Strength of axially loaded screw joints in wooden structures exposed to fire

(1)

Strength of axially loaded screw joints in

wooden structures exposed to fire

An overview of existing knowledge and proposal for test method

Emma Wiklund

Fire Protection Engineer, bachelor's level 2021

Luleå University of Technology

(2)

Strength of axially loaded screw joints in

wooden structures exposed to fire

An overview of existing knowledge and proposal for test method

Emma Wiklund Luleå, April 2021

The fire protection engineering program

Department of Civil, Environmental and Natural Resources engineering Luleå University of Technology

(3)

Title Strength of axially loaded screw joints in wooden structures – An overview of existing knowledge and proposal for test method

Author Emma Wiklund

Internal supervisor Michael Försth

External supervisor Kristoffer Malm TK Botnia

Examiner Alexandra Byström

Bachelor thesis 15 credits

The fire protection engineering program Luleå University of Technology

(4)

I

PREFACE

This work constitutes my bachelor degree thesis in the fire protection engineering program at Luleå University of Technology and was carried out during the fall in 2020 to the winter in 2021.

I hope that my work will spark curiosity and interest in the complexity of fire exposed wooden structures and mainly their connections. The project met great obstacles during the pandemic, but thanks to the wonders of digital communication platforms, a smooth and effective work process could be achieved.

To complete this thesis, I have been dependent on the participation of others. I would therefore like to extend a big thanks to everyone involved.

Thank you Michael Försth, my supervisor, who have been greatly engaged in the topic, my work, and my personal academic development. Without your cheering, positive attitude, and innovative thinking along with your constructive feedback and insights, my work would have been so much harder.

Big thanks to Peter Jacobsson at Martinsons Trä and Kristoffer Malm at TK Botnia for the valuable discussions and feedback of how my work may be applicable in the industry. The feeling that my work can contribute to future development in the industry is unvaluable. I want to thank all the respondents who participated in the interviews for this thesis. To hear your thoughts and experiences in the subject have been interesting and rewarding in the process of carrying out a comprehensive and well-founded thesis.

Also, I would like to thank my current workplace Sweco Sweden, for the opportunity of writing this thesis in an extraordinary work environment with valuable technical assets, motivational surroundings and (most importantly) several needed coffee breaks.

I also would like to thank Ludvig Swedberg for the proofreading of this thesis and for the overall support that you have given me during my work.

Stockholm 2021-04-06

(5)

II

ABSTRACT

Today, distinct approaches are required when designing timber structures and their joints in the event of fire. Considering that the steel fasteners have temperature dependent properties and that the surrounding wooden construction is also affected by the fire due to occurrence of charring during fire exposure, these types of joints are complicated to design.

This study includes an overview of existing knowledge in the subject, where a literature survey has been conducted to enlighten current design procedures, previous research, and available test methods for axially loaded screw joints in wooden structures exposed to fire. The literature survey has been supplemented with an interview survey, in which six respondents’ knowledge in the subject and desires for future studies have played the main role.

The results from this study clearly indicates that the existing knowledge regarding axially loaded screw joints in wooden structures exposed to fire is insufficient. Today’s standards for the design of these types of joints are incomplete with respect to the referred joint arrangement and there are no explicit test methods for determination of the strength of these joints under fire exposure. A few previous studies have been carried out and the results from these have proven to be difficult to interpret since the experimental set-up was not consistent between the fire tests. This complicates the process of determining which factor has the greatest impact on the strength of the joint and the main underlying causes generating a collapse of the structure. The interview survey further strengthens the perception, based on the literature survey, that there is a lack of knowledge in the subject.

(6)

III

SAMMANFATTNING

Idag krävs tydliga tillvägagångssätt vid design av träkonstruktioner och dess förband i händelse av brand. Med hänsyn till att fästelementen av stål har temperaturberoende egenskaper samt att även den omgivande träkonstruktionen påverkas av en eventuell brand och förkolnar vid brandexponering, är dessa typer av förband komplicerade att designa.

Denna studie omfattar en kartläggning av befintlig kunskap i ämnet där en litteraturstudie har genomförts för att belysa aktuella designförfaranden, tidigare forskning och tillgängliga testmetoder för axiellt belastade skruvförband i brandutsatta träkonstruktioner. Litteraturstudien har kompletterats med en intervjustudie där sex respondenters kunskap i ämnet och önskemål för framtida studier har spelat huvudrollen.

Resultatet från denna studie visar tydligt att den befintliga kunskapen om axiellt belastade skruvförband i brandutsatta träkonstruktioner är bristfällig. Dagens standarder för design av dessa typer av förband är ofullständig med hänsyn till direkt brandexponerade skruvar samt att det inte finns explicita provningsmetoder för bestämning av dessa förbands hållfasthet vid brandpåverkan. Ett fåtal tidigare studier i ämnet har genomförts och resultaten från dessa visar sig vara svårtolkade på grund av att försöksuppställningen inte varit konsekvent mellan de genomförda brandprovningarna, vilket därför försvårar processen av att bestämma vilken faktor som har störst inverkan på förbandets hållfasthet samt de viktigaste bakomliggande orsakerna till brott. Intervjustudien stärker ytterligare uppfattningen, baserad på litteraturstudien, om att kunskap i ämnet saknas.

(7)

IV

NOMENCLATURE

Latin upper-case letters

Ad Accidental load [kN/m2]

D Hole diameter [mm]

E_fi,d,t Design effects of actions at a specific time [kN]

𝐹 Applied load [N]

F_ax,α,Rk Pull through resistance/characteristic withdrawal capacity [N]

F_ax,Rk Axial load capacity [N]

F_max Maximum withdrawal load [N]

F_t,Rk Tensile resistance [N]

G_k Permanent load [kN/m2_]

Gk j,sup Unfavorable permanent load [kN/m2]

G_{k j,inf} Favorable permanent load [kN/m2_]

R_d Design load carrying capacity, ambient temperature [kN]

R_{d,t,fi/fi,d,t} Design resistance of screw joint in fire [kN]

R_k Characteristic mechanical resistance of a connection, ambient temperature [kN]

R20 20% fractal value of mechanical resistance [kN]

Qk Variable load [kN/m2]

Latin lower case letters

a_1,2,3 Edge distances of screw [mm]

(11)

VIII

d_h Diameter of screw head [mm]

d₁ Inner threaded diameter of the screw [mm]

fax,k / f1,k Characteristic withdrawal strength perpendicular to the fiber [N/mm2]

fhead,k / f2,k Characteristic pull through parameter [N/mm2]

f_tens,k Characteristic tensile capacity [kN]

k Parameter [-]

k_ax Conversion factor, relationship between screw and fiber direction [-]

kd Dimension factor of the screw [-]

kfi Timber fire coefficient [-]

k_flux Heat flux coefficient for fasteners [-]

k_mod Modification factor, timber [-]

ℓ Total length of screw [mm]

