This may be the author’s version of a work that was submitted/acceptedfor publication in the following source:
Sivapathasundaram, Mayooran & Mahendran, Mahen(2016)Experimental studies of thin-walled steel roof battens subject to pull-through failures.Engineering Structures, 113, pp. 388-406.
This file was downloaded from: https://eprints.qut.edu.au/96569/
c© Consult author(s) regarding copyright matters
This work is covered by copyright. Unless the document is being made available under aCreative Commons Licence, you must assume that re-use is limited to personal use andthat permission from the copyright owner must be obtained for all other uses. If the docu-ment is available under a Creative Commons License (or other specified license) then referto the Licence for details of permitted re-use. It is a condition of access that users recog-nise and abide by the legal requirements associated with these rights. If you believe thatthis work infringes copyright please provide details by email to [email protected]
License: Creative Commons: Attribution-Noncommercial-No DerivativeWorks 2.5
Notice: Please note that this document may not be the Version of Record(i.e. published version) of the work. Author manuscript versions (as Sub-mitted for peer review or as Accepted for publication after peer review) canbe identified by an absence of publisher branding and/or typeset appear-ance. If there is any doubt, please refer to the published source.
https://doi.org/10.1016/j.engstruct.2015.12.016
1
Experimental Studies of Thin-walled Steel Roof Battens
Subject to Pull-through Failures
Mayooran Sivapathasundaram1 and Mahen Mahendran2
ABSTRACT: Despite the increasing usage of thin and high strength steel roof battens in the
roof structures of low-rise buildings, recent cyclones and storms have shown that they
prematurely fail at their screw fastener connections to the rafters or trusses due to the screw
heads pulling through the batten bottom flanges. Such pull-through failures can lead to
catastrophic failures of the entire roofing systems as observed during recent high wind events.
Therefore a detailed experimental study consisting of both small scale and full scale tests was
undertaken to investigate the pull-through failures of roof battens under wind uplift loading in
relation to many critical parameters such as screw fastener tightening, roof batten geometry,
batten thickness, steel grade, screw fastener head size and screw fastener location. Effects of
underside surface and edge details of the screw fastener head, and screw fastener types were
also considered. This paper presents the details of the tests conducted in this study and the
pull-through failure load results obtained from them. Finally it presents the details of suitable
design rules and capacity reduction factors developed in this study that can be used to
accurately determine the design pull-through capacities of steel roof battens under wind uplift
loads.
KEYWORDS: Cold-formed steel structures, Light gauge steel roofing systems, Steel roof
battens, Wind uplift forces, Pull-through failures, Experiments, Design rules, Capacity
reduction factors
1PhD Researcher, Queensland University of Technology (QUT), Brisbane, Australia
2Professor, Queensland University of Technology (QUT), Brisbane, Australia
Corresponding author’s email address: [email protected]
2
1. Introduction
In recent times, lightweight building construction using thin-walled and cold-formed steel
members has become more popular in many countries. This is due to the increased research
efforts on the performance of the entire cold-formed steel building systems [1], isolated
members [2,3] and their connections under the required design actions. Lightweight roofing
systems made of thin and high strength steel roof sheeting and battens are part of this
construction in low-rise buildings. However, the critical load combination of external wind
suction and internal wind pressures that occur during high wind events such as cyclones,
tornadoes and storms often dislocate these lightweight roofing systems partially or even
completely. Past wind damage and research investigations have shown that such severe roof
failures have occurred predominantly due to the premature failures of roof connections [4].
There are two types of connections in the roofing systems. The first connection is between
the roof sheeting and the top flange of the roof batten section and this is mostly referred to as
roof sheeting to batten connection (Figures 1 and 2(a)). The second connection is between the
bottom flange of the roof batten section and the truss or rafter and this is commonly referred
to as roof batten to truss or rafter connection (Figures 1 and 2(c)). In the past, the roof
sheeting to batten connection has often failed and lead to severe failures of the roof structures
during high wind events. Among the roof sheeting to batten connection failures, the screw
fastener head that connects the roof sheeting to the top flange of roof batten pulled through
the thin roof sheeting in most cases and this localised failure is commonly referred to as pull-
through failure (see Figure 2(a)). In other cases, the screw fastener pulled out from the roof
batten, and this is referred to as pull-out failure (see Figure 2(b)).
Past research [5-16] has investigated the roof sheeting to batten connection failures in detail
and developed suitable design rules to accurately determine the connection capacities. These
extensive research efforts have greatly aided to enhance the structural safety and design of the
roof sheeting to batten connections. However, severe roof failures have continued to occur,
and in fact more severely in recent times as they now fail locally at the next level of screw
fastener connections of roof battens to the rafters or trusses. Most of these localised
connection failures were observed in the form of pull-through failures occurring at the bottom
flanges of roof battens as shown in Figure 2(d). These connection failures can cause
3
catastrophic failures of the entire roofing systems by dislocating both roof sheeting and
battens as observed during recent cyclones and tornadoes [17,18].
Despite the severity of roof batten to rafter/truss connection failures, they have not yet been
researched well. In addition, this localised pull-through failure of roof battens associated with
a tearing fracture around the screw fastener head (Figure 2(d)) substantially differs from the
previously investigated pull-through failures of roof sheeting to batten connection that are
mostly related to transverse splitting of thin steel roof sheeting at the screw fastener hole (see
Figure 2(a)). Since these pull-through failure mechanisms differ significantly, the design
rules developed for roof sheeting pull-through failures cannot be used to determine the pull-
through capacities of roof battens. Therefore this research has investigated the pull-through
failures of thin steel roof battens by undertaking an extensive experimental study under
simulated static wind uplift loads. It included investigating the effects of many critical
parameters such as screw fastener tightening, roof batten geometry, batten thickness, steel
grade, screw fastener head size, screw fastener location, underside surface and edge details of
the screw fastener head and screw fastener types. This study has lead to the development of
suitable design rules and capacity reduction factors that can be used to determine the design
pull-through capacities of roof battens more accurately. This paper presents the details of this
experimental study and the results.
2. Experimental Study
A typical roof structure has a multi-span roof batten system subjected to a uniformly
distributed wind uplift load transferred to its top flange via roof sheeting screw fasteners at
100 - 200 mm intervals (Figure 1). The wind uplift loading on the roof batten creates both a
tensile force in the screw fasteners that connect the batten bottom flanges to the rafters or
trusses and a bending moment in the batten cross section. The pull-through failure of roof
battens occurs mainly under these two actions. Therefore a two-span batten system is
considered adequate to represent the multi-span batten systems in laboratory testing. Since
this research is likely to lead to a large number of tests, testing based on small scale tests is
more desirable. Hence a detailed experimental study was first undertaken using both full
scale air-box tests (Figure 3) and a series of small scale tests including not only two-span
batten tests, but also cantilever batten tests and short batten tests (Figure 4) to identify reliable
4
small scale test methods. Two full scale tests and 66 small scale tests were conducted in these
preliminary roof batten tests.
2.1 Preliminary Roof Batten Tests
Two industrial roof battens made of G550 steel (Topspan 4055 and 4075 battens) and 10
gauge screw fasteners were used in the preliminary roof batten tests conducted to identify
reliable small scale test methods. Although they have a common geometric profile (batten
height = 40 mm, top flange width = 32 mm and bottom flange width = 12 mm) (refer Figure
2(c)), they have different base metal thicknesses of 0.55 and 0.75 mm, respectively [19].
Although a two-span batten system with two mid-span uplift loads is considered structurally
adequate to represent a multi-span batten system, the critical central support reaction of a
two-span batten system cannot be measured directly. Simple statics also cannot be used to
estimate it as the screw fastened support conditions are not known adequately. As a solution
to this problem, a modified two-span batten test set-up was also used in which two small load
cells were used with a special fastener arrangement to individually measure the critical
central support fastener reactions (refer Figures 3 and 4(a)).
Since a typical batten to rafter connection consists of two screw fastener connections on the
bottom flanges as shown in Figure 1, the applied load was continued in the tests until both
screw fasteners pulled through the batten bottom flanges. In some cases, the first pull-through
failure load for one bottom flange side was higher than the second pull-through failure load
for the other bottom flange side (refer Figure 5). In other cases, the second pull-through
failure load for the second bottom flange side was higher than the first pull-through failure
load. In all these cases, the pull-through failure loads were determined by averaging the
Individual Fastener Load Measurements (IFLM) as none of them can be neglected. Although
it was rare, in some cases, both pull-through failure loads were equal, ie. an ideal failure
situation to determine the pull-through failure load of roof battens (refer Figure 6).
The pull-through failure loads obtained from these two-span batten tests conducted with
IFLM for two different span values (300 and 450 mm) showed good agreements. In addition,
they showed reasonable agreements with the full-scale test results obtained for a span of 1200
mm. These observations indicated that the bending action of roof battens does not affect the
localised pull-through failures of roof battens significantly. Therefore the possibilities of
5
using other small scale test methods such as 350 mm long cantilever batten and 150 mm long
short batten tests were also assessed by undertaking these tests using suitable test set-ups as
shown in Figures 4(b) and (c), respectively. In the cantilever batten tests, both the fastener
reaction and bending moment at the support were simulated, however, in the short batten
tests, the uplift load was directly applied on the top flange of the roof batten and thus the
bending action of roof batten was not simulated. The accuracy of the load cells in measuring
the fastener loads was evaluated by comparing the addition of the individual fastener loads
with the total applied load (Instron machine load) in the cantilever batten and short batten
tests as seen in Figure 6.
