns have load-carrying capacities at least 2.5 times higher than that of unreinforced timber columns, whilst columns reinforced by carbon fibre sheets exhibit 1.3 times higher load-carrying capacities than unreinforced.

2.4 Composite material reinforcement Repair and reinforcement of the damaged timber members by composite material, such as FRP, has been developed over more than 2 decades. FRP has a remarkable strength-to-weight ratio and leads to light weight strategies when repairing or reinforcing these damaged members. Substantial amounts of effort have been devoted to investigating increasing the strength properties of intact timber members after the application of FRP bonded externally [7-9]. Zhang et al. carried out a series of tests on repairing cracked columns by using FRP wrapping and developed finite element models to simulate the behaviour for parametric studies [10]. The factors considered include (1) the column dimensions, (2) the crack dimensions, (3) whether the crack was filled, (4) FRP properties and (5) FRP spacing. A total of 17 specimens were tested and six different failure modes were observed. Fig. 13 shows the specimens tested and factors considered. The experimental results showed that different combinations of factors, in particular the FRP spacing, will result in different failure modes. It was evidenced in the series of tests that the load-carrying capacity of a column decreases with increase in the length and width of the cracks and the influence of the crack width is more significant. It was also observed that reducing the FRP spacing will increase the recovery of load-carrying capacity of cracked timber columns. Oprişan et al. shows different methods of using FRP composite to strengthen a timber column. They include: (1) FRP fabric with different fibre orientations; (2) FRP strips to provide confinement; (3) FRP strips to share the load; and (4) using embedded FRP rods and fabric to provide confinement [11].

44


Fig. 13 The specimens and different factors considered in the series of tests carried out by Zhang et al. [10]

Fig. 14 Column specimens and carbon reinforcement used by Najm et al. [14]

A series of tests was carried out by Taheri et al. to investigate the buckling response of glulam columns reinforced with FRP sheets with different lengths and end fixity [12]. The reinforcement levels included non-reinforcement (control), fully reinforced, and partially reinforced (the FRP sheet was one-third of the length of the column and attached in the middle of the column). The boundary conditions of the column were pinned-pinned and fixed-fixed ends.

Fig. 15 Ultimate strength and elastic modulus of columns versus fibre content [14]

Fig. 16 Column to foundation specimen Fig. 17 Experimental setup for post-tensioned strengthening with post-tensioned reinforcement and of LVL column [15] external energy dissipater [16]

45

Reinforcement of Timber Structures It was found that columns which were fully reinforced had a higher strength compared with the other conditions. The experimenters concluded that using FRP for partially reinforcing a glulam column is more effective for the pinned-pinned case as the strength of the column reached almost half of the increase in strength of those fully reinforced, but only used one third of reinforcing material. Most FRP composites use organic matrices in manufacturing FRP plates, but since the 90s there has been significant progress in manufacturing FRP with inorganic matrices that are non-toxic, have good fire resistance, and are not affected by UV radiation [13]. A series of tests to investigate the confinement of circular timber columns using inorganic CFRP was carried out by Najm et al. [14]. They tested 40 column specimens in axial compression, two different wrapping methods for the CFRP, spirals and full wrapping. The specimens and the carbon reinforcement used in the tests are shown in Fig. 18. The reinforced column specimens exhibited higher strength and stiffness than the unreinforced specimens. It was also observed that specimens that were fully wrapped had higher strength and stiffness compared with those that were partially reinforced (spiral reinforcement). With respect to strength increase and fibre content, it was observed that the average load-carrying capacity of the column increased with the decrease of the spacing of CFRP, i.e. increase of the amount of CFRP. The same phenomenon can be found for the axial stiffness of the column specimens. In other words, the more CFRP used, the better the mechanical properties the columns will have as can be seen in Fig. 19.

2.5 Post-tensioned strengthening Post-tensioned strengthening is a relatively new development in the seismic field. An extensive experimental campaign was carried out on beam-to-column, column-to-foundation and wall-tofoundation subassemblies for the implementation of LVL hybrid solutions [15, 16]. The design was to use external energy dissipaters together with post-tensioned effect to provide re-centring and energy dissipation capacity of a timber column. Fig. 20 and Fig. 21 show the specimens and experimental setup, respectively. The hysteretic loop (Fig. 22) shows a flag-shape, and it was observed that 4.5% of the storey drift was achieved in the tests; there was no degradation of stiffness and no structural damage after the tests. The residual deformation was still negligible as the post-tensioned mechanism helps the column to recentre when unloading.

2.6 Enlargement of column cross section Enlargement of the cross section of a column will help to reduce the stress within the column so as to reduce the potential for buckling and material yield in compression. In some structures, such as those found in Japanese temples, large section columns will contribute to resisting lateral load by providing restoring forces [17]. Suda, Tasiro & Suzuki [18] proposed to enlarge the column base of existing structures (Fig. 23) to increase the restoring force so as to enhance the aseismic behaviour of traditional temples. Shaking table tests were carried out to investigate the effectiveness of the proposed reinforcement method. The reinforced column shows higher restoring force and larger deformation. This gives the whole structure better lateral force resistance.

46


Fig. 18 Column specimens and carbon reinforcement used by Najm et al. [14]

Fig. 19 Ultimate strength and elastic modulus of columns versus fibre content [14]

Fig. 20 Column to foundation specimen Fig. 21 Experimental setup for post-tensioned strengthening with post-tensioned reinforcement and of LVL column [15] external energy dissipater [16]

Fig. 22 Hysteretic loop of a post-tensioned strengthening LVL column

47


Fig. 23 Enlargement of column base proposed by Suda, Tasiro & Suzuki [18]

3. Reinforcement of timber shear walls This section solely discusses the strategy to reinforce timber shear walls; it is worth noting that repair and strengthening of timber shear walls often are achieved through an intervention on the joints and beams. There are several solutions to strengthening of timber shear walls including [19]:     

to use additional sheathings to reinforce shear walls with steel diagonal elements to reinforce existing sheathing of the shear wall with carbon or high-strength synthetic fibre to reinforce the beams using hardwood inserts to post-tension the walls using prestressing wire.

Fig. 24 The specimen and test setup for CFRP reinforced timber shear walls [21] The first solution is the simplest method and is popular. The effectiveness of this method relies heavily on the stiffness of the fasteners connecting the boards to the frame. The second method is to attach steel diagonal elements to timber frames so as to share the force with the timber shear walls. The first two solutions are relatively straightforward and can be designed by calculation, therefore only limited research efforts have been devoted to these two methods. The remaining solutions ensure that the reinforced timber shear walls will have higher ductility and strength. These methods have attracted more attention in research and are discussed below.

48

Reinforcement of timber columns and shear walls 3.1 Composite material reinforcement A series of research programmes have been carried out on reinforcing timber shear walls using FRP strips [20, 21]. A total of nine specimens were tested in three groups and the CFRP strips were glued on the fibre-plaster board (FPB) attached to the timber frames [20]. Fig. 24 shows the specimen for the tests. The first group (G1) used two CFRP diagonal strips with width of 300 mm glued on to the FPB and also onto the timber frame; whereas the second group (G2) used 600 mm wide CFRP diagonal strips with the other conditions being same as the first group. The third group (G3) has two 300 mm width CFRP strips glued on the FPB but not attached to the timber frame. The experimental results revealed that the third group had the highest elastic resistance (force forming the first crack) although it was found to increase in all the CFRP strengthened test samples. The results from this series of tests showed that the three reinforcement methods do not increase the stiffness but increase the strength. An analytical model has been further developed to approximate the behaviour of timber shear walls reinforced by CFRP strips with satisfactory agreement [21].

3.2 Reinforcement by use of timber Chang, Hsu & Komatsu proposed a new solution to reinforce traditional planked timber shear walls (Fig. 25) after an earthquake by inserting hardwood strips into grooves in beams that accommodate these timber planks [22]. Two different species of hardwood were used, Teak (Tectona grandis) and Padauk (Pterocarpus spp.). The results revealed that the timber shear walls reinforced by Padauk and Teak show a 100% and 60% increase in strength, respectively, compared with unreinforced and intact timber shear walls. The reinforced timber shear walls also exhibit better energy dissipation under cyclic loading.

Fig. 25 Schematic drawing of the reinforcement strategy [22]

3.3 Post-tensioned strengthening Strengthening of timber shear walls by using the post-tensioned technique provides a very unique opportunity to achieve better aseismic behaviour for timber walls. In the experimental campaign described in Section 3.5, two different types of post-tensioned timber shear walls were tested, i.e. the single (Fig. 26) and coupled timber walls (Fig. 27). In the coupled wall specimens, a U-Shaped Flexural Plate (UFP) was developed and adopted to connect two smaller units of shear walls. The hysteretic loop of the coupled walls system shows a promising result as the system has combined good energy dissipation capacity and recentring effect as shown in Fig. 28.

49


Fig. 26 Schematic illustration of post-tensioned timber shear wall

Fig. 27 Schematic illustration of coupled posttensioned timber shear walls

This technique shows good potential in the future for seismic-prone areas. However, to achieve a more robust system, more research should be done to help engineers to deal with long-term creep in columns caused by post-tensioned and stress relaxation.

Fig. 28 Hysteretic loop of coupled walls [15]

4. Discussion The previous sections provide an overview on different methods to repair and reinforce damaged and undamaged timber columns and shear walls, which are tabulated in Tab 2 and Tab 3.

4.1 Reversibility When dealing with architectural heritage, techniques used to repair or reinforce a structural member should be reversible whenever possible. The literature has shown there to be a lack of research into reversible repair techniques for these valuable cultural heritages. Using composite materials, such as FRP and CFRP, with timber tends to be an irreversible technique due to the adhesive used. The screw repair technique proposed by Song et al. is reversible [5], but more research should be carried out to investigate other parameters such as spacing between screws, types of self-tapping screws, the effect of the size of cracks, etc.

50

Reinforcement of timber columns and shear walls Tab. 2 Summary for repair and reinforcement of timber columns

Cracked members

New buildings

Intact members

buildings

Existing

Biodecay members

Prosthesisation

Screw reinforcement

Steel member

Composite material

Posttensioned

Increase strength

NA

X

NA

X

NA

Increase stiffness

NA

X

NA

X

NA

Increase strength

X

NA

NA

X

NA

Increase stiffness

X

NA

NA

X

NA

Increase strength

NA

?

O

O

O

Increase stiffness

NA

?

O

O

O

Increase strength

NA

?

O

O

O

Increase stiffness

NA

?

O

O

O

[3, 23]

[5]

[6]

[7-12, 14]

[15, 16]

members

References

NA: Not applicable; O: Applicable and will increase properties (such as stiffness and strength) X: Applicable and will not increase properties (such as stiffness and strength); ?: Need further research

4.2 Long term behaviour of reinforced structural members Timber is a mechano-absorptive material and therefore creep will need to be considered when longterm loadings are imposed; it is particularly onerous when the moisture content of the timber members constantly varies between high and low levels. Post-tensioned reinforcing techniques tend to introduce high levels of stress into structural members and therefore the long term behaviour of timber columns and shear walls reinforced by this technique should be investigated. The post-tensioned system reviewed in this chapter [15, 16] introduces a compressive stress on the timber column perpendicular to the grain, where beams are connected to the column. This in turn will lead to creep in the material. How this creep will affect the reinforcement will need to be addressed in the future. Another important issue to be considered is ageing and weathering of composite material used in reinforcement and strengthening. Avent tested epoxy reinforced timber connections and pointed out that the dry condition shear strength of epoxy repaired Southern pine was reduced by one-third when the repaired member was exposed to natural weathering conditions [24].

51

Reinforcement of Timber Structures Tab. 3 Summary for repair and reinforcement of timber shear walls

New buildings

Intact members

buildings

Existing

damaged members

Prosthesisation

Composite material

Timber member

Post-tensioned

Increase strength

X

NA

O

NA

Increase stiffness

X

NA

O

NA

Increase strength

NA

O

O

NA

Increase stiffness

NA

X

O

NA

Increase strength

NA

O

O

O

Increase stiffness

NA

X

O

O

[19-21]

[22]

[15, 16]

members

References

NA: Not applicable; O: Applicable and will increase properties (such as stiffness and strength) X: Applicable and will not increase properties (such as stiffness and strength); ?: Need further research

4.3 Compatibility of materials Pizzo et al. compared the values of the thermal expansion coefficients (TECs) of two wood species (Spruce and Iroko) and four different epoxy resins. It showed that minor differences in TEC have been observed between wood in perpendicular-to-the-grain direction and an experimental epoxy adhesive [25]. Other commercial epoxy adhesives showed greater differences in terms of TEC and a proportionally decreasing mechanical compatibility. This implies that when a new material and adhesive are selected to strengthen or reinforce timber columns or shear walls, the TECs of these material needs to be considered and compatibility should be ensured for long term behaviour. Another important issue to consider when selecting new material for strengthening is to consider the compatibility of TEC of timbers so as to ensure the load being transferred from timber to reinforcement in all conditions.

4.4 Fire performance Timber has low thermal conductivity; it is measured at approximately 0.8 (measured in J/h/m2/mm/°C) compared with 12.6 for concrete and 312 for steel [26]. Most epoxies begin to soften at 90o-120oC and the strength rapidly decreases. Therefore it is good practice to inject epoxy into timber to repair the interior regions and the timber will protect and slow down the strength reduction of the epoxy. In the event of fire, the composite material and epoxy are exposed to fire in those

52

Reinforcement of timber columns and shear walls composite material reinforcement methods described previously ([5], [7], [8], [9], [20]). This will lead to the reinforcement and strengthening measurements becoming ineffective. Furthermore, when selecting the composite material and adhesives for reinforcement and repair, one should ensure that no toxic emissions occur during the fire. The strength of steel is halved when exposed at temperature of 600 oC, therefore similar situation can be found in steel reinforcement of timber members [5, 6]. It is therefore important to develop appropriate reinforcement and strengthening methods for timber members in the event of fire.

4.5 Effectiveness of prosthesisation There is no evidence as to how effective the prosthesisation technique is, although this is a widely accepted practice within architectural heritage conservation programmes. Research effort should be invested into experiments as well as into developing design guidelines for this practice, such as the buckling response of timber columns where the lower part of the portion is replaced by new timber with traditional carpentry.

5. Summary An extensive overview has been carried out in this chapter on different repair and reinforcement techniques that should be implemented on timber columns and shear walls under various circumstances. The existing research has shown that reinforcement of columns by screws and composite materials such as FRP are effective although there is a need to investigate the long term performance of these measures. Compared with timber columns, less research has been carried out to explore the strategies to reinforce and repair timber shear walls, however, reinforcement and repair of timber shear walls by employing composite materials or hardwood appear to be effective as have been demonstrated by some authors. This chapter also discusses and analyses the need for future research on the repair and reinforcement of these structural elements.

Acknowledgement The author appreciates permissions granted from Prof. Andy Buchanan and University of Canterbury for Figs. 16-18 and 24, Profs. Suzuki and Suda for Fig. 19, Profs. Weiping Zhang and Xiaobin Song at Tongji University, China, for Figs. 9 and 10.

References [1] [2] [3]

[4] [5]

Yeomans, D.T., The repair of historic timber structures, Thomas Telford, 2003. Keey, R.B., Langrish, T.A., Walker, J.C., Kiln-drying of lumber, Springer Verlag Berlin Heidelberg, New York, 2000. D'Ayala, D., Wang, H., "Conservation Practice of Chinese Timber Structures: ‘No Originality to be Changed’or ‘Conserve as Found’", Journal of Architectural Conservation, Vol. 12, No. 2, 2006, pp. 7-26. NAIT5. PixelSkin02. 2008 [cited 2014 December 30]; Available from: http://nait5.wordpress.com/2008/06/11/pixelskin02/. Song, X., Jiang, R., Zhang, W., Gu, X., Luo, L., "Compressive behavior of longitudinally cracked wood columns retrofitted by self-tapping screws", In: Proceedings of World Conference on Timber Engineering, Auckland, New Zealand, 2012.

53

Reinforcement of Timber Structures [6]

[7] [8] [9] [10]

[11]

[12] [13] [14]

[15]

[16]

[17] [18]

[19] [20]

[21] [22] [23] [24] [25] [26]

54

Tanaka, H., Idota, H., Ono., T., "Evaluation of buckling strength of hybrid timber columns reinforced with steel plates and carbon fiber sheets", In: Proceedings of 9th World Conference on Timber Engineering , Oregon, United States, 2006. Plevris, N., Triantafillou, T.C., "FRP-reinforced wood as structural material", Journal of Materials in Civil Engineering, Vol. 4, No. 3, 1992, pp. 300-317. Triantafillou, T.C., Deskovic, N., "Prestressed FRP sheets as external reinforcement of wood members", Journal of Structural Engineering, Vol. 118, No. 5, 1992, pp. 1270-1284. Ehsani, M., Se, M.L., "Strengthening of old wood with new technology", STRUCTURE, 2004. pp. 19. Zhang, W., Song, X., Gu, X., Tang, H., "Compressive behavior of longitudinally cracked timber columns retrofitted using FRP sheets", Journal of Structural Engineering, Vol. 138, No. 1, 2011. pp. 90-98. Oprişan, G., Ţăranu, N., Enţuc, I.-S., "Strengthening of the timber members using fibre reinforced polymer composites", Buletinul Institului Politehnic DIN IAŞI, Vol. L (LIV)( 1-4), 2004, pp. 67- 75. Taheri, F., Nagaraj, M., Khosravi, P., "Buckling response of glue-laminated columns reinforced with fiber-reinforced plastic sheets", Composite Structures, Vol., 88, No. 3, 2009. pp. 481-490. Davidovits, J., "Geopolymers", Journal of Thermal Analysis and Calorimetry, Vol. 37, No. 8, 1991, pp. 1633-1656. Najm, H., Secaras, J., Balaguru, P., "Compression tests of circular timber column confined with carbon fibers using inorganic matrix", Journal of Materials in Civil Engineering, Vol. 19, No. 2, 2007, pp. 198-204. Palermo, A., Pampanin, S., Fragiacomo, M., Buchanan, A., Deam, B., "Innovative seismic solutions for multi-storey LVL timber buildings", In: Proceedings of 9th World Conference on Timber Engineering, Oregon, United States, 2006. Smith, T., Pampanin, S., Fragiacomo, M., Buchanan, A., "Design and construction of prestressed timber buildings for seismic areas", NZ Timber Design Journal, Vol. 16, No. 3. 2008, pp. 3-10. Suzuki, Y., Maeno, M., "Structural mechanism of traditional wooden frames by dynamic and static tests", Structural Control and Health Monitoring, Vol. 13, No. 1, 2006. pp. 508-522. Suda, T., Tasiro, Y., Suzuki, Y., "Seismic Reinforcement by Restoring Force due to Column Rocking for Traditional Wooden Frame (in Japanese)", In: Proceedings of Disaster Prevention of Historic Cities, 2011. Dobrila, P., Premrov, M., "Reinforcing methods for composite timber frame–fiberboard wall panels", Engineering Structures, Vol. 25, No. 11, 2003, pp. 1369-1376. Premrov, M., Dobrila, P.,Bedenik, B., "Analysis of timber-framed walls coated with CFRP strips strengthened fibre-plaster boards", International Journal of Solids and Structures, Vol. 41, No. 24, 2004, pp. 7035-7048. Premrov, M., Dobrila, P., "Mathematical modelling of timber-framed walls strengthened with CFRP strips", Applied Mathematical Modelling, Vol. 32, No. 5, 2008, pp. 725-737. Chang, W.-S., Hsu, M.-F., Komatsu, K., "A new proposal to reinforce planked timber shear walls", Journal of Wood Science, Vol. 57, No. 6, 2011, pp. 493-500. Pinto, L., Inventory of repair and strengthening methods timber, in Departament d'Enginyeria de la Construcció, Universitat Politècnica de Catalunya: Barcelona, Spain, 2008. Avent, R.R., "Decay, Weathering and Epoxy Repair of Timber", Journal of Structural Engineering, Vol. 111, No. 2, 1985, pp. 328-342. Pizzo, B., Rizzo, G., Lavisci, P., Megna, B., Berti, S., "Comparison of thermal expansion of wood and epoxy adhesives", Holz als Roh-und Werkstoff, Vol. 60, No. 4, 2002, pp. 285-290. Avent, R.R., Issa, C.,A., "Effect of fire on epoxy-repaired timber", Journal of Structural Engineering, Vol. 110, No. 2, 1984, pp. 2858-2875.

Analysis and strengthening of carpentry joints

4 Analysis and strengthening of carpentry joints Jorge M. Branco1, Thierry Descamps2

Summary Joints play a major role in the structural behaviour of old timber frames [1]. Current standards mainly focus on modern dowel-type joints and usually provide little guidance (with the exception of German and Swiss NAs) to designers regarding traditional joints. With few exceptions, see e.g. [2], [3], [4], most of the research undertaken today is focused on the reinforcement of dowel-type connections. When considering old carpentry joints, it is neither realistic nor useful to try to describe the behaviour of each and every type of joint. The discussion here is not an extra attempt to classify or compare joint configurations [5], [6], [7]. Despite the existence of some classification rules which define different types of carpentry joints, their applicability becomes difficult. This is due to the differences in the way joints are fashioned depending, on the geographical location and their age. In view of this, it is mandatory to check the relevance of the calculations as a first step. A limited number of carpentry joints, along with some calculation rules and possible strengthening techniques are presented here.

1.

Timber frameworks and carpentry connections

Timber frameworks are one of the most important and widespread types of timber structures. Their configurations and joints are usually complex and testify to a high-level of craftsmanship and a good understanding of the structural behaviour that has resulted from a long evolutionary process of trial and error. A simplified analysis of (old) timber frameworks, considering only plane parts of the system, is often hard to realize. Nowadays, a considerable number of timber structures require structural intervention due to material decay, improper maintenance of the structure, faulty design or construction, lack of reasonable care in handling of the wood, accidental actions or change of use. While the assessment of old timber structures is complex, it is an essential precursor to the design of the reinforcement of the joints. Owing to a lack of knowledge or time, the species and/or grade assumed are often an overly conservative estimate which can lead to unnecessary replacement, repair and retrofit decisions along with associated superfluous project costs. For the design of the reinforcement of old timber structures or joints, the first step is to understand fully how the structure and the joints work. Old timber structures are usually highly statically indeterminate structures. This means that loads applied to the structure have different pathways to reach the supports. Resolving the indeterminate system involves looking for additional equations that actually express the relative stiffness of all those pathways. To illustrate how the differential stiffness of elements, joints or supports may influence the behaviour of the structure, a simple collar-braced roof is presented in Fig. 1. 1)

Assistant Professor ISISE, Dept. Civil Eng., University of Minho, Guimarães, Portugal

2)

Assistant Professor, URBAINE, Dept. Structural Mech. and Civil Eng., University of Mons, Mons, Belgium

55

Reinforcement of Timber Structures In the absence of buttressed walls, under vertical loads, the collar (or the tie-beam) is under tension because it prevents the roof from spreading. If buttressed walls restrain the feet of the rafters, the collar is in compression. The only difference between these situations is the horizontal stiffness of the supports (zero or infinite). The mass of the walls to resist the outward thrust is not the only influencing factor. Most of the time, principal rafters are connected to wall plates that have to be stiff enough to act as a beam in the horizontal plane spanning between two fixed ends in the walls. If the rafters are notched, for example, with birdsmouth joints, over the plate at the top, the roof can be hung from the ridge purlin, depending on the stiffness of the wall plate. The stiffness determines the ability of the wall plate to act as an additional support. This is valid for most types of carpentry joints as they usually are statically indeterminate. In conclusion, when working on old carpentry joints, it could be useful, when possible, to look at the joint as an assembly of equivalent springs. This model allows a better understanding of how the joints behave and deform and determines where the major stresses will occur. This could help to avoid incorrect positioning of the reinforcement and thereby circumvent poor design.

Fig. 1 Collar-braced roof The main challenges for the structural assessment of carpentry joints are [8]: •

• •

56

Stiffness and strength of joints depend on the type of loading. As an illustration, the rotational stiffness of a joint is mostly different under positive and negative bending. Moreover, within most joints there is an interaction between the different pathways in which the forces are transferred in terms of stiffness and strength. This interaction should be considered to define the mechanical behaviour of the connection. Despite most current standards not declaring any rules for the assessment of the material strength under combined stresses, their appearance in carpentry joints is inevitable. The design of traditional joints essentially involves a check of the contact pressure between the assembled elements. It is not easy to calculate the value of contact pressure in the following situations: unknown contact surfaces and non-uniform stress distributions (because of nonuniform elastic support due to local defects like knots for example).The values of compressive strength of timber are different in the direction parallel and perpendicular to the grain. In order to calculate the strength at any intermediate value of the load angle to the grain Hankinson’s formula, which has been presented in many standards, may be used. SIA 265:2003 [9] suggests a different expression that takes into account a reduction because of the difference between the strength of early wood and latewood. In addition, some standards allow enlarging the real contact surface by taking into account a so-called effective length [10]. Those slight differences about the definition of the compressive strength at an angle to the grain highlight a lack of knowledge, which fortunately, is not of major importance for compression at angles between 30º to 60º (which represent the most common values).


2.

Old carpentry joints

Common traditional carpentry joints found in old timber frames can be categorized in four main types, according to their arrangement and geometry: •

Mortise and Tenon joints: There are countless examples of this type of joint. Tenon joints connect members that usually form an "L" or "T" type configuration. The joint comprises two components: the mortise hole and the tenon tongue. The tenon formed on the end of a member is inserted into a square or rectangular hole cut into the corresponding member. The tenon is cut to fit the mortise hole exactly and usually has shoulders that sit when the joint fully enters the mortise hole. The joint may be pinned or locked into place. In the traditional fashion, the pin hole in the tenon is bored a little closer to the shoulder than in the mortise and the pin pulls the joint together very tightly. This kind of joint is mainly used when the adjoining pieces connect at an angle between 45° to 90°. When the angle between the two jointed elements is different from 90º, the nose of the tenon can be cut off and is called a skewed tenon (see Fig. 3a & 6a).

Fig. 2 (a) Through pinned mortise and tenon (a’) blind pinned mortise and tenon (blind means not going all the way through). (b) Through tenon with outside wedges (flatwise bending of the tenon (b’) wedged and pinned dovetail through mortise and tenon •

Notched joints: This kind of joint is linked to the development of king post and king post-like frames. In order to work successfully, these frames need appropriate joinery at a multitude of locations. A notch is a "V" shaped groove generally perpendicular to the length of the beam, as seen in Fig. 3. Examples where notched joints are used include cases where secure footing is required for the toe of a rafter (or strut) or between the rafter and the king-post. A tenon can be added to the notched joint to essentially keep all the beams coplanar but the notch is what creates the strength of the joint (because it is stiffer than the tenon).

Fig. 3 (a) Notched joint between main rafters and tie-beam.(a’) A skewed tenon may be used to help in keeping all timber pieces co-planar. (b) Peak joint with a notched joint (main rafters and post)

57

Reinforcement of Timber Structures •

Lap joints: In a full lap joint, no material is removed from either of the members to be joined, resulting in a joint whose thickness equals the combined thickness of the two members. The members are held in place by a pin (Fig. 4a). In a half-lap joint, material is removed from each of the members so that the thickness of the resulting joint is the same as that of the thickest member. Most commonly, in half-lap joints, the members are of the same thickness and half the thickness of each is removed. The cogged half-lap joint is a half-lap with additional cogs. The dovetail-lap joint (named after the shape of the tenon being similar to the tail of a dove) is another way to fashion the joint in an attempt to reinforce its tensile strength (Fig. 4c).

Fig. 4 Full lap joint (pinned). (b) Half-lap joint. (b’) Cogged half-lap joint. (c)Through dovetailed lap joint or (c’) wedged dovetailed lap joint if ever the dovetail is embedded in the member •

Scarf joints: Scarf joints (and splice joints), shown in Fig. 5, allow the joining (splicing) of two members end to end [11], [12]. They are mainly used when the material being joined is not available in the length required. This technique is recognised as being the strongest form of unglued member lengthening [13]. The halved-scarf joint is a lap whose surfaces are parallel with the members. It is similar to a half-lap joint with co-axial members. The scarf joint is simply a pair of complementary straight sloping cuts secured to each other with pins (also called pegs).

Fig. 5 (a) Common and simplest halved-scarf joint (or half-lap splice joint). (a’)A lapped dovetail scarf joint is a half-lapped joint in which the lapped portions are shaped like a dovetail joint. (b)Scarf joint. (c)Scarf joint with under-squinted ends. (d) Trait de Jupiter: particular scarf joint with wedges (key)

58

Analysis and strengthening of carpentry joints Another type of scarf joint is known as the Trait-de-Jupiter or Bolt of lightning, in view of its resemblance to lightning. It is more efficient in the presence of a key (or several keys, depending on the number of indentations – see Fig. 23d) made of hardwood to improve contact and to simplify fabrication. From a mechanical point of view, it is an excellent scarf, since the driving of its key separates the twin-tables with a primary mechanical force and closes the under-squinted butts with enormous pressure.

3.

Joint stiffness

Numerous examples demonstrate the excellent performance of old timber constructions during earthquakes or exceptional wind loads. The reason why they are still standing is not only due to their robustness (highly statically indeterminate structures), but also due to the semi-rigid and ductile behaviour of their joints, which allow for the dissipation of energy. Also, thanks to its loadredistribution process, the beams and joints are able to maintain the capacity of the whole structure in spite of partial damage. According to common standards, such as Eurocode 5 [10], the rigidity of elements and joints as well as the eccentricities of the joints have to be taken into account for the computation of the internal forces. However, in order to simplify the analysis, joints are usually designed by assuming an ideally pinned (or rigid) behaviour [14]. It is quite obvious that the assumption of pinned joints is conservative, provided that the joints have enough ductility and are fashioned in a way that their rotation may develop (deformation capacity is sufficient). Nevertheless, in reality most of the carpentry joints are not perfect hinges. Though this is not of major importance for the design of the members, it must be borne in mind that the splitting of timber may occur under low loads (because of component loads perpendicular to the grain). Therefore, in some cases the joints are assumed to be rigid. This is conservative for the joints and results in an uneconomic design. Furthermore, carpentry joints usually have a significant moment-resisting capacity even without any strengthening devices. Test results on full-scale notched joints show that this capacity is a function of the compression level in the rafter, the width of the rafter, the friction, the skew angle and the notch depth [3], [15], [16]. Rotational capacity is positively related to the first three parameters. Undoubtedly, the modelling of the structure taking into account the semi-rigid behaviour of the joints is the best practice:

3.1



A semi-rigid study of a structure suggests taking into account the stiffness of the joints with regard to all the components of the loads (normal, shear and bending). In fact, Descamps et al. [4] have shown that for the computations of the internal forces, the use of the rotational stiffness alone is not enough. Both axial and rotational stiffness have to be introduced in finite element models for an accurate study. The shear stiffness is of less importance.



Uzielli et al. [17] reported on research work in which different assumptions about the joints in an old timber structure were compared. They found a maximum difference of 20% between the computed stresses in a semi-rigid model compared to the experimental results, while the difference increased up to 40% when assuming pinned or rigid joints.

Component method

Design models are available in all standards for estimating the stiffness of dowel-type joints. Unfortunately no information is given about how to get the stiffness of carpentry joints in order to help engineers to gain better results. The component method allows stiffness values to be determined

59

Reinforcement of Timber Structures for joints according to their geometrical and mechanical properties. This method has been used frequently in the field of steel construction and is now applied in research on carpentry joints by several authors [8], [18], [19], [20], [21], [22]. The problem will be explained by considering a skew tenon joint under an axial load (Fig.6). Since different loads paths are possible (the joint is statically indeterminate), the worst case scenario has been chosen (the mortise is longer than the tenon, which is a common way of fabricating this type of joint). The component Fh is balanced through a uniform contact pressure on the shoulders of the tenon (surface (1) in Fig. 6a). The eccentricity of the resulting forces is not considered here. The load transferred up to the axis of the post (loaded perpendicular to the grain) causes a deformation that can be defined by the stiffness Kh: (1) where E90 is the mean value of the modulus of elasticity perpendicular to the grain, A1 is the area of the contact surface (1) and L is the height of the section of the post. Assuming a gap between the end of the tenon and the post, the component Fv is balanced through a uniform contact pressure on the head of the tenon (surface (2) in Fig. 6a). The load transferred to the tenon causes a deformation that can be defined by the stiffness Kv: (2) where Eα is the mean value of the modulus of elasticity at an angle α to the grain, and A2 is the area of the contact surface (2). Fig. 6b gives the equivalent spring model of the joint under an axial load. If M is the moment in the joint and θ is the rotation into the joint (Fig. 6c), the rotational stiffness of the joint is: (3)

Fig. 6 (a) & (b) Skewed tenon joint under an axial load and the equivalent beam and spring model. (c) & (d) Tenon joint under bending and the equivalent beam and spring model

Fig. 7 Definition of a modified modulus of elasticity

60

Analysis and strengthening of carpentry joints The applied moment is balanced through contact pressures on surfaces (1), (2) and (3). Those pressures can be assumed to be uniform or non-uniform. The effect of friction has not been considered. Each of the contact pressures causes a deformation that can be divided in two components. One component of deformation is caused by the material (i), for example, loaded in parallel to the grain and another component of deformation is caused by material (i'), in contact with (i), for example, loaded perpendicular to the grain. The stiffness Ki and Ki’ can be defined as proposed by Meisel et al. Ki,i’ is the equivalent stiffness of two springs Ki and Ki’ in series (in material (i) and (i’), respectively) [8]. Drdácký et al. have proposed another definition of the stiffness Ki based on a well-known model used to calculate the settlement under a rectangular foundation supported by a semi-infinite half space [18]: (4) If i is the deformation at surface contact (i) and (i’), Fig. 6c: (5) For small displacements, such as the surface contact (1): (6) Finally, the rotational stiffness is equal to: (7) Fig. 6d gives the equivalent spring model under bending. Some enhancements of the method, in particular for the definition of the stiffness Ki, have been proposed by Descamps et al. [4]: 

Definition of a "modified modulus of elasticity" that takes into account the edge effect that appears when surface (i) and (i') are next to an edge. C is defined in Fig. 7 where  is the slenderness of the contact area. (8)



New definition of the centre of rotation of the joint. Observations of broken pegs made on real-size tests suggest that the peg is not the centre of rotation of the connection. The assumption has been made that the common position of the centre of rotation corresponds to a minimum of the bending stiffness of the joint (iterative procedure).

Komatsu et al. [20] proposed an enhanced model with a definition of the elastic stiffness Ki that takes into account an effective length for the definition of Ai (as in Eurocode 5 for the check of compression perpendicular to the grain). They have also proposed a definition of the stiffness in the post yielding range. Chang et al. have proposed some enhancements to this model, in particular to take into account initial gaps and slips [21]. Comparisons with experimentation have demonstrated that the enhanced model achieves good results not only for the initial stiffness of the joints with gaps, but also for the initial slip stage which should be regarded as a pin connection in the early stages. The component method can be applied to other types of carpentry joints. Researchers have concentrated its use on tenon [4], [8], [18] and lap joints [20], [21], but its application to notched joints is also possible.

61


4.

Evaluation and reinforcement of carpentry joints

Before any intervention, the first step is the assessment of the existing joints in relation to the material, the strength and the stiffness. Proper assessment of the material (decay) with appropriate techniques is obviously of major importance and therefore the study of recent state-of-the-art is highly recommended [23]. This survey may lead to the replacement of a portion or the whole member. On the other hand, in the case that the member or joint is kept in service and reinforcement is needed, an accurate assessment of the state of conservation of timber is crucial. In the past, the first action taken by carpenters to strengthen joints was based on precise observations of failure modes encountered in real structures and a good understanding of their weakest points. This led to an improvement in the sketching of the joints and one can say that many carpentry joints are an evolution or a reinforcement of older primary joints. For example, a notched joint with a tenon (see Fig. 3a’) can be considered as a “reinforcement” of a tenon joint because the slope of the notch increases the load bearing capacity of the joint. In the past, joints were realised without any metal fasteners such as nails, screws or bolts and their ability to carry the loads was achieved through direct contact and friction. Over the years, various reinforcement techniques such as the use of screws (including self-tapping-screws), metal plates (strips, stirrup...), glued composites (glass or carbon fibres, weft knitted textiles) and glued-in rods or even full injection with fluid adhesives among others have been proposed. It should be noted that special attention has to be paid to any solution that consists of wrapping the joint in an airtight textile (risk of decay). Furthermore, for the restoration of historic buildings, all interventions should be reversible; if not completely, they should not limit further interventions. For this reason, the injection with fluid adhesive directly into the joint is not recommended anymore [24]. Dowel-type fasteners have been used occasionally in timber joints, for example, to counteract any out-of-plane displacements which cannot be counteracted by the joint itself. This practice became common in the 19th century with the development of industrial production methods and the manufacture of low cost fasteners. Nowadays, the strengthening may aim to locally reinforce the material in the joint area, for example, to reinforce the timber in shear or tension perpendicular to grain by means of self-tapping-screws or to avoid the detachment of the connected elements (joints that could not carry any tension loads for example) or to modify locally the pathway followed by the loads into the joint. Particularly, in seismic areas, strengthening can prevent loss of capacity and possible separation of contact surfaces due to the decrease of compression forces, and may maintain a suitable structural behaviour [2]. The first step of any reinforcement intervention is of course the definition of a proper model of the joint to assess its strength and stiffness. Models and reinforcement techniques will be discussed for the most common carpentry joints here after.

4.1

Tenon joints

Tenon joints have a very low stiffness that may cause premature failure of a part or the whole structure caused by large displacements encountered in the joint [2]. The bearing capacity of skewed tenon joints is a function of the angle α of the joint, the length of the tenon and the mortise depth [25], [26], [27]. To check the joint, one may use a simple check on all components of the load that appear on the surfaces as it has been discussed in Section 3.1.

62


Fig. 8 Configuration and force mechanism in a skewed tenon joint Each part (i) or (i') of the surfaces in contact is checked in compression at an angle to the grain. Kock et al. have developed guidelines for design that are suitable for skewed tenon joints, under axial and shear loading [28]. Surfaces A1 and A2 in contact are presented in Fig 8. μH and μV are the coefficients of friction of A1and A2, respectively, hs and ls are the height and the length of the strut, respectively, and tx is the distance between the bottom surface and the loading point of H. The compression loads H and V on A1and A2, respectively, are defined as follows:

(9) (10)

where m is the ratio of Q to F: (11)

4.1.1 Intervention and reinforcement of tenon joints Feio et al. have tested full-scale notched and skewed tenon joints under compression in order to assess the local failure in compression and the slipping of the joint [29]. Failure modes observed in the tested joints are damages due to compression in the brace which are localized at the tenon end or distributed along the full contact length. An out-of-plane bulging of wood under the contact length was observed. In some cases, damages in compression associated with shear failure were observed (Fig. 9). When observed on-site, this type of failure mode mainly highlights a poor design of the joint (with contact areas that are too small) or of the structure (unexpected compression forces in one element). No reinforcement can repair damage in compression perpendicular to the grain and the replacement of the element is required most of the time. 

