Thin-Walled Structures 116 (2017) 234–249

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: www.elsevier.com/locate/tws

Full length article

Experimental study on screwed connections for sheathed CFS structures MARK

with gypsum or cement based panels
⁎
Luigi Fiorino , Tatiana Pali, Bianca Bucciero, Vincenzo Macillo, Maria Teresa Terracciano,
Raffaele Landolfo
Department of Structures for Engineering and Architecture, University of Naples “Federico II”, Naples, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: Some experimental studies on the structural response of constructions made of cold-formed steel (CFS) profile
Non-structural elements are ongoing at University of Naples “Federico II”. Since the connections between panels and CFS frame have a
Gypsum-based panels fundamental role in the global response of CFS constructions, a specific task was devoted to test this kind of
Cement-based panels elements. In particular, the main object of the activity is the experimental characterization of solutions for panel-
Screwed panel-to-steel connections
to-CFS connections commonly used in common practice, with reference to gypsum and cement-based solutions.
Shear tests
In order to define their mechanical properties, 54 tests were carried out in order to assess the effect of panel type,
thickness profile, screw diameter and number of panel layers. The results of this experimental investigation are
discussed in this paper. Furthermore, the experimental values of shear strength obtained in the tests are
compared with the available theoretical predictions and experimental data from literature.

1. Introduction codes. Furthermore, in traditional buildings, drywall systems made of

In the last few years, systems made with frames in cold-formed steel
(CFS) are increasingly used as main load bearing structures of small and
medium size buildings. These systems have high structural performance
together with specific characteristics, such as lightness, speed execution
and ability to meet high standards of performance in terms of safety,
durability and eco-efficiency. For these reasons they represent an
alternative and competitive solution to traditional constructions in
the field of low rise buildings.
Typical CFS structures consist of dry constructive systems, in which
both floors and walls are made of CFS profiles. In particular, the floors
are realized with horizontal load bearing members (joists) and a
sheathing made of gypsum or wood-based panels. Instead, the walls
are made of studs, vertical load bearing members spaced at
300−600 mm, in line with joists. The studs are fastened at each end
to wall tracks, which have the function of supporting the studs laterally
and to distribute loads among the studs. In order to resist to lateral
actions, the bracing function may be absolved by steel X-bracings or
sheathing elements, generally wood, gypsum or cement-based panels or
steel sheets (Fiorino et al., 2012 [1]). The X-braced structures are made
of all-steel materials, and the fasteners commonly used are extensively
codified by EN 1993-1-3 [24], whereas for sheathing-braced structures
only wood-based panel-to-profile screwed connections are covered by
CFS frames are increasingly used as interior partitions and exterior
walls. In partition systems the sheathing is generally made of gypsum
(for indoor application) or cement-based (for outdoor application)
panels and the walls do not represent structural elements. The panels,
both in structural and non-structural systems, are attached to the CFS
frame with specific screws or nails. It is clear that the mechanical
response of such types of systems is strongly influenced by the
interaction between panels and CFS frame through the connections.
In particular, the structural response of a CFS frame sheathed with
panels can be evaluated by considering the mechanical response of its
structural components, e.g. CFS frame, sheathing panels, panel-to-
frame connections, and frame anchors. Among these components,
usually the most important role is represented by panel-to-frame
connections. In particular, in the case of a shear wall system the lateral
resistance can be associated to the weakest failure mechanism of the
wall components. Therefore, for each component it can be defined the
failure mechanism and the smallest associated strength value defines
the wall resistance (Hc):
Hc = min(Hc, frame, Hc, panels, Hc, connections, Hc, anchors ) (1)
where Hc,frame, Hc,panels, Hc,connections, Hc,anchors are the wall resistances
associated to the failure mechanism of CFS frame, sheathing panels,
panel-to-frame connections and frame anchors, respectively. Generally,

⁎
Correspondence to: Department of Structures for Engineering and Architecture, University of Naples "Federico II", P.le V. Tecchio 80, 7th floor, 80125 Naples, Italy.
E-mail addresses: lfiorino@unina.it (L. Fiorino), tatiana.pali@unina.it (T. Pali), bianca.bucciero@hotmail.it (B. Bucciero), vincenzo.macillo@unina.it (V. Macillo),
mariateresa.terracciano@unina.it (M. Teresa Terracciano), landolfo@unina.it (R. Landolfo).

http://dx.doi.org/10.1016/j.tws.2017.03.031
Received 25 January 2017; Received in revised form 15 March 2017; Accepted 28 March 2017
0263-8231/ © 2017 Elsevier Ltd. All rights reserved.
L. Fiorino et al.
the failure mechanism of panel-to-frame connections is the weakest one
and represents the wall resistance (Hc,frame > Hc,connections;
Hc,panels > Hc,connections; Hc,anchors > Hc,connections; = > Hc = Hc,connections).
For the evaluation of the wall panel-to-frame connections lateral
resistance (Hc,connections) all analytical models available in literature
require the shear strength of the single screwed connection; e.g.
according to the lower bound method proposed by Källsner and
Girhammar [2] the wall lateral resistance Hc can be obtained through
the following Eq.:
Hc = n∙Fc
where n is the number of fastener spaces along the top and bottom rails
and Fc is the shear strength of the single connection.
Therefore, it is evident that the accurately assessment of the
structural response of sheathing-braced systems involves the execution
of specific experimental campaigns on panel-to-CFS connections. The
research results allow the identification of the best solutions for
optimizing the structural performance of these kind of systems, also
considering the design requirements provided in the modern design
codes.
In literature, several researches were carried out to study the
response of panel-to-CFS connections subjected to shear loads.
Serrette et al. [3] investigated the behaviour of the screwed
connections along the edge of the panels, considering specimens made
0.88 mm thick steel profiles connected to gypsum, plywood and fibre
board panels. Fiorino et al. [4,5] performed experimental tests with the
aim to compare the response under different loading protocols of
cement, OSB and gypsum–based panels connected to steel profiles with
thickness of 1.00 mm. Miller and Pekoz [6] and Swensen et al. [7]
performed similar connections tests on specimens characterized by
gypsum panels fastened to CFS profiles with thickness of 1.97 mm and
of 0.88 mm, respectively. Vieira and Schafer [8] carried out tests on
specimens characterized by OSB and gypsum panels fastened at
1.81 mm thick profiles, whereas Peterman et al. [9] investigated the
behaviour of the same sheathing materials connected to 0.84 mm,
1.37 mm and 2.46 mm thick steel profiles. Lange and Naujoks [10]
performed tests on connections between many type of sheathing,
including gypsum and cement-based panels and 1.50 mm and 2.00
thick profiles. Ye et al. [11,12] performed a series of shear tests on
screwed connections by considering different steel thickness (0.90 and
1.20 mm) and sheathing materials (gypsum wallboard, Bolivian mag-
nesium board and calcium silicate board and oriented strand board).
Chen et al. [13] performed monotonic tests on screwed connections
with single or double-layer gypsum sheathing panels fastened to CFS
profiles with thickness of 1.00 mm in order to compare the connection
response at ambient and elevated temperatures. Fülöp and Dubina [14]
performed a series of experiments in order to evaluate the mechanical
properties of the connections between OSB panels fastened at 1.50 mm
thick profiles. Nithyadharan and Kalyanaraman [15] presented the
details of an experimental study on behaviour and strength of the
screwed connections between CFS profiles with thickness of 1.20 mm
and calcium silicate boards, under monotonic and cyclic shear loadings.
Okasha [16] investigated the performance of connections between CFS
profiles and different wood-based sheathings varying the profile thick-
ness (0.84 mm, 1.11 mm, 1.37 mm and 1.73 mm) and the steel grade
(Grade 230 and Grade 345).
The available researches are mainly focused on connections invol-
ving steel thickness higher than 0.8 mm. In the recent years, in the
common practice the use of CSF profiles having thicknesses lower than
0.8 mm is widespread, especially for non-structural elements, such as
partition walls or ceiling. For this reason, in order to overcome the lack
of information a specific experimental campaign on panel-to-CFS
connections with steel profiles having small thickness (0.6 mm and
0.8 mm) was performed. This study is a part of activities on CFS
constructions currently conducted at University of Naples “Federico II”,
where, in last years, several researches involving experimental tests on
Thin-Walled Structures 116 (2017) 234–249
CFS walls and components were already developed (Fiorino et al.
[17–19]; Macillo et al. [20]; Iuorio et al. [21,22]).
In the present paper, the experimental characterization of different
solutions for panel-to-CFS connections with steel profiles having small
thickness is discussed. With reference to gypsum and cement-based
solutions the study is focused on the evaluation of the effect of panel
type, thickness profile, screw diameter and number of panel layers. In
the following sections, information about the experimental program,
specimen typologies, test set-up, instrumentation and loading protocol
are provided together with the summary and the comparison of the
(2)
results of the performed tests.
Finally, the test results are compared with the available theoretical
predictions and the experimental data from literature.
2. The experimental program
The experimental program involved different panel-to-CFS connec-
tions. The connections under investigation were made of different
panels (gypsum plasterboard, gypsum-fibre board, impact resistant
special gypsum board and cement-based board), which are fastened
to DX51D+Z steel grade CFS profiles with self-tapping or self-drilling
screws.
In order to characterize the properties of the steel material ancillary
tests were performed. Steel coupons were subjected to conventional
tension tests according to EN ISO 6892-1 [23]. In particular, the tested
material coupons are distinguished in four series obtained by four
different DX51D+Z steel grade coils with nominal thicknesses of 0.60
and 0.80 mm. In EN 1993-1-3 [24], for DX51D+Z steel grade minimum
values of 140 N/mm2 for yield strength and 270 N/mm2 for ultimate
tensile strength are given. In EN 10327 [25] a value for the yield
strength is not given, whereas for tensile strength a range from 270 to
500 N/mm2 is provided.
Table 1 shows the mechanical properties of tested material and the
experimental results obtained for the performed tests. In particular,
Table 1 shows the experimental yield (fy,exp), the ultimate stress (fu,exp),
the peak deformation (εp) corresponding to the ultimate stress for each
test and the elastic modulus (E) obtained as the ratio between the
recorded stress σ and strain ε on the linear branch of the response curve
with the corresponding average values, standard deviation and coeffi-
cient of variation. Besides, Table 1 shows the ratio between experi-
mental and nominal values of yield and ultimate stress.
It can be noticed that the average experimental value of the yield
stress was 20% higher than nominal value. Also the results in terms of
average experimental ultimate stress showed an increment respect to
nominal values of 16%, in average.
The specimens were grouped in 9 series, in which 6 nominally
identical specimens per each series were tested. Each series represented
a specific combination of panel (typology), number of panel layers,
screw (typology) and profile (thickness). The experimental program is
summarized in Table 2.
The series label adopted in Table 2 defines the specimen configura-
tion. Namely, the first group of characters indicate the panel type; the
second group identifies if the specimen is characterized by single (S) or
double (D) layer of panels, third group indicates the screws diameter;
the last group represents the stud profile thickness. For example, the
following specimen notation “GWB-S-35-6” stands for: standard gyp-
sum boards, single layer of panels, 3.5 diameter screws, 0.6 mm thick
stud profiles. The screw typologies adopted in the connections tests are
shown in Fig. 1 and the panel typologies in Fig. 2. In particular, the
panels are manufactured according to the standards EN 520 [26], EN
12467 [27], and EN 15283 [28] for gypsum-based, cement-based and
gypsum-fibre board panels, respectively. In order to take into account
the fabrication variability, each specimen was assembled with panels
cut from a specific panel (one panel corresponds to one specimen).
235
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