ℓ_d Penetration length of screw, initial [mm]

ℓ_ef Penetration length of the threaded part of the screw [mm]

ℓg Threaded length of the screw [mm]

n Number of screws [-]

nef Effective number of screws [-]

t Panel thickness [mm]

t Time [min]

t_d,fi/fi,dDesign fire resistance of unprotected connection [min]

t_req/fi,requRequired time of resistance [min]

Greek letters

(12)

IX

β₀ Design one-dimensional charring rate [mm/min]

β_n Design notional charring rate [mm/min]

γ_M Partial factor, timber [-]

γM,fi Partial safety factor for timber in fire [-]

θcr,d Design value of the critical temperature of the material [°C]

θ_d Design value of the material temperature [°C]

θg Gas temperature in the fire compartment [°C]

η Conversion factor [-]

η0 The degree of utilization in normal temperatures [-]

ηfi Reduction factor for the design load, fire temperatures [-]

ρ Characteristic density [kg/m3_]

ρ_a Density [kg/m3_]

(13)

X

ABBREVIATIONS

CEN European Committee for standardization EEA European Economic Area

EU European Union

EKS Europeiska konstruktionsstandarder (European construction standards)

FPL Forest product laboratory

ISO International Standard Organization LVL Laminated veneer lumber

(14)

1

1 INTRODUCTION

This chapter contains the introduction to the conducted study made in this thesis. Background, aim, and delimitations are presented along with the questions at hand and their motives.

1.1 Background

Today, due to rapid and significant climate change, an environmental awareness has grown strong in society. This consciousness takes hold, for example, by attempting to make more sustainable choices both as an individual and in business related activities. The construction industry also strives for a more sustainable approach to building, both in material selection and the construction process. Timber, known to be a renewable material, contributes to a reduced carbon dioxide content in the atmosphere (EIA, 2019) when used as an alternative to less sustainable building materials and is a popular building material in today’s constructions. The fact that the ban on building timber structures in more than two stories was lifted in 1994 (Brandskyddsföreningen, n.d.) and the inclusion of an increased construction in timber in the “January agreement” (Government of Sweden, 2019), may also have had a relevant impact on the increased demand in timber structures.

Timber constructions have a relatively negative history out of a fire perspective but the knowledge of how these types of structures are to be dimensioned, with aspect to a reliable fire resistance, have increased over time (Svenskt Trä, 2014). Timber is a combustible material and is weak to fire; therefore a thorough design of timber elements and constructions is crucial for keeping stabilizing properties during a potential fire scenario. As important is the design of the joints that keep the structure together: A satisfactory design of the timber elements is not enough for the structure’s total strength if the joints do not fulfill an equivalent strength in case of fire. The design of the joints is not only crucial in fire design, but also for design in ambient temperature. Früwald Hansson (2011) points out that in 23% of the cases where a timber structure collapsed in ambient temperatures, failure in joints was the cause. Of these joint failures, 57% were of dowel-type joints i.e. nails, screws, bolts, dowels etc.

(15)

2

brown area is fresh wood and the dark brown area corresponds to charred wood. The heat transfer from the heated fastener and the fire itself is illustrated as red arrows in Figure 1.

Figure 1 Heat transfer via conduction from a metal fastener into a timber element exposed to fire. Illustration: Emma Wiklund

1.1.1 Designs

There are several different mounting designs for fasteners in timber structures, one of them is shown in Figure 2 where a horizontally mounted screw is supplemented with a screw in a 45° angle. In this design, the inclined screw is longer than the horizontally mounted screw due to the greater penetration length needed due to inclination.

Figure 2 Structure with both horizontal and inclined mounted fasteners. Illustration: Emma Wiklund

(16)

3

force and bending moment. Additionally, a greater friction between the wooden elements, due to the increased contact pressure created from the additional inclined fastener, occurs in the load situation. Figure 3 shows a principle sketch where an external axial force yields bending moment and lateral force in a horizontal screw, whereas the inclined screw is exposed to additional axial force.

In the construction shown in Figure 2, a potential fire can occur at the side from where the fasteners are mounted, which may lead to an increased risk of failure in the connection due to the exposure of the connections, because of the increased charring of the fire exposed side and the heat transfer through the fastener itself.

1.1.2 Connector types

Screws are a common fastener type used in timber constructions due to their advantages in simple mounting and that no predrilling is needed, like for fasteners with higher diameter i.e. bolts, nails and dowels. Figure 4 shows different types of fasteners used as timber connectors.

(17)

4

Several studies have been made where timber structures with fastener joints i.e. bolts, nails, dowels, screws etc., have been tested in order to evaluate the fire resistance in these types of constructions. Buchanan & Abu (2017) points out that research projects have been conducted to determine the fire resistance in lightweight constructions with nailed connections. The authors also enlighten the absence of knowledge in fire performance of screwed connections, but also compares the screws capacity in a fire scenario as,

“[…] but it will generally be better than for nails.”

(Buchanan & Abu, 2017)

based on its similar properties to nails and the better gripping capacity due to the screws threaded shaft.

Axially loaded screw joints in wooden structures exposed to fire is a relatively uncharted area, where only simplified rules for the load capacity of these connections can be found in Eurocode 5.

1.2 Aim and objectives

The aim of this report is to present an overview of current design procedures, results from earlier research and studies regarding the strength of axially loaded screw joints in timber structures in a fire scenario. The expectation is to present the current state of the art of the subject: To enable further research an identification of needs and gaps in knowledge will be illuminated along with proposals of experimental methods for testing and verification to meet the identified needs.

1.3 Research questions

This thesis is expected to answer the following research questions:

1. What studies and research have been done regarding the strength of axially loaded screws in timber constructions exposed to fire?

2. How does the construction industry design axially loaded screw joints today? 3. What needs exist for further studies?

4. What experimental method would be suitable for carrying out further tests?

1.4 Delimitations

The following delimitations are applied for this study:

(18)

5

1.5 Notes to the reader

In this report, different screw types occur in the presented illustrations e.g. double threaded- and fully threaded screws. The fully threaded screws are illustrated in the available standard documents and therefore, illustrations from these documents include fully threaded screws. Double threaded screws are a common screw type used in the industry and therefore the illustrations describing structure designs include double threaded screws. Note that the calculation procedures presented in this report is applicable for both screw types regarding dimensions and thread length etc.

(19)

6

2 RESEARCH STRATEGY

This study was conducted as a normative study, described by Björklund & Paulsson (2012) to be an approach used when some knowledge and understanding exists in the research area and the aim is to guide and suggest further actions in the subject. Therefore, both literature survey and interviews are part of the research strategy for this study to enlighten today’s knowledge and needs for further research. Below, each research question is presented followed by the research strategy chosen to answer them.

1. What studies and research have been done regarding the strength of axially loaded screws in timber constructions exposed to fire?