63


Fig. 9 Typical experimental failure patterns: (a) joint collapsed in compression, with uniform distribution of damage, (b) joint collapsed in compression, with out-of-plane bulging, and (c) combined failure in compression and shear parallel to the grain at the tenon [29] 

To ensure correct strength and stiffness of the joint, it is important to keep all the surfaces of the joints in contact. In the case of reverse loads or because of high shrinkage of the wood elements, joints may develop gaps. One traditional reinforcement technique consists in placing a wooden wedge to ensure perfect contact between the tenon and the mortise (Fig.10). This wooden wedge should be made of hardwood (for strength and stiffness) and its moisture content (MC) should be as close as possible to that of the reinforced wooden elements in order to avoid any shrinkage of the wedge.

Fig. 10 Wooden wedge to ensure the contact between the tenon and the mortise



Pinned tenon joints also have a very low bearing capacity in tension as only the pin acts. If the element has to be replaced, a traditional reinforcement technique consists in fashioning the joint with a dovetail tenon to increase the strength in tension (Fig. 2b’). If the element remains in service, a binding strip may be used as reinforcement in tension. The strip is screwed onto the edge of the supporting beam to avoid any crack (Fig. 11). For the design of the fastening of the strip, Eurocode 5 expressions for double shear in timber-to-steel connections can be utilized. One may check the tensile stress in steel and the compression perpendicular to the grain under the strip as follows: (12)

64

Analysis and strengthening of carpentry joints where

is the tensile strength of steel and

is the shear strength of the fastener. To

avoid any tensile forces in the element, a steel wire may be used as presented in Fig.11b. If broken pins are observed, replacing the wooden pins by steel ones is not suitable as it.

Fig. 11 Reinforcement of a tenon joint in tension

4.2

Notched joints

In a notched joint, the slope of the notch should minimize the angle between the stresses and the grain direction for both connected elements, hence increasing the crushing resistance of the joint.

Fig. 12 Force mechanisms in a notched joint (tie beam and rafter for example) with the contact surface at the front (a) or at the rear (b) Based on simple geometric considerations, it is possible to demonstrate that the ideal configuration of the notched joint is the one reported in Fig. 12a, where the angle of the notch is half of the angle . According to Götz et al [30] and German and Italian standards [31], [32], the depth of the notch, tv, should not exceed h/4 for  50º and h/6 for >60º (linear interpolation between those values is proposed). Friction forces and geometric imperfections are not considered. Based on these assumptions, the axial force is easily resolved into two component forces F1 and F2 perpendicular to the two surfaces of the notch (Figure 12a): (13)

65


(14) If b is the width of the timber elements, the compression at an angle to the grain direction on the notch and the shear in the frontal shear plane must be checked: (15) (16)

For the rear face under compression: (17) where d is the length of the rear compressed surface of the notch. This last verification, which is often neglected, can be of importance because of the risk of high stresses on a surface of limited length. Parisi et al. proposed the following empirical rule to calculate d [2]: (18)

It is important to draw attention to the assumptions that have been made. Assuming that there is no friction is quite far from reality. Friction may increase the stresses in the shear surface. The check of shear stresses in front of the notch must be done with caution. In consequence of the fragile nature of the shear failure, for instance, some standards on earthquake resistant structures (e.g.[31]) adopted a higher partial factor for material properties (equal to 1,3) in the quantification of fv,d. The notched joint with the contact surface at the rear aims to increase the shear strength of the joints by increasing the shear surface in front of the notch (Fig.11b). To fashion the joint easily, a notch is made perpendicular to the direction of one of the two members. In this case the joint strength decreases because the slope of the notch does not minimize the angle between the stresses and the grain direction for both connected elements anymore. The compression at an angle to the grain direction at the notch and the shear in the frontal shear plane must be checked: (19) (20)

Owing to the eccentricity between the load F and the contact surface, the joint will turn (open) and a crack can appear. Even with a gap of 1 or 2mm between the two connected members which prevent the nose bearing of the rafter, it is almost impossible to avoid the splitting of the rafter. Double step joints, whose geometry results from the sum of two different single steps joints, increase the shear surface without a major risk of splitting (Fig. 13). To fashion the joint, the rear notch must be deeper than the one at the front. To ensure that the joint works well, precision is required so that all surfaces will be in contact (which is not easy to get especially when it is done manually). According to Italian standard: and [32]. Whatever the skew angle of the connection, for a double notched joint, Götz et al. recommends and [30], [32].

66


Fig. 13 Double notched joint. Strengthening of notched joint also concerns the friction-based behaviour of the joints in its own plane, preventing the separation of friction surfaces due to the decrease of compression forces (a notched joint has no tensile strength). A series of monotonic and cyclic tests on unstrengthened and strengthened joints has been performed by Branco et al. in order to study the initial behaviour of the connection, as well as its sensitivity to a few parameters [3]. Even without any strengthening devices, notched joints usually have a significant moment-resisting capacity. This capacity depends on the axial compression load in the rafter and on the skew angle [2], [3]. Moreover, it is obvious that the height of the rafter [33], the friction [2], the existence of an additional tenon [30] and the moisture content [16] are also important.

4.2.1 Intervention and reinforcement for notched joints 

If the wooden elements do not perfectly match in the notched area (lack of precision in the cutting of the members or because of shrinking), the placement of wooden wedges is recommended to ensure a perfect contact between connected surfaces with a clear increase in the load-carrying capacity of the joint [34].

Fig. 14 Wooden wedges used to ensure a perfect contact between elements. Gaps in the front notch (a) and rear surface (b) and possible use of wedges, (a’) and (b’), respectively’ 

The strengthening of existing notched joints mainly aims to avoid shear failure in the front portion of the notch. Most of the time, an end beam repair is required because of decay and a wooden prosthesis must be used to replace the degraded material (Fig 15).The prosthesis is mechanically jointed to sound wood (or resin). One may use the following check: where

is the effective number of fasteners and

(21) is their load-carrying capacity [10].

Using inclined fasteners increases the load-carrying capacity of the prosthesis.

67


Fig. 15 Notched joints reinforced to shear stresses in the frontal part of the notch with a screwed prosthesis 

In times past, binding strips, stirrups and bolts were used in seismic regions to avoid the dismantling of the connected members under reverse loads (Fig. 16). When metal elements were used in the original construction of the joints, or, added later, the intervention usually included the substitution of the connectors (nails, bolts, etc.) by new ones and the treatment of the metal.

Fig. 16 Examples of notched joints with metal devices The strengthening techniques used presently look to reproduce the old techniques even when using new fasteners like screws and self-tapping screws (see Fig. 17). These kinds of interventions affect the stiffness of the joint, which should be checked too.

Fig. 17 Contemporary strengthening interventions on notched joints reproducing old techniques



68

Strengthened joints with metal devices were tested by Branco et al. under monotonic and cyclic loading [35]. The purpose was to uncover any advantages and drawbacks in the behaviour of the joint and of the strengthening as well as to look at different types of

Analysis and strengthening of carpentry joints strengthening. The four types of strengthened joints tested are modern implementations of traditional techniques (Fig. 18):

Fig. 18 Traditional strengthening techniques of notched joints: (a) metal stirrups, (b) internal bolt; (c) binding strip (d) tension ties o

o

o

Metal stirrups placed in pairs on two sides of the joint have been a popular reinforcement in the past. Each stirrup was composed of two steel plates welded in a V-shape and bolted with seven bolts of 10mm of diameter. Each prong was 50mm wide and 5mm thick. The steel rod (12mm of diameter) was fixed by a nut at both ends and secured by using a special rectangular-shape washer (70×30mm² and 5mm thick). The rod was located at mid-joint and mid-width and normal to the axis of the tie beam. A notch has been cut in the rafter to ease the contact between wood and the washer. Metal binding strips were also frequently used in the past, particularly to strengthen joints at an angle of 30° [33]. Two updated versions were considered:  The joint was bound by a steel ribbon (50mm wide, 5mm thick) located at midconnection, normal to the tie-beam.  The joint was bound with two steel plates located in the bottom surface of the tie beam and upper surface of the rafter (40×159mm², 10mm thick) tightened through two rods of 12mm. One may notice that this solution enables a control of the tightening force during the strengthened lifetime.

All have been analysed for α=30º and α=60º except the rigid binding strip (α=30° only - Fig. 18c). Force-displacement curves under monotonic loading for unstrengthened and strengthened joints (whatever the α values) are presented in Fig. 19. It was observed that all the strengthening techniques used have resulted in improved joint behaviour.

Fig. 19 Force-displacement curves under monotonic loading for unstrengthened and strengthened connections with α=30º (a) and α=60º (b) Despite the amount of tests done, results are still insufficient to propose design equations for all the tested configurations and reinforcements. There is still an evident lack of results and scientific data

69

Reinforcement of Timber Structures about this topic, which clearly points out the lack of research in this field. However, some interesting observations can be drawn. For α= 30°, one may notice: o

All the strengthening techniques used have increased the stiffness, in particular for the positive loading direction and the maximum load for both directions. One should check if this modification of the stiffness has any consequence on the whole structure.

o

Large improvement of the ductility, especially under a negative loading (the reinforced joint does not behave in a brittle way anymore).

o

Reinforcement with stirrups and binding strips are similar from the point of view of the maximum load reached. However binding strips have a lower ductility under negative forces.

o

Among all strengthening techniques tested, the least efficient regarding both maximum force and stiffness is the solution with the tension ties.

For α= 60°:

4.3

o

The behaviour of the unreinforced joint is ductile. Nonetheless, a significant increase of the ductility of the reinforced joint has been observed.

o

The same conclusion can be given for the increase in maximum load (whatever be the technique used).

o

No significant influence on initial stiffness has been observed.

o

The efficiency of the reinforcement is almost the same for all techniques under positive loading. For a negative loading, the metal stirrup is the most efficient technique whereas the solutions with tension ties and bolt are similar.

Lap joints

The pin used in a full lap joint is of major importance as the joint cannot carry any loads without it. Based only on the strength of the pin, the efficiency of the full lap joint is of course very low. The half-lap joint is a first improvement of the joint given that it carries the loads by contact in addition to the pin. In the half-lap joint, both the connected elements are half weakened, however the joints may be fashioned differently to avoid weakening the members (one-third of the height instead of a half). As a result the two pieces do not sit flush, which can limit the use of the joint. This configuration is useful when both members bear a larger load. The lap joint presented in Fig. 4b’ (between the half-lap joint and the cogging joint) is configured to increase capacity for both members by adding bearing surfaces. The lower supporting beam has less material removed compared to a half-lap. The side housings provide better support for the upper girder and lessen shear problems [6]. To check the joint, the total load can be resolved into equivalent loads acting on contact surfaces. Each component is checked with regards to the strength in compression at an angle to the grain.

4.3.1 Intervention and reinforcement for lap joints

70



If the wooden elements do not match perfectly in the notched area (due to lack of precision in the cutting of the members or because of shrinking), the placement of wooden wedges should be recommended to ensure perfect contact and, consequently, the attainment of full loadcarrying capacity of the joint.



If a shear failure is observed in one of the elements, it can be reinforced with fully threaded

Analysis and strengthening of carpentry joints self-tapping screws (screws are better than bolts in case of shrinkage). Fully threaded screws are required as they have to suspend the lower part of the beam from the upper part the lower part of the beam to the upper part. Inclined screws may achieve a better load-carrying capacity of the reinforced joint than screws screwed perpendicular to the grain direction. To guarantee an efficient reinforcement, the distance a between the screw and the notch should be minimized but with .

Fig. 20 Reinforcement in tension perpendicular to grain in lap joints: (1) reinforcement with selftapping screws perpendicular to the grain. (1’) reinforcement with inclined screws Design equations developed for notched beams can be used. The design tensile force, to be carried by the reinforcement, can be determined according to [36]: (22) Where

is the design shear force and



In dovetail-lap joints, the peg keeps all of the wooden pieces together and prevents the formation of gaps. When a gap occurs, the contact surfaces become smaller and so, the contact pressure becomes larger. One traditional reinforcement technique consisted in the placement of a wooden wedge to ensure perfect contact (see notched joints).



In dovetail-lap joints loaded in tension, the splitting of timber is a common failure mode. The traditional reinforcement of those joints consists in adding fasteners (bolts, nails, screws, etc.) restoring the shear mechanism provided by the pin (Fig. 21). The design of this strengthening technique is based on the calculation of the shear resistance of the new fasteners. This intervention affects the stiffness of the joint (displacement of the centre of rotation). Binding strips or steel wire may also be used (see tenon joints).

4.4

Scarf joints

Two timber members connected using scarf jointing techniques cannot match the strength and stiffness of a single member of the same dimensions. Besides the type of scarf joint, the actual size of the elements, the strength of the wood and other factors have a substantial effect on the assembled member’s strength and stiffness. All the different types of existing scarf joints are a proof of all attempts made by carpenters to strengthen the joints and fit particular requirements. For example, the scarf joint (Fig. 5b) is an improvement of the half-scarf joint (Fig. 5a) that works better in shear because less material is removed from each of the members and there is no sharp angle. If the half-lap is horizontal (and loads are vertical, see Fig. 23a), the maximum moment it can carry is one-quarter of the moment of a solid beam, because the half-lap has the width of the beam but one-half of the height. For a vertical half-lap (the scarf is face-halved and loads are vertical too), the width is one half of the solid beam but the height of the half-lap is equal, so the maximum moment it can carry is one-half of

71


Fig. 21 Traditional reinforcement of dovetail-lap joints under tension loads by adding wooden dowels

Fig. 22 Scarf joints reinforced with metal connectors and plates.

Fig. 23 (a) Scarf joint reinforcement perpendicular to the grain with self-tapping screws. (b) Reinforcement of bending strength (weak axis) with a cog (half cogged scarf joint). (c)Face-halved scarf joint. (d) Multiple scarf joint with under-squinted ends.

Fig. 24 Scarf joint reinforced with glued in rods: steel rods are glued in both timber members and connected with a long nut (Credits: Pascal Lemlyn. Restauration du Moulin de l’abbaye de la Paix Dieu, Institut du patrimoine Wallon, Belgique).

72

Analysis and strengthening of carpentry joints what a solid beam will transfer. A study carried out by TRADA suggested that the limiting moment capacity of scarf joints (which behave better in bending than the half-lap scarf) is equal to only a third of the strength of the unjointed beam [37]. For the design, one must check all components of the load that appear on the surfaces in contact. However, it should be noted that very few research campaigns have been conducted on the reinforcement of scarf joints or even on their design. Four types of reinforced scarf joints have been tested by Hirst et al. [12]. The beam sections were 200x150 mm², all pin holes were 19 mm in diameter and in keeping with traditional practice, the pin holes on one side of the scarf were offset by 3 mm with respect to the other side of the joint (tightening the joint when the pins were driven through the joint). Once again, the results are still not sufficient to discuss about the advantages and drawbacks of different techniques or to propose a design equation for reinforced scarf joints.

4.4.1 Intervention and reinforcement for scarf joints

5.



If the wooden elements do not perfectly match in the notched area (due to a lack of precision in the cutting of the members or because of shrinking), the placement of wooden wedges should be recommended to ensure a perfect contact and hence the achievement of the full load-carrying capacity of the joint.



The easiest way to reinforce a scarf joint is achieved by adding metal fasteners (screws or bolts). In ancient times, wooden pegs were used. In case of high loads, lateral metal plates can be added to improve the load-bearing capacity of the joint and to increase the stiffness. Both types of reinforcements are used in restoration works (Fig.22).



Under bending, the rule of thumb that the weak point is the risk of premature splitting of wood is encountered here too (joints cut with right angles are less suitable). From this point of view, scarf joints are better than halved-scarf joints. Self-tapping screws can also be used to strengthen the joint (Fig. 23a). The design equations used for notched beams can be used (see lap joints). Under tension only, reinforcement screws can be driven only in the overlapping area (Fig. 23d). This reinforcement can be checked using Johansen's equations assuming that the tensile load is completely carried by the screws.



In the case of the Trait-de-Jupiter it is common to add metal connectors passing through the joint depth to reinforce the joint (Fig. 24). Another solution with glued in rods is presented in Fig 24.

Conclusions

When working on old timber structures, the fact that the structure has stood for decades or centuries without failure may not be sufficient proof of the load bearing capacity for the future (new imposed loads etc.). Joints greatly influence the response of the whole structure. Their characterization (the strength, the stiffness and the ability to be reinforced) still remains a big challenge. The design of traditional joints essentially involves a check of the contact pressure between the assembled elements. Even if seemingly trivial, assessment of old carpentry joints still remains a difficult task. As an illustration, the slight difference in the definition of the compressive strength (which is of major importance) at an angle to the grain mentioned in different standards, underscore a basic point that has to be clarified anyway by further research and later on by the revision of current

73

Reinforcement of Timber Structures standards. Moreover, not only the strength, but also the stiffness of the joint has to be considered as it can influence the force distribution within the structure. If the decay of timber elements is too large, then replacement is clearly the only solution. If repairs are necessary, specific reliable on-site assessment techniques are required to determine the appropriate level of intervention needed. This point remains very important to evaluate the replacement, repair and retrofit solutions along with the associated project costs. It should be noted that there is still a noticeable lack of scientific results and design rules regarding the reinforcement of old carpentry joints. This clearly points out the lack of research in this field. Unfortunately, this lack of information in addition to difficulties in assessment and definition of grading protocols for old timber elements often lead to unnecessary replacements. Further studies in the area are deemed necessary in order to establish reliable design models, to set detailing rules and to provide recommendations for future rehabilitation or strengthening interventions, among others. Because of the wide variety of carpentry joint geometries in existence, studying them with an exhaustive approach is neither realistic nor useful. For an accurate study, a good understanding of how the joint works and how the loads are balanced is the key point. As seen in this review, some information about strength, stiffness and reinforcement of common joints exists; even though scientific data is still missing and complementary research is needed. To achieve competency, engineers need specific tools such as the ones defined for the design of dowelled joints. Hopefully, the most important outcomes of existing (and ongoing) research will be integrated into the revised version of Eurocode 5.

6.

References

[1] Descamps T., Léoskool L., Laplume D., Van Parys L., Aira J.R., “Sensitivity of timber

hyperstatic frames to the stiffness of step and ridge joints”, In: Proceedings of the 13th World Conference on Timber Engineering, Quebec, Canada, 2014. [2] Parisi M., Piazza M., “Mechanics of plain and retrofitted traditional timber connections”, Journal of Structural Engineering, Vol. 126, No.12, 2000, pp. 1395–1403. [3] Branco J.M., Piazza M., Cruz P.J.S., “Experimental evaluation of different strengthening techniques of traditional timber connections”, Engineering Structures, Vol. 33, No. 8, 2011, pp. 2259-2270, http://hdl.handle.net/1822/13592. [4] Descamps T., Noël J., “Semi-rigid analysis of old timber frames: definition of equivalent springs for joints modeling. Enhancement of the method, numerical and experimental validation”, International Review of Mechanical Engineering, Vol. 3, No. 2, 2009, pp. 230-239. [5] Gerner M., Les assemblages des ossatures et charpentes en bois, Group Eyrolles, Paris, 190 pages, 2012. ISBN: 978 -2 -212-13620 - 3 [6] Sobon J.A., Historic American timber Joinery, a graphic guide, Timber Framers Guild, , Becket, MA 01223, 56 pages, 2012. [7] Seike, k., The Art of Japanese Joinery, Weatherhill/Tankosha Publ., New York, 128 pages, 1977 [8] Meisel A., Moosbrugger T., Schickhofer G., “Survey and Realistic Modelling of Ancient Austrian Roof Structures”, In: Proceedings of Conservation of Heritage Structure (CSHM-3), Ottawa, Canada, 2010. [9] SIA 265: Constructions en Bois. Swiss Society of Engineers and Architects, 2003. [10] EN 1995-1:2005, Eurocode 5: Design of timber structures - Part 1-1: Common rules and rules for buildings. Brussels, CEN, European Committee for Standardization, 2005 [11] Hewett C.A., English Historic Carpentry. London & Chichester: Phillimore & Co. Ltd., 1980, p.270.

74

Analysis and strengthening of carpentry joints [12] Hirst E., Brett A., Thomson A. Walker P., Harris R., The Structural Performance of Traditional Oak Tension & Scarf Joints, In: Proceedings of the 10th World Conference on Timber Engineering, Miyazaki, Japan, 2008. [13] Thelandersson, S., Larsen, H. J. Timber Engineering. Chichester: John Wiley & Sons Ltd., 346 pages, 2003. [14] Larsen H., Jensen J., “Influence of semi-rigidity of joints on the behaviour of timber structures”, Progress in Structural Engineering and Materials, Vol. 2, No. 3, 2000, pp. 276–277. [15] Candelpergher L., Sperimentazione, modellazione numerica e caratterizzazione sintetica del comportamento di collegamenti lignei tradizionali con elementi metallici, Master’s thesis, Università degli Studi di Trento, 1999. [16] Palma P., Cruz H., “Mechanical behaviour of traditional timber carpentry joints in service conditions - results of monotonic tests”, In: Proceedings of “From material to Structure Mechanical behaviour and failures of the timber structures,” XVI International Symposium, ICOMOS IWC, 2007. [17] Uzielli L., Il manuale del Legno Strutturale, Vol IV - Interventi sulle strutture, Mancosu, Rome, Italy, in Italian, 2004. [18] Drdácký M., Wald F., Sokol, Z., “Sensitivity of historic timber structures to their joint response”. In: Proceedings of the 40th Anniversary Congress of the IASS, Madrid, 1999. [19] Descamps T., Lambion J., Laplume D. (2006), “Timber Structures: Rotational stiffness of carpentry joints”, In: Proceedings of the 9th World Conference on Timber Engineering, Portland, USA, 2006. [20] Komatsu K., Kitamori A., Jung K. and Mori T., “Estimation of the Mechanical Properties of Mud Shear Walls Subjecting to Lateral Shear Force”, In: Proceedings of the 11th Int. Conference on Non-conventional Materials and Technologies, Bath, UK, 2009. [21] Chang W.-S., Hsu M.-F., Komatsu K., “Rotational performance of traditional Nuki joints with gap I: theory and verification”. Journal of Wood Sciences, Vol. 52, 2006, pp. 58–62. [22] Wald F., Mares Z., Sokol M., Drdácký F, “Component Method for Historical Timber Joints”, In: The Paramount Role of Joints into the Reliable Response of Structures. NATO Science Series Vol. 4, 2000, pp. 417-424. [23] Kasal B., Tannert T., "In Situ Assessment of Structural Timber", In: RILEM State-of-the-Art Reports, Vol. 7, 2011, p. 129. [24] UNI 11138, Cultural heritage - Wooden artefacts - Building load bearing structures - Criteria for the preliminary evaluation, the design and the execution of works. UNI Milano, 2004. [25] Aman R., West H., Cormier D. "An evaluation of loose tenon joint strength", Forest Products Journal, Vol. 58, No. 3, 2008, pp. 61–64. [26] Judd J., Fonseca F., Walker C., Thorley P., "Tensile strength of varied-angle mortise and tenon connections in timber frames", Journal of Structural Engineering, Vol. 137, No. 5, 2012, pp. 636–644. [27] Likos E., Haviarova E., Eckelman C., Erdil Y., Ozcifci A., "Effect of tenon geometry, grain orientation, and shoulder on bending moment capacity and moment rotation characteristics of mortise and tenon joints", Wood Fiber Sciences, Vol. 44, No. 4, 2012, pp. 1–8. [28] Kock H., Eisenhut L., Seim W., "Multi-mode failure of form-fitting timber connections – Experimentaland numerical studies on the tapered tenon joint", Engineering Structures ,Vol. 48, 2013, pp, 727–738. [29] Feio A.O., Lourenço P.B., Machado J.S., "Testing and modeling of a traditional timber mortise and tenon joint", Materials and Structures, Vol. 47, 2014, pp. 213–225. [30] Götz K.-H., Hoor D., Möhler K., Natterer J., "Construire en Bois - Choisir, concevoir, realiser". Presses Polytechniques et Universitaires Romandes, Lausanne, Switzerland, 1993. [31] DIN 1052, Entwurf, Berechnung und Bemessung von Holzbauwerk. Allgemeine bemessungsregeln und bemessungsregeln fur den hochbau, 2004. [32] C.T.E., Documento Básico SEM. Seguridad estructural – Estructuras de madera. A código técnico de la edificación, ministerio de vivienda, 2006

75

Reinforcement of Timber Structures [33] Branco J, Cruz P, Varum H, Piazza M., "Portuguese traditional timber trusses. Static and dynamic behaviour". Technical Report E-19/05, Guimarães, Portugal. (in Portuguese), 2005. [34] Derinaldis P.P, Tampone G., "The Failure of the Timber Structures Caused by Incorrect DesignExecution of the Joints. Two Cases Study", In: ICOMOS IWC, XVI International symposium – Florence, Venice and Vicenza, 2007. [35] Branco J.M., Influence of the joints stiffness in the monotonic and cyclic behaviour of traditional timber trusses. Assessment of the efficacy of different strengthening techniques, PhD thesis, University of Minho and University of Trento, 2008. [36] DIN EN 1995-1:2005, NCI NA 6.8.3. National German Annex to Eurocode 5: Design of timber structures - Part 1-1: Common rules and rules for buildings.Brussels, CEN, European Committee for Standardization, 2005 [37] Yeomans D., The Repair of Historic Timber Structures, Thomas Telford Publishing, London pp. 147-150, 2003. [38] EN 1998-1:2004. Eurocode 8: Design of structures for earthquake resistance- Part 1: General rules. Brussels, CEN, European Committee for Standardization, 2004 [40] Tampone G., "Mechanical Failures of the Timber Structural Systems", In: ICOMOS IWC, XVI International Symposium – Florence, Venice and Vicenza, 2007.

76

Reinforcement of connections with dowel type fasteners

5 Reinforcement of connections with dowel-type fasteners Laurent Bléron1, Damien Lathuillière1, Thierry Descamps2, Jean-François Bocquet1

Summary A good design of connections with dowel type fasteners is essential to ensure a safe design for the whole structure and a cost effective solution. With reinforced connections, engineers can achieve better capacity of the connection and safer designs by increasing the ductility of the connection (so-called "capacity design" with a ductile connection to be the weak point). This paper presents an overview of the reinforcement of dowel type connections. Most of the failures encountered in dowel type connections occur because of an excess of shear stresses or tensile stresses perpendicular to the grain in the connection area. Because of the large amount of possible details and reinforcement techniques, this report mainly focuses on the reinforcement of dowel-type connections when the dowels are mainly loaded parallel to the grain. For such connections, modes of failure of un-reinforced and reinforced connections, updated design models and the effect of the reinforcement on the ductility of the connection will be presented. Among the various types of reinforcement techniques available, nowadays, self-tapping screws have found a wide field of application. They will be the main focus of this chapter.

1.

Introduction

In timber structures, the design of dowel type connections may govern the overall design of structural members. Because of the low load bearing capacity of one fastener (when considered alone) relative to the total load supported, designers have to design for numerous fasteners and deal with regulations relating to the minimum spacing requirements. Because of the weakness of timber in shear and in tension perpendicular to the grain, spacing requirements are mandatory to avoid any failure of connections at very low load rates (compared to the ultimate load of the global structure). Spacing requirements may lead to a global over sizing of the timber members and so, a non-effective solution. Designing large connections (with large spacing between the dowels) may result in uneconomical solutions and potentially lead to problems caused by shrinkage that is the result of variations in the timbers moisture content. In that situation, the use of some types of reinforcement is an opportunity to ensure a better design. Most of the failures encountered in dowel type connections occur because of an excess of shear stresses or tensile stresses perpendicular to the grain in the connection area. Because of the embedment of the fastener in the wood, very complex stress states appear in the wood along the contact surface (including shear stress and tension perpendicular to the grain). The relative importance of those stresses depends on the angle between load and grain and different failure modes may occur as presented in Fig. 1. 1)

University of Lorraine, ENSTIB-LERMAB, Epinal, France

2)

University of Mons, Mons, Belgium

77


Fig. 1 Failed embedment specimens tested at different loads to the grain angles: (a) dowel loaded parallel to the grain, (d) dowel loaded perpendicular to the grain [1] Various reinforcement techniques have been proposed among them, truss plates [2], [3], fibreglass [4], [5], threaded rods [6], glued in rods and self-tapping screws (STS) [7] or glued-on wood-based panels [8]. All of them mainly aim to avoid splitting of the timber in the connection area. Some examples are presented in Fig. 2.

Fig. 2 (a) reinforcement using glued-on boards (b) reinforcement by punched metal plate fasteners and (c) nail plates [8] (d) multiaxial stitch bonded fabric [9] Haller et al. [9] have used multiaxial stitch bonded and biaxial weft knitted textiles (Fig. d). This investigation concluded that even light textile reinforcements may significantly increase the strength, the stiffness and the ductility of doweled connections. It appeared from embedding tests that loop-like fibre placements achieved the highest embedding strength and stiffness, whereas biaxial knitted fabrics result in more ductile connection behaviour. However, special attention has to be paid to any solution involving the wrapping the timber connections in an airtight textile because of the risk of decay. Reinforcement helps to overcome timber weaknesses by increasing the shear strength and the tensile strength perpendicular to the grain and by stopping the propagation of cracks. This report mainly focuses on the reinforcement of dowel-type connections when the dowels are loaded parallel to the grain (Fig. 1a). For such connections, modes of failure of unreinforced and reinforced connections, proposed design models and the effect of the reinforcement on the ductility of the connection will be discussed. Situations where dowels are loaded perpendicular to the grain are not developed to a large extent here after. One may notice that in that situation reinforcements (screws, glued in rods or a surface reinforcement as a plate) prevent crack propagation (Fig. 3). This kind of reinforcement has been studied by several authors, among others, Franke et al. [10], Borth et al. [11], Schoenmarkers et al.

78

Reinforcement of connections with dowel type fasteners [12], Jensen et al. [13] & Leijten [14]. Schoenmakers et al. [15] have studied the reinforcement by screws and tested several configurations (different relative heights h and loaded edge distances he) as presented in Fig. 3. Blass et al. proposed a design model that assumes that timber has no tensile strength perpendicular to the grain [16].

Fig. 3 Reinforcement with single plate (a) and one main plate and stitch plates (b) to prevent cracks. Reinforcement with screws (c) Initially developed for assembling elements (member or connections), STS have found a new field of application with the reinforcement of timber beams or timber connections. The benefits of using STS are generally their efficiency (as reinforcement), their low costs (cheaper than other reinforcement materials) and the fact that they are easy to use (the setup is less complex than for glued in rods for example). Moreover, the use of STS does not require any surface preparation or pre-drilling (for the smaller diameters) and new screw-guns enable fast and effective installation of STS. Reinforced with STS, connections may behave plastically until failure (because any splitting of wood is prevented). STS, contrary to traditional wood screws, are fully threaded and made with high tensile strength steel. They can bear high axial loads (even with a small diameter, reducing the danger of the splitting of wood) and have high withdrawal strength due to their improved thread geometry. Due to the rapid growth in the use of STS for reinforcements, they will be the main focus of this chapter.

2.

Reinforcement of dowel-type connection with self-tapping screws

2.1

Modes of failure of an unreinforced dowel-type connection

The different failure modes of a connection made with dowel-type fasteners are presented in Fig. 4. Five failure modes are indicated depending on the connection geometry: edge and end distances of the fasteners, spacing, the fastener diameter (in relation to the thickness of timber) and the number of fasteners. The failure mode (a) is ductile because it is characterized by the embedment of the fasteners into the wood, the others are not [8], [17], [18]. Timber may show a tendency to split in the connection area before the embedding strength is reached according to timber thickness, diameter and number of dowels, load to the grain angle, spacing as well as the end and edge distances of the dowels.

2.2

Self-tapping screws

STS with the thread running along the full length (Fig. 5i) and screwed perpendicular to the grain increase the local strength in shear and tension perpendicular to the grain so they counteract the tendency of timber to split. This reinforcement is achieved by ensuring that the tensile stresses

79

Reinforcement of Timber Structures perpendicular to the grain (or shear stresses) are mainly supported by the screws and not by the timber (because the screws embedded into wood are much stiffer than wood) [19]. By protecting against any premature brittle failure with this technique, the connection failure modes coincide with the ductile failure modes described by Johansen’s theory (no need for a reduced effective number of fasteners) [20], [21].

Fig. 4 Different failures modes of unreinforced dowel-type connections

2.3

Self-tapping screws

STS with the thread running along the full length (Fig. 5i) and screwed perpendicular to the grain increase the local strength in shear and tension perpendicular to the grain so they counteract the tendency of timber to split. This reinforcement is achieved by ensuring that the tensile stresses perpendicular to the grain (or shear stresses) are mainly supported by the screws and not by the timber (because the screws embedded into wood are much stiffer than wood) [19]. By protecting against any premature brittle failure with this technique, the connection failure modes coincide with the ductile failure modes described by Johansen’s theory (no need for a reduced effective number of fasteners) [20], [21]. STS may also increase the load-carrying capacity of one fastener beyond the values defined by Johansen’s yield theory by increasing the embedding strength. For that purpose, STS screwed perpendicular into the grain and perpendicular to the fasteners (to be reinforced) are the best. The tendency of timber to split is minimised by using STS and as a consequence the load carrying capacity will be increased (Fig. 5ii and 5iii).

Fig. 5 (i) full threaded self-tapping screws (STS), (ii) reinforcement of timber with STS reducing the wood's tendency to split (iii) reinforcement of timber and fastener embedding strength by placing the self-tapping screws in contact with the dowel-type fasteners

80

Reinforcement of connections with dowel type fasteners Because of the high tensile strength of steel STS and their special thread, the connection behaviour is no longer governed by the low tension perpendicular to the grain and shear strength of timber. Bejtka et al. & Blass et al. first studied this topic [19], [16]. Timber splitting is prevented when the axial load-carrying capacity of each screw is larger than 30% of the lateral load-carrying capacity per shear plane of each dowel (calculated according to Johansen´s yield theory). The screws of course prevent the splitting of timber, but also increase the embedding strength of the fasteners. Some experimental results are presented in Fig. 6. Two important points can be observed, firstly the increase of the load bearing capacity and secondly the important ductility of the reinforced connection.

Fig. 6 Typical load-displacement-curves of non-reinforced and reinforced connections [8] (a) non-reinforced – (b) reinforced with screws without contact with the dowel-type fasteners – (c) reinforced with screws with contact with the dowel-type fasteners

Fig. 7 (a) example of bolted connections with a slotted in steel plate, (b) example of reinforced connection layout with reduced spacing, (c) full size sample test [24] Bejtka et al. [19] studied the influence of the position of the reinforcement in relation to the fasteners that have to be reinforced. Echavarria [22] has studied the efficiency of STS in stopping the spread of the cracks (Fig. 5ii) and the related increase in load carrying capacity of the fastener. This study has been made for a single bolt assembly. The main benefit of this method is the reduced end-distance without any reduction of the load-carrying capacity of the connection, which leads to a more compact connection. Tab. 1 shows the ratio between the reinforced and non-reinforced load-carrying capacity based on tests performed by Echavarria for different geometrical configurations. One can notice that

81

Reinforcement of Timber Structures the effect of the reinforcement is maximal for small e/d (small loaded end distance/dowel diameter – see Fig. 5). Bocquet et al. [26] have studied block shear failures (Fig. 4d) of dowelled connections in bending and have proposed design models. Different connections with rectangular and circular patterns have been studied. Regarding the block shear failure, the tests conducted showed that an optimal number of reinforcements and an appropriate positioning of STS are required in order to achieve excellent results (efficient reinforcement). On the one hand, a minimum amount of reinforcement uniformly distributed over the connection area is necessary in order to prevent splitting. On the other hand, too many reinforcements can affect the mechanical behaviour of the connection by weakening the timber beams (reduced cross section). This is especially true if the reinforcements are located where the stresses in timber are the highest. Tab. 1 Ratio of reinforced and non-reinforced load-carrying capacity [14] Bolt diameter

Thickness e/d

[mm]

[mm]

Reinforced loadcarrying capacity [kN]

Non-reinforced load-carrying capacity [kN]

Ratio of reinforced and non-reinforced loadcarrying capacity [%]

15.9

2

38

9.8

4.4

120%

15.9

3

38

12.7

12.1

4.7%

15.9

4

38

17.4

17.3

0.8%

15.9

5

38

18.4

18.2

1.0%

Gehloff et al. [24] enabled the use of STS both to reduce the edge distance of bolts and to increase the bending capacity of a moment resisting timber connection (Fig. 7). It was observed that STS prevented early splitting of the members even though small edge distances of bolt were applied (even for stocky dowels). An important decrease of the edge distance (up to a half) in relation to the reference connection (Fig. 7a) resulted in an increase in ultimate moment capacity by a factor of about 1.3. A significant increase in connection capacity could not be achieved when placing the STS near the bolts. One may also notice that based on cyclic loadings, placing the STS close to the bolts reduce the standard deviation and might have a positive impact on the energy dissipation.

2.4

Proposed design models

Bejtka et al. [19] have developed a calculation model based on Johansen’s yield theory. The STS are placed in contact with the dowels, perpendicular to the dowel axis and to the grain direction (Fig.iii). They are loaded just as the dowels themselves perpendicular to their axis. As for Johansen´s yield theory, all components (STS, timber and fasteners) have ideal elastic-plastic behaviour. The STS buckle perpendicular to their axes and in the force direction, when the dowel load component FVE reaches the load-carrying capacity RVE of the STS. In this case, the screw behaves as a “soft” support (Fig. 8).

82

Reinforcement of connections with dowel type fasteners

Fig. 8 (a) one dowel reinforced by one STS; (b) two dowels reinforced by one STS Alternatively, for FVE < RVE, the screw does not move and represents a rigid support for the dowel. This consideration leads to four sub-failure modes for each failure mode in timber-to-timber connections and two sub-failure modes for each failure mode in steel-to-timber connections in Johansen´s yield theory. As an example, Fig. 9a and Fig. 9b represent, for steel to timber connections, the sub-failure modes of failure mode 3 (inner steel plate with two plastic hinges). The efficiency of the reinforcement depends on its position compared to the position of the plastic hinge. According to the distance p of the STS relative to the dowel and the distance x between the shear plane and the plastic hinge in the unreinforced connection (see Fig. 9), the STS does not reinforce the connection (if p>x) or on the contrary the reinforcement increases the resistance of the connection (if p 0.8 for applications in Service Class 3, as defined in Eurocode 5.