Table 1
Production values and experimental results for tested steel material.

Production values Experimental results

Coil Thickness fy,n fu,n ε LABEL fy,exp fu,exp εp E fy,exp/fy,n fu,exp/fu,n

type [mm] [MPa] [MPa] [%] [MPa] [MPa] [%] [MPa]

1 0.60 312.2 377.4 24.6 COIL 1_01 347.4 440.1 19.9 203,328 1.11 1.17
COIL 1_02 348.6 444.6 15.7 189,025 1.12 1.18
COIL 1_03 350.3 445.3 16.4 192,557 1.12 1.18
2 0.60 276.5 358.0 27.3 COIL 2_01 346.9 435.6 22.8 197,921 1.25 1.22
COIL 2_02 361.4 455.3 17.7 203,596 1.31 1.27
COIL 2_03 356.8 449.0 16.4 191,293 1.29 1.25
3 0.60 276.5 358.0 27.3 COIL 3_01 341.8 415.1 21.4 197,528 1.24 1.16
COIL 3_02 317.0 386.1 22.8 201,907 1.15 1.08
COIL 3_03 318.6 387.8 15.9 182,302 1.15 1.08
4 0.80 300.0 370.0 34.0 COIL 4_01 366.0 413.1 18.2 198,529 1.22 1.12
COIL 4_02 369.8 414.4 21.5 172,862 1.23 1.12
COIL 4_03 367.4 411.0 20.9 190,523 1.22 1.11
Average values 349.3 424.8 19.2 193,447 1.20 1.16
St. deviation 17.3 23.5 2.7 9126 0.07 0.06
C.o.V. 0.05 0.06 14.2 0.05 0.06 0.05

3. Test on panel-to-CFS connections
3.1. Specimens and test set-up
The monotonic tests on panel-to-CFS connections were performed
taking into account the prescriptions of EN 520 [26], which gives
requirements and test methods for gypsum plasterboards used in
buildings. In particular, Section 5.13 of EN 520 describes a method
for performing shear tests on panel-to-wood frame connections. There-
fore, specimens and test set-up given in EN 520 were adapted to the
case of cold-formed steel profiles in this experimental campaign.
The specimen consisted of one (in case of single layer) or two (in
case of double layer) sheathing panels screwed to the opposite flanges
of two cold-formed steel profiles. The single panel had dimensions
equal to 300×100 mm (length×width). The adopted fasteners were
self-tapping or self-drilling screws characterized by different head,
diameter, length and thread depending on the used panel type. The
profiles were 75×50×7.5 mm (outside-to-outside web depth×outside-
to-outside flange size×outside-to-outside lip size) lipped channel stud
sections fabricated with DX51D+Z steel grade.
Each panel was connected to the profiles by two screws with an
edge loading distance (defined as the distance from the centre of the
screw to the adjacent edge of the connected panel in parallel direction
to the load transfer) equal to 15 mm, which represents the typical edge
distance used for connecting the wall stud profiles to the sheathing
panels. The shear load was applied to the connection specimens by
means of two 10 mm thick S275JR steel holder, each of which was
bolted to the stud profiles with four M8 8.8 grade bolts. In order to
avoid specimen damages during the clamping of the testing machine,
the bolts between steel holders and steel profiles were tightened only
after the clamping of the steel holders in machine wedge grips, in such a
way the specimen was not subjected to loads during this operation.
Fig. 3 shows the drawings of the test set-up.
3.2. Test instrumentation and loading protocol
Tests were performed by using a universal test machine available at
the laboratory of the Department of Structures for Engineering and
Architecture of University of Naples “Federico II” and four linear
variable differential transducers (LVDTs) were used to measure the
relative displacements between panels and steel profiles, as shown in
Fig. 4.
The tests were performed in monotonic regime by subjecting the
connection specimens to progressive displacements up to failure. The
displacement-controlled test procedure involved displacements at a rate
of 0.15 mm/s and data recorded with a sampling frequency of 5 Hz.

Table 2
The experimental program.

Series number Label Panel type / thickness No. panel layers Screw type Screw diameter Stud thickness No. of tests
[mm] [mm] [mm] [mm]

1 GWB-S-35-6 Standard gypsum board / 12.5 Single Bugle head screw 3.5 0.6 6
2 RGWB-S-39-6 Impact resistant gypsum board / Single Wafer head with full thread screw 3.9 0.6 6
12.5
3 GFB-S-39-6 Gypsum fibre board / 12.5 Single Flat head 3.9 0.6 6
with milling ribs screw
4 CP-S-42-8 Outdoor cement board / 12.5 Single Wafer head 4.2 0.8 6
with milling ribs screw
5 GWB-D-35-6 Standard gypsum board / 12.5 Double Bugle head screw 3.5 0.6 6
6 GFB-D-35-6 Gypsum fibre board / 12.5 Double Flat head 3.5 0.6 6
with milling ribs screw
7 RGWB-S-39-8 Impact resistant gypsum board / Single Wafer head with partial thread screw 3.9 0.8 6
12.5
8 GFB-S-35-6 Gypsum fibre board / 12.5 Single Flat head with 3.5 0.6 6
milling ribs screw
9 GFB-D-39-6 Gypsum fibre board / 12.5 Double Flat head 3.9 0.6 6
with milling ribs screw
Total no. of tests 54

236
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

Fig. 1. Screws typologies adopted in the connections tests.

