Engineering Structures 232 (2021) 111791

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Evaluation of composite action in cross laminated timber-concrete

composite beams with CFRP reinforcing bar and plate connectors using
Digital Image Correlation (DIC)
S.A. Hadigheh a, *, R. McDougall a, b, C. Wiseman a, c, L. Reid a, d
a
School of Civil Engineering, Faculty of Engineering, The University of Sydney, Sydney, New South Wales 2006, Australia
b
Arup, Sydney, New South Wales 2000, Australia
c
Northrop Consulting Engineers, Sydney, New South Wales 2000, Australia
d
GHD, Sydney, New South Wales 2000, Australia

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

Keywords: This research proposes new types of shear connectors made of carbon fibre reinforced polymer (CFRP) com­
Timber-concrete composite (TCC) posites for effective stress transfer between the timber and concrete sections in cross laminated timber (CLT)-
Cross laminated timber (CLT) concrete composite beams. New shear connectors are designed and made of bidirectional carbon fibre reinforced
Fibre reinforced polymer (FRP)
polymer (CFRP) composite plates and crossed CFRP reinforcing bars. The mechanical performance, bending
Composite action
Bending stiffness
stiffness, ductility, and interfacial slippage and strain of timber-concrete composite (TCC) beams with CFRP
Digital Image Correlation (DIC) connectors are compared with those with steel plate and screw systems through four-point flexure testing. Local
Interfacial slip slip, interfacial slip and strain behaviour are comparatively analysed for connections with equivalent axial
Cost analysis stiffness using a Digital Image Correlation (DIC) based technique, so that relative composite behaviour could be
determined. Furthermore, a cost evaluation is undertaken to compare the feasibility of proposed shear con­
nectors for construction of TCC systems. Results from flexural tests demonstrate that CFRP rod specimens
experience higher ultimate load and bending stiffness in elastic loading stage and ductility in failure while
slippage at serviceability and ultimate load is minimal. These results demonstrate that CFRP reinforcing bars can
be used as an alternative to existing steel plate/screw systems. Although CFRP plate connectors show lower
ultimate strength, bending stiffness and ductility, the performance of the system can be further improved by
using sufficient anchorage systems at the end of CFRP plate within the concrete.

1. Introduction concrete composite (TCC) systems, the timber element is connected to

an overlaying concrete member primarily to promote tension within the
Over the past two decades, engineered timber products have expe­ timber and compression within the concrete.
rienced accelerated growth within construction industry due to a wide TCC beams offer the advantage of improved span length, increased
variety of desired characteristics. With a high strength to weight ratio, acoustic performance and higher performance during fire when
foundation loads are reduced, which contribute to less construction time compared with regular engineered timber products. The additional
and effort. Likewise, methods of prefabrication allow timber to be stiffness provided by concrete allows for better deflection and vibration
competitive in terms of hoisting and transportation loads [1]. With an control. This is achieved by use of a connection at the interface of timber
organic cellular structure, timber is inclined to higher strength parallel and concrete to allow composite action to occur between materials [2].
to grain, similar to that of reinforced concrete. However, due to its lower The degree of composite action ultimately dictates the structural per­
stiffness, timber is generally governed by serviceability criteria such as formance of TCC beams, which depends on the effectiveness of the shear
vibration and deflection in structural design. As a result, combining connector. A stiff connection is required at the interface of materials to
timber and concrete as a composite structural system can provide an limit slip and transmit the shear loads between timber and concrete
alternative solution to satisfy serviceability requirements. For timber- sections [3]. Similarly, sufficient load capacity is required by

* Corresponding author.
E-mail address: ali.hadigheh@sydney.edu.au (S.A. Hadigheh).

https://doi.org/10.1016/j.engstruct.2020.111791
Received 30 March 2020; Received in revised form 7 October 2020; Accepted 21 December 2020
Available online 23 January 2021
0141-0296/© 2021 Elsevier Ltd. All rights reserved.
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

connections to withstand the interlayer shear forces and maximise capacity.