3.2

Adhesive strength of bonded rods

Pull-pull or pull-compression configurations are commonly used to evaluate the strength of glued-in rods in timber parallel to grain, whereas pull-beam or pull-pile foundation configurations can be additionally used with rods inserted perpendicular to grain [39], [40]. Moreover, while some authors have preferred to use a one-sided test set-up [22], [41], others have chosen the two-sided set-up [42]– [44]. Jensen et al. [45] have accurately described the mechanics of the two test set-ups. However, to date there is no agreement among structural engineers about the correct approach for the design of glued-in connections, as also evidenced by the lack of a standard procedure adopted at European level for this kind of joint. However, although non-standard, the configurations mentioned above have provided evidence that the various design factors that affect the bond performance include anchorage length, rod diameter, bondline thickness, rod-to-grain angle, edge distance, adhesive stiffness, and wood density. In addition, it is relatively easy to provide a single direct pull out test configuration for simplified on-site testing and proof loading. The testing procedure reported here is biased towards the way the load can be applied on site to gluedin assemblies and it is intended as a method for assessing the overall performance of a joint within a specific bonding process used in a workshop. Within this context, the expression ‘bonding process’ refers to all the elements of the process, application procedures and designer’s specifications. Hence the evaluated performances are more related to verifying the correct application of the product, according to the designer’s or the manufacturer’s instructions, in an on-site bonding situation. This implies that tests according to the reported procedure can also be carried out directly on site. Bonded wood specimens with the internally inserted rods are prepared similarly to the test configurations mentioned above. This results in the process of rods having to be fixed into drilled holes where the diameter of the hole exceeds the diameter of the rod, and where the annulus is filled with the adhesive. Rods are centred into the hole by means of ancillary small devices, such as collars at the top and bottom of the bonded rod length. In consideration of the fact that the described procedure is also intended for testing specimens directly on site, they can be prepared as separate and individual samples to be tested later, or as an integral part of the structure. In this way the prepared samples are truly representative of the work to be carried out. The rods must protrude from the bonded area by a suitable length (in the order of 240 mm) in order to allow a hollow centre cylindrical hydraulic ram to fit over the extended anchor and receive a suitable collet and collet rod gripping holder (Fig. 12).

123


a)

b)

Fig. 12 (a) General view of the equipment for on-site tensile pull out testing; (b) typical assembly for the hollow ram rod and collet gripping device for one possible rod as described within the text.

c)

(c) Easily assembled lightweight equipment for on-site testing of specimen anchorages.

Fig. 13 Test arrangement (in the small picture at the right-top corner) and output of the test as Temperature vs. time curve for three commercial adhesives and the ambient climate Following the minimum recommended cure time for the specific adhesive the load is applied using the hollow ram assembly described above and the use of a suitable hydraulic single action pump with certificated gauge (Fig. 12). This is a relatively easy-to-perform on-site test, provided that the following procedures are followed: a) ensure that the hydraulic ram lower face is set at a 90° angle to

124

Adhesives for on-site bonding: characteristics, testing and prospects the bonded rod (normally a levelling bridge is used in order to avoid any distortion of the adhesively bonded rod during the loading procedure); b) ensure that the load is applied evenly to minimise any uneven or excessive speed during the incremental loading. In this respect, a valid procedure would be that of loading by increments of 20% of the expected strength of the joint, and waiting a pre-defined time between individual increments, in order to assess if there is any loss at each and every individual increase of load. This is usually observed by checking any movements of the gauge between increases of load. Pull-out values below the expected minimum loads may indicate that an error has (or errors have) occurred during the original bonding process, or that the anticipated pre-calculated load was excessive.

3.3

Verification of the on-site mixing

Structural adhesives used for timber structures are bi-component products which have to be mixed immediately prior to their application. Mixing procedures are relatively easy to carry out. In practice, the total contents of a can, usually Part B, should be poured into the container which holds Part A. The two components should be thoroughly mixed with a suitable flat bladed spatula ensuring that the mix is homogeneous and does not entrap excessive amounts of air. This procedure requires care and good practice to ensure a thorough mix. However, it is well known that mixing errors sometimes occur, even with the most experienced operatives. It is also known that the mechanical properties of the mixed epoxy systems, including properties of bonded wood joints, are at a maximum when the correct (stoichiometric) ratio between the two components is produced. Nevertheless these properties may decrease appreciably if: a) one of the two components (A or B) is lacking; or b) when the mixing of the two components is incomplete. The latter is evidenced by incomplete curing in some zones of the retained sample of the mixed batch. A very simple and effective method to verify the mixing on site of individual batches of a mixed adhesive is by filling one or more small cylindrical standard size pots (approximately 20 cc in volume) (Fig. 13). They are then individually placed in an insulated box to ensure approximate adiabatic conditions during the initial cure phase. The pots are left to cure in pre-determined temperature conditions (for instance, the same temperature as that in the workshop and/or work site), while a thermocouple placed inside the pot measures and records the exothermic temperature rise during the curing reaction. The curing adhesive should be left in situ during which time the temperature will dissipate. When ambient temperature is reached the cylinders of cured adhesive may be removed and will be ready for compressive testing. The cylinders need to dwell for a certain time until a sufficient compressive strength is reached. They may then be compressively tested in a laboratory. Both the cure schedule and the compressive strength can be compared to the corresponding values declared by the adhesive producer, thus verifying the correct mixing of the product during use. It is worth noting that while compressive strengths are commonly found in the manufacturer’s technical data sheets, which provide a time/temperature comparison against various compressive strengths, they do not normally include the cure schedule. Therefore, this information has to be specifically sought from the adhesive manufacturer. It is also worth stating here that the cure schedule of a product depends not only on its specific adhesive composition but also on the ambient temperature at which the composition is mixed, the initial temperature of the two components of the mixture prior to mixing, and finally on the volume of an individual mix.

125


4.

Effects related to long-term duration of loads

Several investigations have shown that rods glued in timber (in this present section, glued-in rods are considered as representative of the wood assemblies prepared on site) and subjected to long-term duration of load may exhibit a decrease of mechanical performance over a long period of time [1], [15], [16]. In these cases, experimental data usually conform to the damage accumulation modelling for timber, described by the Madison curve [46]. In fact, if the bonded joint is sustained during the entire service life and the ambient temperatures do not substantially exceed the usual temperatures in indoor conditions (normally 20°C) the long-term behaviour of those bonded joints is similar to the creep behaviour of wood3. This was observed for epoxies in [16] for a constant very humid climate of 85% RH at 20°C, where the strength decrease for the epoxy-bonded assemblies could be described roughly by the usual damage behaviour of timber (conversely, they also observed a strength reduction for the polyurethane-bonded joints). In Eurocode 5 this strength reduction is taken into account by means of the modification factor, kmod, used to derive the design values. The values of k mod approximately conform to the behaviour of the stress level versus the logarithm of time to failure for timber, according to the linear Madison curve [47]. These values are defined through consideration of load duration classes: permanent (e.g. self-weight), long-term, medium-term, short-term (e.g. snow and wind), and instantaneous (e.g. accidental load). kmod values are not considered dependent on the adhesive type, and they are the same for the climatic Service Classes 1 and 2, which are, in the case of timber elements, the usually accepted classes for bonding on site. However, several tests also showed that the mechanical performance of bonded-in rods may depend on both the temperature and the environmental humidity in which the same bonded joints are kept. Assemblies exposed to cyclically varying climates in terms of temperature and relative humidity revealed that ambient temperatures above approximately 50°C caused significant short- and long-term strength losses in joints bonded with epoxies, whereas a temperature above 40°C and humid environments both considerably reduced the strength in joints bonded with polyurethanes [16]. All behaviours described so far are due to a concurrence of effects: a) Timber is subjected to creep, whose entity and behaviour strongly depends on both the environmental humidity and the loading history (this aspect is extensively covered in wood-related literature and requires no further discussion at this juncture) [48]. b) Visco-elastic adhesives such as epoxies and polyurethanes also suffer from creep: its level of importance is related to how close the service temperature is to the T g of those products (the closer it is, the larger the creep effect), to the humidity level (humidity can potentially act as a swelling agent), and to the adhesive cross-link density. The resulting progressive deformation will continue until rupture or yielding. According to ASTM D2990-01 [49], a creep displacement versus time curve can be considered as being in three stages: (i) primary creep, (ii) secondary creep and (iii) tertiary creep. Fig. 14 shows the creep displacement for two commercial products tested: a) well above its Tg (Product a), and b) just below it (Product b): the three stages of creep are well evident in both products (a) and (b), although to a much different extent (it is worth noting that the whole duration of the experiment was limited, as both Products were tested at a temperature close to their Tg).

3

Creep is usually defined as the progressive deformation that visco-elastic solids evidence under constant load.

126

Adhesives for on-site bonding: characteristics, testing and prospects Fig. 14 Creep displacement versus test duration time curve at a 20% of the adhesive strength loads for two commercial products

In the primary creep stage, the creep rate starts at a relatively high value, but decreases rapidly with time, which may be due to the slippage and reorientation of polymer chains under persistent stress. After a certain period, the creep rate reaches a steady-state value in the secondary creep stage, in which the normal duration is relatively long. Finally, the material transcends into the tertiary creep stage, where the creep rate increases rapidly and final creep failure occurs. However this general behaviour is only shown when an adhesive is under or close to its glass transition temperature. The converse is that when substantially above its Tg, a polymer hardly shows any primary and secondary creep, but the tertiary creep is very obvious. c) The occurrence of delamination or cracks taking place at the adherend interfaces may appreciably decrease the bondline duration, mainly under appreciable levels of load. The extent of this reduction may be very high: as previously shown [1] specimens subjected to accelerated ageing, constituted by realistic temperature and relative humidity cyclic variations, showed failures after only 9 months of weathering. The tension load in those specimens was 20% of their average pull-out strength (this corresponds to two or three times the actual in-practice working load values). However, cracks and fissures in both the timber and the adhesive were observed after weathering: as also recognised by the same authors, these occurrences were due to the very small ratio of the adhesive anchorage length to rod diameter (both anchorage length and diameter measured 10 mm) and the fact that the bond-line was not concealed inside the timber element. From the adhesive standpoint, it is possible to increase the T g of room-temperature-curing products by partially reformulating a suitable and targeted product component formulation. It was shown in [50] that the addition of liquid rubber as nano-fillers to the basis of a commercial adhesive induced a strain recovery under constant load, due to the restriction in the slippage, reorientation and motion of polymer chains, thus increasing the Tg of the modified product (from 32°C to 43°C). As a consequence, the addition of these nano-rubber particles reduced the initial creep rate in the primary creep stage, and a much more stable steady-state creep rate within the second creep stage was also observed in the pure adhesive subjected to creep4.

4

It is of note that the creep behaviour obtained from experiments using adhesive-timber joints is better (more precautionary) than that obtained from the adhesives alone under the same conditions.

127


5.

Future challenges

On-going research is required to a) increase the current performance of existing adhesives, which are increasingly used at their limits, and b) to fulfil the expectations and requirements of a growing number of end users. Those issues which seem to be of particular interest and importance are: - Temperature-related issues: The majority of current commercial 2 part (also referred to as 2K) epoxy adhesives have a Glass Transition Temperature, Tg, of 50-70°C or less. These values are very close to those attainable in summer time within roof structures in the countries of Southern Europe. A class of products able to cure at room temperature but characterised by higher values of Tg would be an interesting field of research activity. Some attempts to overcome this limitation by using nanofillers have already been introduced, but much more effort is required to make this improvement widespread and easier to manage. In principle the challenge is to improve the glass transition temperature whilst still controlling the modulus of elasticity of the cured adhesives. However, there are on-going observed experiences of in-service temperature conditions substantially exceeding the theoretical and measured Tg of the installed adhesive. To date, ongoing observations on these on-site moment resisting joints do not indicate any sign of structural demise. - Plasticity/rigidity of adhesives as related to long-term effects of loading: Products to be used on site need to have achieved a balance between a plastic and a rigid behaviour. In effect, a flexible elastic or ductile plastic behaviour is necessary to follow the dimensional variations undergone by wood following changes in environmental humidity. At the same time, a stable behaviour allowing an acceptable resistance to loads applied over a long period is also required. On the other hand, in most of the wood adhesives commercially available to date, appreciable creep behaviour is usually also associated with ductility. The current solution is to use additional fillers such as micro/nanofibers and micro/nanoplates, or other additives which are able to substantially alter this general behaviour. Any of the current ductile products, although more able to follow the wood deformations, are also responsible for the reduced performances in tests with long-term duration of load. While Section 4 above has shown how nanotechnologies can be successfully used to overcome this inconvenience, unfortunately only a few products fulfilling this requirement are commercially available. Further research and product development is required to make these characteristic improvements possible. - Smartness: Considering that maintenance costs are increasing, an ideal series of products to be used on site should be ‘active’, in the sense that they should change one of their easy-to-detect properties (e.g. aesthetic) when they are no longer active, for instance owing to delamination. As an alternative, or eventually as an additional requirement, they should be self-repairing, e.g., able to self-repair limited delamination (self-healing effect). Analogously, multifunctional polymers, which are optimised for performance in a range of application areas (surfactants, coatings, dispersants etc.), can be considered as an advanced challenge to face with regard to wood adhesives. In effect this could even include direct combination with integrated sensors to be used for the active monitoring of timber structures. - To be monocomponent: In practice, a monocomponent adhesive is easier to manage in workshops than a bi-component one. However, in the case of thick, or very thick, bondlines it is difficult for the interface reaction between the timber and the adhesive to continue reacting into the thick portion of the glue line (similar to that involving the OH groups in wood) to extend towards the bulk of the resin,

128

Adhesives for on-site bonding: characteristics, testing and prospects mainly when inert fillers are used. A potential solution could be the use of a specific type of latent catalyst or catalysts. - Acceptance criteria for products to be used on site: Although this aspect does not require a research effort, acceptance criteria as a standard is nevertheless important for adhesive manufactures, designers, and end-users. A good basis for establishing the requirements to be fulfilled is described previously in Section 3. To date, an attempt to establish the requirements of this class of product is under discussion at CEN meetings. This acceptance criteria process could include an evaluation of bonding qualities of the products in existence within existing structures.

6.

References

[1]

Custódio, J. Broughton, J., Cruz, H., “Rehabilitation of timber structures: novel test method to assess the durability of bonded-in rod connections,” Materials and Structures, Vol. 45, No. 1–2, 2012, pp. 199–221. Radovic, R., Goth, H., “Entwicklung und Stand eines Verfahrens zur Sanierung von Fugen im Brettschichtholz,” Bauen mit Holz, Vol. 9, 1992. EN 302-1, Adhesives for load-bearing timber structures - Test methods - Part 1: Determination of longitudinal tensile shear strength. Brussels: European Committee for Standardization CEN, 2013. EN 302-2, Adhesives for load-bearing timber structures - Test methods - Part 2: Determination of resistance to delamination. Brussels: European Committee for Standardization CEN, 2013. EN 302-3, Adhesives for load-bearing timber structures - Test methods - Part 3: Determination of the effect of acid damage to wood fibres by temperature and humidity cycling on the transverse tensile strength. Brussels: European Committee for Standardization CEN, 2013. EN 302-4, Adhesives for load-bearing timber structures - Test methods - Part 4: Determination of the effects of wood shrinkage on the shear strength. Brussels: European Committee for Standardization CEN, 2013. EN 302-5, Adhesives for load-bearing structures - Test methods - Part 5: Determination of maximum assembly time under referenced conditions. Brussels: European Committee for Standardization CEN, 2013. EN 302-6, Adhesives for load-bearing timber structures - Test methods - Part 6: Determination of the minimum pressing time under referenced conditions. Brussels: European Committee for Standardization CEN, 2013. EN 302-7, Adhesives for load-bearing timber structures - Test methods - Part 7: Determination of the working life under referenced conditions. Brussels: European Committee for Standardization CEN, 2013. EN 301, Adhesives, phenolic and aminoplastic, for load-bearing timber structures Classification and performance requirements. Brussels: European Committee for Standardization CEN, 2013. EN 1995-1-1, Eurocode 5: Design of timber structures - Part 1-1: General - Common rules and rules for buildings. Brussels: European Committee for Standardization CEN, 2008. Broughton, J.G., Hutchinson, A.R., “Pull-out behaviour of steel rods bonded into timber,” Materials and Structures, Vol. 34, No. 2, 2001, pp. 100–109. Kemmsies, M., Comparison of Pull-out Strengths of 12 Adhesives for Glued-in Rods for Timber Structures, SP Swedish National Testing and Research Institut, Boras, Sweden, SP Report 1999:20, 1999. E. Serrano, “Glued-in rods for timber structures. An experimental study of softening behaviour”, Materials and Structures, Vol. 34, No. 4, 2001, pp. 228–234.

[2] [3]

[4] [5]

[6]

[7]

[8]

[9]

[10]

[11] [12] [13]

[14]

129

Reinforcement of Timber Structures [15] Bainbridge, R., Mettem, C., Harvey, K., Ansell, M., “Bonded-in rod connections for timber structures. Development of design methods and test observations,” Int. Journal of Adhesion & Adhesives, Vol. 22, No. 1, 2002, pp. 47–59. [16] Aicher, S., Dill-Langer, G., “Influence of moisture, temperature and load duration on performance of glued-in rods,” In: PRO 22: International RILEM Symposium on Joints in Timber Structures, Stuttgart: RILEM Publications, 2001, pp. 383–392. [17] Raftery, G.M., Harte, A.M., Rodd, P.D., “Bond quality at the FRP–wood interface using woodlaminating adhesives”, Int. Journal of Adhesion and Adhesives, Vol. 29, No. 2, 2009, pp. 101– 110. [18] Broughton, J., Hutchinson, A., “Effect of timber moisture content on bonded-in rods,” Construction and Building Materials, Vol. 15, No. 1, 2001, pp. 17–25. [19] Sterley, M., Gustafsson, P.J., “Shear fracture characterization of green-glued polyurethane wood adhesive bonds at various moisture and gluing conditions,” Wood Material Science & Engineering, Vol. 7, No. 2, 2012, pp. 93–100. [20] Pizzo, B., Compatibilità, durabilità e reversibilità nel restauro delle strutture lignee. Diagnosi del degrado, tecniche e materiali per il consolidamento, Ph.D. Thesis, University of Palermo, 1999. [21] B. Pizzo and P. Lavisci, “Un approccio alla valutazione quantitativa della compatibilità nel restauro strutturale del legno,” in Atti del Convegno “Dalla Reversibilità alla Compatibilità”, Conegliano, 13-14 giugno, Firenze: Nardini, 2003, pp. 77–87. [22] Feligioni, L., Lavisci, P., Duchanois, G., De Ciechi, M., Spinelli, P., “Influence of glue rheology and joint thickness on the strength of bonded-in rods,” Holz Roh Werkstoff, Vol. 61, No. 4, 2003, pp. 281–287. [23] Pizzi, A., “Resorcinol adhesives” (chap. 29), In: Pizzi, A. and Mittal, K.L. (Eds.), Handbook of Adhesive Technology (2nd Ed.), New York: Marcel Dekker, 2003. [24] Chui, Y.H., Ni, C., “Stress distributions in glued wood lap joints subjected to an axial force”, In: Proceedings 5th World Conference on Timber Engineering, Lausanne, 1998. [25] Lavisci, P., Berti, S., Pizzo, B., Triboulot, P., Zanuttini, R., “A shear test for structural adhesives used in the consolidation of old timber,” Holz Roh Werkstoff, Vol. 59, No. 1–2, 2001, pp. 145– 152. [26] Pizzo, B., Lavisci, P., Misani, C., Triboulot, P., Macchioni, N., “Measuring the shear strength ratio of glued joints within the same specimen,” Holz Roh Werkstoff, Vol. 61, No. 4, 2003, pp. 273–280. [27] ASTM D3931-08, Test Method for Determining Strength of Gap-Filling Adhesive Bonds in Shear by Compression Loading. West Conshohocken, PA: ASTM International, 2008. [28] ASTM D2559-12a, Specification for Adhesives for Bonded Structural Wood Products for Use Under Exterior Exposure Conditions. West Conshohocken, PA: ASTM International, 2012. [29] Strickler, M.D., “Adhesive durability: specimen designs for accelerated tests,” Forest Products Journal, Vol. 14, No. 1, 1968, pp. 84–90. [30] Serrano, E., “A numerical study of the shear-strength-predicting capabilities of test specimens for wood–adhesive bonds,” Int. Journal of Adhesion & Adhesives, Vol. 24, No. 1, 2004, pp. 23– 35. [31] Broker, F.V., Kuhl, J., “Untersuchungen an zellolosefasergefüllten Epoxidharzen zur Sanierung breiter Risse in Bauholz,” Bauen mit Holz, Vol. 93, no. 9, 1991. [32] Okkonen, E.A., River, B.H., “Factors affecting the strength of block-shear specimens,” Forest Products Journal, Vol. 39, no. 1, 1989, pp. 43–50. [33] Ceccotti, A., Mannucci, M., Uzielli, L. “Effetti del riassorbimento di umidità sul comportamento ad estrazione di barre di acciaio ancorate nel legno mediante resina epossidica,” In: Proc. 2nd Congresso Nazionale sul Restauro del Legno, Firenze, Firenze, 1990. [34] Raknes, E., “Durability of structural wood adhesives after 30 years ageing,” Holz Roh Werkstoff, Vol. 55, No. 2–4, 197, pp. 83–90.

130

Adhesives for on-site bonding: characteristics, testing and prospects [35] River, B.H., Ebewele, R.O., Myers, G.E., “Failure mechanisms in wood joints bonded with urea-formaldehyde adhesives,” Holz Roh Werkstoff, Vol. 52, No. 3, 1994, pp. 179–184. [36] Pizzo, B., Lavisci, P., Misani, C., Triboulot, P., “The compatibility of structural adhesives with wood,” Holz Roh Werkstoff, Vol. 61, No. 4, 2003, pp. 288–290. [37] Green, D.W., Winandy, J.E., Kretschmann, D.E., “Mechanical properties of wood,” in Wood Handbook. Wood as an Engineering Material, Madison, WI: Department of Agriculture, Forest Service, Forest Products Laboratory, 1999. [38] Pizzo, B., Macchioni, N., Lavisci, P., De Ciechi, M., “On-site consolidation systems of old timber structures,” In: Interaction between science, technology and architecture in timber construction, C. Bertolini Cestari, T. Marzi, E. Seip, and P. Touliatos, Eds. Paris: Elsevier, Heritage series, 2004, pp. 323–352. [39] Tlustochowicz, G., Serrano, E., Steiger, R., “State-of-the-art review on timber connections with glued-in steel rods,” Materials and Structures, Vol. 44, No. 5, 2011, pp. 997–1020. [40] Serrano, E., Steiger, R., Lavisci, P., “Glued-in rods,” in Core Document of the COST Action E34. Bonding of Timber, M. Dunky, B. Källander, M. Properzi, K. Richter, and M. Van Leemput, Eds. Vienna: Univerität für Bodenkultur, 2008, pp. 31–39. [41] Del Senno, M., Piazza, M., Tomasi, R., “Axial glued-in steel timber joints. Experimental and numerical analysis,” Holz Roh Werkstoff, Vol. 62, No. 2, 2004, pp. 137–146. [42] Steiger, R., Gehri, E., Widmann, R., “Pull-out strength of axially loaded steel rods bonded in glulam parallel to the grain,” Mater. Struct., vol. 40, no. 1, pp. 69–78, 2007. [43] Chans, D.O., Cimadevila, J.E., Gutiérrez, E.M., “Influence of the geometric and material characteristics on the strength of glued joints made in chestnut timber,” Materials & Design, Vol. 30, No. 4, 2009, pp. 1325–1332. [44] Serrano, E., “Glued-in rods for timber structures. A 3D model and finite element parameter studies,” Int. Journal of Adhesion & Adhesives, Vol. 21, No. 2, 2001, pp. 115–127. [45] Jensen, J.L., Koizumi, A., Sasaki, T., Tamura, Y., Iijima, Y., “Axially loaded glued-in hardwood dowels,” Wood Science & Technology, Vol. 35, No. 1–2, 2001, pp. 73–83. [46] Hoffmeyer P., Sørensen, J.D., “Duration of load revisited,” Wood Science and Technology, Vol. 41, No. 8, 2007, pp. 687–711. [47] Morlier, P., Ranta-Maunus, A., “DOL effect of different sized timber beams,” Holz Roh Werkstoff, Vol. 56, No. 5, 1998, pp. 279–284. [48] Morlier, P., Creep in Timber Structures. London: Chapman & Hall (Spon), 1994. [49] ASTM D2990-09, Test Methods for Tensile, Compressive, and Flexural Creep and CreepRupture of Plastics. West Conshohocken, PA: ASTM International, 2009. [50] Ahmad, Z., Ansell, M.P., Smedley, D., Md Tahir, P., “The Effect of Long Term Loading on Epoxy-Based Adhesive Reinforced with Nano-Particles for In Situ Timber Bonding,” Advanced Materials Research, Vol. 545, 2012, pp. 111–118.

131


132

Reinforcement with glued-in rods

8 Reinforcement with glued-in rods René Steiger1, Erik Serrano2, Mislav Stepinac3, Vlatka Rajčić4, Caoimhe O’Neill5, Daniel McPolin6, Robert Widmann7

Summary Glued-in rods (GiR) have been successfully used for both constructing new and strengthening existing timber structures. The research and development of connecting and strengthening timber structural elements with GiR has been going on since the 1980s. However, agreement regarding design criteria for these applications has not been reached. Today, some few technical approvals for specific adhesives suitable to GiR exist, but an approach for the design of connections or reinforcement with GiR has not been included in the European design code EN 1995 so far. Therefore, it is desired to gather the current state of knowledge to enable application in practice of the existing and documented knowledge and experience. This state-of-the-art review (STAR) summarises results from research done regarding connections and reinforcement with GiR. The review considers manufacturing methods, mechanisms and parameters governing the performance and strength of GiR, theoretical approaches and existing design recommendations. For GiR applied as reinforcement similar rules and requirements apply as for GiR being used as connectors.

1.

Introduction

Glued-in rods (GiR) are an effective way of producing stiff, high-capacity connections in timber structures. In addition GiR have been successfully used for almost 30 years for in-situ repair and strengthening of structures, as well as for new construction works. GiR are used for column foundations, moment-resisting connections in beams and frame corners, as shear connectors and for strengthening structural elements when extensively loaded perpendicular to grain and in shear. Early examples of their use also include the connection of windmill blades made from glued laminated timber (glulam) [1, 2].

1)

PhD, Senior Scientist, Empa, Dübendorf, Switzerland

2)

Professor, PhD, Linnæus University, Växjö, Sweden

3)

PhD Candidate, University of Zagreb, Croatia

4)

Professor, PhD, University of Zagreb, Croatia

5)

PhD Candidate, Queen’s University, Belfast, UK

6)

Lecturer, Queen’s University, Belfast, UK

7)

Researcher, Empa, Dübendorf, Switzerland

133

Reinforcement of Timber Structures Most applications have used the GiR connections/reinforcement with metal bars glued into softwood. In practice, glulam made from softwood in combination with rods with metric threads is the most commonly used combination. Great experience has been gathered in the repair and strengthening of beams made of solid timber, both softwood and hardwood, and in connecting concrete slabs to floor beams. For applications where corrosion or weight could be of concern, the use of pultruded FRP rods is quite common. Some investigations have also aimed at the use of reinforcement bars (rebars), e.g. [3, 4]. Basically all types of adhesives useful for wood bonding have also been tried for GiR, but one and two-component epoxies, PUR and resorcinol types are those most frequently used in practice. Specific adhesive products have been formulated to fulfil the needs of GiR connections/reinforcement with timber, which offer much better performance with respect to strength and durability. A large number of parameters impact the strength of GiR. Hence, the challenge is to adequately account for these in design and to provide quality control measures to guarantee a reliable load bearing behaviour of GiR, which are usually assigned high loads by the designer.

2.

Application – Gluing-in the rods

2.1

Variants

There are several possibilities on how to glue rods into the wood [5]. Most often, a hole is drilled into the timber member with a diameter that exceeds the nominal diameter of the rod by 1 mm to 4 mm. This will result in glue line thicknesses from less than 1 mm to 2 mm. Thin glue lines are usually preferred over thick glue lines as many adhesives perform better the thinner the glue line is made and, in addition the necessary quantity of the expensive adhesive is reduced. In general the holes can be drilled in any direction relative to the grain. An important step after drilling is to clean the hole thoroughly. If pressurized air is used for this purpose it has to be verified that the air is free of oil-dust. If the rods can be set into holes with openings situated at the top of an element an easy variant is to first pour a defined quantity of adhesive into the hole and then to set the rod (Fig. 1(a)). Depending on the viscosity and the open time of the adhesive the rods may sink into the adhesive-filled hole due to their own weight or the rods have to be pushed into the adhesive filled hole. A disadvantage of this method is that there is no adequate control of the glue line quality in terms of assuring that the adhesive fills all cavities completely and no voids are present in the glue line. Another often used technique for setting the rod is to drill a second hole, this second hole being drilled perpendicular to the hole drilled for the rod. This hole should lead to the lower end of the rod and thus the adhesive can be injected under pressure from the bottom (Fig. 1(b)). For every rod the injection of adhesive will be continued until it can be observed that the adhesive pours out at the top of the hole that contains the rod or at another hole positioned at the desired position. The rod has to be fixed while the adhesive is injected. If the opening between rod and hole is sealed (for example by means of a moulded part or super glue), it is also possible to set the rods in a horizontal or overhead configuration as shown in Fig. 1(c) and (d).

134

Reinforcement with glued-in rods (a)

aa

(b)

a a

bb

(c)

b b

cc

c c

d d

(d)

d d

Fig.1 Variants for the application of GiR

(a)

(b)

(c)

(d)

(e)

Fig. 2 For optimum performance avoid: unwanted inclination of drilled hole (a), inclined setting of rod in hole (b), eccentric position of rod in hole (c), incomplete insertion of rod in hole (d), voids in glue line (e)

In the literature other variants of the application of GiR can be found. One of these uses a concentric continuous hole in the rod for the injection of the adhesive [6]. In another variant the rod is drilled into an adhesive filled hole with a diameter equal to or smaller than the nominal diameter of the rod. This procedure can be regarded as a combination of glued-in and drilled-in rods technology. However, today these two latter methods are not of significant importance for practical applications of GiR. 2.2

Quality control

As for all glued connections quality control of the manufacturing process is of great importance. The following parameters have to be checked when GiR connections or reinforcements are applied: Material 

Timber: strength class, moisture content (MC) 135

Reinforcement of Timber Structures 

Adhesive: proven to be suitable for gluing in rods, technical specifications, climatic conditions, open time, curing time



Rod: correct geometry, correct type/strength according to design (an overstrength rod can have adverse effects in this application), corrosion, condition of surface (free of oil and/or lubricants)

Application 

Hole: position (including edge and rod distances), diameter, depth, inclination, straightness, cleanliness (Fig. 2a)



Complete and centered positioning of rod in the hole (Fig. 2b-d). Depending on glue line thickness the use of spacers and/or centering devices like e.g. plastic or metal rings or a countersink at the bottom of the hole might be required.



Adhesive: application according to manufacturer specifications, control of filling level, presence of voids (Fig. 2e)

3.

Key parameters

The following parameters impact the load bearing capacity of GiR connections/reinforcement [5]:

Geometry 

Ratios of area of wood, adhesive area and rod area



Absolute size of the anchoring zone (represented by hole diameter d h and anchorage length  )



Slenderness ratio, which is defined as    / d h



Number of rods, edge distances and rod-to-rod distances



Rod-to-grain angle (including unintentional deviations from planned angle due to production process, definition of a tolerance-range)

Material stiffness 

Moduli of elasticity (MOE) and shear moduli of rod, adhesive and wood



Ratios of MOE to shear modulus for each material (especially important for the wood material, this being strongly orthotropic)

Material strength

136



Strength of the wood (especially shear strength and tensile or compression strength perpendicular to the grain). Note that the strength of wood is influenced by the density and that solid timber and glulam are usually assigned to strength classes according to EN 338 [7, 8] or EN 14080 [9] respectively. (Also applies to engineered wood products!)



Cohesive and adhesive strength of the adhesive



Ultimate strength of the rod material (for steel rods the yield strength is also important).

Reinforcement with glued-in rods Fracture mechanical properties of wood and adhesive 

Fracture energy and fracture softening characteristics

Variability of all properties 

Irregularities, i.e. deviation from nominal properties



Variations in mechanical properties of wood, rod and adhesive

Loading conditions 

Direction of external load on the rod in relation to its axis (pull-out, shearing) and reaction forces on the specimen that counteract the external load in the tests (Fig. 3)



Load duration (static)



Number of load cycles, frequency and amplitude (dynamic)

Other parameters 

Wood species



Special features to reduce stress peaks and/or to guarantee for a ductile failure mode



Manufacturing practice (curing time and pressure, surface characteristics etc.)



Quality control. F F/2

F

F/2

F

(a) pull - pull

(b) pull - compression F

F

e F/2

F/2

F/2

F/2

e

Pile foundations

(c) pull - beam

(d) pull - pile foundation

Fig.3 Different types of loading GiR specimens may be subjected to in tests of axially loaded rods[5, 10]

4.

Adhesives

A variety of adhesives have been tested to glue in rods. In early years, traditional wood adhesives based on phenol-resorcinol (PRF) or epoxies (EPX) were used, while later work has included also the use of polyurethanes (PUR) for instance. In 1999, Kemmsies made an investigation regarding the suitability of 12 different adhesives and adhesive/sealants [11]. In the experiments conducted within a large European research project in the late 1990s, (GIROD), three types of adhesives were used and compared [12]: PRF, EPX and PUR. From the tests and analyses it was concluded that the adhesives

137

Reinforcement of Timber Structures revealed increasing strength in pull-out tests in the following order: fibre reinforced PRF, PUR and EPX. EPX adhesives develop a strong bond with both steel and the wood, resulting in the wood becoming the weakest link of the connection. Thus the fracture properties of the wood or the wood/adhesive interface are decisive for the pull-out strength. Characterising an adhesive only by terms like EPX or PUR is not sufficient. There are many adhesives available of each type and they ‘‘can show all types of constitutive behaviour’’ (regarding EPX: [13]). The pull-out strength of the GiR is obviously related to the adhesive type, but also to the used wood species, since different adherents may develop different bonding strength with different adhesives [14]. Generally speaking, and to a varying degree depending on the specific adhesive used, the bonding strength can be affected by shrinkage during initial hardening, by the adhesive’s sensitivity to elevated temperatures, by its limited gap-filling qualities and by the sensitivity to moisture content changes [14]. All these effects have to be taken into account in design [5, 15]. Adhesives for GiR connections must, in addition to good strength and durability, have acceptable creep and creep-rupture properties. In order to assess these properties tests based on existing methods (e.g. longitudinal shear strength according to EN 302-1) were developed [16]. However, up to now there are no specific standards or guidelines available on how to test adhesives for GiR. The choice of adhesive is not independent of the method used to produce the connections. The main parameters of concern are the adhesion to the wood, the mechanical link to the rod (interlocking), the thickness of the glue line and the properties (e.g. viscosity) of the bonding agent [5]. The adhesive should demonstrate good gap-filling properties. For the connections with GiR there are many failure locations and modes which can be decisive for the load bearing capacity (see 5.3). The adhesive might be chosen during the design of the connection taking into account geometrical properties, requests of application methods and with the aim of avoiding a brittle failure mode finally making sure that the adhesive bond will not be the weakest link of the connection [17] in order to profit from the full capacity in shear strength that wood offers. In countries like Sweden, UK, Switzerland, Germany [18] and New Zealand [19] the most commonly used adhesives for connections and reinforcement with GiR are 2-component PUR and EPX.

5.

Mechanics, failure modes, design philosophy

5.1

Mechanical behaviour of GiR connections

Mostly, our knowledge about the mechanical performance of GiR connections is based on practical experience and design formulas developed by curve-fitting of empirical data [5]. The majority of the studies performed have focused on axial pull-out strength of a single GiR and its dependency on various material and/or geometrical parameters. During axial pulling, the load transfer between timber and rod is governed by shear of the adhesive. Depending on the strength of the adhesive and the surface characteristics of the rod and its surface treatment, the anchorage between the threaded rod and the adhesive may act as a mechanical connection [20, 21] similar to screws [22, 23]. Some design codes (e.g. [24, 25]) even do not allow to use rods lacking a threaded surface, since a pure adhesive bond is suspected not to be able to guarantee a reliable and durable force transfer. The force transferring mechanism is also influenced by the ratio of the diameter of the hole to the diameter of the rod, i.e. the bond line thickness. In some sources it is claimed that GiR connections act like a combination of glued and mechanical connections

138

Reinforcement with glued-in rods [13, 26, 27]. For rods inserted in undersized holes, it can be expected that the connection strength to a large extent results from the mechanical interaction between the wood and the thread of the rod [28]. One major advantage with the GiR connections is the transfer of forces directly into the inner part of the members’ cross-sections [29]. The connection is actually a hybrid one, made up of typically three different materials (wood, adhesive, rod) with different stiffness and strength properties [14] which have to work under loading simultaneously. This severely complicates the analysis of the connections and is one of the reasons for today’s lack of full understanding of the behaviour of this connection type as well as for agreeing on a design model.