The pull-through failure loads were determined by averaging the individual fastener loads for
the tests conducted with IFLM. However, cantilever and short batten tests can be conducted
without IFLM, in which case, the pull-through failure loads were estimated based on the total
applied load. In most cases, the first failure load was higher than the second failure load or
there was only one peak failure load (refer Figure 7). In such cases, the pull-through failure
load was considered as half of the first peak failure load. However, in some other cases, the
second failure load was higher than the first failure load, and, the pull-through failure load
was obtained by considering the average of the first and second failure loads (refer Figure 7).
Such detailed analyses have shown that the two-span tests should be conducted with IFLM
whilst the short and cantilever batten tests can be conducted without IFLM, however, the
above recommended estimation methods should be used to obtain the pull-through failure
loads accurately. Based on these preliminary roof batten tests it was recommended to use 150
mm long short batten tests in parametric experimental studies in combination with some two-
span batten tests. Complete details of these full scale and small scale test methods and the
results are given in [20].
2.2 Main Roof Batten Tests
Since there is a need to investigate the effects of many critical parameters such as screw
fastener tightening, roof batten geometry, batten thickness, steel grade, screw fastener head
size, screw fastener location, underside surface and edge details of the screw fastener head,
and screw fastener types, suitable batten specimens were also fabricated at the Queensland
University of Technology (QUT) workshop and used in the main roof batten tests. The main
roof batten tests were undertaken using QUT made roof battens and the recommended small
6
scale test methods (two-span batten tests and short batten tests) to determine the effects of
critical parameters mentioned earlier on the pull-through failures of roof battens. The tests
were conducted for each of these parameters to clearly investigate their effects on the batten
pull-through failures. However, since roof batten pull-through failures involve many critical
parameters, a suitable testing sequence was formulated to reduce the number of tests needed
to obtain decisive findings. Therefore the main roof batten tests were conducted in two
phases (Phases 1 and 2 main roof batten tests) by categorising the critical parameters into two
groups based on their anticipated importance on pull-through failures.
Phase 1 main roof batten tests were first conducted for the parameters such as screw fastener
tightening, batten height and web angle as their effects on the batten pull-through failures
were expected to be less significant. In this series of tests, a range of G550 steel roof battens
with two different thicknesses of 0.55 and 0.75 mm, three different batten heights of 40, 60
and 80 mm and three different web angles of 70o, 81o and 90o was fabricated at the QUT
workshop and then tested using 10g screw fastener connections. The level of screw fastener
tightening was chosen as the first parameter in Phase 1 main roof batten tests to determine a
suitable level of screw fastener tightening to be used constantly in the remaining tests. Since
the batten pull-through failure is highly localised to the screw fastener region, the effects of
overall deformations due to the geometrical parameters such as top flange width, batten
height and web angle were anticipated to be less significant. However, the batten height and
web angle were included in Phase 1 main roof batten tests to verify this assumption.
Following Phase 1 main roof batten tests, suitable default values were chosen for these less
significant parameters, which were then maintained in Phase 2 main roof batten tests.
Phase 2 main roof batten tests were conducted for the critical parameters such as steel grade,
batten thickness, screw fastener head size, bottom flange width and screw fastener location as
their effects on the pull-through failures of roof battens were considered to be more
significant. The steel grade, batten thickness and screw fastener head size were regarded as
critical parameters as they could directly influence localised deformations, pull-through
failure load and failure mode. In addition, the effects of bottom flange width and screw
fastener location were also anticipated to be critical parameters and included in Phase 2 tests,
as they can influence the localised deformations around the fastener head before a pull-
through failure occurs. One hundred and sixty-seven tests were conducted in this series of
roof batten tests. Fifty-one different batten configurations composed of three different steel
7
grades (G550 & G500 and G300), four different batten thicknesses (0.55, 0.75/0.80,
0.95/1.00 and 1.15 mm), three different bottom flange widths (15, 20 and 25 mm) and two
different screw fastener head sizes (10g and 12g) were used in these tests. For each of these
configurations, a minimum of three tests was conducted, however, in some cases, five tests
were also conducted to reduce the effect of experimental variations.
Phase 1 main roof batten tests were conducted using a two-span batten test set-up (Figure 8)
with Individual Fastener Load Measurements (IFLM) whilst Phase 2 main roof batten tests
were mostly conducted using short batten tests (Figure 9) since there was an inevitable need
for a large number of tests. These short batten tests were conducted without IFLM and the
pull-through failure loads were estimated using the calculation methods recommended in
[20]. Some two-span batten tests were also undertaken in Phase 2 main roof batten tests to
verify the accuracy of the pull-through failure loads obtained from the short batten tests.
Effect of Screw Fastener Tightening
In the roof batten tests, certain level of screw fastener tightening should be used in all the
tests as otherwise it may affect the final test outcomes. It is important to know the required
level of screw fastener tightening. Therefore the screw fastener tightening was selected as the
first critical parameter in Phase 1 main roof batten tests. The batten screw fasteners are
mostly drilled using torque adjustable electrical screw drivers in the current construction
practice until they stop automatically after reaching a particular level of screw fastener
tightness. In most of the cases it is the maximum level of screw fastener tightening that can
be mechanically afforded by the screw driver. Therefore the chances of loose screw fastener
connections are very low. However, screw fasteners can be overtightened and can possibly
affect the overall geometrical deformation of the batten during loading. However, it is not
established yet whether this can alter the batten pull-through failure load significantly. Hence
the purpose of this set of tests was to check whether overtightening of the batten screw
fasteners significantly affects the pull-through capacity of thin steel roof battens.
Since the relationship between the applied torque and the resultant pretension in the screw
fasteners is complicated due to the existence of frictional forces, it was decided to directly
monitor the pretensions (also fastener reactions) using small load cells (refer Figure 8). This
arrangement allowed the variation of the level of screw fastener tightening required in this set
8
of tests and then maintaining a constant level of screw fastener tightening in the subsequent
series of tests. As it was identified that an initial pretension of 100 N (0.1 kN) simulated the
real batten screw fastener connection behaviour accurately and it did not cause any
significant premature damage to the steel batten bottom flange under the screw fastener head
[20], it was considered to represent a gentle level of screw fastener tightening. A pretension
of 1000 N (1.0 kN) was chosen to represent an overtightening situation, which is 10 times
larger than the default pretension of 100 N.
Effect of Roof Batten Height
Currently available steel roof battens have a wide range of batten heights spanning from 40 to
120 mm. Although an increased batten height can enhance the bending capacity significantly,
its effects on the localised pull-through failures occurring in the batten bottom flanges are
unknown. Hence three batten heights (40, 60 and 80 mm) were included in this study to
determine the effect of batten height on the pull-through failure loads (Figure 10).
Effect of Roof Batten Web Angle
Two batten web angles of 70o and 90o were chosen in addition to the most commonly used
batten web angle of 81o to investigate the effect of batten web angle on the batten pull-
through failure loads (Figure 11).
Effect of Steel Grade
Although the roof battens used in Australia are commonly made of high strength (G550 and
G500) steels, they are often made of low strength (G300) steels in Europe. In addition, since
roof batten pull-through failures involve large localised deformations and yielding of the
bottom flange around the screw fastener head, the effects due to varying levels of ductility
available in these steels must be investigated. Hence in this study, the roof battens were made
of both low strength (G300) and high strength (G550 and G500) steels.
9
Effect of Batten Thickness (t)
The current steel roof battens are available in a range of batten thicknesses from 0.48 to 1.20
mm (base metal thickness). However, like other variables, the effect of batten thickness has
not been investigated. Therefore three steel batten thicknesses were chosen from both high
and low strength steels to determine their effects on the batten pull-through failure loads.
0.55, 0.75 and 0.95 mm thick battens were fabricated from G550 steel whilst 0.55, 0.80 and
1.00 mm thick battens were fabricated from G300 steel. In addition to these battens, 1.15 mm
thick battens made of G500 steel were also included in the tests.
Effect of Screw Fastener Head Size (d)
Three different screw fasteners, namely 10, 12 and 14 gauge Teks screw fasteners, are
commonly used to fasten the roof battens to the trusses or rafters. Although the 10g and 12g
Teks screw fasteners have different screw fastener head diameter sizes of 11 and 14.5 mm,
respectively, the 12g and 14g Teks screw fasteners have the same screw fastener head
diameter size of 14.5 mm despite the difference in the thickness of their screw fastener heads.
Hence only 10g and 12g Teks screw fasteners were used in this study.
Effect of Roof Batten Bottom Flange Width (b)
The bottom flange width of the currently available roof battens varies within a range of 12 to
26.2 mm. Hence three different batten bottom flange widths such as 15, 20 and 25 mm were
chosen in this study (see Figure 12). In this series of tests, the screw fasteners were centrally
located in the bottom flanges of roof battens as recommended by their manufacturers for a
more desirable roof batten performance [19].