spacing between connectors. Low strength may be overcome by Carbon fibre reinforced polymer (CFRP) composites have been uti­
reducing the spacing between connectors. However, this leads to lised for a range of civil engineering purposes from construction of new
increased costs and an effective bending stiffness much higher than structures to strengthening of existing building structures [17,18].
required [4]. Characterised by high strength in tension and shear, high stiffness, high
Full composite action between two materials is achieved when there resistance to environmental degradation and automated manufacturing
is no slip at the material interface, resulting in a shared neutral axis. This techniques for production of various shapes and profiles [19,20,27,28],
allows a continuous strain distribution via perfect transfer of strain be­ FRP composites can be used as a suitable alternative for more efficient
tween the materials. As shear connectors resist the slip at the material connections in TCC systems. In a preliminary research, Hadigheh and
interface and transfer the strain between layers, the connector controls Dias-da-Costa [2] showed that CFRP sheets can provide further strength
the degree of composite action. A consequence of complete composite and stiffness to the TCC system. The load-displacement results during
action, however, is that the mechanism of failure of a TCC is due to any the tests showed that the specimens with CFRP inserted inside the
constituent component of the beam undergoing localised failure. This concrete notch could carry high shear stresses due to the combined
means that beam as a whole can experience a brittle failure in the timber contribution of concrete in shear and CFRP in tension. The proposed
or concrete section, or undergo connection failure [5]. The desired CFRP connection was also able to increase ductility and prolong the post
mechanism of failure is by ductile means as there is obvious deformation yielding stage before failure. However, fabrication of proposed FRP
in the lead up to failure, allowing for adequate mitigation steps or safe connector was time consuming and not efficient in practical
decommissioning. As a result, it is favourable to control the mechanism applications.
of failure and localise it through a ductile connector, as in partially In this research, new CFRP composite connectors are proposed as
composite beams with ductile connectors. Ideally this would necessitate alternatives to existing connections in TCC systems. This is carried out
a beam that is highly composite, though not wholly, so that the failure by comparative analyses with current technologies in terms of con­
mechanism will be through the connection as composite action breaks structability and mechanical strength criteria. The aim is to directly
down [6]. Hence, a ductile failure of the connection leads to a ductile compare steel plate to CFRP plates and steel screws to CFRP reinforcing
failure of the entire system. Therefore, an ideal shear connector in a TCC bars. All connectors are experimentally investigated by flexural testing
beam is one that has a high stiffness when loaded in the elastic stage, and and a comprehensive evaluation in terms of mechanical performance,
ductile when loaded in the plastic range [7]. However, serviceability the failure mechanism, bending stiffness and ductility, interfacial slip,
design criteria including stiffness, vibration and long-term deflections strain and cost analyses on each connection system is carried out. In
generally govern the design of TCC floors. This is inherently due to addition, composite actions in each beam is investigated through
TCC’s comparatively low stiffness when compared to conventional application of the digital image correlation (DIC).
methods of construction. Because of this, the ultimate strength therefore
does not always cater as a governing design factor. 2. Experimental program
Research has contributed to the development of a range of connec­
tions, including perforated steel plate (PSP) [3,6,8], inclined screws The two common connections currently used in construction are the
[1,4,5,9], notched sections in the timber [2,9,10], nail plate connectors cross (45◦ ) screws and the steel plate connectors. In this research, the
[8,11,12], and a continuous adhesive layer [13]. These methods exert cross-screw connection is replicated by placing CFRP rebars within the
varying degrees of efficiency in terms of stiffness, strength and ductility. connection with similar configuration and a comparable axial stiffness
Perforated steel plates have been successful in achieving composite ac­ (EA) to the steel screws. In addition, steel plates are compared directly to
tion, albeit at considerable time and monetary expense due to the bi-directional CFRP plates.
extensive gluing process. A recent study [3] evaluated stiffness and
strength of PSP systems when tested both independently and in com­
bination with reinforcing bars. This system offered to solve the out of 2.1. Detail of specimens
plane buckling phenomena previously identified by Clouston et al. [14]
by utilising two parallel plates to promote better flexural rigidity. The The constructability of the CFRP connections was a major factor
proposed reinforcement arrangements did not provide significant im­ considered in the overall design to prevent tedious and costly
provements to shear strength, as the improved lateral stability meant manufacturing of the connections. The installation of the CFRP rein­
that failure was now conditioned by the shear stresses in the wood/ forcement and plates is required to be in a comparable time and cost
adhesive interface [15]. The use of steel screws inclined at 45◦ in the range to the steel connection systems. Table 1 summarises the type of
plane of the beam has the benefit of easy installation, and both the connection that is used in each TCC beam.
withdrawal strength and tensile strength of the screws contributing to C24 grade C160L cross laminated timber (CLT) beams made of Eu­
shear strength. In spite of this, the limited strength of screws requires ropean Spruce with a moisture content of 12–16% and cross section
multiple connectors to achieve desired connection stiffness and limit dimensions of 100 × 160 mm2 were used for TCC beams. The predom­
slip. Without inclination of the screws to contribute tensile strength to inant orientation of grain in the timber laminates was in the plane of the
the shear plane, vertical screws offer a relatively low slip modulus [4], beam in order to maximise flexural performance, as shown in Fig. 1a.
and a resulting lack of composite action. Marchi et al. [5] further The elastic modulus of CLT sections parallel to grain was 12.5 GPa.
examined the introduction of a second steel screw inclined to create a Table 2 provides a summary of the material properties of CLT timber
crossed screw formation, noting that whilst the second screw adds shear specimens.
strength in the other direction, it introduces a new failure mechanism to In TCC design, the overlying concrete element serves to improve
TCC beams with conical expulsion of concrete with insufficient cover. It
appears that most of existing connections are not entirely able to satisfy Table 1
all these criteria simultaneously. Steel screws and epoxy were used as Material properties.
connectors for timber beams reinforced with U-shaped steel sections on Beam Type of connection Codes
the upper and lower flange of the timber component. It was found [16] CLT No connector TB1 & 2
that beams connected with epoxy reached a significantly higher capacity CTC steel screw cross connection TCC CTC Steel screw SS1 & 2
than steel screw connections, albeit at high cost and time delay [3,16]. CFRP reinforcement cross connection TCC CFRP reinforcing bar CR1 & 2
Cracks were prematurely developed on the tension side of the beam due Steel plate connection TCC Steel plate SP1 & 2
CFRP plate connection TCC Bi-directional CFRP plate CP1 & 2
to the presence of screws, thereby limiting the members ultimate

2
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 1. (a) Dimension of CLT and lamina (L: grain parallel to beam length), (b) detail of connectors for timber-concrete (CTC) screws in TCC beams.

overall bending stiffness which allows for a higher ultimate strength to threaded shaft at 4 mm diameter to prevent splitting of the timber
be achieved when compared to the design of CLT beams of comparable (Fig. 2a). The 45◦ angle of the screws was achieved by using a set square
size. In design of concrete section, it must be ensured that the depth of and holding the drill to the correct angle during the predrilling, with the
concrete provides sufficient shear and bending capacity to prevent local screws then simply following the guide hole when being drilled in. When
failure, as well as effective connection cover to prevent concrete the angle of the screws was measured the angle was within ±5◦ meaning
crushing. High elastic modulus of the concrete enables for a neutral axis that the accuracy was quite high for the simplistic method used.
to coincide with the interface of materials with a relatively shallow Perforated steel plates are highly efficient shear connectors due to
depth. It is therefore essential to optimise both material elasticity and their high strength yet ductile response. With perforations into both the
effective depth so that the neutral axis remains at the interface of ma­ timber and concrete regions of the connection, the bonding between
terials for a desired slip modulus in connections. Consequently, the mix materials can be enhanced which thereby limits the potential of
design required a 14-day curing strength of approximately 16 MPa for debonding failure. The use of plates however, does not replicate the
the concrete with maximum aggregate size of 7 mm nominal diameter same ease of onsite installation capabilities offered by screws, and are
(Table 2). To determine the compressive strength of the concrete, 4 therefore more suited to prefabricated construction methods. The LBV
concrete cylinders were cast at time of batching and tested at the day of perforated plates were cut to 80 mm × 80 mm panels and were installed
flexural test. into a continuous 5 mm slot in the CLT beam using an epoxy resin (see
To avoid water seepage and minimise the effects of friction at the Fig. 3 and Table 2).
timber-concrete interface, a 0.2 mm black builder’s plastic was lain CFRP rebars were used to connect timber and concrete in CR speci­
under all concrete poured. Holes for the connections were cut in the mens. 8 mm CFRP rebars were selected due to their comparable axial
plastic before placing it, the plastic was then taped down to the beam to stiffness to CTC screws. The rebars were made from a pultrusion
hold it in place while placing forms. Comparatively, simple form boards moulding process which laid fibres unidirectionally along the length of
were constructed around the CLT beams using screws, extending 75 mm the reinforcement. Properties of the rebars are presented in Table 2. The
above the timber for which concrete was to be placed inside. ridges which replicate deformed bar (N) were provided by a stitching
7 mm Rothoblaas connectors for timber-concrete (CTC) screws were stand that pulled the fibres together at intervals along the reinforcement
chosen as connectors in SS specimens. CTC connectors are made up of as in Fig. 2b. The ribbing of the bar allowed for great adhesion to both
three distinct regions along its shaft. The bottom threads are designed to concrete and the epoxy for timber embedment. The 160 mm long CFRP
be embedded within timber and top threads in concrete. The specific rebars were epoxied into 12 mm holes at 45◦ .
length of bottom thread ensures that nominal slip modulus is achieved in The bidirectional CFRP plates with 80 mm × 80 mm dimensions
construction by minimum embedment. These two regions are separated were used in CP specimens (Fig. 3b). The CFRP plates were designed to
by a localised thickening of the shaft which contains no thread, in which be a direct comparison to the steel plates and for this reason they were
is idealised to be at the shear interface of materials. At the advice of the the same dimensions as the steel plate. To make this comparison, both
manufacturer’s design information, the specific ratio of timber to con­ CFRP plate and steel plate specimens had similar embedment length in
crete (b1/b2 in Fig. 1) was determined. These values provided the timber and concrete. This was chosen so that the neutral axis of the
greatest efficiency of shear development for CTC screws, and as a means connector coincided with the concrete-timber interface, as is typically
of direct comparison to CFRP rebar was kept constant between these considered in design of steel plate TCC systems [14]. The bidirectional
samples. The dimensions and properties of the screws can be seen in fibre orientation in the plates was aligned from corner to corner (45◦ to
Fig. 1b and Table 2. The screws were installed according to the manu­ the length of the beam) to allow for transferring both shear and tension
facturer’s instructions, with predrilling the full length of the bottom in the connection. Table 2 provides more information about the

