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 qut.copyright@qut.edu.au

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: m.mahendran@qut.edu.au

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