Effect of Screw Fastener Location (b’)
Although it is widely recommended by the batten manufacturers to locate the screw fasteners
at the middle of the bottom flanges, there is a high tendency in reality for the screw fasteners
to be located eccentrically in the bottom flanges, in particular when roof battens with wider
bottom flanges are employed. Therefore the effect of screw fastener location was also
evaluated. The roof battens with wider bottom flanges (20 and 25 mm) made of G550 and
10
G300 steels from three different thicknesses were tested using 10g screw fastener
connections. Although the term screw fastener location means the distance between the
centre of screw fastener hole to the end of the curved section, the distance between the high
stress point (the most closest edge point of the screw head) to the end of the curved section
(b’) as shown in Figure 13(a) is used in this research as it also accounts for the size of the
screw fastener head. The b’ value of 2 mm was used in this set of tests instead of the
conventional b’ values of 4.5 and 7 mm that are possible when 10g screw fasteners (diameter
of 11 mm) are centrally located for 20 and 25 mm bottom flange battens, respectively. For
example, Figures 13(b) and (c) show the roof battens with 25 mm bottom flange when the
screw fastener was centrally located (b’ = 7 mm) and when the screw fastener was located
closer to the bottom flange-web corner (b’ = 2 mm), respectively.
Effect of Underside Surface and Edge Details of the Screw Fastener Head
The batten screw fasteners are available at present with and without seal washers. They are
mostly used without seal washers as they are needed only for the purpose of providing water
tightness when exposed to outside conditions (rain). This is unlikely in most of the real
situations as they are protected by steel roof sheeting. However to contain seal washers firmly
under their screw fastener heads, they include a groove shape with an underside edge as
shown in Figure 14. Since batten pull-through failures involve large localised deformations
and stress concentrations occurring in the screw fastener head vicinity, the sharpness of the
underside edge can substantially affect the roof batten pull-through capacity. If the sharpness
of the underside edge is high, it can cause a tearing fracture more easily compared to another
screw fastener head with a less sharp underside edge and thus it is more likely to lead to
lower pull-through capacities of roof battens in such situations.
In the past, some screw fasteners with sharp underside edges (refer Figure 14(a)) were used to
fasten the steel battens without understanding their detrimental effects on the pull-through
failures of roof battens. Therefore it was considered to determine the effect of underside
surface and edge details on the roof batten pull-through failures if such screw fasteners are
still being used. However, it was identified that the currently available batten screw fasteners
do not have any notable sharp underside edges, although they have slightly different groove
sizes to accommodate various types of seal washers (refer Figure 14(b)). Therefore no tests
11
were conducted additionally for this purpose, as their effects on the pull-through failures of
roof battens were anticipated insignificant.
Effect of Screw Fastener Type
Since timber trusses and rafters are also being used in addition to steel roof members, the
batten screw fasteners used for these applications differ from the commonly used Teks screw
fasteners in the steel roofing systems. They are commonly referred to as Type 17 screw
fasteners. However to eliminate the difficulties associated with the use of two types of batten
screw fasteners, the industry has recently introduced a new screw fastener known as
BattenZips (Figure 15) which can be used to fasten the roof battens to both timber and steel
trusses or rafters [21]. Although the groove sizes and thread lengths of the BattenZips differ
from the conventional Teks screw fasteners, the underside edges are very similar to each
other. In other words, the underside edges of the BattenZips are also not too sharp like the
commonly used Teks screw fasteners (Figure 14(b)) to cause any significant impact on the
pull-through failures of roof battens.
Since BattenZips are mostly used to fasten the steel battens to the timber truss or rafter, there
is no need for seal washers or any groove shapes to accommodate them. In addition, longer
threads are required to increase the pull-out strengths of their connections to timber. More
importantly, the screw head size of BattenZips is the same as that of 12g Teks screw fasteners
(14.5 mm). Since 12g Teks screw fasteners have been selected already for the proposed tests,
BattenZips were not used in this study.
3. Test Results and Discussions
The pull-through failure loads obtained from Phase 1 main roof batten tests undertaken to
investigate the effect of screw fastener tightening on the batten pull-through failures are
presented in Table 1. The batten pull-through failure loads decreased with increasing level of
screw fastener tightening. However, the effect was not significant compared to the implied
experimental variation of ±15% [22] for the connection tests. The effect was less than 12%
even for an increment of 10 times the default pretension value of 100 N. This could have
possibly occurred due to the premature damage caused by the overtightening to the batten
bottom flange sheeting under the screw fastener head. Since it was also identified from the
12
preliminary roof batten tests conducted using professional FS 2700 Makita electrical screw
driver that their pull-through failure loads matched reasonably well with the pull-through
failure loads obtained from the batten tests undertaken with a pretension value of 100 N [20],
it was decided to use this pretension (screw fastener tightening) value for the remaining main
roof batten tests.
Table 2 presents the results from Phase 1 main roof batten tests conducted to investigate the
effect of batten height on the pull-through failures of roof battens. Although the test pull-
through failure loads obtained from the 0.55 mm roof batten tests showed that the failure load
of 60 mm height batten increased by 14% compared to the 40 mm height batten test results, it
was almost the same for 80 mm height batten tests. In contrast, the pull-through failure loads
from the 0.75 mm roof batten tests increased with increasing batten height by 12% and 22%,
respectively, for 60 and 80 mm height battens. In addition to these mixed performances,
when the implied experimental variation of ± 15% [22] for connections tests is considered,
three out of four cases proved that the batten height did not affect the batten pull-through
failure load significantly. Therefore it was decided to adapt the mostly used batten height of
40 mm for the remaining main roof batten tests.
The test results from Phase 1 main roof batten tests on the effect of batten web angle are
presented in Table 3. Test results obtained from both 0.55 and 0.75 mm batten tests showed
that it did not affect the batten pull-through failure load significantly and the variations were
less than ± 12%. Therefore it was decided to use the most commonly used batten web angle
of 81o in Phase 2 main roof batten tests.
The test parameters investigated in Phase 1 main roof batten tests such as the level of screw
fastener tightening used, batten height and web angle have shown insignificant effects on the
batten pull-through failure behaviour and loads. Therefore the default values of these
parameters were used in Phase 2 main roof batten tests. They are: 100 N of initial pretension
(screw fastener tightening), batten height of 40 mm and web angle of 81o to represent more
realistic values and mostly used steel batten configurations in practice.
Table 4 summarises the mean pull-through failure loads obtained from Phase 2 main roof
batten tests conducted to investigate the effect of high and low strength steels. In order to
have a wide range of test results, the tests were undertaken using 18 different batten test
13
configurations by combining three different batten thicknesses (0.55, 0.75 or 0.80, 0.95 or
1.00 mm), three different batten bottom flange widths (15, 20 and 25 mm) and two different
screw fastener head sizes (10g and 12g) as shown in Table 4. Figure 16 shows the typical
load versus displacement curves obtained from G550 and G300, 0.55 mm batten tests.
Although low strength (G300) steel batten tests have shown larger deformations due to high
ductility, the pull-through failure loads were almost in the same range as obtained from high
strength (G550) steel batten tests. This finding implies an important fact that a unified design
rule cannot be derived to determine the pull-through capacities of these two different grade
steel battens. In addition, they behaved differently to each other when the screw fastener head
size was varied. This fact also supports the argument that the final design rule should be
derived separately for high and low strength steel battens.
Tensile coupon tests were conducted to determine the important mechanical parameters such
as elastic modulus of steel (E), yield strength (fy) and ultimate tensile strength (fu) required
for design and finite element modeling purposes. The averages of three tensile test results
obtained for coupons taken in the longitudinal direction were used, and the results are
summarized in Table 5.
The mean pull-through failure loads from Phase 2 main roof batten tests conducted to
investigate the effect of batten thickness are presented in Table 6. The effects of varying
thicknesses were found to be very significant and the batten pull-through failure loads
increased rapidly with thickness. The results obtained from Phase 2 main roof batten tests
undertaken to investigate the effect of screw fastener head size are presented and compared in
Table 7. Although high strength (G550) steel battens did not show significant increments,
low strength (G300) steel batten pull-through failure loads increased with the screw fastener
head size. The results from Phase 2 main roof batten tests undertaken to investigate the effect
of batten bottom flange width are presented in Table 8. The pull-through failure loads
obtained from these tests slightly decreased with the batten bottom flange width. However,
the effect was insignificant compared to the overall experimental variations. Test results
showed that the screw fastener locations (smaller b’ values) closer to the bottom flange-web
corner increased the pull-through failure loads. However, the overall effect was not
significant compared to experimental variations as seen in Tables 9 to 15. For example, G550
0.55 mm battens with 20 mm bottom flange width provided pull-through failure loads of
14
1.99, 1.84 and 1.80 kN for b’ value of 4.5 mm, and, 2.03, 1.92 and 1.98 kN for b’ value of 2
mm (Table 9).