3
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Table 2 moisture content less than 18%. The material properties of XEPOX F
Properties of materials used in the experiments. epoxy adhesive are shown in Table 2. High strength and viscosity of
Class C24 timber C160L CLT epoxy adhesive was essential in the experimental design to ensure that
failure occurred in the connections and not in the bonding between
Strength properties Stiffness properties
elements.
Bending, fm,k (MPa) 24.0 Mean modulus of elasticity 11.0
parallel, E0,mean (GPa)
Tension parallel, ft,0,k 14.0 5%-percentile modulus of 7.4 2.2. Design of TCC beams
(MPa) elasticity parallel, E0.05 (GPa)
Tension perpendicular, 0.4 Mean modulus of elasticity 0.37
The design of the TCC beams was performed using the Gamma
ft,90,k (MPa) perpendicular, E90,mean (GPa)
Compression parallel, fc,0,k 21.0 Mean shear modulus, Gmean 0.69 Method procedure outlined in Eurocode 5 - Annex B [21]. This design
(MPa) (GPa) method assumes that the TCC consists of two separate timber and con­
Compression 5.3 Density 5%-percentile, pk (kg/ 350.0 crete beams subjected to Euler-Bernoulli bending, attached by a flexible
perpendicular, fc,90,k m3) connection. This method relies on the use of an effective bending stiff­
(MPa)
Shear, fv,k (MPa) 2.5 Density mean value, pmean (kg/ 420.0
ness derived from the properties of timber and concrete, and the Gamma
m3) Method approximate solution to composite action beams (Fig. 4):
Concrete ∑
2
Fine aggregate (%) 41 Crystalline silica (%) >1 (EI)eff = (Ei Ii + γ i Ei Ai a2i ) (1)
Coarse aggregate (%) 33 Water (%) 11 i=1
Portland cement (%) 14 Compressive strength, f’cm 16
(MPa) for i layers of the timber and the concrete, elastic modulus (E), moment
Connector for Timber-Concrete (CTC) Screws of inertia (I), cross-sectional area (A), and distance between the centroid
Nominal diameter, d1 7 Characteristic yield moment 20 of a member and the effective neutral axis of the composite section (a).
(mm) My,k (kN.mm) The value for gamma ranges from 0 (no composite action) to 1 (full
Tip diameter, d2 (mm) 4.6 Characteristic withdrawal- 11.3 composite action) are calculated by:
resistance parameter fax,k
(MPa) 1
Head diameter, dk (mm) 9.5 Characteristic tensile strength 20.0 γ= 2 (2)
1 + πK ′EA
ftens,k (MPa) L2
Screw length, L (mm) 160 Withdrawal-resistance 10
characteristic - concrete at 45◦ for slip modulus K′ and beam span L.
orientation (kN) Using the nominal CTC slip modulus provided by the manufacturer
Timber embedment length, 110 Min. distance from timber edge 35 and nominal material properties outlined in Table 2, the TCC was
b2 (mm) (mm)
Threaded section length 40 Coefficient of friction 0.25
designed specifically so that the beam neutral axis coincided with the
embedded in concrete, material interface. With this specific design, the main shear plane is
b1 (mm) resisted by the trial shear connectors. The design then was checked for
Shank diameter, ds (mm) 5.0 Elongation (%) 0.2 bending, axial and shear stress capacities, and serviceability deflection,
LVB Steel Plates both in timber and concrete layers and the beam as a whole. With the
Nominal thickness (mm) 1.5 Max. net section holes 4 selection of dimensions for the connectors there was care to make the
Square side dimension 80 Nominal hole diameter (mm) 5 steel screw/CFRP rebar and steel plate/CFRP plate pairings have a
(mm)
Tensile strength (kN) 26.7 Hole spacing distance (mm) 20
constant axial stiffness (EA). This was done using the nominal material
elastic moduli outlined in Table 2, diameter of shafts for screws and
CFRP Rebar
rods, and the cross-sectional area of plates accounting for perforations.
Nominal diameter (mm) 8 Tensile failure stress (MPa) 2378
Rebar length (mm) 160 Elastic modulus (GPa) 143
Fig. 5 illustrates configuration of tested beams.
Timber embedment length 110 Elongation (%) 1.42 It is very important to minimise any unwanted eccentricities in
(mm) testing that may arise from material deformation or non-centred shear
Min. distance from timber 35 Density (gr/mm3) 1.6–1.8 connectors as these can introduce additional stresses from beam torsion
edge (mm) × 10-3
and reflect a lower flexural capacity. This is further a consequence of
CFRP Plate poor TCC design where shear acts through the timber layer and not at
Nominal thickness (mm) 1.5 Elastic modulus (GPa) 58.44 the timber-concrete interface, resulting in timber panel debonding and
Square side dimension 80 Ultimate Tensile Strength 686.9
(mm) (MPa)
therefore out-of-plane movement. This was carefully considered during
Resin content (%) 30 the design and construction of TCC beams.
Epoxy Adhesive
Strength of adhesive bond 25 Flexural strength (MPa) 38 2.3. Test set-up
(MPa)
Tensile strength (MPa) 30 Elastic modulus in flexure 2600
(MPa)
The TCC beams were tested in flexure to gain an understanding of
Elastic modulus in tension 4600 Thermal expansion coefficient 6.0 × 10 5 instantaneous beam capacity and the mechanical properties of connec­
(MPa) (m/m.℃) tors. Flexure testing were carried out via four-point loading using a
Elastic modulus in 3937 Total shrinkage (%) 0.18 Dartec hydraulic loading machine. This method allowed for a constant
compression (MPa)
bending moment between loading points, and a linearly decreasing
Shear strength (MPa) 36
bending moment to occur from the loading point to the supports on each
end. Shear force was zero between loading points and was constant
properties of CFRP plates. between loading and supports. As a result, three connections between
The embedment of the CFRP rebar, LBV plate and CFRP plate into loading points experienced an equivalent bending moment with zero
timber was achieved by using Xepox-F two-part epoxy adhesive with 7 shear, and two connections outside of the loading points carried a
days curing time. The epoxy was specifically designed for structural combined bending and shear force. In all specimens tested, a nominal
usage in the bonding of steel TCC connectors to engineered wood with spacing of 217 mm was used between each connector. This was chosen
to simplify the design of the overall TCC system, both for ease of

4
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 2. (a) Installation of CTC screws, (b) screws and CFRP rebars after installation.