5.2

Theoretical approaches to describe the behaviour of the adhesive bond

The adhesive bond line (by bond line or glue line is meant the adhesive layer plus the interface between adhesive and adherends) plays a major role in the overall behaviour of the GiR. Different approaches to describe the laws governing the behaviour of adhesive connections can be found in literature: (a) traditional strength analyses, (b) analyses based on linear elastic fracture mechanics (LEFM) and, finally, (c) so called non-linear fracture mechanics (NLFM) analyses [5]. In a traditional strength analysis, one tries to predict the stress (and strain) distribution in the GiR for a given loading situation, and then applies some failure criterion for this distribution. The failure criterion can be based on stress or strain, involving also multi-dimensional criteria. The approach will give a prediction of the load bearing capacity of the GiR, and also give a prediction of the stiffness. The stress (and strain) distribution can be determined with analytical or numerical methods, the former e.g. according to the Volkersen theory [30-33]. When using the framework of LEFM, the situation of loading a connection with a pre-existing crack is considered. The crack introduces a stress (and strain) singularity, and thus a traditional single point maximum stress criterion is not useful. Instead the crack driving force, also known as the energy release rate is calculated. The energy release rate is defined as the amount of (elastic) energy released during crack propagation. The critical energy release rate of the connection, Gc, is the amount of energy needed to increase the crack area. By assuming that failure of the connection takes place when the strain energy released is equal to the critical energy release rate of the connection, the load bearing capacity can be calculated [34]. NLFM provides a framework that takes into account not only the strength of the bond line (like in a strength analysis) neither only the fracture energy of the connection (like in the LEFM approach), but in fact accounts for both [34]. NLFM provides a framework that can be said to include both the framework of traditional strength analysis and LEFM. In traditional strength analysis it is assumed that the strength of the material is limited and that the fracture energy is either zero or infinite, the latter for the case of perfect plasticity. If a crack exists, such traditional strength analyses methods will fail since infinite stress (or strain) will be predicted. The framework of LEFM will, as mentioned above, only be applicable to the cases with a pre-existing crack. LEFM assumes finite fracture energy but an infinite strength of the material. NLFM takes into account not only the limited strength of the bond line but assumes also that fracture energy is limited. Thus NLFM can be used for any situation, independent of whether a crack exists or not. In NLFM the stress-strain relation used in conventional approaches is exchanged by a non-linear stress-displacement relation, such that the bond line, after stress has reached the strength of the material can still transfer load. This post peak-stress load transferring capacity diminishes with increasing displacement (normal opening or shear slip across the bond line) and will eventually reach

139

Reinforcement of Timber Structures zero. Thus, a typical stress versus displacement relation involves both an ascending part (typically the linear elastic response) and a post peak-stress descending part, known as strain softening [34]. The choice of theory to be applied basically depends on the predicted failure characteristics (brittle or ductile) of the adhesive bond, related to the properties of the bonding agent, to the size and shape of the connection and to the stiffness of the adherents [34]. For ductile adhesive bonds, stress based approaches can be useful. For very brittle adhesive bonds, an approach based on LEFM can be appropriate, and in theory, a NLFM-approach can be used for both these cases and any in-between situation. It must be emphasized that the failure characteristic of the bond line (brittle or ductile) depends on material (strength and stiffness of timber, type and strength of adhesive), geometry (surface and thickness of bond line) and loading conditions. The main tendency in improving the theoretical basis for analysis of connections with GiR follows Gustafsson’s further development of the Volkersen theory taking into consideration the NLFM. Since the GiR connection is a type of adhesive connection, usually those theories are the base for the stress analysis. The broad description of available theories and the historical development of them are available in many sources, for instance in [13].

5.3

Failure modes

The GiR connection acts like a chain consisting of the serial links “rod”, “adhesive” and “wood” [8], the load bearing capacity and failure mode being influenced by the parameters listed in chapter 3. The following failure modes are relevant for a single rod. Although such connections are of little interest in practice, they form the basis for research and the design of group of rods.

1. Failure of the rod due to a. material failure b. buckling of the rod in case of compression loading

2. Pull-out of the rod due to a. adhesive failure at the steel-adhesive interface (in case of lack of rods without profiled surface) b. cohesive failure in the adhesive c. adhesive failure at the wood-adhesive interface d. cohesive failure in the wood close to the bond line

3. Pull-out of wood-plug 4. Splitting failure of the wood due to a. short edge distances b. the rod being not set perfectly parallel to the grain c. excessive perpendicular to the grain loading

5. Tensile failure in the net or gross wood cross-section In addition to these failure modes for single-rod connections, the following are of interest for multiple rod connections:

140

Reinforcement with glued-in rods 6. Splitting failure due to short rod-to-rod distance 7. Group pull-out Splitting due to shrinkage or excessive shear stresses, and especially due to the stress peaks that are typically formed at the end of the rod [5, 22, 31] can be prevented by transversely reinforcing the connection, e.g. by means of self-tapping screws or threaded steel bars glued into drilled holes [35], crossing potential crack lines, approximately 50 mm from the end of the member [36]. Other possibilities to overcome the peaks in the shear stress distribution are to countersink the drill hole or to widen its diameter at the face end [10]. In references [4, 37] it is suggested to shift the anchorage zone to the inner part of the member (i.e. away from the surface) by either applying no adhesive at the face end of the drill hole or by turning off the thread of the bar over a certain length in order to prevent indentation and shear force transfer there. Successful experiments with widened bottom parts of the drill hole which allow the adhesive to spread in bulbs are reported in [38]. Since moisture induced stresses increase the risk of splitting, the application of GiR is usually restricted to service classes 1 and 2 [39].

5.4

Design philosophy

Depending on the design philosophy basically each of the aforementioned links can be assigned to be the weakest one. Whilst it is straightforward to calculate the tensile strength of the rod, at least in cases where the material quality is clearly defined and is not influenced by too high variations, the load bearing capacity in the wood and in the adhesive as well as in the interfaces is more difficult to estimate. In practice, the failure load for each of the failure modes must be assessed and the design philosophy set, in order that a chosen failure mode can be ensured or prevented respectively. It has to be clearly differentiated between experimental investigations and design for practice. In the first case the connections are designed such that usually the wood should be the weakest link (in order to identify the maximum load bearing capacity of the connection being subject of investigation). In the second case assigning the rod to be the weakest link allows for ductility and robustness. Several design approaches have been suggested [5]. One approach could be to ensure that a connection fails in a ductile failure mode, such as by failure in the steel, which of course must allow large plastic strains to develop with constant or monotonically increasing load capacity until final collapse [37, 40]. Some design codes (e.g. the Swiss design code SIA 265:2012 [21]) prescribe this type of ductile failure, which is favourable for any design case, regardless of materials in use and regardless of the possibility of seismic actions. In case of multiple rod connections it is even of greater importance to aim for a ductile failure mode. Only if the steel rods are the weakest link a uniform distribution of the load among all rods of the connection is possible [37]. Plastic deformations in the steel rod can develop only if there is sufficient free length for elongation. To achieve this, a part of the rod near the surface of the timber should be left unbonded [2, 4, 20, 41, 42] and if possible also necked down to a slightly smaller diameter by turning off the thread [4, 41]. This helps to prevent mechanical interlocking in this particular part of the anchorage zone and to “force” plastic deformations to develop in this zone [4, 37, 43]. With respect to ductility there is certainly an advantage in using mild steel with large yield capacity. For GiR connections in high strength timber like beech or ash rods of quality 8.8 may be indicated. This is also the case when (in experimental investigations) pull-out failures are to be achieved in order to derive the optimal anchorage length, to check performance of a specific adhesive or to study the influence of parameters like wood density or shear strength of the wood.

141

Reinforcement of Timber Structures It is worthwhile mentioning that no matter what failure mode is intended the engineer has to be able to assess all of the above failure modes, in order to perform the design [5]. The adhesive used, in any case, shall not be the weakest link, because this would not allow for making use of the full capacity the glued-in rod connection provides. There is thus no contradiction in performing large test series intended to assess the pull-out strength of GiR, even if the practising engineer would rather choose a failure mode based on plastic failure taking place in the rod. In order to optimize performance of GiR connections (1) the transfer of stresses should be steady, (2) deviations between force and grain direction should be small, (3) both the rod(s) and the timber should have similar stiffness (i.e. ETimber  ATimber  ERod  ARod , which in case of steel rods results in

ATimber  16 to 20  ASteel ) and (4) the deformation in rod and timber should be in similar range and not exceed the ultimate deformation capacity (2 to 3 ‰ for Norway spruce) [37, 41].

6.

Design of GiR connections

6.1

Background

Over the past 25 years, despite many national research projects, European projects, COST Actions (e.g. E13, E34) and constant practical application of GiR there is still no universal standard for the design thereof [44, 45]. This mainly seems to originate from the many different design approaches available in the literature for defining the behaviour of the adhesive connections and due to the fact that a large number of parameters impact the design. An early design approach was published in 1988 by Riberholt [46], who proposed an equation for the estimation of the pull-out strength of an axially loaded single GiR. In the 1990’s a considerable amount of experimental work was done and different design methods were presented (see below). Certain design methods were introduced into national design standards and in 1997 a proposal was implemented in a pre-version of the Eurocode, the pre-standard prEN 1995-2 [47]. Although not being exclusively related to the design of timber bridges, the design rules for GiR had to be included in part 2 of EN 1995 since, at that time work on prEN 1995-1-1 had already been finalized and it was not possible anymore to amend this part of prEN 1995. In 1998, the European GIROD project was launched. The main objective of this project was to establish design rules and the project result was a new calculation model based on the generalized Volkersen theory (GIROD Project Report 2002, [48]). This resulted in a proposal to be implemented in the pre-standard prEN 1995-2, Annex C [49]. During the CEN/TC 250/SC 5 meeting in 2003 it was decided to discard the Annex C. Delegates argued that the proposed code text did not meet the actual status of research (e.g. [50-52]). Recently both past and actual research has been re-visited with the purpose to propose a design approach that could replace several national design rules. Proposals and design rules developed during the years are shown in Fig. 4. A calculation model has to take into account all relevant parameters that impact the load bearing capacity of glued-in rods (see chapter 3). Although there are numerous studies and calculation methods, and although in an earlier version of EN 1995 design methods exists, the basic problem is still which method to accept and to implement in EN 1995. It is clear that a lack of a common European design approach is a serious obstacle to the exploitation of the GiR connection [44]. For more than ten years many research efforts and research programs have contributed to the knowledge about GiR and attempted to provide the information required to prepare design rules which would allow an increased, more advanced and more reliable use of GiR in timber structures

142

Reinforcement with glued-in rods [53]. Stepinac et al. [54] obtained, via online survey, information about the practical use of GiR and about problems the designer faces when designing this connection. Results were as expected: Available design rules were characterized as unreliable and unsatisfying. The most commonly applied design approaches turned out to be the ones in prEN 1995-2, Annex C [49] and in DIN 1052 [24]. Beside the fact that a unified design approach still does not exist, the key reservations with the available design rules are [54]: 

Definition of rod spacing and edge distances are not reliable for rods under tension and shear load.



Design rules (and requirements in rod spacing and edge distances) often are too conservative.



Ductility should be treated as a key issue.



There are no reliable rules for multiple rod connections.



The duration of load (DOL) effect is not accounted for.



There are no design rules for the case of interacting axial load and transverse load.



The influence of load-to-grain angle is not addressed.



Some of the available design approaches contain non user-friendly formulae and/or parameters which are difficult to assess.

Fig. 4 Standards and proposals containing design rules to estimate the pull-out strength of GiR and researchers involved in the development in the last 25 years.

6.2

Comparison of design rules

Most of the available design equations are focused only on the pull-out strength of single axially loaded GiR, since a lot of research has been carried out exclusively dealing with pulling-out of single rods. In Sections 6.3 and 6.4 calculation models for rods set perpendicular to the grain and rules for multiple rods are introduced briefly. In the present Section rules commonly applied for the design of GiR are compared. Diagrams in this Section in general show graphs on characteristic level, except when stated in the caption of the respective figure.

143

Reinforcement of Timber Structures 6.2.1

Axially loaded single GiR parallel to the grain

Tlustochowicz et al. [5] and Stepinac et al. [54] explained in detail proposals and design rules published in the last 25 years. In this state-of-the-art review six design rules and methods which are most commonly applied are analysed and explained in detail: Riberholt equation, 1998 [46]:

Rax, k  f w1  c  d  lg

(1)

prEN 1995-2, 2003 [49]:

Rax, k    dequ  la  f ax, k  tan   / 

(2)

GIROD equation, 2003 [48]:

Pf   f    d  l  tan  /  

(3)

Proposal by Gehri, Steiger, Widmann, 2007 [43]:

Fax, mean  fv,0, mean    dh  l

(4)

New Zealand Design Guide, 2007 [36]:

Qk  6,73  kb  ke  km  l / d 0,86  d / 201,62  h / d 0,5  e / d 0,5 Rax, d    d  lad  f k1, d

DIN 1052:2008 [24] and CNR DT 206/2007 [55]:

(5) (6)

where: Rax,k / Pf / Qk

characteristic value of axial resistance [N], [kN]

Rax,d

design value of axial strength [N], [kN]

Fax,mean

mean value axial resistance [N], [kN]

l / la / lg / lad

glued-in length / effective anchorage length [mm]

d

nominal diameter of the rod [mm]

dh / h

diameter of the drill hole [mm]

dequ

equivalent diameter [mm]

e

edge distance [mm]

kb / km / ke

bar type factor / moisture factor / epoxy factor

ω

stiffness ratio of the connection

τf

local shear strength of the bond line [N/mm2]

fw1 / fv,α,k / fv,k / fax,k / fk1,d

strength parameter / characteristic value of the shear strength of the wood / design value of the shear strength of wood across the grain / characteristic value of the shear strength of the wood at the angle between the rod and grain direction / design value of the bond line strength [N/mm2]

fv,0,mean

nominal shear strength parallel to the grain of a single axially loaded rod [N/mm2].

d = diameter of rod dh

d l

l = anchorage length

dh = strength diameterof ofaxially hole loaded single GiR. Fig. 5 Geometric parameters influencing the pull-out e = glue line thickness

144

Reinforcement with glued-in rods The pull-out strength depends primarily on the interfacial layer and shear strength parameter which is influenced by mechanical and geometrical properties of three different materials. Hence, a simplified calculation model for axial loading could be similar to the one of screws:

Rax,k    d  l  f v,k

(7)

where: Rax,k

characteristic value of pull-out strength

l

anchorage length

d

diameter

fv,k

shear strength parameter.

However, the mechanics of GiR are complex, so any attempted simplification from the designer’s point of view would be helpful since making the design of GiR straight forward but might result in uneconomic connection design. A closer look at the simplified equation reveals several unanswered questions such as: Which diameter (diameter of rod, diameter of hole or equivalent diameter) and anchorage length (length of bonded rod or equivalent anchorage length) to use? Can the geometry of the hole be described by the slenderness ratio    / d ? Which parameters must be included in the shear strength parameter (timber density, MC of timber, MOE of timber, rod and adhesive, rod surface, rod material, type of adhesive, slenderness ratio, geometrical factors, etc.)? This (among other things) is the reason for present standards and proposals differing significantly (Fig. 6 and Fig. 7). From experts discussions it can be concluded that the most common design rules like the ones in prEN 1995-2 [49], the former DIN 1052 [24] are on the “safe side” while equations proposed in various scientific papers in most cases relying on experimental data derived from tests on specific connection systems deliver much higher values for the pull-out strength. The glue line thickness e is considered only in some formulas. Some standards propose a maximum value of 2 mm [24], [57], [21] but do not provide answers for thinner glue-lines. Differences and the influence on the calculated load bearing capacity are shown in Fig. 7. Fig. 8 and Fig. 9 show the characteristic value of the pull-out strength of one single axially loaded rod estimated on basis of different design rules whereby the diameter of the rod and the anchorage length were varied. Problems occur when defining these two parameters in the equation. The diameter d is sometimes the diameter of the rod [46], [24], the diameter of the drill hole [43] or an equivalent diameter [59], [56]. A similar problem applies for the definition of the anchorage length. The former prEN 1995-2 equation [49], which was based on the GIROD project findings, included several different parameters. Some of these parameters e.g. fracture mechanics parameters, cannot be determined easily by engineers in practice. The influence of wood density has been subject of several studies (e.g. [46], [56], [43], [59]) (Fig. 10). The opinions in the matter of influence of density on the pull-out strength of glued-in rods differ. The recommendations given in [47] for the design of GiR connections indicate that the axial strength of glued-in rods depends on the density of the wooden element. It could be expected that such a relation exists, having in mind that it has been demonstrated that the pull-out strength of nailed and screwed connections is dependent on the density of the wooden member [23, 60-62]. On the other hand, the correlation between density and strength of wood in general is poor [63].

145


Fig. 6 Comparison of the pull-out strength [kN] derived with different design approaches ([36], [43], [49], [53], [24], [56], [48], [47], [46]), (EPX, l=200 mm, ρk=370 kg/m3 (MC 6 mm, pre-drilling is required (depending on e.g. density of the timber). Their axial loadcarrying capacity is mostly limited by the pull-through capacity of the screw head. Thus the additional use of adequately designed washers is commonly meaningful. The combination of threaded and smooth shank and the use of mild- or low carbon steel allow using them for loading in shear and tension or a combination of both. Constraints in the load-carrying capacity of primarily axially-loaded screws were overcome with the development of self-tapping screws. In contrast to traditional screws which have their threaded part turned down from the original rod diameter, the thread of self-tapping screws is produced by rolling or forging a wire rod around the shank, which consequently features a smaller diameter when compared to the outer cross sectional thread diameter (see also Fig. 1). Self-tapping screws mostly feature a continuous thread over the whole length. This leads to a more uniform load transfer between the screw and the wood material as well as a considerably enhanced axial load-carrying capacity, the type of loading for which they are optimized. During manufacture, their thread is hardened, leading to an increased bending and torsion capacity, but also to a more brittle failure mechanism. In combination with the development of optimized drill tips and threads, self-tapping screws, featuring diameters up to 14 mm and lengths up to 1000 mm, are produced and applied today, see Fig. 1 and Fig. 2.

Fig. 1 Different forms of drill tips and threads (with and without shank cutter) as well as variations of screw heads; comparison with DIN-screw (third from right), from [5]

Fig. 2 Development of screw length (and implicitly load-carrying capacity) from traditional wood screws to self-tapping screws, from [6]

162

Reinforcement with self-tapping screws and threaded rods An extension of these geometric limits is possible by the application of threaded rods. These are a modification of self-tapping, fully threaded screws featuring screw threads over the full length. These can reach diameters of up to 20 mm and lengths of up to 3000 mm. Threaded rods with screw threads need pre-drilling with the core diameter and a coating and/or lubricant to reduce friction stresses when driving them in. In dependency of the length of the screw or threaded rod, their axial load-carrying capacity may be limited by the tensile capacity of the steel or, if loaded in compression, by buckling. Following EN 1995-1-1 [3] and investigations made in e.g. [4], a group of axially loaded screws may also fail in block shear. Requirements for self-tapping screws are for example given in EN 14592 [7]. Here, the nominal diameter equals the thread diameter d which has to be 2.4 mm ≤ d ≤ 24 mm (practical range: 8 mm, 10 mm and 12 mm for screws, 16 mm and 20 mm for threaded rods). The core diameter d1 has to be 0.6 ∙ d ≤ d1 ≤ 0.9 ∙ d (0.6 ∙ d ≤ d1 ≤ 0.75 ∙ d according to EN 1995-1-1 [3]). The minimum thread length is restricted to lg ≥ 4 ∙ d (minimum embedment depth or effective penetration (anchoring) length, lef, according to EN 1995-1-1 [3] is: for axially loaded screws 6 ∙ d and for laterally loaded screws 4 ∙ d,), see also Fig. 3.

l

d

dh

d1

lg

d d1 dh l lg

… nominal (thread) diameter … core diameter … head diameter … length … thread length

Fig. 3 Geometry parameters for screws The mechanical properties are defined by the characteristic values of (i) the yield moment, (ii) the withdrawal capacity of the threaded part, (iii) the tear-off capacity of the screw head, (iv) the pullthrough capacity of the screw head, (v) the tensile strength of the screw, and (vi) the torsional strength of the screw,. The mechanical properties can be derived from tests or from equations given in e.g. EN 1995-1-1 [3]. The use of other screws as specified in EN 14592 [7] is allowed, provided their applicability is proven by a technical approval. These used to be issued by national building authorities; currently most national approvals are being converted into European Technical Approvals (ETAs). Between the products available on the market, there are a variety of head and thread forms and differences in shank-, tip- and thread-diameter as well as different ratios between thread to core diameter, see Fig. 1. However, a comprehensive comparative study on self-tapping screws of five different manufactures, tested at angles between screw axis and fibre direction of 90°, 45° and 0°, showed only minor differences in the withdrawal capacities. The range of ± 10 % is in-line with the technical approvals. More relevant are the differences in the practical application of the screws, see e.g. [5]. Meanwhile, many regulations in technical approvals have been adapted between different approvals, facilitating design and comparability. Nevertheless, there are a number of approvals in which special rules are given that must be followed. For example, there are different rules in the following areas: 

minimum spacings and distances, minimum member thickness requirements;



axial withdrawal capacities (as a function of characteristic (5 %-quantile) density);



permissible angles between screw axis and grain direction;

163

Reinforcement of Timber Structures 

wood species (mostly softwoods);



tensile capacity;



stability of the screws, i.e. buckling failure when loaded in compression;



stiffness values (Kser, Kax,ser, Ku).

Because of these differences, specified screws shall not be substituted by other screws.

2.

General rules on the application of self-tapping screws

The European basis for the design of self-tapping screws is the design concept given in Eurocode 5 (EN 1995-1-1 [3]) in combination with the provisions given in the European Technical Approvals (ETAs). For applications which are not covered by Eurocode 5, design approaches can be given in the National Annexes to Eurocode 5 (as non-contradictory, complementary information (NCCI)). Some ETAs, e.g. [8], [9] and [10], also feature annexes containing design provisions for certain applications. One advantage of self-tapping screws is that they do not require pre-drilling, given the density of the timber is not too high (e.g. ρ < 500 kg/m3, given for most softwoods). However, research results in [11] indicate, that for the application of self-tapping screws in timber with temperature below zero, pre-drilling may be required to prevent splitting of the timber which becomes more brittle at these temperatures. In more recent technical approvals, pre-drilling is allowed. Here, the borehole diameter shall not be greater than the core diameter of the screw. Pre-drilling can have a positive effect on the precision of the screw positioning. This is of special importance if the screws are positioned at an angle to the grain, which is challenging if carried out free-handed, especially if screws of small diameter (e.g. d = 6 mm) and high slenderness are used. For this, placement devices or CNC machinery can be used. Some ETAs allow to reduce the minimum spacing if pre-drilling is applied. Most self-tapping screws feature a special drill-tip which allows the use of reduced spacings and distances even without pre-drilling. For such screws, most technical approvals (e.g. [8], [9] and [10]) contain the spacing and distance requirements for axially loaded screws given in Tab. 1 (see also Fig. 4). In the case of screws positioned at an angle between grain and screw axis, the centroid of the threaded part of the screw in the respective timber member shall be used as reference for determining spacings and distances (see Fig. 4 and e.g. [8], [9], [10] and [12]). In the case of reinforcement against tensile stresses perpendicular to the grain, the distance between the reinforcement and the (mostly localized) area of stress peaks should be minimized, e.g. by using the minimum of a3,c (see e.g. [12]). Tab. 1 Typical minimum spacings and minimum distance requirements given in technical approvals for axially loaded self-tapping screws

a1

a2

a3,c

a4,c

5∙d a

5∙d a

5∙d

4∙d b

a

may be reduced to 2.5∙d, if the condition a1·a2 ≥ 25·d² is fulfilled.

b

in some cases, 3∙d is allowed.

164

Reinforcement with self-tapping screws and threaded rods

Fig. 4 Definition of spacings, end and edge distances for axially loaded screws According to Eurocode 5 (EN 1995-1-1 [3]), a reduction in cross-section, caused by screws do not need to be considered in cases of d ≤ 6 mm (without pre-drilling) and for screws placed in areas under compressive stresses. For nominal diameters d > 6 mm, a reduction of the cross-section shall be taken into account. This is typically realized by considering only the cross-sectional area of the screws, cut at an angle α at the timber cross section featuring the highest number of penetrating screws (e.g. [12]). However, tension parallel to grain tests on butt-joints with inclined fully threaded self-tapping screws showed, that the net cross section is better approximated by the timber area between the projected screws, see Fig. 5, right. Thus, the sole consideration of the screw holes as loss in net cross section was shown to be not sufficient [13]. Recent research results [14] indicate that a reduction of the crosssection should also be considered in members under compression, even if the holes are filled with a material of higher stiffness than the wood. This is explained by stress peaks developing in the vicinity of the fasteners. The exception is glued-in rods since the glue-line shall lead to a reduction in stress peaks.

Fig. 5 Butt-joint with self-tapping screws and proposal for definition of net cross section, from [13] The characteristic axial load-carrying capacity of the screw, Fax,Rk, is determined as the minimum of the characteristic tensile capacity, Ft,Rk, (or characteristic buckling capacity, Fki,Rk) and the characteristic withdrawal capacity, Fax,α,Rk, of the screws, see Eq. (1). The characteristic withdrawal capacity, Fax,α,Rk, is calculated as the product of the characteristic withdrawal parameter, fax,k, (given in the technical approval; conforming with the characteristic withrawal strength multiplied by π = 3.14), the nominal diameter, d, and the effective anchorage length, ℓef. In the case that screws are positioned at an angle α between grain and screw axis, most approvals contain an equation with which the withdrawal capacity can be determined according to the values for α = 90°. Some technical approvals allow to use the values determined for α = 90° in case of angles α ≥ 45°.

165


 Ft , Rk or Fki , Rk  Fax, Rk  min    f ax,k  d  lef 

(1)

The slip moduli, Kser (slip moduli in shear) and Kax,ser (axial slip moduli), are needed e.g. in the case of mechanically jointed (e.g. doubled) beams which are designed with the γ-method or shear analogy (see chapter 6). Experiments to determine the axial slip modulus, Kax,ser, show large variations in results (e.g. [15] and [16] contain values approximately twice as high as in [17]). This variety of results is also observed in different technical approvals, which report different values of the slip modulus. Eq. (2), which is based on tests on screws from different manufacturers featuring maximum penetration lengths of 120 mm [17]), is generally reported in many technical approvals. Other technical approvals report different equations delivering higher results.

K ax, ser  780  d 0.2  lef

0.4

(2)

Test results reported in [15] and [18] show that the slip moduli are also significantly dependent on the angle α between screw axis and grain direction. In contrast to the withdrawal capacity the slip moduli increase with decreasing angle α. In [18], tests on screwed-in rods featuring penetration lengths of 200 mm and 400 mm are reported that indicate a disproportionate (above-average) increase of the axial stiffness when doubling the penetration length.

3.

Self-tapping screws as reinforcement to carry tensile stresses perpendicular to the grain

Screws under axial loading are very stiff, favoring their use as reinforcing elements. In the following subsections, typical applications of self-tapping screws as reinforcement to carry tensile stresses perpendicular to the grain in timber structures are presented. Within the approaches presented, the tensile capacity perpendicular to the grain of the timber is neglected, i.e. a cracked tension zone is assumed. This is different from to the approach taken by [19] in which only the force components, exceeding the tensile strength perpendicular to the grain of the timber, are considered for the design of the reinforcement.

3.1

Reinforcement of connections with a tensile force component perpendicular to the grain

The approach to design the reinforcement of connections with a tensile force component perpendicular to the grain explained in the following is standardized [20], previous sources include [21] and [22]. The approach is explained in [23] as well as [24], previous works include [25]. The design tensile force perpendicular to the grain, Ft,90,d, is the resultant of the tensile stresses perpendicular to the grain on the plane defined by the loaded edge distance to the center of the most distant fastener, he, (see e.g. [26]). According to beam theory, the connection force component perpendicular to the grain results in a step in the shear force distribution. The tensile force perpendicular to the grain, Ft,90,d, see Eq. (3), is determined from this change in shear stress by integration of the shear stress in the area between the row of fasteners considered and the unloaded edge, as indicated by the shaded area in Fig. 6 right. A derivation of this approach can be found in e.g. [27].

166






Ft ,90,d  1  3   2  2   3  Fv ,Ed with :   he h Fv ,Ed

(3)

design value of the shear force component perpendicular to the grain

The design approach for reinforcement of connections with a tensile force component perpendicular to the grain can also be translated to the reinforcement of girder hangers or the mortise part of a dovetail connection.

Fig. 6 Reinforced cross-connection: reinforcement (left) and distribution of shear stresses and shear flow (right, see also [27]) The highest tensile stresses perpendicular to the grain occur in direct vicinity of the fasteners. Therefore, the distance between the screw and the fasteners, a1, should be minimal. It is ideal to place the reinforcement (also) in between the fasteners, as shown in Fig. 6 left. With respect to further geometric considerations for the reinforcement, the specifications given in 0 apply. The load-carrying capacity of the screw is determined in dependence of the length of the screw between the row of fasteners considered and the unloaded edge, ℓad,t. The thread of the reinforcement should at least cover 75 % of the beam height, see e.g. [28]. In all other cases, the tensile stresses perpendicular to the grain at the screw tip have to be verified as well.

3.2

Reinforcement of notched members

The approach to design the reinforcement of notches in members with rectangular cross-section explained in the following is standardized [20], previous sources include [21] and [22]. The approach is explained in [23], [24] as well as [12]. Preceding works include e.g. [25]. The tensile force perpendicular to the grain, Ft,90,d, can be approximated by integration of the shear stress in the area between the inner and outer corner of the notch (see shaded area in Fig. 7 right). A derivation of this approach can be found in [27]. A more detailed analysis of the magnitude of the tensile stresses perpendicular to the grain around the notch, using plate theory, has shown that these stresses are increased due to the eccentricity between the support and the inner edge of the notch ([29], [30]). For relationships x ≤ hef / 3 (see Fig. 7), the tensile force perpendicular to the grain, Ft,90,d, can be sufficiently estimated by applying an increase factor of 1.3. The design tensile force, Ft,90,d, to be carried by the reinforcement, can be determined according to Eq. (4).

167

Reinforcement of Timber Structures Ft ,90,d  1.3 Vd  [3  1     2  1    ], for x  hef 3 2

with:

3

  hef h

(4)

Vd design value of the shear force The design approach for notched members can also be used for the reinforcement of the tenon part of a dovetail connection, as e.g. applied by [31]. Due to the limited distribution length of the tensile stresses perpendicular to the grain outside the corner of the notch, the distance between the screw and the notch, a3,c, should be minimized, see Fig. 7 left. Only one row of screws at a distance a3,c should be considered, hence it should be aimed at placing all necessary screws in the first row, utilizing – if necessary – the minimum possible distance a2 between the screws (see Fig. 4). The load-carrying capacity of the screw is determined in dependence of the smaller of both anchorage lengths, with ℓef = ℓad.

Fig. 7 Notched beam: reinforcement (left) and distribution of shear stresses (right) Since end-grain is exposed and commonly not sealed at a notch, the superposition of moisture induced stresses and load-dependent tensile stresses perpendicular to the grain around holes can be significant [32]. Therefore, some authors recommend that notched members should always be reinforced, see e.g. [12], [33]. Experiments on notched members reinforced with screws have shown the potential of a crack developing from the corner of the notch. The crack development ends just after it has crossed the screw. The reason for this can be seen in the very small deformation capacity of timber before exceeding the tensile strength perpendicular to grain. Although it could be considered a visual deficiency, such a crack is not a sign of reduced load-carrying capacity of the notched beam. Recent research [34] indicates that Eq. (4) might not be conservative for all cases since it does not account for shear failure of the reinforced notch. The authors state that a combination of reinforcement to cover both shear and tensile stresses perpendicular to the grain, has a positive effect on the structural behavior of reinforced notched beams. Recent developments also include self-tapping screws with changing angle of the thread along the screw. Such screws can be used to induce compressive stresses perpendicular to the grain, thereby reducing the effect of load-dependent or moisture induced tensile stresses perpendicular to the grain [35]. The maximum acceptable displacement between the thread and the wood material is limited, implying that this approach could be most adequate for details with localized areas of high tensile stresses perpendicular to the grain. So far, the influence of relaxation on the long-term occurrence of the imposed compressive stresses perpendicular to the grain could not be clarified.

168

Reinforcement with self-tapping screws and threaded rods 3.3

Reinforcement of members with holes

The approach to design the reinforcement of members with holes explained in the following is standardized [20], previous sources include [21] and [22]. The approach is explained in [23] as well as [24], research on this approach is presented in [36]. Modifications to the approach are explained in [37] and [38]. Modifications constitute the limitation of the permissible relative dimensions of the hole in dependency of the type of reinforcement (e.g. hd ≤ 0.3 ∙ h in case of reinforcement with screws). Information on the behavior of holes in timber beams in general can be found in e.g. [39], [40], [41], [42] and [43]. The tensile force perpendicular to the grain, Ft,90,V,d, see Eq. (5), can be approximated by integration of the shear stress between the axis of the member and the corner of the hole under tensile stresses perpendicular to the grain as indicated by the shaded area in Fig. 8 right. A derivation of this approach can be found in [27]. Experimental results given in [40] also show a tensile failure perpendicular to the grain at holes in areas without shear. From the experimental results, a tensile force component due to bending moment, Ft,90,M,d, was determined which has to be added to the force component due to shear. In the case of circular holes, the parameter hr may be increased by 0.15 ∙ hd to account for the fact that potential fracture will occur at a position described by an angle of 45° from the center line of the hole. The applicable effective anchorage length, ℓef, is equal to the applicable distance hr from the edge of hole to the upper/lower edge of the member, given the length of the screw after the potential failure plane exceeds hr. Ft ,90, d  Ft ,90,V , d  Ft ,90, M , d  with:

Vd

Vd  hd 4h

 h 2 M  3  d2   0.008  d hr h  

design value of the shear force at the edge of the hole

M d design value of the bending moment at the edge of the hole a hd

hole length hole depth

h rl

distance from lower edge of hole to bottom of member

(5)

h ru distance from upper edge of hole to top of member  min hro ; hru  for rectangular holes hr =   min hro  0,15  hd ; hru  0,15  hd  for round holes lA support distance of a hole

Fig. 8 Beam with hole: reinforcement (left) and distribution of shear stresses (right) The superposition of moisture induced stresses and load-dependent tensile stresses perpendicular to the grain around holes can be significant [32], in particular if the end-grain in the holes is not sealed. Therefore, some authors recommend that members with holes should always be reinforced, see e.g. [33]. In most cases, the necessity to reinforce holes will be given by the geometrical boundary

169

Reinforcement of Timber Structures conditions for unreinforced holes (hd ≤ 0.15 ∙ h). With respect to geometric considerations for the reinforcement, e.g. distances, the specifications given in 0 apply. It shall be emphasized that besides the reinforcement with screws also a verification of the shear strength of the timber in the vicinity of the hole is required. The distribution of the shear stresses in the vicinity of the hole deviates considerably (polynomial stress distribution instead of parabolic stress distribution), its maxima can reach significantly higher values compared to the values determined according to beam theory. A description as well as an associated design equation is given in [24], see Eq. (6). In [44] it is recommended to apply the same verification for round holes as well. In the same publication, a method is described to verify the bending stresses above or below rectangular holes, including the additional longitudinal stresses from the frame action (lever of the shear force) around the hole (see also [40]). 0.2

1,5 Vd 1,5 Vd  a  h   1,84  1     d   b   h  hd  h h b   h  hd       max factor to take account for the increased shear stresses in the area of the edge of the hole

max   max  with:

(6)

description of terms see Eq. (5) h d may be replaced by 0.7  h d in case of round holes

3.4

Reinforcement of double tapered, curved and pitched cambered beams

Double tapered, curved and pitched cambered beams mostly feature beam depths which exceed the maximum length of self-tapping screws. An alternative reinforcement is given by threaded rods which are produced in lengths of up to 3000 mm and installed by aid of pre-drilling with the core diameter. Threaded rods featuring screw threads constitute a modification of self-tapping, fully threaded screws, i.e. the main specifications given above apply. A standardized approach to design the reinforcement of curved and pitched cambered beams is given in [20], previous sources include [21] and [22]. The approach is explained in [23]. It is differentiated between reinforcement to carry the full tensile stresses perpendicular to the grain and reinforcement to only carry the tensile stresses perpendicular to the grain from climatic conditions, i.e. the moisture induced tensile stresses perpendicular to the grain. The first approach is based on an integration of the sum of tensile stresses perpendicular to the grain in the plane of zero longitudinal stresses. The equations given in [3] (based on [45]) only provide the maximum tensile stresses perpendicular to the grain in the apex. Depending on the form and loading of the beam, the tensile stresses perpendicular to the grain decrease with increasing distance from the apex. They even spread to some extent into the straight parts of the beam, see Fig 9 and [46]. In [47] and [48], these results were verified and extended to beams with mechanically fixed apex, see Fig 9 below right. For simplification, reinforcement in the inner quarters of the area exposed to tensile stresses perpendicular to the grain is designed for the full tensile stresses perpendicular to the grain, see Eq. (7).

Ft ,90,d 

170

t ,90,d  b  a1 n

(7)

Reinforcement with self-tapping screws and threaded rods with:

Ft ,90,d design tensile force perpendicular to grain in the reinforcement t ,90,d design tensile stress perpendicular to grain b a1

beam width, in [mm] distance between fasteners in grain direction

n

number of reinforcing fasteners within a1

Fig. 9 Distribution of tensile stresses perpendicular to the grain over the beam height and length in double tapered (above left), curved (above right), pitched cambered (below left) and cuved beam with mechanically fixed apex, i.e. secondary apexes (below right) In the outer quarters, the tensile stresses perpendicular to the grain are assumed to reach 2/3 of the maximum tensile stresses perpendicular to the grain determined with Eq. (7), see Fig 9.. The spacing between the reinforcement is limited to a1 ≤ 0.75 ∙ hap (hap as height of the apex) to ensure that the whole area exposed to tensile stresses perpendicular to the grain is covered by reinforcement. Even if the requirements regarding systematic, load-dependent tensile stresses perpendicular to the grain can be met, see [3], it is state of the art to reinforce double tapered, curved and pitched cambered beams. Reason is the superposition of the load-dependent stresses with moisture induced stresses perpendicular to the grain due to e.g. changing climatic conditions or a drying of the beam after the opening of the building, see e.g. [49], [50]. In the lack of a method to reliably predict the magnitude of tensile stresses perpendicular to the grain, it is custom to apply reinforcement if the maximum load-dependent tensile stresses perpendicular to the grain exceeded 60 % of the design tensile strength of the timber member perpendicular to the grain. Some authors recommend that double tapered, curved and pitched cambered beams are always reinforced, independent from the magnitude of tensile stresses perpendicular to the grain, see e.g. [33]. Eq. (8) represents one approach to design reinforcements to carry moisture induced tensile stresses perpendicular to the grain ([20], [21] and [44]). It is based on the assumption that 1/4 of the tensile stresses perpendicular to the grain from external loads are carried by the reinforcement. It is also based on the assumption that the potential magnitude of moisture induced stresses increases with increasing member width, i.e. increases due to decelerated adaption of timber moisture content in the interior of the cross-section. For a member width b = 160 mm, Eq. (8) lends ¼ of the full reinforcement determined by Eq. (7), less reinforcement for smaller member widths and more reinforcement for larger member widths.

