See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/357604875

Fire performance of hybrid mass timber beam-end connections with

perpendicular-to-wood grain reinforcement

Article in Journal of Structural Fire Engineering · January 2022

DOI: 10.1108/JSFE-06-2021-0036

CITATIONS READS

2 347

3 authors, including:

Sam Salem George Hadjisophocleous

Lakehead University Thunder Bay Campus Carleton University
45 PUBLICATIONS 73 CITATIONS 222 PUBLICATIONS 1,832 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fire Resistance Testing of Glulam Frame Connections View project

Fire Behaviour of Concrete Structural Elements View project

All content following this page was uploaded by Sam Salem on 15 January 2022.

The user has requested enhancement of the downloaded file.

 The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/2040-2317.htm

Fire performance of hybrid mass Fire resistance

of mass timber
timber beam-end connections with connections

perpendicular-to-wood
grain reinforcement
Oluwamuyiwa Okunrounmu Received 2 June 2021
Revised 3 December 2021
Department of Civil and Environmental Engineering, Carleton University, Accepted 10 December 2021
Ottawa, Canada
Osama (Sam) Salem
Department of Civil Engineering, Faculty of Engineering, Lakehead University,
Thunder Bay, Canada, and
George Hadjisophocleous
Department of Civil and Environmental Engineering, Carleton University,
Ottawa, Canada

Abstract
Purpose – The fire resistance of timber structures is heavily dependent on the fire behaviour of the
connections between its structural elements. The experimental study presented in this paper aimed to
investigate the fire performance of glued-laminated timber beam connections reinforced perpendicular-to-wood
grain with self-tapping screws (STS).
Design/methodology/approach – Two full-size fire experiments were conducted on glulam beam-end
connections loaded in flexure bending. Two connection configurations, each utilizing four steel bolts arranged
in two different patterns, were reinforced perpendicular to wood grain using STS. The bolt heads and nuts and
the steel plate top and bottom edges were fire protected using wood plugs and strips, respectively. Each
connection configuration was loaded to 100% of the ultimate design load of the weakest unreinforced
configuration. The test assemblies were exposed to elevated temperatures that followed the CAN/ULC-S101
standard fire time–temperature curve.
Findings – The experimental results show that the influence of the STS was significant as it prevented the
occurrence of wood splitting and row shear-out and as a result, increased the fire resistance time of the
connections. The time to failure of both connection configurations exceeded the minimum fire resistance rating
specified as 45 min for combustible construction in applicable building codes.
Originality/value – The experimental data show the effectiveness of a simple fire protection system (i.e. wood
plugs and strips) along with the utilization of STS on the rotational behaviour, charring rate, fire resistance time
and failure mode of the proposed hybrid mass timber beam-end connection configurations.
Keywords Fire resistance, Mass timber, Hybrid wood-steel-wood connections, Self-tapping screws
Paper type Research paper

1. Introduction
Wood is a material used in construction for centuries, ever since the history of mankind
(Smith and Snow, 2008). With the introduction of steel and concrete as building materials, the
use of wood had declined; however, over the past few decades, the usage of wood has
significantly increased owing to its sustainability, renewability and availability of