3.3. Test results 6. du : conventional ultimate displacement corresponding to a load

A typical experimental response in terms of load vs. displacement
curve obtained by shear tests on connections is shown in Fig. 5.
Parameters used to describe the experimental behaviour are:
1. Fp : maximum load;
2. dp : displacement corresponding to Fp ;
3. Fe : conventional elastic limit load, equal to 0.40Fp on the ascending
part of the response curve;
4. de : displacement corresponding to Fe ;
5. ke : conventional elastic stiffness, equal to Fe / de ;
equal to 0.80Fp on the post-peak branch of the response curve;
7. μ : ductility, equal to du / de ;
8. E : dissipated energy, defined as the area under the response curve
for displacements not more than the conventional ultimate displace-
ment du .
The above mentioned parameters are defined on the load (F ) vs.
displacement (d ) curve obtained by each test. The load F is the unit
load recorded for a single tested connection set equal to F =Ftot /2 , where
Ftot is the total recorded load and 2 is the total number of the tested
connections, whereas the displacement d is the average of the measures

Fig. 2. Panel typologies adopted in the connections tests.

237
L. Fiorino et al.
Fig. 3. Drawings of the test set-up.
recorded by the only two LVDTs representative of connections reaching
the rupture. Notice that only Fp, dp, de, du and E can be considered
independent parameters, whereas Fe, ke and μ are dependent variables.
The load vs. displacement curves obtained for each connection
series are shown in Fig. 6 and in Fig. 7, for the specimens with one and
two panel layers, respectively.
In order to compare the results of different tested connection
typologies, the representative experimental curves of one panel layer
connection typologies are shown in Fig. 8. Instead, the representative
curves of one and two panel layers connections having same sheathing
material, steel thickness and screw diameter are compared in Fig. 9.
It can be noted that for specimens with single panel layer the
solution with standard gypsum board (GWB-S-35-6) was the least
238
Thin-Walled Structures 116 (2017) 234–249
Fig. 4. Instrumentation for panel-to-steel connections. (a) Single panel layer. (b) Double
panel layer.
Fig. 5. Typical load vs. displacement response curve.
resistant, whereas the higher value of strength was obtained for
connections with gypsum fibre board (GFB-S-39-6 and GFB-S-35-6).
Connections with impact resistant gypsum board and outdoor cement
board gave intermediate values of strength. In addition, it can be
observed that the strength increased with the number of panel layers.
Also in the case of double layer of panels, the solution with standard
gypsum board panels (GWB-D-35-6) was the least resistant and the
connections with gypsum fibre board panels (GFB-D-35-6 and GFB-D-
39-6) were the most resistant.
Tables 3 and 4 provide the values of the experimental results for all
tested connections together to average values, standard deviation and
coefficient of variation (C.o.V.) for nominally identical specimens. In
addition, Tables provide the experimental characteristic values of the
maximum and conventional elastic limit loads, which are defined
according to EN 1990 [29] through the following relationship:
Xk = mX {1−knVX } (3)
where:
Xk is the characteristic value;
mx is the mean of the n sample results;
kn is function of the test number (n), which was assumed equal to
2.18 (value given by [29] for n=6 and Vx not know);
VX is the coefficient of variation of the n sample results.
From the examination of all results in terms of statistical dispersion,
it can be noticed that the parameter relevant to the description of the
initial shear response (de) is the most scattered value, the strength (Fp) is
the lowest scattered, and the displacement parameters related to the
ultimate behaviour (dp, du) show an intermediate dispersion. In fact, the
C.o.V. of de is in the range from 0.12 to 0.90, the C.o.V. of Fp ranges
from 0.03 and 0.19, and the C.o.V. for dp and du is in the range from
0.16 to 0.45. This difference can be mainly attributed to the higher
sensibility of the first part of the response to the imperfections due to
the mounting process.
In this regards, for the case of single panel layer the presence of the
coating paper seems to play a key role, increasing the effect of the
L. Fiorino et al.
Fig. 6. Experimental load vs. displacement curves for one panel layer connection tests. (a) GWB-S-35-6. (b)RGWB-S-39-6. (c) GFB-S-39-6. (d) CP-S-42-8. (e) RGWB-S-39-8. (f) GFB-S-35-6.
imperfections. This appears significant if the results in terms of C.o.V.
are grouped for papered and unpapered specimens. Indeed, usually the
C.o.V. of papered specimens is higher than unpapered ones. This
difference becomes apparent for the parameter representing the initial
response, i.e. in the case of de the C.o.V. ranges from 0.31 to 0.90 for
papered specimens, whereas it ranges from 0.12 to 0.27 for those
unpapered.
For all tested connections the failure mechanism was affected by the
edge loading distance, which was equal to 15 mm. In fact, for this value
of the edge loading distance, the observed failure mechanism for all
239
Thin-Walled Structures 116 (2017) 234–249
tests was the tilting of the screws and breaking of the panel edge, as
shown in Figs. 10 and 11 for connections with single and double panel
layer, respectively. Nevertheless, the extension of the panel portion
affected by the rupture depends on the panel type. In particular, for
connections with standard (GWB) or impact resistant (RGWB) gypsum
boards, the panel portion affected by the rupture can be obtained by
considering a diffusion of about 90° starting from the screw (Figs. 10
and 11), whereas for the solutions with gypsum fibre (GFB) or outdoor
cement (CP) boards, the rupture involved an higher panel portion. The
following sections present the comparison among the different tested
L. Fiorino et al.
Fig. 7. Experimental load vs. displacement curves for two panel layers connection tests. (a)GWB-D-35-6. (b) GFB-D-35-6. (c) GFB-D-39-6.
Fig. 8. Representative experimental load vs. displacement curves for one panel layer
connection tests.
connections.
4. Comparison of test results
4.1. Comparison among solutions with single layer of panels
In this section the behaviour of different connections with single
panel layer used for CFS walls (GWB-S-35-6, GFB-S-39-6, RGWB-S-39-6
and CP-S-42-8) is examined. The results obtained for specimens GFB-S-
35-6 and RGWB-S-39-8 are neglected in this section because these tests
were performed only to evaluate the effect of the screw diameter (GFB-
240
Thin-Walled Structures 116 (2017) 234–249
Fig. 9. Representative experimental load vs. displacement curves for two panel layers
connection tests.
S-35-6 vs GFB-S-39-6) and profile thickness (RGWB-S-39-8 vs RGWB-S-
39-6). These solutions are not usually adopted in common practice.
The average values of the maximum load Fp, conventional elastic
stiffness ke, ductility μ and absorbed energy E are shown in Fig. 12. In
the same figures, the standard deviation of these values is illustrated.
From the examination of results in terms of the maximum load Fp
(Fig. 12.a) it can be noted that the solution with gypsum fibre board
(GFB-S-39-6) was the most resistant (0.71 kN in average), whereas the
solution with standard gypsum board (GWB-S-35-6) was the least
resistant (0.27 kN in average), and the remaining series showed similar
values of strength in the range defined by the previous configurations
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

Table 3
Experimental results for specimens characterized of one panel layer.

Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E [kNmm]

GWB-S-35-6_01 0.11 0.30 0.36 0.28 1.97 3.20 10.51 0.70

GWB-S-35-6_02 0.10 0.10 1.02 0.26 1.05 2.81 28.14 0.61
GWB-S-35-6_03 0.12 0.06 2.13 0.30 0.80 2.26 40.44 0.58
GWB-S-35-6_04 0.10 0.07 1.41 0.24 2.21 3.26 47.26 0.69
GWB-S-35-6_05 0.11 0.08 1.32 0.28 1.48 2.27 26.90 0.51
GWB-S-35-6_06 0.11 0.04 2.53 0.26 0.72 2.72 65.35 1.26
Average values 0.11 0.11 1.46 0.27 1.37 2.76 36.43 0.73
Standard deviation 0.01 0.10 0.78 0.02 0.62 0.43 18.98 0.27
C.o.V. 0.07 0.90 0.53 0.07 0.45 0.16 0.52 0.37
Characteristic values 0.09 0.23
Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E [kNmm]
RGWB-S-39-6_01 0.21 0.08 2.83 0.54 0.66 2.01 26.45 0.95
RGWB-S-39-6_02 0.19 0.13 1.50 0.47 0.83 2.22 17.67 0.86
RGWB-S-39-6_03 0.16 0.09 1.80 0.41 1.69 2.29 25.43 0.77
RGWB-S-39-6_04 0.18 0.07 2.63 0.45 0.76 2.10 30.84 0.80
RGWB-S-39-6_05 0.13 0.08 1.67 0.33 0.73 1.79 22.55 0.51
RGWB-S-39-6_06 0.14 0.05 2.72 0.34 1.04 2.09 42.05 0.65
Average values 0.17 0.08 2.19 0.42 0.95 2.08 27.50 0.76
Standard deviation 0.03 0.03 0.60 0.08 0.39 0.17 8.36 0.16
C.o.V. 0.19 0.31 0.27 0.19 0.41 0.08 0.30 0.21
Characteristic values 0.10 0.25
Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E [kNmm]
GFB-S-39-6_01 0.29 0.18 1.60 0.73 1.27 1.27 10.24 1.10
GFB-S-39-6_02 0.28 0.14 2.02 0.69 1.11 1.77 12.99 1.01
GFB-S-39-6_03 0.30 0.17 1.74 0.74 1.03 1.39 8.16 0.79
GFB-S-39-6_04 0.28 0.22 1.32 0.71 1.32 1.83 8.49 1.01
GFB-S-39-6_05 0.28 0.10 2.75 0.71 0.80 1.49 14.43 0.86
GFB-S-39-6_06 0.27 0.13 2.07 0.68 0.79 1.04 7.89 0.52
Average values 0.28 0.16 1.92 0.71 1.05 1.46 10.37 0.88
Standard deviation 0.01 0.04 0.49 0.02 0.23 0.30 2.75 0.21
C.o.V. 0.03 0.26 0.26 0.03 0.21 0.21 0.27 0.24
Characteristic values 0.26 0.66
Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E[kNmm]
CP-S-42-8_01 0.19 0.13 1.46 0.47 1.32 2.15 16.73 0.81
CP-S-42-8_02 0.19 0.10 1.90 0.49 1.37 1.86 18.21 0.72
CP-S-42-8_03 0.17 0.13 1.25 0.41 1.54 2.48 18.66 0.85
CP-S-42-8_04 0.19 0.12 1.57 0.48 1.72 2.56 20.89 1.02
CP-S-42-8_05 0.17 0.10 1.72 0.42 1.63 3.98 40.24 1.44
CP-S-42-8_06 0.19 0.12 1.52 0.47 2.28 3.62 29.25 1.41
Average values 0.18 0.12 1.57 0.46 1.64 2.77 24.00 1.04
Standard deviation 0.01 0.01 0.22 0.03 0.35 0.84 9.12 0.31
C.o.V. 0.07 0.12 0.14 0.07 0.21 0.30 0.38 0.30
Characteristic values 0.16 0.39
Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E [kNmm]
RGWB-S-39-8_01 0.14 0.02 7.83 0.35 0.96 1.99 111.99 0.61
RGWB-S-39-8_02 0.18 0.07 2.71 0.46 1.71 3.57 53.03 1.39
RGWB-S-39-8_03 0.16 0.11 1.39 0.40 1.28 2.70 23.54 0.88
RGWB-S-39-8_04 0.21 0.09 2.50 0.53 1.41 2.12 24.83 0.88
RGWB-S-39-8_05 0.15 0.09 1.67 0.37 0.85 2.50 27.91 0.74
RGWB-S-39-8_06 0.22 0.18 1.22 0.55 1.56 2.25 12.57 0.92
Average values 0.18 0.09 2.89 0.44 1.30 2.52 42.31 0.91
Standard deviation 0.03 0.05 2.50 0.08 0.34 0.57 36.66 0.27
C.o.V. 0.19 0.58 0.86 0.19 0.26 0.23 0.87 0.29
Characteristic values 0.10 0.26
Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E [kNmm]
GFB-S-35-6_01 0.27 0.11 2.49 0.69 1.16 1.51 13.73 1.63
GFB-S-35-6_02 0.28 0.19 1.46 0.70 1.96 2.78 14.64 3.02
GFB-S-35-6_03 0.27 0.17 1.59 0.69 1.43 2.26 13.08 2.51
GFB-S-35-6_04 0.24 0.18 1.30 0.59 2.27 2.96 16.30 2.83
GFB-S-35-6_05 0.23 0.25 0.90 0.57 1.65 2.25 8.85 1.94
GFB-S-35-6_06 0.25 0.23 1.09 0.63 2.01 2.61 11.23 2.57
Average values 0.26 0.19 1.47 0.64 1.75 2.40 12.97 2.42
Standard deviation 0.02 0.05 0.56 0.05 0.41 0.52 2.63 0.53
C.o.V. 0.08 0.27 0.38 0.08 0.23 0.22 0.20 0.22
Characteristic values 0.21 0.53

(0.46 kN and 0.42 kN in average for CP-S-42-8 and RGWB-S-39-6,
respectively). Therefore, the maximum load increased of about 2.63,
1.69 and 1.54 times by using a gypsum fibre board (GFB-S-39-6)
connection rather than standard gypsum board (GWB-S-35-6), impact
resistant gypsum board (RGWB-S-39-6) and outdoor cement board (CP-
S-42-8) connections, respectively.
Finally, from the examination of results in terms of resistance
(gypsum fibre board more resistant than gypsum wall board), it can
be noticed that higher values of strength correspond to higher values of
panel bearing capacity (embedment strength fh, see Section 5) and
wider extension of the panel portion affected by the rupture, even if the
last aspect seems to have a secondary role.

241
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

Table 4
Experimental results for specimens characterized of two panel layers.

Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E [kNmm]

GWB-D-35-6_01 0.20 0.33 0.60 0.49 4.07 8.12 24.76 3.27

GWB-D-35-6_02 0.18 0.23 0.80 0.45 2.46 4.30 18.95 1.60
GWB-D-35-6_03 0.19 0.29 0.67 0.48 3.82 5.69 19.65 2.28
GWB-D-35-6_04 0.20 0.32 0.61 0.50 3.37 5.85 18.04 2.40
GWB-D-35-6_05 0.18 0.28 0.62 0.44 2.75 4.01 14.08 1.47
GWB-D-35-6_06 0.18 0.19 0.94 0.45 3.85 7.64 40.19 2.99
Average values 0.19 0.27 0.71 0.47 3.39 5.94 22.61 2.34
Standard deviation 0.01 0.05 0.14 0.02 0.65 1.68 9.27 0.72
C.o.V. 0.05 0.20 0.19 0.05 0.19 0.28 0.41 0.31
Characteristic values 0.17 0.42
Label Fe [kN] de [mm] ke [kN/mm] Fp[kN] dp [mm] du [mm] μ E [kNmm]
GFB-D-35-6_01 0.37 0.69 0.54 0.93 2.25 3.58 5.17 2.34
GFB-D-35-6_02 0.34 0.23 1.50 0.85 1.31 2.77 12.26 1.89
GFB-D-35-6_03 0.37 0.20 1.86 0.93 1.76 2.92 14.57 2.20
GFB-D-35-6_04 0.32 0.12 2.75 0.80 1.01 3.34 28.56 2.34
GFB-D-35-6_05 0.40 0.20 1.96 0.99 1.60 2.13 10.51 1.68
GFB-D-35-6_06 0.34 0.19 1.80 0.86 1.36 2.80 14.66 1.94
Average values 0.36 0.27 1.73 0.89 1.55 2.92 14.29 2.06
Standard deviation 0.03 0.21 0.72 0.07 0.43 0.50 7.82 0.27
C.o.V. 0.08 0.77 0.41 0.08 0.28 0.17 0.55 0.13
Characteristic values 0.30 0.74
Label Fe [kN] de [mm] ke [kN/mm] Fp [kN] dp [mm] du [mm] μ E [kNmm]
GFB-D-39-6_01 0.32 0.30 1.08 0.81 1.50 2.09 6.95 1.26
GFB-D-39-6_02 0.32 0.19 1.72 0.80 1.73 3.26 17.47 2.20
GFB-D-39-6_03 0.32 0.32 1.00 0.79 2.21 3.09 9.80 1.91
GFB-D-39-6_04 0.30 0.21 1.43 0.76 1.75 2.85 13.47 1.78
GFB-D-39-6_05 0.36 0.17 2.16 0.91 1.16 2.88 17.07 2.19
GFB-D-39-6_06 0.33 0.21 1.55 0.83 1.19 2.08 9.72 1.36
Average values 0.33 0.23 1.49 0.82 1.59 2.71 12.41 1.79
Standard deviation 0.02 0.06 0.43 0.05 0.40 0.51 4.30 0.40
C.o.V. 0.06 0.26 0.29 0.06 0.25 0.19 0.35 0.22
Characteristic values 0.28 0.70