As recommended in [20] and as mentioned in Section 2.2 of this paper, although Phase 2
main roof batten tests were mostly conducted using short batten tests, some two-span batten
tests using 650 mm long batten specimens (span of 300 mm) were also conducted to re-
confirm the accuracy of the short batten test results. Six types of steel roof battens such as
G550 0.55 mm, G550 0.75 mm, G550 0.95 mm, G300 0.55 mm, G300 0.80 mm and G300
1.00 mm roof battens were included in this series of tests. To reduce the number of tests, the
mostly used roof batten section (15 mm bottom flange width, 40 mm height, 32 mm top
flange width and 85 mm overall width) was used in these tests with 10g screw fastener
connections and, two similar tests were conducted for each type of steel roof battens. The
mean pull-through failure loads obtained from these tests were compared with the mean pull-
through failure loads obtained from the short batten tests and, the results are presented in
Table 16. The close agreements observed between these two sets of results reaffirm the
accuracy of the pull-through failure loads determined from Phase 2 main short batten tests.
The pull-through failure modes observed during the main roof batten tests conducted using
different batten configurations were similar and for example, some of them are shown in
Figures 17 and 18, from Phases 1 and 2 main roof batten tests, respectively. The localised
pull-through failures of roof battens always initiated at the hot stress point, located at the
screw fastener head edge point closest to the batten web (refer Figure 17) and then moved in
either direction by tearing around the edge of the screw fastener head.
4. Proposed Design Rules
The preliminary roof batten tests conducted to identify suitable small scale test methods
showed that the current design capacity equations largely overestimated the pull-through
capacities (Pnov) of roof battens [20]. The batten pull-through failure loads obtained from the
main roof batten tests were also compared with the pull-through failure loads predicted using
design capacity equations available in the current Australian/New Zealand cold-formed steel
structures design standard AS/NZS 4600: 2005 (Equation 1) [24], North American
specification for the design of cold-formed steel structural members AISI S100: 2012
(Equation 2) [25] and Eurocode 3 Part 1-3: 2006 (Equation 3) [26]. Table 17 shows the
15
significant overestimation of pull-through failure loads by these design equations. Since
AS/NZS 4600 and AISI S100 recommend the use of reduced tensile strengths for low ductile
steels such as G550 steel, the overestimation percentages were also calculated using 75% of
the minimum tensile strength of 550 MPa. However, Equations 1 and 2 were still found to
significantly overestimate the pull-through failure loads (Table 17). Therefore the necessity
of suitable design rules is quite important and urgent to accurately determine the batten pull-
through capacity.
Pnov = 1.5 t dw fu (1)
for 0.5 < t < 1.5 mm, where t - thickness of sheet in contact with screw head, dw - the greater
of screw head and washer diameters (8 < dw < 12.5 mm) and fu - the tensile strength of the
sheet in contact with the screw head in MPa.
Pnov = 1.5 t d’w Fu (2)
where t - thickness of member in contact with screw head or washer, d’w - the effective pull-
over (pull-through) strength diameter and Fu - the tensile strength of the member in contact
with screw head or washer.
Pnov = t dw fu (3)
where t - thickness of the thinner connected part or sheet (0.5 ≤ t ≤ 1.5 mm), dw - diameter of
the washer or the fastener head and fu - ultimate tensile strength of the thinnest sheet which is
next to the screw fastener head (fu ≤ 550 MPa).
The test results obtained from the main roof batten tests identified the critical parameters that
markedly affect the batten pull-through failures. Since high strength (G550 and G500) and
low strength (G300) steel battens have shown different pull-through failure behaviours, it is
not possible to obtain a unified design rule to predict their pull-through capacities accurately.
Therefore the design rules were developed separately for high strength (G550 and G500) and
low strength (G300) steel roof battens. All the individual test results obtained from Phase 2
main tests (refer Tables 9-15) were used in deriving suitable design equations in order to
include the possible variations observed between similar tests. As shown in the last section,
16
the critical parameters identified from Phase 2 main roof batten tests are batten thickness (t),
screw fastener head diameter (d) and ultimate tensile strength of steel (fu) and hence they
were used to develop the final design capacity equations.
The basic dimensionless formula was first obtained by dividing the batten pull-through
capacity (Pnov) by the product of critical parameters t, d and fu. Then one graph was plotted
using the values obtained from this basic dimensionless formula against the values obtained
from another suitable dimensionless formula that relates the ratio of d/t. This suitable
dimensionless formula was derived using curve fitting technique. The line of best fit was
obtained by referring higher coefficients of determination (R-squared values). Figures 19 (a)
and (b) show the graphs obtained for high and low strength steel battens, respectively.
However, since the experimental variations associated with these specific connection failures
appear to be more significant, achieving higher R-squared values (≥ 0.95) was found to be
difficult. As a solution to this, suitable residual plots (residuals versus fitted values (estimated
responses)) were plotted to reaffirm the accuracies of curve fitting techniques used to derive
the design rules. Figures 20(a) and (b) show the residual plots obtained for high strength
(G550 and G500) and low strength (G300) steel battens, respectively. Since the residuals are
distributed randomly around the zero residual horizontal line in both cases, they can be
considered as reasonably accurate [23]. The mathematical relationships obtained from
Figures 19(a) and (b) were then used to develop the following design rules (Equations 4 and
5).
G550 and G500 steel roof battens:
Pnov = 8.13 t1.98 d0.02 fu (4)
G300 steel roof battens:
Pnov = 2.87 t1.38 d0.62 fu (5)
The pull-through failure loads predicted using these two design equations were compared
with the test results to determine their accuracies in predicting the batten pull-through
capacities. Tables 9-12 present the comparisons made for high strength G550 0.55, 0.75, 0.95
and G500 1.15 mm steel battens, respectively. Tables 13-15 present the comparisons made
for low strength G300 0.55, 0.80 and 1.00 mm steel battens, respectively. Equation 4 has
17
predicted the pull-through capacities of high strength (G550 and G500) steel battens with an
error margin of -27 to +25% whilst Equation 5 has predicted the pull-through capacities of
low strength (G300) steel battens with an error margin of -16 to +18%. However, it should be
noted here that these limits were observed for a few tests only and in most other cases, the
proposed design equations have accurately predicted the pull-through capacities of roof
battens.
Considering the nature of steel roof batten pull-through failures (tearing fracture/failure
mode) and the uneven load sharing issues between the two screw fasteners in the tests, these
observed differences are possible and acceptable. Since both screw fasteners are eccentrically
located from the loading point (top flange), it is likely to create a non-uniform stress
distribution around the screw fastener head, which indicates the possibilities of increased
experimental variations in this type of screw fastener connection tests. In addition, since
G550 steel exhibits less ductile behaviour, it is more likely to affect the stress redistribution
around the screw fastener head edge, which can further increase the possible experimental
variations. This could be the reason for the increased experimental differences observed in
the less ductile G550 steel roof batten tests compared to the more ductile G300 steel roof
batten tests. Since these experimental variations related to steel roof batten tests seem
inevitable due to the many reasons as discussed above, they are also likely to cause similar
levels of differences in the comparisons. In addition, when the implied experimental variation
of ± 15% for cold-formed steel connections available in the current test standard AISI S905
[22] is considered, the overestimation percentages in predicting pull-through failure loads of
9.5% and 1.4% for high strength (G550 and G500) and low strength (G300) steel battens,
respectively, can be considered acceptable.
Equation 4 can be used for G550 & G500 steel roof battens made of 0.55 to 1.15 mm
thicknesses and Equation 5 can be used for G300 steel roof battens made of 0.55 to 1.00 mm
thicknesses. Further, these two design equations are applicable to the roof battens that are
made of batten bottom flange widths of 15 to 25 mm, batten web angles of 70o to 90o, batten
heights of 40 to 80 mm, screw fastener locations of 0.25 to 7 mm from the web to bottom
flange corner and, fastened using screw fastener head diameters of 11 to 14.5 mm. A
minimum tensile strength of 550 MPa is recommended to be used with the proposed design
equation (Equation 4) for G550 0.55, 0.75 and 0.95 mm steel roof battens. A minimum
tensile strength of 520 MPa for G500 1.15 mm steel roof battens and a minimum tensile
18
strength of 340 MPa for G300 0.55, 0.80 and 1.00 mm steel roof battens are recommended to
be used with the proposed design equations (Equations 4 and 5) respectively.
The observed pull-through failure behaviour of high strength steel battens differed from that
of low strength steel battens due to reduced ductility and lack of load sharing between
fasteners in the case of high strength steel battens. Hence the pull-through failure dominated
by tearing/fracture appears to be less dependent on the screw fastener head diameter (d) for
high strength steel battens (see Equation 4 with d0.02). Equation 4 can therefore be simplified
further without including the screw fastener head diameter (d).
5. Capacity Reduction Factors
The nominal batten pull-through capacity can be determined using the proposed design rules.
However, since the proposed design rules were developed using available test data, possible
variations expected in the material (cold-formed steel), fabrication processes of the test
specimens and testing methods should be included in the design in terms of suitable capacity
reduction factors to accurately determine the design pull-through capacities of roof battens.
The North American specification for the design of cold-formed steel structural members
AISI S100 [25] recommends the following equation to determine the capacity reduction
factor (Φ).