The research study presented in this paper has been supported in part by the NSERC Discovery grants
Journal of Structural Fire
awarded to the second and third authors. The authors of this paper would like to thank lab technologist Engineering
Cory Hubbard for his great assistance in the Civil Engineering’s Structures Laboratory at Lakehead © Emerald Publishing Limited
2040-2317
University. DOI 10.1108/JSFE-06-2021-0036
JSFE innovative, high quality engineered products. Wood construction can be categorized as either
mass timber or light frame construction. In mass timber construction, the structural members
are known to have a very good and predictable passive fire resistance. This is due to the
charring rate (i.e. the rate of advance of char into the wood) which after an initially high rate it
decreases to a slower steady rate over the remainder of the fire exposure duration as
compared to members used in light frame construction that can burn out much faster
(Buchanan and Abu, 2017).
Fire occurrence in buildings is unpredictable and its effects on structural elements can be
very destructive. In fire conditions, connections between main structural elements, such as
beams and columns, are often the weakest component of any timber structure. The most
common type of connections used in mass timber construction are the wood-steel-wood
(WSW) connections owing to their ease of assembling and aesthetic appearance as well as
reasonable fire resistance. Their strength and fire resistance, however, can be impaired due to
the reduction in the cross-section of the connected member, owing to the existence of a slotted
cut in the middle of the cross-section needed to accommodate the steel connecting plate.
Furthermore, the presence of metal fasteners in contact with the concealed steel plate allows
the transmission of heat into the core of the timber section increasing the charring rate and
thus, decreasing their fire resistance (Peng et al., 2010).
A considerable technical challenge of timber construction is the analysis and safe design
of timber connections in fire due to the many parameters involved, such as the different
thermo-physical and thermo-mechanical properties of the materials (i.e. steel and wood),
variety of wood species, applied load ratios and fire protections, the geometry of the
connection, fastener types and their layout (Buchanan, 2000). Thus, the behaviour of a timber
connection subjected to fire is complex and not easily predictable. Over the past decade, a
number of experimental studies were conducted to investigate the fire performance of timber
connections (Owusu, 2019; Palma et al., 2013; Peng et al., 2012; Petrycki and Salem, 2019).
However, to fully understand and assess the fire performance of such connections, further
experimental studies need to be conducted.
In the experimental study conducted by Peng et al. (2012) on the fire behaviour of bolted
wood connections subjected to direct tension, 22 specimens comprising 16 wood-steel-wood
(WSW) and 6 steel-wood-steel (SWS) connections were exposed to standard fire. Factors that
affected the fire performance of the experimentally tested connections, such as wood member
thickness, fastener diameter, number of fasteners, end distance, load ratio and fire protection
application were all examined. The results show that the application of a single 15.9-mm thick
Type-X gypsum board increased the fire resistance time of the connection by 30 min, while a
double 12.7-mm thick plywood board increased the fire resistance time by 15 min only. The
fire resistance times of both WSW and SWS connections were less than 51 and 25 min,
respectively. Generally, the governing failure mode of the tested connections was bolt hole
elongation followed by splitting of the wood members.
In a more recent study conducted by Owusu (2019), the performance of unprotected and
minimally protected WSW beam-end connections exposed to the elevated temperatures of the
standard fire was experimentally investigated. Four full-size tests were conducted on 1,600-
mm long glulam black spruce-pine beams with cross-sectional dimensions of
184 mm 3 362 mm (two at normal temperature, and two in standard fire conditions). The
two test specimens exposed to the standard fire were also subjected to a monotonic load equal
to 100% of the ultimate design load of the weakest connection configuration throughout the
fire test duration. The steel connecting components (i.e. steel bolt heads and nuts, and steel
plate top and bottom edges) in one of the two specimens tested in fire were protected with
same-species wood plugs and strips, respectively. Tests results showed that the two
predominant failure modes of the experimentally tested WSW bolted connections at both
normal and elevated temperatures were perpendicular-to-wood grain tensile splitting and
row shear-out (both identified as brittle failures). The aformentioned failure modes played a Fire resistance
major role in reducing the moment-resisting capacity of the WSW bolted connections. At of mass timber
normal temperature, the two unprotected connection configurations tested in the said study
attained a maximum bending moment of approximately 7% more than that of similar but
connections
protected configuration with beam cross-section with removed material to accommodate the
protecting wood plugs and strips. At elevated temperatures, the time to failure of the
minimally fire protected connection configuration was 23 min more than that of the similar
but unprotected configuration. It was also observed that splitting of wood occurred along the
glue-line plane aligned with the top row of bolts in the glulam beam section and thus, this
failure further limited the time to failure of the unprotected connection configuration
(Owusu, 2019).
To strengthen timber connections against brittle failures, a few experimental studies
(Zhang et al., 2019; Viau and Doudak, 2021) were conducted by employing different forms of
reinforcements at normal temperature. The study conducted by Zhang et al. (2019) showed
that self-tapping screws (STS) can be used to enhance the bending strength, control crack
propagation and significantly prevent splitting failure in dowel-type moment-resisting
connections. The results of Viau and Doudak (2021) who investigated the behaviour of bolted
connections under blast loads show that reinforcing the connections with STS prevented
brittle splitting failure at the connection and changed the failure mode to a combination of
bolt yielding and wood crushing. Furthermore, the authors of the later study reported that
employing STS promote more ductile behaviour when compared to that of unreinforced
connections. In a very recent study conducted by Letourneau-Gagnon et al. (2021) on the fire
performance of STS in mass timber sections, the thermo-mechanical behaviour and fire
resistance of STS installed in mass timber sections were experimentally and numerically
studied. Their results show that the smaller heated area of the exposed surface of the screws
limited the thermal conduction along the screw shank and maintained a relatively low
thermal profile of the screws for most of the two-hour standard fire exposure.
In a study by Lam et al. (2010), experimental tests were conducted to evaluate the
performance of moment-resisting bolted timber connections with STS acting as
perpendicular-to-wood grain reinforcement. In that study, it was observed that employing
STS as reinforcement greatly increased the ductility and overall performance of the
connections. However, at elevated temperatures, Palma et al. (2013) reported that utilizing
STS as perpendicular-to-wood grain reinforcement in connections can significantly increase
the charring rate of the wood if prematurely exposed to fire as a result of being too long and
hence, decreasing the fire resistance of the connections. In contrast, a recent study conducted
by Petrycki and Salem (2019) on the fire performance of WSW bolted connections with and
without perpendicular-to-wood grain reinforcements revealed that employing STS as
perpendicular-to-wood grain reinforcement increased the time to failure of the connections in
standard fire conditions. This is mainly attributed to the shorter STS utilized in their
specimens, which allowed thicker wood cover to exist at the bottom side of the glulam beams
compared to the longer STS used in the connections studied by Palma et al. (2013).
Accordingly, further studies need to be conducted to better understand the influence of
STS on the fire performance of timber connections as relatively limited data are available in
the literature. In addition, these studies will provide data and design methodologies for
inclusion in codes and standards. For example, despite the benefits of employing STS in
timber connections, no design provisions are provided in the Canadian Engineering design in
wood standard CSA O86-19 (2019) to determine the moment-resisting capacity of STS-
reinforced connections.
The study presented in this paper shows the results of two full-size fire experiments on the
behaviour of WSW bolted connections reinforced with STS and minimally fire protected
using wood plugs and strips to cover the steel bolt heads and nuts and steel plate top and
JSFE bottom edges, respectively. Each connection employed four steel bolts that were arranged in
two different bolt patterns. The two fire experiments discussed in this paper are part of a
larger experimental campaign that aims to investigate the influence of STS as perpendicular-
to-wood grain reinforcement on the fire behaviour of WSW connections with different
number of bolts in varying bolt patterns. The results of the two experiments presented in this
study were compared to those of identical connection configurations but without STS
experimentally tested in a previous related study (Owusu, 2019). The new test results
presented in this paper provide a better understanding of the rotational behaviour and fire
resistance of such type of WSW connections when reinforced with STS and provided with
minimal fire protection (i.e. wood plugs and strips).