2. How does the construction industry design axially loaded screw joints today? 3. What needs exist for further studies?

4. What experimental method would be suitable for carrying out further tests?

To answer the first research question, a mapping of today’s knowledge in the field of timber constructions was necessary. A collection of literature regarding basic timber structures in normal temperatures and timber structures exposed to fire temperatures was made. Fire exposed timber structures their fasteners i.e. screw and dressing were under scrutiny.

Interviews complemented the literature survey to answer the second and third research question. Later, the fourth research question was able to be answered using inspiration from earlier test methods as well as recommendations from interviewed respondents and the supervisors of this work.

2.1 Literature survey

Initially a literature survey was conducted to collect necessary information for the study. The literature survey included national and international standards in building and testing as well as earlier reports and articles related to the subject. This literature was collected to get a theoretical basis of the study.

2.1.1 Data processing

Through the literature survey secondary data was obtained. Secondary data is described by Björklund & Paulsson (2012) as information produced for a different purpose than the existing purpose of the current thesis.

(20)

7

2.2 Interview survey

Interviews were conducted with representatives of several disciplines to collect primary data and map the knowledge in the field of current research. Interview guidelines and inspiration from former interviews in scientific research were applied to gather an appropriate basis regarding interview technique.

2.2.1 Respondents

The respondents were chosen by their associated organization and profession with the aim to get a broad vision on today’s knowledge and potential future needs. Interviews were conducted with representatives from standard organizations and research institutes whose primary mission is to develop and investigate contemporary construction methods. Constructors who apply today’s knowledge in their structural design were also interviewed, to identify specific knowledge gaps. The associated organizations involved in this thesis are presented in Table 1 with respective abbreviations used in the thesis. In Table 2, a description of each respondent is presented.

Table 1 Organizations and abbreviations

Organization City, country Abbreviation

Chalmers University of Technology Gothenburg, Sweden Chalmers

Forest Product Laboratory Madison, United states FPL

Martinsons trä Umeå, Sweden Martinsons

Research Institutes of Sweden Borås, Sweden RISE

Swedish Standard Institute Stockholm, Sweden SIS

Tallinn University of Technology Tallinn, Estonia TalTech

(21)

8

Table 2 Description of each respondent

Respondent Organization Profession

Alar Just RISE Researcher

TalTech Professor in timber structures Researcher

SIS Author Eurocode 5

Robert Jockwer Chalmers Professor in timber structures Researcher

SIS Author Eurocode 5

Laura Hasburgh FPL Researcher

Kristoffer Malm TK Botnia Constructor consultant

Annika Stenmark SIS Standardization project manager in timber Peter Jacobsson Martinsons Trä Development manager

2.2.2 Interview guide

An interview guide was created before the interviews took place in order to minimize the risk of leading questions, which can be the case with no pre-defined nor objective approach from the interviewer (Ejvegård, 2003). The interview guide included predefined questions, their order, and possible supplementary questions to these to clarify the interview flow. The interview guide is presented in APPENDIX A.

The interviews were semi-structured where the predefined questions were asked to the respondent with no given answer options. This semi-structured interview technique generated free answering from the respondents and the possibility of question modification and supplementary questions, depending on the development of the interview. Because of this strategy no interview was the same as the other.

2.2.3 Invitation letter

An invitation letter including information regarding the project and the interview procedure, was sent to the respondents to increase the will of participation but also gave the respondents a possibility to prepare before the interview.

The invitation letter can be reviewed in APPENDIX B.

2.2.4 Interview execution

(22)

9

information like the one presented in the invitation letter and was then followed by the actual interview. The interviews were recorded and then transcribed, with permission from the respondents.

2.2.5 Data processing

The data obtained from the interviews are classified as primary data e.g. data specifically produced for this thesis (Björklund & Paulsson, 2012).

The information collected from the interviews were processed by finding relationships in the answers from the respondents. The research question regarding reaching a suitable research strategy for future tests was considered by comparing and evaluate the suggested testing methods proposed by the respondents.

2.3 Research strategy criticism

According to Björklund & Paulsson (2012) it’s important to evaluate the chosen research strategy and constantly show awareness regarding the selected strategy. The awareness is fundamental in the process of reaching the aim of the study given the available resources. The credibility of the study can be described in three measurements: validity, reliability and objectivity (Björklund & Paulsson, 2012).

Validity is defined as:

“Validity, the extent to which a measuring instrument measures what is intended to be measured.”

Translation of The Swedish National Encyclopedia (n.d.,a)

(23)

10 Reliability is defined as:

“Reliability, in the behavioral sciences measures of how strongly or reliably measured values e.g. a test or experiment is.”

Translation of The Swedish National Encyclopedia (n.d.,b)

To obtain reliable results the collected literature was criticized before the information was used. The literature used in this thesis consisted of scientific articles, research reports, government documents etc. to get a reliability in the results from the literature survey. However, there may be some existing literature that have not been collected: This literature could potentially have affected the results from the survey. The results from the conducted interviews increased the reliability of the study because of the breadth of the respondents but, as well as potential missed literature, there may be additional knowledge in people that have not been interviewed. Objectivity is defined as:

“Objectivity, impartiality (opposite subjectivity), a term used in philosophy, social sciences and general debate in several meanings.”

Translation of The Swedish National Encyclopedia (n.d.,c)

(24)

11

3 THEORY

In the following chapter the fundamental information for this thesis is presented. Description of today’s building standards, related to this thesis, regarding timber, screws, fire, and test methods are presented.

3.1 EN Eurocodes

The Eurocodes are reference design codes in structural design i.e. building, geotechnical, fire, earthquakes, performance, and temporary structures. The member states of the European Union (EU) and the European Economic Area (EEA) are required to accept the EN Eurocodes in their structural design regulations to obtain uniform levels of safety in Europe. Also, the internal market of structural products and technical solutions benefits by the usage of the Eurocodes due to elimination of disparities that hinder free circulation within the market (European Commission, n.d.). In Table 3, each part of the EN Eurocodes with their content is presented.

Table 3 Content of each part of the EN Eurocodes (CEN, 2002b)

Part of EN Eurocodes Content

EN 1990 Basis of structural design

EN 1991 Actions on structures

EN 1992 Design of concrete structures

EN 1993 Design of steel structures

EN 1994 Design of composite steel and concrete structures

EN 1995 Design of timber structures

EN 1996 Design of masonry structures

EN 1997 Geotechnical design

EN 1998 Design of structures for earthquake resistance

EN 1999 Design of aluminum structures

(25)

12

Timber structures and their connections are described in EN 1995-1-1 and EN 1995-1-2 (CEN, 2004a, 2004b) with respect to both normal temperatures and fire situations. Further information about the different designs of timber structures and their connections are presented in sections 3.3 and 3.4 of this report.