Fig. 10. Breaking of the sheathing edge for connections with single layer of panels. (a) GWB-S-35-6. (b) RGWB-S-39-6. (c) GFB-S-39-6. (d) CP-S-42-8. (e) RGWB-S-39-8. (f) GFB-S-35-6.

242
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

Fig. 11. Breaking of the sheathing edge for connections with double layer of panels. (a)GWB-D-35-6. (b)GFB-D-35-6. (c) GFB-D-39-6.

Fig. 12. Average values of Fp, ke, μ and E obtained for solutions with single layer of panel. (a) Maximum load. (b) Conventional elastic stiffness. (c) Ductility. (d) Absorbed energy.

For the conventional elastic stiffness ke (Fig. 12.b), it was observed
that the solutions with impact resistant gypsum board (RGWB-S-39-8)
and gypsum fibre board (GFB-S-39-6) had the highest values (2.19 kN/
mm in average for RGWB-S-39-6 and 1.92 kN/mm in average for GFB-
S-39-6), whereas the solutions with outdoor cement board (CP-S-42-8)
and with standard gypsum board (GWB-S-35-6) had the lowest values
(1.57 kN/mm in average for CP-S-42-8 and 1.46 kN/mm in average for
GWB-S-35-6). Respect to the case of strength (Fp), the comparison in
terms of stiffness (ke) presents less quantitative differences among
different series, with a ratio between maximum and minimum value not
more than 1.5.
For the ductility μ (Fig. 12.c), the lowest value (10.37 in average)
was obtained for the specimens with gypsum fibre board (GFB-S-39-6)
and the highest value (36.43 in average) was observed for the speci-
mens with standard gypsum board (GWB-S-35-6), whereas for the other
solutions were obtained average values of 24.00 and 27.50 for outdoor
cement board (CP-S-42-8) and impact resistant gypsum board (RGWB-
S-39-6), respectively. Therefore, the performance in terms of ductility
presented a comparative result among the different solutions opposite
than that obtained for the strength, i.e. standard gypsum board series
(GWB-S-35-6) with minimum strength and maximum ductility and
gypsum fibre board series (GFB-S-39-6) with maximum strength and
minimum ductility. In addition, the difference of ductility among the
series is significant, with an average value for standard gypsum board

243
L. Fiorino et al.
Fig. 13. Average values of Fp, ke, μ and E obtained for solutions with double layer of panels. (a)Maximum load. (b) Conventional elastic stiffness. (c) Ductility. (d) Absorbed energy.
series (GWB-S-35-6) 3.5 times higher than for gypsum fibre board series
(GFB-S-39-6).
Also the comparison in terms of absorbed energy (Fig. 12.d) showed
not apparent quantitative differences, with average values in the range
from 0.73 kNmm for standard gypsum board series (GWB-S-35-6) to
1.04 kNmm for outdoor cement board series (CP-S-42-8) and a
maximum-to-minimum ratio equal to 1.4.
4.2. Comparison among solutions with double layer of panels
The comparison among the results of series with two layers of
panels is investigated in this section. This comparison is limited to two
solutions only: GWB-D-35-6 and GFB-D-39-6. The results obtained for
specimens GFB-D-35-6 are neglected in this section because these tests
were performed only to evaluate the effect of the number panel layers
(GFB-D-35-6 vs GFB-S-35-6). This solution is not usually adopted in
common practice.
Fig. 13 shows the average values of the maximum load Fp
(Fig. 13.a), conventional elastic stiffness ke (Fig. 13.b), ductility μ
(Fig. 13.c) and absorbed energy E (Fig. 13.d).
From an examination of the results, it can be noticed that, as for
single layer of panels (GWB-S-35-6 and GFB-S-39-6), also for solutions
with double layer, gypsum fibre board (GFB-D-39-6) was more resistant
and stiffer but less ductile than standard gypsum board (GWB-D-35-6).
In fact, the average values of strength, stiffness and ductility were
0.82 kN, 1.49 kN/mm, 12.41 and 0.47 kN, 0.71 kN/mm, 22.61 for
gypsum fibre board (GFB-D-39-6) and standard gypsum board (GWB-
D-35-6), respectively. The only difference was represented by the
comparison of the absorbed energy, which for gypsum fibre board
(GFB-D-39-6) (1.79 kN mm in average) was lower than standard
gypsum board (GWB-D-35-6) (2.34 kN mm in average).
4.3. Effect of the profile thickness
The effect of profile thickness can be investigated by comparing the
results obtained for the series RGWB-S-39-6 and RGWB-S-39-8, which
represent connections between single layer of impact resistant gypsum
boards and profiles having thickness equal to 0.6 mm and 0.8 mm,
respectively.
The average values of the maximum load Fp, conventional elastic
stiffness ke, ductility μ and absorbed energy E are shown in Fig. 14. In
the same figures, the standard deviation of these values is illustrated.
The results show as the increment of the steel profile thickness
244
Thin-Walled Structures 116 (2017) 234–249
produced a better response of connections, especially for stiffness, i.e.
2.19 kN/mm in average for thickness of 0.6 mm (RGWB-S-39-6) and
2.89 kN/mm in average for thickness of 0.8 mm (RGWB-S-39-8), and
ductility, i.e. 27.50 in average for thickness of 0.6 mm (RGWB-S-39-6)
and 42.31 in average for thickness of 0.8 mm (RGWB-S-39-8). In fact,
from the 0.6 mm to 0.8 mm thick steel the conventional elastic stiffness
and ductility improved by 1.32 and 1.54 times, respectively, whereas
strength (0.42 kN in average for RGWB-S-39-6 and 0.44 kN in average
for RGWB-S-39-8) and absorbed energy (0.76 kN mm in average for
RGWB-S-39-6 and 0.91 kN mm for in average RGWB-S-39-8) improved
by 1.05 and 1.20 times, respectively. Finally, it can be noted that the
variation of steel thickness from 0.6 to 0.8 mm does not influence
significantly the behaviour of screwed connections, especially for the
strength. This result can be explained by the observation that no
apparent variation in terms of restraint against screw tilting occurred
for thickness profile in the range from 0.6 to 0.8 mm.
4.4. Effect of the screw diameter
In this experimental research two different screw diameters, 3.5 mm
and 3.9 mm, were included in the scope of study on gypsum fibre board
connections with single (GFB-S-35-6 and GFB-S-39-6) and double (GFB-
D-35-6 and GFB-D-39-6) layer of panels.
The average values of the maximum load Fp, conventional elastic
stiffness ke, ductility μ and absorbed energy E are shown in Fig. 15. In
the same figures, the standard deviation of these values is illustrated.
Observation of the connections with single layer of panels revealed
that the effect of screw diameter variation for the strength was linear
(0.64 kN in average for GFB-S-35-6 and 0.71 kN in average for GFB-S-
39-6) with a ratio of strength between 3.9 mm and 3.5 mm equal to
1.11, which results equal to the diameter's ratio 3.9/3.5=1.11. The
effect of screw diameter variation was not very high for stiffness and
ductility, with the stiffness increased by 1.31 times with the use of
higher diameter (1.47 kN/mm in average for GFB-S-35-6 and 1.92 kN/
mm in average for GFB-S-39-6) and the ductility decreased by 1.25
times with the use of higher diameter (12.97 in average for GFB-S-35-6
and 10.37 in average for GFB-S-39-6). On contrary, the difference of
absorbed energy was significant, with reduction of 2.75 with the use of
higher diameter (2.42 kN mm for GFB-S-35-6 and 0.88 kN mm GFB-S-
39-6).
The effect of screw diameter variation for connections with double
layer of panels was small for all parameters, with the strength, stiffness,
ductility and absorbed energy decreased with the use of higher
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