2. Experimental program
Two full-size fire experiments were conducted on glulam beam-end WSW connections loaded
perpendicular to wood grain to develop bending moment at the connection. Two connection
configurations with different bolt patterns were considered. Both test specimens were
reinforced perpendicular-to-wood grain with STS and exposed to elevated temperatures that
follow the CAN/ULC-S101 standard fire time–temperature curve (CAN/ULC-S101, 2019). The
specimens were loaded to 100% of the ultimate design load of the weakest unreinforced
connection configuration, based on a previous related study conducted by Owusu (2019) in
the same fire testing facility. This load level was deliberately chosen to assess the influence of
STS on the fire resistance time and rotational behaviour of the tested WSW connections in
comparison to the results of identical but unreinforced connections reported by Owusu (2019).
The applied load was kept constant throughout the duration of the fire test until the test was
terminated when the specimen could not sustain the applied load and/or a significant load
drop with no possible recovery was maintained. The test set-up, loading procedure, and
glulam beam section and its steel connecting components of the fire experiments of this study
were identical to those employed in the study carried out by Owusu (2019). The geometry of
the tested connections was designed to satisfy requirements in the Canadian Engineering
design in wood standard CSA O86-19 (CAN/CSA O86, 2019), and the fire tests were conducted
in accordance with CAN/ULC-S101 fire endurance testing standard (CAN/ULC S101, 2019).

2.1 Materials
2.1.1 Glulam beams. The glulam beam sections used in this study were prepared from
Canadian black spruce–pine (S-P) with a stress grade of 24f-EX and an architectural grade.
The cross-sectional dimensions of the 1,600-mm long glulam beams used in this study were
184 mm (width) 3 362 mm (depth). The beam sections were conditioned indoor at 208C and
approximately 60% relative humidity (RH) prior to testing. Thus, the beam sections
experienced minimal shrinkage due to the reduction of their moisture content that was
recorded at an average value of approximately 10% right prior testing.
The test beams were supported at one end by a sturdy steel column which was fire
protected using several layers of ceramic fibre blankets to ensure its reusage in several
subsequent fire tests. Table 1 lists the mechanical properties of the glulam sections as
reported in the Canadian Construction Materials Centre (CCMC) Evaluation report 13216-R
(CCMC, 2018).
2.1.2 Steel T-stub connectors and steel bolts. In this study, the steel T-stub connectors used
were manufactured from a 12.7-mm (1/2 in) thick steel plate of grade 300 W. The connector
had a height of 302 mm which was 60 mm less than the depth of the beam cross-section to
allow the insertion of two 30-mm deep wood strips that were employed as protection to the
steel plate top and bottom edges. Figures 1 and 2 show the dimensions of the two steel T-stub
connectors utilized in the connection configurations with two different bolt patterns (i.e. P1 Fire resistance
and P2) examined. The steel bolts used were 19.1-mm (3/4 in) diameter A325 M high-strength of mass timber
structural steel bolts and thus, all bolt holes on the T-stub connectors were sized 1–2 mm (1/16
in) larger than the bolt diameter as per CAN/CSA O86-19 (CAN/CSA O86, 2019). Therefore, all
connections
the steel connectors’ holes were 20.6 mm (13/16 in) in diameter. Also, a weld size of 6 mm was
used for all fillet welds made for the fabrication of the steel T-stub connectors.
2.1.3 STS reinforcements. SWG ASSY VG plus CSK STS were employed to reinforce the
connection specimens perpendicular to wood grain. The fully threaded STS were made of
carbon steel and had a bending yield strength of 1,015 MPa and unfactored shear strength of
641 MPa. The screws were 300 mm long and had an 8-mm outer thread diameter. Table 2

Property Strength (MPa)

Compression parallel-to-grain 33.0

Compression perpendicular-to-grain 7.5
Tension parallel-to-grain 20.4
Longitudinal shear 2.2 Table 1.
Flexural bending 30.7 Mechanical properties
Modulus of elasticity 13,100 of the glulam beam
Source(s): Adopted from CCMC (2018) sections

Figure 1.
Steel T-stub connector
for the connection
configuration with the
first bolt pattern P1 (all
dimensions in mm)
JSFE

Figure 2.
Steel T-stub connector
for the connection
configuration with the
second bolt pattern P2
(all dimensions in mm)

L Lthread Ltip dhead dmin da th dp

d mm Bit

Table 2. 8 300 283 8 14.8 5 9 4.6 19 AW 40

Specification for STS Source(s): Adopted from CCMC (2018)

provides the full dimensions of the STS as reproduced from the Canadian Construction
Materials Centre (CCMC) Evaluation report 13677-R (CCMC, 2014). Figure 3 shows the
geometry of the STS used in this study.