Fig. 3. (a) Details of perforated steel plates (dimensions are in mm), (b) bidirectional CFRP plate, and (c) installation of steel and CFRP plates.

construction and to ensure that there was directly comparable consis­ span and two loading points. In addition, two 50 mm LVDTs were placed
tency between all samples. With the intention of this study necessitating on either side of the TCC beam to monitor the relative slippage between
a distinct focus on the connections, this meant the only changing vari­ the centre of timber and concrete elements. These were noted as N1, N2
able was the connections themselves. The constant moment of section of and S1, S2 for the Northern and Southern ends of the specimen,
the four-point bending tests additionally provided consistent loading respectively.
conditions between samples. The loading was in accordance to ISO 6891-1983 [22]. This method
Each beam was placed on half rounds as supports, and had a four- allows the connections to be engaged by partial elastic loading and
point bending loading rig applying force as per the setup shown in unloading before being loaded up to ultimate failure. Loading was
Fig. 6. For the measurement of vertical displacement, three Linear applied to 40% of the estimated maximum load (Fest) and held for 30 s.
Variable Displacement Transducers (LVDTs) were placed below the mid- The load was then reduced to 0.1Fest and held for a further 30 s.

5
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 4. Approximation of a TCC system as a strain distribution with two neutral axes based on Gamma method in EN 1995: Eurocode 5 [21].

Fig. 5. Detail of TCC beams: (a) SS, (b) CR, (a) SP, and (d) CP specimens (dimensions are in mm).

Thereafter, the load was increased until the ultimate load or slip of 15 2.4. Digital image correlation (DIC)
mm was achieved. A typical load versus time graph adopting this
method is highlighted through Fig. 7a. This was done for each beam with Digital Image Correlation presents a non-destructive optical tracking
loading rates and holding points shown in Table 3. Likewise, this system method that allows for 2D measurement of relative changes in sequen­
has an idealised load-deformation curve to reflect the loading and tial images. The advantage of using DIC is threefold; relative cheapness
unloading of the specimen Fig. 7b. compared to use of strain gauges and LVDTs; high precision of mea­
During each test, the shape of the load-deformation curve was surement; and potential to obtain entire field measurement distribu­
compared to the idealised curve seen in Fig. 7 to check for any potential tions. Obtaining entire field measurements allows for a clear
errors in loading that would lead to erroneous results. Provided the understanding of stress localisations in testing, thereby providing
initial loading phase to Hold Point #1 had not already been completed, invaluable failure information. Consequently, any non-conformity in
the beam could be unloaded with reasonable expectation that there had material or loading is high visible in this form of analysis.
been no permanent damage, and the error in setup could be diagnosed For all specimens, DIC was utilised during the testing to measure the
and fixed. displacement, strain, interface slip, and separation of concrete and
timber and also identify failure mechanisms during the tests. In order to
perform accurate Digital Image Correlation, a regular pattern of dots

6
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 6. Test set-up for TCC beam testing in flexure (dimensions are in mm).

Fig. 7. (a) standard loading procedure, and (b) idealised load-deformation curve as per ISO 6891-1983 [22].

megapixels and a 35.9 × 24.0 mm2 full frame image sensor, with a 24
Table 3
mm to 200 mm wide angle lens fitted on a tripod was used for imaging.
Loading procedure adopted for the beams.
The quality of video obtained by this camera was 1080p, with a frame
Beam Holding points Holding point Loading rate rate of 30fps. The camera was backed by LED flood lighting to remove
#1 (kN) #2 (kN) (mm/min)
shadowing and create a better image quality for the analysis.
CLT 20 5 1.45 The DIC analysis required a series of photos to be taken from each
CTC steel screw cross 45 10 1.74 video of the beam being tested, with a frame pulled every 32 s. This low
connection TCC
CFRP reinforcement cross 30 8 1.16
frequency of photos was permissible due to the slow progression of the
connection TCC specimen tests, albeit particular care was taken so that pertinent aspects
Steel plate connection TCC 30 8 1.16 of each test, such as cracking and the failure mechanism, were reflected
CFRP plate connection TCC 30 8 1.16 in these photos. Each photo was then loaded as a “Deformation Image”
into the 2D image correlation program GOM Correlate [23], and the
relative movement of each painted point to every other one was
was spray painted onto each beam (Fig. 8). This was done using perfo­
calculated.
rated paper temporarily stuck to the front face of each beam with spray
By defining the target surface as the front face of the beam only,
adhesive and then applying a coat of paint by spray. A Canon EOS 6D
displacement and strain distributions in both the x and y axes were
Mark II digital camera, which has a maximum resolution of 26.2

7
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 8. (a) DIC set-up, (b) tracking points during loading, and (c) analysis surface and placement of deviation labels.

determined. By specific placement of “Deviation Labels” above and obtained were compared to values measured by linear displacement
below the interface (at left support, left loading point, beam midpoint, gauges, seen at three locations at the base of the beam in Fig. 8c. Direct
right loading point and right support), these values were tracked with comparison of these results is covered in detail in Section 3.4.
respect to time and exported for further analysis of interfacial slip.
Additionally, two more labels were placed at the extreme compressive 3. Results and discussion
and tensile fibres of the left loading point, midpoint and right loading
point of the TCC. Using this information, strain distributions along the This section presents the results from flexural beam tests.
depth of section could be determined at multiple locations throughout
the TCC. Fig. 8c summarises the beam surface and the placement of
labels. As a means of validation of DIC analysis in this study, the results

Table 4
Ultimate and failure loads and their corresponding mid-span deflections, and deflection limit loads.
Beam Ultimate Failure Increase in ΔM at 0.4Fult, ΔM at Fult, ΔM at Ff, Ductility, Effective bending Load at ΔM/ Load at ΔM/
load, Fult (kN) load, Ff capacity from Δ0.4Fult (mm) Δult (mm) Δf (mm) µ* stiffness, EIeff × 108 500 = 2.8 180 = 7.8
(kN) TB (%) (kN.mm2) mm (kN) mm (kN)

TB1 65.0 63.6 – 4.0 15.9 17.3 3.2 2.4 18.1 50.2
TB2 59.8 47.1 – 3.4 18.2 29.7 4.2 2.6 19.3 44.7
Ave. 62.4 55.3 – 3.7 17.1 23.5 3.7 2.5 18.7 47.5
SS1 71.8 71.7 15.0 4.0 21.6 21.9 4.3 4.8 29.3 54.5
SS2 74.5 60.7 19.4 3.2 16.9 25.0 4.2 5.7 24.2 65.4
Ave. 73.2 66.2 17.2 3.6 19.2 23.5 4.3 5.2 26.7 59.9
CR1 77.3 61.8 23.8 2.5 16.1 35.4 5.2 6.0 33.5 57.8
CR2 78.3 60.2 25.4 2.6 12.5 33.2 3.8 5.7 32.7 65.9
Ave. 77.8 61.0 24.6 2.6 14.3 34.3 4.5 5.9 33.1 61.8
SP1 70.6 63.1 13.2 2.9 15.1 19.9 4.2 6.8 34.9 57.1
SP2 71.8 56.4 15.0 2.2 15.1 31.7 5.4 6.3 36.4 65.7
Ave. 71.2 59.7 14.1 2.6 15.1 25.8 4.7 6.5 35.7 61.4
CP1 70.3 66.3 12.6 2.5 14.3 15.7 4.5 5.5 32.2 58.0
CP2 65.0 59.7 4.1 2.3 11.4 12.7 4.0 4.8 33.1 60.3
Ave. 67.6 63.0 8.3 2.4 12.8 14.2 4.2 5.2 32.7 59.2

*Ductility of the connections are determined based on definition is incorporated in EN 12,512 [24]. Yield slip, Δy, is determined as 1.25Δ0.4Fult based on proposed
methodology of CSIRO Australia (see [25] for more information).