Fig. 14. Average values of Fp, ke, μ and E obtained for RGWB connections with 0.6 mm and 0.8 mm thickness profile. (a) Maximum load. (b) Conventional elastic stiffness. (c) Ductility.
(d) Absorbed energy.

diameter by 1.09, 1.16, 1.15 and 1.15, respectively. Therefore, it seems
that the increasing of the number of panel layers reduced the effect of
screw diameter on the connection response.
4.5. Effect of the number panel layers
In the experimental program were carried out tests on specimens
characterized by one or two layers of panels. Therefore, it may be
interesting to compare the results obtained on test specimens nominally

Fig. 15. Average values of Fp, ke, μ and E obtained for GFB with 3.5 and 3.9 diameter screws. (a) Maximum load. (b) Conventional elastic stiffness. (c) Ductility. (d) Absorbed energy.

245
L. Fiorino et al.
Fig. 16. Average values of Fp, ke, μ and E obtained with specimens made by one or two layer of panels. (a) Maximum load. (b) Conventional elastic stiffness. (c) Ductility. (d) Absorbed
energy.
identical made by one or two panels (GWB-S-35-6 vs GWB-D-35-6, GFB-
S-35-6 vs GFB-D-35-6, GFB-S-39-6 vs GFB-D-39-6). The effect of the
number panel layers is shown in Fig. 16, in which average values of the
maximum load Fp (Fig. 16.a), of the conventional elastic stiffness ke
(Fig. 16.b), of the ductility (Fig. 16.c) and of the absorbed energy
(Fig. 16.d) are shown. In addition, the standard deviation is shown in
the same figure.
For all cases, it was observed that the average values of strength
increased with the increase of the number of panel layers (0.27 kN for
GWB-S-35-6 and 0.47 kN for GWB-D-35-6, with a ratio of strength
between double and single layer equal to 1.74; 0.64 kN for GFB-S-35-6
and 0.89 kN for GFB-D-35-6, with a ratio of strength between double
and single layer equal to 1.39; 0.71 kN for GFB-S-39-6 and 0.82 kN for
GWB-D-39-6, with a ratio of strength between double and single layer
equal to 1.15). Therefore, it can be noticed that the increasing of
strength decreased for more resistant connections.
For the other parameters (ke, μ, E), the comparison between single
and double layer showed enough clear results for connections with
standard gypsum board (GWB-S-35-6 vs GWB-D-35-6) and gypsum fibre
board with 3.9 mm diameter screws (GFB-S-39-6 vs GFB-D-39-6).
Indeed, for standard gypsum board connections, stiffness and ductility
for the specimens with two panels were lower than those obtained for
connections with one panel (ke=1.46 kN/mm in average for GWB-S-35-
6 vs ke=0.71 kN/mm in average for GWB-D-35-6, with a ratio between
single and double layer equal to 2.06; μ=36.43 in average for GWB-S-
35-6 vs μ=22.61 in average for GWB-D-35-6, with a ratio between
single and double layer equal to 1.61), whereas the absorbed energy
increased significantly its average value when the solution with double
246
Thin-Walled Structures 116 (2017) 234–249
layer of panels was used (E=0.73 kN mm for GWB-S-35-6 vs
E=2.34 kN mm for GWB-D-35-6, with a ratio between double and
single layer equal to 3.21).
Also for the case of gypsum fibre board with 3.9 mm diameter
screws (GFB-S-39-6 vs GFB-D-39-6) the stiffness was lower and the
absorbed energy was higher when the solution with double layer of
panels was used (ke=1.92 kN/mm in average for GFB-S-39-6 vs
ke=1.49 kN/mm in average for GFB-D-39-6, with a ratio between
single and double layer equal to 1.29; E=0.88 kN mm in average for
GFB-S-39-6 vs E=1.79 kN mm in average for GFB-D-39-6, with a ratio
between double and single layer equal to 2.03), whereas the results in
terms of ductility showed not very large differences (variations of
average values not more than 20%).
On contrary, the case of gypsum fibre board with 3.5 mm diameter
screws (GFB-S-35-6 and GFB-D-35-6) did not exhibit a clear result of the
effect of the layers number. In fact, the differences in terms of average
values of stiffness, ductility and absorbed energy were in the range from
10% to 18%.
5. Comparison between test results and theoretical predictions
The characteristic shear strength of screwed connections (Fv,Rk) can
be evaluated through relationships for wood panel-to-steel profile
connections given by EN 1995 Part 1-1 [30]. In the case of connection
between timber and thin gauge steel elements in which the failure mode
is due to screw tilting, the code provides the following expression to
evaluate the characteristic shear strength (Fv,Rk):
Fv, Rk =0. 4fh, k t1d (4)
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

Table 5
Shear strength of tested screwed connection.

Label fh,av Fp,av,exp Fp,av,th Fp,av,exp / Fp,av,th Fd Fp,av,exp / Fd

[N/mm2] [kN] [kN] [-] [kN] [-]

GWB-S-35-6 11.82 0.27 0.21 1.31 0.10 2.67

RGWB-S-39-6 22.89 0.42 0.45 0.94 0.22 1.94
GFB-S-39-6 34.56 0.71 0.67 1.05 0.33 2.15
RGWB-S-39-8 22.89 0.44 0.45 0.99 0.22 2.03
GFB-S-35-6 38.23 0.64 0.67 0.96 0.33 1.97