2.2 Test assembly details and fabrication process

For each test assembly, a T-stub steel connector was used to connect the glulam beam to a
supporting steel column. A slotted cut to accommodate the 12.7-mm thick web of the T-stub
connector was prepared at one end of the beam with approximately 15 mm width to allow for
a fabrication tolerance of 1–2 mm as required in CAN/CSA O86-19 design standard (CAN/
CSA O86, 2019). Circular holes were drilled into the glulam beam face with the use of a 30-mm
diameter spade bit to embed the bolt heads and nuts, as well as to accommodate the
wood plugs used as fire protection to the heads and nuts of the bolts. The inner holes for the
19.1-mm diameter bolts were drilled using a 20-mm diameter drill bit to allow for
approximately 1 mm fabrication tolerance as specified in CAN/CSA O86-19 design standard
 Fire resistance
of mass timber
connections

Figure 3.
Dimension and details
of the self-tapping
screw
(8.0 mm 3 300 mm)

(CAN/CSA O86, 2019). In each connection configuration, the steel connector was sandwiched
between two wood side sections, each with a thickness of approximately 84.5 mm, and
fastened together using four steel bolts. The spacings of the fasteners (i.e. end distance, edge
distance and spacing between bolts) used in the glulam beam-end connections were designed
to meet the requirements as per CAN/CSA O86-19 design standard (CAN/CSA O86, 2019).
Thereafter, the glulam beam section was connected to the supporting steel column via the
T-stub connector using another four 19.1-mm diameter short steel bolts. The heads and nuts
of the four steel bolts inserted into the glulam beam section as well as the top and bottom
edges of the knife plate of the T-stub connector were fire protected using glued-in wood plugs
and strips, respectively, that were made of the same wood material as that of the glulam beam.
Figure 4a and b shows the dimensions of the tested WSW connections with the two different
bolt patterns (i.e. P1 and P2), respectively, along with the layout of the installed STS as
perpendicular-to-wood grain reinforcements.
The two rows of bolts in the connection configuration with the first bolt pattern (P1) were
symmetrically positioned near the top and bottom sides of the beam cross-section. While in
the connection configuration with the second bolt pattern (P2), the bottom row of bolts was
shifted upward to the mid-height of the beam section to further contribute to the moment-
resisting capacity of the connection.

2.3 Installation of reinforcements

Before each connection was assembled for testing, fully threaded STS were installed
perpendicular-to-wood grain and parallel to the embedded steel knife plate through most of
the depth of the beam section to strengthen the connection and prevent splitting failure. For
the installation of the fully threaded long STS, it is a common practice to pre-drill small-
diameter pilot holes (approximately 30–40% of the screw diameter) for the STS to avoid
either wood splitting or damage of the screw head which could be caused due to large friction
forces that might be generated while driving in the screws. Accordingly, pilot holes were pre-
drilled with the aid of a 3-mm diameter drill bit to a 280 mm depth into the beam section to aid
the precision of the screw positioning and to minimize the development of friction forces. Six
(6) reinforcing screws were installed from the top to near bottom of each beam section at one
end where the connection is located. The STS were concealed at the centre of the wood side
JSFE

Figure 4.
The glulam beam-end
connection
configurations with the
STS layout

sections and arranged in two rows, each of three screws, on each side of the embedded steel
knife plate. In the beam longitudinal direction, one pair of screws was centred between the
beam end and the first column of bolts; the second pair of screws was centred between the
first and second columns of bolts; and the last pair was placed at 50 mm beyond the second
column of bolts into the beam length. As previously discussed, in a study conducted by Palma
et al. (2013), it was found that significant increase of the charring rate in the wood connections
can be attributed to the use of reinforcing screws that run almost the full depth of the wood
beam (with only 30 mm wood cover on the bottom side of the beam). Therefore, the screws
used in the study presented in this paper were 300 mm in length to allow for slightly more
than 60 mm wood cover on the bottom side of the beam. The STS layout as installed within
the connection area for the configuration with the first bolt pattern (P1) is shown in Figure 5.
2.4 Instrumentation Fire resistance
Temperatures were measured within the connection area from both the front and back faces of mass timber
of the beam. These thermal measurements were taken by twelve (12) metal sheathed, mineral
insulated Type-K thermocouples (TC) cables with sheath outer diameter of 1.5 mm that were
connections
placed at different locations and depths inside the glulam beam section. The holes prepared in
the glulam beam section to insert the thermocouple cables were created using a portable drill
press capable of accurately drilling to the required depth. The hole diameter was the same as
that of the thermocouple cables so that a snug fit of the cables was achieved to avoid any heat
that may penetrate inside the holes surrounding the thermocouple cables, which might
influence the thermal measurements. The thermocouples were inserted in the pre-drilled
holes to depths of 20, 30, 40, 60 and 85.7 mm from the beam fire-exposed surfaces (i.e. front or
back surfaces). Eight (8) thermocouples were placed at the front face of the beam and four (4)
thermocouples from the back face of the beam, as shown in Figure 6. The location and the
depth at which each thermocouple was inserted inside the glulam section are also shown in
Figure 6. At the beam front face, thermocouples TC1 and TC8 were installed to measure the
temperatures of the bolt heads under the 30-mm thick wood plugs, while another two
thermocouples (TC2 and TC6) were installed at 85.7 mm depth to measure the temperature of
the embedded steel knife plate. At the beam back face, thermocouples TC11 and TC12
measured the temperatures of the bolt nuts under the 20-mm thick wood plugs.
Thermocouples TC4 and TC5 were used to measure the temperatures at 60 mm depth