8
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

3.1. Mechanical performance of TCC beams layers of timber between loading points and supports (Fig. 10c). In both
tests, tensile failure at the extreme tension fibres occurred below the
The results of four-point flexure testing for all TCC samples is shown right loading point. CR2 however, experienced delamination between
in Table 4. It presents the ultimate and failure loads, mid-span deflection timber interlayers below the left support, and propagation of tensile pull
at ultimate and failure loads, and deflection limit loads as an indication apart was beginning to occur between butted laminates in the extreme
of bending stiffness. Ultimate and failure loads were determined based tensile timber lamella. Moreover, the ultimate failure for both CR1 and
on the definitions presented in Fig. 9. With consistent geometric and CR2 was through shear failure of the compressive CFRP rebar located
material properties for timber and concrete components between spec­ between loading points and supports, as seen through Fig. 11c. CR
imens, the variation of response was primarily induced by the difference connections experienced shear failure at the interface of timber and
in stiffness, strength and ductile response of connections. Table 4 dem­ concrete. Such failure mechanism can be explained by the localisation of
onstrates that the CFRP rebar had the highest loading capacity from a forces at this location, as increased loading ultimately led to the disen­
purely strength perspective with a capacity increase of 25% compared to gagement of materials and a loss of composite action. At this point, the
timber samples. Performance of the CFRP rebar was closely followed by internal moments within timber and concrete imposed concentrated
the steel screws, with an average capacity increase of 17% compared to shear forces through the connections. The snapped end of the
timber. The plate connections performed comparably in terms of ca­ compression rebar remained embedded within the concrete. The epoxy
pacity, 13% exceeding the strength of the exclusively timber beam embedded CR connections appeared to have bonded well with the
samples. Overall, the cross connections demonstrated higher ultimate timber as no sign of deformation or debonding was present upon
load. observation of exposed connections. Here, the shear failure of connec­
To understand the mechanical performance of specimens, it is tions may be characterised as a brittle failure with no signs of plastic
important to investigate the failure mode of each connection. The visual deformations within the CR itself. In CFRP rebar specimens, no sign of
examination was conducted on the connections furthest from the the uplift of the concrete from timber was observed.
midpoint, reflective of the most extreme destructive case. As loading Plate connected TCCs did not undergo yielding, and ultimate failure
approached ultimate failure, TB1 experienced flexural shear cracking was characterised by failure in the extreme tensile fibres of timber and
between loading and supports where shear forces were maximum. Ul­ cover cracking above connections. Upon post testing analysis of the
timate failure occurred suddenly as cross layers delaminated at cross beams, no sign of debonding between plates and timber was observed.
panels within the bottom cross grain lamella, as seen in Fig. 10a. Concrete was present throughout the entire of the SP perforations,
Alternatively, TB2 experienced more pronounced flexural shear cracks however no aggregates had set within these perforations. This is because
between loading and supports before ultimate failure. Ultimately, this the 7 mm aggregates were slightly larger than the 5 mm ϕ perforations.
led to a more ductile response for TB2, as cracking of timber interlayers It may be suggested that bonding performance for SP connections could
reduced the global stiffness of CLT, and allowed for higher be improved more through modification of relative aggregate to perfo­
displacements. ration sizes. In both CP and SP beams, cracks propagated from the
As beams with steel screw (SS1 and 2) approached ultimate failure, bottom edge of connectors outwardly to front and back faces of the TCC
flexural shear cracks within the timber cross layers began to propagate (Fig. 11d and e) in the shear-flexure zone of the beams. This may be
from below the left loading point towards the left support Fig. 10b. Both explained by the maximum rotation experienced close to supports, as
SS beams reflected failure of concrete about the connections located the localisation of bearing stress from the connections into concrete was
within the critical shear region, between loading points and supports ultimately greater than the bearing capacity of concrete at this point. In
(Fig. 11a). Observations indicated that the compressive screw within the critical shear zone, the de-bonding of CFRP plates from the concrete
this region began to experience yielding as permanent deformations was observed and poor anchorage was evident by the smooth concrete
were evident upon careful removal of cracked concrete sections. This surface (Fig. 11f). Moreover, analysis of CP strain behaviour of the
means that screw connections were beginning to plasticise prior to beams highlighted that premature de-bonding of connections occurred
failure. Evidence of yielding can be identified through Fig. 11b, where within concrete which resulted in lower capacity in CP TCC beams. This
right compressive screw was bent towards the timber as a result of indicates the requirement of improving shear development of CP con­
combined bending and shear actions. Investigating videos after nections using sufficient anchoring system.
completion of tests revealed that the interface between concrete and A viable TCC connection must demonstrate both high stiffness and
timber from the load point toward the end of the beam was gradually ductility. Fig. 12 shows the load-deflection at midspan for each test
increasing during the last stages of the loading. specimen, with the test results for TB1 and TB2 included as a repre­
CR1 and 2 beams also experienced flexural shear cracking in cross sentation of no composite action. As a further representation of
displacement, the load-deflection graphs underneath the left and right
loading points for each TCC specimen are also presented in Fig. 12. As it
is expected, the absence of composite action in TB1 and TB2 beams led
to the lowest ultimate load and stiffness. By comparing corresponding
loads at the deflection limits, beams with CFRP cross connection and
steel plate (CR and SP) exhibited 30% higher stiffness compared with TB
specimens. Steel screw and CFRP plate also showed 25% higher stiffness
than TB beams. This is indicative of the increased beam stiffness of each
TCC system due to the presence of composite behaviour. The high
stiffness of CFRP/steel plate and screw can meet the requirement of
deflection limits in design of TCC beams. Considering the load and
deflection results, TCC beams with CFRP rebar connectors displayed
both highest strength and stiffness.
Timber and concrete are both inherently brittle in bending. There­
fore, it is important to design connectors in TCC beams in order to
exhibit high ductility and improve overall performance of the beam.
Design codes such as Eurocode 5 [21] allows for a redistribution of in­
Fig. 9. Definition of parameters for the estimation of load-displacement ternal forces if ductile connections are used [25]. Ductility ratio of TCC
response of TCC beams. beams was determined based on definition incorporated in EN 12512

9
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 10. Failure modes of: (a) TB1, (b) SS1, (c) CR2, (d) SP2, and (e) CP1 beams.