Φ = CΦ (Mm Fm Pm) e-X (6)
x = βo (VM2 + VF
2 + CPVP2 + VQ
2) (1/2)
Suitable values provided within brackets for the following terms were found from the same
design specification: CΦ = Calibration coefficient (1.52), Mm = Mean value of material factor
(1.1), Fm = Mean value of fabrication factor (1.0), Pm = Mean value of the professional factor
for tested component, βo = Target reliability index (3.5) for connections, VM = Coefficient of
variation of material factor (0.1), VF = Coefficient of variation of fabrication factor (0.1), CP
= Correction factor = (1+1/n)m/(m-2) for n ≥ 4 where n = number of tests and m = degrees of
freedom = n-1, VP = Coefficient of variation of test results, but not less than 6.5% and VQ =
Coefficient of variation of load effect (0.21). The values of Pm, CP and VP were calculated
19
based on the test results, and the calculated capacity reduction factors (Φ) of 0.63 and 0.67
are presented in Table 18.
The current Australian/New Zealand cold-formed steel structures design standard AS/NZS
4600 [24] recommends a capacity reduction factor of 0.5 to determine the design pull-through
(pull-over) capacity of screwed connections. As this capacity reduction factor should be
applicable to a wide range of screw fastener connections related to different types of pull-
through failures, a more conservative value of 0.5 is recommended. However, the capacity
reduction factor determined using an extensive amount of test data for a specific connection
failure type will be possibly higher than this common reduction factor of 0.5. Mahaarachchi
and Mahendran [16] recommended a higher capacity reduction factor of 0.6 that can be used
to determine the pull-through capacity in the design of trapezoidal roof sheeting with closely
spaced ribs. In this research study of roof batten pull-through failures, a common reduction
factor of 0.6 is proposed to accurately determine the design pull-through capacity of roof
battens using the proposed design rules (Equations 4 and 5).
6. Conclusions
This paper has presented the details of an experimental study undertaken using both small
scale and full scale tests to examine the pull-through failures of thin-walled steel roof battens
under simulated wind uplift loading. This study investigated the effects of many critical
parameters such as screw fastener tightening, batten height, web angle, steel grade, batten
thickness, screw fastener head size, screw fastener location, batten bottom flange width,
underside and edge details of the screw fastener head, and screw fastener types on the roof
batten pull-through failure behaviour and capacity. The main roof batten tests were
undertaken in two phases in order to reduce the number of tests required to obtain useful
conclusions. Phase 1 tests were conducted for less significant parameters such as screw
fastener tightening, batten height and web angle, whilst Phase 2 tests were undertaken for
more significant parameters such as steel grade, batten thickness, screw fastener head size,
screw fastener location and batten bottom flange width. These tests showed that the three
most critical parameters are steel batten thickness and grade and screw fastener head
diameter. The effects of these three critical parameters were considered in the development of
suitable design rules that can be used to determine the pull-through failure capacity of roof
battens accurately. Since high strength (G550 and G500) and low strength (G300) steel
20
battens showed contrasting behaviour in relation to the batten pull-through failures, the
design rules were derived separately. Finally appropriate capacity reduction factors were also
determined for the use with the proposed design rules. A common capacity reduction factor
of 0.6 is proposed to determine the design pull-through capacities of roof battens. This study
also provides the fundamental information and data for predicting the pull-through capacities
of steel roof battens under cyclic wind uplift loads. The design rules developed in this paper
for static pull-through failures can be used to include the fatigue effects caused by fluctuating
cyclonic wind loading through the use of a conservative reduction factor of 0.5.
Acknowledgements
The authors would like to thank Australian Research Council (DP120103366) for their
financial support and Queensland University of Technology for providing the necessary
facilities to conduct this research.
References
[1] Dubina, D., Stratan, A. and Nagy, Z. (2009), Full-scale Tests on Cold-formed Steel
Pitched-roof Portal Frames with Bolted Joints, Advanced Steel Construction, 5(2), pp. 175-
194.
[2] Moen, C. D. and Schafer, B. W. (2009), Elastic Buckling of Cold-formed Steel Columns
and Beams with Holes, Engineering Structures, 31, pp. 2812-2824.
[3] Keerthan, P. and Mahendran, M. (2010), Experimental Studies on the Shear Behaviour
and Strength of LiteSteel Beams, Engineering Structures, 32, pp. 3235-3247.
[4] Beck, V. R., and Stevens, L. K. (1979), Wind Loading Failures of Corrugated Roof
Cladding, Civil Engineering Transactions, Institution of Engineers, Australia, 21(1), pp. 45-
56.
[5] Mahendran, M. (1990a), Static Behaviour of Corrugated Roofing under Simulated Wind
Loading, Civil Engineering Transactions, Institution of Engineers, Australia, 32(4), pp. 211-
218.
21
[6] Mahendran, M. (1990b), Fatigue Behaviour of Corrugated Roofing under Cyclic Wind
Loading, Civil Engineering Transactions, Institution of Engineers, Australia, 32(4), pp. 219-
226.
[7] Mahendran, M. (1994), Behaviour and Design of Crest-fixed Profiled Steel Roof
Claddings under Wind Uplift, Engineering Structures, 16(5), pp. 368-376.
[8] Mahendran, M. (1995), Towards an Appropriate Fatigue Loading Sequence for Roof
Claddings in Cyclone Prone Areas, Engineering Structures, 17(7), pp. 476-484.
[9] Mahendran, M. (1997), Review of Current Test Methods for Screwed Connections,
Journal of Structural Engineering, 123, pp. 321-325.
[10] Xu, Y. L. and Reardon, G. F. (1993), Test of Screw Fastened Profiled Roofing Sheets
Subject to Simulated Wind Uplift, Engineering Structures, 15(6), pp. 423-430.
[11] Jancauskas, E. D., Mahendran, M. and Walker, G. R. (1994), Computer Simulation of
the Fatigue Behaviour of Roof Cladding during the Passage of a Tropical Cyclone, Journal of
Wind Engineering and Industrial Aerodynamics, 51(2), pp. 215-227.
[12] Xu, Y. L. (1995), Determination of Wind-induced Fatigue Loading on Roof Cladding,
Journal of Engineering Mechanics, 121, pp. 956-963.
[13] Mahendran, M. and Tang, R. B. (1998), Pull-out Strength of Steel Roof and Wall
Cladding Systems, Journal of Structural Engineering, 124(10), pp. 1192-1201.
[14] Mahendran, M. and Mahaarachchi, D. (2002), Cyclic Pull-out Strength of Screwed
Connections in Steel Roof and Wall Cladding Systems Using Thin Steel Battens, Journal of
Structural Engineering, 128(6), pp. 771-778.
[15] Mahaarachchi, D. and Mahendran, M. (2004), Finite Element Analysis and Design of
Crest-fixed Trapezoidal Steel Claddings with Wide Pans Subject to Pull-through Failures,
Engineering Structures, 26(11), pp. 1547-1559.
22
[16] Mahaarachchi, D. and Mahendran, M. (2009), Wind Uplift Strength of Trapezoidal Steel
Cladding with Closely Spaced Ribs, Journal of Wind Engineering and Industrial
Aerodynamics, 97, pp. 140-150.
[17] Boughton, G. N. and Falck, D. J. (2007), Tropical Cyclone George Damage to Buildings
in the Port Hedland Area, Technical Report 52, Cyclone Testing Station, James Cook
University, Townsville, Australia.
[18] Boughton, G. N. and Falck, D. J. (2008), Shoalwater and Roleystone WA Tornadoes
Wind Damage to Buildings, Technical Report 54, Cyclone Testing Station, James Cook
University, Townsville, Australia.
[19] Lysaght Topspan Design Manual (2012), BlueScope Steel Limited, Accessed September
01, 2012, http://www.lysaght.com/
[20] Sivapathasundaram, M. and Mahendran, M. (2015), Development of Suitable Test
Methods for the Screw Connections in Cold-formed Steel Roof Battens, Journal of Structural
Engineering (accepted on 24th November 2015, in press).
[21] Buildex Technical Data (2014), ITW Buildex Inc., Accessed September 10, 2014,
http://www.buildex.com.au/
[22] American Iron and Steel Institute (AISI) (2008), Test Methods for Mechanically
Fastened Cold-formed Steel Connections AISI S905, Washington, DC, USA.
[23] Minitab Online Documentation (2015), Minitab Inc., Accessed January 10, 2015,
http://support.minitab.com/
[24] Standards Australia (2005): AS/NZS 4600, Cold-formed Steel Structures, Sydney,
Australia.
[25] American Iron and Steel Institute (AISI) (2012), Current North American Specification
for the Design of Cold-formed Steel Structural Members AISI S100, Washington, DC, USA.
23
[26] EN 1993-1-3 (Eurocode 3) (2006), Design of Steel Structures - Part 1-3. General Rules -
Supplementary Rules for Cold-formed Members and Sheeting, European Committee for
Standardization, Brussels, Belgium.
[27] Dolamune Kankanamge, N. and Mahendran, M. (2011), Mechanical Properties of Cold-
formed Steels at Elevated Temperatures, Thin-walled Structures, 49(1), pp. 26-44.