3.2 Structural fire design according to EN 1991-1-2 – Actions on

structures

When designing a structure in the fire situation, several factors should be considered and analyzed. According to EN 1991-1-2 “Design of timber structures – Structural fire design” (CEN, 2002a) the fire design procedure include: a selection of an appropriate fire scenarios and specification of the design fire itself, followed by calculation of the temperature in the structural elements along with their structural response when exposed to fire temperatures.

When selecting a design fire, consideration of the further usage must be made e.g. a specific structural fire resistance may be specified by national authorities, which often leads to the usage of the standard fire curve ISO 834. This fire curve is used when analyzing the temperature and fire resistance of the structural member.

The ISO 834 fire curve is described by EN 1991-1-2 (CEN, 2002a) as

𝜃𝑔= 20 + 345 𝑙𝑜𝑔10(8𝑡 + 1)

(1)

where 𝜃_𝑔 is the gas temperature and 𝑡 is the time given in minutes.

When analyzing the mechanical response of the structure, the time 𝑡 used in the temperature analysis shall be used. When verifying the fire resistance of the structure, there are three possible parameters that can be determined e.g. time, strength or temperature. In accordance to EN 1991-1-2 (CEN, 2002a), the resistance can be verified as follows:

Either by

𝑡𝑓𝑖,𝑑 ≥ 𝑡𝑓𝑖,𝑟𝑒𝑞𝑢 (2)

where 𝑡𝑓𝑖,𝑑 is the design value of the fire resistance and 𝑡𝑓𝑖,𝑟𝑒𝑞𝑢 is the required fire resistance,

or

(26)

13

where R_fi,d,t is the design value of the fire resistance at the time t, and E_fi,d,t is the design value of the relevant effects of actions in the fire situation at the time t.

Or by

𝜃𝑑≤ 𝜃𝑐𝑟,𝑑 (4)

where 𝜃_𝑑 is the design value of the material temperature and 𝜃_{𝑐𝑟,𝑑} is the design value of the critical temperature of the material.

3.2.1 Fire load design

A fire scenario is classified as an accidental situation and determination of the design value of the relevant effects of actions in the fire situation, Efi,d,t, is made by a combination of the mechanical actions on the structure in accordance to EN 1990 “Basis of structural design” (CEN, 2002b). The load combination for accidental design situation, described in EN 1990 (CEN, 2002b) include the design values of the factors:

- 𝐺_𝑘, the permanent load e.g. the self-weight of the structure - 𝐴_𝑑, the accidental load e.g. fire- or explosion loads

- 𝑄_𝑘, the variable load e.g. snow loads or wind actions - 𝜓, combination factor for the variable load.

The simplified load combination to determine the design value of actions in fire situation, given by the Swedish national annex EKS 11 (Boverket, 2019) is described as:

𝐸𝑓𝑖,𝑑 = 𝐺𝑘𝑗,𝑠𝑢𝑝+ 𝐺𝑘𝑗,𝑖𝑛𝑓+ 𝐴𝑑+ 𝜓1,1𝑄𝑘,1+ 𝜓2,𝑖𝑄𝑘,𝑖 (5)

where

𝐺𝑘𝑗,𝑠𝑢𝑝 is the unfavorable1 permanent load [kN/m2]

𝐺𝑘𝑗,𝑖𝑛𝑓 is the favorable2 permanent load [kN/m2]

𝐴_𝑑 is the design value of accidental action [kN/m2_]

𝜓_1,1/𝜓_2,𝑖 are the combination factors for the variable loads [-]

(27)

14

𝑄_𝑘,1/𝑄_𝑘,𝑖 are the cooperating variable loads [kN/m2_]

The combination factors mentioned above describes which of the accompanying variable loads, to the known accidental load, is superior (𝜓1,1) vs. inferior (𝜓2,𝑖) and their expected simultaneous occurrence on the structure.

3.3 Timber

When designing timber structures, the properties of the wood at elevated temperatures are important to consider because of its special properties. The main properties of timber exposed to fire are charring and the reduction in strength and stiffness (Östman et.al., 2010).

Relevant to this study is the properties of charring in wood, when exposed to fire, due to its potential effect on the connectors. Also, a heated screws possible effects on the charring of the surrounding wood are relevant. When timber is heated the moisture content decreases, and a rapid pyrolysis process is created at around 300 °C, which then leaves a remaining layer of char (Bartlett et.al., 2019). Inside the charred and pyrolysis zone, the fresh wood is exposed to slightly lower temperatures around 100-200 °C which yields both decomposition of the cellulose and evaporation of the contained water and thereby a slightly lower density (Friquin, 2011). The transport of heat in the timber increases the charred layer, which leads to a reduction in strength of the structure. Figure 5 shows an illustration of timber exposed to fire with fresh-, pyrolysis- and charred zone respectively.

Figure 5 Zones in fire exposed timber. Illustration: Emma Wiklund

3.3.1 Fire design procedure according to EN 1995-1-2 – Design of timber structures

(28)

15

“the position of the 300-degree isotherm”, meaning that the charred layer is expected to exist outside this isotherm, and the pyrolysis zone inside the isotherm at just below 300 °C where the material properties are decreased. See Figure 6 for an illustration of the position of the 300-degree isotherm, the surrounding layers and temperatures.

Figure 6 Position of the 300-degree isotherm in fire exposed timber. Illustration: Emma Wiklund

In order to calculate the mechanical resistance for timber structures exposed to elevated temperatures, the effective cross-section of the uncharred timber shall be determined either by using the reduced cross-section method or the reduced properties method given by EN 1995-1-2 (CEN, 2004b). These two methods depend on the charring rate of the timber and are validated for elements without the potential influence of steel connectors.

The recommended design procedure, of the two above mentioned, is according to EN 1995-1-2 (CEN, 2004b) the reduced cross-section method. This method considers the charred zone and the pyrolysis zone as non-load bearing, meaning that the load-bearing capacity of these zones is assumed to have dropped to zero. This will affect a potential connector placed in these zones.

3.4 Screws as a timber connector

In this section, relevant theory is presented along with existing design and test methods for screws in timber structures.

(29)

16

3.4.1 Geometry

According to the standard EN 14592 “Timber structures – Dowel-type fasteners – Requirements” (CEN, 2008, s. 14), screws shall have a nominal diameter, d, greater than 2.4 mm and less than 24 mm. Also, the inner threaded diameter, d1, shall not extend 90% of the nominal diameter nor be less than 60% of the nominal diameter. The standard claims that the screws shall be threaded over a minimum length, lg, of 4 d. Figure 7 shows an illustrative sketch

over the required geometry of a screw.

Figure 7 - Geometry of screws (EN 14592, 2008)

3.4.2 Mechanical strength and stiffness

According to EN 14592 (CEN, 2008), the mechanical strength for axially loaded screws can be determined either by testing according to methods described in EN 1382 and EN 1383 (see section 4.1.2), or be calculated as described in EN 1995-1-1.