171


Ft ,90,d  with:

t ,90,d  b2  a1 640  n Ft ,90,d design tensile force perpendicular to grain in the reinforcement t ,90,d design tensile stress perpendicular to grain a1

distance between fasteners in grain direction

n

number of reinforcing fasteners within a1

(8)

In the case of reinforcement to carry the tensile stresses perpendicular to the grain from climatic conditions, the spacing between the reinforcement should be kept constant. It is recommended that the spacing is limited to the member depth [44]. Recent research has begun to examine the question of the influence of reinforcement on the magnitude of moisture induced stresses since reinforcement restricts the free shrinkage or swelling of the timber beam. Experimental studies (short-term tests) and analytical considerations, presented in [51] and [52], indicate that a reduction of timber moisture content of 3 % to 4 % around threaded rods, positioned perpendicular to the grain, can lead to critical stresses with respect to moisture induced cracks. In addition, a substantial mutual influence of adjacent reinforcing elements could be identified. Experiments on drying glulam members reinforced by screwed-in threaded rods placed perpendicular to the grain are presented in [53]. The results indicate that moisture induced stresses in the timber can lead to forces in the reinforcement in the order of the steel capacity of the rods.

3.5

Reinforcement of connections with a tensile force component parallel to the grain

The load-carrying capacity of connections with multiple fasteners in one row parallel to the grain can be lower than the sum of the load-carrying capacities of the single fasteners (depending on the fastener distances parallel to the grain). This is due to the splitting forces (tensile stresses perpendicular to the grain) induced by the fasteners. The tendency to splitting increases with decreasing spacing of the fasteners parallel to the grain, a1 (Fig. 10). In the codes, this is accounted for by an effective number of fasteners, nef, [3] (based on [54]). Placing self-tapping screws with continuous thread perpendicular to the fastener axis and to the grain direction may prevent splitting, i.e. nef = n may be used, see Fig. 10. The closer the screw is placed to the dowel-type fastener, the better the effect. Since the splitting force is highest close to the joint between two connected members, the screws should be positioned with a minimum edge distance, a4,c, see Fig. 10 right. In [44] it is recommended to design the self-tapping screws for an axial force of 30 % of the load transferred by each dowel-type fastener and shear plane.

Fig. 10 Reinforcement of dowel-type connections According to [55], [56] as well as [57], the load-carrying capacity of dowel-type connections, reinforced with self-tapping screws with continuous thread can be further increased, if the screw is in direct contact with the dowel-type fastener. This necessitates very exact positioning which is not always feasible. Pre-drilling can have a positive effect on the precision of screw positioning.

172


4.

Self-tapping screws as reinforcement to carry compressive stresses perpendicular to the grain

Structural details in which the timber is loaded in compression perpendicular to the grain are very common, e.g. beam supports or sills / sole plates. The combination of high loads to be transferred over localized areas and low capacities in compression perpendicular to the grain can make it difficult to meet the associated verifications. Fully threaded, self-tapping screws are a means to improve the stress dispersion into the timber, see Fig. 11. Research on this type of reinforcement is presented in [56], background information as well as a design approach is presented in [58].

Fig. 11 Reinforcement of support areas The load-carrying capacity of a reinforced support can be determined under the assumption of an interaction between the timber under compressive stresses perpendicular to the grain and the screws under compression. This assumption is valid if certain deformations of the loaded edge are accepted. In addition it should be verified that the compression capacity perpendicular to the grain of the timber is not exceeded at the screw tips (transition between reinforced and unreinforced section). For this, an angle of stress distribution of 45° may be applied. On the safe side, this angle should be measured from the screw heads, see Fig. 9. Here, the factor kc,90 shall be taken as kc,90 = 1.0.   Fc,90, Rd  nS  Fax , Rd F90, Rd  min  b  lef  f c,90,d   with:

F90, Rd design force perpendicular to grain with reinforcement Fc,90, Rd design compression perp. to grain capacity without reinforcement, with Fc,90, Rd  kc,90  b  l  f c,90, d (see [2])

(8)

Fax , Rd design axial load-carrying capacity according to Eq. (1) (without Rt , Rk ) f c,90,d design compression perp. to grain strength kc,90

factor adjusting compression perp. to grain strength to real design situations (see [3])

nS

total number of screws

b

beam width (support width  b)

l l ef

length effective length

The compression force must be evenly distributed to all screws and the compression stresses at the screw heads have to be absorbed by the bearing material. These two requirements can only be met by a hard bearing material. This can be realized in form of a hard intermediate layer from e.g. steel, designed in adequate thickness and thus capable to transfer the load uniformly. The screws shall be equally distributed over the bearing area and the screw heads shall be on one line with the surface of the timber member. The distance requirements are the same as for screws in tension, see Tab. 1 and Fig. 4.

173


5.

Self-tapping screws as reinforcement to carry shear stresses in gluedlaminated timber and cross-laminated timber

5.1

Reinforcement of glued-laminated timber against shear stresses

The shear strength of timber is in the range of five times the magnitude of tension perpendicular to the grain strength. However, there can be applications in which the shear stresses exceed the shear capacity of a timber beam. Examples are double tapered or pitched cambered beams, where the changing depth leads to high shear stresses in the area of the supports, see Fig. 12. With respect to an economic use of reinforcing elements it is of interest, whether a proportionate distribution of shear stresses between the timber beam and the shear reinforcement can be achieved in the unfractured state. In [51] (see also [59] and [60]), an analytical approach is proposed to determine the loadcarrying capacity of timber beams in the uncracked state, featuring shear reinforcement like screws or threaded rods. This approach is based on common theoretical concepts and constitutive equations for material properties and enables the incorporation of the semi-rigid composite action between the reinforcement and wood material, as well as the interaction of shear stresses and stresses perpendicular to the grain. The applicability and accuracy of the approach is verified by laboratory tests, also taking into account research carried out and presented in [18]. It appears that the redistribution of load from the timber to the shear reinforcement is comparatively low. Considering the uncracked state, comparative calculations indicate that, under realistic construction conditions, an increase in shear capacity of up to 20 % is feasible.

Fig. 12 Pitched cambered beam – potential reinforcement zones (from [59]) Since wood is characterized by very brittle failure mechanisms both in shear and tension perpendicular to the grain, it is beneficial to design the corresponding reinforcements so that they are able to also carry the corresponding stresses in the fractured state, see e.g. [61]. The shear analogy ([62], [63], [64]) represents an applicable approach to calculate the semi-rigid composite action between both sections in the fractured state. A numerical study on highly stressed shapes of glulam beams (see [51] and [59]), featuring the minimum required reinforcement to carry the stresses that are released in the case of cracking shows, that the maximum increase in bending stresses between the intact state and the fractured state is in the range of one third.

5.2

Reinforcement of cross-laminated timber against shear stresses

The load-carrying capacity of cross-laminated timber (CLT) elements can be limited by the rolling shear strength, which is in the range of only 1/3 of the shear strength of softwood parallel to the grain, see Fig. 13. Possible examples are CLT elements loaded out of plane under high concentrated loads or point supports. Fully threaded, self-tapping screws used as shear reinforcement can have a distinct positive influence on the shear capacity of CLT elements. The reason is that CLT elements feature significantly larger shear deformations than GLT elements. This can mainly be explained with the low rolling shear modulus of the transverse layers (GR ≈ 0.1 ∙ G), see Fig. 13. This results in a larger share of the screws in the proportionate distribution of shear stresses between the timber beam and the shear reinforcement.

174


Fig. 13 Shear reinforcement of CLT and GLT – deformation and strength capacities, from [60] In [65] and [66] a design concept is proposed which is validated by means of experiments, see also [67]. The concept is based on a strut-and-tie model and takes into account the positive influence of the interaction of compression perpendicular to the grain and rolling shear stresses, by a factor kR,90, as well as the load-carrying capacity of the screws in direction of the potential fracture plane, see Eq. (10) and Fig. 14. In [67] it could also be shown that in the case of biaxial load transfer, additional effects are activated, leading to an increase in the rolling shear capacity compared to that of beam elements.

f R,k  k R,90  f R,k 

with:

Rax,k

2

(10)

2  a1  a2,ef

Rax ,k 2 1  0,35  c,90 k R,90  min  and c,90  1,20 2  a1  a2,ef  f R,k

characteristic load-carrying capacity of the reinforced CLT under stress

[N/mm 2 ]

f R,k

characteristic rolling shear capacity (according to technical approvals)

[N/mm 2 ]

Rax ,k

characteristic load-carrying capacity of a screw parallel to its axis

[N]

a1

distance of the screws parallel to the load-bearing direction

[mm]

a2,ef

effective distance of the screws perpendicular to the load-bearing direction

[mm]

l ef

effective embedment length of the screws for calculation of R ax,k

[mm]

k R,90

parameter for the consideration of the stress interaction

[-]

Fig. 12 Design concept for shear reinforcement in CLT on the basis of a strut and tie model, adapted from [65]

175


6.

Self-tapping screws to realize mechanically jointed beams

Fully threaded, self-tapping screws can also be used to strengthen existing beams by mechanically joining an additional cross-section to the existing beam. Screws feature significantly higher axial slip moduli, Kax,ser, compared to slip moduli in shear, Kser, see Fig. 15. Therefore they should be positioned at an angle to the joint (e.g. 45°), see Fig. 16. Models for calculation of these slip moduli can be found in e.g. [17], see Eq. (2) and [68] and [69].

Fig. 15 Comparsion of axial slip moduli Kax,ser (see Eq. (2)) and slip moduli in shear Kser (see [3]) for different screw diameters and lengths

Fig. 16 Example of two cross-sections, mechanically jointed by screws incl. schematic illustration of composition of shear stresses and longitudinal stresses A common, standardized method [3] to design such beams is the “gamma-method” [70]. The coefficient γ symbolizes the efficiency of the connection (γ = 1 → rigid connection, γ = 0 → no connection). It is used to determine an effective moment of inertia of the beam with which the stresses in the beam as well as the shear flow in the joint can be determined. For cases, in which the basic assumptions for the “gamma-method” are not met (e.g. continuous joint stiffness and cross-section, single-span or symmetric continuous beams), an alternative is given by the “shear analogy method” [62], [63] and [64]. The timber members to be connected by screws should be fixed in their position so that no gap occurs in the joint when the screws are driven in. If the screws are positioned so that they are loaded in axial tension, they could close a potential gap due to a rope-effect. Another option is to place the screws crosswise, i.e. one screw is mainly loaded in axial tension, and the other screw is mainly loaded in axial compression. The shear flow in the joint can be distributed to the screws using a triangle of

176

Reinforcement with self-tapping screws and threaded rods forces. When determining the joint stiffness of a mechanically jointed beam, the slip moduli of the threaded parts of the screws in both beams have to be accounted for. For common types of mechanically jointed beams with self-tapping screws set at an angle between screw axis and fibre direction of 45°, values in the range of γ = 0.8 can be obtained [8]-[10].

7.

Conclusions and remarks

Fully-threaded, self-tapping screws represent the latest developments in screwing technology for timber engineering, providing a significant load-carrying capacity if loaded in axial direction. The screws enable stiff connections but limited plastic potential. Apart from their obvious application as fasteners in timber connections, fully-threaded, self-tapping screws feature a high potential for numerous reinforcement applications. An overview of possibilities and related design procedures is given, including relevant literature for background information. For numerous applications, design procedures exist, which have already been clarified to an extent satisfying engineering needs, e.g. reinforcement taking the full tensile stresses perpendicular to the grain. Other applications are well developed but require some additional research, e.g. reinforcement approaches which consider proportional load sharing between the timber and the reinforcement, including the modelling of stiffness properties (Kser, Kax,ser, Ku), as well as reinforcement against compressive stresses perpendicular to the grain. However, there is a need for further efforts in research, e.g. considering the potentially harmful effect of reinforcement restricting the free shrinkage or swelling of the timber. Although fully threaded, self-tapping screws have undoubtedly a great potential for use in timber structures, should their potential application as reinforcement always be considered with care. For centuries, excellent timber structures have been designed without the need for reinforcement. Wood features a multitude of positive characteristics in view of its application as building material, i.e. a general reinforcement of timber elements – as known from concrete structures – is not necessary. Although recent developments have largely increased the range and forms of structures to be realized with timber products, good design of timber structures should still aim at minimizing stresses for which timber only features small capacities and brittle failure mechanisms (e.g. tensile stresses perpendicular to the grain and shear), thereby avoiding or at least minimizing the necessity for reinforcement.

8.

References

[1]

Steiger, R., Serrano, E. et al., “Glued-in rods”, in: Harte, A., Dietsch, P. eds. Reinforcement of Timber Structures, Shaker, Aachen, Germany, 2015 (this publication)

[2]

DIN 7998:1975-02, Threads and Thread Ends for Wood Screws, DIN, Berlin, 1975.

[3]

EN 1995-1-1:2004-11, Eurocode 5: Design of timber structures – Part 1-1: General – Common rules and rules for buildings, CEN, Brussels, 2004.

[4]

Mahlknecht, U., Brandner, R., Ringhofer, A., Schickhofer G., „Resistance and Failure Modes of Axially Loaded Groups of Screws“, in: Proceedings of the RILEM Timber Structures Conference „Materials and Joints in Timber Structures – Recent Advancement of Technology“, Stuttgart, Germany, 2013.

[5]

Pirnbacher, G., Schickhofer G., „Schrauben im Vergleich – eine empirische Betrachtung“, in:

177

Reinforcement of Timber Structures Proceedings of the 6. Grazer Holzbau-Fachtagung (6. GraHFT’07) „Verbindungstechnik im Ingenieurholzbau“, F 1–22, Graz, Austria, 2007. [6]

Dietsch, P., Winter, S., “Eurocode 5 – Future Developments towards a More Comprehensive Code on Timber Structures“, Structural Engineering International, Vol. 21, No. 2, 2012, pp. 223-231.

[7]

EN 14592:2012-05, Timber structures – Dowel type fasteners – Requirements, CEN, Brussels, 2012.

[8]

ETA-12/0062, SFS intec AG: Self-tapping screws for use in timber constructions, European Technical Approval, OIB, p. 17, 18.06.2012.

[9]

ETA-11/0112, SPAX International GmbH & Co. KG: Self-tapping screws for use in timber constructions, European Technical Approval, ETA-Danmark, p. 84, 05.09.2012.

[10] ETA-11/0190, Adolf Würth GmbH & Co. KG: Self-tapping screws for use in timber constructions, European Technical Approval, DIBt, p. 99, 27.06.2013. [11] Pirnbacher, G., Brandner, R., Schickhofer, G., „Base Parameters of Self-Tapping Screws“, CIBW18 / 42-7-1, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 42, Duebendorf, Switzerland, 2009. [12] Colling, F., Holzbau – Grundlagen und Bemessung nach EC5, 3. Auflage, Springer Vieweg, ISBN 978-3-8348-1789-1, 2012. [13] Ringhofer, A., Brandner, R., Schickhofer, G., „Entwicklung einer optimierten Schraubengeometrie für hochbeanspruchte Stahl-Holz-Verbindungen“, Bautechnik, Vol. 91, No. 1, 2014, pp. 31-37 [14] Enders-Comberg, M., Blaß, H.J., “Influence of fasteners in the compression area of timber members”, CIB-W18 / 46-7-8, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 46, Vancouver, Canada, 2013. [15] Ringhofer, A., Grabner, M., Brandner, R., Schickhofer, G., „SGSC 3.2.1_1 – Prüftechnische Ermittlung des Tragverhaltens der Einzelschraube in der BSP-Schmalfläche“, Internal Research Report, holz.bau forschungs gmbh, Institute of Timber Engineering and Wood Technology, Graz University of Technology, 2013. [16] Krenn, H., Schickhofer, G., „Joints with inclined screws and steel plates as outer members“, CIB-W18 / 42-7-2, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 42, Duebendorf, Switzerland, 2009. [17] Blaß, H. J., Bejtka, I., Uibel, T., „Tragfähigkeit von Verbindungen mit selbstbohrenden Holzschrauben mit Vollgewinde“, Forschungsbericht der Versuchsanstalt für Stahl, Holz und Steine, Abt. Ingenieurholzbau, Universität Karlsruhe (TH), 2006. [18] Blaß, H.-J., Krüger, O., Schubverstärkung von Holz mit Holzschrauben und Gewindestangen, Karlsruher Berichte zum Ingenieurholzbau, Band 15, Universitätsverlag Karlsruhe, 2010. [19] Brüninghoff, H., Schmidt, C., Wiegand, T., „Praxisnahe Empfehlungen zur Reduzierung von Querzugrissen bei geleimten Satteldachbindern aus Brettschichtholz“, Bauen mit Holz, Vol. 95, No. 11, 1993, pp. 928-937. [20] DIN EN 1995-1-1/NA:2010-12, National Annex – Nationally determined parameters – Eurocode 5: Design of timber structures – Part 1-1: General – Common rules and rules for buildings, DIN, Berlin, 2010. [21] DIN 1052:2008-12, Entwurf, Berechnung und Bemessung von Holzbauwerken – Allgemeine Bemessungsregeln und Bemessungsregeln für den Hochbau, DIN, Berlin, 2008.

178

Reinforcement with self-tapping screws and threaded rods [22] DIN 1052:2004-08, Entwurf, Berechnung und Bemessung von Holzbauwerken – Allgemeine Bemessungsregeln und Bemessungsregeln für den Hochbau, DIN, Berlin, 2004. [23] Blaß, H.J., Steck, G., „Querzugverstärkungen von Holzbauteilen: Teil 1 – Teil 3“, Bauen mit Holz, Vol. 101, No. 3, pp. 42-46, No. 4, pp. 44-49, No. 5, pp. 46-50, 1999. [24] Blaß, H.J., Bejtka, J., “Reinforcements perpendicular to the grain using self-tapping screws”, in: Proceedings of the 8th World Conference on Timber Engineering, Lahti, Finland, 2004. [25] Möhler, K., Siebert, W., „Untersuchungen zur Erhöhung der Querzugfestigkeit in gefährdeten Bereichen“, Bauen mit Holz, Vol. 86, No. 6, 1984, pp. 388-393. [26] Ehlbeck, J., Görlacher, R., Werner, H., “Determination of Perpendicular-to-Grain Tensile Stresses in Joints with Dowel Type-Fasteners”, CIB-W18 / 22-7-2, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 22, Berlin, West Germany, 1989. [27] Kreuzinger, H., „Holzbau“, in: Zilch, K., Diederichs, C.J., Katzenbach, R. (eds.), Handbuch für Bauingenieure, Springer, Berlin, 2002. [28] Mahlknecht, U., Brandner, R., „focus_sts 3.1.2_1: Untersuchungen des mechanischen Verhaltens von Schrauben – Verbindungsmittelgruppen in VH, BSH und BSP“, Internal Research Report, holz.bau forschungs gmbh, Institute of Timber Engineering and Wood Technology, Graz University of Technology, 2013. [29] Henrici, D., Beitrag zur Spannungsermittlung in ausgeklinkten Biegeträgern aus Holz, Dissertation, Technische Universität München, 1984. [30] Henrici, D., „Beitrag zur Bemessung ausgeklinkter Brettschichtholzträger“, Bauen mit Holz, Vol. 92, No. 11, 1990, pp. 806-811. [31] Tannert, T., Lam, F.; “Self-tapping screws as reinforcement for rounded dovetail connections”, Structural Control and Health Monitoring, Vol. 16, No. 3, 2009, pp. 374-384. [32] Gustafsson, P.J., “Notched beams and holes in glulam beams”, in: Blaß, H.J., Aune, P., Choo, B.S., et al. (eds). Timber Engineering STEP 1 – Basis of design, material properties, structural components and joints, Centrum Hout, Almere, 1995 [33] Neuhaus, H.; Ingenieurholzbau – Grundlagen, Bemessung, Nachweise, Beispiele, 3. Auflage, Vieweg&Teubner, ISBN 978-3-8348-1286-5, 2010. [34] Jockwer, R., Frangi, A., Steiger, R., Serrano, E., “Enhanced design approach for reinforced notched beams”, CIB-W18/ 46-6-1, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 46, Vancouver, Canada, 2013. [35] Steilner, M., Blaß, H.J., „Vorspannen von Holz mit Vollgewindeschrauben“, Proceedings 17. Internationales Holzbau-Forum, Garmisch-Partenkirchen, 2011. [36] Blaß, H.J., Bejtka, I., Querzugverstärkungen in gefährdeten Bereichen mit selbstbohrenden Holzschrauben, Forschungsbericht der Versuchsanstalt für Stahl, Holz und Steine, Abt. Ingenieurholzbau, Universität Karlsruhe (TH), 2003. [37] Aicher, S., Höfflin, L., “Glulam Beams with Holes Reinforced by Steel Bars”, CIB-W18 / 4212-1, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 42, Duebendorf, Switzerland, 2009. [38] Aicher, S., “Glulam Beams with Internally and Externally Reinforced Holes – Test, Detailing and Design”, CIB-W18 / 44-12-4, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 44, Alghero, Italy, 2011.

179

Reinforcement of Timber Structures [39] Kolb, H., Frech, P., „Untersuchungen an durchbrochenen Bindern aus Brettschichtholz“, Holz als Roh- und Werkstoff, Vol. 35, No. 4, pp. 125-134 [40] Kolb, H., Epple, A., Verstärkungen von durchbrochenen Brettschichtholzbindern, Schlussbericht zum Forschungsvorhaben I.4 – 34810, Forschungs- und Materialprüfungsanstalt Baden-Württemberg, Stuttgart, 1985. [41] Aicher, S., Höfflin, L., “Glulam Beams with Round Holes – a Comparison of Different Design Approaches vs. Test Data”, CIB-W18 / 35-12-1, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 35, Kyoto, Japan, 2002. [42] Aicher, S., Höfflin, L., “Design of rectangular holes in glulam beams”, Otto-Graf-Journal, Vol. 14, 2003, pp. 211-229. [43] Aicher, S., Höfflin, L., Reinhardt, H.-W., „Runde Durchbrüche in Biegeträgern aus Brettschichtholz, Teil 2: Tragfähigkeit und Bemessung“, Bautechnik, Vol. 84, No. 12, 2007, pp. 867-880. [44] Blaß, H.J., Ehlbeck, J., Kreuzinger, H., Steck, G., Erläuterungen zu DIN 1052:2004-08, Bruderverlag, Karlsruhe, 2004. [45] Blumer, H., Spannungsberechnungen an anisotropen Kreisbogenscheiben und Sattelträgern konstanter Dicke, Lehrstuhl für Ingenieurholzbau und Baukonstruktionen, Universität Karlsruhe, 1972/1979. [46] Ehlbeck, J, Kürth, J., „Influence of perpendicular-to-grain stressed volume on the load-carrying capacity of curved and tapered glulam beams“, CIB-W18 / 24-12-2, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 24, Oxford, United Kingdom, 1991. [47] Moser, M., Theoretische Betrachtung der Berechnungsgrundlagen für auf Querzug beanspruchte BSH-Bauteile, Master Thesis, Institute of Timber Engineering and Wood Technology, Graz University of Technology, 2012. [48] Dietsch, P., Winter, S.; Untersuchung von nicht in DIN EN 1995-1-1 geregelten Formen von Satteldachträgern im Hinblick auf den Nachweis der Querzugspannungen, Research Report, Lehrstuhl für Holzbau und Baukonstruktion, Technische Universität München, 2014. [49] Gowda, S, Ranta-Maunus, A., Curved and cambered glulam beams - Part 1: Short term load Tests, Research Notes 1500, Technical Research Centre of Finland, Espoo, 1993. [50] Dietsch, P., Gamper, A., Merk, M., Winter, S.; "Building Climate – long-term measurements to determine the effect on the moisture gradient in large-span timber structures", CIB-W18 / 4511-1, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 45, Växjö, Sweden, 2012. [51] Dietsch, P., Einsatz und Berechnung von Schubverstärkungen für Brettschichtholzbauteile, Dissertation, Technische Universität München, 2012. [52] Dietsch, P., Kreuzinger, H., Winter, S.; “Effects of changes in moisture content in reinforced glulam beams”, in: Proceedings of the 13th World Conference on Timber Engineering, Quebec, Canada, 2014. [53] Wallner, B., Versuchstechnische Evaluierung feuchteinduzierter Kräfte in Brettschichtholz verursacht durch das Einbringen von Schraubstangen, Master Thesis, Institute of Timber Engineering and Wood Technology, Graz University of Technology, 2012. [54] Jorissen, A.J.M., Double shear timber connections with dowel type fasteners, Dissertation, Delft University of Technology, Netherlands, 1998. [55] Bejtka, I., Blaß, H.J., “Self-tapping screws as reinforcements in connections with dowel-type fasteners”, CIB-W18 / 38-7-4, Proceedings of the international council for research and

180

Reinforcement with self-tapping screws and threaded rods innovation in building and construction, Working commission W18 – timber structures, Meeting 38, Karlsruhe, Germany, 2005. [56] Bejtka, I., Verstärkung von Bauteilen aus Holz mit Vollgewindeschrauben, Dissertation, Band 2 der Reihe Karlsruher Berichte zum Ingenieurholzbau, KIT Scientific Publishing, Karlsruhe, 2006. [57] Kobel, P., Modelling of strengthened connections for large span truss structures, Master Thesis, Institute of Structural Engineering, ETH Zürich, Switzerland & Division of Structural Engineering at Lund University/LTH, Sweden2011. [58] Bejtka, I., Blaß, H.J., “Self-tapping screws as reinforcements in beam supports”, CIB-W18/ 397-2, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 39, Florence, Italy, 2006. [59] Dietsch, P., Kreuzinger, H., Winter, S., “Design of shear reinforcement for timber beams”, CIBW18/ 46-7-9, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 46, Vancouver, Canada, 2013. [60] Dietsch, P., Mestek, P., Winter, S., "Analytischer Ansatz zur Erfassung von Tragfähigkeitssteigerungen infolge von Schubverstärkungen in Bauteilen aus Brettschichtholz und Brettsperrholz", Bautechnik, Vol. 89, No. 6, 2012, pp. 402-414. [61] Dietsch, P., "Robustness of large-span timber roof structures – Structural aspects", Engineering Structures, Vol. 33, No. 11, 2011, pp. 3106-3112. [62] Kreuzinger, H., „Platten, Scheiben und Schalen – ein Berechnungsmodell für gängige Statikprogramme“, Bauen mit Holz, Vol. 101, No. 1, 1999, pp. 34-39. [63] Kreuzinger, H., „Verbundkonstruktionen“, in: Holzbaukalender 2002, Bruderverlag, Karlsruhe, 2001, pp. 598-621. [64] Kreuzinger, H., “Mechanically Jointed Beams: Possibilities of Analysis and some special Problems”, CIB-W18 / 34-12-7, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 34, Venice, Italy, 2001. [65] Mestek, P., Kreuzinger, H., Winter, S., “Design Concept for CLT – Reinforced with SelfTapping Screws”, CIB-W18/ 44-7-2, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 44, Alghero, Italy, 2011. [66] Mestek, P., Winter, S., „Punktstützung von Brettsperrholzkonstruktionen – Schubverstärkungen mit Vollgewindeschrauben“, Bauingenieur, Vol. 86, No. 12, 2011, pp. 529-540. [67] Mestek, P., Punktgestützte Flächentragwerke aus Brettsperrholz (BSP) – Schubbemessung unter Berücksichtigung von Schubverstärkungen, Dissertation, Technische Universität München, 2011. [68] Kevarinmäki, A., “Joints with inclined screws”, CIB-W18 / 35-7-4, Proceedings of the international council for research and innovation in building and construction, Working commission W18 – timber structures, Meeting 35, Kyoto, Japan, 2002. [69] Tomasi, R., Crosatti, A., Piazza, M. "Theoretical and experimental analysis of timber-to-timber joints connected with inclined screws", Construction and Building Materials, Vol. 24, 2010, pp. 1560-1571. [70] Möhler, K., Über das Tragverhalten von Biegeträgern und Druckstäben mit zusammengesetzten Querschnitten und nachgiebigen Verbindungsmitteln, Habilitation, Universität Karlsruhe, 1956.

181


182

FRP reinforcement of timber structures

10 FRP reinforcement of timber structures Kay-Uwe Schober1, Annette M. Harte2, Robert Kliger3, Robert Jockwer4, Qingfeng Xu5, Jian-Fei Chen6

Summary Timber engineering has advanced over recent decades to offer an alternative to traditional materials and methods. The bonding of fibre reinforced plastics (FRP) with adhesives to timber structures for repair and strengthening has many advantages. However, the lack of established design rules has strongly restrained the use of FRP strengthening in many situations, where these could be a preferable option to most traditional techniques. A significant body of research has been carried out in recent years on the performance of FRP reinforced timber and engineered wood products. This chapter gives a State of the Art summary of material formulations, application areas, design approaches and quality control issues for practical engineers to introduce on-site bonding of FRP to timber as a new way in design for structural repair and rehabilitation.

1.

Introduction

Fibre reinforced polymer materials combining high strength fibres and a resin matrix have a wide variety of industrial applications due to their high strength-to-weight ratio and ease of handling. Their versatility is reflected in the construction industry where they have been widely used for many years, especially for strengthening of concrete structures [1],[2]. More recently, the techniques have also been extended to timber structures [3]-[6]. FRP reinforcement in the form of pultruded rods or plates and woven fabrics has an expanding and particularly effective application in structural repair of timber structures. They may be inserted at critical locations for enhanced loading capacity, or for load transmission when damaged beam-ends are cut off and replaced with new timber or epoxy mortar. The reinforcement techniques are usually based on the use of adhesives on site, and use procedures that are common for the repair or the

1)

Professor of Timber Engineering and Structural Design, Mainz University of Applied Sciences, Mainz, Germany 2)

Senior Lecturer in Engineering, National University of Ireland, Galway, Ireland

3)

Professor of Structural Engineering, Chalmers University of Technology, Gothenburg, Sweden

4)

Postdoctoral Researcher in Timber Engineering, ETH Swiss Federal Institute of Technology, Zurich, Switzerland 5)

Deputy Chief Engineer, Shanghai Research Institute of Building Sciences, Shanghai, China

6)

Professor of Civil and Structural Engineering, Queens University Belfast, Belfast, Northern Ireland, UK

183

Reinforcement of Timber Structures upgrading of concrete and metallic structures. Such techniques minimize the disturbance to the building and its occupants during the intervention. However, some concerns have prevented the wider use of adhesives, particularly in historical timber structures, where sufficient reliability cannot yet be guaranteed. One reason is that a long service life has not yet been fully proven for synthetic adhesives, since the oldest bonded joints are only around sixty years and greater ages cannot be simulated by existing accelerated ageing tests. Fibre reinforced polymers and structural adhesives have been used to repair or strengthen structural members for many years. This approach is nevertheless difficult to extend to less common wood species, less favourable environmental conditions, new adhesive formulations and other variables. The lack of established design rules available to engineers and other decision makers has significantly restrained the use of FRP strengthening techniques in many situations where these could otherwise be a preferable option to most traditional techniques or to the total replacement of timber members. In this chapter, the materials used in FRP reinforcement of timber structures are discussed, a design approach for bondline delamination is presented, current and potential applications of FRP for reinforcement of timber structures are described, design rules are outlined and finally relevant quality control procedures for on-site bonding are summarised.

2.

Materials

2.1

FRP reinforcement materials

FRP materials are composites comprising fibres that provide the load-bearing capacity and stiffness, embedded in a polymeric resin that transfers loads between fibres and provides protection for the fibres. They are available in a wide variety of forms, and have properties that vary considerably depending on the fibre material, volume fraction and orientation. Typical properties of the common fibres and polymers are given in Tab. 1 Tab. 1 Fibre and polymer properties [1] Material

Modulus of Elasticity [GPa]

Tensile Strength [MPa]

Failure Strain [%]

CTE [10-6 °C-1]

Density [g/cm3]

E-glass

70-80

2000-4800

3.5-4.5

5.0-5.4

2.5-2.6

Carbon (HM)

390-760

2400-3400

0.5-0.8

-1.45

1.85-1.90

Carbon (HS)

240-280

4100-5100

1.60-1.73

-0.6 - -0.9

1.75

Aramid

62-180

3600-3800

1.9-5.5

-2.0

1.44-1.47

Basalt

82-110

860-3450

5.5

3.15

1.52-2.7

Polymer

2.7-3.6

40-82

1.4-5.2

30-54

1.10-1.25

CTE: Coefficient of thermal expansion; HM: High modulus; HS: High strength

For structural reinforcement, two main forms of FRP are generally used, namely, pultruded rods or plates and fabrics. For internal reinforcement, pultruded rods and plates are bonded into slots or grooves formed in the timber element. For external reinforcement, FRP plates or fabric materials are used.

184

FRP reinforcement of timber structures 2.2

Adhesives

The reinforcement of timber with FRP is normally implemented by adhesive bonding. For bonding of pultruded FRP plates, the adhesive is applied to the timber substrate and the FRP is then applied to the adhesive under pressure. FRP fabric reinforcement is generally applied using a wet lay-up method, whereby the fabric is impregnated with the adhesive first and then this is applied to the timber. Epoxy based adhesives have been used in most cases for on-site repair jobs, most formulations were developed for other materials. These adhesives are generally too rigid for bonding timber and there is no chemical bonding or suitable mechanical anchorage in wood. The bond surface is prone to failure because of dimensional changes in the wood induced by moisture content variations, even for indoor applications such as those defined under Eurocode 5 Service Class 2. More recently, epoxy formulations have been developed specifically for use with wood. On-site application of adhesives may be difficult and the quality of adhesive bond is not easy to evaluate. Since properties of reinforced elements very much depend on construction quality, relevant procedures for applying and controlling adhesive quality[9],[10],[10] have to be followed. The selection of the adhesive for bonding of FRP to timber has to be undertaken with great care. The adhesive must be capable of bonding with both the FRP and timber and should have adequate strength. Four categories of adhesive are available: epoxies, polyurethanes, polyesters, phenolics and aminoplastics [11]. While many of these adhesive types have been shown to provide satisfactory bonding performance when used in a controlled environment [12],[16], two-part cold-cure epoxy adhesives have generally been found to be most suitable for on-site bonding as they have good gapfilling properties, are thixotropic and have low curing shrinkage. Additional information on adhesives for on-site bonding: characteristics, testing, applications can be found in [17]. A wide variety of epoxy formulations is available and compatibility with the adherents must be verified by the manufacturer. On site, the moisture content and the surface quality of the timber are difficult to control and it may not be possible to achieve the desired bonding pressure. Thick bond lines in the order 1-3 mm are common. Careful surface preparation is essential in order to achieve good bond strength and durability. Surfaces to be bonded should be dry, free from contaminants, such as release agents and dust, and have sufficient surface roughness. For the FRP this involves abrasion followed by solvent cleaning or else removal of the peel ply if one is provided. Surface preparation should be carried out immediately prior to bonding while preparation and application of the adhesive should be in accordance with the manufacturer’s instruction. The importance of using experienced operatives cannot be underestimated. Quality control measures should be implemented at each stage of the process [18].

2.3

Bond behaviour of FRP-timber interface

Central to successful reinforcement is the integrity of the bond between the FRP and timber substrate [1],[19]. Over the last two decades, a number of studies were undertaken by different researchers on bond behaviour of FRP-to-wood and many useful results were acquired, including failure modes, stress distribution and local bond-slip relationship. In the early studies on the behaviour of FRP to wood bonds, the test configuration, shown in Fig. 1(a), was a modified form of the ISO 6238 and ASTM D905-03 block shear test [12],[13],[20]-[23]. In those tests, the shear strength of the bond was derived as an average stress over the bonded plate length. As the surface of the FRP was sandwiched between two pieces of timber, FRP plate surface strains were difficult to monitor, not to mention the shear stress distribution and bond-slip responses.

185

Reinforcement of Timber Structures SILVA et al. [24] conducted four-point bending tests on timber-FRP joints with the strengthening techniques of near-surface mounted (NSM) and externally bonded reinforcement (EBR) which was convenient to get the FRP strain distribution, shear stress distribution and bond-slip responses. On the basis of SILVA’s tests, Juvandes and Barbos [25] analyzed the maximum anchor strength of the composite and the maximum composite strain, and proposed the effective bonding length for the EBR and NSM reinforcements Wan et al. [26] introduced single-lap FRP-to-timber joint shear tests, shown in Fig. 1(b), in which the strength of the bond between the FRP and timber was examined. The shear tests were conducted on softwood (Pine) which was strengthened with carbon FRP with the main test variables being the FRP bond length and the growth characteristics of the timber. Extensive strain gauging of the FRP has enabled the onset and propagation of debonding cracks to be monitored. Failure modes of the joints and effective bond length were identified based on the test results. Wan et al. [19] also conducted a series of tests on 86 single-shear FRP-to-timber joints to study on FRP-to-timber bonded interfaces. The test parameters included adhesive type, FRP plate type and timber species. Test results showed that all softwood joints failed predominantly in the timber while the hardwood joints exhibited failure at different interfacial positions. Load-slip, strain and bond stress distribution, and bond stress-slip responses were consistent with those of FRP-to-concrete bonds. A test-based theoretical bond stressslip law utilizing the J-integral method was proposed and could be implemented in analytical and numerical models. wood

FRP b)

a) Fig.1 (a) Shear block test configuration; (b) single lap joint shear test The FRP composites mentioned above were all FRP sheets or strips. There was much less research on bond behaviour of wood and FRP bars. Lorenzis et al. [27] conducted pull-out tests to study the bond performance of FRP rods epoxied into glulam timber. The test variables were bonded length, surface configuration of the rod and direction of the wood fibres with respect to the longitudinal axis of the joint. The observed failure modes were cohesive failure in timber for rods glued-in parallel to the grain and adhesion failure for rods glued-in perpendicular to the grain. In addition to the experimental investigations, finite element simulation on bond behaviour of FRP-totimber was developed in recent years. Valipour and Crews [28] proposed a novel force-based element in the framework of the total secant approach for nonlinear analysis of timber beams strengthened with FRP sheet (bar), including bond-slip effects. The formulation takes account of material nonlinearities and preserves the continuity of slip shear. Further, a composite Simpson integration scheme with a finite difference scheme was employed to calculate the bond shear forces along the element. It was concluded that for the considered cases, bond shear-slip between the FRP sheet (bar) and timber has a minor effect on the ultimate loading capacity as well as the load-deflection response

186

FRP reinforcement of timber structures of timber beams strengthened with FRP sheets (bars). The assumption of perfect bond between the FRP and timber beam was observed to be acceptable in most cases. In summary, the research on the bond behaviour of FRP-to-timber is still in its infancy. Further investigations should be carried out to determine the influence of the various factors affecting the bond between FRP and wood including timber properties and specimens, thicknesses (diameters) and specimens of FRP sheets (bars), and thickness of adhesive. The local bond-slip relationship for FRPto-timber which can be directly used for finite element analysis is definitely the research focus in the future.