[24] which was defined by the ratio of the displacement at ultimate load,
Ke a ( 2 )
Δult, to the yield slip, Δy (Fig. 9) EI eff = 3L − 4a2 (4)
48
Δult
μ= (3) ΔF
Δy Ke = (5)
Δδmid− span
Yield slip, Δy, is estimated as 1.25Δ0.4Fult based on proposed meth­
odology of CSIRO Australia (see [25] for more information). By ana­ where a is the distance from loading point to support, ke is the ratio
lysing ductility of TCC beams using data in Table 4 and Eq. (3), LVB steel between the loading and the mid-span deflection increments, and L is
plate and CFRP rods demonstrated to have the highest ductility followed the clear span of the beam. The elastic bending stiffness was obtained for
by CTC screws and CFRP plate. The ductile behaviour of steel plate and 0.1Fult and 0.4Fmax and their corresponding mid-span deflections. A
CFRP rod connections could be due to the confining effect of concrete summary of EIeff for each TCC beam is presented in Table 4.
around perforated steel plate and the rope effect of laterally loaded rods, In service loading conditions, the steel plate and CFRP reinforcement
respectively. The latter engaged the axial capacity of the rod, generating connections exhibited the highest effective bending stiffness due to the
a normal force that pulled the concrete and timber sections together and samples’ resistance to deflection. Indicative of the connections’ perfor­
produced friction in the shear plane around the ridges of the CFRP rod. mance, both the steel plate and CFRP reinforcement showed efficient
The lower ductility of CP beams may be the result of relatively low transfer of load and high composite action. Alternatively, the steel screw
plasticisation of CFRP plates prior to ultimate TCC failure and insuffi­ and CFRP plate connections performed less effectively, due to low
cient bond between plate and concrete. This behaviour can be modified stiffness and lack of adhesion to concrete respectively. In addition,
using appropriate anchorage systems at the end of CFRP plates within further review of the testing videos revealed higher interfacial slip in
the concrete. these beams, that was not seen in the steel plate or CFRP reinforcement
specimens, indicative of lower force transfer.
3.2. Bending stiffness of beam samples
3.3. Interfacial slip and strain analyses of composite action
In order to investigate the effectiveness of connections in distributing
load between the concrete and timber layers, the overall beam’s bending High composite action occurs when there is minimal interfacial slip
stiffness can be evaluated. This method of analysis takes into consider­ between the timber and concrete and when the overall beam neutral axis
ation the impact of multiple connections, which is a more realistic coincides with this interface. This can be achieved by using a connection
interpretation of connection’s performance. The effective elastic that exhibits a high stiffness and shear capacity. Connection stiffness can
bending stiffness of the beams, EIeff, was calculated by Eq. (4) based on be represented by its ability to resist interfacial slip between materials.
the experimental load-deflection curves: Similarly, the relative strain in timber and concrete is indicative of

10
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 11. (a and b) failure of CTC screws, (c) shear failure of CFRP rebar in CR beams, (d) connection bearing cracks in SP1 and 2, (e) cover cracking above connection
in CP2 beam, and (f) debonding of CFRP plate from the concrete in CP2 beam.

connection stiffness. In the context of TCCs under flexure, composite transducer equivalent. The maximum separation of curves is less than
action can be extrapolated through analysis of the beams’ strain profile 2.25 mm, 1.97 mm and 1.11 mm for points at right load, midspan and
at specific locations. Therefore, interfacial slip and strain behaviour is left load, respectively. However, the overall shape of each graph is
considered in this section. In order to compare the effectiveness of comparable while there is relatively less deviation in measurement at
connections based on composite action, the stiffness of connections must the ultimate load of each specimen. This small difference can be due to
be comparable. Theoretically, connections with comparable axial (EA) the fact that LVDT and DIC measurements are recorded from middle of
and bending (EI) stiffness should experience the same strain under the the section and the side of the beam, respectively (Fig. 14). In addition,
equivalent loading. Through this design philosophy, it was idealised that the camera was fixed while beam was deflecting, resulting in an angle
connection stiffness was directly comparable for inclined connections, between the measurement point and the position of the camera creating
and plate connections. pseudo-deflection [26]. The point where each DIC measurement began
In the previous sections, slip after ultimate load has been primarily to record more displacement than the transducer, from around 10 mm
discussed due to its effect on the mechanism of TCC failure. When deflection, was a reflection of the increasing impact of this pseudo-
considering serviceability loading however, it is more important to deflection. Consequently, it seems that there can be a constant offset
consider slip leading up to the ultimate load. It has been shown that TCC allowance added to the DIC measurements to account for scaling to
beams with low interfacial slip at initial loading stages experience minimise pseudo-deflection, in order to correlate the two sets of results.
higher composite action [8]. Therefore, DIC analysis was performed for Despite of this negligible difference, DIC results were strictly used in the
each specimen to measure the relative slip at both supports, loading context of relative measurements, specifically relative slip and strain
points and the beam midspan for serviceability loads. Fig. 13 illustrates profiles, thereby negating any offset in result.
strain map for steel plate connections. Relative slip was considered as relative displacement between the
Throughout analysis conducted in this study there was a presump­ concrete and timber. For this analysis, only loading from the start for the
tion that DIC analysis provided accurate results. To verify accuracy of reloading phase up to the ultimate load was considered. Positive relative
DIC, analysis of CR2′ s deflection was conducted at the extreme bottom slip means that the concrete has moved further to the right of the mid­
fibre, underneath the two loading points and the midspan. In each case, span than timber, and negative likewise to the left. CR and SP specimens
the deflection points were tracked directly above each of the three showed minimal slip until 40% of ultimate load, where the composite
displacement transducers as a means to minimise disparity due to beam action started to reduce and the TCC beam experienced slipping at the
deformation. This is compared directly to the equivalent deflection for interface (Fig. 15). For loads approaching the ultimate load, similarly
each transducer in Fig. 14. steel plate and CFRP reinforcement cross connections experienced low
Based off observation of Fig. 14, during the loading phase all DIC interfacial slip at midspan and all supports and loading points. CP and SS
measurements recorded slightly less vertical displacement than their beams experienced greater slippage during both initial loading (up to

11
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 12. Load versus deflection at (a) midspan, (b) right side LVDT, and (c) left side LVDT of TCC beams.

Fig. 13. DIC analysis of the beams (a) SP1 at 40% Fult, and (b) SP2.

0.4Fult) and ultimate load between timber and concrete. The high left concrete and CFRP plate was not sufficient.
support slip and subsequently a shift in degree of composite action It is important to note that the serviceability performance (from 10%
occurred at 40 kN for CR1 beam. Upon further review of CR1 testing to 40%) of each beam in Fig. 15 closely mirrors the performance seen in
video, the sudden increase in relative slip at both ends coincided with the effective bending stiffness calculations in Section 3.3. Both steel
the first flexural shear cracks in timber. Flexural shear cracks in concrete plate and CFRP reinforcement show considerable resistance to slip,
occurred in SS1 below loading points prior to global failure of the TCC at corresponding to their high effective bending stiffness. Similarly, the
ultimate load. This was reflected by the scattering of measurement lower effective bending stiffness of steel screws and CFRP plate allows
experienced at this point. CP beams experienced consistently increasing for greater interfacial slip in these loading conditions. With this in mind,
slip during loading. Such performance may be attributed to the lack of DIC represents a valid method of analysis in the context of TCC systems.
adhesion between CFRP plate and concrete previously discussed and An examination of strain distributions can reflect the degree of
identified through Fig. 11f. Sliding of concrete reflects poor composite composite action provided by a certain connection. DIC analysis was
behaviour, and is suggestive that internal force transfer between conducted for each specimen to obtain strain distributions at 10%, 40%

12
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 14. Comparison of DIC and transducer results for CR2 deflection.