1
Figure 1. Typical steel roof structure and its connections
Figure 2. Roof connection failures (a) Roof sheeting pull-through failures (b) Pull-out
failures (c) Roof batten to rafter connection and (d) Roof batten pull-through failures
Batten Bottom
Flange
Screw
Fastener Head
Purlin/Rafter
Top Flange Width
(32 mm)
Height (40 mm) Screw
Fastener Head
(11 mm)
Roof Batten
Rafter
Bottom Flange Width (12 mm)
Screw
Fastener
Roof
Sheeting
Transverse
Splitting Screw
Fastener
Roof
Batten
Bending of Steel
around the
Fastener Hole
(a) (b)
(c) (d)
Sheeting
Rafter
Batten
Batten to Truss/Rafter Connection
Batten
Sheeting
Rafter
Sheeting to
Batten
Connection
Batten
Sheeting
Truss/Rafter
2
Figure 3. Full scale air-box tests [20]
Figure 4. Small scale test methods (a) Two-span batten tests (b) Cantilever batten tests and
(c) Short batten tests [20]
(a) (b) (c)
Applied
Load Fastener
Reaction
Batten
Applied
Load
Batten Fastener
Reaction
Loading
Beam
Fastener
Reaction
Instron
Batten
Applied
Load
‘C’ Section
Rafter
Load
Cell 450 mm 150 mm
150 mm
350 mm
Two-span Roof Batten
1200 mm
‘C’ Section
Rafter
Roof
Sheeting
Load Cell
Air-box
Critical Batten to
Rafter Connection Air Pump
Pressure
Transducer
750 mm
3
Figure 5. Typical load versus displacement curves from two-span batten tests with IFLM
[20]
Figure 6. Typical load versus displacement curves from short batten tests with IFLM [20]
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5
Ap
pli
ed/F
aste
ner
Lo
ad (
kN
)
Displacement (mm)
Instron Load (Total) Individual Load Cell 1
Individual Load Cell 2
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6
Ap
pli
ed/F
aste
ner
Lo
ad (
kN
)
Displacement (mm)
Instron Load (Total) Individual Load Cell 1
Individual Load Cell 2 Addition of Individual Load Cells 1&2
First Pull-through Failure Second Pull-through Failure
Second Pull-through
Failure First Pull-through
Failure
4
Figure 7. Typical load versus displacement curves from short batten tests without IFLM [20]
Figure 8. Phase 1 main roof batten tests
0
1
2
3
4
5
0 1 2 3 4
Ap
pli
ed L
oad
(k
N)
Displacement (mm)
Test 1 Test 2
QUT Roof Batten
Fastener Reaction
Applied Load
‘C’ Section Rafter
Screw
Fastener Head
300 mm
Small
Load Cell
First Pull-through Failure
Second Pull-through Failure
5
Figure 9. Phase 2 main roof batten tests
Figure 10. Roof batten heights
Figure 11. Roof batten web angles
81o 90o 70o
15 mm 15 mm
15 mm
32 mm 32 mm
32 mm
40mm 40mm
40mm
80 mm
60 mm
40 mm
81o 81o
81o 15 mm
15 mm 15 mm
85 mm 85 mm 85 mm
QUT Roof Batten
Fastener Reaction
Applied Load
‘C’ Section Rafter
Screw
Fastener Head
150 mm
6
Figure 12. Roof batten bottom flange widths
Figure 13. Screw fastener location
Figure 14. Underside surface and edge details of the screw fastener head
Buildex 10g
Screw
Bremick 10g
Screw
Zenith10g
Screw
(a) (b)
Sharp underside
edge
b
b’
d
25 mm 25 mm
7 mm 2 mm
11 mm 11 mm
(a) (b) (c)
End of curved
section
Screw hole
High stress
point
25 mm 15 mm 20 mm
12.5 mm 10 mm 7.5 mm
81o
81o 81o 40mm 40mm
40mm
32 mm 32 mm
32 mm
7
Figure 15. Screw fastener types
Figure 16. Load versus displacement curves from G300 and G550 0.55 mm short batten tests
0
1
2
3
4
5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Ap
pli
ed L
oad
(k
N)
Displacement (mm)
Test-1 (G550 0.55 mm) Test-2 (G550 0.55 mm) Test-3 (G550 0.55 mm)
Test-1 (G300 0.55 mm) Test-2 (G300 0.55 mm) Test-3 (G300 0.55 mm)
Buildex
BattenZips
Buildex 12g
Screws
Buildex 10g
Screws
14.5 mm
14.5 mm
11 mm
8
Figure 17. Pull-through failure modes observed in Phase 1 main roof batten tests (a) Screw
tightening (0.1 kN), (b) Screw tightening (1.0 kN), (c) Height (80 mm), (d) Height (60 mm),
(e) Web angle (90o) and (f) Web angle (70o)
Hot Stress
Point
(a) (b)
(c) (d)
(e) (f)
9
Figure 18. Pull-through failure modes observed in Phase 2 main roof batten tests (a) G550
0.95 mm batten with 10g screw, (b) G300 1.00 mm batten with 10g screw, (c) G550 0.75 mm
batten with 20 mm bottom flange width, (d) G300 1.00 mm batten with 20 mm bottom flange
width (e) G550 0.55 mm batten with 12g screw and (f) G300 0.80 mm batten with 12g screw
(a) (b)
(c) (d)
(e) (f)
10
(a)
(b)
Figure 19. Data fittings (a) G550 & G500 steel battens and (b) G300 steel battens
R² = 0.8653
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 5 10 15 20 25 30
Pn
ov/(
tdf u
)
d/t
R² = 0.7268
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0 5 10 15 20 25 30
Pn
ov/(
tdf u
)
d/t
11
(a)
(b)
Figure 20. Residual plots (a) G550 & G500 steel battens and (b) G300 steel battens
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0 1 2 3 4 5 6 7
Res
idu
als
Fitted Values
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0 1 2 3 4 5 6
Res
idu
als
Fitted Values
1
Table 1. Effect of screw fastener tightening on pull-through failure load
Batten
Thickness (mm)
Screw
Tightening (N)
Test 1
(kN)
Test 2
(kN)
Test 3
(kN)
Test 4
(kN)
Mean
(kN)
COV
0.55 100 2.07 1.90 1.75 2.41 2.03 0.14
1000 1.42 1.66 2.29 ... 1.79 0.25
0.75 100 3.02 3.27 3.17 3.27 3.18 0.04
1000 2.72 3.09 2.69 ... 2.83 0.08
Table 2. Effect of batten height on pull-through failure load
Batten Thickness
(mm)
Height
(mm)
Test 1
(kN)
Test 2
(kN)
Test 3
(kN)
Test 4
(kN)
Mean
(kN)
COV
0.55 40 2.07 1.90 1.75 2.41 2.03 0.14
60 2.60 2.32 2.02 ... 2.31 0.13
80 2.14 2.00 2.02 ... 2.05 0.04
0.75 40 3.02 3.27 3.17 3.27 3.18 0.04
60 3.36 3.57 3.77 ... 3.57 0.06
80 3.45 4.39 3.81 ... 3.88 0.12
Table 3. Effect of batten web angle on pull-through failure load
Batten Thickness
(mm)
Web Angle
(degree)
Test 1
(kN)
Test 2
(kN)
Test 3
(kN)
Test 4
(kN)
Mean
(kN)
COV
0.55 70 1.71 1.88 1.82 ... 1.80 0.05
81 2.07 1.90 1.75 2.41 2.03 0.14
90 2.09 1.97 1.98 ... 2.01 0.03
0.75 70 3.27 3.07 3.90 ... 3.41 0.13
81 3.02 3.27 3.17 3.27 3.18 0.04
90 2.93 3.84 2.35 ... 3.04 0.25
2
Table 4. Effect of steel grade on pull-through failure load
Bottom Flange Width
(BFW) (mm) and
Screw Fastener Size
*Batten
Thickness
(mm)
Pull-through Failure Load (kN)
G550/G500 G300
Mean COV Mean COV
BFW-15, 10g 0.55 2.07 0.08 2.18 0.05
0.75/0.80 3.56 0.11 3.70 0.03
0.95/1.00 4.60 0.02 4.55 0.03
1.15 6.80 0.02 ... ...
BFW-20, 10g 0.55 1.88 0.05 2.10 0.02
0.75/0.80 3.53 0.16 3.48 0.05
0.95/1.00 4.88 0.07 4.39 0.05
1.15 6.58 0.10 ... ...
BFW-25, 10g 0.55 1.73 0.08 1.87 0.03
0.75/0.80 3.38 0.11 3.14 0.09
0.95/1.00 4.58 0.08 4.48 0.02
1.15 6.48 0.05 ... ...