Figure 5.
STS layout for the
connection with the
first bolt pattern (P1)
installed from the top
side of the beam

Figure 6.
Thermocouples layout
for the connection
configuration with the
first bolt pattern (P1)
JSFE inside the beam section; thermocouples TC3 and TC7 were installed at 40 mm depth; and
thermocouples TC9 and TC10 were installed at only 20 mm depth to measure the internal
temperatures of wood so that the calculations for the actual charring rates can be executed.
Figure 6a and b shows the thermocouples layout for both the front and back faces of the
beam, respectively. Table 3 shows the test matrix details.

2.5 Test set-up and procedure

The fire tests were conducted in the large-size horizontal natural gas-fuelled furnace located at
Lakehead University Fire Testing and Research Laboratory (LUFTRL), Thunder Bay, Ontario,
Canada. Two high-capacity gas burners are installed in opposite corners of the furnace to allow
uniform circulation of the hot gases within the furnace volume. Gas input into the furnace is
regulated via automatic gas actuator that is controlled by the computer system and logic
circuits that are built in an advanced control panel connected to the furnace. Through the touch
screen Human–Machine Interface (HMI) that is integrated in the furnace’s control panel, the
CAN/ULC-S101 (CAN/ULC S101, 2019) standard fire time–temperature curve was programmed
and precisely followed (with ±2% deviation from the standard fire time-temperature curve). To
apply the transverse load near the free end of the cantilevered glulam beam, a hydraulic
cylinder mounted above the furnace to a strong steel loading frame was utilized. The distance
between where the transverse load was applied and the beam end attached to the supporting
column was measured at 1400 mm, as shown in Figure 7.
Each specimen was gradually pre-loaded in four increments each of 25% of the total load to
allow for stabilization of deflections after each load increment. The fire tests were conducted in
accordance with CAN/ULC-S101 fire endurance testing standard (CAN/ULC S101, 2019). The
followed fire endurance testing standard prescribes the procedure for loading the specimen before
and during the fire test as well as the failure criterion; however, it does not include instructions for
any specific test set-up for beam connection fire tests. The test commenced at least 30 min after the
full load was applied in line with the said fire testing standard. During the test, the load applied on
the specimen was monitored and maintained constant with the aid of a load cell and the hydraulic
cylinder connected to a manually controlled hydraulic pump. A draw-wire displacement

Initial wood cover

Connection Bolt Applied Number of Number of self- underneath the bottom row
configuration ID pattern load (kN.m) bolts tapping screws of bolts (mm)

4BP1 P1 14.8 4 6 91
Table 3. 4BP2 P2 14.8 4 6 181
Test matrix details Note(s): 4B for four bolts; and P1 and P2 for the first and second bolt pattern, respectively

Figure 7.
A schematic diagram
of the test set-up for the
connection
configuration with the
first bolt pattern (P1)
transducer installed outside the furnace was used to continuously measure the vertical Fire resistance
displacements of the beam free end throughout the test. The measured vertical deflections of the of mass timber
beam were then used to calculate the connection rotations with respect to time.
In both fire tests, the specimens were exposed to elevated temperatures that followed the
connections
CAN/ULC-S101 standard time–temperature curve (CAN/ULC S101, 2019) on all sides except
the top one, imitating the existence of a slab on top of the beam. Accordingly, the top side of the
beam, the steel loading post, and the supporting steel column were all fire protected using
several layers of 25-mm thick ceramic fibre blanket. Figure 8 shows one of the test assemblies
installed inside the furnace. The two test assemblies with different bolt patterns were loaded to
100% of the ultimate design load of the weakest unreinforced connection configuration
experimentally tested in a previous related study conducted by Owusu (2019). The failure
criterion for the connection was assumed to occur when the vertical deflection of the beam
corresponds with a connection rotational limit of the magnitude of 0.1 radians. However, both
fire tests continued beyond this point until the test assembly could not sustain the applied load
and/or a significant load drop with no recovery was observed. Thereafter, the test was
terminated and the cooling phase of the furnace commenced. The time to failure and the thermal
measurements of each test specimen were recorded, and in the subsequent phase the recorded
temperatures were used to calculate the average charring rate of the wood at the connection
location.

3. Test results and discussion

The experimental results presented in this paper describe the effects of the bolt patterns and
STS utilization on the time to failure, failure mode and rotational behaviour of glulam beam-
end connections in WSW configuration with their metal connecting components (i.e. bolt
heads and nuts and steel plate top and bottom edges) fire protected using same-species wood
plugs and strips, respectively.

3.1 Load and connection rotational limit

The applied load, failure criterion, and the connection rotation with respect to time curves for
the connection configuration with the second bolt pattern (P2) that lasted longer in standard
fire condition are all illustrated in Figure 9. The failure criterion for the connections was
determined when the vertical deflection of the beam free end corresponded with the
maximum rotational limit of the connection set at 0.1 radians. However, the tests continued
beyond this point for additionally few minutes until the test specimen could not sustain the
applied load and/or a significant load drop with no recovery was observed. This is
represented on the time-load curve by a noticeable fall with no recovery, as shown in Figure 9.