3.4.3 Design procedure according to EN 1995-1-1

The design value of the load carrying capacity, R_d, in a structure at normal temperatures shall, according to EN 1995-1-1 “Design of timber structures – Common rules and rules for building” (CEN, 2004a), be calculated as

Rd= kmod

Rk

γM

(6)

(30)

17

Table 4 Recommended partial factors 𝛾𝑀 for material properties and resistances (CEN, 2004a)

Table 5 Values of 𝑘𝑚𝑜𝑑 (CEN, 2004a)

(31)

18

(1) The withdrawal failure of the threaded part of the screw i.e. the screw is being pulled out of the timber.

(2) The pull-through failure of the screw head i.e. the screw head is pulled though the wood.

(3) The tensile failure of the screw i.e. the screw stretches and then breaks.

The given failure modes mentioned above are illustrated in Figure 8Fel! Hittar inte referenskälla..

Figure 8 Failure modes for axially loaded screws. Illustration: Emma Wiklund

Also, the standard describes minimum spacing between screws and ends according to Table 6 provided by a timber thickness t ≥ 12d. The minimum penetration length of the screw should be 6d. Figure 9 explains the parameters in Table 6 and shows an illustrative sketch over placements and distances of screws, in a connection of timber elements relevant to this research.

Table 6 Minimum distances between screws and ends (CEN, 2004a) Minimum screw

spacing in a plane parallel to the fiber

a1

Minimum screw spacing perpendicular to a plane

parallel to the fiber

a2

Minimum end distance to the threaded screws

center of gravity*

a1,CG

Minimum edge distance to the threaded screws

center of gravity*

a2,CG

7d 5d 10d 4d

(32)

19

Figure 9 Spacing, end and edge distance of screws (CEN, 2004a)

For a connection with a group of screws loaded by a force component parallel to the screw length, the effective number of screws, n_ef, should be taken as

nef= n0.9

(7)

where n is the number of screws acting together in a connection. The effective number of screws can be described as a reduction factor i.e. the strength of a group of connectors is less than the sum of its individual strengths, because of the eventual splitting of the wood (Jorissen, 1998).

Characteristic withdrawal capacity

The withdrawal capacity of the screw is the maximum amount of withdrawal force that can be held in the connection. The withdrawal capacity can be calculated as described in EN 1995-1-1 (CEN, 2004a), as described in this section.

For connection with screws in accordance to EN 14592 (CEN, 2008), (see Figure 7 for definitions) and the geometry of the screws fulfill both criteria in (8) where d1 is the inner threaded diameter of the screw,

12 mm ≥ d ≥ 6 mm

(8)

0. 75 ≥ d1

d ≥ 0. 6

(33)

20

Fax,α,Rk =

neffax,kdℓefkd

1.2cos2_{α + sin}2_α (9)

where α is the angle between the screw axis and the fiber direction (α ≥ 30°), n_ef is the effective number of screws, d is the nominal diameter of the screw and ℓ_ef is the penetration length of the threaded part of the screw. f_ax,k is the characteristic withdrawal strength i.e. the mechanical property indicating its ability to resist the withdrawal force. It can either be tested according to EN 1382 or be calculated as

fax,k= 0.52d−0.5ℓef−0.1ρ0.8 (10)

where ρ is the characteristic density of the wood. The dimension factor of the screw, k_d, is

kd= min {d/8₁

(11)

where d is expressed in millimeters but the factor kd, is dimensionless.

If the requirements presented in (8) are not fulfilled the characteristic withdrawal capacity should be calculated as Fax,α,Rk = neffax,kdℓef 1.2cos2_{α + sin}2_α( ρ ρa ) 0.8 (12)

where ρ_a is the density used at testing of the screw’s withdrawal capacity.

A numerical example for calculating the characteristic withdrawal capacity is presented in APPENDIX C.

Characteristic pull-through resistance

The pull-through resistance of a screw is its capability to withstand actions without being pulled through the wood.

The characteristic pull through resistance of an axially loaded screw, Fax,α,Rk, should according to EN 1995-1-1 (CEN, 2004a), be taken as

(34)

21

where f_head,k is the characteristic pull though parameter determined by EN 1383 and d_h is the diameter of the screw head.

A numerical example for calculating the characteristic pull-through resistance is presented in APPENDIX C.

Characteristic tensile resistance

The tensile resistance of a screw is its capability to withstand actions without being snapped off. The characteristic tensile resistance of an axially loaded screw, F_t,Rk, should, according to EN 1995-1-1 (CEN, 2004a), be taken as

Ft,Rk= nefftens,k

(14)

where f_tens,k is the characteristic tensile capacity of the screw determined by EN 14592 and n_ef is the effective number of screws.

A numerical example for calculating the characteristic tensile resistance is presented in APPENDIX C.

Axial load capacity

The axial load capacity, Fax,Rk, of the screw is the lowest value of the three above mentioned failure modes, see also Figure 8, i.e. withdrawal, pull-through or tensile capacity

𝐹𝑎𝑥,𝑅𝑘 = 𝑚𝑖𝑛 {

𝐹𝑎𝑥,𝛼,𝑅𝑘

𝐹𝑡,𝑅𝑘 (15)

3.4.4 Fire design procedure according to EN 1995-1-2

There are two types of fire design described in EN 1995-1-2 (CEN, 2004b), load dependent resistance and time resistance. Both design procedures are presented in this section.

3.4.4.1 Load resistance

(35)

22

Figure 10 Cross-section and definition of distances (EN 1995-1-2)

The design resistance, R_d,t,fi, of the screw at a given time, t, should be calculated as

Rd,t,fi = η

R20

γM,fi

(16)

where γ_M,fi is the partial safety factor for timber under fire exposure, with recommended value of 1.0. The 20% fractal value of a mechanical resistance, R20, should be calculated as

R20= kfiRk

(17)

where Rk is the characteristic mechanical resistance of a connection at normal temperature without the effect of load duration and moisture (kmod= 1). kfi = 1,05 according to Table 2.1 in EN 1995-1-2 (CEN, 2004b).

For connections corresponding to Figure 10, where

a₂ ≥ a₁ + 40 a₃ ≥ a₁ + 20

(36)

23

Figure 11 Calculation procedure for the conversion factor, 𝜂(EN 1995-1-2)

The calculation procedure presented above includes td,fi, which is the required fire resistance period expressed in minutes.

A numerical example for calculating the fire design resistance is presented in APPENDIX C.

3.4.4.2 Time of resistance

The time of fire resistance for a connection, td,fi, can be calculated as described in EN 1995-1-2 (CEN, 1995-1-2004b): td,fi = − 1 kln ηfiη0kmodγM,fi γMkfi (18)

where 𝑘 is a parameter given in Table 7. Note that this parameter is validated up to 20 min.

Table 7 Value of the parameter k (CEN, 2004b)

𝑘_𝑚𝑜𝑑 is a strength modification factor given with respect to the wooden material, service class3 and the load duration class4_{. The value of the strength modification factor can be reviewed in} EN 1995-1-1, table 3.1 (CEN, 2004a) or in section 3.4.3 Table 5 in this report.