where:
fh,k is the characteristic embedment strength of the timber element;
t1 is the smaller of the thickness of the timber side member or the
penetration depth; Fig. 17. Experimental vs. theoretical prediction.
d is the fastener diameter.
Since the embedment strength (fh,k) does not represent a require-
The comparison between experimental (Fp,exp = rei) and theoretical
ment for relevant fabrication standards [26–28], only average values of
values (Fp,av,th = rti) is shown in Fig. 17.
this parameter are available for standard gypsum (GWB), impact
The mean value correction factor b and the error terms δi can be
resistant gypsum (RGWB) and gypsum fibre (GFB) board panels. For
evaluated as follows:
this reason, Eq. (4) can be used for the evaluation of the average
maximum load (Fp,av): ∑i re, irt , i
b= = 1.0072
∑i rt2, i (7)
Fp, av = 0.4fh, av t1d (5)
rei
in which t1 and d were set equal to the panel thickness and nominal δi =
brti (8)
screw diameter, respectively.
In particular, the values of average embedment strengths were Since the value of b was equal to 1.0072 and all points (rti, rei) are
assumed on the basis of available experimental results (Table 5). No near the bisector, the resistance function can be considered exact.
data for the embedment strength were available for outdoor cement On the basis of the value of b and δi the coefficient of variation of the
board (CP) panels. As a consequence, it was not possible to predict the error terms δi (Vδ) can be calculated with following relationship:
shear strength for this connection typology. The results of the theore-
tical prediction are illustrated in Table 5. The theoretical prediction Vδ = exp(s∆2 )−1 = 0.174 (9)
provides good results for connections with impact resistant gypsum where:
(RGWB) and gypsum fibre (GFB) board panels. In particular, the
n n
theoretical approach slightly overestimates the experimental results 1 1
s∆2 = ∑ (∆i − ∆)2 = 0.03;∆i = ln(δi );∆ = ∑ ∆i .
for connections made of impact resistant gypsum (RGWB) and gypsum n −1 i =1
n i =1
fibre board with 3.5 mm screw diameter (GFB-S-35-6). Instead, for
connections made of standard gypsum board (GWB) and gypsum fibre Since the resistance function is a product function of the variables,
board with 3.9 mm screw diameter (GFB-S-39-6) the theoretical pre- the mean value E(r) may be obtained by:
diction underestimates the experimental shear strength. E (r ) = bgrt ( Xm ) (10)
In order to evaluate the design value of the shear strength, the
statistical procedure given in EN 1990 Annex D [29] was used. The Finally, the design value for a limited number of tests (n < 100) is
following assumptions are made: given by the following relationship:
⎡
• the resistance is a function of a number of independent variables X; Fd = b 0.4 fh, av t1dexp⎢ −kd ,∞
ln(Vrt2+1)
− kd , n
ln(Vδ2+1)
• a sufficient number of test results is available; ⎢
⎣ ln(Vr2+1) ln(Vr2+1)
• all relevant geometrical and material properties are measured; −0. 5ln(Vr2+1)]
• there is no correlation (statistical dependence) between the vari- (11)
ables in the resistance function; where:
• all variables follow either a normal or a log-normal distribution. kd,∞ is the value of kd , n for n = ∞ equal to 3.04;
kd , n is the characteristic fractile factor, equal to 3.44 for test number
The standard procedure is structured in seven steps: (1) develop a n = 6;
design model, (2) compare experimental and theoretical values, (3) j
estimate the mean value correction factor b, (4) estimate the coefficient Vrt2 = ∑ Vxi2 = V f2 + Vt2 + Vd2 = 0.02;
of variation of the errors, (5) analyse compatibility, (6) determine the i =1
coefficients of variation VXi of the basic variables and (7) determine the
characteristic value rk or the design value rd of the resistance. Vr2 = Vδ2 + Vrt2 = 0.05;
The first step consists in developing a design model for a theoretical
in which Vxi are the coefficients of variations of the basic variables, that
resistance rt represented by a resistance function:
in the examined case are assumed as follow:
rt = grt ( X ) (6) Vd, the coefficient of variation for screws diameter, assumed equal
to 0.05;
For the resistance model it was assumed Eq. (5), in which the Vt, the coefficient of variation for panels thickness, assumed equal to
variables embedment strength (fh,av), panel thickness (t) and screw 0.05;
diameter (d) were assumed statistically independent. Vfh, the coefficient of variation for embedment strength, assumed

247
L. Fiorino et al. Thin-Walled Structures 116 (2017) 234–249

equal to 0.10.
Design values of shear strength are shown in Table 5 together the
ratios between experimental and design results. It can be possible to
observe that the experimental results are about 2 times the design
strength values obtained by the procedure given in the EN 1990 for the
solutions with impact resistant gypsum board (RGWB) and gypsum
fibre board (GFB), whereas for the connections with standard gypsum
board (GWB) the ratio between experimental and design strength is
higher (2.67).
Peterman et al. [9]), the results obtained in the presented experimental
campaign are lower than available experimental data (0.27 vs. 0.53 kN
or 0.27 vs. 0.51 kN considering Swensen et al. [7] or Peterman et al. [9]
results).
7. Conclusions
The main conclusions obtained from the study object of this
framework are the following:

1. For the adopted edge loading distance, equal to 15 mm, the

6. Comparison between test results and available experimental
data
In literature there are not test results on screwed connections with
thickness profiles equal to 0.6 mm, whereas several tests were carried
out on connections involving gypsum panels and steel thickness of
about 0.8 mm (Serrette et al. [3], Swensen et al. [7] and Peterman et al.
[9]). These test results were used for a comparison with the experi-
mental results presented in the current paper, considering that the
variation of steel thickness from 0.6 to 0.8 mm does not influence
significantly the behaviour of screwed connections, especially for the
strength, that mainly depends on the panel thickness, panel type, and
screw diameter, as discussed in the previous sections. Also the edge
loading distance plays a significant role on the connection shear
resistance. In particular, the shear response is significantly affected by
edge failure mode for small edge distance (usually less than about
20 mm), whereas it is affected by bearing failure for higher values of
the edge distance, which corresponds to higher load capacity. In the
presented experimental campaign, an edge distance equal to 15 mm
was used, which represents the minimum edge distance used in
common practice for connecting CFS profiles and sheathing panels.
As a result, the value of the edge distance equal to 15 mm was selected
in order to obtain experimental results on the safe side (underestima-
tion of mechanical properties). The results of the comparison are shown
in Table 6.
From the comparison, it can be noticed that when connections with
same screw diameter, panel type, panel thickness and edge distance are
considered (presented experimental campaign vs. Serrette et al. [3]),
the results obtained in the presented experimental campaign are higher
than available experimental data (0.27 vs. 0.18 kN for standard gypsum
board and 0.64 vs. 0.26 kN for gypsum fibre board). This finding is
supported also by the different failure modes occurred in the compared
experimental campaigns, i.e. tilting of the screws and breaking of the
sheathing edge in the presented experimental campaign vs. breaking of
the sheathing edge in Serrette et al. [3]. The different failure modes and
strength values could be justified by different mechanical properties in
terms of embedment strength of the panels tested in the compared
experimental tests, but this possible justification cannot be confirmed
because the value of the embedment strength is not given in Serrette
et al. [3]. On the other hand, when standard gypsum board connections
with higher screw diameter, edge distance and panel thickness are
compared (presented experimental campaign vs. Swensen et al. [7] and
observed failure mechanism for all tests was the tilting of the screws
and breaking of the panel edge.
2. The parameter relevant to the description of the initial shear
response, i.e. the stiffness, is the most scattered value (C.o.V. from
0.12 to 0.90), the maximum strength is the lowest scattered
parameter (C.o.V. from 0.03 to 0.19), and the displacement para-
meters related to the ultimate behaviour show an intermediate
dispersion (C.o.V. from 0.16 to 0.45). This difference can be mainly
attributed to the higher sensibility of the first part of the response to
the imperfections due to the mounting process.
3. For the solutions with single layer of panels, the connection with
gypsum fibre board was the most resistant, whereas the connection
with standard gypsum board was the least resistant, and the
remaining connections (impact resistant gypsum board and outdoor
cement board) showed similar values of strength. The solutions with
impact resistant gypsum board and gypsum fibre board were the
stiffest, whereas the solutions with outdoor cement board and
standard gypsum board had the lowest values of stiffness. For the
ductility, the lowest value was obtained for the connection with
gypsum fibre board and the highest value was observed for the
connection with standard gypsum board, whereas for the other
solutions were obtained intermediate values. The comparison in
terms of absorbed energy showed no significant quantitative
differences among different connection typologies.
4. In case of double layer of panels the shear strength can be doubled
for solutions with standard gypsum boards, whereas the shear
strength can be increased of 1.4 times for connections with gypsum
fibre board.
5. The increment of the steel profile thickness did not produce
apparent increasing of strength for the investigated case (variation
of steel thickness from 0.6 to 0.8 mm for connections with impact
resistant gypsum board).
6. The increment of the screw diameter did not produce apparent
increasing of strength for the investigated case (variation of screw
diameter from 3.5 to 3.9 mm for connections with single layer of
gypsum fibre boards).
7. For 3.9 mm diameter screwed gypsum fibre board (GFB) connec-
tions the shear strength obtained according to the theoretical
prediction given by EN 1995 Part 1-1 [30] is very similar to the
experimental value, whereas the prediction slightly overestimates
the experimental results for impact resistant gypsum (RGWB)

Table 6
Test results obtained in the presented research vs. available experimental data.

Author Sheathing material Steel thickness Screw Edge distance Fp,av Failure mode
[mm] [mm] [mm] [mm] [kN]

Current research 12.5 - gypsum 0.60 3.5 15 0.27 T, BE

12.5 - gypsum fibre 0.60 3.9 15 0.71 T, BE
12.5 - gypsum fibre 0.60 3.5 15 0.64 T, BE
Serrette et al. [3] 12.7 - gypsum 0.88 3.5 15 0.18 BE
12.7 - gypsum fibre 0.88 3.5 15 0.26 BE
Swensen et al. [7] 15.9 - gypsum 0.88 4.2 89 0.53 C
Peterman et al. [9] 12.7 - gypsum 0.84 3.5 38 0.51 B, PT
BE: Breaking of the sheathing Edge; C: Crushing of panel; B: fastener bearing against sheathing; PT = fastener pull-through; T = tilting of the screws.