2.4

FRP damage and delamination by loads

2.4.1 Global fracture criteria for composite design A realistic design approach to account for the fracture and delamination in FRP strengthened timber structures needs appropriate material models describing physically-based failure criteria of the anisotropic and non-linear properties of these composite structures. Global fracture criteria with complete stress interaction were the first failure criteria developed [29]. The main advantage is the ease of use for analytical, numerical and design approaches due to a single scalar equation of failure for unidirectional laminates. These hypotheses contain no information about the fracture mode of composites in three-dimensional stress states; the fibre-parallel fracture plane is unknown. Stress combinations like (21, 2) and (31, 3) are set to equal and do not consider the importance of loads perpendicular to the fibre within the fibre plane which are critical due to the dimensions of most structural elements. The most common criterion used in FE-codes has been developed by TSAI and WU, shown in Eq. (1). The symmetry of material properties is considered by strength coefficients which can be easily transformed in their invariances.

F

D0 D D 2 2 2  1  90  2  90  3  1  2  3 F11 F22 F22 F11 F22 F22  F12 1 2  F12 1 3  2 F23 2 3 

 122 f v2,90,0



 132 f v2,90,0



 232 f v2,90,90

(1)

1

with

F11  1  f t ,0 f c ,0 

F1  1 f t ,0  1 f c ,0

F22  1  f t ,90 f c ,90 

F2  1 f t ,90  1 f c ,90

F12  0

2 F23 

D0  f c ,0  f t ,0

D90  f c ,90  f t ,90

2 1  f c ,90 f t ,90 f v2,90,90

ft,0

tensile strength in fibre direction

fc,0

compressive strength in fibre direction

ft,90

tensile strength perpendicular to fibre direction

fc,90

compressive strength perpendicular to fibre direction

fv,90,0

shear strength in laminate area

fv,90,90

shear strength perpendicular to laminate area

187

Reinforcement of Timber Structures 1, 2, 3 are normal stresses and 23 = 32, 13 = 31, 12 = 21 shear stresses related to the (xl, x2, x3) coordinate system of the unidirectional laminate area, where (xl, x2)-area describes the laminate area and x3 the thickness direction (Fig. 2).

Fig. 2 Stresses in unidirectional laminates With this hypothesis most of fracture modes can be described but fibre fracture (FF) and interlaminar fibre fracture (IFF) cannot be identified. Furthermore, the strength increase cannot be represented by a maximum criterion but only by an interactive criterion. Therefore, for FRPs the combination of the general Tsai-Wu criterion with the hypothesis of maximum stresses is recommended, otherwise stress states greater than the single value of the strength will be accepted.

2.4.2 Failure-mode based strength criteria for composite design Each failure mechanism is governed by one mode-associated strength. Failure conditions are required to be simply formulated, numerically robust, and physically-based and to allow a simple determination of the highest mode stress effort (design driving). The most important investigations have been made by Puck in a further development of the classical Mohr theory and Hashin’s failure criteria [30]. The key statement in his proposal is that failure in a fibre parallel fracture plane under tension perpendicular to the plane is caused by the tension stresses and also by the shear stresses nt and n1 (see 2), while compression forces perpendicular to the fibre plane increase the shear resistance. A further enhancement of this criterion was the introduction of an additional term of the internal Coulomb friction to account for stress stiffening below the fracture limit due to compression. For the case of tension reinforcement, the general fracture criterion of Puck is given by an easy to modify 7parameter model identifying FF and IFF for design issues. A more sophisticated version has been developed in the form of the mathematically simpler Puck-Knaust-IFF-failure criteria [30].

F

 22   32   2 3   232 f t ,90 f c ,90

 1 1   212   312     2   3   2 f v ,90,0  f t ,90 f c ,90 

 1 1 1    12    1 f t ,0, IFF f c ,0, IFF  f t ,0, IFF f c ,0, IFF 

(2)

F 1 

with F() = fracture plane

2.4.3 Practical application for engineering problems The definition of failure is an issue during failure analysis. In particular, initial failure like delamination has to be defined in the right way. The classification of failure criteria regarding their original aim has to be assessed before their use:

188

FRP reinforcement of timber structures  Analysis range (first ply or post failure)  Physical model (micro, macro or component level)  Mathematical approach (limit, polynomial tensor or physically based) One of the biggest problems for circulation of knowledge on failure of FRP is the non-availability of respective post-processing software codes. The Tsai-Wu criterion is implemented in most software e.g. ANSYS® or ABAQUS® being able to analyze the composite structures. Due to the same mathematical layout of the parabolic criterion of the Puck-Knaust and the Tsai-Wu criterion in form of tensor polynomials it is possible to implement the Puck-Knaust criterion in commercial FE-codes for an appropriate design of composite structures by enhancing the included failure criteria. The complete mathematical expression of the modification and implementation for engineering design problems is given in Tab. 2.

Tab. 2 Strength indices for Puck-Knaust modified Tsai-Wu criteria for FRPs Strength index

FF modification

IFF modification

Strength index

FF modification

IFF modification

ft,x

ft,0

1,3 ft,0

fv,xy

> 104 MPa

fv,0,90

fc,x

fc,0

1,3 fc,0

fv,yz

> 104 MPa

ft,y

> 104 MPa

ft,90

fv,zx

> 104 MPa

fv,90,90

fc,y

> 104 MPa

fc,90

Cxy

0

0

ft,z

> 104 MPa

ft,90

Cyz

0

–1

fc,z

> 104 MPa

fc,90

Czx

0

0

1/ 3 f t ,90 f c ,90

Due to the mathematically similar definition of the Tsai-Wu criteria and the Pick-Knaust criteria by tensor polynomials, the implemented algorithm in FE-codes can be modified using Tab. 2 for the specific of fiber reinforced polymers and own stress and failure analyses.

2.4.4 Numerical aspects of bond-line delamination Decohesion along interfaces plays an important role in a wide variety of failure processes in structures when using chemical bonding as the optimal form of combining two surfaces with each other. The analysis of delamination initiation is based on stresses and interaction criteria of the interlaminar stresses in conjunction with a characteristic distance as a function of geometry and material properties. Crack propagation is predicted by a fracture mechanics (FM) approach. This avoids the difficulties associated with a stress singularity at the crack front but requires the presence of a preexisting delamination. When used in isolation, neither the strength-based approach nor the FM approach is adequate for a progressive delamination failure analysis. For the combination of anisotropic materials like wood with FRP reinforcement in an interface damage law, the interaction between shear and normal stresses has to be considered. This problem can be solved by using a delamination analysis and an exponential interface damage law [32]. The definition of traction and separation depend on the finite element and the applied material model.

189

Reinforcement of Timber Structures Decohesion response is specified in terms of a surface potential () relating the interface tractions and the relative normal and tangential displacements n and t across the interface. In most of the computations all cohesive surfaces are taken to have identical cohesive properties which simplifies the surface potential [33] to

    2  ( )  e  c  n 1  exp   n  t2    n  t   e

Euler constant

c

maximum normal traction at the interface

n

normal separation where the maximum normal traction is attained with t = 0

t

shear separation where the maximum shear traction is attained

(3)

The potential leads by derivation on the displacements to the stresses if  < max where the traction components T are coupled to both normal and tangential crack opening displacements. The main advantage of the cohesive zone modelling is that, when it is known where fracture may occur a priori, a cohesive zone may be placed anywhere along element interfaces in that area, to take these effects into account. Furthermore, using decohesion elements, both onset and propagation of delamination can be simulated without previous knowledge of crack location and propagation direction and is therefore suitable for structural design and evaluation of composite beams.

2.5

FRP damage by fire

When FRP is subjected to elevated temperatures, the strength and stiffness are reduced due to the loss of mechanical integrity of the polymer matrix [3]. The two-part epoxy adhesives normally used in bonding have glass transition temperatures less than 100°C. For externally bonded FRP reinforcement, the exposure to high temperatures in a fire situation can therefore lead to both loss in the reinforcing effect of the FRP and degradation of the bond between the FRP and timber. In order to prevent this situation, the innate insulating properties of timber can be used to protect the reinforcement by providing adequate timber covering to the FRP. Martin and Tingley [34] undertook a test program in which FRP reinforced and unreinforced control glulam beams were subjected to standard fire exposure and mechanical loads equal to or in excess of the design loads. The beams were reinforced with three different types of FRP plates, bonded either externally to the soffit or between the bottom two laminations. The results showed that internally reinforced beams had a 44% higher fire endurance compared to the externally reinforced specimens. No difference in fire performance was found between the different types of FRP. Williamson [35] showed that glulam beams with externally bonded FRP can be designed to achieve a 1-hour rating. Fire exposure testing on glulam beams with beam end repairs comprising short epoxy resin replacement beams attached to the main beams by means of steel reinforcing bars [36] showed that these repairs could give one hour fire resistance. In this case, the steel bars had a minimum cover of 60 mm to the surface.

190

FRP reinforcement of timber structures

3.

Applications

3.1

Beam end reinforcement

In order to make the interventions minimally intrusive, the connecting elements are confined to the corners of the beam and their length is reduced to the minimum necessary. The operating procedures depend on the specific requirements of the site. This starts with the propping of the beams followed by removal of the decayed portion of the timber, usually terminated in an inclined cut (3). The new section of wood is shaped to match the external dimensions of the original decayed section. After the preparation of internal holes or external grooves between the original and newly introduced wood for the positioning of the connecting elements (4), the holes or groove are partially filled with adhesive and the reinforcing elements are inserted. The last stage is the insertion of a final wood fillet, to hide the grooves (5) followed by removal of the beam supports after the complete curing of the adhesive.

Fig. 3 Inclined cut-decayed end

Fig. 4 Linking grooves

Fig. 5 Insertion of final wedge

Epoxy resins have been used in specific cases to repair timber that has deteriorated from decay or insect attack and certain structural deficiencies in existing construction. The compressive strength and filling capabilities of epoxies can aid in repairing timber structures. The tensile, shear, and bond strength of epoxies as structural wood adhesives are, in certain cases, limited and are further subject to variability due to conditions of use. Mechanical reinforcement should be used in conjunction with epoxies for repairs intended to develop shear capacity. For structural rehabilitation and restoration the decayed ends are cut off and replaced by timber or polymer concrete where the shear and tension forces resulting from loads are transmitted by GFRP reinforcement with a PU or EP resin filling compound between the new and the existing part (Fig. 6).

Fig. 6 Shear reinforcement with GFRP

In some cases the decayed part of the structure will be replaced by a polymer concrete (PC) supplement using rapid-setting organic polymers as binders and small grade aggregates such as sand or gravel. The application of PC in structural rehabilitation and restoration has several advantages. Historically significant structures can be protected by minimum disturbance of the construction and minimum replacement of decayed parts. The appearance of the timber structure will not be changed. The section design can be easily executed by timber formwork on the level of the necessary construction height. All work can be undertaken from the top side so suspended ceilings will remain unaffected. The floor below the reconstruction work can be used with some restrictions. The method is suitable for inhabited floors and restoration of complicated timber joints. The full load-carrying capacity is achieved after one day with an increase of the structural performance by 140% for timber on-timber supplements and 30% for timber-on-PC supplements as reported by Schober et al. [32],[33].

191

Reinforcement of Timber Structures 3.2

Tension reinforcement perpendicular to the grain

The tensile strength of the timber perpendicular to the grain is considerably lower compared to the tensile strength parallel to the grain. High tension stresses perpendicular to the grain stresses occur at notches, holes or curved beams and beams with variable height. Efficient reinforcement techniques are required for increasing or maintaining the load-carrying capacity of these structures. The reinforcement should offer high strength and stiffness and should create a more ductile failure of the structure. Reinforcement made of FRP offers high strength and stiffness along its fibre direction; however, it commonly does not lead to more ductility of the structure. There is relatively little information on research and applications regarding FRP reinforcement of regions with tension perpendicular to the grain reported in literature due to the existence of other reinforcing techniques offering more ductility, like e.g. self-tapping screws. In general, the reinforcement should be applied close to the region of high tensile stresses perpendicular to the grain. As discussed by Hallström and Grenestedt [37], the use of internal reinforcement may be necessary for internal members that are not accessible at their external sides. For wrapping of FRP fabrics around the timber member good accessibility of the member at all sides is crucial. Studies on glued laminated timber (glulam) beams with holes reinforced by means of glass fibre reinforced polymers (GFRP) are reported by in Hallström and Grenestedt [37],[38]. It was aimed at reducing the stress singularity in the corner of the hole by changing the anisotropy by means of the reinforcement. However, the theoretical decrease of stresses at the corner of the hole was considerably lower than the level of increase of load-carrying capacity observed in the tests. The effect of load redistribution from the timber to the GFRP was confirmed in comparative FE simulations. Results of tests on notched glulam beams reinforced by means of GFRP plates are reported by Coureau et al. [39]. The reinforced notched beams showed higher load-carrying capacity compared to reference tests on unreinforced beams. Higher load-carrying capacities were achieved in the tests with an increase in the width of the FRP plates. Failure of the reinforced notched beams was accompanied by delamination of the FRP plates. Debonding of the FRP laminates was also observed in tests performed by Jockwer [40]. In the case of perpendicular to the grain reinforcement, debonding occurred in the lower beam section due to insufficient bond line area, whereas in the tests reinforced with an angle of 45° to the grain debonding occurred in the upper beam section near the support due to the high relative stress difference between the timber in compression perpendicular to the grain and the CFRP in tension in fibre direction. In both types of configuration of the CFRP, it was not possible to prevent crack initiation in the timber at the notch corner by the reinforcement. The two configurations are illustrated in Fig. 7.

Fig. 7 FRP reinforcement of notched beams

192

FRP reinforcement of timber structures Tests on curved and pitched cambered beams reinforced by means of GFRP are reported by Enquist et al. [41]. The reinforcement perpendicular to the grain led to a considerable increase in load-carrying capacity compared to unreinforced beams. Tensile failure in the GFRP and debonding occurred in the tests. Radial reinforcement repairs for curved bending members can also be accomplished by placing GFRP rods in oversized holes filled with epoxy PC. Here, the radial stresses are transmitted through the epoxy PC in shear. The need for further research on the application of GFRP as reinforcement perp. to grain includes [41]:  Long-term tests to cover the effects from variation in moisture and ambient conditions  Tests on full-size beams in order to cover scale effects.

3.3

Bending reinforcement

The reinforcement of timber members in bending using pultruded FRP rods, strips, plates or other structural shapes and fabric wraps has been the subject of a large number of research programmes during the last 50 years. The reinforcement can be deployed internally by bonding rods or strips in grooves cut into the tension and compression faces of the member or externally by bonding FRP plates to the tension face but not on the compression face due to the risk of buckling failure in the FRP. Experimental campaigns to investigate the reinforcement of solid timber beams and glulam with CFRP reinforcement by several researchers [42]-[47] have demonstrated that the use of a small percentage of reinforcement in the order of 1.5-2.5% can result increases in the bending strength and stiffness of up to 90% and 100%, respectively. In addition, reinforced beams were shown to have less variability in their properties than unreinforced beams. With increasing amounts of tensile reinforcement, the loaddeflection response becomes ductile due to compression yielding at high strain levels. Due to the high costs associated with CFRP materials, several investigators [47]-[50] have studied the flexural reinforcement of timber members with GFRP. As for the case of CFRP, large increases in strength were reported when using small percentages of reinforcement. However, the increases in stiffness were less significant due to the lower stiffness of the GFRP material. Decreased variability in the performance of the reinforced members was found. Limited ductility was achieved, which increased with increasing percentage of reinforcement. More recently, flexural reinforcement of timber using basalt fibre reinforced polymer (BFRP) has been undertaken with promising results [52].

3.4

Shear reinforcement

Various studies on the reinforcement and repair of regions with high shear stresses in glulam beams are reported in the literature. The studies were often performed more in order to compare different reinforcement techniques and to evaluate the reinforcing effect rather than to validate or verify design procedures. In general the tests can be roughly separated whether the reinforcing was applied externally or internally. External reinforcement was commonly made by means of FRP plates, fabrics or roving. Internal reinforcement was commonly made by FRP rods. The different techniques and relevant parameters are shown below.

V

β

V

Fig. 8 Externally applied FRP plates

V

β

a1

V

Fig.9 Internal applied FRP rods

193

Reinforcement of Timber Structures A good summary of the literature regarding shear strengthening of timber beams by means FRP is given by Andre [4]. The main focus in this work is put on the properties of different fibre materials. The behaviour of mechanically connected GFRP plates as shear reinforcement of timber beams was evaluated by Akbiyik et al. [53]. However, it is not possible to restore the stiffness of the undamaged beams by this reinforcement method. An approach for the calculation of the stresses in the timber of a beam reinforced by means of FRP panels is proposed and evaluated by Triantafillou [5],[54]. The approach is verified in a series of experiments on small size specimens. The main focus of the study was the required area fraction of the FRP panels and the relative height of the reinforcement in order to reduce the shear stresses in the timber. The approach is valid only for undamaged timber beams with no checks in the cross-section and, hence, not adequate for the repair of structures. As discussed in [5],[54] the reinforcing effect of the FRP depends on its stiffness in the grain direction of the timber. Hence, the use of multidirectional FRP fabrics or unidirectional FRP roving with small angle β relative to the grain direction of the timber are beneficial both for strength and stiffness of the reinforced beam. The advantages of internal reinforcement by means of FRP rods are the good aesthetics of the wooden surface of the beam and the possibility of reinforcing beams with reduced accessibility from the sides. Hence, FRP rods are a good method for repairing beams in situ [55]. A large test series on damaged beams of railway bridges reinforced by means of GFRP rods installed with β = 90° is reported by Radford et al. [6] and Burgers et al. [56]. A considerable increase of the shear stiffness of damaged beams was observed in the tests. Compared to external FRP wraps internal FRP rods show a lower stiffness but also less material use and an easier installation. The latter point is of special importance if beams are only accessible from the bottom surface. GFRP rods with β = 90° were tested in small and medium scale glulam beams by Gentry [57]. A general good reinforcing effect and an increase of approximately 50% of the load-carrying capacity in shear were observed. The method of reinforcing beams by means of FRP rods with β = 90° is also referred to as the Z-spiking method. Tests on glulam beams reinforced with GFRP dowels in shear and bending were performed by Svecova and Eden [58]. The internal shear reinforcement was applied with different distance a1 between the bars partly along the whole length of the beam and partly only in the region of high shear force near the support. The rods were installed perpendicular to the grain at β = 90°. It was possible to increase the load-carrying capacity, to reduce the variability of the beams, to increase the deflection before failure and hence, introduce some kind of ductility to the beams. A dowel distance equal to the beam depth is recommended by the authors based on tests. An example of an application of CFRP wraps in practice was given by Lauber [58]. A rafter was reinforced in shear by wrapping it with CFRP fabric in a U-shape. The advantage of fabric compared to plates is better flexibility during installation and its possibility to adapt to other connecting elements. An exact surface preparation of the existing beam is necessary in order to smooth the surface and to create good bond between timber and CFRP. It is emphasized that constant ambient conditions and constant moisture content of the structure are required in order to prevent delamination and to guarantee the long term strength. Sonti and Gangarao [60] found FRP wraps to be adequate for increasing the strength and stiffness of timber beams in infrastructural application. Widmann et al. [61],[62] performed a test series on full size beam specimen with shear cracks reinforced by means of CFRP roving with β = 45°. A considerable increase of the stiffness of the damaged beam was observed at that fibre angle.

194

FRP reinforcement of timber structures 3.5

Pre-stressed FRP

The reinforcing efficiency of FRP materials can be improved by prestressing. In passive or slack reinforcement, the load-carrying capacity of the FRP is often not reached as failure commonly occurs in the timber element. In the case of tensile reinforcement of beams, a tensile force is usually applied to the FRP sheet or rod by means of hydraulic jacks before pressure bonding to the timber element. The eccentric prestress induces significant compressive stresses in the bottom of the beam, which oppose the tensile stresses due to the external loads. In this way, the bending strength of the beam is increased. In addition, higher compressive stresses in the timber under the action of the external loads give rise to a plastic response. Prestressing of the beam also influences the deformation of the member. When the jacking force is released, the beam cambers in the upward direction. This deflection can be offset against the deflection due to the external loads thereby giving an apparent increase in flexural stiffness. Several researchers have investigated the feasibility of prestressing timber beams with FRP plates. As far back as 1992, Triantafillou and Deskovic [63] undertook a small scale testing programme, in which one timber beam was reinforced with 2.5% prestressed CFRP plate bonded to the beam soffit. A strength gain of 40% compared to 16% for passive reinforcement was achieved. More recently, Dagher et al. [64] tested 6.7 m long glulam beams with 1% GFRP plate reinforcement bonded to the soffit using a PRF adhesive. A prestressing force of 30% of the ultimate tensile strength was applied to the FRP plate using hydraulic jacks. The strength of the prestressed beams was found to be 95% higher than unreinforced beams and 38% higher than passively reinforced beams. The average precamber of the beams was 10.9 mm, which can be offset against the allowable deflection. Rodd and Pope [65] investigated the behaviour of a GFRP prestressed glulam beam with a bumper lamination glued to the bottom face of the FRP plate. Lehmann et al. [65] reported increases in bending resistance of 30% in solid timber beams when prestressed with 0.3% CFRP. The precamber was about 30% of the allowable deflection. A key limitation is the fact that the prestressing force needs to be anchored at the ends of the beams. Load transfer takes place over a short length at the end of the bond line and delamination may occur due to the local stress concentration. Brunner and Schnueriger [67] described a gradiented prestressing device developed by EMPA as a possible solution to this problem. In this method, the curing of the epoxy adhesive is carried out in a controlled fashion using heat to accelerate curing. Starting at midspan and gradually moving towards the support while reducing the prestressing force, ensures that the prestressing force is anchored over a longer length and the force at the end is reduced. An alternative approach [64] is to release the jacking force immediately after applying the bonding pressure. Prestressing of FRP using hydraulic jacks may not be suitable for on-site applications. An alternative method proposed by Negrão et al. [65] involves the pre-cambering of the timber before installing the FRP reinforcement as shown in 9. This is achieved using an adjustable prop located at the centre of the beam. This approach has the additional advantage of inducing a triangular bending moment distribution in the beam due to the prop force resulting in a low, constant shear stress in the glue line. This method may not be suitable for large beams as the prop force required may be too great.

Fig. 10 On-site procedure for FRP prestressing of timber [65]

195

Reinforcement of Timber Structures Another method, most recently developed at Chalmers University of Technology, is the stepwise prestressing of the laminate, which greatly reduces stress concentrations at the ends of the laminate, making mechanical anchorage unnecessary, [67]. This method differs from the gradiented prestressing device developed by EMPA in that it is based on the stepwise introduction of the prestressing force in the laminate, rather than the gradual release of the pre-stressing force. Special equipment was also developed to pull the laminate in such a way that the entire pre-stressing force is gradually introduced into the laminate in a discrete or continuous manner by distributing the total force over ten steps. Numerical and experimental studies show that it is possible to reduce the shear and peeling stresses in the bond line to levels below 1 and 0.2 MPa respectively, for a pre-stressing force of 100 kN. These values are well below the shear and tensile strength perpendicular to the grain of wood and adhesives. Design models for FRP prestressed timber beams have been presented by a number of researchers. Brunner and Schnueriger [67] describe a calculation model for the case of FRP prestressed plate bonded to the bottom of the beam using an iterative approach for the moment capacity and a linear elastic perfectly plastic model for timber in compression. Triantafillou and Deskovic [63] considered the same situation but used a bilinear model with a falling branch post-yield for compressive response. Brady and Harte [69] presented a closed form expression for the moment capacity of prestressed glued-laminated beams incorporating a bumper lamination covering the FRP, where the timber in compression was modelled using a bilinear model with a falling branch post-yield. McConnell et al. [70] presents a theoretical model for glulam members post-tensioned with BFRP tendons. The long-term performance of prestressed timber requires further examination before this method can be recommended for use.

4.

Design

4.1

Flexural strengthening

Flexural reinforcement can be placed on the tension and compression faces of the member and can be in the form of externally bonded plates (EBP) or near surface mounted (NSM) rods, plates or strips. Externally bonded plates are not recommended for compression reinforcement due to the likelihood of buckling. Analysis of reinforced timber flexural members is based on a classical strength of materials approach. The analysis is based on the following assumptions:  The member cross-section is symmetric in the plane of bending  Plane sections remain plane  De-bonding or slippage does not occur between the FRP and wood  The FRP material is linear elastic to failure in tension and compression  The timber is linear elastic to failure in tension and nonlinear in compression Various constitutive models have been used to model the non-linear behaviour of timber in compression. The bilinear Bazan-Buchanan model [71] assumes linear elastic behaviour up to the yield point followed by a falling branch with a negative slope. This model has been found by a

196

FRP reinforcement of timber structures number of authors [45],[48],[69] to match well with experimental results. In many cases, data on the slope of the falling branch is not available and, in that case, a simplified linear elastic, perfectly plastic model has been used [43],[49]. A quadratic approximation has also been used successfully [46]. In order to determine the ultimate moment capacity, all possible failure modes must be considered. In a large number of test programmes over the last twenty years, it has been found that failure of the FRP reinforcement is unlikely to occur and, in practise, only two failure modes need to be considered. These are:  Mode 1: Failure of the timber in tension while in compression the response is linear elastic  Mode 2: Failure of the timber in tension after the onset of compressive yielding These two scenarios are illustrated below for beams with NSM reinforcement on both faces. For failure mode 1, the maximum tensile strain in the timber, ε2, reaches the ultimate value while the maximum compressive strain, ε1, is less than the yield strain.

 2   tu ; 1   cy

(4)

where εtu is the ultimate tensile strain for the timber and εcy is the timber compressive yield strain

Fig. 11 Failure mode I

Fig. 12 Failure mode 2

The analysis is based on strain compatibility and force equilibrium. Equilibrium of the axial forces acting on the reinforced section requires that

Ff 1  Ff 2  Fcw  Ftw  0

(5)

where Ff1 and Ff2 are the forces in the compressive and tensile reinforcement, respectively, and Fcw and Ftw are the total compressive and tensile forces in the wood, respectively. As all of the strain terms are linearly related, Eq. (5) reduces to

E f Af 1 

h  hNA  h f 1 h  hNA

  E f  Ew  Af 2 

hNA  h f 2 h  hNA

1 1 h2   Ewb   h  hNA    Ewb  NA  0 2 2 h  hNA

(6)

where the dimensions are defined in Fig. 11, Ef and Ew are the elastic moduli of the FRP and timber, respectively, Af1 and Af2 are the areas of tensile and compressive reinforcement, respectively, and εf1 and εf2 are the corresponding strain terms. Solving this quadratic equation gives the location of the neutral axis, hNA. Increasing the axial stiffness EA of the FRP in tension causes the neutral axis to move down, while increasing that of the FRP in compression causes the neutral axis to move up. Knowing the location of the neutral axis, the ultimate moment capacity of the reinforced section is determined by taking moments of the forces about the neutral axis

197


2 2 M u  Ff 1   h  hNA  h f 1   Ff 2   hNA  h f 2   Fcw   h  hNA   Ftw  hNA 3 3

(7)

For failure Mode 2, the maximum tensile strain in the timber, ε2, reaches the ultimate value while the maximum compressive strain, ε1, is greater than the yield strain (Fig. 12).

Ff 1  Ff 2  Fcw1  Fcw2  Ftw  0

(8)

This again gives a quadratic expression which can be solved to give the location of the neutral axis. The ultimate moment capacity is then determined by taking moments of the normal forces about the neutral axis

2 2 2 M u  Ff 1   h  hNA  h f 1   Ff 2   hNA  h f 2   Fcw1  hcy  Fcw2  hcy 2  Ftw  hNA 3 3 3

(9)

The extent of compression yielding in the timber depends on the relative magnitudes of εtu and εcy and on the amount of tension and compression reinforcement. Increasing the amount of reinforcement on the tension side leads to more ductile behaviour and increased strength. Ideally, the timber beam should be unloaded prior to application of the FRP reinforcement. If this is not possible, the existing strain in the structure before FRP strengthening takes place must be taken into account. Fundamental to the design process is the selection of appropriate design values for the material properties. While some countries have developed national guidelines, there is currently no European Standard specifying the design of FRP structures or FRP reinforcement for timber, steel or concrete structures. The Italian National Research Council has developed guidelines [2],[3] for FRP reinforcement of existing structures. Using a limit state approach, the design value of a property is expressed as

Xd   Xk /  m

(10)

where η is a conversion factor to account for environmental and creep effects, Xk is the characteristic value and γm is the partial material factor. Some of the recommended values for the conversion factor η from CNR-200 [2] are summarised in Tab. 3. CNR-DT200 provides different partial material factors for FRP depending on the failure modes and the types of FRP system used. For FRP rupture, γm is given as 1.1 for certified systems and 1.25 for uncertified systems. Tab. 3 FRP design conversion factors  [2] Exposure

Loading mode

FRP InternalExternalAggressive Creep Fatigue Glass / Epoxy 0.75

0.85

0.95

0.30 0.50

Aramid / Epoxy 0.65

0.75

0.85

0.50 0.50

Carbon / Epoxy 0.50

0.70

0.85

0.80 0.50

198

FRP reinforcement of timber structures A major challenge in the design of reinforcement for existing timber structures is the determination of the properties of the timber for use in design calculations. This is particularly difficult for older structures. Non-destructive testing carried out during the assessment of the structure prior to reinforcement is necessary in order to get characteristic values for design. For timber, the design value of a property is expressed as

X d  kmod  X k n  /  m

(11)

where kmod is the modification factor for service classes and load-duration and Xk(n) is the characteristic value of the property from on-site tests. It has been found by several researchers [47],[48],[50] that reinforced timber beams fail in tension at a higher stress than unreinforced beams. This is because the reinforcement in the tension zone bridges defects such as knots and constrains crack opening so that the timber can carry a higher load before failure. Gentile et al. [48] found that an enhancement in tensile strength of between 18 and 46% could be achieved depending on the unreinforced timber strength. For the serviceability limit state, the materials are assumed to behave elastically. The flexural stiffness EI, of the reinforced beam can be determined using the transformed section method. Depending on the situation, a gravity load test may be carried out on the unreinforced beam to measure unreinforced flexural stiffness. From this the mean modulus of elasticity Ew for the timber parallel to grain and the modular ratio n can be calculated, where

n  E f / Ew

(12)

In the transformed section method, an equivalent section in which the FRP reinforcement is transformed to an equivalent area of timber is considered. The bending stiffness for the reinforced section is then found by multiplying the second moment of the transformed section by the modulus of elasticity of the timber. The maximum increase in flexural stiffness can be achieved by placing half of the reinforcement on the compressive face and half on the tensile face [43], however, the increases compared to reinforcing only on the tensile face may not be significant enough to justify the additional material and labour costs. Increasing the modular ratio will result in greater increase in the flexural stiffness. An efficient method for investigating different reinforcing configurations is finite element modelling incorporating nonlinear material behaviour for timber [42],[72].

4.2

Shear strengthening

4.2.1 Shear strengthening by means of external reinforcement The design of shear reinforcement is different for intact cross-sections without cracks and for damaged members. In the former case, both the timber and the FRP elements carry the forces whereas in the latter case only the FRP element carries the load in the regions of the cracked member. As a consequence the design procedures are different for both cases. Triantafillou [54] summarized the design procedure for the case where both timber and FRP elements carry loads proportionally to their stiffness in direction of the beams axis. In addition to Eq. (12), the area fraction of the FRP αfrp can be used for simplification.

 frp 

Afrp Atimber



2  t  h frp bh

(13)

where the geometric quantities are defined in Fig. 13 and Fig. 14.

199


h/2 t

t

hfrp h

hfrp h/2

b

Fig. 13 Intact timber cross-section

b

Fig. 14 Cracked timber cross-section

The maximum shear stress in the timber, due to the shear force V, can be calculated by choosing the timber as the reference material:

 timber ,max 

V Scomp ,max I comp bcomp

(14)

where the effective geometrical properties of the composite cross-section are denoted by the subscript comp. The effective width of the composite cross-section is:

  h bcomp  b  2 tn  btimber  1  n  frp   h frp  

(15)

The effective moment of inertia of the composite cross-section is: I comp

2 3    h frp  b h 3 2 n t h frp    I timber 1  n   frp   12 12  h   

(16)

and the effective static moment of area in the centre of the composite cross-section is:

Scomp ,max  b

h h h   hh  2nt frp frp  Stimber ,max 1  n frp  frp  24 2 4 h  

(17)

Hence, the stresses in the timber according to Eq. (14) can be calculated as follows:

 timber ,max

h frp V Stimber ,max 1  n h  frp 1    2 h btimber I timber 1  n  frp 1  n  h frp    h  frp h frp  

(18)

And consequently the maximum shear stresses in the FRP are:

 frp,max  n  timber ,max

(19)

Using Eq. (12) the amount (area fraction αfrp) and properties (n) of the FRP can be adjusted in order to reduce the shear stresses in the timber and satisfy the shear strength of the FRP. For a good reinforcing effect FRP roving should be applied with an angle smaller than 90° with respect to the grain direction of the timber. Widmann et al. [61] reported good performance with an angle of 45° between FRP fibre and grain of the timber. In addition, the bondline stresses have to be checked.

200

FRP reinforcement of timber structures In the situation of a damaged beam by shear cracking in the cross-section centre (Fig. 14), shear is transferred only by the FRP. For a perfect reinforcement the full shear stress has to be carried by the FRP. However, in reality the reinforcing effect connecting the upper and lower beam parts will not be perfectly stiff and, hence, a reduction of the transferred shear forces will be the result. Reinforcement with low stiffness, like FRP roving perpendicular to the grain, will not be able to transfer loads between the timber members and will not contribute to the stiffness of the beam. The two beam parts will act separately as individual members. Hence, an inclined application as described by Widmann et al. [61] should be preferred. An estimation of the maximum stresses acting in the FRP can be made assuming perfectly stiff reinforcement, where shear stresses at the crack location are carried in full by the FRP. Stress peaks occur in the bond line between FRP and timber in vicinity of the shear crack. Securing an intact bond line is crucial for achieving the optimal reinforcing effect. Special consideration has to be paid to the long term behaviour of the bond line under the influence of moisture variations and changes in ambient conditions. The slenderness of the FRP sheets together with the high stiffness can cause high tension stresses in the laminate in the case of swelling of the timber or buckling due to compression in the case of shrinkage of the timber.

4.2.2 Shear strengthening by means of internal reinforcement Internal reinforcement with FRP rods act locally along the beam axis. In the undamaged cross-section the distribution of shear stresses is affected only in the reinforced area. For design of undamaged cross-sections, the internal FRP reinforcement should be neglected. Similar considerations as for reinforcement by means of self-tapping screws [73] can be made also for FRP rods. Due to the low relative portion of FRP compared to the timber cross-section the impact of the reinforcement on stiffness can be considered to be low. Nevertheless, FRP rods can be an adequate measure for reinforcing damaged beams with shear cracks. The FRP rods should be designed to carry the forces acting between the upper and lower timber parts. In the literature, most tests have been made with FRP rods installed perpendicular to the grain, however, an inclination should be chosen in order to benefit from the high tensile strength of the reinforcing elements. Svecova and Eden [58] observed a good reinforcing effect with a dowel distance equal to the beam depth in their tests.

4.3

Long-term deformations in timber and composites

4.3.1 Overview Timber structures are subject to long-term deformations due to creep, which are exacerbated in the presence of moisture and especially moisture variation, known as mechano-sorptive creep. By reinforcing timber with material that has superior properties when it comes to short- and long-term stiffness, the long-term behaviour could be improved and deflection could be reduced. The first recordings of accelerated creep in wood due to varying humidity conditions were described in the 1960s [74]. These recordings were later verified by performing bending tests on small wood specimens in both cyclic and constant humidity [75]. Specimens subjected to constant humidity show an almost constant creep rate. For cyclic humidity the deflections varies with the drying and wetting cycles, but the total deflection increases for every cycle. For higher stress rates, the rate of deflection increases significantly in the presence of cyclic humidity. Generally, the fibres in composite materials are not expected to creep significantly. It is widely known that the creep for specimens loaded in the fibre direction is negligible [76]. On the other hand, the resins and adhesives exhibit marked rheological properties (viscous properties) experiencing

201

Reinforcement of Timber Structures significant creep, which are strongly influenced by temperature. To improve the long-term properties of two-component epoxy type adhesives, a study was made where dog-bone shape epoxy specimens were cast [77]. Half of the specimens were reinforced with 0.5% carbon fibres. These specimens had a creep rate 38% lower than the unreinforced specimens.

4.3.2 Experimental creep test on solid wood strengthened with composites Creep testing of unreinforced and reinforced beam specimens was carried out at Chalmers University of Technology [78]. In the study, 24 specimens were used with the cross-section 45x70 mm and length of 1.1 m. In order to minimise the variability of the material properties, the specimens were well defined in terms of the origin of the raw material and sawing pattern [79]. The specimens were assigned to four different groups with approximately the same mean value and standard deviation in modulus of elasticity. The different reinforcement schemes are shown in Tab. 4. Two types of CFRP were used, namely, CFK 150/2000 (CFRP 165) by S&P and Carbodur H514 (CFRP 300) by Sika. No creep was expected to occur in the CFRP reinforcement or in the steel as it was loaded in the direction of the fibres. The adhesive used was in all cases S&P Resin 220 Epoxy. In order to allow moisture exchange on three parts of the cross-section, the tension surface of unstrengthened timber specimens were sealed with epoxy. Tab. 4 Creep test specimens and reference beam, sealed with epoxy on the tension side.