3_{Describes the environmental conditions affecting the structures moisture content, such as temperature and}

relative humidity (CEN, 2004a).

(37)

24

𝛾_𝑀 is the partial factor for material properties as described in Table 4 in section 3.4.3 in this report.

𝜂₀ is the degree of utilization in normal temperatures i.e. a percentage that corresponds to how much of the load bearing capacity is used, may not extend 1.0.

𝜂_𝑓𝑖 is the reduction factor for the design load in the fire situation, and can be determined by choosing the smallest value for the load combinations described in EN 1990 (CEN, 2002b) as:

𝜂𝑓𝑖 = 𝑚𝑖𝑛 { 𝐺𝑘+ 𝜓𝑓𝑖𝑄𝑘,1 𝛾𝐺𝐺𝑘+ 𝛾𝑄,1𝑄𝑘,1 𝐺𝑘+ 𝜓𝑓𝑖𝑄𝑘,1 𝜉𝛾𝐺𝐺𝑘+ 𝛾𝑄,1𝑄𝑘,1 (19)

where 𝐺𝑘 is the characteristic value of permanent action, 𝑄𝑘,1 is the characteristic value of the leading variable action, 𝛾𝐺 is the partial factor for permanent actions, 𝛾𝑄,1 is the partial factor for variable action, 𝜓_𝑓𝑖 is the combination factor of variable actions in the fire situation (𝜓_1,1 or 𝜓_2,1) and ξ is a reduction factor for unfavorable permanent actions.

According to EN 1995-1-2 (CEN, 2004b), the time of fire resistance for unprotected screws is 15 min with the restriction that the screw diameter is equal to or extend 3.5 mm. The time of fire resistances for different connections are presented in Table 8.

Table 8 Fire resistances of unprotected connections with side members of wood (CEN, 2004b)

A greater time of fire resistance, than the above given values, can be achieved by increasing the following dimensions with the factor a_fi:

-

The thickness and width of the side members

(38)

25 where

afi= βnkflux(treq− td,fi)

(20)

The charring rate β_n of the wood is determined according Table 9 below. k_flux is a coefficient that consider an increased heat flux through the fastener and should be of value 1.5, t_req is the required time of fire resistance and t_d,fi is the value taken from Table 8 above.

Table 9 Design charring rates of timber-based panels (CEN, 2004b)

(39)

26

4 RESULTS

In this section, the results obtained by literature review e.g. previous research, current test methods, guides for test planning, is presented. Also, of the results from the interview survey is presented.

4.1 Literature review

In this section, a summary of the available relevant literature regarding screw joints in fire exposed timber structures is presented. Previous research related to the subject is summarized and current test methods are presented along with a resume of how the test planning process can be approached.

The most important result from the literature survey is that there is a great lack of information published about axially loaded screws in wooden structures subjected to fire. Today’s knowledge regarding this subject is obviously limited and therefore the literature review couldn’t be conducted in a desired way.

4.1.1 Previous research

In this subsection, the most relevant research is presented that covers the strength of laterally loaded screw joints, dimension and protection variations of axially loaded screw joints and studies about the fasteners’ influence on charring.

Puong Hock (2006) have studied the fire resistance of screwed connections in laminated veneer lumber (LVL) timber exposed to lateral forces. The top view of the fire test set-up is shown in Figure 12, where the fire is located below the structure in the figure. Thermocouples were placed at three locations inside the specimen to measure the temperature of the fastener inside the wood, the temperature between the wood elements and the surface temperature of the wood. The exact position of the thermocouples are not described nor illustrated in the report. In fire situations, the connections were held under constant tensile forces while being exposed to an ISO 834 fire curve.

(40)

27

Conclusion made by Puong Hock (2006) were that the cap is the weakest point of the screw joint which led to failure, where the cutting of groves ends and creates the weakest point as the load is concentrated to this area. Also, Puong Hock (2006) observed that a failure along the screw was caused by a splitting of the LVL parallel to the screw direction.

Research conducted by Werther et.al. (2014) analyzed the fire resistance of fully threaded screw couple mounted in angle. In this study, the couple of screws were placed in 45° angle and therefore the screws were exposed to both shear and axial force.

Werther et.al. (2014) determined the load capacity in shear in both cold condition and fire situation. The test set-up used in this study is shown in Figure 13 and the screw mounting is presented in Figure 14.

Figure 13 Set-up for fire tests (Werther et.al., 2014)

(41)

28

The results obtained by Werther et.al. (2014) showed that the failure mode of the screw joint was a withdrawal of the screws from both the primary- and secondary beam, which led to displacement and later failure. The results also showed that a long screw performed better in shear force resistance under fire temperatures in comparison with a shorter screw. An efficiency of 0.84 (fire situation versus cold condition) was recorded for the longer screw, whilst the shorter screws protected with wood or gypsum only reached an efficiency of 0.51-0.59. Efficiency results from this study are presented in Table 10 along with type of screw and their placement in reference to timber element edges. Werther et.al. (2014) claim that the high efficiency of the long screw is generated by the temperature distribution along the screw, to the timber, which yields a large pull-out resistance over a long part of the screw hence the heat is distributed over a greater timber area. However, determination of axial force capacity in the screws alone was not conducted.

Table 10 Efficiency in fire situation for different types of screws and their placement (Werther et.al., 2014)

Test S1 S2 S3 S4

Screw dimensions [mm] 6 x 180 mm 6 x 180 mm 6 x 180 mm 12 x 350 mm

Coverage of screw head 25 mm solid

wood board 25 mm solid wood board 15 mm gypsum board 25 mm solid

wood board None

Time of fire exposure [min] 30 60 30 30

Distance from side edge on top

[mm]

36 64 32 73

(42)

29

Several studies have been made where charring around fasteners have been the main topic. Norén (1996) made a conclusion that theoretical methods can be applied for calculation of load bearing capacity of nails exposed to fire and shows that there is a considerable reduction in embedding strength5_{of the wood in fire situations. However, Puong Hock (2006) points out} that the calculation procedure provided by Norén (1996) only is applicable on single nails, therefore the test results with several nails in joints may differ from the calculated results because of the increased charring rate caused by heat from the nails. Also, the mechanical properties of the wood at elevated temperatures exclude the embedding strength and only covers properties in compression and tension strength.

Carling (1989) summarized results conducted by Markku & Kallioniemi (1979, 1983) where they studied the charring around nails in timber structures exposed to fire. The results showed that the presence of nails affected the charring depth of the timber, when comparing the charred zone to an area where there were no nails. The results also showed that the penetration depth of the charred zone depended on the dimensions of the nails. Large nails caused a charring in the immediate vicinity of the nail inside the timber whereas the small nails generated a greater charring depth on the fire exposed side of the timber. Figure 15 shows a schematic illustration of the charring depth in vicinity of different sized nail, as described by Markku & Kallioniemi (1979, 1983).