BFW-15, 12g 0.55 2.08 0.10 2.66 0.05
0.75/0.80 3.46 0.04 4.19 0.02
0.95/1.00 4.32 0.02 5.49 0.02
BFW-20, 12g 0.55 1.83 0.11 2.61 0.05
0.75/0.80 3.19 0.06 4.24 0.01
0.95/1.00 4.16 0.03 5.49 0.01
BFW-25, 12g 0.55 1.62 0.05 2.36 0.01
0.75/0.80 3.12 0.06 4.03 0.02
0.95/1.00 4.35 0.01 5.16 0.03
Note: *G550: 0.55, 0.75, 0.95 mm, G500: 1.15 mm & G300: 0.55, 0.80, 1.00 mm
Table 5. Mechanical Properties from Uniaxial Tensile Tests
Steel
Grade
Batten
Thickness
(mm)
Elastic
Modulus (E)
(MPa)
Yield Stress
(fy)
(MPa)
Ultimate Stress
(fu)
(MPa)
G550 0.55 214000 710 710
0.75 225000 700 700
0.95 217000 615 615 *G500 1.15 213000 569 589
G300 0.55 200000 365 381
0.80 200000 360 380
1.00 200000 323 363
Note: *[27]
3
Table 6. Effect of batten thickness on pull-through failure load
Pull-through Failure Load (kN)
Steel
Grade
BFW
(mm) Screw
Fastener Size
Batten Thickness (mm)
0.55 0.75/0.80 0.95/1.00 1.15
G550
&
G500
15 10g 2.07 3.56 4.60 6.80
12g 2.08 3.46 4.32 ...
20 10g 1.88 3.53 4.88 6.58
12g 1.83 3.19 4.16 ...
25 10g 1.73 3.38 4.58 6.48
12g 1.62 3.12 4.35 ...
G300 15 10g 2.18 3.70 4.55 ...
12g 2.66 4.19 5.49 ...
20 10g 2.10 3.48 4.39 ...
12g 2.61 4.24 5.49 ...
25 10g 1.87 3.14 4.48 ...
12g 2.36 4.03 5.16 ...
Table 7. Effect of screw fastener head size on pull-through failure load
Pull-through Failure Load (kN)
Steel
Grade BFW
(mm)
Batten Thickness
(mm)
Screw Size (Gauge)
10g 12g
G550 15 0.55 2.07 2.08
0.75 3.56 3.46
0.95 4.60 4.32
20 0.55 1.88 1.83
0.75 3.53 3.19
0.95 4.88 4.16
25 0.55 1.73 1.62
0.75 3.38 3.12
0.95 4.58 4.35
G300 15 0.55 2.18 2.66
0.80 3.70 4.19
1.00 4.55 5.49
20 0.55 2.10 2.61
0.80 3.48 4.24
1.00 4.39 5.49
25 0.55 1.87 2.36
0.80 3.14 4.03
1.00 4.48 5.16
4
Table 8. Effect of batten bottom flange width on pull-through failure load
Pull-through Failure Load (kN)
Steel
Grade
Batten Thickness
(mm) Screw
Fastener Size
Bottom Flange Width (mm)
15 20 25
G550
&
G500
0.55 10g 2.07 1.88 1.73
12g 2.08 1.83 1.62
0.75 10g 3.56 3.53 3.38
12g 3.46 3.19 3.12
0.95 10g 4.60 4.88 4.58
12g 4.32 4.16 4.35
1.15 10g 6.80 6.58 6.48
G300 0.55 10g 2.18 2.10 1.87
12g 2.66 2.61 2.36
0.80 10g 3.70 3.48 3.14
12g 4.19 4.24 4.03
1.00 10g 4.55 4.39 4.48
12g 5.45 5.49 5.16
5
Table 9. Comparison of pull-through failure loads from G550 0.55 mm batten tests
and Equation 4
Bottom
Flange
Width
(b) (mm)
Screw
Fastener
Location
(b’) (mm)
Screw
Fastener
Diameter
(d) (mm)
Pnov,
Test
(kN)
Pnov,
Eqn. 4
(kN)
Pnov,
Test/
Pnov,
Eqn. 4
15 2.0 11.0 2.26 1.85 1.22
15 2.0 11.0 1.92 1.85 1.04
15 2.0 11.0 2.02 1.85 1.09
15 0.25 14.5 2.17 1.86 1.16
15 0.25 14.5 2.23 1.86 1.20
15 0.25 14.5 1.83 1.86 0.98
20 4.5 11.0 1.99 1.85 1.07
20 4.5 11.0 1.84 1.85 0.99
20 4.5 11.0 1.80 1.85 0.97
20 2.0 11.0 2.03 1.85 1.09
20 2.0 11.0 1.92 1.85 1.04
20 2.0 11.0 1.98 1.85 1.07
20 2.75 14.5 2.14 1.86 1.15
20 2.75 14.5 1.61 1.86 0.86
20 2.75 14.5 1.72 1.86 0.92
20 2.75 14.5 1.90 1.86 1.02
20 2.75 14.5 1.80 1.86 0.97
25 7.0 11.0 1.57 1.85 0.85
25 7.0 11.0 1.84 1.85 0.99
25 7.0 11.0 1.77 1.85 0.95
25 2.0 11.0 1.62 1.85 0.87
25 2.0 11.0 2.09 1.85 1.13
25 2.0 11.0 1.99 1.85 1.07
25 2.0 11.0 1.94 1.85 1.05
25 2.0 11.0 2.36 1.85 1.27
25 5.25 14.5 1.52 1.86 0.82
25 5.25 14.5 1.69 1.86 0.91
25 5.25 14.5 1.65 1.86 0.89
Mean 1.02
COV 0.11
Note: Measured base metal thickness (t) and ultimate tensile strength (fu) of G550 0.55 mm
battens are 0.55 mm and 710 MPa.
6
Table 10. Comparison of pull-through failure loads from G550 0.75 mm batten tests
and Equation 4
Bottom
Flange
Width
(b) (mm)
Screw
Fastener
Location
(b’) (mm)
Screw
Fastener
Diameter
(d) (mm)
Pnov,
Test
(kN)
Pnov,
Eqn. 4
(kN)
Pnov,
Test/
Pnov,
Eqn. 4
15 2.0 11 3.17 3.38 0.94
15 2.0 11 3.57 3.38 1.06
15 2.0 11 3.94 3.38 1.17
15 0.25 14.5 3.34 3.40 0.98
15 0.25 14.5 3.43 3.40 1.01
15 0.25 14.5 3.60 3.40 1.06
20 4.5 11 2.74 3.38 0.81
20 4.5 11 3.86 3.38 1.14
20 4.5 11 3.54 3.38 1.05
20 4.5 11 3.96 3.38 1.17
20 2.0 11 3.71 3.38 1.10
20 2.0 11 3.99 3.38 1.18
20 2.0 11 2.54 3.38 0.75
20 2.0 11 4.06 3.38 1.20
20 2.0 11 4.17 3.38 1.23
20 2.75 14.5 3.32 3.40 0.98
20 2.75 14.5 3.26 3.40 0.96
20 2.75 14.5 2.98 3.40 0.88
25 7.0 11 3.75 3.38 1.11
25 7.0 11 3.34 3.38 0.99
25 7.0 11 3.04 3.38 0.90
25 2.0 11 2.83 3.38 0.84
25 2.0 11 3.27 3.38 0.97
25 2.0 11 3.68 3.38 1.09
25 2.0 11 4.21 3.38 1.25
25 2.0 11 3.93 3.38 1.16
25 5.25 14.5 3.10 3.40 0.91
25 5.25 14.5 3.33 3.40 0.98
25 5.25 14.5 2.93 3.40 0.86
Mean 1.03
COV 0.13
Note: Measured base metal thickness (t) and ultimate tensile strength (fu) of G550 0.75 mm
battens are 0.75 mm and 700 MPa.
7
Table 11. Comparison of pull-through failure loads from G550 0.95 mm batten tests
and Equation 4
Bottom
Flange
Width
(b) (mm)
Screw
Fastener
Location
(b’) (mm)
Screw
Fastener
Diameter
(d) (mm)
Pnov,
Test
(kN)
Pnov,
Eqn. 4
(kN)
Pnov,
Test/
Pnov,
Eqn. 4
15 2.0 11.0 4.67 4.74 0.99
15 2.0 11.0 4.63 4.74 0.98
15 2.0 11.0 4.51 4.74 0.95
15 0.25 14.5 4.21 4.77 0.88
15 0.25 14.5 4.41 4.77 0.93
15 0.25 14.5 4.35 4.77 0.91
20 4.5 11.0 5.28 4.74 1.11
20 4.5 11.0 4.66 4.74 0.98
20 4.5 11.0 4.69 4.74 0.99
20 2.0 11.0 5.19 4.74 1.10
20 2.0 11.0 5.06 4.74 1.07
20 2.0 11.0 5.00 4.74 1.06
20 2.75 14.5 4.02 4.77 0.84
20 2.75 14.5 4.04 4.77 0.85
20 2.75 14.5 4.14 4.77 0.87
20 2.75 14.5 4.36 4.77 0.91
20 2.75 14.5 4.23 4.77 0.89
25 7.0 11.0 4.25 4.74 0.90
25 7.0 11.0 4.98 4.74 1.05
25 7.0 11.0 4.51 4.74 0.95
25 2.0 11.0 3.88 4.74 0.82
25 2.0 11.0 3.87 4.74 0.82
25 2.0 11.0 4.93 4.74 1.04
25 2.0 11.0 4.77 4.74 1.01
25 2.0 11.0 5.67 4.74 1.20
25 5.25 14.5 4.37 4.77 0.92
25 5.25 14.5 4.31 4.77 0.90
25 5.25 14.5 4.33 4.77 0.91
25 5.25 14.5 4.40 4.77 0.92
Mean 0.96
COV 0.10
Note: Measured base metal thickness (t) and ultimate tensile strength (fu) of G550 0.95 mm
battens are 0.95 mm and 615 MPa.