Figure 8.
Test setup of the
specimen with the
connection
configuration of the
first bolt pattern (P1)
installed inside the
furnace
JSFE 0.18 11

0.16 10
9
0.14
8
0.12
7
Rotation (rad)

Load (kN)
0.1 6

0.08 5
4
0.06
3
0.04
2
0.02 1
0 0
0 10 20 30 40 50 60
Figure 9. Time (min)
Applied load and
connection rotation
versus time curves for
the configuration with
bolt pattern P2

The average applied transverse load on the tested connections was 10.6 kN, which resulted in
14.8 kN m bending moment on the beam-end connection.

3.2 Charring rates

The average charring rate for each of the two specimens tested in this study was calculated
from Equation (1) using the thermal measurements recorded throughout the fire test by the
thermocouples installed inside the glulam section. The charring rate calculation was based on
the position of the 3008C isotherm inside the glulam section, which implies the times at which
the temperatures measured by thermocouples TC3 and TC7 (installed at 40 mm depth) and
TC9 and TC10 (installed at 20 mm depth) reached 3008C. The charring rates were calculated
by dividing the depth of the intermediate char layer (i.e. 40 – 20 5 20 mm) by the time needed
to reach 3008C at the two different depths (i.e. 40 and 20 mm), as shown in Equation (1). This
was the only method that could be used to determine the average charring rate of the
connections since the glulam beam sections were burnt out completely when it was possible
to open the furnace door after it was completely cooled down. Table 4 summarizes the
calculated charring rates for the two connection configurations tested in this study.

20–40 mm
Connection configuration ID TC3 and TC9 TC7 and TC10

Table 4. 4BPI 0.87 –

Charring rates 0.87 (Avg.)
estimated for the two 4BP2 0.80 0.72
connections tested 0.76 (Avg.)
in fire Note(s): 4B for four bolts; and P1 and P2 for the first and second bolt pattern, respectively
 β ¼ C=t (1)
Fire resistance
of mass timber
Where, β is the rate of charring (mm/min) connections
C is the depth difference of the char front to reach depths of 40 and 20 mm (mm)
t is the time difference for the thermocouples installed at depths of 20 and 40 mm to reach
3008C (min).
Based on the temperatures measured by the thermocouples installed at different depths
inside the glulam sections, it was found that the protection applied to the metal components of
the connections (i.e. bolt heads and nuts, and steel plate top and bottom edges) slowed down
the transfer of heat into the glulam sections and thus, reduced the charring rates compared to
those calculated for identical connection configurations but unreinforced that were tested by
Owusu (2019).

3.3 Fire resistance time and failure modes

The main results of the fire tests are summarized in Table 5. Based on the obtained results, the
time to failure for both connection configurations exceeded 50 min which is higher than the
minimum fire resistance rating specified as 45 min in the current National Building Code of
Canada (NBCC, 2020) for combustible construction. From Table 5, it is observed that the time
to failure of the connection with the second bolt pattern (P2) was recorded at 57 min which is
5 min (about 10.0%) more than that of the connection with the first bolt pattern (P1).
Therefore, the bolt pattern influences the time to failure of the connection to some extent.
Most importantly, and to investigate the influence of the STS on the rotational behaviour and
fire resistance time of the connections, which is the main objective of the current study, the
experimental results presented in this paper for STS-reinforced connections were compared
to those of similar fire tests conducted by Owusu (2019) on identical connections but without
the use of STS. As a result of this comparative study, it was observed that the STS-reinforced
connections performed better in terms of the connection rotation and fire resistance time than
the respective unreinforced connections. The unreinforced connections exhibited faster
increase in their rotations when compared to the reinforced connections, and this indicates
that employing STS increases the stiffness and strength of the connections and enhances
their fire resistance.
The more rapid increase in the rotations of the unreinforced connections compared to
those of respective STS-reinforced connections was mainly due to a delamination of the
glue line aligned with the top row of bolts, which caused premature wood splitting failure at
elevated temperatures. On contrary and due to the utilization of the STS, the reinforced
connections tested in the current study never experienced splitting due to delamination of
the glue lines in the fire tests. This confirms the positive influence of the utilization of STS

Connection configuration Applied load Fire resistance

ID (kN.m) time (min) Failure mode Table 5.
Summary of fire test
4BP1 14.8 52 Excessive rotation due to deflection results of the reinforced
4BP1- unreinforced 14.8 42 Splitting due to glue line delamination at connections in
(Owusu, 2019) the top row of bolts comparison to identical
4BP2 14.8 57 Excessive rotation due to deflection connections but
4BP2- unreinforced 14.8 48 Splitting due to glue line delamination at unreinforced tested by
(Owusu, 2019) the top row of bolts Owusu (2019)
JSFE in reinforcing this type of WSW glulam beam-end connections, which also resulted in
considerable reduction in the rotation values of the reinforced connections throughout the
fire tests. For instance, the failure time of the STS-reinforced connections with bolt pattern
P1 and pattern P2 increased by 10 min and 9 min, respectively, compared to the failure time
of the respective unreinforced connections experimentally tested by Owusu (2019). This
reflects an increase in the fire resistance time of this type of WSW connections by
approximately 24 and 18% in the configurations with bolt patterns P1 and P2, respectively.
Table 5 illustrates a comparison between the STS-reinforced connections of the current
study and the respective unreinforced connections experimentally tested by Owusu (2019)
in terms of their fire resistance time and failure modes.
It is worth mentioning that none of the two STS-reinforced connection configurations
exhibited brittle failure such as wood splitting or shear row-out. This is attributed mainly
to the application of the STS as perpendicular-to-wood grain reinforcement. Also, neither
the steel T-stub connector nor the steel bolts exhibited any sign of deformation in the fire
tests. Only the STS showed different extents of deformations that ranged from mild to
excessive depending on their location within the connection. The first pair of screws
located between the beam end and the first column of bolts exhibited the most noticeable
deformations compared to those exhibited by the screws located farther from the beam
connected end.