Figure 15 Penetration of the charred zone in vicinity of large and small nails (Markku & Kallioniemi, 1979, 1983)

In the research conducted by Puong Hock (2006) regarding charring rate in timber connections, the connectors had an influence of the charring rate of the timber. After the fire tests was performed, the remaining layer of char was removed, and the thickness of the non-affected

5_{The average compressive stress in timber under actions of a linear fastener loaded perpendicular to its axis,}

(43)

30

wood was measured to determine the thickness of wood that had charred during the tests. The average char rate without connectors was 0.67 mm/min and 0.71 mm/min for screw connections. However, Puong Hock (2006) does not describe the position of where the charred layer was measured i.e. in direct vicinity of the screw or between screws. In this research, joints with several screws were tested and no single screws. Also, a slight difference in char rate could be distinguished related to the total enclosure area of steel connectors contributing to heat transfer inside the wood. A greater total enclosure area of the connectors generated a slightly higher char rate. The influence of different parameters of the connector, on temperature distribution in the fastener itself and the charring depth of the timber, was also determined by Werther et.al. (2014). The results generated from Werther et.al. (2014) strengthens the arguments Puong Hock (2006) made, where the enclosure area of the fastener plays a significant role in the charring behavior of the wood. Table 11 presents the results proved by Werther et.al. (2014) regarding the influence of different screw parameters on the temperature within the screw along with the fasteners effect on the charring depth of the wood.

Table 11 Influence of the screw parameters on the connection, derived from fire tests by Werther et.al. (2014)

Parameter Low temperature at the screw tip if:

Low charring depth of the wood if:

Influence of the parameter

Length Long Short Significant

Material Stainless steel Stainless steel Moderate

Diameter Small Small Moderate

Screw head type Not detectable Not detectable Negligible

Norén (1996) points out the importance in placement of the connector with respect to the impact of charring. Charring depth increases more rapidly close to edges of a structure, and therefore connectors close to the edge are more exposed to charring. The author also claims that this phenomenon, in turn, affect potential connectors further away from the edges with respect to the increased load applied on these connectors.

4.1.2 Current test methods

(44)

31

temperature, because of the absence of existing test methods regarding the load bearing capacity in fire situations.

Common for the tests presented in this section is that the timber test pieces used shall be manufactured at an equilibrium moisture content corresponding to 20 ± 2 °C and 65 ± 5 % relative humidity. To verify a constant mass of the timber piece, the weight of the piece is measured at a 6h interval. With a maximum difference in 0,1 % in mass, the timber piece is considered conditioned. Also, timber density and moisture content shall be determined as described in the standards ISO 13061-1 (ISO, 2014a.) and ISO 13061-2 (ISO, 2014b.) respectively. The timber products used in the tests shall be representative for each class or range of product to which they belong.

4.1.2.1 Withdrawal capacity of timber connections – EN 1382

Test methods for the withdrawal capacity of timber connection i.e. screws are described by CEN (2016a) in EN 1382. A summary of the test procedure is presented below.

The mounting of the screws in the timber piece shall follow normal practice and preparation. Figure 16 shows the test specimens and applied load in both parallel and perpendicular fiber direction to determine the withdrawal capacity of the joint. The dimensions presented in Figure 16 is defined as l, the length of the screw and d, the outer thread diameter of the screw. The screws shall be driven into a depth of 8 d to 20 d for both load cases.

Figure 16 Withdrawal timber test specimen - load parallel and perpendicular to fiber respectively CEN (2016a)

(45)

32

The withdrawal parameter, f_ax, also known as the characteristic withdrawal strength used in the theoretical calculation presented in section 3.4.3, is then determined by

fax=

Fmax

d × ℓd (21)

Where ℓd is the premeasured penetration length of the screw in initial position before loading.

4.1.2.2 Pull through resistance of timber fasteners – EN 1383

Test methods for the pull through timber fasteners i.e. screws are described by CEN (2016b) in EN 1383. A summary of the test procedure is presented below.

The moisture content and the density of the timber specimen shall, beyond the moisture content and density determination presented in section 4.1.2, be determined according to EN 322 – “Wood-based panels – Determination of moisture content” (CEN, 1993c) and EN 323 – “Wood-based panels – Determination of density” (CEN, 1993b) .

The dimensions of the timber piece, depending on type, are presented in Table 12.

Table 12 Minimum dimensions of the timber pieces Test piece material Test piece size (minimum) Solid timber 4t x 4t t ≤ 7d

Wood based products 4t x 4t t = panel thickness as produced

The mounting of the screw shall be perpendicular to the surface of the timber piece where the insertion of the specimen shall follow normal practice and preparation.

(46)

33

Figure 17 Test apparatus and test specimen used for determination of the pull through capacity (CEN, 2016b)

The test specimen is placed on a metal plate with a circular hole with diameter, D ≥ (2t + d_h). The screw is clamped into the load device that pulls the screw.

The apparatus assure that the applied force is along the axis of the screw. The applied load shall pull the screw through the timber piece with a continuous movement and the time to reach F_max should be 300 ± 120 s.

Where F_max is the pull through capacity and the pull through parameter, f_head, used in the theoretical calculation presented in section 3.4.3, is determined by

fhead=

Fmax

d_h2

(22)

where d_h is the diameter of the screw head.

To determine the head pull off strength, the same test procedure is used with the amendment that a steel plate is mounted on top of the timber specimen, to isolate the screw head in tension.

4.1.3 Planning of tests – EN 1990

Before a test is conducted, a test plan should be created. By following the recommended content of the test plan described in EN 1990 (CEN, 2002b), a proper test to establish the ultimate resistance of structural members can be made. In accordance to the standard the following steps should be considered when creating a test plan:

(47)

34 - Prediction of test results

- Specification of test specimens and sampling - Loading specifications

- Testing arrangement - Measurements

- Evaluation and reporting of the tests.

EN 1990, section D4 (CEN, 2002b).

In the following section, a description of the content of the recommended test plan is presented and adapted to the aim of this thesis, regarding what experimental method would be suitable for carrying out further tests of axially loaded screw joints in timber at fire exposure.

Objectives and scope

Required properties, influence of certain variation in design parameters and the range of validity of the test should be clearly expressed. Also, the test limitations and required corrections applied to measurements to determine the values corresponding to a full-sized structure, should be presented and specified.

Prediction of test results

All properties and circumstances that can affect the prediction of the test results, e.g. geometrical and material properties, mounting and execution parameters etc., should be considered. Also, the expected failure modes should be described related to the corresponding variable where consideration that a structural member can possess several different failure modes. In case of unknown/unsure prediction of critical failure modes, pilot tests should be conducted and followed by a custom test plan related to the observed failure modes.