Beam types

Strengthening schemes Reinforcement area [mm²] E-Modulus [GPa] Tensile strength [MPa]

Timber

CFRP 165

CFRP 300

Steel

Sealed with epoxy -

CFK 150/2000* 70 165 2310

CarboDur H514** Steel 70 100 300 210 1350 235

* Sto S&P flexural strengthening product (plates) ** Sika Chemicals flexural strengthening product (plates)

The creep tests were conducted in four-point bending at a stress level of 8 MPa. Directly after loading of the specimens the climate was set to 90% RH thereafter the climate was changed between 30% RH and 90% RH in two week cycles, while the room temperature was held constant at 23°C. The mid-point curvature was measured using a LVDT gauge. The results showed that the reinforced specimens had a lower initial deflection than the un-reinforced specimens. This result was expected due to the higher stiffness of the reinforced specimens. The creep results were presented as relative creep, i.e. the deflection divided with initial deflection (after 60 seconds). The mean values of the relative creep from all series are shown in Fig. 15. The variation in relative creep between the wet and dry periods is very high, much higher for the reinforced specimens than for the unreinforced specimens. For the reinforced specimens the variation between 90% and 30% RH is of the same magnitude as the initial deflection. This is caused by the longitudinal shrinkage of the timber material while the reinforcement material does not shrink. This effect is much larger when one-sided reinforcement was used. The conclusion is that the reinforcement material should be distributed on both the tension and compressive side of the beams. The trend lines in Fig. 16 were calculated by taking the sum of the initial value and the final value for

202

FRP reinforcement of timber structures each 2 week period. The trend showed clearly that the reinforcement prevents the mechano-sorptive creep rather well. After the first 4 weeks the increase of the relative creep for the steel and CFRP-300 reinforced specimens is only a fourth of the increase for timber with epoxy.

Fig. 15 Mean relative creep for the four different Fig. 16 Mean relative creep - trend lines for the reinforcement schemes four different reinforcement schemes

4.3.3 Design aspects In order to minimise the mechano-sorptive creep, it is beneficial to strengthen both the tension and compression sides of a timber beam to avoid large moisture-induced movements. EC5 does not provide kdef factors for FRP strengthened beams. However, the main benefit is due to increased short-term stiffness of the strengthened beams, but the long-term creep deformations are the same as for unstrengthened beams. More research is needed to investigate various climate classes and load durations for certain standardised strengthening system. Calculation of deflection for a reinforced beam (with about 2% reinforcement) in a residential house showed that it is possible to increase the span length by as much as 20% or reduce the size of the cross-section compared to an un-reinforced beam.

5.

Quality Control for Bonding on Site

The success of the reinforcement intervention depends on the implementation of adequate quality control measures at each stage of the process. All work must be carried by appropriately experienced personnel, using suitable materials, procedures and equipment in accordance with the Quality Plan. A procedure for quality control on site was developed as part of the EU CRAFT Project LICONS [80]. This was subsequently adopted by COST Action E34 [18]. Draft standards for on-site acceptance testing of mixing and application of adhesives have been prepared by CEN TC 193/SC1/WG11 [7][10]. These groups recommended that the Quality Plan should include the following checks and should specify the frequency of each check / test. 1. Reception of Materials: A record should be kept of all materials delivered to site. The materials should be checked to ensure that they match the specification and the expiry date should be checked. 2.

Inspection and tests a) Elements to be repaired: the timber moisture content should be checked and recorded, the condition of the gluing surface should be checked and the dimensions and locations of holes, slots etc. should be checked and recorded.

203

Reinforcement of Timber Structures b) Mixing and application of adhesives: the following tests should be carried out in accordance with the draft standards [7]-[10]  On-site sampling and measurement of the cure schedule for adhesives  On-site sampling and subsequent laboratory measurement of the shear strength of adhesive joints  On-site sampling and proof loading of the strength of adhesive joints c) Visual inspection of final repair including dimensional checks. Based on these tests, the actual properties obtained on site can be compared to the values used in the design.

6.

Outlook and recommendations

Using bonded fibre-reinforced polymer laminates for the strengthening and repair of wooden structural members has been shown to be an effective and economical method. The high strength and stiffness, light weight and good durability properties of FRP composites, together with advantages offered by adhesive bonding, have made it a suitable alternative for traditional strengthening and repair techniques. FRP materials and adhesives and their properties suitable for various methods are discussed. It is pointed out that careful surface preparation is essential in order to achieve good bond strength and durability. Two-part, cold-cure epoxy adhesives have generally been found to be most suitable for on-site bonding, as they have good gap-filling properties and low curing shrinkage. It is pointed out that careful surface preparation is essential in order to achieve good bond strength and durability. Research on the bond behaviour of FRP to timber is still in its infancy. More research is needed, focusing in particular on various factors affecting the bond between FRP and wood, including combined properties and the local bond-slip relationship for FRP to timber, which can be used directly for finite element analysis. A design approach, which accounts for the fracture and delamination in FRP strengthening, including appropriate material models, is presented. Failure-mode-based strength criteria are discussed and the development made by Puck is shown in terms of the mathematical formulation and strength indices for the Puck-Knaust failure criterion. The applications presented are related to beam-end reinforcements, improving tension strength perpendicular to the grain, shear reinforcements and the pre-stressing of FRP on the tension side of a beam in flexure. Beam-end reinforcement is a very important repair method for the decayed part of a structure and specifically the restoration of historically significant structures. An increase in structural performance of 140% for timber-on-timber supplements has been reported. Studies of the slack reinforcements of beams in flexure and shear reinforcement are briefly summarised, as studies of this kind have been conducted for many years and a large amount of research is available. The pre-stressing of FRP materials on the tension side of a beam in flexure offers the most effective utilisation of these materials; increasing the load-bearing capacity and pre-cambering of existing beams and, by doing so, improving the serviceability limit state which often governs the design. It is recognised that one major challenge when it comes to the design of reinforced timber members in the ultimate limit state is the lack of appropriate properties of timber and of older timber structures in particular. The maximum increase in flexural stiffness (in the serviceability limit state) can be achieved by placing half the reinforcement on the compressive face and half on the tensile face. There

204

FRP reinforcement of timber structures are no creep factors for long-term deflection as defined in design codes for timber beams strengthened in flexure. It is therefore not possible to take advantage of lower mechano-sorptive creep than that in unstrengthened beams. In general, more research is needed on long-term tests and durability performance to cover the effects of variations in moisture and ambient conditions. The need for appropriate quality control for bonding on site is recognised and a quality plan based on various standards is presented.

References [1] [2]

[3]

[4] [5] [6] [7]

[8]

[9]

[10]

[11] [12]

[13] [14]

[15] [16]

Hollaway, L., and Teng, J., Strengthening and Rehabilitation of Civil Infrastructures Using FibreReinforced Polymer (FRP) Composites, Woodhead Publishing Limited, Cambridge, UK, 2008. CNR-DT 200/2004, Guide for design and construction of externally bonded FRP systems for strengthening existing structures. Materials, RC and PC structures, masonry structures, National Research Council. Advisory Committee on Technical Recommendations for Construction, Rome, 2004. CNR-DT 201/2005, Guidelines for the design and construction of externally bonded FRP systems for strengthening existing structures. Timber Structures, National Research Council. Advisory Committee on Technical Recommendations for Construction, Rome, 2007. André, A., Fibres for strengthening of timber structures, Master Thesis, Luleå tekniska universitet, Civil and Environmental Engineering, Structural Engineering, 2006. Triantafillou, T., “Composites: a new possibility for the shear strengthening of concrete, masonry and wood”, Composites Science and Technology, Vol. 58, No.8, 1998, pp. 1285-1295. Radford, D. W., Van Goethem, D., Gutkowski, R.M., Peterson, M. L., “Composite repair of timber structures”, Construction and Building Materials, Vol. 16, No. 7, 2002, pp. 417–425. Greco, A., Maffezzoli, A., Casciaro, G., Caretto, F., “Mechanical properties of basalt fibers and their adhesion to polypropylene matrices”, Composites Part B: Engineering, Vol. 67, 2014, pp. 233-238. CEN TC 193/SC1/WG11, Adhesives for on site assembling or restoration of timber structures. On site acceptance testing: Part 1: Sampling and measurement of the adhesives cure schedule, Doc N20, CEN TC 193, 2007. CEN TC 193/SC1/WG11, Adhesives for on site assembling or restoration of timber structures. On site acceptance testing: Part 2: Verification of the shear strength of an adhesive joint, Doc N21, CEN TC 193, 2007. CEN TC 193/SC1/WG11, Adhesives for on site assembling or restoration of timber structures. On site acceptance testing: Part 3: Verification of the adhesive bond strength using tensile proof loading, Doc N22, CEN TC 193, 2007. Broughton, J.G., Hutchinson, A.R., Review of relevant materials and their requirements for timber repair and restoration, LICONS CRAF_1999-71216, 2003. Raftery G.M., Harte A.M., Rodd P.D., “Bond quality at the FRP–wood interface using wood laminating adhesives”, International Journal of Adhesion and Adhesives, Vol. 29, No. 2, 2009, pp. 101–110. Raftery, G.M., Harte, A.M., Rodd, P.D., “Bonding FRP materials to wood using thin epoxy glue lines”, International Journal of Adhesion and Adhesives, Vol. 29, No. 5, 2009, pp. 580-588. Davalos, J.F., Qiao, P.Z., Trimble, B.S., “Fiber-reinforced composite and wood bonded interfaces: Part1. Durability and shear strength”, Journal of Composite Technology and Research, Vol. 22, No. 4, 2000, pp. 224-231. Harvey K., Ansell, M.P., “Improved timber connections using bonded-in GFRP rods”, In: Proceedings of 6th World Conference on Timber Engineering, Whistler, Canada, 2000. Custodio, J., Broughton, J.G., “Factors influencing bond performance”, In Core Document of COST Action E34, Bonding of timber, University of Natural Resources and Applied Life Sciences, Vienna, 2008.

205

Reinforcement of Timber Structures [17] Pizzo, B., Smedley, D., “Adhesives for on-site bonding: characteristics, testing, applications”, In [18]

[19] [20] [21]

[22] [23]

[24]

[25] [26]

[27] [28]

[29] [30] [31] [32]

[33]

[34] [35] [36] [37] [38]

206

State-of-the-art report on the Reinforcement of Timber Structures, 2015, COST Action FP1101. Smedley, D., Cruz, H., Paula, R., “Quality control on site”, In: Core Document of COST Action E34, Bonding of timber, University of Natural Resources and Applied Life Sciences, Vienna, 2008. Wan J., Smith S., Qiao P., and Chen F., “Experimental Investigation on FRP-to-Timber Bonded Interfaces”, Journal of Composites for Construction, Vol. 18, No. 3, 2013, pp. A4013006:1-9. Barbero E., Davalos J., and Munipalle U., “Bond strength of FRP-wood interface”, Journal of Reinforced Plastics and Composites, Vol. 13, No. 9, 1994, pp. 835-854. Davalos J., Qiao P., and Trimble B., “Fiber-reinforced composite and wood bonded interfaces: Part 1. durability and shear strength”, Journal of Composites Technology and Research, Vol. 22, No. 4, 2000, pp. 224-231. Lopez-Anido R., Gardner D., and Hensley J., “Adhesive bonding of eastern hemlock glulam panels with E-glass/vinyl ester reinforcement”, Forest Products Journal, Vol. 50, No. 11/12, 2000, pp. 43-47. Crews K., and Smith S., “Tests on FRP-strengthened timber joints”, In: Proceedings, Third International Conference on FRP Composites in Civil Engineering, CICE 2006, Miami, USA, pp. 677-680. Silva P., Cachim P., and Juvandes L., “Técnicas avançadas de reforço de estruturas de Madeira com compósitos reforçados com fibras (FRP)”, In: Proceeding 1st Iberic Congress Cimad’04Wood in Construction, Guimarāes, 2004, pp. 613-622. Juvandes L., and Barbos R., “Bond analysis of timber structures strengthened with FRP systems”, Strain, Vol. 48, No. 2, 2012, pp. 124-135. Wan J., Smith S., and Qiao P., “FRP-to-softwood joints: experimental investigation”, In: Proceedings of Fifth International Conference on FRP Composites in Civil Engineering, CICE 2010, Beijing, China, pp. 951-954. Lorenzis L., Scialpi V., and Tegola A., “Analytical and experimental study on bonded-in CFRP bars in glulam timber”, Composites Part B: Engineering, Vol.36, No. 4, 2005, pp. 279-289. Valipour H., and Crews K., “Efficient finite element modeling of timber beams strengthened with bonded fibre reinforced polymers”, Construction and Building Materials, Vol. 25, No. 8, 2011, pp. 3291-3300. Talreja, R., Singh, C.V, Damage and Failure of Composite Materials. Cambridge University Press, 2012. Puck, A., “A Failure Criterion shows the Direction – Further Thoughts of the Design of Laminates”, Kunststoffe German Plastics, Vol. 82, 1992, pp. 29-32. Puck, A., “Praxisgerechte Bruchkriterien für hochbeanspruchte Faser-Kunststoffverbunde”, Kunststoffe, Vol. 82, No. 2, 1992, pp. 149-155. Schober, K.U., Rautenstrauch, K., “On the application of cohesive zone modeling in timber composite structures”, In: Proceedings of the 10th World Conference on Timber Engineering (WCTE 2008), Miyazaki, Japan, 2008. Schober, K.U., Rautenstrauch, K., “Structural Behaviour of Hybrid Timber-Composite Beams”, In: Proceedings Composites & Polycon 2009, American Composites Manufacturers Association, Tampa, FL, USA, 2009. Martin ZA, Tingley DA., “Fire resistance of FRP reinforced glulam beams”, In: Proceedings of World Conference on Timber Engineering, Whistler, Canada, 2000. Williamson T.G., “Fire performance of fiber reinforced polymer glued laminated timber”, In: Proceedings of 9th World Conference on Timber Engineering, Portland, Oregon, 2006. Metten C.J., Page A.V., Robinson G.C., Repair of structural timbers. Part 2 Fire resistant repairs, TRADA, High Wycombe, UK, 1993. Hallström S., Grenestedt, J. L., “Failure analysis of laminated timber beams reinforced with glass fibre composites”, Wood Science and Technology, Vol. 31, No. 1, 1997, pp. 17-34. Hallström S., “Glass fibre reinforced Holes in Laminated timber”, Wood Science and Technology, Vol. 30, 1996, pp. 323-337.

FRP reinforcement of timber structures [39] Coureau J.L., Cuvillier E., Lavergne C., “Strength of locally PGF reinforced end-notched

[40] [41] [42]

[43]

[44] [45] [46]

[47] [48]

[49]

[50] [51]

[52] [53]

[54] [55] [56]

[57] [58]

[59]

beams”, In: Joints in Timber Structures, Aicher S., Reinhardt H.W. (eds) International RILEM Symposium, 2001. Jockwer, R., Structural behaviour of glued laminated timber beams with unreinforced and reinforced notches, PhD Thesis, ETH Zurich, Institute of Structural Engineering, Zurich, 2014. Enquist, B., Gustafsson, P. J., Larsen, H. J., “Glass-fibre reinforcement perpendicular to the grain”, In: Proceeding of International Timber Engineering Conference, London, 1991, pp. 3.242-3.250. Nowak, T.P., Jasienko, J., Czepizak, D., “Experimental tests and numerical analysis of historic bent timber elements reinforced with CFRP strips”, Construction and Building Materials Vol. 40, 2013, pp. 197–206. Kliger, R., Johansson, M., Crocetti, R., “Strengthening timber with CFRP or steel plates – short and long-term performance”, In: Proceedings of World Conference on Timber Engineering, Miyazaki, Japan, 2008. Schober K.U., Rautenstrauch K., “Post-strengthening of timber structures with CFRPs”, Materials and Structures, Vol. 40, 2006, pp.27–35. Borri, A., Corradi, M., Grazini, A., “A method for flexural reinforcement of old wood beams with CFRP materials”, Composites Part B: Engineering, Vol. 36, 2005, pp.143-153. Li, Y.F., Xie, Y.M., Tsai, M.J., “Enhancement of the flexural performance of retrofitted wood beams using CFRP composite sheets”, Construction and Building Materials, Vol. 23, 2009, pp.411-422. Johns K.C., Lacroix, S., “Composite reinforcement of timber in bending”, Canadian Journal of Civil Engineering, Vol. 27, 2000, pp. 899-906. Gentile, C., Svecova, D., Rizkalla, S.H., “Timber beams strengthened with GFRP bars: development and applications”, Journal of Composites for Construction, Vol. 6, No. 1, 2002, pp. 11-20. Hernandez, R., Davalos, J.F., Sonti, S.S., Kim, Y., Moody, R.C., Strength and stiffness of reinforced yellow-poplar glued-laminated beams, Research Paper FPL-RP-554, Department of Agriculture, Forest Service, Forest Products Laboratory, Madison, WI, US, 1997. Raftery, G.M., Harte, A.M (2011): Low-grade glued laminated timber reinforced with FRP plate, Composites Part B: Engineering 42(4):724-735. Alam, P., Ansell, M.P. Smedley, D., “Mechanical repair of timber beams fractured in flexure using bonded-in reinforcements”, Composites Part B: Engineering, Vol. 40, No. 2, 2009, pp.95106. Pengyi, Z., Shijie, S., Chunmei, M., “Strengthening Mechanical Properties of Glulam with Basalt Fiber”, Advances in Natural Science, Vol. 4, No. 2, pp. 130-133. Akbiyik, A., Lamanna, A. J., Hale, W. M., “Feasibility investigation of the shear repair of timber stringers with horizontal splits”, Construction and Building Materials, Vol. 21, No. 5, 2007, pp. 991-1000. Triantafillou, T., “Shear Reinforcement of Wood Using FRP Materials”, Journal of Materials in Civil Engineering, Vol. 9, No. 2, 1997, pp. 65-69. Alam, P., Ansell, M. P., Smedley, D. “Mechanical repair of timber beams fractured in flexure using bonded-in reinforcements”, Composites PartB: Engineering, 40(2), pp. 95-106. Burgers, T., Gutkowski, R., Balogh, J., and Radford, D., “Repair of Full-Scale Timber Bridge Chord Members by Shear Spiking,” Journal of Bridge Engineering, Vol. 13, No. 4, 2008, pp. 310-318. Gentry, T. R., "Performance of glued-laminated timbers with FRP shear and flexural reinforcement.", Journal of Composites for Construction, Vol. 15, No. 5, 2011, pp. 861-870. Svecova, D., Eden, R.J., ”Flexural and shear strengthening of timber beams using glass fibre reinforced polymer bars -an experimental investigation”, Canadian Journal of Civil Engineering, Vol. 31, 2004, pp. 45-55. Lauber, B., “Ertüchtigung von Holztragwerken - Bespiele aus der Praxis”, In: Proceedings of the Holzbautag Biel, Biel, 2008.

207

Reinforcement of Timber Structures [60] Sonti, S.S., GangaRao, H.V.S., “Strength and Stiffness Evaluations of Wood Laminates with

[61]

[62]

[63] [64]

[65]

[66] [67]

[68]

[69]

[70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80]

208

Composite Wraps”, In: Proceedings of the 50th Annual Conference, Composites Institute, The Society of Plastics Industry Inc., 1995. Widmann, R., Jockwer, R., Frei, R., Haeni, R., “Comparison of different techniques for the strengthening of glulam members”, In: Enhance mechanical properties of timber, engineered wood products and timber structures: COST Action FP1004 Early Stage Researchers Conference, Zagreb, Croatia, 2012, pp. 57-62. Widmann,R., Tannert,T., Frei,R., “Comparison of different techniques for the shear strengthening of glulam members”, In: Proc. of Second Conference on Smart Monitoring, Assessment and Rehabilitation of Civil Structures (SMAR 2013), Empa, Istanbul, Turkey, 2013, 8 pp. Triantafillou, T.C., Deskovic, N., “Prestressed FRP sheets as external reinforcement of wood members”, Journal of Structural Engineering, Vol.118, No. 5, 1992, pp. 1270-1283. Dagher, H., Gray, H., Davids, W., Silva-Hernandez, R., Nader, J., “Variable prestressing of FRP-reinforced glulam beams: methodology and behaviour”, In: Proceedings of World Conference on Timber Engineering (WCTE 2010), Riva del Garda, Italy, 2010. Rodd, P.D., Pope, D.J., “Prestressing as a means of better utilising low quality wood in glued laminated beams”, In: Proceedings of the International Conference on Forest Products, Daejoen, Korea, 2003. Negrão, J., Brunner, M., Lehmann, M., “Pre-stressing of timber”, In: COST E34 - Bonding of Wood - WG1: Bonding on site - Core Document, K. Richter and H. Cruz eds., 2008. Brunner, M., Schnueriger, M., “FRP-prestressed timber”, In: Proceedings of 8th International Symposium on Fiber Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-8), Patras, Greece, 2007 Haghani, R., Al-Emrani, M. A new method and device for application of bonded prestressed FRP laminates. Proceedings of the second international conference on advances in civil and structural engineering, ISBN 978-1-63248-035-4, 2014, Kuala Lampur, Malaysia. Brady, J.F., Harte, A.M., “Flexural reinforcement of glue-laminated timber beams using prestressed FRP plates”, In: Proceedings of 4th International Conference on Advanced Composites in Construction (ACIC), Edinburgh, UK, 2008. McConnell, E., McPolin, D., Taylor, S., “Post-tensioning of glulam timber with BFRP tendons”, Proceedings of ICE- Construction Materials, 2014, p. 1-9. Buchanan, A., “Bending strength of lumber”, ASCE Journal of Structural Engineering, Vol. 116, No. 5, 1990, pp. 1213-1229. Raftery, G.M., Harte A.M., “Nonlinear numerical modeling of FRP reinforced glued laminated timber”, Composites: Part B, Vol. 52, 2013, pp. 40-50. Dietsch, P., “Einsatz und Berechnung von Schubverstärkungen für Brettschichtholzbau-teile”, PhD Thesis, Technical University of Munich, Munich, 2012. Armstrong, L.D., Kingston R.S.T., “Effect of moisture changes on creep in wood”, Nature, Vol. 185, 1960, pp.862-863. Hearmon R.F.S., Paton J.M., “Moisture content changes and creep of wood”, Forest Products Journal, Vol. 14, 1964, pp. 357-359. Lou, Y., Schapery, R.A., “Viscoelastic characterization of a nonlinear fiber-reinforced plastic”, Journal of Composite Materials, Vol. 5, No. 2, 1971, pp. 208-234. Miravalles, M., Dharmawan, I., The creep behaviour of adhesives. A numerical and experimental investigation. Chalmers Uni. of Technology. Master’s thesis, 2007. Hansson, S., Karlsson, K., Moisture-related creep of reinforced timber. Theoretical studies and laboratory tests, Chalmers Uni. of Technology. Master’s thesis, 2007. Perstorper, M., Pellicane, P.J., Kliger, I.R., Johansson, G., “Quality of timber products from Norway spruce”, Wood Science and Technology, Vol. 29, 1995, pp. 157-170. Anon., Low intrusion conservation systems for timber structures (LICONS), http://www.licons.org/, 2006.

Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings

11 Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings Tanja Marzi1

Summary In this chapter the latest technological innovations in nano-structured materials for the reinforcement and protection of timber structures are presented. Starting from the definition of nanotechnologies applied in the construction field, the chapter briefly describes nano-materials already existing on the market, nano-coatings and their wood surface protection functions, classification and compatibility with the different wood species, focusing on their potential usage and their application in the fields of architecture, civil engineering and cultural heritage. The second part of the chapter describes an experimental research on the application of a polymeric resin reinforced with carbon nanotubes on historic timber structures belonging to cultural heritage sites.

1. Introduction Nanosciences and nanotechnologies represent a new scientific and technological approach for manipulating material structure and behaviour at the atomic and molecular scales, where properties differ significantly from those observed at a larger scale. Descending from the normal scale to the infinitely small (the prefix "nano" means 10-9), we enter the domain of quantum physics: the use of nano-particles makes it possible to obtain materials with new chemical, physical and mechanical properties and to increase the original performance of conventional materials (e.g. carbon atoms connected to form nanotubes, can produce materials that are stronger than steel) [1]. Nanotechnology involves multidisciplinary areas of investigation, which in recent years have affected all sectors of industry with significant economic implications. For the European Union, nanotechnologies represent one of the major fields of scientific development for the near future and have been identified as priority areas for European economic and industrial development. The growing interest in the potential applications of nanotechnology is confirmed at international level in the strategic documents for the planning of scientific research and technological innovation funding. In 2008, the European Commission also adopted the "Code of Conduct for responsible research in the field of nanotechnology." In the absence of a comprehensive structured legislation, this voluntary code aims at promoting integrated, safe, ethical and responsible research, recalling some principles (i.e. "sustainability" and "precaution") on which the EU States are invited to take concrete action. As nanotechnology is an emerging research field, there is a great debate regarding to what extent nanotechnology will benefit or pose risks for human health and environment and many research projects at international level are specifically focusing on these issues. 1)

Arch. Ph.D., Post-doc Research Fellow, Politecnico di Torino, Department of Architecture and Design, Torino, Italy

209

Reinforcement of Timber Structures Hundreds of products containing nano-materials are already in use in many sectors of life (including health, information society, industry, energy, transport and space). In particular, in the field of construction and cultural heritage, nanotechnologies are providing a significant boost to innovation in traditional processes and products. Some applications in relevant international architecture are contributing to their diffusion. At present, the most promising applications relate to nano-structured coatings. Together with the common performance requirements (long term stability, durability and weather resistance, good adhesion to the substrate, transparency, sustainability of the production process, etc..), they introduce additional functionality such as self-cleaning, photocatalysis, water resistance, fire resistance, scratch resistance, graffiti resistance, antibacterial coatings [2,3]. Currently, the nanotechnology industry is rapidly developing in the field of construction and cultural heritage [4]. However, it should be highlighted that most research has been directed at materials such as concrete and metal. To date, several ready-to-use nano-products for wood protection are available on the market but very little research has focused on the reinforcement of timber structures. The last part of this chapter focuses on this specific topic, describing also experimental research carried out at Politecnico di Torino on the application of a polymeric resin reinforced with carbon nanotubes on historic timber structures. The research addressed the definition and assessment of a methodology for preparation and application of the nano-composite. The aim was to verify whether there was an increased mechanical resistance in comparison with traditional reinforcement methods.

2. Nanotechnologies for wood in Cultural Heritage Due to its anatomical features, wood is considered a natural nano-structured composite material, anatomically similar to strong piping bonded with a thermoplastic matrix, the lignin, and equipped with strong dissipative capacity with regards to fracturing energy (Fig. 1). It can be seen as a polymeric composite of cellulose, hemicellulose, protein and lignin, as well as, at the nanoscale level, a cellulosic fibrillar composite [5]. This natural nano-composite can potentially offer important applications in the field of nanotechnologies mainly as nano-materials derived from forest products (i.e. nano-cellulose, cellulose nano-composites) or as nanotechnology incorporated into traditional forestbased products (i.e. nano-coating to enhance wood durability) [6]. Wooden cultural heritage is widespread, with different formal, colour and structural features that make it unique. Architectural heritage built in wood is an important sector of our cultural heritage, including different typologies (such as floors, roof trusses, bridges, etc.). On the topic of their conservation it has become widely accepted that such structures should be maintained and preserved, with interventions respectful of their original conception, but also of their material: wood [7, 8]. Wood frequently discolours as a result of exposure to ultraviolet (UV) light, moisture and bioorganisms. Treatments introduced by nanotechnologies can improve its durability or stability, which are negative factors associated with timber exposure to the environment. Despite the highly innovative character of these techniques (some still experimental), recent international case studies of applications of nano-structured materials on historic timber structures belonging to cultural heritage sites are beginning to be known (Fig. 2).

210


a)

b) Fig. 1 Different levels of observation of wood (a) (Artwork by Mark Harrington, copyright University of Canterbury, 1996) and variability of wood structure at different scales of observation (b) (reelaborated from R.Zanuttini)

a)

b)

Fig. 2 Examples of nano-coatings application on relevant architectures: pavilion at Hakone Open-air Museum (a) and timber structures of traditional Japanese homes (b) (from [13])

3. Nano-coatings for wood protection Nano-coatings can improve the performance and functionality of wood, extending its stability, which is often a factor that limits its use. On the other hand, it is noted that traditional products and

211

Reinforcement of Timber Structures treatments available on the market, aimed at maintaining wood durability, are often highly toxic for humans and for the environment. There is therefore the need for new non-toxic products. Nano-materials are usually incorporated into coatings, in an aqueous, organic or polymeric medium and different nano-impregnation of timber is possible. An essential aspect is that nano-particles are well dispersed in a suitable medium to avoid aggregation. Nano-particles are less than 100 nm in size and they protect or enhance the properties of the substrate. Their high ratio of surface area to mass ensures that a loading of only a few percent by weight in coatings can significantly enhance chemical, thermal and physical properties [9]. A survey was carried out on nano-based coatings specifically addressed to wood that are already present on the international market [10]. Among the main fields of application of these products there is water repellency, followed by UV rays protection (Fig. 3). It should be highlighted that there is a wide choice of nano-particles which offer different functionality or multi-functionality. Some nanomaterials, such as silica (SiO2) and titania (TiO2), are multi-functional and there is frequently a choice of nano-materials for a given application. There are also health and safety issues associated with the abrasion and wear of nano-materials, releasing potentially hazardous particles into the atmosphere [11]. In-life and end-of-life health and safety factors should be considered in developing nano-materials for timber [9].

Fig. 3 Main fields of application of nano-based coatings addressed to wood already present on the market: waterproofing (Clay, SiO2, CeO2, TiO2); UV protection (TiO2, ZnO, SiO2, CeO2, Fe2O3, clays); biotic decay protection (Ag, ZnO, Cu); fire resistance (SiO2, Clay, TiO2); selfcleaning/photocatalitic (TiO2, ZnO); anti-scratching, hardness (TiO2, Al2O3, SiO2, clays, lime) (from [10])

3.1. Waterproofing Hydrophobic (water-repellent) nano-coatings are among the promising developments for facade treatments and weather protection of timber. These products mainly consist of silica nanoparticles, by exploiting the so-called "lotus effect", properties of water repellence and at the same time breathability of the surface can be obtained (Fig. 4). It is important to highlight that this kind of nanostructured coating, which is nearly VOC-free and based on a water-borne silane system, has good water repellent functions but, essential for wood, it is open-vapour.

212


Fig. 4 Waterproofing and “lotus effect” on timber surfaces treated with nano-coatings based on silica nanoparticles. In the central image it is possible to notice the difference between untreated wood (top) and wood treated with nano-coating (bottom) (from [12]-right, [13]-left and T. Marzi centre])

Fig. 5 Durability test of pine board impregnation after 1 year: board protected with nano-coating (left) and unprotected board (right) (from [12])

Fig. 6 Durability test of larch wood used for outdoor furniture in the Alpine area impregnated with different UV absorber coatings (2 traditional and 4 nano-coatings). In the image is visible the difference after 2 months of outdoor exposure. (from [21])

213


3.2. UV Protection Ultraviolet rays can change wood structure in the sub-surface layer, causing lignin degradation and discoloration. A nanoscale protective barrier can increase resistance to ultraviolet radiation, preserving the appearance and the original colours of wood, which can be very important when dealing with cultural heritage (Fig. 5). The small size of the particles makes it possible to offer high protection without affecting the transparency of the coating. The water based nature of this typology of products, which are volatile organic compound (VOC) free, is a functional and environmentally friendly solution against rot and moss build-up [12]. Water absorption is significantly reduced as well. By using this technology, the product reflects the most damaging waves of the sun’s spectrum, protecting wood for a longer time from solar radiation.

Furthermore, nanoscale solutions are able to concentrate more active substances in a smaller volume of liquid. Therefore, it is possible to use less product to achieve better durability [12]. Nano-scale titanium dioxide is a well-known catalyst and photocatalytic material. The associated literature is extensive [11, 14, 15], analysing its applications as a self-cleaning, UV-absorbing, or sterilising agent. A key factor is that titania is not degraded by the oxidative reactions and its photocatalytic action is perpetuated. Furthermore, nano-titania has a very high ratio of surface area to weight, enhancing its photocatalytic action. Recent research has examined the photochemical stability of water-based acrylic paints containing anatase and rutile nano-titania [16]. Research carried out at the Université Laval Québec investigated the properties of nano-coatings using zinc oxide, silica, alumina and titanium dioxide active nano-particles dispersed in acrylic-based, waterborne, solidcolour stains for exterior wood [17, 18, 19]. A specific research carried out at the Department of Architecture and Design of Politecnico di Torino (EU Project Alcotra 2007 “Savoir Bois”) has studied the durability of wood surfaces of different wood species typical of the Alpine area exposed to outdoor extreme conditions (Fig. 6). Wooden outdoor furniture realized by local artisans of the Monviso valleys were coated with traditional UV absorbant coatings and nano-coatings with a view to assessing the best solution that would require the minimal maintenance. Natural weathering tests were carried out in outdoor exposure in the mountains for a period of two years [20]. Best results were achieved with a polyurethane resin modified with silica nanoparticles from sol-gel.

3.3. Biotic decay Improved protection from biotic decay of wood can be obtained with the use of nano-materials used as a protective treatment and as a barrier to moisture. In particular, products containing silver nanoparticles are particularly effective for their antibacterial, anti-microbial and anti-mould properties and nano-silver is nowadays a constituent of several commercial wood preservatives. Wood can be attacked by a variety of insects, fungi and other organisms. These decay factors can successfully be tackled by using preparations based on nano-silver. Specimens coated with nano-silver

214

Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings leave active nanoparticles on the surface. When bacteria come into contact with the protected surface, the silver binds within the bacteria and causes their enzymes to breakdown, stopping the bacteria and causing them to die (Fig. 7). Silver is a well-known biocidal agent and it has been available for over 100 years in colloidal nanoform [22, 23]. Silver-containing compounds which release silver ions into a moist environment differ from nano-silver particles, which may act as a catalyst and are often held in a polymeric matrix as a coating. The release of nano-silver by leaching from paints [24] is often associated with conversion into less toxic compounds such as silver sulphide, so health risks are reduced. Nano-silver has been widely utilised to reduce the biodegradation of wood and also to improve resistance to termite attack [25, 26].

Fig. 7 Scheme of silver nanoparticles action.

Fig. 8 Durability test carried out on wood composite panels protected with nano-silver coating and unprotected (from [12])

Fig. 9 Test carried out on wood treated with nanostructured coating to increase fire resistance (top wood) and untreated wood element (bottom) before and after fire action (from [29])

3.4. Fire resistance Fire safety is an important concern in all types of construction and nanostructured coatings can be effective also as fire retardants. When heated, wood undergoes thermal degradation and combustion to produce gases, vapours, tars and char. In order to improve the resistance to fire, timber products are commonly treated with fire retardants that are typically coated onto the surface of wood (painted, spayed or dipped), or impregnated into the wood structure using vacuum-pressure techniques or other techniques as plasma treatments. Fire retardants can usually provide thermal insulation, absorb the surrounding heat by endothermal reactions, or increase the thermal conductivity of wood in order to dissipate the heat from the wood surface [27].

215

Reinforcement of Timber Structures Traditional fire-retardant treatments, such as inorganic salts, may affect wood in various ways by: increasing hygroscopicity, reducing strength, and because they lead to dimensional stability changes depending on treatment, wood degradation, corrosion of metal fasteners, adhesion problems, increased abrasiveness, leaching of treatments. They also contribute to the production of smoke and gases toxic and dangerous for people’s health [28]. The growing awareness of environmental issues and consumer safety of fire retardant products means that some traditional products (i.e. boron and formaldehyde) are likely to decline. The toxicity of fire retardants plays an important role in health and safety legislation. A wide range of fire retardant treatment systems for wood, including nano-composites and layer by layer applications, have been studied throughout recent years and many others are under development (Fig. 9). Fire resistance can be increased through the use of high performance thin film fire retardant coatings based on nanoparticles of titanium dioxide and silicon dioxide. These products, that may be sprayed using standard airless equipment, have been formulated to retard the flame spread across a wide variety of materials and also to suppress the generation of smoke. The two-part epoxy coating creates a highly adherent fire retardant barrier on wood. In the event of fire, it produces water and gases, which snuff out oxygen and provides a cooling effect at the flame front. A dense char is formed, which further protects the surface from combustion [26]. The high thermal conductivity of nano-silver coatings has also been tested to improve heat transfer in wood and enhance fire resistance [28]. Nano-silver treatment clearly showed potential in improving some of the fire-retarding properties in solid wood products. Such coating may delay thermal degradation and carbonization by reducing the accumulation of heat by transferring it. Other important aspects of fire retardant nano-coatings should still be studied in detail, such as toxic gases and smoke as well as hygroscopicity.

3.5. Deacidification Lime has been used traditionally for the treatment of wood in order to provide anti-microbial properties. It acts as a reflective coating and enhances aesthetic qualities. Nano-limes have been developed and patented at the University of Florence for the consolidation and conservation of limestones and lime-based wall paintings [30, 31]. Other researchers [32] have examined the role of nanoparticles of calcium and magnesium hydroxide and carbonate for neutralising acidification processes in paper and wood. Specific research concerning the treatment of the wood was dedicated to the conservation of Vasa warship (Fig. 10) with nano-limes. The almost complete hull of the Vasa was recovered from Stockholm harbour in 1961 after 333 years of immersion. The wood was treated initially with polyethylene glycol to prevent shrinkage, but subsequent acidification resulted from penetration of hydrogen sulfide in the atmosphere catalysed by iron elements present in the timber. A dispersion of nano-lime particles in 2-propanol was impregnated into the surface of oak and pine samples from the ship and controlled deacidification took place with the formation of calcium sulfate. Excess nanolime was converted to nano-calcium carbonate, which also acted as alkaline reservoir for deacidification [33].

216

Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings Recent work at the University of Bath has examined the effect of application of nano-lime to the end-grain of Welsh woods, which are used in flooring applications to improve indentation hardness [9]. The anti-microbial and antireflective qualities of nano-lime together with its deacidifying and breathable qualities could potentially be the basis for multi-functional coatings for wood.

Fig.10 Vasa warship in Stockholm (photo P. Isotalo)

3.6. Environment and Nanotechnology Nanosized particles could replace current chemical treatments. It was shown that preserving wood with direct impregnation of titanium dioxide, zinc oxide and other particles improves wood durability. This will become an important issue especially considering that many countries are already banning current preservative-treatments of wood that are considered as toxic [5]. Surface treatment nanocoatings can also significantly reduce the use of chemical cleaners and maintenance costs. As many of the coatings are water based, they are solvent and VOC free. Furthermore, in contrast to water soluble biocides, nanoparticles have the advantage that they can be fixed and bounded more effectively in the coating matrix, and the threat of leaching into the environment is lower, with supposed lower ecological risks. Therefore, nanostructured coatings have the ability to reduce energy costs and improve indoor air quality. The use of engineered nano-materials offers advantages as well as disadvantages from a sustainability perspective. Life Cycle Assessment (LCA) is a suitable method to assess the environmental performance of a product or process. A comprehensive material LCA is necessary in order to understand potential risks [34]. But so far studies applying LCA to the area of nanotechnology have been scarce, probably because LCA frameworks have a list of issues that need further precision and investigation in order to be applicable to nanotechnologies [35]. A critical aspect is that at present there are no long term studies available, even though many research projects (e.g. EU 7FP “NanoHouse” Project) have been specifically addressed to the assessment of potential environmental impact of nanotechnologies and nanoparticles’ release from nano-based coatings [36]. Other studies have been focused on techniques for durability assessment, such as the use of the nanoindentation technique in conjunction with atomic force microscopy for durability investigation and service life prediction of nanocomposite coatings for wood [19]. Further research to assess and model environmental impacts on wood surfaces and timber structures is still needed.