248
L. Fiorino et al.
connections and for 3.5 mm diameter screwed gypsum fibre board
connections. On contrary, for connections with standard gypsum
board panels (GWB) the theoretical prediction underestimates the
experimental shear strength.
8. On the basis of obtained results, the suggested design values of the
edge failure shear strength per single screw with an edge loading
distance of 15 mm obtained according to the statistic procedure
given by EN 1990 [29] are the following:
a. For a single 12.5 mm thick standard gypsum board fastened to a
0.6 mm thick DX51D+Z steel profile with a 3.5 mm diameter screw:
0.10 kN.
b. For a single 12.5 mm thick gypsum fibre board fastened to a 0.6 mm
thick DX51D+Z steel profile with a 3.5 or 3.9 mm diameter screw:
0.33 kN.
c. For a single 12.5 mm thick impact resistant gypsum board fastened
to a 0.6 or 0.8 mm thick DX51D+Z steel profile with a 3.9 mm
diameter screw: 0.22 kN.
Acknowledgments
The study presented in this paper is a part of a wider research
programme coordinated by Prof. Raffaele Landolfo with the technical
and financial support of Knauf Gips KG (Research study agreement
between Knauf Gips KG and the Department of Structures for
Engineering and Architecture of the University of Naples Federico II -
September, 3rd 2013).
Knauf Gips KG is gratefully acknowledged by the authors for the
Consent to Publish the scientific results presented in this paper in the
scope of the above mentioned common research.
References
[1] L. Fiorino, O. Iuorio, R. Landolfo, Sheathed cold-formed steel housing: a seismic
design procedure, Thin-Walled Struct. 47 (2009) 919–930.
[2] B. Källsner, U.A. Girhammar, Plastic models for analysis of fully anchored light-
frame timber shear walls, Eng. Struct. 31 (2009) 2171–2181.
[3] R.L. Serrette, J. Encalada, M. Juadines, H. Nguyen, Static racking behaviour of
plywood, OSB, gypsum, and fiberbond walls with metal framing, J. Struct. Eng. 123
(8) (1997) 1079–1086.
[4] L. Fiorino, G. Della Corte, R. Landolfo, Experimental tests on typical screw
connections for cold-formed steel housing, Eng. Struct. 29 (2007) 1761–1773.
[5] L. Fiorino, O. Iuorio, R. Landolfo, Experimental response of connections between
cold-formed steel profile and cement-based panel. In: Nineteenth International
Specialty Conference on Cold-Formed Steel Structures St. Louis, Missouri, U.S.A,
October 14 & 15: 639-653, 2008.
[6] T.H. Miller, T. Pekoz, Behavior of gypsum-sheathed cold-formed steel wall studs, J.
Struct. Eng. 120 (5) (1994) 1644–1650.
[7] S. Swensen, G.G. Deierlein, E. Miranda, Behavior of screw and adhesive connections
to gypsum wallboard in wood and cold-formed steel-framed wallettes, J. Struct.
249
Thin-Walled Structures 116 (2017) 234–249
Eng. E4015002 (2015).
[8] L. Vieira, B.W. Schafer, Experimental Results for Translational Stiffness of Stud-
Sheathing Assemblies. A supplemental report for AISI-COFS Project on Sheathing
Braced Design of Wall Studs, 2009.
[9] K.D. Peterman, N. Nakata, B.W. Schafer, Hysteretic characterization of cold-formed
steel stud-to-sheathing connections, J. Constr. Steel Res. 101 (2014) 254–264.
[10] J. Lange, B. Naujoks, Behaviour of cold-formed steel shear walls under horizontal
and vertical loads, Thin-Walled Struct. 44 (2006) 1214–1222.
[11] J. Ye, X. Wang, H. Jia, M. Zhao, Cyclic performance of cold-formed steel shear walls
sheathed with double-layer wallboards on both sides, Thin-Walled Struct. 92 (2015)
146–159.
[12] J. Ye, X. Wang, M. Zhao, Experimental study on shear behavior of screw
connections in CFS sheathing, J. Constr. Steel Res. 121 (2016) 1–12.
[13] W. Chen, J. Ye, T. Chen, Design of cold-formed steel screw connections with
gypsum sheathing at ambient and elevated temperatures, MDPI, Appl. Sci. 6 (2016)
248, http://dx.doi.org/10.3390/app6090248.
[14] L.A. Fülöp, D. Dubina, Design criteria for seam and sheeting to framing connections
of cold-formed steel shear panels, J. Struct. Eng. 132 (4) (2006) 582–590.
[15] M. Nithyadharan, V. Kalyanaraman, Experimental study of screw connections in
CFS-calcium silicate board wall panels, Thin-Walled Struct. 49 (2011) 724–731.
[16] A.F. Okasha, Performance of Steel Frame/wood Sheathing Screw Connections
Subjected to Monotonic and Cyclic Loading (M.Sc. thesis), Department of Civil
Engineering and Applied Mechanics, McGill University, Montreal (Canada), 2004.
[17] L. Fiorino, O. Iuorio, R. Landolfo, Seismic analysis of sheathing-braced cold-formed
steel structures, Eng. Struct. 34 (2012) 538–547, http://dx.doi.org/10.1016/j.
engstruct.2011.09.002.
[18] L. Fiorino, O. Iuorio, V. Macillo, R. Landolfo, Performance-based design of sheathed
CFS buildings in seismic area, Thin-Walled Struct. 61 (2012) 248–257, http://dx.
doi.org/10.1016/j.tws.2012.03.022.
[19] L. Fiorino, O. Iuorio, R. Landolfo, Designing CFS structures: the new school bfs in
Naples, Thin-Walled Struct. 78 (2014) 37–47, http://dx.doi.org/10.1016/j.tws.
2013.12.008.
[20] V. Macillo, O. Iuorio, M.T. Terracciano, L. Fiorino, R. Landolfo, Seismic response of
CFS strap-braced stud walls: theoretical study, Thin-Walled Struct. 85 (2014)
301–312, http://dx.doi.org/10.1016/j.tws.2014.09.006.
[21] O. Iuorio, L. Fiorino, R. Landolfo, Testing CFS structures: the new school BFS in
Naples, Thin-Walled Struct. 84 (2014) 275–288, http://dx.doi.org/10.1016/j.tws.
2014.06.006.
[22] O. Iuorio, V. Macillo, M.T. Terracciano, T. Pali, L. Fiorino, R. Landolfo, Seismic
response of Cfs strap-braced stud walls: experimental investigation, Thin-Walled
Struct. 85 (2014) 466–480, http://dx.doi.org/10.1016/j.tws.2014.09.008.
[23] UNI EN ISO 6892-1: 2009. Metallic materials - Tensile testing - Part 1: Method of
test at room temperature. European committee for standardization.
[24] CEN, EN 1993-1-3. Eurocode 3: Design of steel structures—Part 1–3: General rules
— Supplementary rules for cold formed members and sheeting. European
Committee for Standardization, 2006.
[25] CEN, EN 10327: Continuously hot-dip coated strip and sheet of low carbon steels for
cold forming - Technical delivery conditions. European Committee for
Standardization, 2004.
[26] CEN, UNI EN 520, Gypsum plasterboards - Definitions, requirements and test
methods. European Committee for Standardization, 2009.
[27] CEN, EN12467: 2012+ A1: 2016. Fibre-cement flat sheets. Product specification
and test methods.
[28] CEN, EN 15283-2: 2008+ A1: 2009. Gypsum boards with fibrous reinforcement.
Definitions, requirements and test methods. Gypsum fibre boards.
[29] CEN, UNI EN 1990, Eurocode - Basis of structural design. European Committee for
Standardization, 2006.
[30] CEN, EN 1995-1-1. Eurocode 5: Design of timber structures - Part 1-1: General
-Common rules and rules for buildings, 2008.