3.4 Effect of bolt pattern on the connection rotations

The vertical displacements of the beam free end measured throughout the duration of the two
fire tests were used to calculate the rotations of the connections. Figure 10 shows the
connection rotations with respect to time curves for the two fire tests. Generally, both
connection configurations exhibited similar trend of increased rotation with time but at a
different time interval specially near the end of the fire tests. For instance, both configurations
(i.e. with bolt pattern P1 and pattern P2) had a steady linear increase of their rotations for the
first 30 min or so into the fire tests; however, the rotation of the configuration with bolt pattern
P1 started to increase exponentially approximately 5 min earlier than that of the
configuration with bolt pattern P2, as shown in Figure 10. Therefore, the connection
configuration with bolt pattern P2 (4BP2) had a failure time (57 min) that is only 5 min more

Figure 10.
Time–rotation
relationships
developed to illustrate
the effect of bolt
patterns
than that (52 min) of the configuration with bolt pattern P1 (4BP1), which indicates minimal Fire resistance
influence of the bolt pattern on the fire resistance of the connection. of mass timber
Overall, the rotational capacity of the connection is regarded as a measure of the ductility
of the connection and hence, the degree of ductility of the STS-reinforced connections
connections
presented in this study was considerably higher when compared to that of identical
connections but unreinforced tested by Owusu (2019).
Figure 11a and b illustrate how the two STS-reinforced connection configurations
exhibited slight increased rotations throughout the duration of the fire tests.

3.5 Time–temperature curves

Thermal measurements recorded with the use of the twelve (12) metal-sheathed, mineral
insulated Type-K thermocouples (TC) installed at different locations within the beam-
end connection were used to plot the time–temperature curves as illustrated in
Figure 12a and b for the configurations with bolt pattern P1 and that with pattern P2,
respectively. As shown in Figure 12a and b, at the early stage of the fire tests
thermocouples TC9 and TC10 inserted at 20 mm depth from the fire exposed face of the
beam recorded a temperature of 1008C after approximately 5 min into the fire test. As the
temperature continued to increase above 1008C, the moisture in the wood started to
evaporate and as a result, at around 1808C–2008C the wood started to decompose and

Figure 11.
The two different
connection
configurations
undergoing standard
fire testing and
exhibited increased
rotations throughout
the tests
JSFE

Figure 12.
Time–temperature
curves developed from
the thermal
measurements
recorded during the fire
test of the two
connection
configurations (refer to
Figure 6 for TC
location and depth)

subsequently completed charring when wood temperature reached 3008C at

approximately 15 min from the start of the fire test. As wood begins to decompose at
2008C, its strength decreases and results in an increase in the connection rotation. In both
fire tests, it was observed that the temperature of the steel plate as well as that of
thermocouples TC4 and TC5 installed at a depth of 60 mm into the beam section were
considerably below the 3008C. This indicates that the char layer thickness of the beam
section did not reach the 60 mm depth from the fire exposed surface and that the steel
knife plate embedded in the slotted cut at the middle of the beam cross-section with its
top and bottom edges protected from fire using the glued-in wood strips did not
contribute to the charring of the wood section. However, at about 30 min it was found
that the temperatures of the steel bolt head (protected with the 30-mm thick glued-in
wood plugs) measured by TC1 and TC8 reached the 3008C temperature mark. This is
believed to assist in accelerating the charring around the bolt holes later in the fire tests,
which also led to noticeable elongation of the bolt holes that further increased the
rotations of the connection eventually in the fire tests.
 At about 40 min into the fire test, thermocouples TC3 and TC7 installed at a depth of Fire resistance
40 mm from the beam face measured temperatures that correspond with the wood charring of mass timber
temperature (i.e. 3008C), and after a few more minutes (i.e. 45 min) it is believed that the char
layer progressed to the location of the STS at the centre of the wood side member. At the
connections
failure time of the connection, it was found that thermocouples TC9 and TC10 installed at a
depth of only 20 mm inside the wood section measured temperatures as high as the furnace
temperature, whereas thermocouples TC2 and TC6 installed deep in the wood section at the
face of the embedded steel knife plate measured temperatures below 2008C.