and 100% of ultimate load. In each loading case, strain distributions 3.4. Cost analysis
were obtained at both loading points and at the midspan of the TCC
specimen. These load cases were chosen to correspond with the timber In order to compare the efficiency of connections, it is also important
fastener testing procedure in ISO 6891-1983 [22]. It should be noted to consider the cost-impact. In the initial theorisation of the connections,
that for each specimen strain values were calculated relative to the input performance was the main objective to ensure that the connections
values used to construct the DIC mesh. These input values were would be structurally viable, and then through some minor changes the
dependent on camera angle, closeness of camera, lighting, colour and cost of the connections was thought through. Since the geometry of all
arrangement of painted dots which could not be held completely con­ the beams with respect to concrete, timber, spacing of connections and
stant between all test videos. Consequently, the numerical strain values number of connections were the same only the cost of connection was
calculated by the program were relatively arbitrary when compared considered in the cost analysis. It is worthwhile to mention that life cycle
between videos. For consistency of analysis, the numerical strains analysis was not undertaken in this research.
measured have only been used to establish overall shape of the strain For calculation of costs, a flat rate of A$45 per hour for labour was
distributions to allow comparison of samples. From this, a qualitative used, with the relative production rate of each connection recording
interpretation of composite action in relation to each sample type is during construction. The production rate of each connection accounted
possible as in Fig. 16. for drilling of holes, cutting of slits, cutting to size and installation of
At 10% and 40% of ultimate load, strain distributions along the cross epoxy where relevant, based on current base labour hire costs for typical
section of steel screw and CFRP rebar TCC specimens illustrated an construction in Australia. The time allowance for curing of epoxy
effective composite action between concrete and timber (Fig. 16). This, however was not considered in costing, due to inability to quantify. This
as expected, led to a higher overall ultimate loading capacity for the SS however can be planned for in construction practice accordingly as to
and CR systems when compared to their plate connection counterparts. effectively be reduced to zero with adequate scheduling. For screws and
At both 10% and 40% of ultimate load, strain distributions indicated a CFRP rebars, the rate consisted of the time taken to install two individual
neutral axis positioned at the material interface. The advantage of the pieces to create the crossed orientation.
shared neutral axis in the high composite action is that the concrete The steel screw required no application of epoxy, resulting in the
beam takes almost the complete compressive load and the timber beam fastest installation time. Both plate connections must be cut to size, the
takes mostly tensile load in the beam. Ultimately this is highly desirable steel plate only required cutting 80 mm pieces from a 800 mm × 80 mm
for the connection as this maximises the material properties of the TCC larger plate, while the CFRP plate required to be cut diagonally from a
components. 500 mm × 500 mm larger plate to achieve biaxial fibre orientation.
The steel plate and CFRP plate connectors transferred strain highly Moreover, a further 30 mm thick epoxy was assumed to be placed in
effectively across the neutral axis at the right loading point. The each timber slit for both plate connections, in order to provide adequate
culmination of this effective load transfer is minimised interfacial slip. stress development. With these rates used to calculate labour costs and
Despite this high performance at a low load, both SP1 and CP2 ulti­ material cost available from each supplier, the total breakdown of
mately failed underneath the right loading point, as connections were costing was calculated as presented in Table 5.
increasingly detached from the concrete and composite action reduced. For both the plate and the cross connections, the CFRP connectors
This is reflected in the independent action of the timber and concrete at cost more than the steel, while the cross connections have lower cost to
ultimate load shown in Fig. 16. In case of CP2 beam, the concrete strain their plate counterparts. In the case of plate connections, the epoxy
distribution showed almost no composite action, with the neutral axis at provides considerable expense. Alternatively, both the screws and CFRP
the concrete geometric centroid. The timber component had a distinct rebar have the majority of their expense coming from material cost, with
neutral axis mostly closer to the material interface, shifted upwards due low epoxy and labour costs. The steel screw connection is approximately
to the result of interface friction providing some beam restraint. The lack 15.5% cheaper than the CFRP rebar, albeit with a 7.5% lower ultimate
of composite action at only 40% of ultimate load however showed the loading capacity for the specimens tested in this research.
impact of connection slipping completely away from the concrete. In this cost analysis, it was assumed that each TCC sample will un­
dergo the same deterioration over the time. Considering that CFRP
composites generally typically show higher resistance to moisture and
aggressive environments and in this research, showed higher stiffness

13
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 15. Relative slip from midspan for TCC beams.

and strength in the case of CFRP rebars, the higher initial cost for this 4. Conclusions and remarks
type of connectors can be moderated over the operating life of the TCC
beams. In addition, by minimising use of epoxy adhesive in the This research proposed new composite shear connectors for the use
connection or application of mechanical anchorage for installation of in TCC beams using bidirectional CFRP plates and crossed CFRP rein­
CFRP composites, the cost can further be reduced. forcing bars. Performance of new CFRP connectors in CLT-concrete
composite beams under four-point flexural tests was evaluated and
compared with perforated steel plates and crossed steel screws shear
connectors. Connections were evaluated based on structural

14
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Fig. 16. Normalised strain distribution of TCC beams.

performance criteria, including ultimate strength, bending stiffness and of the composite action for each beam was evaluated.
ductility. The paper also presented an application of the Digital Image The beams with CFRP rebar systems showed the highest ultimate
Correlation (DIC) based technique to determine local slip, interfacial load, followed by steel screw and steel plate connectors. Debonding of
slip, and strain at each point along TCC beams through which the degree CFRP plate from the concrete led to high interfacial slip and low

15
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

Table 5 the Australian Research Council’s Discovery Early Career Researcher

Cost analysis for each connection type (costs are in the Australian dollar). Award (DECRA) fellowship scheme (project DE200100406) which was
Connection Rate Connection Epoxy Labour Total received by the first author. Authors would like to thank the School of
type (connections/ cost ($) cost ($) cost ($) cost Civil Engineering at the University of Sydney for supporting this project
hour) ($) and Stora Enso Australia PTY Ltd. for providing CLT beams. The help of
Steel screw 120 6.20 0.00 0.38 6.58 PhD student Mr Ziyuan Wang for DIC analysis of the beams is
CFRP rebar 100 6.42 0.92 0.45 7.79 appreciated.
Steel plate 90 1.68 5.43 0.50 7.61
CFRP plate 70 3.60 5.43 0.64 9.67
Data availability