8
Table 12. Comparison of pull-through failure loads from G500 1.15 mm batten tests
and Equation 4
Bottom
Flange
Width
(b) (mm)
Screw
Fastener
Location
(b’) (mm)
Screw
Fastener
Diameter
(d) (mm)
Pnov,
Test
(kN)
Pnov,
Eqn. 4
(kN)
Pnov,
Test/
Pnov,
Eqn. 4
15 2.0 11.0 6.94 6.63 1.05
15 2.0 11.0 6.87 6.63 1.04
15 2.0 11.0 6.59 6.63 0.99
20 4.5 11.0 5.68 6.63 0.86
20 4.5 11.0 6.78 6.63 1.02
20 4.5 11.0 7.29 6.63 1.10
25 7.0 11.0 6.33 6.63 0.96
25 7.0 11.0 6.18 6.63 0.93
25 7.0 11.0 6.92 6.63 1.04
Mean 1.00
COV 0.07
Note: Measured base metal thickness (t) and ultimate tensile strength (fu) of G500 1.15 mm
battens are 1.15 mm and 589 MPa [27].
9
Table 13. Comparison of pull-through failure loads from G300 0.55 mm batten tests
and Equation 5
Bottom
Flange
Width
(b) (mm)
Screw
Fastener
Location
(b’) (mm)
Screw
Fastener
Diameter
(d) (mm)
Pnov,
Test
(kN)
Pnov,
Eqn. 5
(kN)
Pnov,
Test/
Pnov,
Eqn. 5
15 2.0 11.0 2.23 2.12 1.05
15 2.0 11.0 2.05 2.12 0.97
15 2.0 11.0 2.25 2.12 1.06
15 0.25 14.5 2.77 2.52 1.10
15 0.25 14.5 2.50 2.52 0.99
15 0.25 14.5 2.72 2.52 1.08
20 4.5 11.0 2.14 2.12 1.01
20 4.5 11.0 2.06 2.12 0.97
20 4.5 11.0 2.09 2.12 0.99
20 2.0 11.0 2.36 2.12 1.11
20 2.0 11.0 2.27 2.12 1.07
20 2.0 11.0 2.03 2.12 0.96
20 2.75 14.5 2.74 2.52 1.09
20 2.75 14.5 2.47 2.52 0.98
20 2.75 14.5 2.62 2.52 1.04
25 7.0 11.0 1.93 2.12 0.91
25 7.0 11.0 1.84 2.12 0.87
25 7.0 11.0 1.85 2.12 0.87
25 2.0 11.0 1.98 2.12 0.93
25 2.0 11.0 2.13 2.12 1.01
25 2.0 11.0 2.46 2.12 1.16
25 5.25 14.5 2.34 2.52 0.93
25 5.25 14.5 2.35 2.52 0.93
25 5.25 14.5 2.38 2.52 0.95
Mean 1.00
COV 0.08
Note: Measured base metal thickness (t) and ultimate tensile strength (fu) of G300 0.55 mm
battens are 0.55 mm and 381 MPa.
10
Table 14. Comparison of pull-through failure loads from G300 0.80 mm batten tests
and Equation 5
Bottom
Flange
Width
(b) (mm)
Screw
Fastener
Location
(b’) (mm)
Screw
Fastener
Diameter
(d) (mm)
Pnov,
Test
(kN)
Pnov,
Eqn. 5
(kN)
Pnov,
Test/
Pnov,
Eqn. 5
15 2.0 11.0 3.57 3.54 1.01
15 2.0 11.0 3.80 3.54 1.07
15 2.0 11.0 3.72 3.54 1.05
15 0.25 14.5 4.07 4.21 0.97
15 0.25 14.5 4.27 4.21 1.01
15 0.25 14.5 4.22 4.21 1.00
20 4.5 11.0 3.36 3.54 0.95
20 4.5 11.0 3.69 3.54 1.04
20 4.5 11.0 3.39 3.54 0.96
20 2.0 11.0 3.86 3.54 1.09
20 2.0 11.0 3.64 3.54 1.03
20 2.0 11.0 3.72 3.54 1.05
20 2.75 14.5 4.24 4.21 1.01
20 2.75 14.5 4.29 4.21 1.02
20 2.75 14.5 4.18 4.21 0.99
25 7.0 11.0 3.03 3.54 0.85
25 7.0 11.0 2.92 3.54 0.82
25 7.0 11.0 3.48 3.54 0.98
25 2.0 11.0 3.71 3.54 1.05
25 2.0 11.0 3.81 3.54 1.07
25 2.0 11.0 4.05 3.54 1.14
25 5.25 14.5 3.94 4.21 0.94
25 5.25 14.5 4.02 4.21 0.96
25 5.25 14.5 4.13 4.21 0.98
Mean 1.00
COV 0.07
Note: Measured base metal thickness (t) and ultimate tensile strength (fu) of G300 0.80 mm
battens are 0.80 mm and 380 MPa.
11
Table 15. Comparison of pull-through failure loads from G300 1.00 mm batten tests
and Equation 5
Bottom
Flange
Width
(b) (mm)
Screw
Fastener
Location
(b’) (mm)
Screw
Fastener
Diameter
(d) (mm)
Pnov,
Test
(kN)
Pnov,
Eqn. 5
(kN)
Pnov,
Test/
Pnov,
Eqn. 5
15 2.0 11.0 4.40 4.61 0.95
15 2.0 11.0 4.66 4.61 1.01
15 2.0 11.0 4.58 4.61 0.99
15 0.25 14.5 5.37 5.47 0.98
15 0.25 14.5 5.53 5.47 1.01
15 0.25 14.5 5.58 5.47 1.02
20 4.5 11.0 4.33 4.61 0.94
20 4.5 11.0 4.22 4.61 0.92
20 4.5 11.0 4.62 4.61 1.00
20 2.0 11.0 4.96 4.61 1.08
20 2.0 11.0 4.77 4.61 1.04
20 2.0 11.0 4.98 4.61 1.08
20 2.75 14.5 5.52 5.47 1.01
20 2.75 14.5 5.55 5.47 1.01
20 2.75 14.5 5.40 5.47 0.99
25 7.0 11.0 4.60 4.61 1.00
25 7.0 11.0 4.39 4.61 0.95
25 7.0 11.0 4.45 4.61 0.97
25 2.0 11.0 4.97 4.61 1.08
25 2.0 11.0 4.57 4.61 0.99
25 2.0 11.0 4.74 4.61 1.03
25 5.25 14.5 5.26 5.47 0.96
25 5.25 14.5 5.26 5.47 0.96
25 5.25 14.5 4.96 5.47 0.91
Mean 1.00
COV 0.05
Note: Measured base metal thickness (t) and ultimate tensile strength (fu) of G300 1.00 mm
battens are 1.00 mm and 363 MPa.
12
Table 16. Comparison of Mean Pull-through Failure Loads from Phase 2 Main Short Batten
Tests and Two-span Batten Tests
Batten Type
Phase 2 Main Two-
span Batten Tests
(kN)
Phase 2 Main Short
Batten Tests
(kN)
Two-span / Short
Batten Test
Failure Loads
G550 0.55 mm 1.93 2.07 0.93
G550 0.75 mm 3.91 3.56 1.10
G550 0.95 mm 4.29 4.60 0.93
G300 0.55 mm 2.35 2.18 1.08
G300 0.80 mm 3.70 3.70 1.00
G300 1.00 mm 4.83 4.55 1.06
Mean 1.02
COV 0.07
Table 17. Comparison of pull-through failure loads obtained from main roof batten tests and
Equations 1 to 3
Batten Type,
Screw Size
Tests
(kN)
Equations 1
or 2 (kN)
Overestimation
(%)
Equation 3
(kN)
Overestimation
(%)
G550 0.55 mm,
10g
2.07 6.44
211.3
80.8*
4.30 107.5
G550 0.75 mm,
10g
3.56 8.66 143.3
43.4*
5.78 62.2
G550 0.95 mm,
10g
4.60 9.64 109.6
40.6*
6.43 39.7
G550 0.55 mm,
12g
2.08 8.49 308.3
137.2*
5.66 172.2
G550 0.75 mm,
12g
3.46 11.42 230.0
94.5*
7.61 120.0
G550 0.95 mm,
12g
4.32 12.71 194.2
97.3*
8.47 96.1
Note: *Overestimation percentages were determined using 75% of minimum fu (550 MPa) =
412.5 MPa
Table 18. Capacity reduction factors (Φ)
Steel Grade Mean
(Pm)
COV
(VP)
Number of
Tests
Correction
Factor (CP)
Capacity Reduction
Factor (Φ)
G550 & G500 1.001 0.115 95 1.0325 0.63
G300 1.000 0.065 72 1.0433 0.67