4. Conclusions
In this paper, the fire resistance test results for two full-size glulam beam-end WSW connections
that were minimally fire protected using wood plugs and strips and reinforced with STS were
analysed and presented. The main objective of the conducted two fire tests was to investigate the
effects of bolt pattern and STS utilization on the rotational behaviour, charring rate, fire resistance
time and failure mode of this type of hybrid mass timber beam-end connection. Based on the
obtained new experimental results and the comparison of these results with those of a previous
related study conducted by Owusu (2019) on identical connections but without STS reinforcement,
the following conclusions can be drawn.
(1) The fire resistance time of both connection configurations tested in this study
exceeded the minimum fire resistance rating specified at 45 min in the NBC (NBCC,
2020) for combustible construction.
(2) The tendency of brittle failures, such as wood splitting and row shear-out, frequently
encountered in unreinforced timber connections were sufficiently prevented in the
connections tested in this study due to the utilization of STS as internal reinforcement
for the connections, which in turn increased the strength and the ductility and
enhanced the fire resistance time of the WSW connections when compared to the
respective connection configurations without reinforcement.
(3) The fire resistance time of the STS-reinforced connections increased by 24 and 18% in
the configurations with bolt patterns P1 and P2, respectively, when compared to the
failure time of their respective connection configurations without reinforcement.
(4) Although the new results are based on only two fire experiments due to budget and
time constrains, it can be recommended to utilize STS as perpendicular-to-wood grain
reinforcement in the four-bolt connection configuration with either bolt pattern P1 or
P2 to achieve a 45-min fire resistance rating.

References
Buchanan, A.H. (2000), “Fire performance of timber construction”, Progress in Structural Engineering
and Materials, Vol. 2 No. 2, pp. 278-289.
Buchanan, A.H. and Abu, A.K. (2017), Structural Design for Fire Safety, 2nd ed., John Wiley and Sons,
Chichester.
CAN/CSA O86-19 (2019), Engineering Design in Wood, Canadian Standards AssociationCanada,
Rexdale, Ontario.
CAN/ULC S101-19 (2019), Standard Methods of Fire Endurance Tests of Building Construction and
Materials, 5th ed., Underwriters Laboratories of Canada, Ottawa.
CCMC (2014), Evaluation Report: SWG ASSY VG Plus and SWG ASSY 3.0 Self-Tapping Wood Screws,
Canadian Construction Materials Centre Report No. CCMC 13677-R, National Research Council
of Canada, Ottawa.
JSFE CCMC (2018), Evaluation Report: Nordic Lam, Canadian Construction Materials Centre Report No.
CCMC 13216-R, National Research Council of Canada, Ottawa.
Lam, F., Gehloff, M. and Closen, M. (2010), “Moment-resisting bolted timber connections”, Proceedings
of the Institution of Civil Engineers – Structures and Buildings, Vol. 163 No. 4, pp. 267-274.
Letourneau-Gagnon, M., Dagenais, C. and Blanchet, P. (2021), “Fire performance of self-tapping screws
in tall mass-timber buildings”, Applied Sciences, Vol. 11, p. 3579, doi: 10.3390/app11083579.
NBCC (2020), The National Building Code of Canada. Commission on Building and Fire Codes,
National Research Council of Canada, Ottawa.
Owusu, A. (2019), “Structural performance of hybrid timber connections with varying bolt patterns at
ambient and elevated temperatures”, Master of Applied Science Thesis, Dept. of Civil and
Environmental Engineering, Carleton University, Ottawa.
Palma, P., Frangi, A., Hugi, E., Cachim, P. and Cruz, H. (2013), “Fire resistance tests on steel-to-timber
dowelled connections reinforced with self-drilling screws”, 2nd CILASCI-Ibero-Latin-Amerian
Congresso in Fire Safety Engineering, ETH Library, Coimbra, doi: 10.3929/ETHZ-A-009985513.
Peng, L., Hadjisophocleous, G., Mehaffey, J. and Mohammad, M. (2010), “Fire resistance performance
of unprotected wood–wood–wood and wood–steel–wood connections: a literature review and
new data correlations”, Fire Safety Journal, Vol. 45 Nos 6-8, pp. 392-399.
Peng, L., Hadjisophocleous, G., Mehaffey, J. and Mohammad, M. (2012), “Fire performance of timber
connections, Part 1: fire resistance tests on bolted wood-steel-wood and steel-wood-steel
connections”, Journal of Structural Fire Engineering, Vol. 3 No. 2, pp. 107-132.
Petrycki, A.R. and Salem, O.S. (2019), “Structural fire performance of wood-steel-wood bolted
connections with and without perpendicular-to-wood grain reinforcement”, Journal of Structural
Fire Engineering, Emerald Publishing, doi: 10.1108/JSFE-02-2019-0016.
Smith, I. and Snow, M.A. (2008), “Timber: an ancient construction material with a bright future”, The
Forestry Chronicle, Vol. 84 No. 4, pp. 504-510.
Viau, C. and Doudak, G. (2021), “Behavior and modeling of glulam beams with bolted connections
subjected to shock tube–simulated blast loads”, Journal of Structural Engineering, Vol. 147
No. 1, 04020305, doi: 10.1061/(ASCE)ST.1943-541X.0002876.
Zhang, C., Guo, H., Jung, K., Harris, R. and Chang, W.-S. (2019), “Screw reinforcement on dowel-type
moment-resisting connections with cracks”, Construction and Building Materials, Vol. 215,
pp. 59-72.

Corresponding author
Osama (Sam) Salem can be contacted at: [email protected]

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: [email protected]

View publication stats