217


4. Next-generation fields of application Concerning future sectors of application, there is ongoing international research in the following areas:

4.1. New wood-based composite materials Nanostructured materials should be applied to the preservation and durability not only of wood but also of its derivates that nowadays are considered as the sustainable future of this material. New nanostructured adhesives and resins would allow several improvements. Some examples include an enhancement of thermal insulation properties, photovoltaic properties or the integration of monitoring devices directly deposited within the composite. With the incorporation of nanoscience into woodbased building materials, a new generation of multi-functional, high-performance, low maintenance, durable building materials and components could be achieved [6]. Several research projects have been applied to the study of wood/polymer/nano-clay composites obtaining improvements in wood properties, such as surface hardness, modulus of elasticity and water-repellency (with the addition of hydrophobic nanofillers into the wood structure) [37].

4.2. Monitoring / Maintenance Monitoring of building heritage is a sector of great interest which foresees the use of devices related to nanoscience and nanotechnology (e.g. Nanoelectromechanical systems or NEMS). These devices, already used in other sectors (e.g. medical sector) have the advantage of collecting information and storing it in-situ for a long period, resulting in improvements in the efficiency of the management, maintenance and rehabilitation process. Such approaches can be adapted to assess the structure condition and to produce a "good" or "bad" condition certificate of the construction. Maintenance actions (if applicable) and/or rehabilitation actions will be based on this information. These shall also provide the basis for monitoring after the rehabilitation interventions. The increasing requirements towards safety and security, demand for the application of advanced sensors with greater sensitivity, better specificity, of smaller sizes, lower power dissipation and advanced remote monitoring, without expensive and time consuming supervision can be satisfied (Fig. 11). Nanotechnology is a driving force in the development of advanced sensors, concerning not only the preparation and investigation of smart nano-sized materials for sensor applications (i.e. nanotubes, nanowires, …), but also the combination of their performance with integrated circuit (ICs), micro and nano-optics, Microelectromechanical systems (MEMS) and NEMS, leading to higher levels of integration, and more effective processing and transmitting of the sensor signals. Recently, nanosensors have been tested for relative humidity and temperature monitoring under controlled environmental conditions. Nanocomposites have been prepared and used to realize prototype devices proposed for use in art conservation and museum applications [38]. With specific respect to timber, there is great potential to adopt and integrate existing early-stage nanotechnologies and monitoring devices with timber structures in mind (e.g. for early detection of decay, monitor moisture content, sense UV degradation).

218

Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings There is ongoing research focusing on the use of multifunctional wood-adhesives for structural health monitoring. Tests have been carried out with functionalized carbon nanotubes used as fillers in epoxy resins that can increase the shear strength and fracture toughness of the composite together with interesting conductive properties. Since most of structural timber is glued together from lamella, the electrical modified woodadhesive could be integrated in the production process of composite timber [39, 40, 41]. Recently, there has been a focus on the development of civil structures that have embedded sensors and on-board data processing capabilities. The fusion of these smart technologies into infrastructure (e.g. bridges) is intended to provide better and more timely information on structural behaviour and maintenance [42]. Fig.11 Example of scheme of smart-monitoring for timber structures

4.3. Reinforcement The application of composite nano-materials for reinforcement of timber structures is a technique not yet widespread. Some research has been conducted in recent years on the wooden elements of new constructions [43]. Other research has been focusing on the assessment of epoxy adhesive modified with different nano- and micro-particles for in-situ bonding of FRP timber connections [44]. The application of composite nano-materials to reinforce wooden structures is a technique that is not yet available on the market. It requires therefore to be fully tested before being applied on a large scale. A research project conducted at Politecnico di Torino and specifically focused on the reinforcement of timber structures in reference to the use of these methods in the consolidation and reinforcement of existing timber structures is described in Section 4.4.

4.4. Reinforcement of historic timber structures with carbon nano-composites 4.4.1. Introduction to the research This part of the chapter describes the experimental investigation carried on the application of a polymeric resin reinforced with carbon nanotubes (CNTs) for localised interventions on historic timber structures [10, 45, 46]. The study regards the conservation of a specific wooden joint belonging to traditional constructions and defines and assesses a methodology of preparation and application of the nano-composite. The aim was to verify whether the mechanical resistance increased in comparison with traditional reinforcement methods. Different wood species were considered and the experimentation was carried out with a view to possible in-situ applications of the technique. Laboratory tests were carried out first on small wooden samples, and afterwards on full-scale wooden elements [47].

219

Reinforcement of Timber Structures The use of nano-materials and composite materials to strengthen wood has been suggested recently by several researchers who have studied the effects of carbon nanotubes-based composites on the mechanical characteristics of reinforced wooden elements [43]. In Italy, given the greater relevance of rehabilitation and conservation, research has focused on the methodologies for consolidating existing historical structures. The development of polymers reinforced with nanoparticles is one of the most promising approaches in the field of future engineering applications. The unique properties of some nanoparticles (carbon black [43] and CNTs) and the possibility of combining them with traditional reinforcing elements (glass fibre, carbon fibre or Kevlar) have generated an intense research program in the nanocomposites sector [43, 48, 49]. CNTs have a diameter of several nanometres (1-50 nm) and their length measures several microns (up to 10 μm); they have very good potential for improving the electrical and mechanical properties of polymers, even with addition of 0.1% weight content compared to epoxy resin [43- 50]. The difficulties lie in transferring these remarkable mechanical, thermal and electrical properties to the polymer matrix. Consequently, the correct dispersion of CNTs in the polymer and timber interface is a crucial factor [43, 51]. The difficulty in dispersing CNTs arises from their inherently hydrophobic nature and tendency to agglomerate due to their size and shape [51]. Different techniques for dispersing them in solvents (acetone, ethanol, …) in addition to ultrasound, mechanical shaking or a combination of the two techniques, have already been tried [43,52]. The interface adhesion can be improved by chemically functionalizing the CNTs surfaces; this generates strong covalent-type of bonding [48, 49, 53, 54, 55]. Referring to the experimentation on wood, CNTs potentially provide a number of advantages: - they are morphologically and chemically compatible with polymer resins used as bonding materials, and also with wood. This is due to the fact that CNTs are anatomically similar to strong pipes bonded with a thermoplastic matrix and equipped with strong dissipative capacity with regards to fracturing energy; - they allow the polymer bonding matrix to improve its own inbuilt deformation capacities; - resin-fibre compounds have excellent mechanical characteristics due to the high specific resistance of the fibres, which ensure great cohesive strength combined with high ductility. This produces significant creep resistance within the composite; - the tubular structure of nanofibre has great permeability to vapour: this is a crucial characteristic when dealing with large surfaces treated with glue, especially in the case of wood, since any accumulations of humidity must be easy to disperse in order to avoid biotic degradation; - CNTs have excellent conductive properties. The objective of the study is to set up a nano-material, based on polymeric resin reinforced with CNTs, for the reinforcement of timber structures. The new material could be used as adhesive, surface consolidant (like in the present case) or as a deep impregnation consolidant in combination with vacuum techniques [47].

4.4.2. Materials and methods The first step of the experimentation assessed the efficiency of different impregnation techniques. CNTs were dispersed by means of an ultrasonic probe for different lengths of time in ethanol and

220

Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings acetone. These solvents are supposed to act as a transport medium to bring CNTs directly into the wood microstructure. Wood specimens with section size of 2.5 cm x 2.5 cm were sunk into the suspensions for different times (Fig. 12). Only natural capillarity action was exploited rather than vacuum or other pressure assisted impregnation techniques. This choice was made in view of future on-site application.

Fig. 12 Impregnation test at different times: progression of the solution absorption (from [45]) The following wood species were selected for the analysis: Fir (Abies alba Mill.), Douglas-Fir (Pseudotsuga menziesii), Oak (Quercus robur) and Larch (Larix decidua Mill.). Some samples were obtained from new timber, others from 18th century structures (oak). Multiwall carbon nanotubes (MWCNTs) were used for the process. For economical reasons, the Nanocyl® series 7000 [56], not functionalized (Fig 13), was used to set up the process, while series 3101 functionalized with carboxyl groups were used for mechanical tests (Nanocyl®-3101 series are purified to greater than 95% carbon and then functionalized with COOH groups).

a)

b)

c)

Fig. 13 TEM image (a) and CNTs characteristics of Nanocyl® series 7000 (b) (from [56]. In the subsequent phase a two component solvent-free epoxy resin (Mapei Epojet®) [57] was used for the dispersion in order to obtain a product for mechanical improvement that could be applied on the wood surface by painting, or act as a reinforced glue to connect different timber parts. The premeasured portions (Part A = resin and Part B =hardener) must be mixed together before using them. The mix ratio between part A and B was 4 to 1. This resin was selected because of its low Brookfield viscosity (respectively 500 and 320 mPa.s for part A and B). The efficiency of the impregnation procedure was assessed using the scanning electron microscope (SEM). The overall mechanical improvement of the timber specimen was evaluated through a comparison of mechanical tests performed on untreated samples and on samples impregnated with resin and with CNTs resin coating. The bending strength was evaluated by standardized three point bending tests on small wooden samples, and afterwards on full-scale timber elements [47].

221

Reinforcement of Timber Structures 4.4.3. Carbon nanotubes dispersion In the first stage, the possibility of deep impregnation exploiting natural capillary action in wood has been investigated. Four different suspensions have been prepared with different CNT (Nanocyl® series 7000) concentrations and different solvents (ethanol and acetone). Each wood sample was sunk into these suspensions for twenty-four hours. The specimen was cut and observed with a SEM in order to assess the efficiency of impregnation. The image was compared with untreated wood (Fig. 14a). These tests proved that although the dispersion in the solvent was efficient, and the dimensions of CNTs were compatible with wood porosity, the overall impregnation technique was not working satisfactorily [47], since in some of the dispersions the CNTs remained agglomerated (Fig. 14b).

a)

b)

Fig. 14 SEM image of larch sample (a); Clustering of CNTs in first dispersion (b) [47] Therefore, in a second stage, the experimentation focused on the dispersion of CNTs in polymeric resins: a two-component, low viscosity resin was selected for use for CNT dispersion, with the aim of having a product that could be applied on wooden surfaces by painting. The main difference between epoxy resins and the solvents described above pertains to its higher viscosity. Therefore, it was necessary to find the optimal CNTs content assuming a two hour sonication duration. A CNT weight content equal to 0.3% with respect to the resin (A+B fraction) was adopted after some trials. It was decided to disperse first the CNTs in the resin component B, which has the lower viscosity. Then the two components were mixed together with mechanical stirring at room temperature. The mixture was allowed to cure for six hours at 60°C. Unfortunately, the modified procedure was not sufficient to avoid CNTs clustering [58].

4.4.4. Mechanical characterisation of small scale specimen The overall mechanical improvement of the timber specimen has been evaluated by comparison of mechanical tests performed on untreated samples, on samples impregnated with resin and on samples with CNT resin coating. The bending strength was evaluated with standardized three-point bending tests carried out with displacement control (Fig. 15).

222


a)

b)

Fig. 15 Detail of the three-point bending test (a); Load displacement diagrams for four different unreinforced timber species (b) (from [47]) All the coated samples were cured at room temperature for seven days, according to the manufacturer’s recommendations, and to simulate an in-situ intervention. The preliminary results obtained from the samples tested are in some cases very promising, but in other cases more ambiguous. Future investigations must be carried out on a larger number of specimens to assure statistical significance of results. In the case of the 18th century oak samples (Fig. 16a), an increase of about 19% in the peak load was observed when only resin is used for the coating. If CNTs reinforced resin is used for the coating, the gain in the peak load increases to 42%. On the other hand, as far as the Hemlock-Fir samples is concerned (Fig. 16b), an increase of about 45% is obtained in case of resin coating, regardless the presence of a CNTs reinforcing in the resin.

a)

b)

Fig. 16 Load displacement curves for the 18th cent. Oak samples (a) and Hemlock fir samples (b). Comparison between un-reinforced, resin coated, and CNTs reinforced resin coated(from [47])

The results concerning the assessment of the timber retrofitting with CNTs resin coating are positive, since these provide an increase in the flexural strength [47].

4.4.5. Tests on structural elements The methodology was extended to timber structural elements coming from historical structures. Old decayed timber components of Poplar wood (Populus sp.), belonging to a dismantled timber floor of the Venaria Reale castle (Italy) have been selected for the experimental campaign. In particular, those timber elements were strongly affected by brown rot decay and xylophagous insects attack (Anobium, etc.). Both significantly decreased the mechanical resistance of the wood.

223

Reinforcement of Timber Structures Reinforcement and conservation interventions often require the improvement of the load-bearing capacity of a highly decayed timber component, and often partial substitutions are necessary (e.g. beam heads). A common intervention is to put in place a new wood prosthesis, which must be connected with the old, possibly reinforced, remaining part of the structure. Therefore, the effect of strengthening with CNTs resin impregnation has been studied in combination with the placing of new wooden prostheses, and tests were carried out on nine composite samples [10]. The numerical analysis was also able to identify the weakest point of the joints, where most often a crack initiates and propagates [47]. The main connection typologies of timber joints belonging to the historical constructive technique suitable for rehabilitation of timber elements were considered. After a comparison among the suggested solutions of historical treatises; the so-called “Jupiter” joint was selected. This joint, traditionally used to connect two elements subjected to bending, is based on a particular geometric construction that permits the connection of two elements with a profile very similar to a thunderbolt (Fig. 17). A protocol for collecting data has been defined, which is particularly important when the experimental phase is applied to historic timber structures, as e.g. temperature, moisture content, drilling resistance, and tensile strength. For this last parameter, a preliminary finite element analysis was performed, accounting for the presence of interfaces, and assessing the magnitude of the initial stresses that take place when the joint is connected. The same impregnation method used in the first testing phase on small scale specimens was adopted. Functionalized MWCNTs (Nanocyl® 3101) were adopted for the impregnation of the decayed timber elements, dispersed in the same bi-component polymeric resin with low viscosity (used in the measure of 0.75 wt % in respect of the resin).

Fig. 17 Summary of the combinations adopted Fig. 18 Comparison between the resistance drilling for three point bending tests on timber profiles of the same specimen (specimen n. T1/C3) “Jupiter” joints (from [10]) before and after the impregnation with CNTs reinforced resin (from [10])

224

Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings Special efforts were dedicated to obtain a product that could be applied by brush in case of in-situ application. All the “Jupiter” joints were first characterized by visual grading and non-destructive testing (measurement of the drilling resistance). The characterization was also conducted with destructive mechanical tests. Wood moisture content was 13±2%. Non-destructive characterization was repeated after the impregnation procedure by painting. The painting was applied assuming a constant amount of material applied on each sample. Fig. 18 shows a comparison between the resistance drilling profiles prior and after the impregnation. It was possible to detect penetration levels of the CNTs reinforced resin in the wood of ~5 mm. It is reasonable to assume that, based on the state of degradation of the wood and on the amount of painting, much higher penetration depths can be achieved. Destructive four-point bending tests were carried out to assess the mechanical resistance of the samples. The test geometry was selected to have uniform bending in the joint region. Different combinations (Fig. 17), with or without the impregnation of polymeric resins and of polymeric resins reinforced with CNTs, were considered, in order to verify whether there was an improvement of the structural efficiency. As the tests represent a rehabilitation intervention, realized through an old wood-new wood prosthesis, part of the joints were made with new wooden elements of the same wood species (Poplar) (Fig. 19).

4.4.6. Discussion of results Despite the complexity of variations that influence the results of the experimentation (wood species, typology of joint, degree of decay), the tests carried out on structural elements have confirmed the hypothesis of an increase of mechanical performance. The load displacement curves obtained from the beams connected with “Jupiter” joints have demonstrated an improvement of the mechanical resistance of the samples impregnated with resin and CNTs, both compared to decayed wood (from 25% to 35%) and wood impregnated with resin only (Fig. 20). The applied methodology can constitute a basis for the definition of protocols for use in practice and recommendations for standard testing and measurement procedures for polymeric resins reinforced with CNTs for the rehabilitation of timber structures. The proposed intervention technique refers to localised structural interventions on some decayed timber parts or connections and not to an overall intervention. This allows a possible re-tractability of the interventions. The results concerning the assessment of the timber retrofitting with CNTs resin coating are positive, since they provide an increase in the flexural strength which is greater or at least equal to the one obtained with the resin alone. Further investigations are necessary to prove the statistical significance of the results. In addition, permeability of the CNTs resin coating must be studied more in detail, in order to obtain CNTs content that maximizes the coating permeability to vapour. The results obtained appear to be promising, and the research on the procedure is still in progress to improve the innovative technique and considering a larger number of samples [47].

225


Fig. 19 Decayed wooden samples; bending tests and scheme for the mechanical resistance on timber “Jupiter” joints, detail of the brown rot decay (L=1000mm h=100mm B=35mm) (from [10])

Fig. 20 Load displacement diagram: positive effect of reinforcement with resin and CNTs [10]

5. Conclusions Wood, as a natural nano-composite, can potentially offer important applications in the field of nanotechnologies. Together with the performances commonly required (long term stability, durability and weather resistance, good adhesion to the substrate, transparency, sustainability for the production process, etc.), it is possible, through nano-structured coatings to introduce additional surface functionality such as self-cleaning, photocatalysis, water resistance, fire 226


resistance, scratch resistance, graffiti resistance, and antibacterial coatings. The main nano-materials already existing in the market, nano-coatings and their wood surface protection functions, classification and compatibility with wood have been reviewed in this chapter. On-going experimental research projects for potential new fields of application in the sector of architecture, civil engineering and cultural heritage have also been analysed with a special focus on reinforcement of historical timber structures. The results of experimental research activity on the application of a polymeric resin reinforced with CNTs on historical timber structures are promising and constitute a base for future developments of the application of these nano-composite materials. Research to assess and model environmental impacts and the economic viability of using nano-scale materials in wood surfaces and timber structures is still needed in most of the applications.

6. Acknowledgements The research on reinforcement of historic timber structures with carbon nano-composites was carried out with the support of Italian Ministry of University and Research, through a PRIN Project (Use of nanotechnologies in Cultural Heritage for the efficiency of maintenance systems in wooden built heritage: innovative technologies for restoration) and with EU Regional funds through the M.A.N. Project (Nanostructured Architectural Maintenance) coordinated by Prof. C. Bertolini Cestari who is gratefully acknowledged.

7. References [1]

Ashby M.F., Ferreira P.J., Schodek D.L., Nanomaterials, nanotechnologies and design. An introduction for Engineers and Architects, Elsevier, Paris, 2009. [2] Bartos P.J.M., De Miguel Y., Porro A. (editors), 2005, Proceedings of the 2nd International Symposium on Nanotechnology in Construction, 13-16 November 2005, Bilbao (Spain), RILEM, Bagneux. [3] Elvin G., “The Nano Revolution. A science that works on the molecular scale is set to transform the way we build”, Architect Magazine, May 2007. [4] Blee A., Matisons J.G., “Nanoparticles and the conservation of cultural heritage”, Materials Forum, Vol. 32, 2008, pp. 121-128. [5] AA.VV., Nanotechnology for the Forest Products Industry. Vision and Technology Roadmap, Lansdowne, U.S.A., October 17-19, 2004, TAPPI Press, Atlanta. [6] Cai Z., Rudie A., Stark N., Sabo R., Ralph S., “New Products and Product Categories in the Global Forest Sector”, In: Hansen E., Panear R., Vlosky R. (Eds.), The Global Forest Sector: Changes, Practices, and Prospects, CRC Press, Boca Raton, 2013, pp. 129-149. [7] Bertolini Cestari C., Marzi T., Seip E., Touliatos P. (editors), Interaction between Science, Technology and Architecture in timber construction, Elsevier, Paris 2004. [8] Kasal, B., Tannert, T., In Situ Assessment of Structural Timber, Springer, 2010. [9] Ansell M. P., “Multi-functional nano-materials for timber in construction”, Proceedings of the ICE, Construction Materials, Vol. 166, Issue 4, 2013, pp. 248 –256. [10] Marzi T., Impiego di nanotecnologie nei beni culturali per l'efficienza di sistemi manutentivi del costruito in legno: tecnologie innovative di recupero (Nanotechnologies / nanosciences: from wood improvement to the reinforcement of timber and monitoring of interventions), Ph.D. Thesis in Innovation Technology for Built Environment, tutor Prof. Bertolini Cestari C., Politecnico di Torino, 2010. [11] Nowack B. and Bucheli T.D., “Occurrence, behavior and effects of nanoparticles in the

227


[12] [13] [14]

[15] [16]

[17]

[18]

[19]

[20] [21] [22] [23] [24] [25]

[26]

[27] [28] [29] [30]

[31]

[32]

228

environment”, Environmental Pollution, Vol. 150, No. 1, 2007, pp. 5–22. www.nanobiz.com (date of access October 2014) The NanoPhos Casebook: applications from across the world, Jan 2013, http://www.nanophos.com/en/ (date of access October 2014) Auclair N., Riedl B., Blanchard V. and Blanchet P., “Improvement of photoprotection of wood coatings by using inorganic nanoparticles as ultraviolet absorbers”, Forest Products Journal, Vol. 61, No. 1, 2011, pp. 20–27. Blanchard V. and Blanchet P., “Color stability for wood products during use: effects of inorganic nanoparticles”, BioResources, Vol. 6, No. 2, 2011, pp. 1219–1229. Allen N.S., Edge M., Ortega A. et al., “Behaviour of nanoparticle (ultrafine) titanium dioxide pigments and stabilisers on the photooxidative stability of water based acrylic and isocyanate based acrylic coatings”, Polymer Degradation and Stability, Vol. 78, No. 3, 2002, pp. 467-478. Vlad-Cristea M., Riedl B. and Blanchet P, “Enhancing the performance of exterior waterborne coatings for wood by inorganic nanosized UV absorbers”, Progress in Organic Coatings, Vol. 69, 2010, pp. 432–441. Vlad-Cristea M., Riedl B. and Blanchet P, “Effect of addition of nanosized UV absorbers on the physicomechanical and thermal properties of an exterior waterborne stain for wood”, Progress in Organic Coatings, Vol. 72, 2011, pp. 755–762. Vlad-Cristea M., Riedl B., Blanchet P. and Jimenez-Pique E, “Nanocharacterization techniques for investigating the durability of wood coatings”, European Polymer Journal, Vol. 48, 2012, pp. 441–453. Germak C., Bozzola M., “Savoir Bois: culture and furniture for the mountain”, Paesaggio Urbano, Vol. 4, 2013, pp. 42-47. Germak C., Bozzola M., Technical Report of the EU Project Alcotra 2007 “Savoir Bois”, Politecnico di Torino, 2013. Nowack B., Krug H.F. and Height M.,” 120 years of nanosilver history: implications for policy makers”, Environmental Science and Technology, Vol. 45(4), 2011, pp. 1177–1183. Rai M., Yadav A. and Gade A., “Silver nanoparticles as a new generation of antimicrobials”, Biotechnology Advances, Vol. 27(1), 2009, pp. 76–83. Kaegi, R., Sinnet B., Zuleeg S. et al., “Release of Silver nanoparticles from outdoor facades”, Environmental Pollution, Vol. 158(9), 2010, pp. 2900–2905. Kartal S.N., Green F. and Clausen C.A., “Do the unique properties of nano-metals affect leachability or efficacy against fungi and termites?”, International Biodeterioration and Biodegradation, Vol. 63, No. 6, 2009, pp. 490–495. Bertolini C., Crivellaro A., Marciniak M., Marzi T., Socha M., “Nanostructured materials for durability and restoration of wooden surfaces in architecture and civil engineering”, In: Proceedings of 11th World Conference on Timber Engineering, 2010, Vol. III, pp. 705-706. Lowden L.A., Hull T.R., “Flammability behaviour of wood and a review of the methods for its reduction”, Fire Science Reviews, Vol. 2, No. 4, 2013. Taghiyari H.R., “Fire-retarding properties of nano-silver in solid wood”, Wood Science and Technology, Vol. 46, 2012, pp. 939-952. www.percenta.com (date of access October 2014) Dei L. and Salvadori B., “Nanotechnology in cultural heritage conservation: nanometric slaked lime saves architectonic and artistic surfaces from decay”, Journal of Cultural Heritage, Vol. 7, No. 2, 2006, pp.110–115. Daniele V., Taglieri G. and Quaresima R, “The nanolimes in cultural heritage conservation: characterisation and analysis of the carbonatation process”, Journal of Cultural Heritage, Vol. 9, No. 3, 2008, pp. 294–301. Baglioni P. and Giorgi R., “Soft and hard nanomaterials for restoration and conservation of cultural heritage”, Soft Matter, Vol. 2, no. 4, 2006, pp. 293–303.

Nanotechnologies for reinforcement and protection of timber structures: innovative nano-coatings [33] Baglioni P., Chelazzi D., Giorgi R., “Nanoparticles of calcium hydroxide for wood conservation. The deacidification of the Vasa warship”, Langmuir, Vol. 21, 2005, pp. 1074310748. [34] Kaiser J-P, Zuin S., Wick P., “Is nanotechnology revolutionizing the paint and lacquer industry? A critical opinion”, Science of the Total Environment, Vol. 442, 2013, pp. 282-289. [35] Hischier R., Walser T., “Life cycle assessment of engineered nanomaterials: State of the art and strategies to overcome existing gaps”, Science of the Total Environment, Vol. 425, 2012, pp. 271-282. [36] “NanoHouse Dissemination Reports”, http://www-nanohouse-project.eu [37] Hetzer M., De Kee D, “Wood/polymer/nanoclay composites, environmentally friendly sustainable technology: A review”, Chemical Engineering Research and Design, Vol. 86, 2008, pp. 1083-1093. [38] Avella M., Cocca M., Errico M.E., Gentile G., “Nanotechnologies and Nanosensors: Future Applications for the Conservation of Cultural Heritage”, In: J.P. Reithmaier, P. Paunovic, W. Kulish, C. Popov, P. Petkov (edited by), Nanotechnological Basis for Advanced Sensors, NATO Science for Peace and Security Series B: Physics and Biophysics, Springer, 2011. [39] Winkler C., Schwarz U., “Multifunctional Wood-Adhesives for Structural Health Monitoring Purposes”, Materials and Joints in Timber Structures, RILEM Bookseries, Volume 9, 2014, pp. 381-394. [40] Thostenson, E.T., Chou, T. W., “Real-time in situ sensing of damage evolution in advanced fiber composites using carbon nanotube networks”, Nanotechnology, Vol.19, 215713, 2008. [41] Wehnert, F., Heinrich, J., Jansen, I., “Multifunctional adhesives by integration of carbon nanotubes”, In: EURADH 2012 - 9th European Adhesion Conference, 2012. [42] Wipf T., Phares B. M., Ritter M., Literature Review and Assessment of Nanotechnology for Sensing of Timber Transportation Structures Final Report, United States Department of Agricolture, Forest Service, Forest Products Laboratory, General Technical Report FPL–GTR– 210, April 2012. [43] Breuer, O., Sundararaj, U., “Big return from small fibers: A review of polymer/carbon nanotube composites”, Polymer Composites, Vol. 25, No. 6, 2004, pp. 630-645. [44] Ahmad Z., Ansell M.P., Smedley D., “Epoxy adhesives modified with nano- and microparticles for in-situ timber bonding: effect of microstructure on bond integrity”, International Journal of Mechanical and Materials Engineering, Vol. 5, 2010, pp. 59-67. [45] Bertolini Cestari C., Invernizzi S., Marzi T., Tulliani J.M., “Nanotechnologies applied to the restoration and maintenance of wooden built heritage”, In: D’Ayala D., Forde E. (editors), Structural analysis of historical construction. Preserving safety and significance, Proceedings of SAHC2008. VI International Conference, Bath; July 2008, Taylor & Francis, London, 2008, pp. 941-947. [46] Bertolini Cestari C., Invernizzi S., Marzi T., Tulliani J.M., “Use of nanotechnologies and nanosciences in cultural heritage for the efficiency of maintenance systems in wooden built heritage: restoration, conservation, maintenance, monitoring of interventions”, In: Structures en bois dans le patrimoine bâti, Actes des journées techniques intern. Bois, Les cahiers d’ICOMOS France, Icomos, Paris, 2009, pp. 87-91. [47] Bertolini C., Invernizzi S., Marzi T., Tulliani J.M., “The reinforcement of ancient timber-joints with carbon nano-composites”, MECCANICA, Vol. 48, 2013, pp. 1925-1935. [48] Gojny FH, Wichmann MHG, Köpke U, Fiedler B, Schulte K., “Carbon nanotube-reinforced epoxy-composites: enhanced stiffness and fracture toughness at low nanotube content”, Composites Science & Technology , Vol. 64, 2004, pp. 2363–2371. [49] Lau K.T., Lui D., “Effectiveness of using carbon nanotubes as nano-reinforcements for advanced composite structures”, in Carbon, Vol. 40, No. 9, 2002, pp. 1597-1617. [50] Yeh, M.-K., Hsieh, T.-H, Tai, N.-H, “Fabrication and mechanical properties of multi-walled carbon nanotubes/epoxy nanocomposites”, Materials Science and Engineering, Vol. 483–484, 2008, pp. 289–292. [51] Chakraborty AK, Plyhm T, Barbezat M, Necola A, Terrasi GP, “Carbon nanotube (CNT)-epoxy

229


[52]

[53] [54]

[55]

[56] [57] [58]

230

nanocomposites: a systematic investigation of CNT dispersion”, J Nanoparticles Research, Vol. 13, 2011, pp. 6493-6506. Chen H., Jacobs O., Wu W., Rüdiger G., Schädel B., “Effect of dispersion method on tribological properties of carbon nanotube reinforced epoxy resin composites”, Polymer Testing, Vol. 26, 2007, pp. 351–360. Miyagawa H., Drzal L.T., “Thermo-physical impact properties of epoxy nanocomposites reinforced by singlewall carbon nanotubes”, Polymer, Vol. 45, 2004, pp. 5163–5170. Lau K., Lu M., Lam C., Cheung H., Sheng F., Li H. (2005), “Thermal and mechanical properties of single-walled carbon nanotube bundle-reinforced epoxy composites: the role of solvent for nanotube dispersion”, Composites Science & Technology, Vol. 65, 2005, pp. 719– 725. Miyagawa H., Rich M.J., Drzal L.T., “Thermophysical properties of epoxy nanocomposites reinforced by carbon nanotubes and vapour grown carbon fibers”, Thermochim Acta, Vol. 442, 2006, pp. 67–73. http://www.nanocyl.com (date of access October 2014) http://www.mapei.it/Referenze/Multimedia/367_Epojet_GB.pdf (date of access October 2014) Bertolini Cestari C, Invernizzi S., Marzi T., Tulliani J.M., “Nano-technologies/Smart-materials in timber constructions belonging to cultural heritage”, In: Proceedings of 11th World Conference on Timber Engineering, 2010, Vol. IV, pp. 761–762.


Outlook Dietsch, P., Harte, A.M.

Wood features a multitude of positive characteristics in view of its application as building material and it is undisputed that good design of timber structures should still aim at minimizing stresses for which timber only features small capacities and brittle failure mechanisms (e.g. tensile stresses perpendicular to the grain and shear), thereby avoiding or at least minimizing the necessity for reinforcement. Despite this, there is still a multitude of reasons necessitating the structural reinforcement of timber buildings, not only in existing buildings but also for new structures. In this report, current and emerging methods that are available to repair or enhance the structural performance of timber structures are summarised. Criteria for selection and examples are given for the implementation of the various reinforcement methods. For all construction works it has to be verified that essential requirements like mechanical resistance, stability and safety in use are met. The required performance is commonly verified by complying with corresponding harmonized technical rules for the structural design as well as products used in construction works. In cases where harmonized technical rules or technical approvals are not available, an approval in the individual case or comparable have to be sought. Various reinforcement methods presented in this report still lack harmonized technical rules. According to the European position on future standardization, harmonized technical rules shall be prepared for “common design cases” and shall contain “only commonly accepted results of research and validated through sufficient practical experience”. The target audience for such rules is “competent civil, structural and geotechnical engineers, typically qualified professionals able to work independently in relevant fields”. When preparing items for standardization it is important to take into account the approach within the internationally accepted system for standardization. Comparable to the approach taken by scientists when tackling a problem - (1) methods, (2) materials (3) results - the system for standardization is based on a 3-step-pyramid. This pyramid is based on (1) test standards (containing rules on how to test products). Relating to these, product standards (2) are developed (giving strength and stiffness parameters, boundary conditions and rules for production and quality control). The design standards (3) represent the tip of this pyramid (providing design equations and formulating specific requirements in e.g. spacing, edge distance, minimum anchorage length, etc.). When developing design rules it a precondition to also develop (1) test procedures as well as (2) a product standard on the product or system used. Without the latter, rules in a design standard cannot be used since the basic parameters are missing, in other words, the pyramid will not be complete if one element is missing.

231


List of contributors Laurent Bléron


Jean-François Bocquet


Jorge M. Branco

ISISE, Dept. Civil Eng., University of Minho Guimarães, Portugal

Reinhard Brandner

Institute of Timber Engineering Graz University of Technology, Austria

Wen-Shao Chang

Department of Architecture and Civil Engineering, University of Bath, UK

Jian-Fei Chen

Queens University Belfast, Belfast, Northern Ireland, UK

Thierry Descamps

URBAINE, Dept. of Structural Mech. and Civil Eng., University of Mons, Mons, Belgium

Philipp Dietsch

Chair of Timber Structures and Building Technische Universität München, Germany

Bettina Franke

Bern University of Applied Sciences

and

Biel/Bienne, Switzerland Steffen Franke

Bern University of Applied Sciences Biel/Bienne, Switzerland

Alessandra Gubana

University of Udine Udine, Italy

Annette Harte

College of Engineering and Informatics National University of Ireland, Galway, Ireland

Robert Jockwer

ETH Swiss Federal Institute of Technology, Zurich, Switzerland

Robert Kliger

Chalmers University of Technology, Gothenburg, Sweden

232

Wood

Technology

Construction,

Reinforcement of Timber Structures Damien Lathuillière


Tanja Marzi

Politecnico di Torino, Department of Architecture and Design, Torino, Italy

Daniel McPolin

Queen’s University Belfast, Belfast, UK

Caoimhe O’Neill

Queen’s University Belfast, Belfast, Northern Ireland, UK

Maria Adelaide Parisi

Politecnico di Milano, Milano, Italy

Maurizio Piazza

Università degli Studi di Trento Trento, Italy

Benedetto Pizzo

CNR-IVALSA, Trees and Timber Institute, Sesto Fiorentino (FI), Italy

Rajčić Vlatka

University of Zagreb, Zagreb, Croatia

Kay-Uwe Schober

Mainz University of Applied Sciences Mainz, Germany

Erik Serrano

Structural Mechanics Lund University, Sweden

Dave Smedley

Rotafix Ltd. Abercrave, Swansea, UK

René Steiger

Swiss Federal Laboratories for Materials Testing and Research (EMPA), Dübendorf, Switzerland

Mislav Stepinac


Robert Widmann


Qingfeng Xu

Shanghai Research Institute of Building Sciences, Shanghai, China

233


List of reviewers Martin Ansell

Department of Mechanical Engineering, University of Bath, UK

Francisco Arriaga Martitegui

Departamento de Ingeniería y Gestión Forestal y Ambiental, Universidad Politécnica de Madrid, Spain

Jorge Branco

ISISE, Deptartment of Civil Engineering, University of Minho, Guimarães, Portugal

Clara Certari Bertolini

Department of Architecture and Design, Politecnico di Torino, Italy

Helena Cruz

LNEC, Timber Division, Lisbon, Portugal

Dina D’Ayala

Civil, Environmental and Geomatic Engineering, University College London, UK

Philipp Dietsch

Chair of Timber Structures and Technische Universität München, Germany

Steffen Franke

Bern University of Applied Sciences, Biel/Bienne, Switzerland

Annette Harte

College of Engineering and Informatics, National University of Ireland, Galway, Ireland

André Jorissen

Technical University of Eindhoven, Eindhoven, Netherlands

Bo Kasal

Department of Organic Materials and Faculty of Wood Materials, TU Braunschweig, Germany

Robert Kliger

Chalmers University of Technology, Gothenburg, Sweden

Daniel McPolin

Queen’s University Belfast, Belfast, UK

Maurizio Piazza

Università degli Studi di Trento, Trento, Italy

Vlatka Rajčić


Jakub Sandak

National Research Council, Institute of Tree and Timber IVALSA, Italy.

Erik Serrano

Structural Mechanics, Lund University, Sweden

René Steiger

Swiss Federal Laboratories for Materials Testing and Research, (EMPA), Dübendorf, Switzerland

Jean Marc Tulliani

Department of Applied Science and Technology, Politecnico di Torino, Italy

Robert Widmann


David Yeomans

Consulting Engineer, Banbury, UK

234

Building

Construction


235

Reinforcement of Timber Structures

Reinforcement of Timber Structures

Suggest Documents

Self-tapping screws as reinforcement for timber structures - RiuNet

handbook 1 - timber structures

Large-Span Timber Structures

Bending Reinforcement of Full0Scale Timber Beams ...

Earth reinforcement and soil structures

Galvanized Reinforcement In Concrete Structures

Design for Robustness of Timber Structures

Fibres for strengthening of timber structures - CiteSeerX

Basis of Design Principles for Timber Structures

Service life modelling of timber structures | SpringerLink

Development Of Timber Reciprocal Frame Structures ...

Diagnosis of Timber Structures and Archaeological ...

structural health assessment of timber structures

Design of shear reinforcement for timber beams - Lehrstuhl fÃ¼r

Strength Estimation and Reinforcement of Glue-Laminated Timber ...

Nathalie Labonnote Damping in Timber Structures - Core

Woods, Timber structures, Preservative treatment - Scientific ...

Innovation in Timber Architectural Structures and ...

seismic vulnerability assessment for timber roof structures

seismic vulnerability assessment for timber roof structures

Guide to timber cladding - Outdoor Structures

FRP reinforcement in RC structures

Innovative reinforcement for fabric formed concrete structures

the ancient timber structures of the royal palace of naples