composite action. As a result, beams with CFRP plate connectors showed The raw/processed data required to reproduce these findings cannot
lower load carrying capacity. By investigating effective bending stiffness be shared at this time as the data also forms part of an ongoing study.
of the beams in service loading condition, it was revealed that the steel
plate and CFRP reinforcement systems exhibit the highest stiffness References
compared with steel screw and CFRP plate beams. This led to observa­
[1] Mascia NT, Soriano J. Benefits of timber-concrete composite action in rural
tion of minimal slippage at serviceability and ultimate load at CFRP rod
bridges. Mat Struct 2004;37(2):122–8.
and steel plate specimens. [2] Hadigheh SA, Dias-da-Costa D. Shear strength of cross laminated timber-concrete
In terms of ductility of connections, steel plate and CFRP rod systems connections reinforced with carbon fibre polymer composites. In: Wang CM,
Ho JCM, Kitipornchai S, editors. Lecture Notes in Civil Engineering. Springer
demonstrated the highest ductility. The lower ductility of CP connectors
Nature; 2020. p. 179–85.
may be due insufficient cohesion between CFRP plates and surrounding [3] Suárez-Riestra F, Estévez-Cimadevila J, Martín-Gutierrez E, Otero-Chans D.
concrete which led to earlier separation between these two elements. Perforated shear + reinforcement bar connectors in a timber-concrete composite
However, the overall beam’s ductility performance can be improved solution. Analytical and numerical approach. Compos Pt B-Eng. 2019;156:138–47.
[4] Steinberg E, Selle R, Faust T. Connectors for Timber– Lightweight
using sufficient anchorage systems at the end of CFRP plate within the Concrete Composite Structures. J Struct Eng 2003; 129(11): 1538–45.
concrete. Considering the strain distribution along the beam’s cross [5] Marchi L, Scotta R, Pozza L. Experimental and theoretical evaluation of TCC
section, CFRP rebar and steel screw TCC beams illustrated an effective connections with inclined self-tapping screws. Mat Struct 2017;50(3):180.
[6] Shan B, Xiao Y, Zhang WL, Liu B. Mechanical behavior of connections for glubam-
composite action between concrete and timber at 10% and 40% of ul­ concrete composite beams. Constr Build Mater 2017;143:158–68.
timate load. [7] Zhang C, Gauvreau P. Timber-concrete composite systems with ductile
Cost analysis of the connections indicated that steel screw system connections. J Struct Eng 2015;141(7):04014179.
[8] Lukaszewska E, Johnsson H, Fragiacomo M. Performance of connections for
was the most affordable. With effective ductility and load carrying ca­ prefabricated timber–concrete composite floors. Mat Struct 2008;41(9):1533–50.
pacity and relatively lower construction effort, steel screws present as a [9] Khorsandnia N, Valipour HR, Crews K. Experimental and analytical investigation of
cost-effective option for ultimate loading. However, the enhanced short-term behaviour of LVL–concrete composite connections and beams. Constr
Build Mater 2012;37:229–38.
stiffness and load capacity at both serviceability and ultimate loading [10] Deam BL, Fragiacomo M, Buchanan AH. Connections for composite concrete slab
stages demonstrate that CFRP rebar is a desirable connector to provide a and LVL flooring systems. Mat Struct 2008;41(3):495–507.
high capacity and safe alternative, at only a slightly higher price point. [11] Yeoh D, Fragiacomo M, Carradine D. Fatigue behaviour of timber–concrete
composite connections and floor beams. Eng Struct 2013;56:2240–8.
Moving forward there needs to be additional research into the
[12] Dias AMPG, Martins ARD, Simões LMC, Providência PM, Andrade AAM. Statistical
functionality of CFRP in TCC systems prior to full-scale implementation analysis of timber–concrete connections – Mechanical properties. Comput Struct
in construction. Despite promising indication that CFRP rebar presents a 2015;155:67–84.
stiffer connection than the commonly used steel screws or plates, these [13] Brunner M, Romer M, Schnüriger M. Timber-concrete-composite with an adhesive
connector (wet on wet process). Mat Struct 2007;40(1):119–26.
new types of connections should also be evaluated for serviceability [14] Clouston P, Bathon LA, Schreyer A. Shear and bending performance of a novel
design aspects. With that in mind, there is great potential for widened wood-concrete composite system. J Struct Eng 2005;131(9):1404–12.
investigation into varying CFRP connection types to greater utilise this [15] Otero-Chans D, Estévez-Cimadevila J, Suárez-Riestra F, Martín-Gutiérrez E.
Experimental analysis of glued-in steel plates used as shear connectors in Timber-
material. Concrete-Composites. Eng Struct 2018;170:1–10.
[16] Ghazijahani TG, Jiao H, Holloway D. Composite timber beams strengthened by
CRediT authorship contribution statement steel and CFRP. J Compos Constr. 2017;21(1):1–11.
[17] Sarker M, Hadigheh SA, Dias-da-Costa DA. Performance-based characterisation of
CFRP composite deterioration using active infrared thermography. Compos Struct
S.A. Hadigheh: Conceptualization, Data curation, Methodology, 2020; 241(2020): 112134.
Resources, Software, Investigation, Formal analysis, validation, Visual­ [18] Gravina RJ, Hadigheh SA, Setunge S. Interfacial bond strength of resin
impregnated fibre reinforced polymer laminates bonded to concrete using vacuum
ization, Writing - original draft, Supervision, Project administration, and heat: Experimental study. Aust J Struct Eng (AJSE) 2014;15(2):189–201.
Funding acquisition. R. McDougall: Methodology, Software, Valida­ [19] Hadigheh SA, Gravina RJ, Setunge S. Influence of the processing techniques on the
tion, Formal analysis, Investigation, Visualization, Data curation, bond characteristics in externally bonded joints- Experimental and analytical
investigations. Compos Constr 2016;20(3):1–13.
Writing - review & editing. C. Wiseman: Methodology, Software, [20] Hadigheh SA, Kashi S. Effectiveness of vacuum consolidation in bonding FRP to
Validation, Formal analysis, Investigation, Visualization, Data curation, concrete surfaces. Constr Build Mater 2018;2018(187):854–64.
Writing - review & editing. L. Reid: Methodology, Software, Validation, [21] EN 1995: Eurocode 5. Design of timber structures. European Committee for
Standardisation; 1995.
Formal analysis, Investigation, Visualization, Data curation, Writing - [22] ISO 6891-1983. Timber structures — Joints made with mechanical fasteners -
review & editing. General principles for the determination of strength and deformation
characteristics; 1983.
[23] GOM correlation. Technical Documentation as of V8 SR1, Technical
Declaration of Competing Interest Documentation as of V8 SR1, 28122 Braunschweig. Braunschweig, Germany;
2016.
[24] EN 12512—ISO/DIS 16670. Timber structures. Test methods. Cyclic testing of
The authors declare that they have no known competing financial joints made with mechanical fasteners. Brussels: Comité Européen de
interests or personal relationships that could have appeared to influence Normalisation, CEN; 2001.
the work reported in this paper. [25] Jorissen A, Fragiacomo M. General notes on ductility in timber structures. Eng
Struct 2011;2011(33):2987–97.

Acknowledgements

This research was supported by the Australian Government through

16
S.A. Hadigheh et al. Engineering Structures 232 (2021) 111791

[26] Hohmann BP, Bruck P, Esselman TC, Schmidt T. Digital Image Correlation (DIC): [28] Hadigheh SA, Wei Y, Kashi S. Optimisation of CFRP composite recycling process
An advanced nondestructive testing method for life extension of nuclear power based on energy consumption, kinetic behaviour and thermal degradation
plants. Int Atomic Energy Agency (IAEA) 2012. mechanism of recycled carbon fibre. J Clean Prod 2021:JCLP_125994. In press.
[27] Hadigheh SA, Ke F, Kashi S. 3D acid diffusion model for FRP-strengthened
reinforced concrete structures: Long-term durability prediction. Constr Build Mater
2020;261:1–13. https://doi.org/10.1016/j.conbuildmat.2020.120548.

17