Specification of test specimen and sampling

I order to obtain a realistic structure in the test, the properties of the test specimens should be specified. Also, a statically representative of the test procedure should be obtained with the attention paid to differences between the test specimens and the potential factors that could influence the test results. The factors to be considered while specifying the properties of the test specimen include:

(48)

35 - Restrictions of the specimen

Loading specifications

The loading scenario that will be used in the test should be representative to the later use of the test specimen, in both normal and severe conditions. Load conditions should be specified and include loading points, loading history, deformational load/force control, etc. Also, potential interaction between the load apparatus and the structural response of the specimen should be considered. Environmental conditions, e.g. temperatures, restraints and relative humidity should also be specified.

Testing arrangement

Relevant test equipment should be used, where measures to obtain enough strength and stiffness of the loading and supporting equipment, controlled deflection etc., should be payed attention. Measurements

All relevant properties to be measured for each test specimen should be listed including the measurement locations, the accuracy of the measurements and suitable measuring devices. Evaluation and reporting the test

(49)

36

4.2 Interview survey

In this section, the results from the interview survey is presented under subsections created from the interview question.

In total, seven interviews were conducted. One interview per respondent whom are described in Table 2 in section 2.2.1.

4.2.1 General identification and processing of new needs

The respondents clearly had two common ways of identifying new needs, either they receive identified needs from the industry, or the identification is made by themselves.

Stenmark (SIS) stated that the standardization process in Sweden is marked oriented and therefore SIS receives proposals for development projects directly from the industry. The researchers Jockwer (Chalmers), Just (TalTech/RISE) and Hasburgh (FPL) also claimed that in the university environment, needs from the industry is one of the project incomes. However, along with the distribution from the industry, the universities are interested in new needs that trigger further research and the universities are therefore a source of identification themselves. Malm (TK Botnia), pointed out that as a consultant, development by cooperation with the suppliers is common. Mainly because the consultants are the end usage designer of the supplier’s products in construction and can identify problems or development project related to their products.

Later, when the need is identified, the processing of the needs is conducted. At SIS, there is an ongoing iterative process working with development and maintenance of the standards, described by Stenmark. Different working groups within SIS, consisting of representants from specific disciplines, work with development of new standards based on the desires received from the industry. The working groups also keep existing standards available, applicable, and up to date through constant development, modification and in some cases cancellation of existing standards; an equally important phase in the standardization process claims Stenmark.

(50)

37

4.2.2 Current knowledge about axially loaded screws in timber structures exposed to fire

The respondents’ answers were unanimous in terms of current knowledge regarding axially loaded screws in timber structures exposed to fire: It is very limited.

To his knowledge, Jockwer answers, there is little information available about axially loaded fasteners generally, in fire situation in contrast to laterally loaded fasteners. Through his work with the renewed edition of Eurocode 5, the experience is that there is a knowledge gap in the connection chapter regarding axially loaded fasteners exposed to fire, and points out that there is no current development in the new edition regarding this topic. Confirmation of this statement was also made by Just, who claims that the absence in expressed needs from the industry may have caused little activity in this area, since the industry is one of the main resources in needs identification.

In the design of axially loaded screw joints in structures, Malm states that today, the standard design execution only involves the demand of protection of the joints, not the design of the connection itself. This because of the limited information available regarding the unprotected axially loaded screw joints, and the more explored area of fire protection of structures in general. Hasburgh informs that in FPL there is no existing research unit who work with connections within the fire group, also there is no ongoing fire related research within the structures group. Therefore, there is little knowledge regarding fire exposed connections within FPL, however Hasburgh points out that the interest in fire exposed connections is existing but the current available equipment is limited hence no further research can be made today.

Due to the role of project manager at SIS, Stenmark informs that she is not familiar with the technical parts of the ongoing work and therefore respectfully refrains from answering the technical questions.

Although the absence of information regarding the axially loaded fasteners in fire exposed structures, the respondents where all positive about a future development in the subject.

4.2.3 Existing demand for new knowledge regarding axially loaded screws in timber structures exposed to fire

The respondents were unified in their answers in terms of that there is an existing demand for development of the information base regarding knowledge in axially loaded screw joints in timber structures exposed to fire.

(51)

38

conditions is studied first and then studied later in the fire situation, and this area has clearly not yet arrived at the latter part regarding fire performance due to the complicated process of verifying a specific connection with the lack of previous reference material, Jockwer added. Malm expressed that more information is needed because of the cost advantages, compared to more complex fastener designs, that axial screws can possess in terms of both material and mounting costs, and therefore a more explicit design method is desirable. Today, as a constructor, Malm design the capacity of an unprotected axially loaded connection by considering factors such as charring of the wood, the countersink of the screw and if a protective wooden dowel is used or not. This estimation follows by a traditional protection by either gypsum or wooden boards to achieve the desired protection of the connections and thus the load bearing capacity for a certain period of time. Malm predict that if an accepted design method exists for unprotected joints, effectiveness in the design- and mounting process can be achieved without the risk that follows by making estimations in the design.

(52)

39

Figure 18 The presence of protective wood affecting the charring of laterally- and axially loaded screws respectively. Illustration: Emma Wiklund

Hasburgh also expressed the need for studies in the subject and a need for additional knowledge regarding joints within the fire group at FPL and in the industry overall.

4.2.4 Proposed testing method

To meet the identified and existing needs in knowledge regarding axially loaded screw joints in fire exposed wooden structures, the respondents freely discussed potential future testing methods which differed from each other.

Jockwer proposed an approach where the screw connection is first studied in normal temperatures, and thereby verification of the shear and pull out behavior of the screw is conducted. Thereafter followed up by fire tests of the connection where time to failure is measured. However, Jockwer pointed out that ideally the test design is conducted to get a parametric design of the structure where the temperature behavior of the screw in the joint is verified i.e. what the temperature distribution in the screw is and determination of the critical temperature which leads to failure. This temperature dependency may later be applicable in different loading situation to predict the capacity of the joint.

Strength of axially loaded screw joints in wooden structures exposed to fire

Strength of axially loaded screw joints in

wooden structures exposed to fire

An overview of existing knowledge and proposal for test method

Emma Wiklund

Strength of axially loaded screw joints in

wooden structures exposed to fire

An overview of existing knowledge and proposal for test method

PREFACE

ABSTRACT

SAMMANFATTNING

TABLE OF CONTENTS

NOMENCLATURE

ABBREVIATIONS

1 INTRODUCTION

1.1 Background

1.2 Aim and objectives

1.3 Research questions

1.4 Delimitations

1.5 Notes to the reader

2 RESEARCH STRATEGY

2.1 Literature survey

2.2 Interview survey

2.3 Research strategy criticism

3 THEORY

3.1 EN Eurocodes

3.2 Structural fire design according to EN 1991-1-2 – Actions on

structures

3.3 Timber

3.4 Screws as a timber connector

-

4 RESULTS

4.1 Literature review

4.2 Interview survey